Loading...
The URL can be used to link to this page
Your browser does not support the video tag.
Home
My WebLink
About
Study Session – Finance Board Presentation on Prevailing Wag
CITY OF HUNTINGTON BEACH INTERDEPARTMENTAL COMMUNICATION FINANCE OFFICE TO: JOAN FLYNN, CITY CLERK FROM: DAN T.VILLELLA, CPA, FINANCE OFFICER za SUBJECT: FINANCE BOARD PRESENTATION ON PREVAILING WAGE (STUDY SESSION—OCTOBER 17, 2005) DATE: OCTOBER 3, 2005 Attached is a hard copy of the Finance Board's PowerPoint presentation scheduled for the October 17t' Study Session on Prevailing Wage. cn c c § Attachment / y� i N nV) O - ■ cn 4-J o 0- u, a� > , .J .o o-6 o6 u :3 C: V) 4-J Lf) — c (1) U) (U Cn ._ V o — o m o — c u_ N 7 � o L .— -W — LL. r\jU fu O —1 V N � E- 0 m O s- Q u 1-- 4 :3 0 L. V 0) 4- C o � r 0 .ems �..�._ 0Lr) M 0 W C 0- 0 _ > �- o . - N ._ 0 U o V cu o0) 4-J U. E 4- C _ ( „ -J o = 0 — � °' o 0) = a� .. 0 o �.., m o v -�E Q a-' -- V V •— uO m w to U- a--) a) fu E > o :3 4-j U) C: L U O O U. LL CL ® i ( ) U [6 O 0. O O �- 4 to — M 4� L. U o� 4-J U U U (� (� (D c O U 4) 4- J L. Cu U U a C: 4 s .a U (Q .- Lf) • E -- E Q) o L- V 3: V) � p 4) 4-' p (n O L 41 U4) (1) O . U C O > (,a n (n Ste. (D q} E 0- ca_. -0 0 [ � � � 43 .- V (� f V U) L -0 Q) Cal Q .ra > 0 S, �-, •- 1 t' a L - ra J lD n .{-J O 10 V / 0 Q E 0 C: V -W . ._. _0 U (1) 4 U- _v 0 _ Q. V) U �. v U � . . O -0 ,_ O E O w ca L U -O s-- 0- Zj N a) c L 3 a' M to -' 4-1 O .N N V V � O ro s- D � y-- -0 O 4- O O 70 -0 v v 0 0 C: i L .. 0 N -- a a a U U • L • • O h r cn -�o c) rn o 0 0 N O cM 00 O) �_ 1— O) 0) t4 - N J J J N Q L rtrt LL L a 3: z L C: Co _ ) W c c 4-1 0 c�a 0 Z U 0 0 o U EV o � O 0 O c > O co � �- o C 0 .0 (D o > Q o z o 0 0 L _I i vJ Q p_ O O Q O d �. L � 00 00 n Y � X U � U <� 0 U U 0 O O U O C) lL ca OL J O ro U) (iT to 0 -a_ (6 0cc) O 0 N N O W o 0 O =o 5 a) c o c D D 4- U O 0 4- 0 Q Q Q O i 0 C 07 N 3 U cu o a) o a) ca -a) o 0 � u) a) ■ ®■ U) Q c .ems j a_ Q- V c o }' o U o C 000 M N !f) ~ 00 0) M C cm N N O M M O m o -c U C) 00 4-)0 O t . _. a 0 y-- _... � 4-) (u [a � u 4� L. 0) .4.� 1 m � •0 O O — a} .0 L .,�.. L 0) C: -r—, > y- 0 cn C ._. p L o Q > ca V > � E a � � _ tQ ' c6 Ln ( N f T3 — a) L3 -� U) u O -0 : •� a) 70 O O O ,V) V C v J E a) •— E � to E O to .� ua .c N -�-0 S_ v � p -(1-1-)j t„ 4) O 0 >. cu cn L 4- N 4. .. � c� a� o -0 0) 0 c: zj- > 0 c E � O o cn ca a. u 0 (1) 0 (1) V E--•t 4— H . H ■ ■ 0 1- 0 0 0 0 0 -0-0 0 0 0 0 0 -.-0 0 0 0 0 O O 0 0 0 0 LO O [l- oI N N N N (V) N N 00 N N O O O O 0 0 0 O O O O O O O O O O O O O O O O O 0 0 0 0U') 0 0 0 0 0 0 LO Lo O O 00 O O L 00 O LO O �- j CO CIO to LO �- N N to N M L O U) E6 _ Q) 0 *�■ V) u a. Cl 0 0 0 0 0 O o 0 0 o O o o Cl o 0 Co Cl o O o 0 0 O o O o O O o 0 0 0 0 0 a xi66C L66L6 L666 0 '�.., � � N N M tf) 'ti M tf 3 OLO � 00 3 N m C'M Lci to RON3 O C: 0 Ef3 Efl 0 CL 0 0 a.-• L � � � ® 0 c ,.� U c 0 C: ■® V Co a) a) m 0) (U C;) U U m s C: C � to m N a) _� E (0��+ C � ED (Q V (a -i � � � U i>- �: a. D 'C 'C 7�6� U N m U3 N - U) LJL2i � -.-. C C i U) i LL _0 N N C N N (o U JEl. t— U? U) 0 r-i L V) IZ- 4-J ' U) 4-J L > a a3 ca u 3: u 4-J D 0 a) r^ 3: c L 0 . .�.. V) V) 3 4-1 L 07 ru > D U C: (1) C ru > .— > I' ._ D 0 0 07 -- c N s- n 0 Ul 00) E c: E > 04- � ro 4- 0 p E p V a-' 0 u 4- 4- 0 C4 0 O 0 .� O O � � O — 0 L L > -0 � (3 ) 0c C ` cL (1) D � > .- 0 0 0 0 0 a� 4—J 4J �> O O .-j 4-J -0 O 0 fu1-� T 0-6 -o a') �I E 0 o — O M� tr) ru _0 �0 � �_ O 4} O U 4 E O > q- . -J .— u.- U Q U, Ul •— �-' .s O to _ 4-j I (1) H V 0 • 0 � 0- 0 OX Z m ■ (,� i J N V, fa 4. 3: 0 fu (n O p . C- N U 0. ro '0 0) n 0 0 C: u fu C� Cps �03 cu � o - fu E Q. 0) 4-J > fu EZ a) r- N u E -W r) " E ._ Q �caC: 0 > cn > mo x ru V a) o � ca � � L O L. � O LL DC ■ ■ i RerpEort on the Prevailing By Peter Phliip's; Ph.D Professor of Econami Wage Law of Nevada Univers'sty of Utah ri l . .Nil IY,r its j Ico `:ti}an IT e c.±S, r t L ' 1>nr ;Y 1y i��ct7i1tfi1 ' 1 al':IIII5'G1 41 ►..�.:i,dl� +Acatr � .. - - � ♦\.mow. $. rt?°� _ m ding 0-0 tee. to INUNiii tas smog } - Executive Nevada's prevailing wage law was passed in 1937. But support for prevailing wage regulations goes all the way back to the beginning of the state. Nevada's fiat two U.S.Senators,Republicans William Stewart and.lames Nye both supported the first prevailing wage law embedded in the National Eight-Hour Day Act of 1868. Stewart argued that when hours were cut by regulation from around 10 to 12 fours per day to 8, the daily wage should remain that which prevailed in the area. Critics responded that such a provision would result in higher hourly wages for workers on public projects. Stewart contended that workers would be more productive If given reasonable rest from their labors and that higher productivity would justify the higher wage. But that in any case, Stewart argued, the law would result in safer work, longer and better lives and an increase in the " aggregate of human happiness". Critics complained that the law interfered with the operations of a free market and that anyone worth their sett should be ready and wilting to work the longer hours. Strange how—the more things change, the more they remain the same. Today politicianss and others still argue about prevailing wage lawsr--one of the most venerable of labor market regulations. But this argument is embedded in a wider argument over whether the reforms that began after the Civil War, took up speed during the Progressive Bra prior to World War I, and were codified into national policy during the Great Depression were good ideas or not. These reforms included • Free public schooling and compulsory school attendance . • Factory safety inspection • Child labor laws • Social Security • Minimum wages • Overtime protection • Workers compensation • And prevailing wage laws Today some critics want to replace public schooling with school vouchers that may, or may not, cover the cost of private schools. Other critics want OSHA regulations and enforcement to become voluntary. Worldwide there is a debate over the 1 permissibility of child labor. Some want to privatize social security. Others resist raising the minimum wage with Inflation. Some want to change the 40-h0ur workweek to longer hours over a longer period. Some see workers compensation awards a$ overly generous and fraught with fraud. And many of these critics also Wish to eliminate prevailing wage laws. While there ere specific complaints against each of these regulations, there is also a general patternyy-regulations stand in the way of flexibility, they Interfere with the operations of the market; they raise costs. use criticisms sometimes create strange betllellows. School boards-4n the hopes of saving on their construction casts—sometimes ally themselves with critics of prevailing wage regulations---the same critics that in other venues are cdtiical of public schooling. Critics of prevailing wage laws WIN sometimes say they are unneeded now that we have minimum wage laws--yet in other venues, these same critics will resist raising the minimum wage. in INS report,we VA% do six thir►gs. Fitst, we raM examine the history aryl det me svrrvur►ding pr itirtg wage laws,in ger*rai, and Nevada%PTOYOMM wage lava, in particular. Second, we will axamine the trey contention of critics ctf prevailing wa laws--- nan tit pwv g w rag WS ral"e poetic awmi action costs in his s%tiion,we will p4y particular attention to the invisstlgations and contentigns of the Las Ve9as Review Journal. We will find that critics tend to assert that public agemle-$can save liom 15%to 5tr'A on cm,-Wuction costs by eliq*nstir9 prevailing wage laws. We will find that this contention rests on an unlikely assumption and a back-of-tree-envelop calculation. The unlikely assumption is that when 1? prevailing wage ragolabom are eliminated and 2)wages are consequently out by rotighly half, worker produclAd1l y wilt remain completely unaffected. Workers will bring to the job the same skills, the same experience, and the same work ethic as before. Contractors will Trust these workers with the same valuable tools and equipment and materials, And, in short, Productivity will remain entirely unchanged despite a major drop in wages. This is an odd assumption from a camp that also contends that markets work through prices. Mere is a price that fatly by roughly half, yet nothing else in the market happens. in any case, from this assumption, critics make a ample calculation. It labor costs in construction are roughly half of total costs, and laiaor costs are out in half, total costs will fall by 25%. Q.E.D. The problem with this back of the envelope calculation is that labor costs In construction are not half of total costs. They are more likely to run 20% to 250A of total costs—nat includes land purchases or architect and engineering fees. So even assuming that when wages are cut by cif% productivity Is unchanged, for tdkel costs to be cut by WJ6, everyone wit simply have to work for free. 2 But we need not settle for assumptions and simple calculations. In Chapter 2:"The Costs of Public Road and Schaal Construction"we use actual construction costs in states with and without prevailing wage laws. Comparing legal regimes,we find that there is no statistically measurable difference in the cost of public construction in states that apply prevailing wage regulations compared to states that do not have these laws. We do find, however, that there are other ways of saving on public construction. Looking specifically at school construction costs, we find that school districts that build into the teeth of a construction boom are going to pay a premium price to get instance, the unemployment rate in an area is cut in half one. If for si their work d , because of good economic conditions,school boards can expect the cost of building a school of any given size and type to go up around 20%. This is true in states with prevailing wage laws as well, as states without prevailing wage laws. And it is not p g primarily duels higher market wages during tight construction markets. Contractors are pulling wider profit margins, material suppliers are raising their prices, and the coordination protAems multiply when things are tight. We also find that school hoards can create their own tight construction conditions— tiny local over-demand pressures that stretch the capacity of local construction contractors. Double the number of new school construction starts in a local area, and school boards can expect their costs to rise 3.60/a per square foot. So for school 9 trying boards considering tr `n to save costs on local school construction, the best place to look is at their local planning. Counter-cyclical construction saves money. Even stretching out work over a longer time span saves money. But the same data that find these results cannot find a statistically measurable or significant effect associated with eliminating prevailing wage regulations. Now why is that? Chapter 3, "Training and Productivity—the High-Wage, High-Skill, Low-Wage, Low- Skills Growth Path In U.S. Construction"--explains why.This chapter shows that there are two divergent paths of development in U.S.construction. Along one path, few workers are trained, little capital is invested, many workers move out of the industry rather than making careers that generate lifetime experience, and consequently, productivity is low. This sector is primarily associated with open shop contractors. In the U.S.,open shop contractors account for only about 25a/a of all the apprentices graduating to journeyworket status. Furthermore, to the extent apprenticeship training takes place at all in the open shop, it is focused on electricians and plumbers. A comparison of 2000-2001 apprenticeship graduations in Nevada shows that 14 open-shop electrician apprentices turned out as journeyworkers. This 3 compares to 83 graduating apprentices sponwxed by ft Neva& Waumal Electrical Contractors Association jointly with the International Brotherh<W of Electrical WWAsr$. Three open shop plumber Vprenfioes graduated cornparedto 38 Jointly sponsored by the Nevada Mechank*0 Coors Associat" arxi to United Association of Plumbers and Pi{aefitters. In the Case at both crafts, the training process was basically 2W16 torgec, 6 Was compared to 4,in the case of the Joint tabor-management progams. So the tore and more demanding training c xxxses turned out 9%ftianfiaDy mare frs ined Warlmrs. A national survey of"ownW—tbe Wy"oonstrucCm sawices—lound that owners bWW9 time Wowing: • ,01M labor-nw"smerd apprentOes are more ftly to team a%the skfs of their trade • Joint apprentices ere More likely to be properly trained • Joint apprentices are rxm%e MWy to be Ufted to the fullest extent On-The- job Furthermore, owners found that signatory contreolors•--those working wvft Coitedive bafgaining—are more Okety to • Manage the job Pr dy • Eop their workers wish the most current tools of the t • Do the lob r4 t the fast time V.S.govoi r►t data from ft Censts of Construction and the Current P+o ulat ,qvrvpyare consiateM wit"these opolons of owners. 'i*se government data show Mt workers under collective bargaining in construction are: One-ftmird mare likely to have s me college education • Be ewtPod V th to 111%i%wer capital read vvy Per worker 4%firm,those workers • Process 11%to 13% more materials per worker in a given day • Tura out 14%to 18% auxe output per Worker • Generate 17%to 20% more value added per worker in contrast, again according to these government data, open shop workers are • Twice as likely to be high school drop outs • Twice as likely to be young and inexperienced owner-surveys indicate that open shop contractors working with a younger, less- trained and less experienced work force are likely not to do the job right the first time one-third of the time. These are the facts that put the he to the assumption critics of prevailing wage regulations rely on. It is not true that you can cut wages by 50% and expect worker productivity not to fall. Indeed, it does fall and may fall quite dramatically This accounts for the fact that hypothetical cost savings from prevailing wage law repeals are not forthcoming. But the dangers of using a low-skill, low-wage construction industry to build the infrastructure of a state are not limited to unmet promises. One very real danger Is simply the increase in injuries associated with using a less trained and less experienced labor force. Again referring to owner surveys,owners experienced with both open shop and signatory contractors find that • Union workers are more likely to be committed to safety. • Union workers are more likely to receive on-going safety training. • Signatory contractors are more likely to be committed to safety. • Workplace accidents are more liketly to be minimized, and ► Safety investments are more likely to translate into enhanced project value. Using Bureau of Labor statistics data on workplace injuries over the period 1976 to 1999 in Chapter 4 "Construction Safety;we find that among general contractors in states that repealed their prevailing wage laws,less serious injuries went up by 13% while more serious injuries rose by 23%. Among specialty contractors,less serious injuries in states that repealed their prevailing wage laws went up by 61k while more serious injuries went up by 12%. Only among heavy and highway contractors did injuries remain unchanged after the repeal of state prevailing wage laws. But the Basis-Bacon Act---the federal prevailing wage law--still largely regulated heavy and highway work before and after state law repeals. Thus, the dangers of repealing prevailing wage laws are real dangers born primarily by the workers themselves. However,to the extent that this increase in injury rates raises worker compensation 5 claims, the industry and owners wilt e!nd up atsa paying tar a mvie dangerous construction sector of the local economy. When advocates of prevailing wage law repeals make their case that wages will be cut substantially, they rarely mention that benefits wig be cut even more. This is obviously a problem for the construction worker and his or her family. Losing health insurance is a family problem. But to the extent society cares about the health of children, it is a public health problem tom losing pension benefits, c tMinq swat security contributions is a fart*problem. taut to the extent tl'tatt the old age security is a social concern, the loss of pensions and the cutting of social security is a public policy problem as well. Chapter 5, 'The Cruelest Cut"examines what happeris to health insurance and old age insurance when prevailing wage laws are repealed, In comparing Nevada construction workers with Arizona—a state that repealed its prevailing wage law we find that in 1997 (the last year of the Census of Construction? • Nevada construction workers earned$29,812 in wages-30A more than the average wage earner nationally contributing to social security. In contrast, Arizona construction workers earned$23,862 in wages--17% less than the average wage earner contributing to social security. • A ty =1 low wage earner contributing to social security who retires at age 85 can expect to receive$7,176 annually in retirement benefits wt* the typical average wage earner can expect to receive 65% more-. $11,844 per year. Thus, Arizona construction workers begin from behind in two ways as they prepare for retirement. a They make less, so they can save less. o 'TheycantrWe lase to social security, so they will receive less. But that is W the beginnft al the problem far Arizona construction workers. Arizona contractors pay 47% less into company health insurance and pen" contributions, This too Is a double wham". • Arizona workers can expect to (ec)P-%ve foughty hag as much in pension Income to supplement social security caompafed to Nevada constntc workers. 6 Construction is a remarkably unstable, cyclically sensitive, boom-bust industry, knitted together by a system of articulated subcontracting where the issues of coordination and responsibility continually threaten to fall between the cracks. It is an industry of intense competition where, In the public sector, at least,the job goes to the lowest bidder regardless of contractor reputagws, past histories of nonperformance, histories of litigation or noncompliance with safety regulations, building codes or other mandated standards. This creates a delicate environment and a s4 Pery sJgpe upon which the industry develops. While In many cases, the Industry maintains its commitments to training, safety, quality and on-time performance,in other Instances the industry slips into a downward cycle of negative cornpetftim and the sbandonment of the long-term costs of the Industry for short- term competitive gains. When this happens, the industry moves from a high-skill, capital-intensive, high-wage and quality growth path to a law-skill, under-equipped, low-wage and shoddy construction growth path. Safety issues are shoved under the rug and poor workmanship is hidden behind the walls. in the short run,this type of strategy can win in a market that does not measure downstream maintenance costs, does not adequately capture the costs of worker injuries, does not calculate the impact of lost health insurance and old-age pensions and fails to sponsor adequate training. Prevailing wage regulations are about keeping the Nevada construction industry on track, redressing market failures and ensuring that the local construction industry Is capable both now and in the future of building the quality infrastructure upon which all Industries In the state rely. 8 r �,�'� � r 1 r ,r l � � 1 �. a�. � � � _ { ' ���, , �. ... 4 ,y €f ; ` e ��. .� �ems"^�-.., �°i� � � I � f � L � � � � � I DNIS=BACON ACT THE PROMOTES. Of Preservation of the American Standard of Living 0 Fair and Cost-Effective Federal Contracting 0 Adherence to Free Market Principles Productivity in the Construction Industry 0 Quality Intrastructure 0 A Skilled Construction Labor Force Minority Hiring, Training, and Economic Advancement ,1 Healthy Families and Communities T Of COMO What is the Mavis-Baron Act? .... ..... ... ......... ..... . . ... .. . . 5 Whyis such a taw needed today? .... . ....... .... ........ .. .. . . . . b Is the Davis-Bacon wage a"union"wager ............... .. .... ... 6 Is the Davis-Bacon Act expensive for the federal government? ... . .. . t Should the Davis-Baron Act apply only to all federal construction projects financed by Peden d grants?. ....... ... .. . ..., .....,.. 6 Doesn't a cut in wages automatically decrease costs of construction to the federal government? ......... ................. . ....... . 7 How does the Davis-Baran Act improve local economics ....... .. . 8 Have we seen any evidence of what would happen without Davis-Bacon? . ............. . ... .. .... . . . .. .. ......... . .. 8 Could a repeal of Davis-Bacon result in cost shifting to other governmentprograms?. .... . ..... ..... ... ..... .<. ..... 9 Does the Davis-Bacon Art impede or improve the functioning of labor markets? . ............. .. ............. ..... . . . ........ 9 Does Davis-Baron Krause unemployment? . .. ...... ........... ..... 10 Why do construction workers need the Davis-Bacon Act? ........... 10 Wouldn't non-union workers be-nefit from repeal of Davis-Bacon:' .. .. 10 Does the Davis-Bacon Act dLgedininate against minority workers?, ... 11 Are most minority groups for or against Davis-Bacon? ...... ... . .... 12 Is there a connection between safety nand health for workers and communities and the Davis-Baron Act:' . .. .... ...... . . .... .. 12 I:Iow can you argue that Davis-Bacon adheres to market principles? .. 12 References ..... ..... . . .... . . ............... . .. ... . .. ..... .... 12 3 F QUESTIONS Fc t ER Q: Mat is the Dav"ac ACt? A: The Davis-Baran Act--also known as the prevailing wage Inver— "� preserves local area wages and labor standards in the process of letting contracts for federal construction work.Enacted in 1931,the law states that contractors rctors for federal projects rnuA pay their workers no less than the wage rates prevailing in the local area for each craft,as determined by the U.S.Department of Labor(17►OL). Q: why Is sah a law thy? A` For several reasons.The Davis-Bacon Act was originally intended • to encourage the development of a high-skill,high-wage growth path for the labor market in general,and the construction labor market in particular.Where the Davis-Bacon Act is applied,union and non-union contractors win federal construction jobs based on having the most productive,best equipped and best managed workforce. Under the Davis-Bacon Act,local labor standards cannot be artificially depressed by competition for federal constniction contracts.Nonetheless,critics of the Davis-flacon Act argue that the government should use its bargaining power to cut local wage rates.They contend that local wage rates could be exit by as much as fill percent..And they contend that such a race to the bottom can cut public c omstruction costs substantially.But when local wage rates are artificially depressed,as advocated by critics of the Davis-Baron Act,there is now a substantial body of evidence that indicates worker skills,experience and motivation also full off. Contractors no longer compete on the bay of who can best train, best equip and hest;manage a construction crew_Instead,they compete on the basis of who can find the cheapest workers either locally or through importing labor from elsewhere.This puts the quality of construction at risk.It may also lead to substantial cost overruns. Additionally,the absence of ax prevailing wage rate can cause downstream increases in building and road maintenance costs.And it definitely leads to an increase in c onstniction Wuries and a decline in the health and pension coverage of construction workers. This puts pressure on worker compensation costs.Similarly,it puts pressure on social services--as family health needs go urunet and retired workers cannot make ends nest:, ry The philosophy underlying the Davis-Bacon Act is that a community is better off in the long run by encouraging competition that builds skills,builds productivity,pays decent wages and provides for the health and old age needs of its citizens.That is the philosophy and that is what the Travis-Bacon Act does. A. No.The prevailing wage that must be paid on federal pjects is based upon typical wages and benefits paid for construction work in each comnnrnity,regardless of whether those workers are union members.According to the Department of Labor,a whopping 72 percent of wage determinations issued in 2000 were based upon non union scales of labor.A union wage only prevails if the FAIL wage survey proem determines the local union wage to be the prevailing wage. Q I IM DavWhm Ad fop IM -fedWS NO MIMI A: No.As John T.Dunlop, Ph.D.,Secretary of Labor under President Ford and Harvard University professor,concludes,Mvis-Bacota bs at least neutral with inspect to costs The nation's preeminent. economiast on construction,Dunlop observes that productivity is so much greater among high-wage,high-skill workers that often projects using such workers cost less than those using low-wage, low skill workers.Inferior construction requiring repairs,revisions and lengthy delays actually means the federal government could lose money if Davis-bacon is repealed.Other independent studies reach the same conclusion, Opponents of the law who claim that the government would Nave billions per year utilize vastly oversimplified and;fundamentally flawed methods of economic analysis which fall to take into ac- count productivity,safety,community development and other economic forces contributing to the real cost-effectiveness that: Davis-Bacon ofler_� M IN [MVIS48M k# �Iply OM to A M PNJBM '---- C Uy A: Not at all-Since 1931,Congress has on many occasions r+eeaffumred the principle underlying the prevailing wage concept of the Davis-Bacon Act as a futnda mental principle of public-policy(regardless of whether ti a federal construction program is financed directly b3'SMTO or indi- rectly by some other mechanism).Congress has developed many indi- rectly financed programs that provide federal construction assistance in the form of granter,loans,loan guarantees,insumnce,and other so- called innovative financing techniques such as tax credit bonds,state revolving loan fiords,and other credit enhancements that maximize the leverage of limited federal funds to facilitate urgently needed con- sttuction.Time and again in the past 70 years,Congress has applied the prevailing wage provisions in the Davis-Baron Act to newly-tievel- oped federally-assisted programs including those that assist,construc- tion of hospitals,waiter pollution control,highways,mass transporta- tion,airports,and housing.In spite of this continual reaffirmation or the prevailing,wage principle,opponents repeatedly attempt to block its application to various proposed new federally assisted construe- tion programs because it is an"unwarranted expansion of the Davis- Bacon Act."After all these years,consistent application of the Davis- Bacon prevailing wage principle to any new federally-assisted n- struction program can hardly be called-unwarranted."Simply put,the history of the Davis-Bacon Act clearly demonstrates that the specific type of financing mechanism utilized in a federal construction pro- gram is immaterial to the issue of Davis-Bacon scope and application- Q: DOOWt a CK ki VWMW Mt=890ft Code 18MI t A. Absolutely not. om Wage cuts don't automatically translate to pr cement • saving&if you pay someone half the wage you were paying someone else,but this person takes twice as long to do the,fob,you haven't saved a penziyy.And if the job is done so poorly that it mlu im hiritmg someone else to bring it up to standard,you're paging mom,not less. Repeated studies have proven that there is a direct correlation between wage levels and productivity—that well-trained workers produce more value per hour than poorly trained,low-wage workers. For example,a study of 10 states where nearly half of all highway and bridge work In the U.S.is done showed that when high wage workers were paid douUe that of low-wage workers,they built 74.4 mom miles ref mWbM and 3,8 wwm miles of bridges for$557ntiHion le&.s. Furthermore,most analyses fail to take into account the spin-off economic impact of maintaining prevailing wages.When workers' income goes down,they have less money to spend purchasing goods and making investments When businesses close or cut back as a result,lax revenues to the federal government decline and social expenditures rise.It is simply penny-wise and pound-foolish to assume that driving wages down will be of any benefit in helping balance the federal budget. 7 NOW does the 0804MM Act Drove Wd ec 0 1 mMbsy 1�. The government has alw ayrs�been a major purchaser of construction services,and this he local economies.Prevailing wage laws are also pro[ltable community investments.A study clone by Lionel Richman,I.I..B.,N.A.A.and Julius Reich,J.D.showed that in San Bernardino,California,the prevailing wage law generates benefits to the community 2,4 times the amount spent on the actual ec►n- struetion project:.That's because workers spend.part of their income in local shops and restaurants and pay local taxes,which recircet- lates throughout the economy. Furthermore,for the Ilispan c community in San Bernardino,Davis- Bacon gave some permanency to their ecmvnomie gains by support- ing apprenticing,as 48 percent of apprentices were Hispanic. Additional economic benefits of prevailing wage laws include- • Maintaining funding for building trades health and pe milon plans and training programs; • Immunizing the employee from flip need to seek benefits from st iatl programs; • Contributing to the ability of the community to assist the needy;and • Establishing an upwardly mobile track for minority members of the community to advance into higher-paying occupations. Q: Dave we sin My OVMOM d Met Wold hen wit Davis- ? s Yes,and its not a pretty night.There are 12 states that have repealed their own prevailing wage laws over the past two decades,and the consequences have included: • Competitive pressures in the industry leading to lower wages and fewer benefits; • Reductions in and wholesale elimination of apprenticeship training prograntis, • Declines in the quality of the workforce as the best candidates find careers in other industries more appealing*and • Increases In injuries. and deaths on the job as more untrained workers are employed. Repeal of tliaWs prevailing wage law raiused a decline in average eonviruction wages in the state:mid decreased union apprenticeship R training for construction.No other public or private source offset this decline in training.This led to high turnover of nonunion apprentices,despite contractors'efforts to retain workers.Overall, the construction industry lost a significant portion of its human capital. Another study in Iowa showed thatt contractors did not pass on savings to the taxpayer from paying lower wages—instead they lined their own pockets. of BOm SMM Y � �5 A•. Yes.prevailing wage laws often insure that bonafide health,pension ,and educational benefits are included in the DOL prevailing wage rate.('Chew benefits are included in the wage determination).In addition,if construction wages decline significantly,there will be a corresponding rise in the demand for government programs,ranging n --not just _ students to Food Stain le a etude � J iriotn flnatacisi aid for college eunong the families of construction workers but among owners and employees of business patronized by these workers. Furthermore,a current practice in some segments of the industry on private sector projects is for employers to miscl-assil'y workers, enabling irresponsible contractors to avoid paying employment taxes,such as social security,unemployment insurance and workers'compensation.This doses not mean that:costs are lowered —it means that omegas pay for these costs.Without Davis-&~moon, such practices would be extended to government contracts,with sonic employers effectively using tax dollars in staff other taxpasyets. Q: -BWM Aft knpft ar MIS M A� of A. Davis--Bacon promotes sound invesunent in human capital and in our physical infrastructure.As 77ir Wa11,51 reef Journal noted,there are severe shortages of skilled work in construction in many areas of the country.'%`hen wages are cut,the lndttstry's ability to attract and train qualified individuals to work on constxuction projects is hindered even more.An adequate wage is essential to forming human capital within the industry. Because of its cyclical and extremely competitive nature--and our reliance an infrastructure? for economic development and national security—construction labor markets must be protected. t~om ruction workers are trained for their skills.It often takes years of schooling and apprenticeship to gain proper experience—and 9 the importance of training is greater than ever at a time when rapid technological advancements are changing the nature of work in the indu-stry Tb retain a skilled workforce,workers must be paid fairly --by employers who contribute to training programs. Davis-Bacon ensures the proper functioning of labor markets by grounding the indusWs competition in fair wages:making contrav- tors compete on more efficient management and entrepreneurial techniques. 10118 A No.Unemployment is not caused by paving workers the prevailing ■ wage in the Industry—it is caused by other factors in the econonny mach as need for training,rapidly changing technology,under- investment and underconscmtption,and too-high interest rags. Besides,government pe)licy aimed at reducing unemployment by reducing wages simply guarantees even more hardship and economic anxiety for the hard-pressed middle class and working families. Q: Why do + 14Wft WOi hrs so A First,the Davis-flacon :"Act prevents big government and big busine r from undercutting local wages.Making government and Business pay prevailing wages in each community protects local,private industry and apprentice programs.Second,while many people believe that construction workers make above-average income,the typical annual income for construction workers is$35,000,which is Below the median family income in the United States. One important economic outcome of the Davis-bacon Act is that,it provides some stability to uncertain employment,which helps working families.This,in turn,protects one of the last remaining industries in the United States to employ blue-collar workers.With- out it,the.industry will follow the low-wage,low-growth and low- productivity path of other industries. Q* t t Wn WNW fit aw P"al of -B --HT Ar Just the opposite.They would find their wages and benefits driven ` down to even lower levels.A.%noted previmmly,in more-than seven of every ten communities,non-union wage scales are the prevailing 10 wage.So they do not find their employment opportunities hindered by Davis-Bacon,in fact,Davis-Bacon extends to non-union workers many of the protections enjoyed by union members. Without Davis-Bacon,the downward pressure on wage and benefits would weaken or render non-existent the already rmble attempts by non-union contractors to establish apprenticeship and health and safety programs,making non-union constme ion work even more dangerous than it already is. r � na"iiy W S$ n Absolutely not!Opponents of the Act--not otherwise known for ! � their advocacy of civil rights and affirmative auction—have repeatedly charged that Davis-Bacon discrin inates against minorities.This is not only a lie that egregiously misstates both history and contemporary reality--it is also premise..on the pernicious notion that the only way to hire minorities on construc- tion projects is to pay theta less. As a matter of historical.record,Sen.James J. Davis(R-PA,and former secretary of labor under President Harding),Rep.Baron (It-NY)and countless others supported the enactment of the DaAs- Bacon Act because it would give protection to all workers, regardless of race or ethnicity.The overwhelming legislative intent of the Act is clear.all construction workers,including;nunority employees,were rescued from abusive industry practices. Mandating that a fair and livable wage be paid to every worker not only stabilized local wage rates and labor standard for local wags earners anti local contractors but also prevented migratory contracting practices which treated African-American workers as exploitable indentured servants. 7bday,thanks to the Davis-Bacon Act,African Americans,Latinos, .Asian Americans,Native Americans and women are able to secure fair wages for their work on federal projects.In fart,minorities are heavily employed in the construction industry—especially in the unionized sector,whore union apprenticeship program graduate a greater number of minorities than non-union apprenticeship programs. Norman Hill,the Preskent of the A.Phillip Randolph Institute, stated that minority workers are"particularly vulnerable to exploita- tion such as the wage-cutting practict-,which the Davis-Bacon Act- of 1031 is deigned to prevent."Repeal of the Act,would leave mi- nority workers with the twin specters of unemployment.and wage reduction. lI Are rest MMMV SUP UP W 29dW A: The Davis-Bacon Act has long enjoyed the support of minority and women`s groups.The NAACP passed a msolution supporting Davis- Baron enforcement and its role in strengthening of opportunities for minorities through training and apprenticeship programs.Latino, Native American and women's groups have spoken in favor of the law.They recognize Davis-Bacon's role in helping to bring many minority families into the middle class.Davis-Bacon is important to women's pay and equality on the jab,says Tradeswomen Now and Tomorrow tTN' ,a national coalition of tradeswomen advocates and organizations, Q: IS MV a COMMft ba11Wa81 ably am MIM fop wat'ka1'S �NI 4' ilat iha Danrlaan A■ Yes.The skilled,trained and dedicated workers who are hired at ■ prevailing wages are trained to work safely.Better project safety and quality mean fewer risks of environmental or health disasters to communities.By preventing shoddy,unsafe work which can occur from employing poorly trained workers,our society actually saves money on environmental and economic clean-up costs. Q: Now M YOU XW that Dow aa-B coo IdIMM to NOW As Because the government plays by the rules set by the private sector in the free market.Large-scale government spending on highways, bridges,office buildings,harbors,sewage treatment plants,military construction and other rojects has the potential to skew the market and throw it out of balance,with serious consequences for private industry. Davis-Bacon makes government play a neutral role by paying the same average wages paid by the private sector.It imposes no artificial standards,and instead rots market forms. REFERENCES: ,Alen,Steve"Unionized Construction workers am More Productive"The Quarterly Journal of Ec ononviv%(May 1984). Allen,Steven,State Pmralling Wa Laws anti C(amtrudion Costs:A Reply to Professor Thiebl)t" 12 Azari-Rad,liaanld,Anne Beagle and Peter Phillpa"The Effects of the Retread of Utah'rs Prevailing Wage Law on the Construction Labor Market.'Labor Law Reform (Forthcoming:Cornell University Press Spring 19M). Bourdon,Clinton C.anti Raymond E.Levitt;"A Compasuan or Wages and construe.- tion,"(MIT Research Report No.R-78 3). Bureau of National Affairs,-Repeal F M Choice or Davis-Bacon Foes,'Construction Labor Repnri,12-2149,(Vol 4A 1).995). Dunlop,John T.(December 17,1990)Affidavit at par.9 Building and C'nrnstraction 7Vados Department tr.Matlitt,961 F,2d 269(D.C.Cir.1992)(No.9Q-53451). Lockett,Briars,"Specialty Contractors Loltby Lawntgkers to Support Davis-Bacon, Terrorism Insurance"Construction Labor Report[May J). Mandelstamm,Allan B."71ae Effect of Unions on Efficiency in the Residential Con- struction Indtoteyt A Case Study'Indwitrial said Labor Relations Review IS (Ifha6). Ph9lips,Peter,"A Comparison of Public School Construction Cow,In 11we Midwestern States that Have Cltuaged 71teir Prevailing W Laws in the 1996s, Kentucky,Ohio,Mkidgan,"ITniversity of Utah(FebruatyZWO. Prus,Mark J.,"Ptt vding Wage Laws and School Construction Costs,all►Analysis of .' States.". the Mid Atlantic Public School Construction in Maryland and Prepared for the Prince.Grntge's County Cauntiil,Maryland(SUNY Cortland ,lanuary 1999), Richman,Lionel and Reich.J.D.-Me Prevailing Wage Law.A Profitable Community lr(rtntent." Ruttenbetg,Ruth,`The Mavis-Bacon Act:A Respom-se to the Cal' O Institute"'Attack." U.S.Dept of Labor:Employment and Earnings.Vol.41 No.7July 1094. Sheehan,Michael,-State of Oregon;repeal provides no Direct Cost:SaviW Ballot Measure 12 laformation Packet. Sheehan,Michael,"The Economic Impact of A Prevailing War Law for Iowa State Construction PrqJecu4"Fisher,Sheehan and Colton,Scapposse Oregon. Sugarman,Lauren,Tradomtomen Now and Tcxxmrrttw,rlo Chimp Wooten In Thmies, 1657 West Adam,Suite.#401,Chicago,IL 00012, Tomsho,Robert,"Labor Squeeze:Ulth Housing Strong,Builders Often Find Skilled Help Lacking,'Wall Street Journal,Feb 27,1994 p.A-1. U:S.(fit of Commerce,U.S..Certsm Bureau,Income SeMeys IlmneWNHES Diiishon "MEDIAN INCOME FOR 4-PERSON FAMILIES,BY STATE,1999 Calendar Year.- U.S.Dent of Labor,Bureau of Labor Statistics,"A%vmge hourly and weekly earnings of 4Troduction or nortsupervisory workers'on private nonrarm 1trtvrolln by majar industry,seasonally adjusted,"Employment.and Earnings,May 209t,p.7. G.S.Dept of Labor.Bureau of Labor Statistics,Occupational F.nrployment Statistics, '2000 National Ovenpation Employment and Wage F.stimMO."*Cotatruetinn and Extraction occupations." G.S.General Accounting Office Human Resources Division,Feb 7,1994 letter to Senator Craig and Reprrzentatities Stenhotm,Goodling,Valentin and Petri from Div.Director Moue(GAOMEH&N-951?UavwBacon Act). U.S.House of Representatives tr'9th Congress,2nd session HR 17060,-Hours of labor and Wages on Public.Works."liewings Befbw the Committee on Laker,Feb 1% 1927, z PROT%CTINti TAN Li VI Hass+ Builftig&Construction Trades Department.AFIXIO 815 IGth Street,N.W, waslrit"l, D.C.20006 (202)t347-1461 %vww,bctd.org Edwarsi C,Sullivan Jor, ph Maloney President Secretary-Treastuer R ' s Los Angeles l Orange Counties 1626 Beverly Boulevard Building and Construction Los Angeles,CA90026-5784 �W$plPlp Phone(213)4834222 Trades Council (714)827-6791 RICHARD N.SLAWSON Fax(213)483-4419 Executive Secretary Affiliated with the Building&Construction Trades Dept.,AFL-CIO October 17, 2005 RECEIVED Council Members, OAF PUBLIC RE p Huntington Beach City Council CITY FFlC Huntington Beach, California JQIW L FLYNy,CQ Y CL=Mw RE: Public Works' Prevailing Wage Study Session and Review Dear Council Members: The City Council will be participating in a Study Session on Prevailing Wages to be presented by the City Staff at the Huntington Beach Council Chambers this evening. The Craft Unions,who represent the 450,000 Craft Workers,men and women, in California have had many opportunities to review the effects of decent wages, skills and training in our industry and see Prevailing Wages as vital to working families. The California Prevailing Wage helped to balance the wages of Construction Craft Workers in California for over 70 years. The conditions in the construction industry and especially Public Works Construction during that time have not changed the need for Prevailing Wages. Public Works Projects, awarded by the State of California, Local Communities and other awarding bodies are by far,the largest construction employers in the construction industry. Public Works is 27%of the total construction dollars spent each year. Obviously, this has a tremendous impact on the number of contractors that are able to participate in the industry and to a greater extent in maintaining a skilled, Craft workforce. Prevailing Wages were instituted to provide fair wages paid to taxpayers, in their own communities, who work in the construction industry. Study after study have shown that Prevailing Wages are cost effective,by attracting skilled Craftsmen and women;providing employment to local workers; attracting local, experienced and professional contractors to bid public works projects; supporting training and Apprentices in the Industry, and a number of other impacts that warrant the continuance of the requirement. We are providing you with a number of Studies that have been performed by federal, State and local governments, educators, as well as,research by recognized experts in the construction industry relating to Prevailing Wage and its impact on construction costs. We believe that you will find that the benefits of maintaining Prevailing Wages far outweighs any perceived negatives. Please take these research documents into consideration in your deliberations. We would look forward to meeting with you to discuss issues of Prevailing Wages and answer any questions you may have. Please call our office so that a meeting can be arranged. Sincerely, `. Richard Slawson ?Jim Adams Executive Secretary Business Representative Attachments: Prevailing Wage Studies The Davis-Bacon Act: A Response to the CATO Institute Is Attack �RU C T 10,N S l EEpEROoN 0 V � o x q N v Q Cl A O rrt * FL _ o � O o �. F P O .p �NDUSTR� O� C 1 FEB. 1 0 , Building and Construction Trades Department, AFL-CIO 81 S 16th Street, N.W. Washington D.C. 20016 ROBERT A. GEORGINE JOSEPH F. MALONEY President Secretary-Treasurer 3 THE DAVIS-BACON ACT: A RESPONSE TO THE CATO INSTITUTES ATTACK 'So what Is the opposition [to Davis-Bacon] about today? It Is to help these fly-by-night operators operating on the basis of greed against responsible construction work.' Rep. Augustus Hawkins (D-Calif.) May 3, 1988 'One of the myths about the Davls-Bacon Act Is that It only protects highly paid workers.' Rep. Bill Clay(D-Mo.) May 3, 1988 'By protections flowing from the Davis-Bacon Act In part, [the] lot [of minorities] has been Improved dramatically and an otherwise casual labor system has been partially stablized In many areas.' Dr. John T. Dunlop July 15, 1982 1. INTRODUCTION The Cato Institute has recently launched a broadside attack on the Davis-Bacon Act under the guise of advancing the interests of minority construction workers. The Cato Institute contends that Davis-Bacon was a "Jim Crow' law enacted to exclude black workers from federal construction projects, and that Davis-Bacon repeal in 1993 will improve the economic opportunities of minorities. Both arguments are utterly without merit. To support its contention regarding the intent behind Davis-Bacon, the Cato Institute engages in blatant distortion of the Act's legislative history and the entire historical record surrounding the law. The Cato Institute's contentions regarding the effects of Davis-Bacon are grounded in racist assumptions about the worth of minority construction workers, along with a willful disregard for the nature and realities of the construction industry today. As we will show in Section U, a fair reading of the legislative history of the Davis- Bacon Act negates any claim that the purpose of the Act was to exclude minority workers. In Section Ill, we will lay bare the Cato institute's claim that the Davis-Bacon Act reduces minority employment for what it is - a Trojan Horse for unfair contractors seeking to lower the wages and benefits of construction workers. We will further show that the Act benefits minority construction workers by protecting their wages and benefits from being driven down by vicious wage competition, and by providing the underpinning for the apprenticeship programs that are the route through which minority workers have been gaining access to high-skilled, high-paying construction jobs for two decades In Section IV, we will show that the so-called "helper" reform of the Davis-Bacon Act espoused by the Cato Institute will relegate minority workers to low wage, dead-end jobs in the construction industry. And, finally, we will describe the conclusions of leading. construction industry economists, such as Dr. John Dunlop, that the Davis-Bacon Act does not add to Federal budgetary outlays for construction. iL THE DAVIS-BACON ACT WAS ENACTED TO PROTECT ALL WORKERS AND TO DISCRIMINATE AGAINST NONE As opponents of the Davis-Bacon Act have tried to do for the last fifteen years, the Cato Institute claims that the law was passed in order to discriminate against minorities in the construction industry. However, anyone who has actually read the Act's legislative history knows that racial discrimination was in no way a purpose of the law. In attempting to show otherwise, the Cato Institute has manipulated and misrepresented the historical record surrounding the Act. -2- A. The First Prevailing Wage Laws The Cato Institute's misrepresentations start with its version of the purported origins of the Act. It claims that 'the story of Davis-Bacon begins . . . . in 192T' with, the construction of a Veterans'Bureau hospital in the district of Representative Robert Bacon (R-N.Y.). That assertion is simply wrong. The idea of protecting and stabilizing local laborconditions on construction projects was alreadywell-established before Rep.Bacon introduced federal prevailing wage legislation. By the time Congress considered national solutions, no less than seven states had already passed prevailing wage laws to bring stability to their construction industry.' Four other states, as well as the federal government, had enacted statutes giving preference to local contractors and manufacturers'and at least two other states statutorily gave preference to state residents in hiring workers for state construction projects.' Indeed, as early as 1891, the state of Kansas had adopted a law requiring that"not less than the current rate of per diem wages in the locality where the work is performed shall. be paid to laborers, workman, mechanics, and other persons so employed by or on behalf of the State of Kansas.' Thus, in 1927, Rep. Bacon did not originate the notion that a contractorshould be forbidden from destabilizing and degrading local labor standards. Rather, Rep. Bacon's legislation was an outgrowth of efforts across the nation to stabilize local wage rates and protect local contractors.5 B. The Original Davis-Bacon Act Did Not Have a Discriminato!Purpose To support its claim that the purpose of the Davis-Bacon Act was to exclude black workers, the Cato Institute cites only a few phrases, taken completely out-of-context from the voluminous hearings and debates on Davis-Bacon between 1927 and 1935. Significantly, none of the remarks cited by the Cato institute are by the law's sponsors, and are thus of no value in interpreting the purpose of the Act" In fact; there is absolutely no support in the legislative history of the Act for the Cato Institute's charge that exclusion of African-Americans from federal projects was a 'goal of the sponsors.' And the handful of remarks that the Cato Institute cites are selectively manipulated to grossly distort the legislative record. For example, the Cato Institute relies on the remark made by Rep. Upshaw(D-Ga.) in 1927 when Rep.Bacon discussed the previously mentioned Veterans'Bureau hospital in his district._Rep. Upshaw said to Rep. Bacon, 'You will not think that a southern man is more than human if he smiles over the fact of your reaction to that real problem you are confronted with in any community with a superabundance or large aggregation of negro labor.'' The Cato Institute fails to note that Rep. Upshaw was a lame-duck at the time he made that remark, having been defeated in his bid for renomination in 1926, and that he left office almost four years before the Davis-Bacon Act was passed." Thus, Rep. Upshaw was no supporter of the Davis-Bacon Act and, moreover, he was not even a member of Congress when it was debated and passed. -3- Even more importantly, the Cato Institute conspicuously fails to mention that Rep. Upshaw's racist insinuations were immediately and sound reiected by Rep. Bacon, the sponsor of the Davis-Bacon Act. Stating that migratory workers of all races were being brought in to work on the hospital, and replying that the race of the out-of-state workers was of no import, Rep. Bacon responded to Rep. Upshaw: MR. BACON. I just merely mention that fact because that was the fact in this particular case, but the same thing would be true if you should bring in a lot of Mexican laborers or if you brought in any nonunion laborers from any other State . . . In the case that I cite the contractor has also brought in skilled nonunion labor from the South to do this work, some of them negroes and some of them white, but all of them are being paid very much less than the wage scale prevailing in New York State . . . .° Thus, the Cato Institute tries to paint Rep. Bacon as a racist by quoting the remarks of Rep. Upshaw, and leaving out the response of Rep. Bacon: In so doing, the Cato Institute has engaged in a purposeful distortion of the legislative history of the Davis- Bacon Act. When one reviews the overall history of the Act as opposed to the isolated and out-of-context remarks cited by the Cato Institute;it becomes absolutely clear that the Cato Institute has misrepresented the purpose of the law. C. The Purpose of the Davis-Bacon Act Was to Protect Workers of All Races The true purpose of the Davis-Bacon Act was, in fact, to "give local labor and the local contractor a fair opportunity" to participate in federal construction projects." The roots of Davis-Bacon are solidly in the Republican Party and northern - not southern "Jim Crow" - support, including the support of contractors. Rep. Bacon, a banker from New York, first introduced a bill in 1927 to require that local prevailing wage standards be met on federal construction projects. James J. Davis, Secretary of Labor under Presidents Harding, Coolidge and Hoover was elected as a Republican Senator from Pennsylvania in 1930. Shortly thereafter, Sen. Davis introduced a companion prevailing wage bill in the Senate. Support of the legislation was solid and widespread. In 1931, the Secretary of Labor, Assistant Secretary of War and Acting Supervising Architect all appeared at hearings to support the bill." Contractors also supported the idea of protecting local standards as many were being forced to stand by and watch as fly-by-night contractors underbid local employers by paying far below the area's wage rate. Clearly, the law was as concerned with protecting the local contractor as it was with protecting and stabilizing local wage rates. Fair contractors realized the two purposes went "hand-in-glove" with one another. Congress also recognized that both workers and contractors were seeing their -4- opportunities stolen by a "selfish group of contractors" who disregarded local standards and underbid local community-based contractors.- Congress was concerned that local contractors and local wage earners, who were contributors to their community's tax base, were being deprived of the economic benefits associated with the construction of federally assisted projects in their own communities.` The law, in the words of-Sen. Davis, was designed to remedy the situation and "give a square deal to all.M'3 The Cato Institute asks us to believe that the Act's many supporters were using racist code words for black labor when they described the underlying problem to be the use of"cheap, "imported", 'transient', "migratory"and "itinerant"labor to undermine local labor standards. But the record reveals that Rep. Bacon and other supporters were concerned about any degradation of local labor standards by low-wage workers brought from outside the locality, without regard to whether the "itinerant' workers were black or white workers. For example, Rep. Connery of Massachusetts described a hospital in Vermont built by a contractor who "brought in people from out West. . . at wages very much below the prevailing scale of union wages in Vermont," and a post office in St. Louis constructed by a contractor who "brings in labor from South Dakota."" The Cato Institute would also have us believe that the only contractors whose practices the Act was intended to preve nt t were southern contractors who employed black workers. Again, the legislative record demonstrates that this simply was not the case. In testifying in support of the original legislation in 1931, William Green, president of the American Federation of Labor, presented documentation about complaints he received from 16 areas around the country. His testimony makes it quite clear that the contractors who presented problems were not just southern "invaders"of the north. For example, one complaint involved a contractor from Des Moines on a project in Arkansas. Another dealt with a Louisville contractor on a post office facility in the same city. Others involved a Texas contractor on an El Paso project, a New York City contractor on a Utica, New York project, and a Pennsylvania contractor on a New Jersey project." The Davis-Bacon Act was not only intended to stabilize local wage rates and labor standards for local wage earners and local contractors, but also was intended to help migratory workers by preventing unsavory contractors from exploiting them. Representative LaGuardia (R -Progressive- N.Y.) appreciated the fact that the Act would help all workers. As Rep. LaGuardia stated, "[t]he unfair and unethical contractor however, after getting the contract and being paid on such basis, turns around and imports labor from other localities at low and reduced prices, not only exploiting his own workers, but all to the discrimination and disadvantage of labor living in that vicinity.," After witnessing the very project which the Cato Institute alleges gave rise to the Ac4 Rep. LaGuardia further stated., l saw with my own eyes the labor that[the contractor] imported there from the South and the conditions under which they were working. These unfortunate men were huddled in shacks living under the most wretched -5- conditions and being paid wages far below the standard. These unfortunate men were being exploited by the contractor. " Clearly, the Davis-Bacon Act was not designed to discriminate against minority construction workers. Rather it was implemented to protect all workers, be they white, black or any other race. Moreover, the Act sought to protect and stabilize labor standards to aid local communities and contractors who were having their areas raided by a select group of unscrupulous contractors. In suggesting that Davis-Bacon was enacted for a discriminatory purpose, the Cato Institute has deliberately turned a blind eye to the law's true intentions. Indeed, if the Davis-Bacon Act were repealed, as the Cato Institute urges, then today's construction workers would risk sharing the fate of the unfortunate persons, black and white, who were employed by the infamous Alabama contractor in New York in 1927 -- all could be paid wages far below the standard and all could be exploited. In sum, the legislative record demonstrates that Davis-Bacon was enacted to save all construction workers, including minority employees, from abusive practices by mandating that a fair and livable wage be paid to every worker. ill. THE DAVIS-BACON ACT BENEFITS MINt)RITY CONSTRUCTION WORKERS TODAY A. The Davis-Bacon Act Does Not Reduce Minority Employment The Cato Institute also charges that the Davis-Bacon Act reduces employment opportunities for present-day minority construction workers. But an examination of the economic realities facing construction workers today shows that the Cato Institute is engaging in a hollow rhetorical assault upon the Act. The Davis-Bacon Act erects no barrier to the employment of minorities. It does nothing to exclude from federal jobs those contractors, either union or non-union, that employ minority construction workers. It simply requires that those contractors pay all their employees, including minority employees, wages that are no less than the locally prevailing wages. Thus, far from being discriminatory, the Davis-Bacon Act mandates equal and fair treatment of all employees. According to the Cato Institute, the prevailing wage requirement hurts minority workers because if a contractor is required to pay a prevailing wage to its employees, it will not hire minorities. The gist of the Cato Institute's argument is that contractors cannot, and should not, be expected to pay prevailing wages to minority workers. The unstated assumption behind that argument is that minority construction workers do not deserve to be paid the same prevailing wage as their white counterparts. That racist assumption lies at the core of the Cato Institute's argument. The Cato Institute tries to veil its racist premise with economic jargon, contending i that prevailing wages bid minorities out of the labormarket because minority construction employees are "unskilled." The labeling of minority employees as "unskilled' is itself offensive and racist. It also is reminiscent of the discriminatory practice of underclassifying workers (often minorities) into lowerskilled job categories(i.e. "helpprs") to justify the payment of lower wages, even though the workers actually perform more highly skilled work.18 The Cato Institute claims that unskilled minority workers cannot gain employment on projects subject to the Davis-Bacon Act. This is a bald-faced lie. Since 1935 and continuing to the present, the Labor Department has continuously issued prevailing wage determinations for skilled, unskilled and semi-skilled laborers' classifications. For example, the "General Wage Determinations Issued under the Davis-Bacon and Related Acts"1' establishes the job classifications for laborers on Davis-Bacon projects in counties in all fifty states. And who are these workers employed daily on thousands of federal construction contracts subject to the Davis-Bacon Act? They are in large part minority workers, whether they be African-American, Hispanic or Native American.20 Dr. John Dunlop, former Secretary of Labor under President Ford and the nation's preeminent construction industry economist, has de$cribed how the Davis-Bacon Act is integral to the continued employment of these minority laborers on federal construction. For example, he states that under the auspices of the Laborers'-Associated General Contractors Educational and Training Fund, 31,913 construction laborers received training in 1989 from 73 local training funds. Forty percent (401/1o) of these Laborers' trainees were minority and female workers.21 These are the minority and female workers who have gone o n on t employment in 'ob 9jobs protected by the Davis-Bacon Act: If the Cato Institute has its way and Davis-Bacon is repealed, it is these minority workers who will face either job loss or an immediate reduction in wages and fringe benefits.22 We believe it is Dr. Dunlop, and not the Cato Institute who understands nds the effects of Davis-Bacon on minority workers. And the Cato Institute offers virtually no proof to back up its claims. The Cato Institute does not cite any statistical evidence showing that Davis-Bacon reduces minority employment, but relies only on conclusory statements made by those supporting repeal of the Davis-Bacon Act more than a decade ago. The Cato Institute fails to provide any evidence that employment of minority construction workers on projects covered by the Davis-Bacon Act is disproportionately low, or any proof that the employment of minorities on such projects is lower than it would be if the law were not in effect In fact data gathered b the Department of Labor indicates that 9 Y P the Davis-Bacon Act does not reduce minority employment. According to reports submitted by construction employers to the Office of Federal Contract Compliance Programs in 1991,23 the percentage of minorities employed by contractors on federally funded projects (20.12%) was nearly identical to the percentage of minorities employed on non- federal projects (20.56%).2' And, for classifications covered by the Davis-Bacon Act - e i -7- craftworkers, operators and laborers - the percentage of minorities employed by federal contractors was higher than the percentage of minorities employed by contractors who were not working on federally funded projects: % Minorities % Minorities for Federal for Non-Federal Contractors Contractors Craftworkers 17.61 17.41 Operators 26.22 24.4 Laborers 41.18 39.1325 These statistics bell the Cato Institute's assertion that the Davis-Bacon Act has a negative effect on minority employment. Lacking direct evidence that Davis-Bacon has an adverse effect on minority employment, the Cato Institute attempts to prove its'point by asserting that minority employment is higher in non-union than union construction. Its sole source for this charge is Dr. Armand Thieblot, a long-time opponent of the Davis-Bacon Act who has devoted much of his career to its repeal.26 And, not surprisingly, the Cato Institute fails to provide any data on minority employment to back up this claim as well. The available evidence contradicts the Cato Institute's assertion: according to data from the U.S. Department of Labor's Bureau ofApprenticaship and Training(BA7), minority participation in union apprenticeship programs is consistently higher than minority participation in non- union programs.?7 .The same data reveals that the drop-out rate of minorities from apprenticeship programs is much lower in union programs than it is in non-union programs.28 Academic experts reach the same conclusion. In their book Union and Open-Shoo Construction. Dr. Clinton C. Bourden of Harvard University and Dr. Raymond E. Levit of MIT examined the relative degrees of minority entry into the construction industry through the union and non-union sectors, and found that"ftabjulation from data show that minority participation, both in terms of percentages and absolute numbers, is substantiaW higher in union programs."2' Moreover, the Davis-Bacon Act does nothing to require the use of union rather than non-union contractors on federal projects, and a significant percentage of the work on such projects is performed by non-union contractors and employees. It comes as no surprise, however, that the Cato Institute tries to equate Davis-Bacon coverage with union status: that misleading ploy has long been favored by Davis-Bacon foes.3O The tactic ignores the simple fact that the Davis-Bacon Act protects all workers, whether or not they belong to a union." i -S- I' B. The Davis-Bacon Act Protects Mlnority Construction Workers From Exploitation and Wage-Cutting The protections provided by the Davis-Bacon Act are urgentlyneeded byAmerica's • construction workers. The nature of the construction* industry, and the special r characteristics of the government o g nt contracting process, make construction workers on government projects particularly susceptible to wage-cutting competition. The construction industry is intensely competitive. There are hundreds of thousands of contractors,mainly small business ventures, competing forwork on federally funded projects.72 Because it is relatively easy to go into business as a contractor, firms often enter the market to compete for work on a particular project' In the case of government contracts, the government is required by law to award the contract to the i lowest bidder, with a few exceptions. Consequently, the pressure on contractors to minimize their costs is enormous.' And, because the employer in the construction industry has little control over most of its major costs, such as land, materials, equipment, and interest payments, the cost-cutting is invariably directed at wages and other labor costs. v Unfortunately, construction workers are in a poor position to resist these wage- cutting pressures. Construction employees have virtually no job security, as they rarely form the kind of long-term employment relationship that is the norm in most other industries. Instead, they are hired only for the duration of a particular project; and must look for a new employer when that project is finished."" Employment in construction fluctuates widely depending on the season, the economy, and various other circumstances,3Q and construction workers face unemployment rates that are persistently high." For all of these reasons, there is a perpetual pool of unemployed construction workers competing for available jobs. And because of their desperate circumstances, unemployed construction workers can be induced to go wherever the jobs are, and to work at virtually any price. Under these conditions, the offer of work can be and is used to drive down wage levels.J' Without the minimum wage floor provided by Davis-Bacon, all contractors would be forced to bid down the price of labor to match the lowest common denominator. Then federally-funded construction contracts would destroy locally prevailing wage structures as contracts would be awarded to the lowest bidders - the contractors who pay the lowest wages to the most desperate and least skilled workers. The protections provided by the Davis-Bacon Act to wages and benefits are especially important to minority employees. As former Secretary of Labor, and respected labor economist, Ray Marshall has observed, "[t]he workers most often victimized by unscrupulous contractors are the minority workers, whether he or she is Black, Hispanic, a native American or an undocumented worker . . . Davis-Bacon is an integral part of ensuring a decent life for the hardworking men and women in the construction -9- industry.*" Similarly, Norman Hill, President of the A. Philip Randolph Institute, has concluded that minority workers are "particularly vulnerable to exploitation such as the wage-cutting practices which the Davis-Bacon Act of 1931 is designed to prohibit.'`O Given its extremist ideological bent, the Cato Institute would probably deny that there is such a thing as exploitation of workers, but America's construction workers know better. Even with the Davis-Bacon Act in place, exploitation of minority workers goes on today by dishonest contractors. For example, testimony submitted by a Department of Labor official to the Senate Subcommittee on Labor contained a vivid description of just how Davis-Bacon violations can have a particularly harsh impact on minority workers: Violations continue to mount as unscrupulous contractors come on the scene and old contractors take more chances and become more inventive in their efforts to evade the requirements of the Act. Outright falsification and concealment is still found in many cases. One case that involved man forms of cheating employees was a fine [One] Y 9 which took the easy route of employing primarily undocumented workers. These workers will not complain. They represent an ideal work force for those who would exploit labor on government jobs . . . This subcontract was for the fabrication, transportation, and installation of bridge railing on a bridge across the Potomac River. The company employed undocumented workers at rates of $10.00 per day plus food and lodging for work days of 7 to 10 hours daily, 6 and 7 days a week. It should be noted that this contractor is transporting many undocumented aliens from the South Texas area where wage rates are lower, to the Washington, D.C.area with higher prevailing wage rates . . . [One] Arkansas contractor was found owing $7,000 in back wages to employees. Payrolls were falsified to show compliance. . . The employees were all black and another example of a group exploited by an unscrupulous employer." If the protections of the Davis-Bacon Act were removed, many more minority workers would face this type of exploitation. And all construction workers, including all minority construction workers, would be forced to accept lower wages and reduced or no benefits when working on federal construction projects.42 To claim, as does the Cato Institute, that reducing the wages and benefits of minority workers is somehow in their best interest is ludicrous, and smacks of the worst sort of racism and paternalism. Unlike the Cato institute, minorities recognize that their interests are served by the preservation of the Davis-Bacon Act. For example, many representatives of the African- American community have supported Davis-Bacon because of its role in protecting f minority workers. As noted earlier, Norman Hill, President of the A. Philip Randolph -10- Institute, has acknowledged the importance of Davis-Bacon in preventing exploitation of minority construction workers.' Similarly, Rep. Augustus Hawkins, former Chairman of E the House Education and Labor Committee, has been a leading supporter of the Davis- Bacon Act; hailing it as "vital worker protection.'' Significantly, Rep. Hawkins also has described the current role of Davis-Bacon in protecting workers in much the same way as Rep. Bacon and others did in 1931. He has stated: I think we should understand how Davis-Bacon actually operates . . . [S]uddenly, you are faced with a contractor who comes in, bids on a contract, based on lower wages paid its Itinerant workers, and ignores equity, and the moral responsibility to your community. They could have come from thousands of miles away In order to underbid your local people. Now this results, as has been stated In the past, In very # shoddy construction work. Now maybe it won't appear in the first few } years, but over the life of the construction, shoddy workmanship will be revealed. It was to prevent such practices that Davis-Bacon was started in ' the first place back in the 1930'S.45 As this demonstrates,passage Rep. Hawkins clearly doss not believe that 'itinerant' labor is a racist code word or that the original purpose of the Davis-Bacon Act was to prevent black employment on federal construction jobs. Similarly, in a debate on a bill to weaken Davis-Bacon, Rep. Bill Clay of Missouri, a long-time member of Congress and former Chairman of the House Subcommittee on Labor-Management Relations, described how Davis-Bacon is particularly critical in protecting minority workers: One of them s about the Davis-Bacon Act is that it only y protects highly paid construction workers . . . This proposal [to weaken Davis-Bacon] would cause serious problems for minority and lower paid construction workers . . . This [bill]would severely limit the employment standards and opportunities of the segment of the construction work force which is largely composed of minority members . . . It is clear that[this bill] will limit training of young workers and have a devastating impact on the living standard of minority construction laborers- many of whom now struggle to maintain a decent standard of living at backbreaking and dangerous work.iQ Like these distinguished African-Americans, leading organizations representing minorities and women support Davis-Bacon. The NAACP, the National Women's Political Caucus, the Navajo Tribal Council, and the Mexican American Unity Council have all expressly endorsed the Davis-Bacon Act." The Cato Institute's cavalier dismissal of the -17- opinions of minority group representatives and organizations, as if it knows better than they what is best for minority workers, is yet another example of its patronizing and racist attitude. In sum; as is widely recognized by those who, unlike the Cato Institute, truly have the interests of minority workers in mind, the Davis-Bacon Act benefits minority construction workers by ensuring that they are paid fair and decent wages and fringe benefits when they work on federal projects.'' C. The Davis-Bacon Act Benefits Minority Construction Workers by Providing the Framework Within Which the Nation's Apprenticeshlp System Has Grown and Flourished The General Accounting Office (GAO) has found that the "major incentive for construction contractors to use apprentices has been that registered apprentices can be paid less than prevailing rates on federally funded construction projects covered by the Davis-Bacon Act."" Davis-Bacon is, in shoe; the underpinning for the private apprenticeship system in the construction industry. The important interplay between Davis-Bacon and the growth of the private construction apprenticeship system has existed since the 1930's. Under the 1937 PP P Y National Apprenticeship Act, a mechanic or laborer classified as an apprentice or trainee can be paid less than the prevailing wage rate on a project covered by the Davis-Bacon Act if he or she is enrolled and registered in a legitimate training program with the Labor Department's Bureau of Apprenticeship and Training(BAT) or a state agency recognized by the BAT. This interplay between the Davis-Bacon Act and the NationalApprenticeship Act has ensured a skill base that has produced quality and safe construction on federaliy- financed and assisted public buildings and public works. The BAT-approved joint union- management apprenticeship programs have proved for more than a half century their worth in providing the training and skill development our nation must have. Repeal of the Davis-Bacon Act would permit contractors to employ unskilled and semi-skilled workers on projects now covered by the Davis-Bacon Act without registering them in apprenticeship and training programs approved by the BAT or an appropriate state agency. so Repeal of Davis-Bacon and the resultant erosion of the apprenticeship system in the construction industry would have a terrible impact upon minority construction workers because of the "increase [in] access of minorities to the jobs and apprenticeship' in construction since the 1960's.51 The GAO has found that '[sjince 1973, the proportion of minorities in apprenticeship programs has risen by nearly 50 percent to 22.5 percent of apprentices . . . about the same as their representation in the labor force. . . [By/ 1990 the number of minority apprentices had exceeded the previous high reached in 1981."S2 And, the largest number of civilian minority apprentices are found in construction trades, including carpenter (22.3 percent minority), roofer (38.201o); painter (28.901o), operating I ------------ F -12- engineer(32.401o) and cement mason (48.1%).� Thus, formal apprenticeship programs are now the premier path for minor' ty workers into skilled, high wage jobs in construction. That fact has been confirmed by Dr. John Dunlop. Based on his in-depth examinations of the construction industry, Dr. Dunlop has stated. My experience teaches that formal training programs are essential to recruit and train minorities for the construction industry. Indeed, this is how progress has been made. Over •' the past decade, substantial progress has been made in recruiting minorities and now women to the ranks of the trades and also in Placing them into 9 bona fide craft ft apprentice programs. That system deserves the support of our government." It is clear, therefore, that repeal of the Davis-Bacon Act would destroy estroy f private apprenticeship system in the construction industry, and consequently would wipe out the decades of progress made by minority workers in securing high wage, skilled construction jobs through participation in apprenticeship programs. IV. THE SO-CALLED "REFORMS" OF THE REAGAN-BUSH ERA WILL DEPRESS MINORITY EMPLOYMENT ON FEDERAL CONSTRUCTION PROJECTS seeks The heviscerate Institute, like may other opponents of the Davis-Bacon Act, the law by a series of reforms' because they have failed to convince Congress to repeal the Act. The leading such "reform*, begun in the Reagan-Bush era, was the introduction of low- ■ low-wage onto federal construction projects. The result of the helper"reformemeasu a is to provide contractors a source of cheap labor on federal construction projects, for whom they need not provide training that would allow these workers to advance up the skill ladder. Or. Dunlop has studied the so-called helper "reform' in great depth and has concluded that its chief victims will be minority construction workers.5d Dr. Dunlop's examination shows that weakening the Davis-Bacon Actbypermitting the unregulated employment of low-wage,untrained"helpers' on federal construction projects will sharply reduce minority employment on these projects: 1 strongly disagree with the conclusion that allowing contractors to employ the helper -13- classification throughout the entire construction Industry will enhance work opportunities for minorities and women.... Under the new helper regulation, it is my opinion that minorities and women who have achieved employment in the building trades and are currently enrolled in bona fide craft apprenticeship programs will be replaced with lower paid and untrained helpers ...57 Like Dr. Dunlop and many otherneutrals who have examined this Davis- Bacon "reform", United States District Court Judge Harold Greene has described how the use of "helpers' will harm minority employment: "The [helper] regulation adopted by the present Secretary [of Labor Raymond Donovan] is likely to have the effect of allowing contractors to replace higher wage minority laborers with lower wage minority helpers.n58 Judge Greene rejected Secretary Donovan's argument that the "helper" reform would in any way advance job opportunities for women and minorities and specifically found that the it would result in the creation of a "cheap" classification from which such workers would have little hope of escape. Judge Greene stated: The Secretary defends the regulation in part on the ground that it will facilitate non-formal training of women, minorities, and young workers.. . . . In fact', it will assign members of such groups to the lowest classification of workers, and It will likely keep them there on a. permanent or long-term basis.59 Not surprisingly in light of these facts, Congress voted three times in the 102d Congress alone to stop the Secretary of Labor from implementing the helper "reform".60"reform". The debates from these votes show that members of Congress share the view that this "reform" will do nothing but cost African- Americans decent jobs, good pay, benefits and training opportunities. For example, Rep. Craig Washington (D-Tx.) observed 'that construction laborers are the largely minority and female component of the construction workforce . [l]f the new helper class is now used . . . [t]heir wages are cut and they lose health care coverage . . . [C]ontractors can now simply reclassify these minority and female laborers as helpers and cut their wages and eliminate their fringe benefits."Q1 The debate in the Senate was similar.' Sen. Tom Harkin (D-la.) stated '[w]e are going to have a whole subculture of helpers with subminimum wages, low wages with no hope of getting out of it.' In an earlier vote on the helper "reform' in 198r, Rep. Bill Clay pointed out that the "minority members of Congress' believe that the helper -14- rule would create a subclass of low wage, minority workers on federal I construction projects: There are a number of leading minority members of Congress who specifically dispute that the new helper regulation will advance minority employment ... The proposal will allow contractors to substitute helpers for laborers and pay them lower wages for the same work ... It is clear that the Stenholm helper proposal will limit training of young workers and have a devastating impact on the living standard of minority construction laborers.05 A Once again, while the Cato Institute may believe the helper"reform"will advance minority employment, African-American representatives in Congress, such as Rep. Clay, soundly reject that view. V. THE DAVIS-BACON ACT DOES NOT INCREASE FEDERAL CONSTRUCTION COSTS t Finally, the Cato institute argues that the Davis-Bacon Act costs the federal government"billions every year...." Like other critics, the Cato Institute offers no sound economic analysis to support this claim. It is not surprising, because the leading construction industry economists in the nation have concluded that the Davis-Bacon Act is either neutral with respect to costs or r that repeal of the Act may even add dollars to the federal budget in the long- run. The so-called "studies" of Davis-Bacon's impact upon federal costs relied upon by the Act's opponents use an economic methodology that equates wage reductions for construction workers employed on federally- funded construction projects with a dollar-for-dollar savings in federal budgetary outlays for that construction project Dr. Bourdon and Dr. Levitt have described such an approach as a "crude methodology."°Q Dr. Dunlop has described this methodology as "a simplistic formulation which is wanting from an economist's point of-view...."a' Why do leading economists such as Dr. Dunlop describe these r "studies" as "crude" and "simplistic"? It is because they fail to take into account, as Dr. Dunlop says, the "myriad"88 factors which go into the cost of a construction project, including productivity costs arising from low-paid, lower-skilled workers who take longerto perform theirwork, thereby increasing overall labor costs. The productivity factor is a highly significant one in the -15- labor-intensive construction industry and can wipe away any short-term cost benefits derived from lowered wage rates. These "crude"studies also fail to consider that if low-wage workers are utilized on federal jobs, an increased number of highly-paid supervisors must be hired to supervise these workers. This factor also drives up overall construction costs. Productivity is not the only factor ignored by these "studies" advanced by the Act's opponents. It has been shown that when Davis-Bacon is not enforced or is not applicable to a federal construction project any short-term savings which occur from a reduction in wages are offset later on because inferior construction leads to costly long-term maintenance, repair and reconstruction of the public work or building. Dr. Dunlop's economic analysis concludes that Davis-Bacon is at least "neutral with respect to costs"because of the productivity differences between high and low wage workers and the resultant cost of inferior construction.°D He has stated. Its [Department of Labor] conclu§lon was reached by use of a simplistic formulation which Is wanting from an economist's point of view and which 1 find to be totally Insupport- able. Its methodology is based on a formula which utterly fails to take into account many of the real economic factors and forces which contribute to costs on a construction project. It appears that the Department takes the current wage rates and subtracts them from the savings to be gained by the cheaper rates.... [T]he Department multiplies that wage differentiation ... and then arrives at some projected mathematical savings....70 Dr. Dunlop further states: flJn the real world ... the speed of operations is also vital to costs...There is simply no sound basis for gratuitously assuming that lower wage rates in the construction industry generally mean lower costs to the public...." Many other studies have reached the same conclusion as Dr. Dunlop.'? In addition to Dr. Dunlop's study, other studies which have examined directly the factors contributing to construction costs on public jobs show little -16- if any correlation between overall costs on prevailing wage jobs and higher wage levels. For example, one study examined the costs of construction of new secondary schools on a state-by-state basis and ascertained the average cost of construction per classroom in new secondary schools during the period 1968 through 1974. It concluded that the highest costs per classroom were not necessarily in high-wage states with prevailing wage laws. For example, California, a high wage state with a prevailing wage law, was seventeenth from the bottom in costs. Among the states which paid more for the construction of new secondary schools were twelve that had no prevailing wage laws at all." Another study, using 1987-1990 data from the Federal Highway . Administration, examined comparative costs on federally-funded highway construction. That study shows that states in which higher-wage workers are employed to construct highways pay about ten percent less per mile in total costs even though they pay about twenty percent higher in labor costs per mile. This result was attributable to lower man-hours per m'� p ale (14,810) in higher wage states compared to 26,651 man-hours per mile in lower wage states - or a 44% difference." Also, a comprehensive analysis of the impact of Davis-Bacon was prepared by the HUD Inspector General and reported in the HUD Audit Report on Monitoring and Enforcing Labor Standards. it found that when Davis-Bacon is not enforced on federal construction, there is lost revenue in unpaid taxes and higher long-run repair and maintenance costs due to the original, inferior construction which comes from using low-wage, untrained workers. It found: ... [a] direct relationship between labor standards violations and construction deficiencies.... This systematic cheating is costing the public treasury hundreds of millions of dollars, reducing workers earnings, and driving the honest contractor out of business or underground's The HUD Audit Report also demonstrates the relationship of inadequate Davis-Bacon enforcement to poor quality construction: Competitive bidding frequently results in use of less skilled workers paid below prevailing wage rates and shortcut construction methods leading to poor quality work.... Direct correlation between labor violations and poor quality construction on 17 projects are shown In Appendix 3. On these 17 projects, we found violations and -17- construction deficiencies in the same construction trades. Poor workmanship quality, in our opinion, results from use of inexperienced or unskilled workers and short cut construction methods. Roofing short cuts result in leaks and costly roof and ceiling repairs. While short cuts in painting may not be as serious, it does require future maintenance expense by requiring repainting sooner than anticipated. Electrical short cut deficiencies are not as readily detected but may lead to serious problems such as fires and shocks.... Poor quality work leads to excessive maintenance costs and increased risk of defaults and foreclosures.'a In contrast to Dr. Dunlop's conclusions are three papers of the Congressional Budget Office (CBO) which examined various options for modifying the Davis-Bacon Act and attempted to estimate their impact on the Federal budget. The original 1983 CBO paper, while maintaining that "an assessment of costs against benefits must under-lie [sic] any possible legislative action on Davis-Bacon"nevertheless focused its attention primarily on Federal budgetary. costs issues rather than attempting to quantify both costs and benefits associated with the Davis-Bacon Act." in an exhaustive analysis of the first two CBO estimates, the Preliminary Review of Congressional Budget Office CBO Cost Estimates of the Davis- Bacon Act. Workplace Economics, Inc. (1986), concluded that the CBO papers are wholly inadequate as a basis to conclude that repeal of Davis- Bacon will, in fact; save the taxpayer money. It found: o The 1983 and 1986 CBO estimates fail to produce a full and accurate cost-benefit accounting of the impact of Davis-Bacon standards. Neither all the direct nor all the indirect benefits of the legislation were measured, such as positive regional income multiplier effects, the impact of the Act on the quality of construction and the Act's role in aiding the recruitment and training of skilled-labor; o The 1983 and 1986 CBO estimates rely in large part on a 1982 Department of Labor(DOL) analysis which was marked by significant shortcomings in that it failed to provide any quantification of -18- possible offsets to the purported cost savings, such as offsets that may be -associated with a decline in apprenticeship programs. o The CBO failed to take into account that fewer apprenticeship programs will either add to the cost of training workers or, without employer-provided training, will result in a long term shortage of y skilled craftsmen. Fewer apprenticeship programs will also result in the social cost of fewer women and minority craft workers being trained. None of these potential offsets were quantified by eitherthe DOL or the CBO. o The 1983 and 1986 CBO estimates failed to provide any offsets for productivity differences associated with higher versus low wage i workers.78 The Preliminary Review of Congressional Budget Office (CBO) Cost Estimates of the Davis-Bacon Act also describes how the positive multiplier stimulus on local economies provided by spending generated by construction workers employed on federal contracts was not addressed by the CBO. Indeed, despite CBO's concern for the impact of Davis-Bacon on the federal budget, no effort was made to assess the negative tax revenue implications at both the federal and local levels of reductions in Davis-Bacon outlays.71 In the CBO's most recent congressional testimony, it made no attempt to cure the defects of its previous two studies and even conceded that "[hJigherwage rates do not necessarily increase costs.... [I]fthese differences in wages were offset by hiring more skilled and productive workers, no additional construction costs would result.i80 Congress has been persuaded for the last fifteen years that repeal of the Davis-Bacon Act would not result in a reduction of federal budgetary outlays for federal construction projects. As Rep..Hawkins has argued so well, the claim that the Davis-Bacon Act increases costs on federally-funded construction is based upon "imaginary savings." He has stated: This Imaginary savings Is something that is clearly an Illusion. No study has really considered all of the operations In ferms of cost overruns, of the under-handed practices engaged In.the construction Industry by fly by -19- night operators who move into your community where your local people, your local business people and your local wage earners have combined to establish a local wage level. Now is this a service to the taxpayers or is it a cost to that community and a cost to the Federal Government? I think that in talking about so-called savings we ignore the fact that this type of practice would prevail. So what is the opposition[to Davis- Bacon] about today? It is to help these fly-by-night operators operating on the basis of greed against responsible construction work.87 V1. CONCLUSION The Davis-Bacon Act is premised on the principle that all construction workers on federal projects.should be paid a fair and decent wage. The Cato Institute rejects that premise, at least so far as it applies to minority + construction workers. The Cato Institute wants to bring back vicious wage- cutting as the primary basis for competition for federal construction contracts, and apparently has no qualms about also bringing back widespread exploitation of and substandard conditions for workers on such projects. There can be little doubt that minorities would be among the chief victims of such exploitation. Thus, the only opportunity that the Cato Institute wants to open up for minority construction workers is the "opportunity'to be paid substandard and inequitable wages and benefits by unscrupulous contractors. As leading minority group representatives have indicated through their support of Davis- Bacon, minorities are not interested in pursuing that kind of "opportunity." And, regardless of what the ivory-tower ideologues of the Cato Institute obviously think of them, minority construction workers deserve better than that. They deserve fair and living wages, an adequate level of benefits, jobs that provide them with training and genuine opportunity for advancement; and jobs that accord them equal status with other construction workers-in other words, they deserve all of the things that the Davis-Bacon Act provides and makes possible. 1 -20- END NOTES 1. Arizona, Delaware, Kansas, Kentucky, Maryland,Massachusetts and New York had already passed some form of prevailing wage legislation. See Employment of Labor on Federal Construction Work. Hearings on H.R. 7995 and H.R. 8232 Before the Committee !� on Labor. 71st Cong., 2nd Sess. 14 [hereinafter Hearings ; cf. 125 Cong. Rec. 21155 (1979). 2. California, Florida, Maine, Massachusetts and New Hampshire all had enacted such legislation. Hearings. 14-15 (March 3, 1930). 4. 125 Cong. Rec. 21155 (1979). 5. As Representative O'Connor (D.N.Y.) noted, `there is not a progressive State or municipality in the Union that has ■ p ty not had identical legislation of this kind for ears. 74 9 y Cong. Rec. 6512 1930 . 9 ( ) 6. See Sutherland Stat. Const. § 48.15, 48.16 (4th•Ed). 7. Hours of Labor and Wa es on Public Works: Hearings on H.R. 17069 Before the c Q Committee on Labor. 69th Cong, 2d Sess. 2 (February 18, 1927). 8. Biographical Directory of the American Congress 1844-45 (1971). 9. Hearings on H.R. 17069, supra note 7, at 4. 10. 74 Cong. Rec. 6510 (1931) (Remarks of Rep. Bacon). 11. 74 Cong. Rec. 6505 (1930). 12. Id. 13. S. Rep. No. 1445, 71st Cong, 3d Sess., 2 (1931). Sen. Davis asserted that the bell was "fair and just to the employees, the contractors and the Government alike." Id. 14. Hearings on H.R. 16619 Before the House Committee on Labor 71st Cong. 3d Sess. 5,7 (January 31, 1931). 15. Hearings Before the Committee on Manufactures on S. 5904 71st Congress, 3d Sess. 10-13 (Feb. 3, 1931). 16. 74 Cong. Rec. 6510 (1930). 17. 74 Cong. Rec. 6510 (1930). -21- 18. See Nom7an Hill, Minorities, Women and tfie Davis-Bacon Act National Joumal. Sept. 19, 1981, at 4. 19. General Wage Determinations Issued Under the Davis-Bacon Act and Related Statutes. U.S. Department of Labor (Jan. 2, 1987). 20. See July 15, 1982 Affidavit of John Dunlop at 7 15 and December 17, 1990 Affidavit of John Dunlop at 4 9, Building and Construction Trades Department v. Martin 961 F.2d 269 (D.C. Cir. 1992)(No. 90-5345). 21. Dunlop (December 17, 1990), supra note 20. 22. Id. at 18 and 1 11. 23. The reports were submitted by all federal contractors with 50 or more employees, and all other contractors with 100 or more employees. 24. See U.S. Department of Labor, Employment Standards Administration, Office of Federal Contract Compliance Programs, 1991 Utilization for Construction SICS(computer printout dated May 28, 1993). 25. Id. 26. Dr. Thieblot's Prevailing Wage Legislation, (1986)was funded by the Public Service Research Council an organization which has ap peared ppeared in lawsuits challenging the Davis- Bacon Act, along with the non-union contractors' organization, the Associated Builders and Contractors. Other works by Dr. Thieblot were funded by the Associated Builders and Contractors Merit Shop Foundation,including The Little Davis Bacon Acts: Prevailing Wage Laws of the States (1976). Dr. Thieblot's The Davis-Bacon Act(1975) was funded by the Chamber of Commerce. 27. The data shows the following percentages of minority participation in apprenticeship programs for the years 1986-1989: % Minorities in % Minorities in Union Programs Non-Union Programs 1986 15.40 13.47 1987 14.10 12.92 1988 14.13 13.30 1989 16.48 15.36 See U.S. Department of Labor, Bureau of Apprenticeship and Training, Apprentices and Minority Apprentices in Construction Industry By Year of Registration and Types of -22- Sponsors (1993 .) The BAT data represents 60% of the national totals for apprentices, because it does not include the data for the 17 states and territories with Apprenticeship Councils. 28. According to the data, apprenticeship programs had the following rates of cancellations by minority participants in the period 1986-1989: Cancellation Rate Cancellation Rate of Minorities in of Minorities in Union Programs Non-Union Programs 1986 27 42 1987 28 49 1988 30 40 1989 36 47 See U.S. Department of Labor, Bureau of Apprenticeship and Training, Apprentices and Minority Apprentices in Construction Industry By Year of Registration and Noes of Sponsors (1993). 29. Dr. Clinton C. Bourdon & Dr. Raymond E. Levit, Union and Open-Shoo Construction 68 (1980). Bourden and Levit further stated: While this is in large part due to the relative absence of data on minorities in nonunion construction firms, this lack has not prevented some from concluding that 'the open show sector is both more hospitable as a whole to minority employment and being without craft restrictions and union rigidities, more capable of dealing with the problem.'. . . This conclusion is not supported by the only available data on minority participation in union and non-union apprenticeship programs . . . Tabulations from data show that minority participation is substantially higher in union programs. Id. 30. See e.cy.. Hill, supra note 18, at 5 rMany opponents of the Davis-Bacon Act attempt to put the issue in the context of union vs. non-union interests and argue that the law's only supporter is organized labor. This is clearly not true."). 31. Id. 32. See Ray Marshall, "America Still Needs Davis-Bacon', National Journal at 6 (Sept 19, 1981). 33. id. -23- 34. Id. 35. Id. 36. S. Rep. No. 96-259, 96th Cong., 1st Sess. 4 (1979). 37. Marshall, supra note 25, at 6. 38. S. Rep. No. 96-259, 96th Cong., 1st Sess. 4 (1979). 39. Marshall, supra note 25, at 8. 40. Hill, supra note 18, at 4. 41. Statement attached to testimony of the Building and Construction Trades Department, submitted to Subcomm. on Labor of the Senate Comm. on Labor and Human Resources (April 29, 1981). 42. Dunlop (December 17, 1990) at 18 and 1 11, supra note 20. 43. See Hill, supra note 18. 44. 134 Cong. Rec. H2818 (daily ed. May 3, 1988). 45. 134 C ..Con Rec. H2827 (daily arly ed. May 3, 7988). 46. 134 Cong. Rec. H2823 (daily ed. May 3, 1988). 47. Hill, supra note 18, at 5. 48. Smaller minority contractors have also been found to benefit from the Davis-Bacon Act. Smaller federal construction jobs, because of the equality of bidding opportunity provided by Davis-Bacon, serve as entry for small contractors into the construction industry. The smaller contractor may with Y compete P large contractors because of the control on wages. And, because of the greater concentration of minority contractors in the ranks of these smaller contractors, the "entry of minority contractors into the construction industry will be severely curtailed if the Davis-Bacon provisions are lifted from smaller Federal jobs ... . Therefore, retaining Davis-Bacon will not only continue its protection for workers, but will also ensure that smaller contractors, and particularlysmall, minority contractors, are able to successfully compete for federal construction contracts. S. Rep. No. 96.259, supra note 36, at 11. 49. GeneralAccounting Office, GAO/HRD-92-43,Apprenticeship Trainingat 11 (1992). -24- 50. Dunlop (July 15, 1982) at 5 14, supra note 20. 51. General Accounting Office, supra note 49;-at 11. 52. General Accounting Office, supra note 49, at 21. 53. Id. at 30. 54. Dunlop (July 15, 1982) at 7 15, supra note 20. 55. Dunlop (July 15, 1982), supra• note 20; Dunlop (December 17, 1990), supra note 20. 56. Dunlop (July 15, 1982) at 7 15, supra note 20. 57. Dunlop (December 17, 1990) at 9 6 and 17, supra note 20. 58. Building Trades v. Donovan. 553 F. Supp. 352 (D.D.C. 1982), aff'd in part and rev'd in part. 712 F.2d 611 (D.C. Cir. 1983), cent. denied. 464 U.S. 1069 (1984). 59. Id. i 60. See Dire Emergency Supplemental Appropriations Act of 1991, Pub.L.No. 102-27; Departments of Labor, Health and Human Services, Education and Related Agencies Appropriations for 1992; and Supplemental Appropriations, Transfers, and Recessions Act of 1992. 61. 137 Cong. Rec. H1531 (1991). 62. 137 Cong. Rec. S3548 (1991). 63. 137 Cong. Rec. S3628 (1991). 64. 134 Cong. Rec. H2829 (1988). 65. 134 Cong. Rec. H2823 (1988). 66. Union and Open Shop Construction. supra note 29, at 81. 67. Dunlop (July 15, 1982) at 1 11, supra note 20. 68. Id. at 4 12. 69. Id. at 1 12. l -25- 70. Id. at 1 11. 71. Id. at 11 12 and V 13. 72. See Allan B. Mandelstamm, "The Effect of Unions on Efficiency in the Residential Construction Industry. A Case Study", Industrial and Labor Relations Review 18. (1965) (presents a detailed comparison of union and nonunion home building in Michigan and concludes the greater productivity } largely offsets the higher wages paid to union workers); Steven G. Allen "Construction Workers are More Productive", *The Quarterly Jouma/ of Economics (May, 1984) (reports the result of a comprehensive econometric study indicating that unionized construction workers are between 29% and 51% more productive than their nonunion counterparts); Clinton C. Bourdon and Raymond E. Levitt, A Comparison of Wages and Construction Massachusetts Institute of Technology,Research Report No.R-78-3(describes various factors found to contribute to higher productivity among union workers). 73. Steven G. Allen and David Reich, A Case Study of Public School Construction Costs. Center To Protect Workers' Rights (1980). See U.S. Department of Housing and Urban Development Evaluation of the High Cost of Indian Housing at 23-2 7, (1979)("In order to assess how great an effect high wages have on project costs, [HUD] examined the Davis-Bacon wage rates applicable to each project and compared wage rates with dwelling construction costs ... A comparison of average wage rates with average dwelling construction costs shows no correlation between high wages and III . high construction costs. ) ) 74. Intemational Union of Operating Engineers, High Wage vs. Real Cost Analysis: Wage Rate vs. Productivity Report Update (1991). 75. HUD Office of Inspector General, HUD Audit Report On Monitoring and Enforcing Labor Standards at 13. 76. Id. 77. Congressional Budget Office, Modifying the Davis-Bacon Act: Implications for. the Labor Market and the Federal Budget (1983). 78. Workplace Economics, Inc., Preliminary Review of Congressional Budget Office Costs Estimates of the Davis-Bacon Act 7 (1986). -26- 79. Id. 80. Statement of Robert D. Reischauer, Director, Congressional Budget Office, submitted to Subcomm. on Labor Standards of the House Comm. on Education and Labor at 3 (May 4, 1993). 81. 134 Cong. Rec. H2827 (1988). Losing Ground: Lessons from the Repeal of Nine " Little Davis-Bacon" Acts Peter Philips, Garth Mangum Norm Waitzman, and Anne Yeagle Working Paper Economics Department University of Utah February 199-41 THE UNIVERSITY of UTAH Acknowledgements The authors wish to thank Hamid Azari-Rad. Randv Brown. Van Hemever. Matt Hotchkiss, Garti Rav, and Scott Smith — all students at the University of Utah — for help gathering information for this study. We also wish to thank the many government officials at the state and federal levels who helpfully provided data for our analysis. This study was originally funded in 1992 by a gift from Local 3 of the International Union of Operating Engineers. Subsequent funding came from the United Association of Plumbers and Pipe fitters of Utah and from the AFL-CIO. Some material in this volume originally appeared in Hamid Azari-Rad. Peter Philips. and Anne Yeagle. "The Effects of the Repeal of Utah's Prevailing Wage Law on the Labor Market in Construction," in Sheldon Friedman, et al., eds.. Restoring the Promise of American Labor Lai- (Ithaca: Cornell University ILR Press, 1994). 207-22. This is a working paper of the Economics Department of the University of Utah. Inquiries. comments, criticisms and suggestions should be directed to: Peter Philips Professor Economics Department University of Utah Salt Lake City, Utah 84112 TEL 801 585-6465 FAX 801 585-5649 INI'ERNET PHILIPS@ECON.SBS.UTAH.EDU The authors welcome your comments. This working paper will be periodically updated as our research progresses. Updated electronic versions of this working paper are available from the anonymous ftp site at the Economics Department of the University of Utah. Internet users may use ftp (file transfer protocoh commands to obtain an updated version. The File is in Wordperfect 5.1 with extended postscript attachments. It may be printed out in Wordperfect 5.1 or 5.2 on any postscript laser printer. To obtain an electronic copy do the following: (1) log on to an internet terminal; (2) at the prompt type ftp keynes.econ.utah.edu ; (3) at the request for user name type anonymous ; (4) at the password request type you e-mail internet address; (5) once you have entered the University of Utah Economics Department ftp site tN-PC cd tmp (then returril then type binary ; (6) then type is to obtain a fisting of the files in the ftp site. The file you are looking for is called DAVISBAC (7) Type get davisbac c:Ndavisbacmy where the second phrase c.\davisbac.my refers to the drive you want the file to go to and the name you wish the file to have. Copyright 1995. All rights reserved. Peter Philips Garth Mangum Norm W'aitzman Anne Yeagle s Contents Executive Summary, page in The Authors, v I. The History of Prevailing Wage Laws in the United States, 1 Passage of State Prevailing Wage Laws, 2 Passage of The Davis-Bacon Act, 3 Repeals of Some State Prevailing Wage Laws, 6 Efforts to Repeal Other Prevailing Wage Laws, 7 Efforts to Repeal Davis-Bacon, 8 II. The Economic Effects of Davis-Bacon Repeals, 11 Cutthroat Bidding, 11 A Loss of Earnings for All Construction Workers, 16 A Loss of State Tax Revenues, 17 Regression Analysis of the Decline of Construction Worker Earnings, 21 Increased Employment Associated with Lower Wages, 24 The Net Effect of Rcpeals on Government Budgets, 25 Summary, 31 M. The Effect of State Repeals of Prevailing Wage Laws on Training, Black Unemployment, and Minority Participation in Training, 33 The Effect of Repeal on Construction Unions and Wages, 34 The Relation between Repeals and Black Unemployment, 37 A Decline in Training, 40 Market Responses: Training, Turnover, and Careers, 42 National Trends in Registered Apprenticeship Training, 44 Summary, 55 IV. Construction Safety Put at Risk, 57 Why Prevailing Wage Law Repeals Lead to Increased Injury Rates, 58 A Comparison of Injury Rates, 60 The Cost of Injuries, 60 Summary, 63 V. Conclusion, 65 The Effects of Repeal of Prevailing Wage Laws. 65 The Goals of State Prevailing Wage Laws, 65 The Definition of a Prevailing Wage, 65 The Financial Costs of State Repeals, 66 Other Costs of State Repeals, 68 Estimated Effect of a Davis-Bacon Repeal, 73 End Notes, 76 RCiCM7:zs. 92 Figures 2.1 The mix of construction employment in Utah, by contractor type, before and after the repeal of the state's prevailing wage law, 12 2.2 Average cost overruns as a percentage of accepted bids on Utah road consnvcnon, before and after the repeal of the state's prevailing wage law, 14 2.3 The ratio of accepted bids and final cost to the Utah state enginecr's estimate of road construction project cost, before and after repeal of the state's prevailing wage law, 15 2.4 A comparison of annual construction earnings, by status of prevailing wage law, 18 2.3 A comparison of construction earnings in nine repeal states only, before and after repeals (ia 1991 dollars), 19 3.1 Union membership in construction in Utah, 1977-89, 35 3.2 Wages and employment in construction in Utah relative to wages, 36 3.3 The ratio of black to white unemployment in five repeal states, before and after repeals , 38 3.4 Black, white unemployment ratio for states that retained and never had state prevailing wage laws, 39 1961-91 41 3.5 Apprentice ce plumbersas n percentage of journeymen plumbers in Utah, 3.6 Turnover in Utah's construction industry compared with all employment statewide, 45 3.7 Apprenriceship training rates, by state groups, before and after repeals, 48 3.8 Apprenticeship training rates, by state groups, 49 3.9 Minorities as a percentage of all construction apprenrices by state groups, 53 3.10 Ratio of the percentage of minorities in construction as a ratio of the percentage of minorities in the state population, by state groups, 54 4.1 Injury rates in construction by status of prevailing wage law, 61 5.1 Annual income-tax revenue loss, construction cost savings, and resulting effect of repeal on Utah budget, 1987-93, in 1994 dollars, 68 a of minority 5.2 The percentagecea of minority apprentices in construction, divided by the pertag population in the state -the minority reflection percentage-for nine repeal states, 71 u construction cost, and effect of a repeal of the Davis-Bacon Ac t on income-tax revenues,red e$ 5.3 Esrima total budget (billions of dollars), 72 Tables 1.1 Prevailing wage laws, by state, 4 2.1 A simple estimate of Utah tax revenues lost in 1991 as a result of the 1981 state prevailing wage law repeal, 20 model of construction earnings decline, 22 2.2 A description of the data used in regression 2.3 A regression model estimate of the effects of state repeal on construction annual earnings,controlling . in earnings, and for secul ar and evclical trends in earnings, 23 for regional dr$creaca 8 2.4 Effects of wages on employment,controlling for state drfferences in employment, differences in the size of SIC groupings, the direct effects of repeals, and secular and cyclical treads, 26 2.5 Effects of construction earnings I' a employment for an average-size detailed construcaoa • n gs d« ra on standard industrial classification (4-digit SIC) of 3,540 workers per state, 27 2.6 The relation of hypothetical construction-cost srnnp to tax revenues, 29 2.7 Projected effects of a repeal of Davis-Bacon on the federal budget, 30 3.1 Linear regression model of turnover rate in consaueaou in Utah, 1956-91, 46 3.2 Training rates in repeal and never-had states as a percwtage of training rates in states that rcuumd their wage laws, 51 rates states �� rewssad rates in repeal and never-had states as s percentage of training 3.3 Training m g their wage laws, 1975-78 and 1987-90, 52 4.1 Regression model of the effect of state repeals on rnJury rates for plumber and pipe fitters. 62 Executive Summary Like the 1931 federal Davis-Bacon Act, legislation in 41 states has required that the "preVaiIing" w age be paid on state-government-funded construction projects. From 1979 to 1988, however, nine states have repealed their prevailing wage laws. (Nine states never had such a law.) The remaining 32 states din wages. These variations in state expenence provide useful information with have retaine p 8 which to consider probable effects of additional state repeals or the proposed repeal of Davis-Bacon. This study found that state repeals of prevailing wage taws had several effects. First, in Utah, whose expenence was examined most closely, the state budget has not benefited from repeal of the prevailing wage law. The repeal helped dnve down s. In the evade construction fctioore the9 8 earnings repeal the state has Iost substantial income tax and sales tax Utah, construction worker earnings averaged about 125 percent of average non-agricultural earnings. By 1993, construction worker earnings had fallen to 103 percent of the average earnings for Utah workers. This decline in eamings is because of lower wages, but also because of a subsequent shift to a less-skilled construction labor force. Second, also in Utah, the size of total cost overruns on stets road constriction has tripled the decade since repeal in compar'son to the previous decade. The shift to a less-skilled labor force — lowering labor productivity along with wages — and the greater frequency of cost overruns have costs asociated with the repeal. lessened any possible savings in public works constructs general dowsnward trend in real �onstrucnon Third, looking at all the states, and controlling g repeals have cost earnings, variations in state unemployment rates, and regional differences in wages, reps construction workers in the nine states at least S1,477 per vear in earnings, on average (in 1994 dollars). The costs may eventually be higher as the effects of the more recent repeals mature, dnving wages and training down further. Fourth, controlling for a general downward trend in the amount of construction training, variations in state unemployment rates, and regional differences in training availability, the nine state reoeals have reduced construction training in those states by 40 percent. Fifth, minority representation in construction training progrza.'ns has fallen even faster than have the trainingprograms m repeal states. Until the various state repeals, minorty anprcnticcship population. After repeal, minonnes participation mirrored the minority percentage of each slate's became significantly under-represented in construction apprenticeship programs. n rose by 15 percent where state prevailing wage laws Sixth, occupational injuries in constructio were repealed. Based on these findings, we conclude that, if the federal Davis-Bacon Act were repealed: • Federal income rat collections would fall by at least S1 billion per year in real terms every year for the foreseeable future. This sa is because conucaon wage levels would decline across all loytrient levels states and —based on the expencricc of the nine repeal states — con=cnoa ernp would not rise enough to offset this revenue loss. The figure for lost tax revenues may well be higher. If the expenence of the nine states that never had a prevailing wage law is indicative, lost tax revenues from a repeal of Davis-Bacon could rise to S2 billion per year. Whether the losses are S 1 billion or S2 billion, the government cannot count on making them up with its cast savings as a purchaser of construction. The government will not break even. • There would be 76,000 additional workplace injuries in construction annually, with 30.000 of them serious and thus requiring time off from work to recover. As a result, more than 675,000 work days would be lost each year in construction. This could lead to additional workers' compensation costs of about S3 billion per year, of which S300 million would be passed on to the federal government as increased costs on public works. • Utah's experience suggests that repeal of Davis-Bacon would generate a period of significant cost overruns and the increased use of expensive change orders. Although we cannot measure the exact costs of such practices, it is generally accepted that change orders add substantially to construction costs. i iv Davis-Bacon Repeal Effects The :authors Garth Mangum (PhD, Harvard University, 1960) is Max McGraw Professor of Economics and Management at the University of Utah. He is author or coauthor of 30 books and numerous monographs and articles on labor and employment, including Capital and Labor to American Copper, The Operating Engineers: Economic History of a Trade Union, and Labor Struggles in the Post Office and Union Resilience in Troubled Times: the Story of the Operating Engineers 1960-1993. In addition to teaching at several universities and serving as an arbitrator in more than 600 labor disputes, he has served the federal government as senior research analyst of the Presidential Railroad Commission, research director of the Senate subcommittee on employment and manpower, and executive secretary of the National Commission on Technology, Automation and Economic Progress. Peter Philips (PhD, Stanford University, 1980) is a Professor of Economics at the University of Utah. He is co-editor of Three Worlds of Labor Economics and coauthor of the forthcoming Building Retirement Security: the Central Pension Fund of the Operating Engineers. Philips has published widely on the canning and construction indusmes in journals such as Industrial and Labor Relation Review, Industrial Relations, Business History, the Journal of Economic History, and the Cambridge Journal of Economics. Philips has received awards for his teaching, including University of Utah Public Service Professor and Presidential Teaching Scholar. Norm Waitzman (PhD, American University, 1988) is an Assistant Professor of Economics at the University of Utah and the co-author of Business War on the Law: An Analysis of the Benefits of Federal Health and Safety Law Enforcement and The Incidence and Costs of Birth Defects: The value of Prevention. Waitznan's work on the economics of safety and health appears in the American Journal of Public Health, Demography, Inquiry, and the Economic Forum. He is writing a book, to be called Physician Supply and Public Policy. Anne Yeagle is a PhD candidate at the University of Utah who is writing her dissertation on the effects of the 1981 repeal of Utah's prevailing wage law. t could equire private or Although it was not clear at the time whetherg set annexamp e,rGompes believed. In emplovers to honor the eight-hour day, government could state after state, he pleaded for the eight-hour day for government workers and private sector workers employed on public works. Gompers also pleaded for workers to be paid the "current" daily wage so they could afford the reduced work time. Government was being asked to set a good example for the private sector, to show that a refreshed labor force ten or° btwelve hours.d produce in eTight he hours what a fatigued and bedraggled labor force turned out prevailing wage law in its infancy was an attempt to obtain shorter working hours for all labor. nment at The AFL paid attention to public works, however,tru should not trytry to savle money by eroding l levels was a major purchaser of conse lion. The AFL said government o the wages of its citizens. an end to convict labor. Many states employed With similar logic, the AFL called for convicts to pay for their keep. Convicts built roads on chain gangs, operated farms, made textiles, prices and the wages of and sewed garments. Convict-made goods were sold, forcing down working free citizens. Thus, prevailing wage law legislation, at its birth, was embedded in an overarching intent to shorten the grueling working day for all labor, to compel the working poor to make ends meet in some fashion other than by sending their children into the factories, to compel children into workers and better citizens, to compel employers to become be tter 0 schools so that they might adopt techniques that profited on the employment of skilled adult workers rather than unskilled child labor, to present government as an exemplar of good management to abolish the the.e ght- ce of hour day to government employment and on pub works or throw convict labor- government saving tax dollars by grinding n It is not surprising, then, that the first prevailing wage law passed in the United States — in Kansas — was part of an eight-hour-day law. Passage of Stat e Prevailing Wage Laws law in The Kansas Eight-Hour law. Kansas established the first prevailmgc aeration for its Sixth January 1890, the Kansas Bureau of Labor and Industry Statistics, to p p _ Annual Report, distributed a questionnaire to each trade union and s U '. hts of aionof Parsoasr Assembly. In response to a question about needed legislation, the Mold " of contracts for State work to unfair employers."" This against the letting Kansas, replied, a law one of the first reports leading up to the tracts fairly peas to be plea for the state to let out con Y � enactment of a prevailing wage law. In February 1891, the Second Annual Convention of the Kansas�S belFederation cluded e in Topeka, approved a bill concerning state-paid wages. That month, bill, by " ;�,�. Avery of prevailing wage section, called "for an Eight Hour LAw and was brought the Typographical Union No.121, Topeka The bill suited. officer, board, or commission, doing or performing any service Of That in no ease shall any provisions of the act be allowed to furnishing any supplies to the Stan of Kansm under the p performing such veers es od with him (or them) in redact the daily wages paid to emplo, g� cvvtded for :n the ac:. spice or furnish-.g such :applies, ea ac Dahl of the reaucaon of hours provided Qav:5-Bacon Repeal =;Ter-u I. The History of Prevailing Wage Laws in the united States In February 1891, Samuel Gompers, president of the American Federation of Labor, visited Topeka, Kansas, to speak on what the local newspaper called "the great topic of labor." Ten years earlier, the AFL — at its own creation — had laid out legislative aims that included the eight- hour work day, the elimination of child labor, free public schooling, compulsory schooling laws, the eliminarion of convict labor, and prevailing wages on public works. These proposals were based on a belief that the American labor market should consist of highly skilled workers earning decent wages, with time for family, and with children free to earn an education. In pursuit of these aims, Gompers' political strategy in Kansas allied him with the Republican Parry. On the morning of Gompers's arrival, the Alliance Party, known to history as the Populist Parry, withdrew an earlier invitation for him to speak in the hall of the state House of Representatives, which the parry controlled. Gompers, who represented 900,000 workers, had fallen out of favor with the populists, reportedly because of his belief that the trade unions should not form a political party with the Alliance.` The Republicans, who controlled the Kansas Senate, invited Gompers to speak there, and he did. Gompers was in Kansas to focus on the eight-hour day. Like other Americans, Kansans in 1891 typically worked six days per week, ten to twelve hours per day. In the older trades and crafts, such as carriage making and saddle making, where the work pace was slow and under the workers' direction, the long work day was tolerable. In the newer factories producing shoes, textiles, and the like; in the mines; and in the urban putting-out systems in needlework, six-day weeks and twelve-hour days were grueling. The AFL had made its prime objective a shortened work day and work week with as little cut in pay as possible. In his Topeka speech, Gompers declared: Our banner floats high to the breeze and on that banner float is inscribed, "Eight hours work, eight hours rest and eight hours for mental and moral improvement."' At that time, when there were no income supplement programs for the poor, low-income parents worked and had to send their children to work to make ends meet. This practice was later referred to by a North Carolina newspaper editor as "eanng the seed corn." Each generation of poor condemned its offspring to poverty because the children grew up as illiterate as their parents. The prevalence of cheap child labor, which accounted for 5 percent of the manufacturing labor force in 1890 and a larger proportion of service sector workers, kept wages down and forced adult workers to put in the long hours to make ends meet. Gompers wanted regulation to force employers and the poor to adopt a strategy, however painful in the short tun, of a high- wage, high-skilled growth path where children were to school and workers had the skills to Justify wages that would allow for a family life. Gompers said, The Federation endorses the total abolition of child labor under 14 years of age, an eight hour law for all laborers and mechanics employed by the government directly through contractors engaged on public work, and its rigid enforument; protection of life and limb of workmen employed in factories, shops and mines; ...the extcasron of suffrage as well as equal work for equal pay to women....The Federation favors mauttres, not parties.' Table 1.1 Prevailing wage Laws. by State States having Year States without prevaiLng wage taws prevailing wage laws passed Alaska 1931 Georgia Arkansas 1955 Iowa California 1931 Mississippi Connecticut 1935 North Carolina Delaware 1962 North Dakota District of Columbia 1931 South CArohna Hawaii 1955 South Dakota Illinois 1931 Vermont Indiana 1935 Virgmta Kentucky 1940 Maine 1933 Maryland 1945 Yea Year of Massachusetts 1914 States that repealed passed repeal Michigan 1%5 prevailing wage laws Minnesota 1973 Missouri 1957 Alab a= 1941 1980 Montana 1931 Arizona 1912 1984 Nebraska 1923 Colorado 1933 1985 Nevada 1937 Florida 1933 1979 New Jv sey 1913 1911 1985 New Mexico 1937 Idaho as 1891 1987 - New York 1394 1. a 1%8 1988 Ohio 1931 1941 1985 Oklahoma 1`909 U� Hamper 1933 1981 Oregon 1959 ?ennsy lvania 1961 Rhode Island 193J� Tennessee 1953 , Texas 1933 Washington 1945 West V'r 1933 tgtata W;,consm 1931 Wyoming 1967 is listed here, but not inchyied in the count of states. Note: The District of Cohimhta ore: The Dow-Bocow Source: State laws and corrected version of Armand J. Thwbiot. 3r.,Pmailneg wage Legulaa Act, Stan Little Davis-Bacon Acts,"The Walsh-Hooky Act aped T7w Sg* a Contract Act Philadelphia. The '-;'honor School, 1986, p.140. For four years before the 1931 passage of the Davis-Bacon Act, 14 bills were introduced to ing wages in construction. Robert L. Bacon in 1927 introduced the Congress to establish prevail R 17069. The member of Congresx first bill proposing a prevailing wage for construction, H. justified his measure as follows: Davis-Bacon Repeal c:Tmu 4 That in all cases such daily wages shall remain at the minimum rate which was in such cases paid and received prior to the passage of the act.' The eight-hour bill was one of four labor-related bills pending in the legislature. The weekly pay bill, the child-labor bill, and the bill to make the first :Monday in September a holiday, which would become known as Labor Day. In addition, that year the Kansas State Federation of Labor approved a resolution calling "for the abolition of convict labor when in competition with free labor."' The eight-hour bill, Senate Bill 151, failed in the Kansas senate March 6, 1891, with the prevailing wage section removed. But by March 10, when the prevailing wage section was put back in, the bill became law. This first prevailing wage law stated, That not less than the current rate of per diem wages in the locality when the work is performed shall be paid to laborers, workmen, mechanics and other persons so employed by or on behalf of the state of Kansas....' At first, however, the law was not enforced.' Not until 1900, did the Kansas Bureau of Labor and Industry Statistics report enforcement: "there were hundreds of complaints that were attended to by correspondence, and good results obtained.w . Prevailing wage laws in other states. New York was the second state to pass a prevailing ter 385 was amended in 1894 by Chapter 622 to eight-hour law (Chapter ) w w York's ei t ap wage la New ( W t n Kansas however, there include a prevailing wage law for those employed on publ ic works. As , were many violations.10 Laws similar to those in Kansas and New York were passed in Oklahoma (1909), Idaho (1911), Arizona (1912), New Jersey (1913), Massachusetts (1914), and Nebraska (1923) (see table 1.1). These laws established a precedent for the creation of the federal Davis- Bacon prevailing wage law. Passage of The Davis-Bacon Act Three federal laws primarily affect prevailing wages in the United States: the Davis-Bacon Act of 1931 which applies to construction, the Walsh-Healey Public Contracts Act of 1936 which covers employers in manufacturing and supply industries, and the Service Contract Act of 1965 (known as the O'Hara-McNamara Service Act), covering suppliers of personal and business services. These laws attempt to neutralize the effects of government purchases on wage determination in the private sector. The Davis-Bacon Act is the most significant of the three laws A prevailing wage is intended to prevent the federal government from affecting local wages and construction conditions; Davis-Bacon disallows the government from pushing down wages s been a ma jor purchaser of construcno� competitive bidding. The government has always a! P to come 8 P services. As a primary customer of construction services, the government holds the potential to use its bargaining power to force down wage rates. Davis-Bacon Repeal Effects The Government is engaged in building in my district a Veteran's Bureau hospital. Bids were asked for. Several New York contractors bid. and in their bids. of course. theN had to take Into consideration the high labor standards prevailing in the State of New York...The bid. however, was let to a firm from Alabama who had brought some thousand non-union laborers from Alabama into Long Island. N.Y.; into my district. They were herded onto this job, they were housed in shacks, they were paid a very low wage, and the work proceeded...It seemed to Inc that the federal Government should not engage in construction work in any state and undzrininc the labor conditions and the labor wages paid in that State...The least the federal Government can do is comply with the local standards of wages and labor prevailing in the locality where the building construction is to take placa." Hearings for a federal prevailing wage law began in 1927 and continued in 1928 and 1930, but no bill was passed. On March 3, 1931, Bacon's original proposal, which he had reintroduced as H.R. 16619, was signed into law by President Hoover." The Davis-Bacon Act required payment of prevailing wages on federally financed construction projects. The law essentially ruled out bidding on construction worker wages on federally financed construction. The original language was vague, however, and prevailing wages generally were not determined before the acceptance of bids. In 1935, President Roosevelt signed clarifying amendments to the act, which became the basis of the current Davis-Bacon Act. T"ne National Labor Relations Act of 1935 gave the Secretary of Labor authority to set the prevailing wage. In 1935, Roosevelt's Secretary of Labor, Francis Perkins, established the original rules for determining the Davis-Bacon prevailing rates. The prevailing wage was said to be the wage paid to the majority, if a majority existed; if not, the 30-percent rule was used. The 30-percent rule means if 30 percent of the workers in an area are paid the same rate, that rate becomes the prevailing rate there. The 30-percent rule often resulted in the union wage being the prevailing wage. If the 30-percent rule did not apply, because at least 30 percent of the workers in a given occupation in the local labor market did not receive the same wage rate, the average wage rate was paid to workers doing the same job. The prevailing wage was determined this way for 50 y ears. In 1985, President Reagan changed administration of Davis-Bacon, creating the 50-pe►teat rule. The revised regulation reduces the influence of the negotiated ration wage in most areas (see page 9, below). The Tenth Amendment to the Constitution restricts the ability of the federal government to dictate contract terms for the states. Thus, work funded entirely by state or local governments is not covered by Davis-Bacon. Each state, county, or city can establish its own prevailing wage — if it chooses to do so -- through legislation. In 1994, 29 percent of all county-level federal Davis-Bacon prevailing wage rates were taken from union contracts, 48 percent used average wages, and the remaining 23 percent of counties used a mix of union and average wages, depending on the occupation. 4 DaVr7-Bacon Repeal Effects Repeals of Some State Prevailing Wage Laws i Kansas had passed the first prevailing wage law in 1891 and, by 1969, 41 states and the District of Columbia had prevailing wage laws. Several cities also passed local prevailing wage laws affecting construction. However, state governments began experiencing fiscal crises in the late 1970s. In 1978, California voters passed Proposition 13, restricting state expenditures, and the Labor Law Reform Bill failed in Congress. In this political context, many state iegislarures believed that, to save tax dollars, government should use its bargaining power to lower construction costs, even if the probable effect of this action would be the lowering of construction wage rates and a possible effect might be the lowering of quality in the construction industry. More than 51 bills have been introduced in 23 state legislatures to repeal or curtail so-called little Davis-Bacon legislation." Alabama, Arizona, Colorado, Florida, Idaho, New Hampshire, Kansas, Louisiana, and Utah have repealed their prevailing wage laws. Florida. Florida, which passed its prevailing wage taw in 1933, was the first state to repeal. The statute was repealed over the veto of the governor in 1979." One of the most populous counties, Broward, established its own local prevailing wage law and several cities in Broward passed similar laws." Alabans& Alabama was the next state to repeal, in 1980.16 After Alabama's repeal, the entire South. from Virginia to Mississippi, except Tennessee, was without state prevailing wage law. Unsuccessful attempts were made in 1983 and 1984 to reinstate the 1968 Alabama laws. However, prevailing wage laws exist at the local Ievel, such as one in Mobile for city-sponsored construction." Utah. Utah's prevailing wage law had been passed in 1933. Eventually, prevailing rates were set by hearings held in three districts that were created for this purpose. In addition to covering construction, the Utah statute established prevailing rates for piece work. The first indications of intent to repeal the Utah law were heard from the local chapter of the national Association of Building Contractors (ABC) in 1978. (The ABC, nationally and in Utah 8h P sought to represent the interests of non-union contractors.) The Utah ABC outlined its strategy in a letter to other state ABC chapters in 1978: It is our hope that the major argument is favor of repeal would be based on tax savings and is unnecessary government spending, rather than a union verses non-union argument. The ABC lobbying effort became public during the Utah legislative session in 1979. The sponsor of the Utah repeal, Republican Representative S. Garth Jones wrote in the Deseret News The prevailing wage rate is substantially the union pay scale. In 1933 the law was designed to place money into a depressed economy,to increase wages to get the economy moving.The law does the same thing today. But today, the economy is not depressed; inflationlf is the fee lem and the cost of government is too high. Repeals&& the prevailing a8 w enterprise system to establish the wages of tradesmen at a substantial savings to the taxpayers. The prevailing wage law is inflationary. Additionally,the prevailing wage rate discourages non- union contractors from bidding public contracts. it encourages union contractors to bid public contracts. The effect is to force people looking for work to go to union contractors. The law is inconsistent with Utah's Right to Work Iaw.fFeb 23, 1979) Devts-Bacon Repeal EfTecu 6 The first bill to repeal the statute was introduced in 1979, only to be vetoed by Democratic Governor Scott Matheson. In 1981, repeal bills were introduced in 14�t�1att�0onYt9n Utah did repeal succeed that year and it succeeded only after a second veto from The bill was approved on almost straight parry lines — Republicans favoring repeal and Democrats opposed. The Salt Lake Ciry Tribune noted that only one Republican representative, who called himself a lifelong Republican and union member, voted against repeal and broke away from parry lines.'0 aid, "I'm convinced that repeal of this law �s When Matheson vetoed the bill in 1981, he s not in the best interests of working people in the trades whose skills are essential for a vigorous i21 Nonetheless, the Senate overrode the veto 21-7 and the repeal took effect construction industry. P 2 months later. Those in favor of the repeal maintained that the prevailing wage law was inflationary and pro-union. Republican C. McClain (Mac) Haddow sponsored the 1981 repeal bill. He said, "the reserved only as a tool to extend union control. The law is contrary to outmoded and is Y law is outmo P Utah's right-to-work philosophy...."Z` g predicted a 10 to 15 Roger Evershed, president of the Associanon of Buildin Contractors, percent savings on public works projects with repeal. g P z. -eight- Arizona. The next state to repeal was Arizona in 1984. Arizona's statute began as an hour work day in 1912 and, by 1930, became asprevailing wage law. In a court test, the statute nal in September 1979. In November 1984, voters repealed the statute �tuno . . was found unconsu P in a ballot initiative, Proposition 300. Provisions of the ballot initiative prevented communit;es from implementing local prevailing wage statutes. eight-hour law. The prevailing wage law was first enacted in 1911 as an Idaho's a8 Idaho I p 8 rattle was extensively amended until 1965: efforts to repeal it began in 1979. The leg!slature failed to override several vetoes but did repeal the law in 1985." At the same time, overtime pay requirements for more than eight hours of work were repealed.2 -9 q is ,or repeal Colorado. Colorado also repealed its prevailing wage law in 1985. Attempts P began in the late 1970s, but it was not until after the governor had vetoed the bill several times _ that the veto was ovemdden and the repeal passed. Nevertheless, since 1985 at least one its own prevailing wage rate for local construction.30 municipality, Pueblo, established New Hampshire New Hampshire joine d Colorado and Idaho in 1985 when it, too, repealed." Although legislators began in 1979 to try to repeal the prevailing wage law, they did not succeed until 1985. Influenced by reports of inflated costs on a school construcnon job, boat houses passed repeal without the signature of Governor John Sununu Kansas and Louisiana. Kansas, the first to have a state prevailing wage law, repealed it n 19g7." Louisiana followed in 1988 with repeal over the.initial veto of the governor." Efforts to Repeal Other Prevailing Wage Laws The Massachusens ballot initiative In Massachusetts. in 1988, thousands of union members, already active in the presidential election, worked with community groups to help defeat a ballot inidative that would have repealed Lhe rate's 1914 prevailing wage law. The effo^ to block y a: repeal in Massac`�use^S appears also :o have s'.-,wed efforts to repeal other mate pre trig w'a2e Dar3-Bacon Repeal Effect laws until the midterm elections of 1994. Question 2, the repeal initiative and the hottest issue on the ballot that year, was defeated 58 to 42 percent on November 8." The Massachusetts law requires contractors to pay employees on state-financed projects a predetermined wage. Prevailing wage rates are most often based on collective bargaining agreements, which vary by trade and geographical jurisdiction." In 1988, the Association of Building Contractors (ABC) and Citizens for Limited Taxation formed a coalition that spearheaded the repeal effort, with a signature drive run by the "Fair Wage Committee." In March, a report by the Massachusetts Foundation for Economic Research, The Peculiar Prevailing Wage Law, presented the public rationale for a repeal of the state law." The report stated that the many attempts to modify the prevailing wage law were defeated before reaching the governor's desk." Using confidential data collected from a construction contractor, the authors estimated that the prevailing wage law increased construction costs by 14 percent through higher wage costs. The report concluded that, in 1987, the prevailing wage law cost Massachusetts at least S212 million dollars.»39 In August, in response to the report by the Foundation for Economic Research, the Regional Information Croup of Data Resources Inc. presented a contrasting view. Data Resources said the earlier report had used insufficient data and oversimplified analyses.`0 Data Resources maintained that a repeal in 1990 would result in a "total wage loss of S 196 million and a net employment loss of 600." Data Resources concluded that although there would be nominal tax savings with a repeal, the overall impact would be to increase unemployment and lower living standards." By the end of a hard-fought campaign, community support included the Catholic Church; the Jewish Labor Committee; the Massachusetts Nurses Association; the National Women's Political Caucus; and the National Organization for Women.'Z A similar effort in 1994 to repeal by initiative failed on the Oregon ballot. The battleground has shifted back to state legislatures and the U.S. Congress. Efforts to Repeal Davis-Bacon The onset of state efforts to repeal prevailing wage laws coincided with U.S. Senate hearings in 1979 to repeal Davis-Bacon. During the first hearings, Davis-Bacon proponents defended the law with these points: 1. The act prevents the disruption of local wage and construction market conditions by the introduction of federally financed constructi on. 2. The act protects the prevailing living standards of construction workers by discouraging cutthroat competition by construction contractors. 3. The act provides equality of opportunity for contractors who are free to bid on the basis of skill, efficiency, and knowledge, rather thsn on their ability to slash labor standards. 4. The act helps maintain the high quality of the construction labor force and equal employment opportunity in the construction trades by encouraging use of bona fide training programs on federally funded construction." 8 Davis-Bacon Repeal EfTecu Advocates of repeal of Davis-Bacon said: 1. The act has inflated construction costs. 2. The act costs the federal government huge amounts of money. 3. The act is poorly administered. 4. The act is biased toward union contractors and hurts non-union contractors. 5. The act has caused wage inflation. 6. The act discriminates against minorities, because they are disproportionately represented among the low-skilled labor force. 7. The free-market system is suppressed. Although the Davis-Bacon Act was not repealed in 1979, the Reagan administration changed the way the law is administered a few years later. The administration in cammaltereed the 30 wage to percent rule. Until then, the Department of Labor used the modal — most modal wge to determine the prevailing wage for an occupation in a local labor or market, I IfeLhe modalawage the penny accounted for more than 30 percent of all wages group." e was declared the accounted for fewer than 30 percent of all wages, the mean (average) wage prevailing wage. The Reagan administration raised the threshold tto1C5mopercentt befodal wage dre a they ten modal co ld be declared the prevailing wage. union wages tend to an administrative change had the effect the mean or average wage for an occupation. So the Reagan of lowering the prevailing wage in areas where unions were weak. Some of the competing claims for and against Davis-Bacon can be to to well dose st the exper experience of the states — those that have repealed state prevailing wage laws, This study examines nee laws, and states that have never legislated a prevailing wage. of to have such revert the disruption the contentions of Davis-Bacon proponents that prevailing wage laws p local wage and construction labor markets and that prevailing wage laws protectliving n Davis standards and discourage cutthroat competition. This study examines, as well, the en Bacon opponents that the law costs government considerable lsuoms.o money and discriminates ors 1Farst against women and minority construction workers. The studyin what are the effects of prevailing wage laws s have on the safety and health f cotraing and human capital rnstn:c:ion constriction? Second, what effects do these la workers? Dais-Bacon Repeal Efrcc:s Davis-Bacon Repeal Effecu 10 II. The Economic Effects of Davis-Bacon Repeals Cutthroat Bidding As soon as the law was repealed, some of these non-union people {contractors" that had been doing small work around town suddenly just took off. and the union people (contractors] like ourselves, our market share decreased. - President, a union construction company, Salt Lake City, 1993 [Our] company has consisted of my father and my grandfather and me from about 1963. ['We are a double-breasted company.] Company A is a union [general] contractor that hires merit shop companies with no regard to union affiliation. Company B is a non-union merit shop company.... Our industry became very competitive during the mid-eighties, a lot of people are chasing the same type of work. — General contractor, double-breasted company, Salt Lake City, 1993 We've been in business for 51 years. Before that my great-grandfather ran a construction company and so we've always done construction. Right now we're doing mostly mechanical, and we do utilities, Mountain Fuel, water lines, sewer lines, AT&T jobs. We've built homes. We've built golf courses. We've built apartment buildings. In the last probably about eight vears [since the mid-1980s] there's a lot more small companies — little tiny, you know, dad and his three boys. We can't compete against them. We have too much overhead to do that and you get small start-up companies, they're willing to work for nothing for a while and you know they'll go out there for two years and just take these jobs dirt cheap. Sometimes they can`t finish. They'll go broke in the middle but still, we don't want to work for nothing. We'd just rather lock the gate and wait. — Office manager, union construction company, Salt Lake City, 1993 When Utah repealed its prevailing wage law in 1981, the structure of the construction industry changed dramatically. The most obvious effect was the decline of union membership and union contractors. But this was only the most obvious effect. Underlying the decline of union contractors was the rise of the little contractor and increasing turnover of contracnng firms in the business. The industrial organization of the industry changed, with an increased reliance on subcontractors. Comparing the 12 yews prior to repeal to 10 years after repeal, the share of total construction employment accounted for by the typically bigger and more capital-Lritensive general contractors and heavy and highway contractors fell, while the share of total employment accounted for by specialty subcontractors rose (fig. 2 1) With the entry into the market of more contractors and smaller contractors, compeunve pressure to win bids heated up. Urns pushed wages down An operating eag>n_er familiar with the bidding wars stimulated by Utah's prevailing wage law repeal tells how tt�.e bidding affected labor. Davis-Bacon. Rexal c:fecs i 60% 55% --------------------------------------------------- ... U) ...... i ---------------------------- C45% . .................................................... ... ... c 35% ---------------------------•-----•----•----------- Q ............................. -_- 0 3Q% 8 °�° _ ._.._._......._........._. ._ m a. 20% Before 70-8 1 After 62-91 151 Ile Spedatty General Heavy ' ent in Utah, by contractor type,before and after repea Figure 2.1 The mix of construction employm l of the state's prevailing wage law Source: Utah LM1 Snn acc l Rennrt- Table 5. When they repealed Utah's law. a lot of companies went out of business because of the cutthroat competition. .-k lot of companies just bought jobs so they could have a cash floµ to make pav'ments on their -equipment. The design engineers would tell the contractor that let's say the job was going to cost a million dollars. The contractor would still go to there anywa} and low-ball the bid. Then they would turn around to their workers and make their wages fit whatever they had to be to fit the low-ball bid. The general contractors did a lot of bid shopping after the prevailing wage law was repealed. The general contractor would get a bid from the subocon�ectoorbof a he wo ld•Ogo and bac then he the would low-ball the bid. Then, when the general g 1 subcontractor and say yeah I've got the job but you've got to cut your bid to 540,000 to have this job I've got and the sub would go back to the workers and say OK we've got this job but now I've got to cut your wages. See costs of materials and supplies and equipment were stable. The price of bricks and the asphalt didn't go down just because you got this job. So the workers had to make up the difference for all this tow-ball bidding. So basically the employer got their money off the backs es had to fall. of the worker. Whether it was to make money or just to break even, wag 4 — Operating engineer, Bountiful, Utah, 199 process But wages were not the only factor to feel the ow exposed f to practices oflowPballing Government purchasers of construction services were bids and over-muting costs. Average annual cost overruns for the Utah Department of laws repeal was 2 percent of initial accepted bid (Fig. 2.2). Since the Transportation prior to the la p �° This rise in overrun repeal, however, overrun costs have risen to 7.3 percent of the initial bid. costs has come despite the introduction of computers as a tool for contractors in preparing their bids. The cause of these increased overrun costs is the post-repeal tendency for contractors to take r the pressure of increased competition (fig. 2.3). W hen ' ks in the bidding process trade p ,, more risks g the state calls for bids on a project, the state engineer prepares an initial estimate of the projec, s cost. In the decode prior to the repeal of Utah's prevailing wage law, winning bids averaged 91 percent of the state engineer's estimate. After the repeal, winning bids have been, on average, 89 percent of the state engineer's estimate. Contractors are shaving their bids to win state contracts. These lower estimates have not proved to be a windfall for the state. Instead, after Utah's prevailing wage law repeal, final construction costs have been running at 95 percent of the state engineer's initial estimate. This amounts to 6 percentage points abovIe the accepted bids. Prior to Utah's repeal, final costs were running 93 percent of the engineer's estimate, only two points higher than initial accepted bid prices. er for This does not necessarily mean that the pre-repeal construction wase anndately�costs was the state, but it does mean that the relarionship between accepted bid pnC more certain and that contractors promised less before Utah's repeal, but delivered more relanve to the state engineer's cos estimates. ly created uncer-unry in Heightened competition after Utah's repeal has not onacross the board. A union plumber bidding plumber process, but has also lowered Utah consuuca wages describes this: Dads-oacon Re-Pell Effects ca 8.00% ............................................ 7.00% ----------------------------------------- ....... i 6.00% ---------------------------------- -------- ----- ------- c 5.00% a) ...._.... c, -------------- ------------------------------------- 4.00% a) --------------------.............-----------...... ......... ¢' 3.00% 2.o°io ..... ; c 2-00% -- ......_....... ------- 1.00 0.00'/° Before (197"1) After (1982-94) Santa:Lxw.Don °t Twu=rww,fA 6mr 0 b ems' 19s6 tt0 t t Figure 2,7 Average cost overruns as a percentage of accepted bids on Utah road construction,before and after repeal of the state prevailing wage law Cost overruns on the construction of Utah roads averaged 2 percent over accepted bids in the decade before Utah's repeal of its prevailing wage law. In the decade after repeal, average cost overruns rose to 7.3 percent over the accepted bid. Change orders associated with cost overruns are one of the more expensive components of construction costs. ....................................... ............ 1000% ...... ....................................................... . .. ........ ........................................................ .. .... a) > C-L40// 0 0 ....... ........ S 92% ---- ....... 95% ........ ------- ........ Cz 15 889/0 ....... ........ ....... ........ 84% 82% Before 70-81 /After 82-44 Sic:V(Eng's Es:) AcuaV(Eng's Est) Figure 2-3 The ratio of accepted bids and final cost to the Utah state engineer's estimate of road construction project cost, before and after repeal of the state's prevailing wage law -After the Utah repeal of its prevailing wage law, competition among contractors heated up and contractors shaved their bids to win contracts. In the decade before the state repeal, accepted bids averaged 89 percent of the state engineer's estimated project cost on road construction. After repeal, accepted bids fell, on average, to 89 percent of the state engineer's estimates. However, this cutthroat bidding did not cut final project costs as a percentage of the state engineer's estimates. In the decade after the repeal, because or a tripling of cost overruns, the final project costs averaged 95 percent of the state engineer's estimate. After Utah repealed its little Davis-Bacon law I was working on a job as a union plumber. The electricians on the job were non-union. At that time there was terrific pressure on wages and, as I remember, the IBEW [International Brotherhood of Elcctncal Workers] took a big wage cut — something like S3 — from S 16 to S 13. Anyway, the day after the union electricians took that cut, the contractor came on the job and told these non-union guys they would have to take a S3 cut too. There was a lot of animosity, around that but they took the cut anyway. They had to. Our union held off two years before we had to do the same thing the electricians did, and when we took our cut the non-union plumbers' wages fell right along with ours. — Union plumber, Salt lake City, 1994 Utah repealed its prevailing wage law just as the economy was falling into the 1982 recession. Thus, the effects of the repeal initially were tangled up with the effects of the recession. However, some of the nine states that have repealed their prevailing wage laws did so in good times and some in bad times. A comparison across states can somewhat disentangle effects of the business cycle from effects of a repeal. A Loss of Earnings for All Construction Workers Whatever a government might save in construction expenses from the repeal of a prevailing wage la w, the saving has to be balanced against the loss of other revenues. The lower wages paid on government-financed construction have a ripple effect, lowering wages throughout the local construction industry. Construction workers in states that have a law have a higher average annual income than construction workers in states that have repealed a laver, and those workers, in turn, earn more, on average, than do construction workers in states that have never had a prevailing wage law (fig. 2.4). That pattern may be explainable, however, for more than one reason. States that have different prevailing wage law policies may have higher or lower construction earnings for reasons unrelated to the wage law. For instance, repeal states might also be low-wage states in general. It may thus be more useful to isolate earnings data for repeal states only — before and after nine states that repealed the ir (fig. 2.5). Average annual construction-worker earnings to then eP prevailing wage laws from 1979 through 1988 show a drop of 51,835 from 524,317, or about 7 S percent in wages, adjusted for inflation and denominated in 1991 dollars, or 52,016 in 1994 dollars. The nine states are not heavily unionized and a fall of this magnitude cannot be b a fall of union es to the non-union level. � accounted for simply y � In recent years, the average construction unionization rate in the nine states that repealed " their state prevailing wage Iaws has been around 13 percent of the construction labor force. With this level of union coverage, for a fall in the union wage to account for all of the fall in the would have had to have been earning i average wage, at the outset of the repeal, union workers 60 percent more than non-union workers.` Union wage differentials typically are around 10 to 20 percent above non-union wages. Because union wages are not sufficiently high and uaioa coverage not sufficiently wide to account for all the fall to construction wages in these repeal states, we know that non-union workers have had to absorb some share of this average earnings decline. Davis-Bacon Repeal EfTecu 16 If one assumes that the union differential is 20 percent above the non-union wage and, after the repeal, the union wage falls to the non-union wage, both wage rates will have to fall e 'en n waze further to attain an overall 7.5 percent cut in earnings. Assumthe union wageng that -,he owould ha`euto fall to the non-union rate and then they would both fall together, _ by 21 percent and the non-union rate would have to fall by 5 percent to obtain an overall fall of 7.5 percent." In fact, only rarely does the union rate fall entirely to the non-union wage A reasonable assumption would be that the union rate prior above the non-union PateeGivenva 5 non-union rate and after the repeal fell to 10 P percent overall fall in earnings and a 13 percent union membership rate, union wages would have to fall 14 percent and non-union wages would have to fall 6.3 percenttoCobtain as much overall no l of 7.5 percent. In other words, while the union rate would have to f all union rate, the non-union sector of construction workers would have to absorb much of the e wage cut. The effects of state repeals of prevailing wage laws are isolated average percentage $ generate across-the- neither to union workers nor to government-financed construction." They S board cuts in the earnings of all construction workers. A Loss of State Tax Revenues The tax revenue losses that result from lower construction wage levels are surprisingly large. with income Whatever the source of this earnings decline among construction workers, states consa-action taxes have lost tax revenues as a result of this decline in taxable income among workers. And, because this lost income means lost purchasing power, states that have repealed their prevailing wage laws have also lost some sales tax revenues. On average, construction workers account for 5 to 6 percent of a state's labor force. In Utah to 1991, individuals earning aid a marginal income tax rate of about 7 percent. Taking the 31,528 520,000 to 530,000 p construction workers employed in Utah in 1991 and an average per capita decline in income of S1,3351 the total loss of annual income from the Utah construction industry in Utah in 1991 because Utah's 1981 repeal could be calculated as S58 .^. :!!ion (SI,835 ulnas 31,528). Given a on margin al tax rate of 7 percent, 1991 lost state income tax revenues might amount to S4 million e on (in 1991 dollars) (table 2.1). Assuming anmarginal es ropensity to tax rate of 625 per ent lost statesales tax from changes to income of 80 percent these two losses and revenues from this loss of income amount to 52.9 million in 1991.i9 Adding bringing them to 1995 values using the consumer price index yields an estimated loss of S8 2 million in state taxes in Utah in 1991 evaluated in 1995 dollars. T'ne figure of $8.2 million in lost tax revenues may be an overestimate for more reaworsons however. First, if wages fall and labor becomes cheaper, contractors mightrhire rs resulting from So we must consider possible increases in total income of construction es have possible increases in total construction employment 3. a ca fall in wag tion indusm Second, fine lower be-en falling to the United States generally, including the wouldat wages after state repeals may simply reflect a long-term cnontne in are sensitive o es �unemp oyment .aken place anyway. Third, annual earnings ►n constru Earnings rise when unemployment falls and fall when unemployment increases. Because �s�,s•�scon Rcreal F:`ecs S3 525.542 _...-----•------^ a82 ca .......................................... $25 � 52t,t25 o I a� S2 m ---•---- ........ ------ E 2 8 _c o ca H Q ..... --- W Q I .,� ..... ....... CD Have Law After Repeal Never Had Law States by Groups Figure 2.4 A comparison of annual construction earnings, by status of prevailing wage law Source: US DOL Employment and Earnings, 1975-91. -Figure 2.4 groups states into three categories (from left to right). The first bar, on the left, shows average annual income in 1991 dollars for construction workers in all states and pears where a state prevailing wage law was enforced. This includes repeal states prior to repeal. The second bar shows the average annual earnings of construction workers in repeal states after repeal. The third bar represents average annual earnings for construction workers throughout 1975 to 1991 in all states that never had a prevailing wage law. These data provide initial evidence that repealing or never having a prevailing wage law lowers construction income not only on public works but across the entire state construction industry. Jv 0-1 -1Z 24.3 i 1 1 f+ J O c� $2C/ O .0 .. -------------- in �15 — s .. ••---..._.. ...... v $1 G ai Before Repeal After Repeal sa m ox 4na m— States by Groups Csa tr r�rs i37sn a�� I Figure 2S A comparison of construction earnings in nine repeal states only,before and after repeals (in 1991 dollars) In the nine states that repealed their prevailing wage laws between 1979 and 1988, average annual income fell after the repeals (calculated in constant 1991 dollars). This fact does not control for other factors that might have been driving down ie evidence that the repeals forced lower earnings not wages, but it is prima fac just on public works but across the construction labor market. Table 2.1 A simple estimate of Utah tax revenues lost in 1991 as a result of the 1981 stele prevailing wage law rcpe2l Individual construction income prior to repeal (1991 doflc.*s) Individual construction income after repeal S==»82 (1991 dollars) Lost income due to repeal (1991 dollars) S1.835 1991 Utah construction employment 31.528 Total lost income in construction (1991 dollars) S57.353.880 S4.049.772 Lost Utah income tax S2.776.986 Lost Utah sales tax S6 4 2�7t9 Total lost tax revenues Total lost tax revenues in 1995 dollars The average annual construction earnings in 1991 dollars for nine repeat states in the years after 1975 and before each state's repeal was S24.317• In the years after each t 1991 the averse construction earnings fell to S22,482. Utah construction repeal u o g Pe P employment in 1991 was 31528 workers and multiplying these by an annual loss or income of S 1,835 yields a total lost income in Utah consmuction of S57.8 million. Based on Utah's income tax rate of slightly over 7 percent and a sales tax rate of slightly over 6 percent and a marginal propensity to consume taxable items of 80 � - percent, total lost state tax revenues were 56.8 million. In 1995 dollars, This is S8 2 million. unemployment varies by state and year, some of the difference in gamines might be because of variations in the unemployment rate (see figs. 2 4 and 2.5). Last, construction wages vary by region for reasons that are not directly due to the presence or absence of prevailing wage laws. These regional differences in earnings, unemployment, and long-term trends in wages can be accounted for by using linear regression analysis. Regression Analysis of the Decline of Construction Worker Earnings Using linear regression analysts, thi s section takes U.S. Department of Labor employment and rs in states for 1975-91 to re-estimate the constmcnon earnings data for construction worke earnings loss resulting from state repeals of prevailing wage laws. The analysis controls for long- term trends in wages, variations in unemployment, and variation in wages by region of the wage law, (2) country, and then focuses on the effect of (1) never having had a prevailing repealing a prevailing wage law, and (3) raising the threshold for implementing a state prevailing wage law to contracts worth S500,000 or more. U.S. Department of Labor employment and earnings data provide detailed information on annual construction earnings broken down by year, state, and type of construction contractor.90 For 1975-91, there are 27,778 separate observations. The inclusion in these data of information about prevailing-wage law status by state and year and t•ranslanon of all money values into 1991 dollars (using the consumer price index) allows us to test for (1) the effect that never having had a prevailing wage law has on per capita construction earnings, (?) the effect on individual ing a state prevailing wage law, and (3) the effect on individual earrings of earnings of repeal raising the threshold for applying a prevailing wage law. In this test, we control for regional differences in construction earnings, secular trends in ariations in earnings as a result of variations in unemployment, and esrnings,s' cyclical v differences in earnings by detailed contractor type.5' The data used for this test include average earnings across all states, years, and cons-5ucnon trades — S26,645 per year in 1991 dollars (table 2.2).`1 States that never had a prevailing wage law account for 15.6 percent of all the observations. States that repealed theiw account they 10.5 percent of all observations after repeal and 7 8 percent of all observations o repeale d their laws, for a combined total of 18.3 percent. States that had and retained their 6.1 percent of all prevailing wage laws between 1975 and 199fclahami for the with prevailing wage laws but the data set. Maryland and 0 observations in �7' with threshold levels of projects costing rates r period averaged account6for 4 per�olall observations. State-by-state unemployment The results of this regression model esnmanng the effects of state repeals on construction are statistically significant and the overall model has a goodness of fit of 73 percent- earnings in the data set are which means that 73 percent of the overall varisnon in annual earnings explained by the model. The results may be read as follows (see t le(This2.3). is a starting point Begin with a constant amount of annual esrnin SS of S33,005. calculated by the regression model and is typically called the "constant.") Then select a state and a year. In any state for any year we know the status of prevailing wage laws for construction. We use Utah as an example in column (3). Utah once had a prevailing wage law, but, by 1991, that lawhad been repealed. Furthermore, 1991 was 17 yes.*5 titter t^.e beginning of the data set and. DaN15-Bacon Repeal Effects Table 2.2 A description of the data used in regression model of construction earnings decline Observations 27•778 Average Earnings S26.645 Variables Percentage of data Percent of all observations by region South 29.9% Midwest 13.9% Atlantic 10.5% Mountain 1 OS% Corn Belt 103% Pacific 8.8% New England 81% Hawaii 1.1% Alaska 0.4% Other control variables Average state 6.76% unemployment rate Percentage of states with 43% threshold for applying state law of more than S500.000 Legal variable Percentage of all states 183% that are repeal states Percentage of all states 15.6% that never had state law Source: U.S. Bureau of Labor Statistics, Office of Earnings and Employment Statistics. Table 7 3 A regression model estimate of the effects of state repeal on construction annual earnings. controlling for regional differences in earnings and for secular and cyclical trends in earnings Regression Model Examples for 1991 I Variables and Coefficients (in 1991 Dollars) Utah I Mzrviand I Geer::. �) Starting Point: S33.005 S33.005 S33.005 5: .'X �I Regional Control Variables: Alaska S 15.628 Hawaii S7.982 Midwest S4.768 Pacific S4.638 Atlantic S4.617 S4.617 New England S 1.545 Corn Belt 51.010 Mountain -S79 -S79 i South -S2-360 S:_An Trend Control Variables: i Secular Trend -S225 -S3.829 -S3.829 -S:3:9 t Unemployment -S30.23 l -S 1.481 -S 1.78-s _S; Focus on Legal Variables: Never Had Law -S2.960 Repeal -S I.3 50 -S 13 50 Threshold S500.000 -S1.1 74 41.174 Predicted Income: S:6.266 S30.836 i thus, the time variable is set at 17 and Utah is in the mountain states region. Se, all other regional variables to Zero and multiply the mountain states control coefficient by I. Multiply the secular trend control variable by 17 because this is the seventeenth year of the data set. Liultiply the unemployment control by 4.9% because that was the unemployment rate in Utah in 1991. Set the "never had law" variable to zero because Utah did have a prevailing wage law up to 1981 and set the threshold variable to zero, because in 1991 Utah did not have a prevailing wage law (and even when it did, the threshold was below 5500,000). Now, set the repeal variable to 1 and multiply it times the repeal coefficient. Thus, the model now predicts Utah's 1991 construction income to be 526,266. That is 533,005 (the starting point) minus S79 (lower wages in the mountain states) minus 53,829 (secular down trend in real wages) minus 51,481 (associated with unemployment) minus S1,350 (because of Utah's prevailing-wage law repeal). The same exercise yields a predicted income of 530,836 for Maryland in 1991 and 522,345 for Georgia in 1991. Change the year and/or the state and the model predictions change. The RZ statistic of 73 percent indicates that the model fits the data well and that the predicted values are close to the actual s earnings in the various states for the various years. Controlling for all these variables, the model estimates that the effect of the repeal of the nine state prevailing wage laws was a negative S1,350 annual hit on construction earnings. Given average annual earnings of 526,645, this means a decline in earnings of 5.1 percent. This is a low estimate of a repeal's effect on earnings. The effect of a repeal may accumulate with time. The states that never had prevailing wage laws in construction have lower construction wages — after controlling for regional differences in wages and differences in unemployment rates. The model estimates that, in the nine repeal states, construction earnings are S2,960 less than in other states, controlling for other factors. This is an 11 percent reduction in construction eamings so asciated with never having had a prevailing wage law. The simple procedure in the previous section which compares construction earnings in repeal states before and after repeals estimates the repeal effect to have a 7.5 percent negarive effect on earnings. Thus, the range of estimated effects vanes from S.1 percent to 7.5 percent to an 11 percent decline to construction earnings associated with the repeal or absence of prevailing wage laws." Increased Employment Associated with Lower Wages As construction labor becomes cheaper, contractors may alter their crew mix to use more workers who are unskilled Have the nine state repeals of prevailing wage s laws easonalgenerated h��ends�n of employment? Construction employment vanes markedly with the economy. These employment swings can hide the effect of more jobs generated by falling wages. For instance, Utah repealed its prevailing wage law Just as the constriction economy was going into recession. On the surface, it looked like the repeal and wage cuts did not generce more construction employment. Multivariate linear regression analysis can control for these variations and pick out the potentially hidden effect of a repeal, controlling for other factors Table 2.4 presents the results of a generalized least-squares regression test of the hypothesis that, as construction earnings fall, all other things being equal, cons-auction employment will nse oyment The model controls for variatiolovtn unemployment, secular trends in ment associated with the repeal f a state l prevailing wage l •law and any nonwage effect on emp , Davis-Bacon Repeal Effects 24 The focus variable in the model is average annual earnings in construction and the hypothesis is that the relationship between earnings and employment should be negative. As earnings go down, employment might well go up. The regression model also includes control (dummy) variables ;or each state and each detailed industry classification (four-digit SIC, such as, plumbers and pipe fitters, SIC 1711). Thus, the model predicts construction employment in specific states, years, and each construction subclassification, such as plumbing and pipe fitting. In the data set for 1915-91, the average employment in a four-digit subclassification is 3,540 construction workers. The unemployment rate, not surprisingly, negatively affects construction employment and there is a small but statistically significant upward trend in employment. The effect of prevailing wage rate repeals on employment is negative, but this variable is not statistically significant which means the true direct effect of repeals on employment is zero. working throw lower earnings t of state repeals on employment or g through _ However, the indirect effect P , is not zero. The effect of earnings on employment is as theoretically expected. As earnings fail, employment increases and this estimated effect is statistically significant. From this relationship, we can estimate the indirect effect of state prevailing wage laws on employment through the repeals' effects on earnings. Possible employment effects may be calculated for various levels of earnings decline. In table 2.5 column (1) presents hypothetical earnings declines and, in column (2), the results from table 2.4 are used to calculate a predicted increase in the construction industry when it is analyzed at the detail of 4-digit SIC codes (such as plumbers and pipe fitters, SIC 1711). As average annual construction earnings fall from a loss of 5500 to a loss of S3,000, employment in given SIC industry groups rises from 24 new workers to 118 new workers. Given an average employment size of a 4-digit-SIC industry group of 3,540, these hypothetical increases in employment translated in percentage terms to an increase of from 0.7 percent when earnings fall by S500 to an employment increase of 4.0 percent when earnings in construction fall by 53,000 The Net Effect of Repeals on Government Budgets 7he overall effect of state repeals of prevailing wage laws on state expenditures in construction and state tax revenues will depend on the amounts of government cost savings from such a repeal and lost tax revenues from a repeal. Government construction cost savings will depend on three questions: how much lower are wage costs after a repeal, how much lower is worker producnviry at lower wages, and how much construction work does the government purchase? Lost tax revenues will depend on (1) the marginal income tax rate for construction workers earning 520,000 to 540,000 per year, (2) the sales tax rate, (3) the marginal propensity to consume taxable commodities for construction workers earning S20,000 to 540,000 per year, (4) lot, per- capita construction income associated with a repeal, and (5) gained construction employment associated with a repeal. (The S20,000 to S40,000 range encompasses most construction worker; ) Previous estimates of construction cast swings associated with a hypothetical repeal of t:e federal Davis-Baran Act range from 1 to 11 percent." The Congressional Budget Of ice favors an estimate of a 1.5 percent cost savings associated with the wage effect plus a 0.2 percent cost savings because of paperwork associated with Davis-Bacon." The savings may be higher or lower. Dav:s-Bacon Rexal Effects -5 Table 2.4 EfTects of wages on employment, controlling for state differences in etnplovMent. differences in the size of SIC groupings, the direct effects of repeals,and secular and cyclical treads Starting point: 5-500 workers in each SIC group in state Unemployment rate Subtract 211 workers for each percentage- point rise in unemployment Secular trend Add 31 worker increase per year Repeal Subtract 170 workers (no( statistically significant) Average Earnings Subtract 47.1 workers per SI.000 increase in earnings Number of Observations 27.778 Avg. Employment in SIC Group 3540 by State and Year Controlling for state differences in construction employment, differences in the size of four-digit SIC groups (such as plumbing versus electrical), secular trends, and cyclical variations in emplovment in each state—and the direct effect of repeals on employment — a fall in earnings resulting from a fall in wages raises employment in construction. For an average-size SIC group of 3,540 workers, a fall in earnings of S 1,000 per worker would increase employment by 47 workers—or 13 percent. Results for controls for each of the states and SIC ¢roues are omitted from the table. Note P 8 di: An example of a four-digit SIC (Standard Industrial Classification) group is plumbers and pipe fitters, SIC 1711. Table 2 c Effects of construction earnings decline on employment for an average-size detailed construction standard industrial classification (4-digit SIC) of 3,540 workers per state Various Predicted Rise in Pe.-ceniaze Rise Hy rt pothencal E:nplovenc Because of 1n E:.^.play^e^t Earnings A Fall in Annual Because of a Fall Deciines Const-ucnon Earnings in Ea.:.:^gs -F (1) ;2; I (:) -S500 24 -S1.000 47 '=�' -S1S00 71 2.0% -S2.000 94 2.7 40 S2.500 118 3-� -S3.000 141 4.0 04 As repeals force a fall in construction wages and earnings, construction employment rises. The model in table 2.4 indicates that a S500 fall in earnings results in a 0.7 percent rise in employment. An average annual S3,000 drop in earnings would result in a 4 percent rise in employment. This is an "inelastic" demand for labor — the percentage that earnings declines is substantially higher than the resulting percentage rise in employment (for the group). This means that even though employment rises when wages fall, the rise in employment is relatively small compared to the fall in wages. Consequently, overall income to construction workers declines after state repeals. The effect in Utah. In this section, we will simpiv accept all ranges of hypothetical or estimated savings rates from 1 to 11 percent in order to examine our model of lost tax revenues as it applies to Utah (see table 2.6). Rows 1 through 10 of table 2.6 provide half of the information needed to calculate the net effect on Utah's budget balances associated with the repeal of Utah's prevailing wage law in construction. Row 2 shows the level of employment in construction in Utah for 1987 to 1993. Taking from our regression model the value of lost income associated with a repeal of a state prevailing wage law (41,350) and translating that into 1994 dollars, using the consumer price index (41,477), we multiply this lost income times the level of construction employment in Utah for each year. This lost income associated with a repeal, denominated in 1994 dollars, is shown in row 3. Row 4 shows the gained amount of employment associated with a fall in construction wages and earnings because of a repeal. Row 5 shows average construction worker income in each year (in 1994 dollars). Row 6 shows the gained income due to additional workers shown in row 4 multiplied by average construction worker income in row 5. Row 7 reports the difference between GROSS lost income due to lower earnings and gained income due to lower wages. This net lost income is the source of the lost income tax revenues reported in row 8. Utah's income tax rate is flat at 7 percent of income. Utah's sales tax rate is 6.25 percent. For construction workers, it is conservative to assume an 80 percent marginal propensity to consume locally on items subject to sales tax. This means that as a construction worker's income rises by $1,000, that worker will spend $800 on local commodities subject to state sales taxes. This allows for 20 percent of additional income to go to savings or purchases not subject to sales taxes. (Food purchases are subject to sales taxes in Utah.) Row 9 reports lost sales tax revenues as a result of net lost income reported in row 7. Row I0 combines lost income and sales tax revenues. Rows 12 and 13 report in 1994 dollars the value of building and road construction in Utah not covered by the federal Davis-Bacon Act. Roughly 20 percent of road work in Utah is not covered by the federal prevailing wage law. Rows 16 through 21 calculate, again in 1994 dollars, hypothetical levels of construction cost savings associated with Utah's repeal of its prevailing wage law. These hypothetical savings range from 1 to 11 percent of total construction costs. Rows 23 to 28 subtract lost tax revenues from construction cost savings for the various hypothetical levels of cost savings. Rows 23 to 28 show that in Utah, at total construction cost savings of below 3 percent, the repeal of the state's prevailing wage law tended to increase state finance deficits. The loss in tax revenues associated with lost construction worker eamngs exceeded likely gains in construction cost savings. At and above 5 percent in total construction cost savings, the repeal helped up the balance of state finances into the surplus. Using the Congressional Budget Office's estimate of a 1.5 percent increase in construction cost savings plus 0.2 percent in paperwork, the state of Utah would have lost more in tax revenues than it guned in construction cost savings every year since it repealed its prevailing wage law in 1981. The likely of t%l of a Davis-Bacon repeal on federal budgets. For construction workers earning $20,000 to 540,000, federal marginal income tax rates range from 16 to 28 percent. There are no widely significant federal sales taxes. With these changes in mind, and using federal data for construction employment, we can use the above model to estimate the tax revenue effects of a repeal of Davis-Bacon (table 2.7). 28 Davis-Bacon Repeal Effect, Table 2.6 The relation of hypothetical construction-cost savings to tax revenues ,98T 1999 1 !�1 197: x' , Yur 3490: a%) :5E6d .7336 315:3 ioytncnt :567b I ;538.:03.'3U iia!,l;C.:'.5) fTafi.,c 613:31 3 L,.0 two— iS39.J9'.O+a) rS36.193.'a 6) i 6:3 aN a63 49S 56a I a Own d a'8 i Empia m--C rc.gaC S:S.:t) S:5.166 S:3.973 S J •+' 5 (non o » S:5.:06 S:5.3:9 b Gar+adG+eams SI_SS3.a5] Slt."3.130 St-a]1.]+9 S11.56_5S0 S14.C"_a0E SI s.95:.3:' is S:'%981 + (S:3.Sa7.696) (13:.360.a:I) iS36.593.46S) set L=L lt- rt- (S:6.M.591) (Z"3.2:0.366} (S26,IT JS.) (Sl,g9t,33d1 (I_361.Sa3) :-359 130, d �4taenrTaa (S(.tiL1S1) (SL75i.af0) (51,133.465} :St.B':9.6'3) 9 :.ou Sias ias• (Sl.}+A.IEO) (11.:36.17SS) (11.309.619) tS1.a:7,3Fa) (51,61d.0:1) 3)+-:.3E•^.. 10 law(.at'age •S3. 5.0)1) S3 91 a a6d) Q].tal OEa Sl Ss.)91 a'S (T).EE].:SOt a 63 i I I V41W asuu�F.mmed:aasmrLmn 193.J.E06 I72.661.036 Si7,51E.355 St0t.3:5.01I SI:S.'90.3'1 t, �� S9+.a36,6:p S7t0d9.603 13 PUNda r:•:: 7.077 $9.974.176 SI7.123.065 $11.9'r0.161 n+••6"•"C S1! 5110.90E.871 590.631.:17 S113.196.075 1•=66Zt53 S::' 14 Tots! SS53 6T S8'917.779 15 yvveuru NJ Sa+zr+o e+cm%= vc^G"" 1901.31: S1.:11.960 St.;r.6.6� 3t3ai3i 16 t% St.:$533T Si79,lli SI.109.Ci9 ��1 aV 3-7tL9'36 S3.455.01 13.679.963 i S7!66.611 52677,a13 $]�'.�56 La630.'36 I �aa wil- I 15.+59AM S6.133.109 SK SS.7'17,6is 1a,95.6E9 15345. ' 18 S8.5",351 7% $E.00.'59 T6.1S1.96a T7 76J.6Zt f6.3�+.lis Si.o63.7= i 19 $11.099S9a S10J99.E33 :0 57.91Z:a0 S9.9E1,79E =.I56.d09 9% 510361,6i3 30969 a3.a - TI.57I $I l a4:S37 3 `JS .56a •i 11% 90' 59.6'•C.316 S:'199.976 it ! tint Gan is Lai n Tu�ie.muas 'I SJ.;64.595) (S^_O7]99a) (S:.S19.a10) (S`'3I�0)1% (�C70.a9a) (Li75.J30) oa.-:e. :a 3% S:a0.Si0 (13T7.055) Slta,lia (5704716) �I { Sl.l76.SS1 a 3.SS j 5% SZ551.65a SI_UI=1 S':.a0"J61 SLIM,I39 TI? I.t4f :5 M S1,i6Z'� S3.139. .6ro.339 W1 t.a63 S4,150.aT: Sa.193.I l� a97 Sd 1 T4.-31.Or S6,aia.393 S6.bai 371 W".:. 9% S7.173.9m Sa.i91'T: S6J3i.716 •8 11% T9 44a.976 56.656.OaE 59.056-2" f6 Sa l Table 2.7 Projected effects of a repeal of Davis-Bacon on the federal hudget 1 Employment 6.000,000 2 Lost Income S8,862.000,000 (Employment•$1,477) II 3 Gained Employment 107,400 I (Eceployment•1.0179) 4 Avg. Income in 1994 S27,373 II S Gained Income from New S2,939.829,040 I ) Employment 6 Net Lost Income SS,922,170,960 7 Lost Income Tax at Various Marginal Income Tax Rates 8 16% Marginal Rate $947,547.354 9 20% Marginal Rate S1,184,434,192 10 28% Marginal Rate S1.M,207,869 11 Value of Federal S11,328,571,429 Construction 12 Hypothetical Savings in Construction 13 1 % S11SZ15,7I4 14 3% S34S.8S7.143 1S S% S576,428,571 16 7% S807,000.000 17 9% S1.037,S71,429 18 11% S 1,268,142.957 19 Net Gain (Loss) in Budget 16% Margistial Rate 20% Marginal Rate 2896 Mar=iaal Rate 20 1 (S832Z61,639) (S1,469,148,478) (S1,542,922.154) 21 a 3% (S601,690.211) (5838,577,049) (S1,31'2.3S4,726) i 22 ' S�e (5371.118,7g2) I (S608,445,621) (S1,081,779^97) i 23 - 7 ($140.541JS4) (S377,434,192) (S851.207469) Y r 24 9 594.424.475 (S146,S62,763) (S620.636,444) 2S 11% 5320.595.544 S83.708,66S (S390,065.412) With as employment level of 6 million construction workers and an average annual earvic; or S27,000, the lost income from lower wages exceeds the gained income from iac-c2sed empioyment. This results in differing values of last income tax revenues depending on tbt assumed marginal tax rate. with a value for federal construction of 511.5 billion, the hypothetical savings on construction from a repeal depends on the assumed cost-savings race_ At a marginal income tax rate of 16 percent, net budgetary savings from a repeal occur only with construction cost savings rates above 5 percent. At a 20 percent marginal tax rate. with Construction cost savings rates acj�e net budgetary savings from a repeal occur only h g 9Fercent. At a 28 percent marginal tax rate, net budgetary savings from a repeal at-er occur within the range of cost savings between 1 and 11 percent. In short, a repeal of the Davis-Bacon Act will hurt the federal budget deficit. n workers in the united States." Table 2.7, rc;•w There are approximat..ely 6 million �onstructio. 2 shows what would have been the loss in income that these consruction workers would uav_ experienced given the 1994 value (-SI,477) of our regression estimate of the effect or s. := reoeals on cons action income. Row 3 presents an estivate of increased national coast-uc:.ea employment associated wi_' lower wages. Row 4 presents average annual income for consrvc=cn workers in 1994. Row 5 multiplies gained employment in row 3 times average income LZ 70 4 to obtain the increase in total can t. cnon workers' income associated with a h;rpotnen;.zl .repeal of the Davis-Bacon Act. Row 6 subtracts gained workers' income ::om new empi0',= t: fron lot income as a result of lower wages to yield net ins worker income restzlring = a hypothetical repeal. Rows 8 through 10 present lost income tax revenues due to net lost in©�.e at three marginal tax rates of 16, 20 and 28 percent. In fiscal year 1990-91, the test_ government spent 510.491. billion on construction.'o Row I 1 presents this stun in 1994 dnULn t on costs associat-,; .., i con .,uc-i o hypothetical ca.i savings e s i n Rows 13 through I8 pr..s....t levels of hypo ., a repeal of Davis-Bacon. Recall that the Congressional Bucget OF.,ce estiMates total tze sat-WCs to be 1.7 percent, but others have presented savings esi.mates be*we-cm 0.5 percent percent. Rows 20 through 25 present the net effect on the federal budget of hy;)=e,=;, construction cos projected savings at various rates minus tax revenue losses at various mar L tax rates. Rows 20 through 25 show that only at very low marginal tax rates and very a__: construction cost savings rates does the federal budget bene^t from a repeal of Davis-Bar-oa. '•. a marginal tax rate of 20 percent and a consructon cast savings rate of 3 percent, the leer budget Ioses 5838 million annually in 1994 dollars based on the i991 level of federal govern=. exuenditures on constru^on. Summary w led ;o an overheated 'bidding process :.=:- in Utah, the repeal of the sate prevailing wage la e,. added u.Zce air. v to the cost of state cons— ion. L- :he decade before the repeal, cos ems_ on state-financed road cons-,ruction averaged percent o' accepted bids. in the decade repeal, average roac consruction cost ove:,urs rose :o % ?er:_r t of theb'd ' inspec:ion cf the s^o"wed that, hies at a io-- Davis-3accr. Rcpcs'. :::ccs percentage of the state engineer's estimate of project costs but that, after change orders, the projects ended up costing the state a higher percentage of the state engineers project cos; estimate than in the decade prior to repeal. After the Utah repeal, contractors shaved their bids to get state jobs and more than made up for low-ball bids with subsequent change orders. This caused the increased cost overruns. An econometric analysis controlling for variations in regional differences in construction earnings, variations in unemployment rates, and general trends in real earnings showed that the nine state repeals' effects on earnings was a loss of S 1,477 in 1994 dollars. Econometric modeling also showed that construction employment rose in repeal states after repeal by about 1.7 percent. This employment increase appeared controlling for variations in unemployment and long-term trends in construction employment growth. Thus, in assessing the budget effect of repeals of prevailing wage laws, we are able to do two things. First, balancing the overall loss of construction worker income resulting from lower average earnings against the overall gain in construction worker income resulting from higher construction employment, we are able to estimate the change in overall construction worker income and consequently the change in government tax revenues resulting from these repeals. Second, taking a very wide range of hypothetical construction cost savings, we are able to estimate the net gain or loss to government budgets associated with repeals. In Utah, given its structure of income and sales taxes, the state budget benefits from its repeal of the prevailing wage law at construction cost savings at and above 3 percent. At the Congressional Budget Office estimate of a 1.7 percent construction cost savings (including paperwork costs), the state of Utah's budget has annually lost money as a result of the repeal every year since the repeal At the federal level, construction cost savings must be substantially higher to generate any budget benefit from a repeal of the Davis-Bacon Act because of the federal income tax structure At the more conservative estimate of 3 percent construction cost savings with a 20 percent marginal tax rate and the 1991 level of federal construction spending (in 1994 dollars), the federal government would lose SS38 million per year by repealing the Davis-Bacon Act. The jusnfication often eiven for repealing the Davis-Bacon Act is that a repeal would help cut the federal deficit. Tha: is incorrect. A repeal of Davis-Bacon would help raise the federal budget deficit. This is because the purpose and effect of a repeal is to [owe* the cost of wages on federally funded construction projects. But lower wages and earnings will not be isolated to federally financed public works. Earnings would decline across the entire construction labor market and the govemment would lose more in income tax revenues than it will gain n - construction cost savings. ;? Davis-Bacon Repeal Eftec. III. The Effect of State Repeals of Prevailing ;age Laws on Training and Minority Participation in Training This chapter presents a case study of the effects of the repeal in 1981 of Utah's prevailing wage law on unionization, construction earnings, and training. The Utah repeal accelerated the decline in the union share of the state's construction labor market, drove down average construction wages in the state, and decreased union apprenticeship training for construction. No public or pnvate source has offset the decline in training. In response to the decline in union membership and training, contractors have reduced turnover in order to retain skilled workers and to minimize screening and training costs. In response not only to the decline in construction wages but also to the coincident decline in health and pension benefits, however, experienced construcnon workers are leaving their trades for careers in other industries. Thus, while construction firm th turnover is on the decline, tur nover in the industry is on e This chapter examines also whether the Utah experience i rise.b'n training can be generalized to the eight other states that have repealed their prevailing wage laws in construction. The U.S. or Bureau of A renticeship Training keeps state-by-state records on registered Department of Lab PP union and non-union apprenticeship programs in construction. These records suggest that what happened in Utah is typical of what has happened in other states after repeal of their prevailing wage laws. The ratio of apprentices to journeymen in construction is higher in states that retain their prevailing wage laws compared with states that never had such a law. The rate of apprenticeship training in states that repealed their prevailing wage laws was substantially higher before the repeal compared with after the repeal. This remains true even when one controls for regional differences in training rates, the effect of unemployment, and long-term trends in training. workers in Utah in construction, but nationally there are There are not many minority o (":Minority" here refers to nonwhites, male and female.) Some have argued that prevailing wage will open job o ortunines for unskilled minority workers and lower the law repeals PP p P 1 unemployment rate of minorities, relative to whites. However, there is no evidence to support this claim. Black-white unemployment ratios rose in repeal states after repeals. Black-white prevailing wage laws have never had g � unemptoyment ratios tend to be slightly higher in states that ha P probably compared to states that have retained their laws. While repealing prevailing wage laws p y has not caused black-white unemployment up. There is no evidence to suggest that p P Ymeat ratios to go a repeal of the Davis-Bacon Act would cause black-white unemployment ratios to decline. The repeal of prevailing wage laws has especially hurt the training of minorities. There are proportionately more minorities trained as construction apprentices in states that retain their wage comp e laws aced with states that have never had such laws. In repeal states• the prevailing w P proportion of minorities trained in construction apprenticeship programs declines substarina11Y the repeals. This remains true after controlling for regional differences in relative training after p e are independent t o f stet rates, unemployment, and long-term trends in rrunoriry training which P repeals of prevailing wage laws. The decline in minority participation in construction apprenticeships after repeal is tied i� a decline in unionization. Union apprenticeship programs tend to be large. Apprenticeship or in order to broaden the experiences coordinators move apprentices from contractor to contract of the apprentice. Tvpic Ily, because nen-•anion apprenticeship programs tie True apprentice to one contractor, tie non-union programs tend :o be small, s:ngle-'_1r7n prcgrarns, as opposed to larger. Davis-Bacoa Repeal ::Tectj 33 joint programs. At the same time, affirmative action regulation of apprenticeship programs applies only to programs having five or more apprentices. With the repeal of prevailing wage laws, not only does formal apprenticeship training decline, but also remaining apprentices are found more often in smaller apprenticeship programs. Thus, one effect of state repeals of prevailing wage laws has been to move more apprenticeship training out from under the oversight of affirmative action regulation. The result has been a substantial decline in minonty participation in the remaining apprenticeship training. The Effect of Repeal on Construction Unions and Wages When Utah repealed its prevailing wage law in construction, wages became a focus of competition between contractors bidding on state jobs. Many union contractor went non-union or double-breasted (with union and non-union subsidiaries) to match or beat the lower wages of non-union contractors, and other union contractors lost market share. Because construction employment was falling, many union members went non-union with their traditional employers to stay employed. The vice president of a large industrial and commercial general contracting firm in Utah noted that, after the repeal, There were a lot of union workers that carried their card in their shoe. They worked open shop until a union job came available. A lot of folks all of a sudden started to find homes over there [in the open shop] and never came back (personal interview, May 15, 1993). q Y conse uentl , in the short-run, at least, contractors that remained union did not have a significant labor productivity advantage over many of the newly non-union contractors. This effectively forced remaining union contractors out of much of the construction market. 62 With the decline of union contractors, Utah construction union membership fell (fig. 3.1). The decline in membership was accelerated by the 1982 recession. Union membership appeared to recover from the recession, but many dues-paying members were working open shop. With the onset of the next downturn in Utah constriction in 1986, union membership began to fall steadily. These data are consistent with the story that union members working in the open shop eventually found a home there and quit paying their union dues. With the repeal of the prevailing wage law and the resulting decline in unionization in Utah, average wages in construc tion fell relative to the average Utah wage (fig. 3.2). Construction - wages, which had ranged from 120 to 125 percent of the average Utah wage before the construction boom of the 1970s, exceeded 130 percent during the boom. When construction high end employment growth stopped in the late 1970s, consaucaon wage s fell back toward the of their normal premium over average Utah wages. But with the repeal of the prevailing wage law, construction wages fell to a new lower range of 110 to 115 percent of the average wage m decline in consaveaon wages and not isolated to union earnings Utah. Thi s is an across-the-board . nor the earnings of construction labor on public works. This relative decline to consavcnon earnings in Utah is consistent with the overall decline in construction wages following repeal (chapter II). The data for Utah actually underestimate the effect of Utah's repeal on construction workm earnings, in part because the data do not include the change in value of benefits. u Davis-Bacon Repeal Effects 5500 Law 8DE-:a ec in 1981 "7..........--•................................................... 9� vj%le:101� r 4500' - .._.... --- ------ - ...................._._....._............................ c 0 4000 .................... ................•----- ----.-............._... ..----- C 1982 Recession 3500 ..................................... .......................•• ........... .......................... Union Members "Carry Their Cards in Their Shoes" `o 2 2500 ..-------•------------------------------------------------- . ...__....................... ------ - - E Eventually, Members "Find z 2000-.------------------------ a Home in the Open Shop' 1500 . 19� 1979 1981 19 3 1985 1987 1989 Quarterly Membership Totals Figure 3.1 Union membership in construction in Utah, 1977-89 Source: Utab State Building and Construction Trades does records. Union membership began to decline with the prevailing wage law repeal and the onset of the 1982 recession. Membership recovered somewhat in 1983 but not as fast as overall construction employment. With the 1985 downturn in Utah construction employment, union membership began a steady decline to less than half its late-1970s peak. Employment Stagnates �t 35 I E ,a 130 ----.-----•--•----•........ ......... r 1 52 a) a 1 30 o .... ---....w..................... I --^ 125/o --M-- .o � � L-25 ------------------- ................. ... ................. 20 <Co U ' G �, Wages m a Cu 115% .... . .... Fall ...... .. 15 a ' CD 10 110% . . 90 50 55 60 65 7Q Law Repealed i— Wages @°A L1T median--&- C nsnc=an Endoy i Figure 3.2 Wages and employment in construction in Utah relative to wages Source: Utah Job Security, Division of Labor Market Information, Annual Report, table 5. Construction employment in Utah grew rapidly in the 1970s, but growth stopped in the 1980s and cyclical fluctuations became more pronounced. wages that ranged between 120 and 125% of the Utah median wage prior to the construction boom of the 1970s, rose above 130% of Utah's median wage during the boom. As the boom ended, wages moved rp e peal of Utah's prevailing wage law in 1981, wages down to their normal range. with the plummeted. Typically, unionized cons ruction workers receive better health and pension benefits han do non- unionized workers. Lower benefits, particularly health and pension benefits, conmbute to c!e increase in overall labor turnover in and out off the s essttroalnedustry and 'less experienced d This reased la or occupational turnover, we will see, led to y g force. The Relation between Repeals and Black unemployment It has been argued that the Davis-Bacon Act was passed, in part, to restrict southern blacks from northern construction job opportunities. It is further claimed that the current high and rising ratio of black unemployment rates relarive to white unemployment rates is partly due to restrictions that prevailing wage laws impose on the ability of unskilled black labor to compete with better skilled white labor. From these beliefs, it is argued that a repeal of the Davis-Bacon Act wouid lower black unemployment relative to white unemployment by opening up jobs for less-skilled black labor.63 d by the available evidence. Black unemployment These arguments are not directly supporte rates are separately collected for only five of the nine states that have repealed their state prevailing wage laws. Arizona, Idaho, New Hampshire, and Utah do not have large-enough black populations to generate meaningful unemployment statistics. However, Alabama, Colorado, Florida, Kansas, and Louisiana-do have sufficient black populations to test the above argument. The ratio of black-to-white unemployment for five repeal states can be shown In all ca ses, unemployment rates for white and blacks and white males and black males (fig• black unemployment rates are more than twice the rate of white unemployment. Before the repeal o and :he of state prevailing wage laws, however, the maleblack-to-white both less than theuneir unemployment ratios after overall black-to-white unemployment ratio were both these states repealed their prevailing wage laws. This does not mean that the repeals caused the black-to-white unemployment ratios to rise Black-to-white unemployment ratios were using across the country in the 1980s in repeal states and elsewhere. The rise in the black-to-white unemployment ratios simply reflects this time trend.` that never By comparing the states that retain their prevailing wage laws a trends those n black-to-white had prevailing wage laws, we can eliminate the effect of time a male black-to-white unemployment ratios. The black-to-white uuempto the ment ratio and y unemployment ratio are both lower for stases with prevailing wage laws ;,ompared to states Without prevailing wage laws— averaging unemployment rates across states amend yeost ars frroonm 1 and to 1992 (fig.3.4).6' The male unemployment ratios in figure 3A are statistically they are not different. These data do not support the proposition that a repeal of the Davis-Bacon Act would ameliorate in any significant way the relative unemployment of blacks to whites. �a�a-�acon Rcpesi �:few 3 2.5J Q aa) 2 ! a � E 03 ....... ...... m ------- Cu � 0.5 After Repeals 0 Before Repeals Males All Figure 33 The ratio of black-to-white unemployment in five repeal states before and after repeals Pv pea Source: US DOL Gicarranhicif Profile of P:mjLSme"t and Unem loym 1974-92. Five repeal states—Alabama, Colorado, Florida, Kansas, and Louisiana —have sufficient black populations to report a separate black unemplovment rate and a black male unemployment rate. In these five states, in the decade prior to repeals, the ratio of black to white unemployment rates was 2.43. After repeals, the ratio rose to 2.61 which means black unemployment was even higher in relation to white unempioyment For males, the black-to-wbite unemployment ratio was 2.0 before repeals and 2.60 after repeals. These ratios are based on unemployment rates for the entire state not simply construction. If repeals opened job opportunities for blacks, the effect is bidden. Black-wbite unemployment ratios rose throughout the 1980s and the rise is not due directly to the repeals. 3 ------ j 0 2.5 -----• c ' c I 2 ...... o ' I C m i U Cz Never Had Law Retained Law Q Males All Figure 3.4 Black-to-white unemplovmcnt ratio for states that retained and that never bad sate prevailing wage laws Source: US DOL Geograpbical profile of employment and unemployment 1974-92- Comparing the black-to-white unemployment ratio in states that retained their state se states prevailing wage laws throughout the last 25 eeeRettwith o! astro g time ratio in e trend thatishoa►s up at never bad state prevailing wage laws eliminates th in before-and-after analysis. The male black-to-white unemployment ratio is slightly ared with states that higher in the states that never bad prevailing wage laws comp nce is not statistically significant- The overall black-to-White retained theirs. The differ e unemployment ratio is signir<cantiv greater in the states that never having bad a of female unemployment dierr4estials, which are prevailing wage 12w, but this is because y unlike! to be significantly alfected by construction employment pa A Decline in Training With the decline in union membership and in relative wages, training for construction in union apprenticeships and through vocational schools—declined in Utah. Union apprenticeships are tied to the availability of union jobs. For instance, unionized plumbers and pipe fitters in Utah, the United Association of Journeymen and Apprentices of the Plumbing and Pipe Fitting Industry of the United States and Canada, historically have attempted to maintain apprenticeship rates at 10 t (fig. 3.5 . As emp loyment ment in the state Y to 15 percent of the number of union journeymen plumbers ( g ) P boomed in the 1970s,however, the union could not meet the demand for journeymen from union increased apprenticeship rates to a peak of 25 percent in uentl the union aPP . contractors. Consequently, 1975. The boom persisted, but the backlog had been remedied. So the union lowered its apprenticeship rate back to normal ranges by 1978. Employment during the construction boom peaked in 1979 and membership in the plumbers and pipefitters union peaked in 1981. With the repeal of the Utah prevailing wage law, the union dropped its apprenticeship rate a histo rical low. Union membership fell slightly in 1982 and began a steeper to 10 percent,P decline in 1983. Faced with these sustained declines in membership, the union cut its apprenticeship rate even lower in 1986 and thereafter. Unions hit harder by declines in . ente rs union Utah membership have scaled back their apprenticeship programs further. The carp , locals 184 and 1498 of the United Brotherhood of Carpenters and Joiners of America, which graduated seventy in a class in 1977, graduated five in 1992. The Utah International Union of Bricklayers and Allied Craftsmen suspended its apprenticeship program altogether. The decline in union apprenticeship training in Utah has not been offset by a rise in other sources of training. Because the repeal of Utah's prevailing wage law was motivated by a desire to limit state expenditures, state legislators were not eager to raise funding for state-sponsored vocational training. Although the number of vocational graduates in construction grew in the 1970s, the construction labor force grew more rapidly. Thus, while the 1970s was the heyday of vocational training at Salt Lake Community College, vocational graduates as a percentage of the construction labor force had already begun to decline." The steady decline in state-supported vocational training as a percentage of the construction labor force through good times and bad supports the notion that the state has simply tried to get out of the business of vocational training in construction. The fall in union membership and wages has made construction a less attractive career. t the same time, unions are less able to train construction workers. As unions are weakened and community colleges drift toward academic offerings, the capacity to respond smoothly to an upsurge in construction jobs is undercut. And federally sponsored Job Corps vocationsl training is not in a position to fill in the 8aP• Federal revenues pay for Job Corps training in Utah at the Weber Basin and Clearfieid centers. Federal funding in real terms for these centers his not expanded, but the Weber Basm Job Corps Center, which draws predominantly from the Utah population, has significantly cut Ies construction worker training throughout the 1980s. This center committed itself to changing from an all-male student population in 1980 to 50 percent female by 1990. To accommodate this switch, training for traditionally male occupations such as construction, have been scaled back .o Davis-Bacon Repeal Effects k"J Peaks in ------- I 200 ----------------- .............. ........ Above Normal ................. ............. i 1CO 2 55 0/6 Training Rate 4, 200le --------------------------------- ........ -0 Cz ....... ------------------------------------—900 ......................... ---------- --- ------ :ts vi ................. Boo .................. ----------- -----------4 100 -27 .................... ------ 700 z 5%- g ,,/WJLSOI-------------------------------------------- Rate 600 9 0 1971 1981 -4*- Plumbers-number Apprenuces-percent Figure 3.5 Apprentict plumbers as 2 percentage of journeymen plumbers in Utah, 1%1-91 Source: Utah plumbers and pipe fitters local's membership records. The plumbers' union in Utah has historically attempted to train apprentiCtS at a rate or-10 to 15 percent of their journeymen members. As employment boomed in the 1970s, the union could not meet journeyman demand and consequently expanded apprenticeship training rapidly. As the numbers of journeymen grew to meet demand, apprenticeship training was reduced to DorM21 rates. But with the repeal of the state prevailing wage 12w in 1981, -union membership declined and apprenticeship training rates were Cut to 211-tiMt lows. to accommodate new offerings in traditionally female occupations, such as office management and clerical work. Cement masonry and heavy-equipment training have been eliminated, and instruction in carpentry, painting, and brick laying has been cut in half The Clearfield Center has graduated approximately 100 construction trainees per year since the early 1970s. Fewer Clearfield graduates go into the Utah labor market, compared with Weber Basin graduates, because most of Clearfield's students are from out of state. Perhaps 10 percent of Clearfield's graduates go into the Utah labor market, but this percentage rises during periods of local labor shortage. It is estimated, however, that at most only 25 percent of Clearfield's graduates will stay in Utah. Even without union pressure, it is possible that a shortage of skilled construction workers in Utah will raise wages and induce a new generation of young people to enter construction vocational training for the industry. Nonetheless Utah is now in a building boom — when wages would normally rise — and annual earnings in construction relative to annual earnings for all Utahns continue to fall. In 1993, the most recent year for which data are available, the construction earnings premium fell to a historic new low of 103 percent of the average annual earnings for all non-agricultural worker in Utah. Utah is now in a building boom, one that has come quickly. High-quality training programs, which take time to create, are not in place to meet the demand. This adds an additional lag to the usual time it takes to train a skilled laborer. Utah's current boom has relied partly on using a less-skilled labor force (which partly accounts for the lower construction earnings premium) and partly on travelers from California, which is currently in a construction lull. Whether the Utah construction industry can rely, in the long rim, on training systems for construction worker in California remains to be seen. A pick-up in California construction would quickly bleed away the skilled workers Utah is now attracting. This is one difference between state repeals of prevailing wage laws and a federal repeal of the Davis-Bacon Act. If construction cycles are not ible, if a state is lucky, for one state to freely ride on the training synchronized, it is at least poss Davis-Bacon would create a nationwide decline in training. systems of another state. A repeal of Da Under such a circumstance free riding on the training of another area would not be an option. Market Responses: Training, Turnover, and Careers The market in Utah has not successfully made up for the decline in union and state-sponsored n of Bui lding Contractors (ABC) has Associatio 8 training. At the national level, the non-union for hours contributions to a training fund. attempted to replicate the union system of bargaining Y It is difficult, however, to induce ABC's member contractor to include general training costs in their bids. Each contractor fears that his competitors will not include training costs. Thus, its an attempt to be the low-cost bidder, ABC contractors,often refrain from including training costs despite the ABC initiative. Consequently, very little ABC training has occurred in Utah. ip programs operate, however, in the licensed trades of In Utah, non-union apprenticesh n electricians and plumbers. In 1992, there were 846 non-union licensed apprentice Utah and 2,068 non-union journeymen. Thou, there are 4 apprentices for every 10 journeymen in the non-union sector. In contrast, there were 123 apprentices and 607 journeymen sector,in P union sector in 1992, or 2 apprentices for every 10 journeymen. In the begin at around S6 per hour with no benefits. Over a four-year period, the state mandates that Davis-Bacon Repeal EfPecu the apprentice wage rise to 80 percent of a journeyman's pay. In the union sector, apprentices begin at S7 per hour whin an addirioral S3 in benefits. Their wages rise to S14 per hour plus S3 in benefits over five years. Non-union apprentices are sponsored by a particular contractor that oversees on-the-job training, and these apprentices take classwork at a participating community college. Union apprentices work under the direction of an apprenticeship coordinator, rotate among employers for on-the-job training, and take classes at community colleges and union apprenticeship centers. Roughly 90 to 95 percent of the union apprentices complete their programs and graduate to journeymen status, while only 15 to 20 percent of the non-union apprenticesgraduate.raduate. Given these rates, in four years, out of 846 non-union apprentices, we should expect 125 to 170 journeymen to be graduated. In five years in the union sector, out of 123 apprentices, 110 to 115 apprentices would graduate to journeymen electrician. Thus, while the non-union sector accounts for more than 85 percent of all electrician apprentices, it accounts for about 60 percent of journeymen graduates. Economic theory is consistent with this pattern wherein non-union apprentices are paid less en and graduate at a lower rate than union apprentices. Economic theory Posits that in the absence of marketwide institutions or government subsidies, individual workers will have to pay for their own on-the-job training when the skills leamed are general to an industry and not specific and unique to the activities of a particular firm. The worker-learner pays for training by accepting a wage that is lower than the value to the firm of that worker's marginal product. By working for less than the worker's worth to the employer, the worker pays the employer for on-the-job training. That beginning non-union electrical apprentices cam S6 per hour while union apprentices earn S10 per hour (including benefits) is consistent with the theoretical proposition that non-union apprentices pay for their own training by taking a discounted wage below their marginal value to the contractor. Because the employer does not a much for non-union training, the theory suggests that the pay employer has no stake in the worker's training. If the worker leaves, the employer does not Icse any investment in the worker's human capital. So, the employer will tolerate high levels of turnover. Because the worker is receiving less than what the worker can earn in other jobs with no on-the-job training, the worker may be tempted to exit jobs with training when current personal budget needs became pressing. So, on both the employer side and the worker side, turnover is tolerated in the non-union sector. This view is consistent with the higher turnover rates among non-union apprentices, but other factors also contribute to the roughly 20 to 90 percent differential in non-union to union graduation rasa. Because the non-union employer prices new hands at discounted wages that shield the employer from investing in the human capital of new workers, the employer does not screen new workers extensively to forestall subsequent turnover The employer's failure to preselect new workers for aptitudes and attitudes consistent with a long-term attachment to construcnon work adds to the turnover among non-union consaucnon apprentices. in contrast, the joint apprenticeship boards of unions and union cony-uwrs do considerable preselection for aptitude and attitude before letting a candidate into an apprenticeship program. This is because the union contractors and unions will invest in the union apprentices' training.' In the non-union sector, workers may also leave apprenticeships if it becomes apparent that the employer offering training at a discounted wage is not delivering on that training promise to tram Because em-21cvers are able to discount wa?a of app,.-ennces below -.^eir current word to 43 lava-�acoa Rexa: �TKJ the employer, it is tempting to engage in bait-and-switch tactics whereby training is promised but not delivered. By saving on training costs, the employer can earn an additional profit from employing green hands at discounted wages. In the union sector, because employers and union journeymen invest in the training of the apprentices, bait-and-switch tactics are less attractive. Because the apprentices' wage is not discounted as much below what they could earn elsewhere, the apprentices are not as tempted to leave. Thus, the non-union sector must begin training five apprentices to graduate one journeyman, while the ratio in the union sector is close to one to one. While non-union contractors tolerate high levels of turnover among apprentices, with the decline in training and union membership, non-union Utah contractors have sought to reduce the turnover among trained journeymen. There has been a long-term decline in labor turnover in construction (fig. 3.6). This long-term decline can be explained with a pooled cross-sectional, time-series linear regression model, as can the differences to turnover rates in Utah by contractor type from 1956 to 1991 (table 3.1). Not surprisingly, this model shows that turnover was higher in years in which variations in monthly construction employment were great. It also shows that contractors with larger crews tolerated proportionately more turnover. Contractors employing more-expensive labor sought to reduce turnover. When union membership was a high percentage of the construction labor force, turnover was higher simply because contractors losing one good worker could turn to the hiring hall for a reasonable substitute at little additional cost. When vocational schools were graduating a large number of construction-trained students relative to the Utah construction labor market, contractors tolerated more turnover because the market had proportionately more trained substitutes. The numbers of union membership and vocational graduates have been on the decline, however. Thus, this regression model shows that, over time, contractors have responded by reducing the turnover among journeymen . Although turnover at the firm level has been on the decline, workers may be entering and leaving construction at higher rates than 20 years ago. In 1970, Utah construction workers, on average, were years Y 42 old. B 1990, bef ore the recent construction boom had begun in Utah, the age had fallen to 33 years.70 Much of this decline may be due to the construction expansion in the 1970s, which brought in a new generation of younger workers. But the decline in age may also be a result of both the decline in health and retirement benefits and the decline to relative wages associated with the decline in unions. Although non-union contractors increasingly are providing health and retirement benefits, especially to their key people, the health benefits tend to be more expensive for a given level of care and the retirement 401 K plans lack the insurance component associated with union-defined benefit plans National Trends in Registered Apprenticeship Training The U.S. Department of Labor, Bureau of Apprenticeship Training, monitors registered Data are avulable apprenticeship programs-- union and non-union construction industry.to the _ for 1975-78 and 1987-90. Not all states have2�9 states to the Bureau ofdid reportAegistered copprenticeshipasavcaoa ng for all years during these periods. NonethBless, apprentices for every one of those years. The states included 6 states that eventually repealed 44 Davis-Bacon Repeal EfFecu 24'a .-- ----------•-•--• ..... .._..... 2 0��� .f,� 1" r ...._. -.--- ................................................. 18 \ lr .. .-. .. 1-.................................................._..._. 1 'r ......................��. ..-. - ,�o � -------------- . ............................. ........................................ �o� - - 8�a ----------------------- .-----•-•------•---•---•---- ...................................................••.--........._... __. --...... ... ....... 40% 65 70 15 80 85 °0 50 60 1 YEARS i [--t- ent Cona-act Corot "'�" Building All Employm —S— Heavy & H —4-- Specialty ion �ad Figure 3.6 Turnover in Utah's construct ��u` Informationket Annu>i Report, table ode Source: Utah Job Security, Division of Labor journ eymen in U a ion hiriag halls declines and the number or As the number o6 trained }ours - reducing turnover. non-union journeymen declines, firms respond by g Table 3.1 Linear regression model of turnover rate in construction in Utah. 1956-91 Source: Utah Job Security, Annual Report. table 5, Dependent variable = firm turnover in construction' Actual Standardized. Variable' C'oeff icient Coefficient Union Members' 1.76 .24 New Vocational Graduates` 2.45 .20 Real Wage -.076 -•62 Seasonality '_.12 .15 Workers per Contractor .052 .40 (Constant) -1.88 ---------------- ' The actual variable is ln(turnover/0-turnover)) to meet the technical requirement in linear regressions of being an unbounded dependent variable. ° All independent variables are stadstically significant at the 1% level. ` As a percentage of the construction labor force. Adjusted R2 Square = 0.24 Number of Cases = 351 Contractor Type = 4-digit SIC _ Contractors in Utah have tolerated higher labor turnover when union membership has been a high percentage a of the lab or force and wbes new vocational school graduates have been plentiful. Turnover has been more common is rears when monthly employment has fluctuated a IoL Contractors have been more willing to tolerate turnover among lower- paid workers and have bad to accept higher levels of turnover among larger work crews. Standardized coefficients indicate that worker skill and crew size bave bad the largest effect on variations in employer turnover rates, while the availability of both union members and new vocational graduates have bad larger effects than seasonal fluctuations in employment. their prevailing wage laws, 4 states that never had prevailing wage laws, and 19 states that retained a state prevailing wage law throughout the period. These 29 states can be divided into the categories "repeal," "never-had," and "retained-law," for cornpanson (figs. 3 7 and 3 8). .No state had repealed its prevailing wage law by 1978 By the end of the first quarter of 1987, all nine repeal states had passed their repeals except Louisiana which repealed in 1988. The data for 1987 are for the summer of 1987, after Kansas had repealed in that year. In the "before pe riod, states that had prevailing wage laws — those that retained such a law and those that had not yet repealed theirs — typically trained a higher percentage of registered apprentices than the states that never had a prevailing wage law. For unknown reasons, the year 1976 is an exception to this pattern. During this pre-repeal period, the states that would eventually repeal their laws had as high or higher trainingrates compared with the states that kept their laws throughout the period. By 1987, traini ng rates had fallen for all sates, but they had fallen least in states that had retained their prevailing wage laws. By 1989, the states that had repealed their prevailing wage laws had training rates as low as the states that never had prevailing wage laws. This is clear evidence that repealing state prevailing wage laws lower formal apprenticeship training. A simple analysis can help isolate the effect on training of repealing sate prevailing wage laws from a general downward trend in construction apprenticeship training. Apprenticeship training rates for states that repeal their prevailing wage laws in the late 1970s and 1980s are presented as a percentage of the training rates of states that retained their prevailing wage Iaws (table 3.2, col. 2). Throughout the 1970s, before repeals, the repeal sates had training rates that were at or above the average training rates for sates that had and would keep their prevailing wage laws. After the repeals in the late 1980s, the repeal states had training rates that fell to as little as 63 percent of the training rates of states that kept their prevailing wage laws. By 1990, the repeal states had relative training rates that were as low as the states that never had prevailing wage laws. Thus, while training in construction has been falling for all states, the fall for repeal states has been the most precipitous and — setting time trends aside — the repeal states matched the training rates of the retaining states prior to repeal and fell to the rates of sates never having had prevailing wage laws after the repeal. cession analvsis can Unlike the simple analysis Just presented, however, a multiple linear peg control for other factors, such as differences in state unemployment rates or regional differences in training (table 3.3). The dependent variable in the analysis is a transformation of the training r each state where the training rate is calculated as registered apprentices as a percentage rate for , of all construction employees in a state and year. For technical reasons associated with the regression analysis, the actual dependent variable is the natural log of the assumptions of linear gr y odds ratio of the training rate where the odds ratio is calculated as (the percent trained) divided by (one minus the percent trained).' In the regression model, regional differences in training rates are controlled for with the regions corresponding to standard Bureau of Labor Statistics regional categorizacons Unemployment differences are controlled for by state and year. The data are for the years 1975- 78 and 1987-90. The focus variable is REPEAL, a dummy variable equalling 1 once a state repeals its prevailing wage law. A second focus variable is NEVERHAD which equals zero for in all states except for those nine states that never had a sate prevailing wage lawconstruction For those s VEZ ates, N-E �iAD equals L T'r.er- are 297 ccservanons in the data set. Califon :a Day.'s•33con Repeal Effecs 47 5.50% W a, 5.00% - ................................... ....... - - o` ........................................... 3: 4.50% - .................................... c4.00% - .•...- After Repeals ........-•-••-_... 3.50% ................................. .... ....._....----....._......................._..._ •� I a3.00% ............................................ .................. --...------- 2.50% ------•• R ..... .. ._.. Before epeals j CD j . cn 2.00% .... ......... -- ----- ° ............................................ .......................... 1.00% 75 76 77 78 87 88 89 90 Year -♦- Repeal - Never Had -a Retained i Figure 3.7 Apprenticeship training rates, by state groups, before and after repeals Source: U.S. Department of Labor, Bureaus of Labor Statistics and Apprenticeship 'Training. This figure shows apprentices as a percentage of all construction workers in 29 states grouped by state treatment of prevailing wage law. In the four years before the repeal of state prevailing wage laws, states that would eventually repeal their laws had high apprenticeship training rates. States that would retain their prevailing wage laws also had high training rates. Except in 1976, states that never had prevailing wage laws in construction had relatively low training rates. In all state groupings,training rates in the late 1980s were lower than training rates in the late 1970s. However, after the several state repeals, those states that retained their prevailing wage laws had relatively higher training rates. Those states that repealed their prevailing wage laws eventually had training rates that matched the states that had never had prevailing wage laws. .. ........................................... ..................................•. 4.15% ........................................................ a) 400%/ ................................. r-28 .................. ............................. . ......(n 3.0%_ .............................. 2.8% 2.5%- ..... ...... ...... M-66 ...... 2.0%-/ ...... ...... CL ...... ...... ------ ...... ...... ...... 0.5%_ 0.0% Repeal States: States which: States which: Repeal 2tI=!;. "airs.191f,79 WC W Figure 3.8 Apprenticeship training rates, by state Source: U.S. Department of Labor Bureaus of Labor Statistics and Apprenticeship Tr2iCiVg- States are grouped here into four categories, repeal states before and after their repeals of prevailing wage laws, states that retained their prevailing wage 12WS, and states that never had prevailing wage laws. This simple pattern shows that repealing or not having prevailing wage laws reduces formal training in Construction. (Part of this before-and- after picture is due to an overall downward trend in registered apprenticeship rates in construction overtime.) Repeals hurt appreaLictshil) training because repeals hurt unions- Non-union construction contractors do less training and less formal, high quality training. Delaware, the District of Columbia, Hawaii, and Rhode Island are omitted from the analysis because they did not report to the Bureau of Apprenticeship Training of the U.S. Department of Labor during the second period of our analysis. The model is a good fit of the data with an adjusted R, of 45 percent, and all variables are statistically significant. The focus variable in the regression analysis REPEAL — a marker for states that repealed their prevailing wage laws — is negative. This means that — controlling for unemployment, time trends and regional differences in training — when states repeal their prevailing wage laws, the training rate goes down. At the mean training rate for the entire data set of 3.7 percent, this model indicates that repeals drove down training rates to around 2.1 percent. The NEVERHAD variable, marking states that have never had a prevailing wage law, is also negative and statistically significant but smaller than the REPEAL variable. This is because of a close correlation (about 40 percent) between never having had a prevailing wage law and being a southern state. This means the analysis could not fully distinguish between the hypothesis that training rates in the South were low because many of these states never had prevailing wage laws and the hypothesis that other reasons associated with being a southern state caused training rates to be low. The REPEAL effect was easier to pick up compared to the NEVERHAD effect, simply because the repeal states presented information about their training rates before and after each state repealed its prevailing wage law. Thus, looking at training rates from a variety of measures and methods of analysis, it is clear that state repeals of prevailing wage laws have significantly lowered formal, organized, and quality training of construction workers. The effect is to lower training rates by about 40 percent. When apprenticeship training falls as a result of repeals of state prevailing wage laws, minority participation in apprenticeship programs falls even farther (fig. 3.9). Minorities comprise almost 20 percent of all construction apprentices in the repeal states in the years before repeal of state prevailing wage laws. In the same states, after repeal of their prevailing wage laws, minority participation in apprenticeship programs falls to just under 13 percent of all apprentices. While construction apprenticeship training is falling in these states by around 40 percent, the share of minorities in this downsized training also falls by about 36 percent. One reason for the decline in minority training is the decline in union training. In figure 3.9, the share of minorities in apprenticeship training appears the same for states that retain their prevailing wage laws and states that never had such laws, but this is an illusion. Many of the states that have never adopted prevailing wage laws are in the South where there is a high percentage of minorities in the overall stue population (fig. 3.10). We account for that factor with the ratio of the minority percentage in construction apprenticeship programs, divided by the minority percentage in the state population. This ratio is 100 percent if the two percentages are equal. We call this the "minority reflection percentage" because it measures whether minorities in apprenticeships reflect minorities in the state population. In the repeal states before repeal, the minority reflection percentage was 107 percent, which means that the construction apprenticeship programs slightly over-represented minorities. After repeal, minority representation in apprenticeships fell to 85 percent of minority representation in the state population. In the states that retained their prevailing wage laws throughout the period under review, minority representation in apprenticeships just about mirrored minority representation in the state population (a ratio of 102 percent). But, in states that never 50 Davis-Bacon Repeal Effects Table 3.2 Training rates in repeal and never-had states as a percentage of training rates in states that retained their wage laws Reveal States States Never Having Had Law (1) 1 -,(2) (3) 1975 106% I 85% 1976 112% 1090/0 1977 100% ` 86`7c 1978 97c7o 819c 1987 87r7c 74% 1988 687o 1 527a 1989 70°/0 70% 1990 63% 60% w have training rates which never had prevailing wage la s g r Except in 19776, the states that fall from 86 percent of the training rates of states that retain their prevailing wage laws to 60 percent of the training rates of states that such laws. Repeal states mirror the I re repeats of s rate training races of retaining states prior to their De als several. After the se re . prevailing wage laws—from 1978 to 198S— the average training rate in repeal states falls to 63 percent of the training rates its states retaining the laws. This is a simple way of viewing the roughly 40 percent drop in registered construction apprenticeship training caused by state repeals of their prevailing wage laws. Table 3-3 Training rotes in repeal and never-had states as a percentage of training rates in states that retained their wage laws, 1975-78 and 1987-90 Source: US Department of Labor and Bureaus of Labor Staristics and Apprenticeship Training Dependent Variable= Log of the Odds Ratio of the Percent Appr.ntices Effect on fndeoendent variables Percent Trained Region 1 -1.1 1 Region 2 -0.99 Region 3 -0.77 Region 4 -0.81 Region 5 -1.18 Region 6 -1.l0 Region 7 -0.53 Region 8 -0.55 Time trend -0.02 State unemployment rate 0.04 Marker for states never having had law (NEVERHAD) -0.13 Marker for states once they repealed their law (REPEAL) -0•44 Constant Adjusted R2 =0.45 Number of Cases =297 Years 19.75-78 and 1987-90 All variables are statistically significant at the 1% level except the marker for states never having had a prevailing wage law. That variable is significant at hte 10% level. Region 1: CT MA NH RI VT ME Region 2: NY NJ DC PA DE MD Region 3: WI IL IN OH MI Region 4: ND SD MO MN KS IA NE Region 5: WV VA NC SC GA FL Region 6: TX OK NM AZ AS MS LA AR TN KY Region 7: MT WY CO UT M Region 8: CA OR WA NV This multiple linear-regression model shows that — even when controlling for regioaal differences in training rates, variations in state uoemploymeat over time, and as overall decline in apprenticeship training — repeals of state prevailing wage laws signifieand lower apprenticeship training rates. �oarora -aoea 20.0% _. - .. ...... ........•...........•... i$.0 /o I r�29 lyevw r ad Law F........................... :'.ems l aw .... 15.1% � 1544/6 A.�er"aoeaa 14.0% ro.221 -66 ..... c 8 12.00 ... --- ...... g. .. O u' 10.0% _ ..... c 8.0% _ -- C. -- CL 6.0% __. -- -- I i m 4.0% i 2.0% 0.00/0 �Repeal States: States whit: States which: Repea �::ram.-- �..R.,a7s.7e,K'a6 7 s s, 53 renticeship Figure 3.9 Minorities as a percentage of all construction apprentices by state groups,u Training- Source: Source: U.S. Department of Labor 1:3ureaus of Labor Statistics and App onstruction, mio In repeal states, before repeal of their prevailing wage laws in c ai.I participation in registered apprenticeship programs averaged 19.1 percent off ail apprentices. After the repeals, minority participa I n fell to 12-4 percent of all apprentices. The n_28 and n=66 refer to the number of state-year observations in eacb group. States that kept their prevailing wage taws and states that never had prevailing wage laws had roughly the same rate of minority participation througaout 1975-78 and 1987-90. On average, however, populations of the states that never had prevailing wage laws had much higher proportions of minonties. 1200 o !Bebre Rao"I .... r,eovn ...................................................... 100°0 10T/ Attar r-28 102'/. ver Clad law 0) n.221 cc ... ... ... ..... C C C 60% ..... .. •-- E �% 20% Repeal States: States which: States which: Repeal States: 1 QC%=Minority App. Reflect State Pop. Figure 3.10 Ratio of the percentage of minorities in construction as a ratio of the percentage of minorities in the state population, by state groups Source: U.S. Department of Labor, BAT and BLS. A ratio of 100% would show that the proportion of minority apprenticeships in each group of states reflects the minority as part of the state population. Minority participation in construction apprenticeships mirrored the state population in repeal states prior to repeal and in states that retained their prevailing wage laws throughout 1990. After the repeals,however, minority participation in apprenticeships in repeal states fell to levels that seriously under-represented minorities and resembled the under- representation characteristic of states that never bad prevailing wage laws. NOD-union apprenticeship programs tend to be small and do not fall within the oversight of affirmative action guidelines — which may be why the repeals have led to an under- representation of minorities in apprenticeships. 4 had prevailing wage laws, minority representation rates averaged 83 percent throughout the period. Thus, both repealing states prior to repeal and "retaining" states throughout the period had minority participation in construction apprenticeships that mirrored the state population. In contrast, both repealing states after the repeal and states which never had prevailing wage laws had substantially under-represented minority participation in construction apprenticeships. Summary Employment in construction is inherently unstable, because the industry fluctuates cyclically and seasonally — and firms expand and contract their employment as they win and lose job bids. Unions have acted like a flywheel in the industry, creating career workers when there were only casual jobs. Unions did this by facilitating the movement of journeymen from employer to employer and minimizing the employers' transaction and screening casts for the training. Unions also lowered training turnover by providing a mechanism whereby employers and journeymen could rationally invest in the human capital of apprentices. This raised the wages of apprentices so they would stay with training and induced the union and employers to promote the passage of apprentices to journeymen in order to preserve their investment. Unions also encouraged the career attachment of trained journeymen by providing relatively high wages and additional wages in the form of health and retirement insurance, which are increasingly attractive to workers as they age. By creating career jobs in a casual labor market, unions created the institutions needed to make human capital investment a rational market activity. With the decline of unions in Utah, the formation and preservation of human capital skills have become less-rational. Self-investment by apprentices becomes more precarious as the differential between the apprentices' wage and alternative wages in other industries widens. It simply becomes more reasonable for apprentices to leave construction if unforeseen personal budget problems emerge. The high turnover among non-union apprentices represents in the aggregate a considerable loss of human capital to the construction industry, even though it is not a loss the employer or the state pays for directly. With the lowering of construction wages, it becomes reasonable for young construction workers to limit the amount of human capital they invest in themselves. With the worker's lower stake in construction skills and with the disappearance of wages in the form of health and old-age insurance, it becomes more reasonable for journeymen construction workers to abandon the construction field when they start families This represents an additional loss of built-up human capital. Contractors in Utah have attempted to minimize the effect of this increased skill volatility within the industry by encouraging firm attachment. Still, despite initiatives, such as profit- sharing, 401K plans, and health insurance, designed to attach key workers to a firm, construction turnover remains well above the average for the Utah labor market. In short, union decline has meant the decline of the career worker within Utah construction, a diminution in incentives to invest in constriction skills, and an increased loss of accumulated human capital as apprentices and journeymen leave the trades. Although the loss of human capital and career jobs in this industry does not appear as a private cost on the ledgers of any contractor, the industry and society at large pay a price for the loss of financially secure occupations in construction. Not only is quality in the industry put at risk when human capital stocks are allowed to dwindle, but the SS Davis-Bacon Repeal E:fecu quality of social life is imperiled when we dismantle the institutions that generate stable jobs out of unstable working conditions. This instability is mirrored in the continuing decline of construction wages in Utah. Despite elanive a return to boom times in Utah construction, construction workeras had o piggy-back r boom on the to average annual earnings in the state. Utah's construction ifornia construction workers. Whether Utah can continue this free ride is uncertain. training of Cal ide from the effects of a federal repeal of Davis-Bacon. What is certain is that there is no free r Experience from state repeals indicates that formal apprenticeship training in constnicnon will fall by about 40 percent if Davis-Bacon is repealed. If state experiencesare wprdiawse,minority hurt minority workers most. In states that repealed their prevailing age oration to participation in apprenticeship programs fell from reflecting each state's minority p p significantly under-representing minorities. This pattern is consistent with states that have never s had prevailing wage laws. Although states that retain their prev states�etha have never had participation in apprenticeships that reflects their state populations, prevailing revailin wage laws have minority participation rates that are only about 80 percent of the rates in which minorities are present in the state population. From chapter II, we have seen that a repeal of the Davis-Bacon Act will lower construction stent with the case study of Utah presented in this wages and earnings. That finding is consi chapter. We have also seen that a repeal will significantly reduce training in bs market becomes � n and as the Jo nstru cao well be that as the stock of human capital falls in � aining casual and turbulent, more minority workers will obtain Jobs. But they will not obtain tr will not be entering as they do now in the states that retain their prevailing wage laws and they into occupations that offer a middle-class income with benefits. Se Da%ii-Bacon Repeal Ef zcw IN'. Construction Safety Put at Risk Construction is dangerous work. In fact, it is the nation's most dangerous Indust y According to the U.S. Bureau of Labor Statistics: More than 900 construction workers are killed each year — 3 to 5 per workday • 510,500 work-related injuries and illnesses occur annually — almost 2,000 cases per "day. • day.74 0 cases involve lost work days, for a total of 4.6 million days lost from work per year. A recitano ru n of the hazards associated with construction work, however, cannot ignore the substantial variability of across of accidents and their q .job saes and institutional environments. Accidents and injuries are the product of a complex interaction between worker environment, and injuries will be either fostered or limited, depending on how well nis and enviro , interaction promotes safety. This chapter focuses o the safety t of rather e repeal of thanoveralls health, at this wage laws on injuries in construction. The focus on safety juncture, is strictly a concession to the paucity of reliable data on illnesses related to construction. Why might the repeal of a state prevailing wage law affect the safety record in construction? worker and How does the presence or absence of such a law alter the important For instance fconstrucII n rice of inu environment? Certain parameters are key to the incidence 1 �' Stresses associated with and inexperience Ste.. work is more dangerous when workers are untrained P the possible avenues for grievance all feed i a lack of job security, the pace f o and Pnto the o w critical interaction of work and envi rkronment on any job site. training declined as te's prevailing wage law, g In Utah, following the 1981 repeal of the sta p the construction labor market was going into recession (see chapter III). The lack of training and widespread use of inexperienced workers vegan to surface as the cons=ucuon economy rebounded. One experienced pipe fitter recalls of that era. Contractors were using inexperienced people with no training. They had no tde. g program to begin with, they were hiring people off the sa with no experience i =e n the cde. What fie` would do is everyone that got hired on one project that did not have a history or work experience on a construction job, they had to wear a rod sticker on their hard 'hat. They had to rc- wear that for 30 days. Well evervwhcrs you would looked the,n were red sticr'xh� ��� 1 estimate that about 40 or SO% of the people had one o "hamburger kids." — Pipe Etter, Salt Lake City, 1994 Lack of training and inexperience are not the only sources of work injurees. In Utah there was a greater sense of job insecurity after the repeal of the attire's prevailing wage law and the ing related decline in union work. Without union security, ex-unione�rkrQr o th tr P � e*<Zerience found themselves taking chances th-y would not have p u:.icn worker wi:o was forced to .ake w�o;x in the open shop recalls: oa,,s•3acoc Rcreal ElTecu I got hurt in 1986. There was a great deal of pushing to get the job done. I was working with an older man that came out of retirement. He was about 70 years old. We were waiting for a cherry-picker to move some pipe. We were waiting for a couple of hours, because they laid off some operators. After two hours of waiting, two hours of superintendents eyeballing us, I went and walked under the piece of pipe, which weighed 253 lbs. I carried it over to the structure, but I didn't see because the snow was covering a hole in the ground. I stepped in it, it was about 14 inches deep and 2 feet across. I pulled muscles in my back, pulled some discs in my back. What I was thinking of at the time was, I can't afford to lose this job. All these guys walking by me looking at me, I thought we better get this pipe in there some way. I was nervous, I should not have done it but I did. — Union pipe fitter, Salt Lake City, 1994 u Why Prevailing Wage Repeals Lead to Increased Injury Rates why repeal safety p We can postulate, based on studies of s ry and health in the construction industry, y, re of the state prevailing wage laws is associated with increases in injury rates. Take as the first premise these telling facts: • The rate of injuries "decreases substantially as length of service increases."73 • Large, experienced employers in construction have injury rates that are 80% below small- to-medium-size contractors. Repeals of state prevailing wage laws have altered construction labor markets in those states in several ways that affect job site safety: 1. The bidding process has become cutthroat. 2. Workers are less likely to make a career of construction work. 3. Even as experienced workers are leaving the industry in increasing numbers, apprenticeship training has declined. Cutthroat competitiveness in contracting. In Utah, the repeal of the state's prevailing wage law led to a burgeoning of start-up contractors with limited track records (chapter II). These new entrants joined existing contractors in a heated bidding process for state contracts that resulted in lower bids, but ultimately higher costs, as a percentage of the state engineer's estimate of the job cost. Cutthroat competitiveness, in other words, resulted in increased cost overruns. Inexperience at the firm level, small size, and cost pressures all contribute to compromised safety on the job. Because of their relative inexperience, new firms tend to face greater on-site coordinanon problems than firms with longer track records. Such problems can add to costs, but also directly endanger safety.Problems in coordination,perhaps related to delivery of materials and equipment, or in scheduling work with subcontractors, lead to greater u ncertainry with respect to true construction schedule. Uncertainty is a breeder of safety risk, as workers can less easily anticipate and plan for the daily contingencies of work. New entrants in the industry also are generally smaller in size than established firms. Smaller firms have worse safety records than larger firths, in part because of greater laxity of enforcement of safety rules and the relarive absence of formal safety programs. Of greatest importance, however, is the firm's reacuon to increased pressure to cut costs to sa Davis-Bacon Repeal Effects the face of intensified competition and cost ovemans. There is a tendency to speed up work and cut back on safeguards in the face of such pressures. Workforce turnover. When state predvail it�n�de wage retain workers for llong-term careers (see increased significantly, as the industry fount chapter III). Repeals resulted in a decline in the union share of the construction labor market, driving down average construction wages in the state and decreasing union apprenticeship training hi for construction. In response to the decline to union members and toeminimizedscreening t , contractors and training attempted to reduce turnover — to retain skilled workers costs. Still, the decline in wages and in health and pension benefits drove experienced construction workers from their trades for careers in other industries. Thus, while construction firm turnover is on the decline, tumover in the whole industry is on the rise. on federally funded Davis-Bacon projects are more likely to be union 'Those who now work Yng are likely trained because of the demanding naturo esn technology,e, civi and they are morelikely to have to know more about new processes andchanges graduated from certified apprenticeship programs. In states that retain their prevailing wags law — compared with those that never had such a law or repealed such a law — the proportion of construction workers receiving training is higher and injury rates are lower. A decline in wages and benefits leads to a flood of well as a decline in skilled, expenenced workers as inexpenenced workers into the industry work safely needed to supervise the recruits and to assure that they y eriance is a major Decline in the skill base of the construct ion labor mark& Exp . srruc determinant of safe work performance — tion and productivity.in rai ograms,ing of skilled of which are workers is normally conducted through apprenticeship tr g p g operated by unions and employers through joint trust funds.o kerstlearn from experegral part of ience wwhile learning on the job while property supervised. In thajentic� are trained to identify and correct on a variety of projects. Among other things, apprentices ergonomic problems, to detect physical hazards, and to detect the presence. or ate P release of a about safe and acslth hazards, aPP P r hazardous chemicals. Knowicuge abo safety measures, and hazard communication methods are all important elements that apprenticeship programs provide. renticeship programs decline and � When little Davis-Bacon acts are repealed, training and app i the skill base of workers erodes (chapter • ncentives to continue and Without employer apprenticeship programs, knowledge of proper safety health Procedures declines as well.to er Summary. The combination of these factors — cutthroat ��P tition, a the safety hint iated skill and and an erosion of career attachments to the industry rove and know experience base of the construction labor force. Workers become more tnjury-p as the workforce becomes less skilled less about the kinds of risks they are taking. Furthermore, sty to mply and its wages in construction decline, workers are forced to take moreeedaue oftenspresst their make a living. Furthermore, contractors caught in the competitive sp P workers to speed up and take more chances. Workers are put at increased risk in an already hazardous industry. Davis-Sacon Rc^eal E recu �a A Comparison of Injury Rates The U.S. Bureau of Labor Statistics' annual Occupational Injuries and Illness Survey reports accidents by state and year. Construction injuries vary by the type of work being done. We will analyze these BLS data for plumbers and pipe fitters employed by specialty contractors in the Standard Industrial Classification (SIC) 171. This specialty trade has injury rates in the mid-range of rates for construction and this trade is often employed on public works. For pipe fitters in 1978-91, states that had state prevailing wage laws averaged 13.83 injuries for every 100 workers employed (fig. 4.1). In addition, in the states that repealed prevailing wage laws, injury rates for plumbers and pipe fitters before repeal was slightly less (I3.54 per 100 workers) than the injury rates in other states with state prevailing wage laws. By contrast, states that never had state prevailing wage laws had higher injury rates (14.74 per 100 workers) and the repeal states, after they repealed the prevailing wage Iaws had the highest injury rases of 15.41 per 100 workers. These increases in injuries resulted in a similar increase in workdays lost per worker." It is possible that injury rates might differ between states for reasons other than changes in legal status. The union pipe fitter who got hurt in Utah in 1986 slipped partly because of snowy conditions. Perhaps factors associated with safety unrelated to repeal coincidentally worsened after repeal. We controlled for factors such as regional differences in weather, time trends in injury rates, and the effects of unemployment in a multiple regression analysis of construction injuries among plumbers. This approach permitted us to isolate the effect on safety of changes strictly associated with the repeal of state prevailing wage laws. We modeled injury cases per worker as a fiutction of geographic regions, the unemployment rate, a time trend, and the legal status of state prevailing wage laws (table 4.1). Three measures of injury rates are reported: injury cases per worker (col. 2); serious injury cases per worker, defined as injury cases that required time off from work (col. 3); and the number of lost work days per worker col. 4 . In all three models, our focus variable, the act of repealing a state Y P ( ) prevailing wage, has a positive coefficient. This means that as the states repealed their prevailing wage laws, injury rates went up according to all three measures. In our model, the dependent variables are logged. This allows for a straightforward interpretation of the repeal variable as a percent increase in injury rates. So, as these states repealed their laws, the injury case rate went up by 14 percent, the serious injury case rate went up by 15 percent and the work days lost per worker per year went up by 12 percent. All of these findings are statistically significant All other things being equal, states that have never had prevailing wage laws also have higher injury rates for plumbers and pipe fitters in the construction industry. In terms of injuries per worker and serious injuries per worker, our results indicate that states that never had prevailing wage laws affecting construction had a =nstically significant 5 to 9 percent higher rate compared with states that have prevailing wage laws.' The Cost of Injuries The costs of injuries in the construction industry are staggering. Of the nation's S62 billion spent on workers' compensation, approximately 30%goes for construction-related injunes and illnesses, Davis-Bacon Repeal Effects 60 i .............................................� !4�t1i 1 A. OC% 12.00°O '0 10.0C°o I .. -- .. O 8.00°o 6.00% 4.00% 1 --- 2.00 0.00% Before Re-Peal Have Law Never Had Law After Reoea! Figure 4.1 injury rates in construction by status of prevailing wage taw Source:. US. Department of Labor, BLS. Injury rates in construction were relatively low in the nine repeal states prior to repeal (13.54 percent). After the various repeals, injury rates, on average, rose to 15.41 percent. In the 32 states that have retained prevailing wage taws, injury races have been and remain relatively low. In nine states that have never had state prevailing wage laws. injury rates were and remain relatively high. The notation "a" refers to ttie numbers of st2te-7ear observations in each group. For instance, there were :-0 state--fear combinations for states that had prevailing wage laws in 1978-91. I r Table 4.1 Regression model of the effect of state repeals on injury rates for plumbers and pipe fitters Source: US DOL, BLS. Dependent variable: log of injury rate for plumbers and pipe fitters (by year and state) Serious Days Cases Cases Lost Per Worker Per Worker Per Worker (1) (2) (3) (4) (Constant) -1.21 -2.16 -6.85 III Region 1 -039 -0.41 -0.10' Region 2 -027 -029 0.14' Region 3 -0.46 _0.70 -035 Region 4 -0.40 -0.65 -0.44 Region 5 -034 -0.49 -0.29 Region 6 -013 -0.40 -0.13' Region 7 -032 -0.64 -0A3 Region 8 -0.18 -0.26 -025 Time Trend -0.02 0.00' 0.01 Unemployment -0.18 -0.19 -0.04' Never Had Law 0.09 0.07 0.05' Repealed Law 0.14 0.15 0.12 2 Adjusted R 35�b 49% 16% J Observations 350 313 350 Not statistically significant. (Regions are standards BLS categories). In columns (2), (3) and (4), we report three models of injury rates, the first for injury I cases per worker (2), the second for serious injuries per worker (3) and the last for days lost per worker (3). Controlling for regional differences in injury rates, general trends in injuries over time and variations in state unemployment rates, all three types of injuries are higher in states that bave repealed their prevailing wage laws and states that never had such laws. In repeal states, injury rates climb from 12 to 15 percent compared to the rates prior to repeals. or roughly 520 billion. This, for a construction labor force which represents but 5 to 6 Percent of the whole U.S. labor force. In addition to the direct costs of workers' come.^.saaon. there are numerous industry-related indirect costs connected to work-related injuries or dear^.s These include job shutdowns and retraining of workers. According to the Construction Industry Institute, "even when the estimates of claims are deleted from cost a,dat indirect costs still exceed the direct costs. of the nine states that repealed Based on the our regression model of the experience their prevailing wage laws, we project that national injury r es79 will increase by around 15°o if the Davis-Bacon Act is repea led. What this will mean to terms of safety is: • There will be 30,000 new cases of lost-time injuries each year, accounting for 675,000 days lost from work. • Workers' compensation costs will increase by about S3 billion per year. • Because Davis-Bacon construction accounts for approximately 10epercent of all construction, nsmctlon costs it is estimated that repeal of the Davis-Bacon Act would in by S300 million per year indirect, workers'-compensation-related costs alone, and indirect costs would double this figure. The numbers might prove larger, because a Davis-Bacon repeal in the wake of state repeals may have a larger impact on the construction industry. Summary The institutional context of work is critical to worker health and safety. State prevailing wage laws on the surface, have little to do with worker health and safety. But such repeal has fundamentally altered` an institutional context that was more conducive to workplace safety Repeals of state prevailing wage laws, therefore, have had hidden effects. Because the bidding process becomes overheated; because contractors, as a group, take less respoe sworkforrY or b g training and safety; because workers feel less secure on the job; and because becomes less attached to and experienced with construction work,put a risk by repeal n� becomes more dangerous. Safety in an already relatively dangerous industry p prevailing wage laws. Darts-Bacon Repeal 63 I A.A Davis-Bacon Repeal Effects V. Conclusion The Effects of the Repeal of Prevailing Wage Laws The federal system of government in the United States is sometimes called "democracy's diverse experiences of the 50 states afford a valuable window for assessing the workshop." The dive and failures of public policies. Between 1979 and 1988, nine states repealed their successes prevailing wage laws regulating the construction of public works. These legislative changes enable us to examine the before-and-after pictures of the effects of such repeals. Nine other states public construction, while the remaining thirty-two never had prevailing wage laws governing states retained prevailing wage laws. These "never-had" states and "retaining" states give us additional perspectives on what it means to keep or repeal prevailing wage laws. Legislators are often forced to act on theory; this is one instance where they can act on face of the last 20 years in the application and removal of state and experience. The experience ffers insight into n A prospective effects of further prevailing wage laws on public construction o state repeals or the proposed repeal of the federal Davis-Barn The Goals of State Prevailing Wage Laws Prevailing wage laws were first enacted at the state level. Kansas passed the first prevailing wage law on public works in 1891 as part of legislation mandating the eight-hour work day. Prevailing can citizens wage laws were central to a larger effort to improve worki ng conditions for and t}iiszht-hour The notion was that child labor laws should enable children to be in sc hool work day should help allow workers time to spend with their families. The proponents of prevailing wage legislation wanted to prevent the government from using its purchasing power to undermine the wages of its citizens. It was believed that the government should set an example, by paying the wages prevailing in a locality for each occupauorl hired by government contractors to build public projects. Massachusetts, Nebraska, New Jersey, Before the Great Depression, Arizona, Idaho, Kansas, New York, and Oklahoma passed prevailing wage n A Soolattihereafter, buildingstate 8 addiuo al states construction. In 1931, Congress passed the Davis-Barn adopted prevailing wage laws. After World War [I-and un: til 1982,s 15 moreant states pas g prevailing wage laws. All of these laws raised the question wage? The Definition of a Prevailing Wage Wages in local labor markets often have a peculiar distribution. Particularlymarket where there Lm is often the most unions, but also in other circumstances, the highest wage in a local labor e occurs most often, however, it fofl�n� commonly found wage rate. Even when the highest _ however few or many — be the average wage simply because the lower wag es occupation will bring the average wage down. Prevailing wage laws are intended to get the govern:^ent out of he business of pulling down ,5 j)avv-Bacon Repeal E c" e pays the average wage, I will automatically undercut the wages. The dilemma is that if the star p y most commonly found wage. Alternatively, if government pays the highest wage found, it w111 always be pulling the average wage up. When is the highest wage sufficiently common that it should be called the prevailing wage rate, even though it will never be the average wage? In the federal law, this dilemma was resolved by a threshold rule. This rule stated that if the most commonly found wage rate, to the penny, accounted for more than 30 percent of all wages oat for an occupation in a local labor market, that was the prevailing wage even though wa s not the average wage. On the other hand, if the most commonly found wage rate accounted for less than 30 percent of all wages for an occupation in a local area, the average wage rate prevailed. In 1985, the Reagan administration revised the rule and raised the threshold to 50 percent. Today, Davis-Bacon wage rates are the average rate for an occupation in a local labor market except, in roughly one-third of the cases, where 50 percent of the wages in that area are precisely the same. If more than half of all workers in an occupation in an area make the same wage, that wage rate -- even if it is above the average — is said to prevail. But two-thirds of the time the average wage prevails. Modern opposition to prevailing wage laws is usually founded on one of two objections. Some people oppose the idea of the government agreeing in advance to pay the average wage rate for workers in specific occupations in a local area This cnticism is completely at odds with the original purpose of prevailing wage legislation, which was to prevent the government from hiring labor at below-standard rates. Other critics object to paying a prevailing wage that is greater than the average wage in the locality. The premise of this second objection has lost a great deal of its force in recent decades. As a result of the adoption of the 50 percent threshold, and the additional fact that unionization in the construction labor market has fallen from 70 percent to about 25 percent in the last three decades, there are far fewer cases in which the wages rates determined as prevailing are greater than the average rate. The Financial Costs of State Repeals Lower wages for all construction workers. Supporters of Utah's 1981 repeal of its prevailing wage law recognized that repeal would lower construction wages. They maintained,however, that the money saved on public works construction justified the government's indirectly lowering the wages and earnings of some of its citizens. And, indeed, construction earnings did fall. In Utah, construction workers, who through the 1950s, 1960s. and 1970s earned 120 to 130 percent of the average non-agricultural wage in the state, saw their wages fall steadily after repeal. By 1993, Utah construction workers were earning only 103 percent of the average annual earnings in Utah, even though Utah was then experiencing a massive construction boom, in which construction wages normally go up. This earnings decline affected all Utah construction workers — whether union or non-union, whether employed on publicly or privately financed projects. Taking the nine repeal states as a whole, the average annual earnings of construction workers in these states fell from 524,317(in 1991 dollars) per year before the repeals to S22,148 after th4 repeals. This is simple but compelling evidence that repeals of state prevailing wage laws have lowered construction wages. A more complex analysis confirms this general observant whilen. ncontruollple ng linear fo the businessanalysis, we isolated the earnings effects of the state repeal Davis-Bacon Repeal E1Teets 66 cycle, regional differences in wages and unempiovment, and long-term trends in earnings ar:d employment that are not associated with repeals of prevailing wage laws. We found gnat in 199 t dollars) per worker every year s repeals co cons—,uction workers in those states Sl,�-- i ction earnings a,:butable to each since sate repeal. This was about a 5 percent drop in constru . slate's repeal of its prevailing wage law on public works. A slight increase v construction employmenL Proponents of state repeals -maintained that the lowering of wages would be offset by an increase in construction employment. While high- paid, high-skilled work ers would be hurt by a repeal, it was believed, low-paid, low-skilled workers would have more job opportunities in construction. increase in o t an would leas o . labor o Repeal proponents were right that cheaper construction 1 construction employment. Again using regression analysis, we found that the repeal sates experienced a 1.7 percent increase in construction emplomnthetstandpointawould lof workers howevea without these repeals. This was an unfavorable trade-of from as their wag es fell by 5 percent overall while their employment rose by le we 11 ss -than 2 percent. It turned out to be a tough trade-off for government budget-watchers as Lost tax revenues. As a group, construction workers lost income, because their wages dropped by 5 perce nt and th eir total employment rose by less than 2 percent. This caused the government to lose substantial tax revenues. In recent years, the state it has lost p to S5 milIion annually in sales tax and income tax revenues because it repeated its prevailing wage law in construction. Increased construction cost overruns. Cos overruns are a hidden cost of repealing in Utah resulted from an over-heated bidding process prevailing wage laws. In Utah, the overruns in which contractors, uncertain about each others labor costs and confronted with the enrry of many cart-up construction companies, shaved ids oneir sate jobds in a desperate came in lower than ever government contracts. After the repeal, winning bids before, but the final job costs were a higher percentage id sub oI al emwo d��ge change (chapter 2, fig. 2.3). Having underbid jobs, contractors to ;,;tiers to get the jobs done or simply walk away he from ndof state eb-, and leave the roads plein sthee 10 pick up the pieces. In Utah, cost overruns o years after repeal, compared with the 10 years before. The bottom linefor Utah's budget The Congressional Budget Office esdmaat sf that, should the federal Davis-Bacon Act be repealed, the federal government might save a to percent on its construction costs. This savings might even be less." Using an even more conservative by repealing what Utah saved in construction costs its prevailing figure of 3 percent to estimate cite — broke even. tat wage law, we calculate that the Utah se budget al most — but not q Balancing construction cost savings a8�n lost tax revenues, in two of the years since 1987 the Utah budget saved more money in construction costs than it lost in tax revenues five o 5�1) g years since 1987, the state lost more in tax revenues than it saved in construction 8 In either case, the net savings or losses were small compared with the lost earnings of Umh's we workers — and the industry — were to lose more citizens (table 2.6, row 3). But construction than money when these repeals were enacted. Devts-Bacac Repeal Effects S4 $3-00 ..... ..... .. . .. .......... .......................................... $2 $1 ............................_......... .......................................... c o •------ -------------•----............-----•-----•--........---....----•--•- ($2) ------•- .................................................__................. + ($3) -------- ............................................_...._ Lost Taxes Saved Costs Budget Loss Figure 5.1 Average annual income-tax revenue loss and construction cost savings and net effect of repeal for Utab, 1987 to 1993, in 1994 dollars Source: Table 2.6. On average, the repeal of Utah's prevailing wage law has cost the state budget 5400,000 per year from 1987 to 1993. This figure has been rising and reached S1 million in 1993. Should the federal prevailing wage law be repealed, the gap between lost federal tax dollars and construction cost savings will be greater. This is partly because a Davis-Bacon repeal would affect more construction and more workers, but also because the federal government income tax rate is higher than Utah's. The higher the income tax rateare taxed, the greater the taxes if incomes fall. Other Costs of State Repeals A less-skilled labor Jorce. Unions and union contractors do the lion's share of worker training in the construction industry. Some very large non-union contractors do their own training, but most non-union contractors hire out-of-work union-trained construction workers and workers who have learned a trade on a catch-as-catch-can basis. Most non-union contractors are not big enough to afford to train and retain their own labor force. Contractors, understandably, are afraid that in the first slack period, the workers they trained will leave them and work for their competitors. Unions historically have compensated for this market failure by inducing union contractors to share the burden of training and to share each other's apprentices. In Utah, the repeal of the prevailing wage law led to a dramatic decline in union apprenticeship programs because the repeal led to a dramatic decline in local construction unions. Having repealed the prevailing wage law, the state was not inclined to pour money into local training centers to make up the difference. At first after the communitycolleges and vocational g u repeal in 1981, the Utah consrrtiction economy limped along in the trough of a business cyc le so the absence of quality training systems was not strongly missed. Non-union contractors hired out-of-work union members and the older generation of construction workers provided a relatively skilled labor force in the open shop of non-union construction. In the last three years, however, Utah has experienced a massive construction boom. Few training systems were in place to meet this boom. Utah has filled the gap by relying on traveiing construction workers from California, which is in a construction slump. Utah has also relied on a less-skilled labor force. Whether Utah will be able to continue to rely on California workers ll remains to be seen; if California's economy picks up, many of the skilled California travelers likely return home to the increased wagcs there. Utah's experience with declining availability of construction training was not unique. Comparing repeals to the decade after repeals, union and non-union the decade before apprenticeship rates in construction fell by more than half in the nine states that repealed their ned their prevailing wage laws did not lose ground prevailing wage laws. States that retai states that never had revailing wage laws had relatively low training apprenticeship training and st P rates in construction throughout the period. The repeal of prevailing wage laws thus had the indirect effect of reducing training and hindering the formation of a skilled labor forca. When unions declined in the wake of repeal, only state government could have picked up the pieces. The cost of expanded state-financed . vocational training is a hidden cost of repealing prevailing wage laws. So far, it is a hidden cost that few repeal states have been willing to Pay. A faltering stock of human skills is Slowed economic gains by minority works. g construction is not the only nonmonetary cost that resulted from state repeats of prevailing wage laws. Construction used to be one of the few blue-calla occupations left where a worker lacking a college education could earn amiddle-class income. Nationwide, IIa ker was a road out of poverty income in 1994 was S27,500. Becoming a skilled construction for minority workers. Before the nine state repeals, participation by minority group members — male and female nonwhites -- in construction apprennceships mirrored the minority populations in each state. In the repeal states before the repeal of their prevailing wage laws, minonties accounted for t9 Davie-Bacon Res-cal Efrectj almost 20 percent of all construction apprentices. After repeal, minonry participation fell to 12.5 percent of all construction apprentices. Thus, after these repeals, minorities became significantly under-represented in construction apprenticeships. One reason for this decline is that union apprenticeship programs usually enrolled dozens of apprentices. Non-union apprenticeship programs tied to single employers tended to be smaller, often involving no more than one, two, or three apprentices. Affirmative action regulations do not cover apprenticeship programs of fewer than five apprentices. So the union programs had to fill out affirmative action plans and follow affirmative action guidelines, while the smaller programs did not. When the repeals drove the union programs into decline, minonry workers lost the most. For instance, the percentage of minority apprentices in construction, which reflected the minority proportion in each state's state population before repeal, declined in the repeal states (fig. 5.2). Minority construction workers may still enter the industry but they are less likely to receive full formal training in the absence of prevailing wage legislation. Although it has been suggested that repeal of Davis-Bacon would lower black unemployment relative to white unemployment by opening up jobs for less-skilled black labor,`' the data do not support such a claim (see chapter I 3, figs. 3.3 and 3.4). Thus, repeal means that minority workers will begin construction work in unskilled jobs and get their training, if at all, on a catch-as-catch-tan basis. Furthermore, minorities will enter an industry that is less able to provide a secure blue-collar, middle-class income. Repealing prevailing wage laws has therefore cut off an important road for minorities into the middle class. training, worker are less roductive; without safety training, they are at greater Without skills trot P I 8 risk of injury in an already dangerous profession. Increased work-related injury rats. All construction workers in the nine repeal states have been Put at increased physical risk by the repeal of the several state prevailing wage laws. Injury rates in construction in the nine repeal states have nsen by 15 percent after repeal, even controlling for other factors such as unemployment, trends in construction safety, and differences in work safety experiences by region. This finding is consistent with other research. The Department of Labor found that the rate of injuries "decreases substantially as length of service increases."" If the experience in these states can be extended to the nation, a repeal of Davis-Bacon would result in 76,000 additional construction workplace injuries annually. About 30,000 of these injuries would be serious, requiring time off to recover. More than 675,000 work days would be " lost. These new injuries would occur because workers would be less well-trained and because they would have fewer on-the-job protections against contractors who are in a hurry. Workers, of course, suffer directly from these occupational injuries— info conayasctor�s well- Workers, and in their wallets. Increased injury rates also lead to Increased costs must pay higher worker's compensation premiums. c as consumers higher of worker's compensation compensation ction semces. local, state, and federal governments pay a share premiums. Darns-Bacon Repeal E:Tecu 70 120°�a ................................ .............-•--- 100 a� c� ........... ............ ---....----- i 80% a`) ............ 60% -...----- ............ a 40o/o o - - -------- ` 20% 0% Before Repeal After Repeal 1000/6--Minority App. Reflect State Pop. Figure 5.2 The percentage of minority apprentices in construction, divided by the percentage of minority population in the state — the minority reflection percentage — for nine repeal states Source: Table 3.8. In the nine repeal states where separate data were available on minority populations, in the decade before repeal, minority apprentices were slightly over represented relative to their proportion of the state population. The minority reflection percentage was 107 percent. In the decade after the repeals, the minority reflection percentage fell to 85 percent, indicating significantly under- represented minorities. Billions of Dollars 50.5 $0.0 w 0 ----- ......... m ($0.5) ......... v � Lost Taxes Saved Costs Net Loss Figure 53 Estimated effect of a repeal of the Davis-Bacon Act on income-tax revenues, construction costs and total budget Source: Table 2.7. The Congressional Budget Office estimates that the federal government would save a total of 1.7 percent in construction costs from a repeal of Davis-Bacon. This chart uses the more conservative cost savings estimate of 3 percent. At a 3 percent construction cost savings, with a marginal income tax rate of 20 percent and federal construction expenditures at their 1991 level (in 1994 dollars), a repeal of Davis-Bacon would cost the federal government Si.2 billion in income tax revenues. The federal government would save S346 million in construction costs and the federal budget would lose, on net. $838 million. Estimated Effect of a Davis-Bacon Repeal Democracy's workshop has given us an opportunity The nine states that repealed the.'r "little" Davis-Bacon Acts offer a chance to estimate the likely consequences of the repeal of the federal Davis-Bacon Act. Based on this study, we project the following. First, construction earnings would drop if the federal law was repealed. Collecrvely, for all construction workers, this would mean a loss of almost S5 billion per year in real terms every year. As a result of lower wages in construction, federal income tax collections would fall by federally urchased construction almost cted cost savings on fe y p roughly S1 billion per year. Prod e g certainly would be less. (fig. 5.3). Second, we estimate that formal training in construction could fall by 40 percent. The industry would move from one of skilled blue-collar workers earning a middle-class income to a much-less-skilled labor force earning substantially lower wages. Nl;nonry access to good training likely would fall even farther than overall training rates. Contractors would be using more construction workers and paying less for them, but the less-skilled workers would be building less and adding less value to building projects. Purchasers of construction services would not necessarily profit from lower-wage labor if that labor is also less skilled. This is a potential lose-lose situation. Utah was able to patch together a large-enough construction labor force after its repeal of prevailing wage law. Contractors in Utah rode freely on the training systems i Davis-Bacon place in California. But the country as a whole cannot go on a simil tie�nitedeStates will otbe is repealed and construction training nationally declines sharply,, small state like Utah turning to California for its rescue. Nationally, there will be no place to turn. Is the federal government prepared to spend the money to establish its own apprenticeship programs in construction? Alternately, will the government induce or require contractors to join into cooperative training programs? If prevailing wage legislation is repealed, it is likely that some additional measures will be needed to ensure occupational training for the cons mcnon industry. Last, but not least, we estimate that the construction job site would produce 30,000 These injuries would add a large but mil-undetermined us injuries earl J additional steno ) yearly. financial cost to the ultimate price of repeal. It goes without saying that the public benefits from a bidding process that lowers construction costs. But the bidding process trust be kept within certain bounds, to p . as could lead to increased — rather than decreased — public and societal costs consequences that to Competitive pressures tempt contractors to cut corners on quality. States and communities emp y rs to assure that quality is maintained. Histoncally, unions have assumed the role building inspectors of "building inspector" for safety and training in the construction industry. le o unions Employment in construction is inherently unstabie because the industry The role f P win and lose fluctuates cyclically and seasonally. Firms expand and contract employment as they job bids. A worker rarely has a long-term attachment to employer in the industry, and L construction union may be the only stable, work-related institution op ertt not. oer ut of i Construction unions act like a flywheel in the industry, creating careerpo zasual labor market. Unions do this by facilitating the movement of joumevmen from employer to employer and minimiIIng the employers' ;rans.a.caon and screening cow. Davis-Bacon Repeal Effect Unions lower training turnover by providing a way for employers and journeymen to rationally invest in the human capital of apprentices. Collectively bargained agreements create wage incentives for apprentices to stay with training programs, and also cause their employers to promote the workers' passage to journeyman status. Unions also encourage the career attachment of trained journeymen by providing relatively high wages and health and retirement insurance, which is increasingly attractive to workers as they age. By creating career jobs in a • � casual labor market, unions create the institutions needed to make human capital investment a rational market activity. With the lowering of construction wages, young construction workers will limit the amount of human capital they invest in themselves. With a lower stake in construction skills and the disappearance of wages in the form of health and old-age insurance, it becomes more reasonable for many journeyman construction workers to abandon construction work entirely when they start families. This is an additional loss of built-up human capital. The loss of a career. Contractors have attempted to minimize the effect of this increased skill volatility in the industry by encouraging attachment of workers to their firms. Still, despite initiatives such as profit-sharing, 401K plans, and health insurance to bind key workers to the v remains high. It appears that the decline of unions has been turnover firm, construction firm to gh associated with the decline of the career worker in construction, a diminution in incentives to invest in construction skills, and an increasing loss of accumulated human capital as apprentices and journeymen leave the trades. ' The loss of human capital and career jobs in this industry does not appearas a private cost on the ledgers of any single contractor. Nonetheless, the industry and society at large pay a price for the loss of middle-class occupations in construction. Not only is quality in the industry at risk e but the soli of our society is imperiled when human capital stocks are allowed to dwindl , q ty when we dismantle the institutions that generate stable employment out of unstable working conditions. The construction industry is turbulent. Caught in a perennial boom-bust cycle, characterized by fleeting relationships between small contractors and subcontractors, and driven by short-term strategies of free-riding on the training of others, the construction industry is a market failure waiting to happen. The turmoil in the construction labor market has traditionally been tempered prevailingby - e legislation and labor unions. Absent these institutions, it is unclear how — waggl or whether — the market will regularly and carefully tram workers, or assure safety and health on the job site, or provide training opporturuties for minority workers, or offer the incomes needed to make construction an attractive career. Government pond more, the ases account for r20 percenent t of all construction in the United States. For the last six decades thew s contributed to the stability in construction labor markets by requiring on act rsto Pato use wage rates that already prevail in a local areas. Today, voices are urging government its purchasing powers to reduce construction costs at the expanse of worker incomes. Such a strategy has a very real cost for workers, the industry, and the government. When nine states chose this path, the results were significantly lower construction wages, slightly higher construction employment, a tripling of cost overruns on public works, an across-the-board 15-percent increase Dsvts-Bacon Repeal E(Tecu 74 in construction injuries, a 40 percent decrease in apprenticeship training, and an even further decline in minority apprenticeship training. All this was sacrcf;ced to save an estimated 1 i percent in state construction costs. Even that savings was squandered by the loss in state tax revenues from an impoverished construction labor force — a poor bargain indeed. I Dads-Bacon Repeal Effects End Notes 1. Topeka State Journal, February 24, 1891, col 4. p. 4. 2. Topeka Daily Capital, February 25, 1891. p.1. 3. Topeka State Journal, February 25, 1891, col. 3-4, p.1. 4. Sixth Annual Report of the Bureau of Labor and Industrial Statistics,126. 5. Sixth Annual, 215. 6. Sixth Annual, 124. 7. L. 1891 Ch. 114 p.192-193. 8. In reply to the question of needed legislation,most workers polled by the Kansas Bureau of Labor and Industry. cited a lack of enforcement of the eight-hour law or complained of long hours. (Twelfth Annual Report of the Kansas Bureau of Labor and Industry Statistics, 1896, 88-89) Sunilarly,the Kansas Bureau of Labor and Industrial Statistics repotted complaints of noncompliance with the eight-hour law. P.E. Cook, of the A.R.U. Local No.57, stated that the '[eight hour] law is being openly violated by the corporate and private parties on public work....' (Fourteenth Annual Report of the Kansas Bureau of Labor and Industry Statistics, 1898, 204.) 9. Sixteenth Annual Report,272-78. From July 1900 to June 1901, 33 cases were won upholding the law. 10. Thirreenth Annual Report of the Bureau of Labor Statistics of the State of New York, 1895, 515-37;Foureeemh Annual Report of the Bureau of Labor Statistics of the State of New York, 1896, 802. 11. U.S. Congress, 1927. Although it has recently been asserted that the Alabama workers were black there is oo direct evidence supporting this claim. (See George F. Will,"It's ume to repeal the Davis-Bacon Act,'Deseret. eye. Feb. 5, 1995.) 12.Hearings Before the Committee on Labor, House of RepresentativesSeventy-First Congress. January 31. 1931 Bacons proposal was re-introduced in 1930 as H.R. 9232 by Congressman Elliot W. Sproul from Illinois, while Bacon proposed a complementary bill. 13. Armand J. Thieblot Jr.,Prevailing Wage Legislahon:The Davis-Bacon Act.State "Little Davis-Bacon Acu. The Walsh-Healey Act, and The Service Contmet Act. Phi adelphu The Wharton School,1986, p. 8. I4. Laws of Florida 1979 ch79-14 HB730. 15. Thieblot, 163. 16. Laws of Alabama 1979 Act no. 79-123 H362 Letson. 17. Thieblot, 151. 18. Mark Erlich, Labor at the Ballot Box: The Massachuiens Prevailing Wage Campaign of 1988 (Philadelphua Temple University Press, 1989). 33. 76 Davis-Bacon Repeal Effecu 19 Nelson,Richard R. 'State labor legislation enacted in 1981' U.S. Department of Labor Employment Standards Administration Division of State Employment Standards,1. 20. Salt Lake City Tribune, Jan. 23, 1981, 92. 21. Salt Lake City Tribune January 23, 1981 B2. 22. Salt Lake City Tribune Jan. 16. 1981, A6, col 1. 23. Salt Lake City Tribune, January 16, 1981 (A6, col 1). 24. Arizona laws 1984 S.C.R. No.1001. 25. The Phoenix Gazette, 'Little Davis-Bacon Act is Ruled Unconsutuuonal,' Sept. 12, 1979, p. A-1. 26. Thieblot, 12. 27. Idaho Session Laws Ch 3 HB7. 28. Thiebiot, 165. 29. Colorado Revised Statutes 1985 Article 16,8-16-101. 30. Thieblot, 159-60. 31. New Hampshire Ch I17(SB71). 32. Thieblot, 184. 33. Kansas Ch 186 S.B.112. 34. Louisiana Act No.18 H.B.Z. 35. Erlich, 4 and 6. 36. Erlich, 7. 37. Regional Information Group of Data Resources,'Executtve Summary of the Study of the Economic Impact of Repeal of the Massachusetts Prevailing Wage Law' Lexingtim MA. August 1988. 38. The Massachusetts Foundation for Economic Research, Thr Peculiar Prevailing Wage Law (I I I Cabot Street Needham, MA 02194), March 1988. 39. Mas sachusetts Foundation for Economic Research. 10. 40. Data Resources, L 41. Data Resources, 4. 42. Erlich, 101, I11, and I12.. Davis-Bacon Repeal Effects 43. Senator Harrison Williams, United State Senate Heanng Before the Subcommittee on Housing and Urban Affairs,First Session on Oversight to Examine the.4dminutration of the Davis-Bacon,4ct, Washington.CPO,1979, 1-2. 44. With the increasingly prevalence of market-recovery agreements between unions and contractors, more often there are multiple union wage rates for a single occupation in a local labor market. This means that, even when unionization rates are above 50 percent, there is not always a single union wage rate that counts for 50 percent of workers in the market. Thus it is even less likely that the union rate will be the prevailing rate. 45. Current Population Survey. 46. Let us first investigate this issue Using an example. Assume that union wages are S 16.30 per hour and non- union wages are S10 per hour. When average wages decline by 7.5 percent, and if non-union wages remain the same, what is the percentage decline in union wagesl W. r W. r, W� 7.5% Union Non-Union overall Wage Wage decline in Percent Percent W,R Decline Decline S16.30 0.13 510.00 0.87 510.82 510.00 S 10.00 0.13 S 10.00 1 0.97 I S 10.00 60.3% 0% Where W. is the union wage, r, is the percentage of the construction workforce that is unionized, W.is the non- union wage,r, is the percentage of the construction workforce that is not unionized, and Wes, is the average wage in construction. This table shows that the percentage decline in union wages must be 60 percent. 47. Assume that non-union wages are S10 per hour and there is initially a wage differential of 20 percent between the union and non-uruon workers. This implies that the union wages are S 12 per hour. If unions represent 13 percent of the construction labor force,average wages decline by 7.5 percent,and the wage differential is eradicated, what is the percentage decline in non-union wagesl W 7 5% decline Percent Percent W.� r, W. r. .., in W., Decline Decline in Umon in Non- Wage Union `� S 12.00 0.13 S 10.00 0.87 310.26 59.49 S 10.00 0.13 S 10.00 0.97 S 10.00 59.49 0.13 59.49 0.87 S9 49 21% roe percentage dec1me in union wages must be 5 percent ttrom 110 per hour to T9.4 per hour). 48. Assume there is an initial wage differential of 20 percent between the union and non-union sectors. After the repeal of a prevailing wage law,assume the union-non-union wage differsnual decreases to 10 percent. Non let us investigate the effect of a 7.5 percent overall fall in wages. Assume that non-union wages are S10 per hour and the wage differential between the union and non-union workers falls to 10 percent.This means that the union wages are about S 1 i per hour. If average wages decline by 7.5 percent and the wage differential remains unchanged, what is the percentage decline in union and non-union wages? 78 Davis-Bacon Repeal EtTecu W. W� wen 7 5% Percent Percent Ca Ce decline in Decline Decline Won to Union to Non- Wage Union W S 12.00 0.13 S 10.00 0.87 S I0 26 S9 49 S11.00 0.13 S10.00 0.87 S10.13 510.31 0.13 59.37 0.87 S9 49 14 1% I 6 3% The percentage decline in union wages is 14 percent (from S 12 per hour to S 10.31 per hour), and the percentage 6 percent from S 10 r hour to S9.37 per hour). wages is ( per decline in non-union g Pe 49. Unless described differently, figures are given in 1991 dollar amounts. 50.The data are provided in four-digit detail of the Standard Industnal Classification(SIC)code.Data are from U S. ice Office of Earnings and Employment Statistics, Data Analysis Department of Labor, Bureau of Labor Statist , Section, Special Tape XC4057, provided by Darrell E. Cart. 51. 'Secular' trends refers to trends in earnings that are not due to fluctuations in the business cycle nor due to the state repeals of prevailing wage laws. 52.Controlling for contractor type is a conservative procedure.Overall construction earnings may decline as a result of ea s within trades mstead worker . We are focusung on the decliner�8 of a shift in the mix of construction type of any decline resulting from a shift to the mix of trades. Additional earnings losses may be calculated associated with a shift to a mix of less skilled workers. This is one reason why the regression estimate of earnings decline is lower than the simple estimate which does include the effect of changing crew mixes within the states construction industries. Unemployment rates are for each state for each year. te 53. This is an annual earnings average by SIC group. When e-�urto. are weighted_ $hd by the number of workers :n each group, earnings fall to slightly below 525,000. 54. Technical derails: This regression model was tested on 27,778 observations. Control (dummy) variables tor .6 Included in the regression model but not reported ' cations we7•e 13r code SIC classifications d induce ) detained 4-digit standar try in the table. Each coefficient reported in column(2)of the table is stausucally significant except the control for the mountain states region. (This means that the estimated rsponal effect on annual construction earnings for the mountain region of-S79 is small and probably not difTerew frem zero.) The unemployment rates for 1991 for the example states shown in table 2.3 were 4.9 for Utah. 5.0 for Georgia. and 5.9 for Maryland. The model is it geueralized leastquare weighted regression with the weig etng s wei ted re ht b the square root of the annual average employment in each SIC industry for that year. The Rt is 0.73, which means the model is a good fit of the data 55. The model also estimates a negative effect on anau;l aw ags of S1173 associated with raising the threshold for construction contracts covered by a prevailing wage law to S500.000 or more. This suggests that at some pout raising the threshold has a similar effect to repealing the law altogetber However,thus result is based on expenc=0 from only two states,Maryland with a S500,000 threshold and Oklahoma with a S600.000 threshold- ach leads us could 004 be find negative effects on earnings from lower thresholds in the S 100.000 to S400,000 tang cautious about this result. A conservative interpretauon of this tesulimay u�v prv�r�ssrvelly aegnuve�eOfTKu a mm.mum Impact on constriction earnings while thresbolda above on earnings. '9 Davis-Bacon Repeal Effects 56. The Standard Industrial Classification (SIC) code for construction consists of detailed categories of general contractors such as commercial and residential general contractors, detailed categories of heavy and highway contractors, and detailed categories of specialty subcontractors such as masonry and carpentry 57. Robert D. Reischauer, Congressional Budget Office Testimony, before the Subcommittee on Labor Standards, Occupational Health and Safety,Committee on Education and Labor. U.S. House of Representatives, May 4, 1993, p. 4. 58. Robert D. Reischauer,Congressional Budget Office Testimony before the Subcomrruttee on Labor Standards, Occupational Health and Safety,Committee on Education and Labor,U.S. House of Representatives, May 4, 1993. p. 4-5. 59. U.S. Department of Labor, Bureau of Statistics, Employment and Earnings. December, 1994, Table 3-3 for November 1994, p. 55. 60. U.S. Department of Commerce, Economics and Statistics Administration, Bureau of the Census, Government Finances. 1990-91. Series GFl91-5, Table 8, column 6, p. 9. 61. Some of the material in this chapter concerning the Utah use originally appeared in Hamid Azari-Rad, Peter Philips and Anne Yeagle, 'The Effects of the Repeal of Utah's Prevailing Wage Law on the Labor Market in Construction,' in Sheldon Friedman,et al.,eds.,Restoring the Promise of American Labor Law,Cornell University, ILR Press, Ithaca,New York, 1994, 207-22. 62.These data are based on quarwrly per capita dues contributions to the Utah AFL-CIO Building and Construction Trades Council.These payments underestimate union membership because of under-reporting of membership from participating locals as well as other exemptions and withdrawals of locals. 63. George F. will, 'It's time to repeal the Davis-Bacon Act',Deseret News, February 5, 1995. 64. Indeed, an unreported multiple linear regression model tested whether changes in the male black-white unemployment ratio could be associated with state repeals of prevailing wage laws or the fact that a state never had such a law. This model controlled for time fronds in the male black-to-white unemployment rauo, regional differences in unemployment ratios, and changes in the level of unemployment. While the morel found a strong time trend in the black-to-white unemployment ratio and significant regions! differences in the ratio, there was oo statistically significant relationship between either the repeal of prevailing wage laws or the complete absence of prevailing wage laws and the black-to-white unemployment rauo. In short,there is no stsustical connection one way or the other between the status of prevailing wage Laws and the mlauve unemployment of blacks and whites. 65. These state demographic unemployment rates are fiom U.S. Depa=ent of Labor,Bureau of Labor Stausucs, Geographic Profile of Employment and Unemployment, 1974 to 1992. 66. Data available on request. 67. Utah Department of Employment Security, Labor Market Information and Research,Annual Report of Labor Market Information, 1993, table 5, Salt Lake City, 1994. 68. Hamid Azah-Rad,Peter Philips and Anne Yeagle, 'Tie Effects of the Repeal of Utah's Prevailing wage Lair on the Labor Market in Construction,' in Sheldon Fnedman at al.,eds..Restoring the Promise of American Labor Law. Cornell Univeniry, ILR Press, 1994, 207-22. 69. U.S. Bureau of the Census. 1970 Crnrui of Population. 80 Davis-Bacon Repeal Effects 70. U.S. Bureau of the Census, 1970 Census of Population. 71. Figures 3.7 and 3.8 include all states for which any data are available. except California. Deiaware, the Dist.-1ct of Columbia, Hawaii, and Rhode Island — for which there are no Bureau of Apprenticeship Training data for the second penod. We exclude these states and the District of Columbia (for the same reason). 12. We do not know what accounts for the unusually high training rate for 'never-had' states in 1976. This anomaly disappears when average training rates by decades are compared. 73. This transformation into the log of an odds ratio meets the normality assumptions of Iinear regression analysis The technique used is generalized least-squares regression, with the regression weighted by the square root of (percent trained) times (one minus percent trained) times (state employment). 74. Latent illnesses resulting from exposure to toxic materials are responsible for an uncounted and thus undetermined additional number of injuries and illnesses — the costs of which are borne as reduced productivity, ruined lives for workers and their families,and burdens on workers'compensation and other social secunry systems. For a mix of reasons, there are no reliable estimates on the number of such illnesses. 75.C.Culver,M. Marshall,and C.Connolly,Construetion Accidents: The Wor*srs'Compensation Data Bast, 198J- 1988, Washington, DC, OSHA Office of Construcuon Engineenng, 1992. 76. In fipm 4.1, n refers to the number of observations in each state-law category. For instance, there were 230 state-year combinations for states that had prevailing wage laws throughout the period. 77. In the case of lost workdays per injury,the reported result is of the expected sign,but not statistically significant. 78. Jimmic Hinze, Indirect Costs ofConstrvetiorr Accidents. Seattle: The University of Washington, 1992, 14 79. Because of small numbers, there are no reliable estimates on how repeal would affect death rates. Thus, we cannot calculate the projected increase in fatalities due to repeal. If, however, they were to be affected at the same magnitude as are injuries, we would expect an increase of 130 to 150 fatalities per year. $0. Utah. pepart:sreat of Transportatwrt Final �*+-'-' Prooeaa.d for hymns, 19'ft3.74 doss publrsi+ea in 1923 and 19%report. 81. The saving ae so small because tabs cam on public worts an oaO raug* 23 pssmt of laud co+a. If you as tbow labor corn ley 10 Patna.you have cut toad Caere by only 2.3 perana. 82. coorp F. WiilL'trs silos to repeal the Dare-8rooa Ad',Dasaw N*.R Few 3. 1995. 83 Charles Culver,Michael Marshall,and Constance Conaolty.Construenon Accidents:The Worriurs'Canpensoeaw Data Base, 1985-1988. Office of Construction and Engmacr & OSHA, 1992. al Davis-Bacon Repeal Effects References Chapter III Azair-Rad, Hamid, Peter Philips, and Anne Yeagle, "The Effects of the Repeal of Utah's Prevailing Wage Law on the Labor Market in Construction" in Sheldon Friedman, et al., eds., Restoring the Promise of American Labor Law, Cornell University, ELR Press, Ithaca, New York, 1994, 207-22. Jones, S. Garth. 1979. Letter to the Editor. Deseret News, February 23, 1979. Thieblot, Armand J, Jr. 1986. Prevailing Wage Legislation. U.S. Bureau of Apprenticeship Training, State and National Apprenticeship Program Survey (SNAPS) 1975 to 1978, and Apprenticeship Registration Actions (AMS Report 18) 1987 0 1990. U.S., Bureau of the Census. 1970 Census of Population. U.S., Bureau of the Census. 1987 Census of Construction. U.S., Bureau of the Census, 1989-1991 Current Population Survey. U.S. Department of Labor, Geographical Profile of Employment and Unemployment, 1974-92. U.S., House of Representatives, 69th Congress, 2nd Session, H.R. 17069,"Hours of Labor and Wages on Public Works", Hearings Before the Committee on Labor, Feb 18, 1927. Utah, House of Representatives, House Audio Tapes, Feb 1979. 82 Davis-Bacon Repeal ElTecu f t > OPINION Cutting wages won't save tag money By DAN CURTiN lower than California was high-wage Mis on public construction would have to meet he very foundation of California's souri. the local wage standards,known as pre- economy is crumbling. There is little credible evidence sup- vailing wages,if they wanted to bid for the Once a model for the world,that porting the one-dimensional theory that work.it was called the Davis-Bacon Act, foundation—infrastructure investment lower wages equal lower costs.But there and it was authored by two Republicans —has failed to keep up with California's is growing evidence advancing the age-old and signed into law by a Republican presi- needs, wisdom of"you get what you pay for." dent. Worse still,while relying on infrastruc- Recent studies show that repeal of pre- If we repeal prevailing-wage laws,gov- ture—roads,bridges,schools,sewage vailing wages results in a rapid decline in ernment will end up rewarding not the systems and the like—largely built in the productivity,a substantial increase in best contractor,but the contractor that's 1960s,we have deferred maintenance to costly injuries and a virtual disappearance best at reducing wages to rock bottom and the point of critical danger.In a state of apprenticeship training resulting in a shifting health and pension costs to the known for earthquakes and floods,this is a less skilled work force.Reduced wages taxpayer. and increased use of out-of-area workers Government is the largest player in the problem,a big problem. g Pr H P Gov.Pete Wilson has proposed the also cause a serious drop in local and state construction industry,responsible for wrong answer.Avoiding the hard fiscal tax collection. more than 25 percent of all construction. choices that need to be made,the gover- Just after Christmas,Superior Court Without prevailing wages,competition nor has taken the simplistic approach and Judge William Cahill granted a preliminary will be based on wages,not productivity or called for lowering the wages of Cafifor- injunction blocking the governor's propos- quality and experience.You would quickly nia's skilled construction workers. With this action,Wilson has joined a chorus of congressional leaders,corporate CEOs and policy elites who,like Pavlov's _ dogs trained to salivate at the sound of a - bell,believe lower wages are the answer to all our economic woes. Using the smoke screen of California's budget crisis.Wilson and Republican lead- ers in the Legislature,including Poway Assemblyman Jan Goldsmith,have offered legislation with a medley of proposals al- tering the law that protects wages andr� benefits for California's construction _ workers. Their intent has been to lower wages. r Aft he national level,congressional Re- publican leaders were even more aggres- sive.They made several attempts to re- peal prevailing-wage laws altokether. Their efforts failed. Undaunted by these setbacks,Wilson : has recently proposed administrative changes to the state prevailing-wage law. Q.� •\, His administration claims this proposal will save taxpayer money by lowering wages as much as 20 percent.In the tine 1 \ pk`lnt,even they admit that savings from lower wages will only shave a few percent from total project costs. t But they're wrong.Taxpayer money won't be saved.If you cut 20 percent from wages,what you cut is the 20 percent of a construction worker's wages that goes to Health insurance,pension benefits and ap- prentice-training funds.These costs won't be saved,they'll be shifted. Public hospitals,already facing their own budget crisis,will see more patients, and we'll all pay higher taxes to foot the bill.Those who maintain health insurance will pay higher premiums to cover the un- insured A construction worker who re- >v�uszxaaisra tires without a pension ends up on welfare and other public assistance.We'll pay the ais.The judge determinedlhat the admin- see low-wage,out-of-area workers drive taxes. istration violated the state Administrative wages to rock bottom. There's another problem with the gov- Procedure Act by not basing its proposed San Diego,with its proximity to the eiitor's proposal.The law works just fine amendments on"adequate information border,is particularly susceptible to this as is.Not only has it protected local busi- concerning the need for,and cone- low-wage pressure.If government back- ilesses,it's proved that in construction, quences of,its proposed changes." pedals on wages,safety and quality,the decent wages can actually reduce costs. The administration also failed to assess rest of the industry will trample itself in a A four-year study in the late 1980s "the impact of the proposals on California stampede to the lowest possible stand- showed California construction workers business enterprises and individuals,in- ards. with decent pay and working conditions cluding the creation and elimination of If we dismantle prevailing-wage laws, built a mile of highway for less than$1 jobs." we'll succeed only in making one more million.A low wage,keeping Florida con- Over 60 years ago,our policy-makers blue-collar profession incapable of sup- struction workers on the ragged edge of had the right answer.They didn't look for porting a worker's family.We won't save poverty,resulted in a mile costing more the easy fix at the expense of working peo- taxpayer money,but we wig end up sacri- than$1.5 million. ple.When low-wage,itinerant contractors ficing construction quality.Our estab- In Virginia,another low-wage state,it's began to undercut local,established busi- fished contractors will either bow to the closer to$2 million per mile.The only ma- nesses,government decided it didn't want reduced wage and quality standards or go jor highway-construction state to build at to destroy local economies in the process out of business,This is a lesson we of building public infrastructure. learned,and oa government leaders un- CURTIN Is director of the California State To stop the decline in wages,a wage derstood.60 years ago.It would be a Council of Carpenters. floor was established.Contractors bidding shame to have to team it ad over again. The Icano ic attark ..on wor rust stop A j By OANIEL W CURTIN Y ,� U S AMMCA lost its sens2 ? ' Good a3 The attack on the Amen= ��+��++ ;worker maunues anaonted.Like -` wages are deltlsiam of a junkie.our economy g m e to has acceatea he seemiagiy addictiv g- .. not the } = iezhat lower*wages are good.For 30 •- "� years after Worid War II-amenca in ' 1 ctudee'wortting people in its ecoaamie problem � expansion.We were the envy of the 1 worid.No longer. Dame s12. Since the early`Sims.wages for worst- G'711'ZZ7Z ers have•been on steady decline.This is no suet.dsk any economist.ask any T dida all I 'en C tes OI worx e.. ?resf a ttai an - T costs won't be savecL the ll be shifted. -� es,�te. �o me problem with great ! � I - b empathy;and?resident Clanton success- Punlic aospimis will see more oaumts Fully rode the issue in 123I`.'et.wages I and we'll pay higher taxes to loot the bill cm=Ue to snae.Even college graduates Those of us who maintain health iasur-I - ! ums and hi ' e. ante:villa er uremr gh� are now seeing thel. wages decline. pay hfgn ?ic.+ang iLimseif of the canvas aver his ! er medical fees to cover the nmnsarrri-A presiaentai llmockout Gov.?ete Wilson carpenter that retires without a pension I as;usv Yven this iow-wage ti end a ; ends up on weifar.and other pundc as- I miteftii shove. He's prouosea new reguia sistance.We pay these taxes. raps to change ?revatiing wages" for or ay siashiug consamction woritere wag _onst<^sc^on workers buu=9 sc oois. governor all srmpiy maxe one j -.1--a`vays.hospitais and;aiis.;Then WVU more"51ue-collar"profession incapable .on hays c an a tie really :leans lower '- ;Mvaiiing .vases."Ali=the name"of oI SUIiDOrtiag 3 tvOrKPi"S fafatly.:t won't ;oc ernment etimenc. and:vita"an eye I save*atpayer money,or reduce the rfs- :owrard saving t=- avers• :coney."Sort-, ing cost of health car.or solve the Cr^rSiS I Torernor.lower wages won't save money is our schaois.or make a dent=the or:-axe government:Wore ef�ent n i mime problem or acing a moral comuass ccns.:acron.l�s sail and produc vir, ibacX.o our television and entertainment: _ tat::aunt. Mausz:7. cu_-ies snow twat a decent wage paid -ks:ae cuts worker wages. the governor :o California s construuction worriers offers a.5 aercent ncome tax cut:hat ouuas a;rule of�igawav for'.ess than 31 will p:.mariiy benefit he rrc ana the cv w i epnmo - T ::u.uion. ;o nee.Ii n 1oraa..ti ... wed mot;.a's a .ax c.f.they don't:.een. ,onsti^:ction workers on:tie rag;ea edge From_983 t!irougn 1989 Amer=^- of pover^-,:esutts:n a mile costing more chest`Jo.000 families increased their :ran 31. million.In Seconaar,-school weaita on average from S8 mfllian to cnnst;^acaon.California:vas m the iow million for a totai zicrease of aimost$2 est u'2ird of buiiding cost despite aigner trllion.This mcnease alone was more wages than most other states.mother j hap:he entire wealth of 60 amt=of stuav shows lower wages means more ac amer.:.'s famiiios-almost`a�.:l-0 cidents.more cost overrans.and a de- ( families.whose wreaith acttlailY aeciined cr ease h training,pardcniariy among in the same period.As if to throw sait in nfnonties.It concludes wet construction he wounds.some of amerlca's weaithi- cost savings are uncertmn.but reaticed est c:^y:ens have peen changmgtaeir citi tax revenues wom lower-paid constrsc- zensam to avoid paying taxes won workers are real. The construction inaustry is already There nave been two clear-es=to lean and hungr?. Competition is tierce. this:o•ee market sires song—lower wag Toaav's construction wage ads already es for vorkiag americans and vast been oushed beiow 1965 levels-improved creases for our millionaires-It is premse- tec nmogy has aramaticaily decreased ly this massive swift of weaith u oIn :he:work:ours needed on a job site. ; working americans to the tic ;hat has igrl^sidiled-veil=rained workers are become the defining:ssue of our age.its .afar:tie moaern construction maustr^ time:o stop eouatmg economic 3 eac7 ,.ees. Good.wages are not the problem wittt:ow wages ana layolfs.a s time m < in construction:good wages are the an- stop providing tax breAKS to people who swer. don"-:are wbai dag they fly on•. eir A cut in prevailing wages will not ;yachts-it's ithe to;member hat most guarantee the best contractor gets the Amer:: s worn for a living. Vl= L ;on.: will guarantee that the contractor ne"doing poorly,amerca's acing best at redumng wages,a inanaring bene- pooraY,in spite of what'Nail Strxt and dts and cutting corners will get the bid- the bond markets tell Lin::c.the_0 percent savings the rover- - - nor :rPdicm will be the 30 percent of a =I^enters wage package that goes into S- Cantei,m. G1mn is the ai e=of die, :lean`~:assurance and pensions These j Caaiomia State Counc:i ei Carpenters. 43 ; ^ 0 303 0 c m 5D ea0C3 y 4 o m �n � :� �✓� _ a E" L O C O y A q y y O y ' q O ".•. .+ •.q,.. cam. y cyo= off q 13 J y q Gr v V m C q L)to O C UC ^•• ^ y C y .+ .. U a = 3 > cJi o° >��_ �� U2to v oc 3 o .c q q a 3 .c .., m a q 3 m af.J� _ C _n3 i 3� = G ?+scc 5,.- > > v v = - y .0 _ c q v_•a _ q ^C;�-' U a J r q•' v q .'� ' v1 yx .y y .>, .:a - COO _ .- J _ , C_ .� CD .., m •7 0 � of � >v m cC ' a o W > O tw ej r � � i .� ,. _ - -�� Jam_. .� J •^ r; J 7�Q' `-'J _� � ,� � � - —_ - �_ � = =— , � � �� _ I � 3 Z I - ... n ;J cc = tD CIS .. - _ •U _ n U ,r J J n 2 — w � � +e :� _. ,3 y..O _ .' _ :� �.='� 7 .. - � :� � O ,n _ _ =3 � U n .. t'O;, + J J O - U _ •r. •r. _ CZ ZI) 71 I PON-" s — a 12 J — POW PNMEMO PON Oy 7 Uy —ca 71 _ - - v v , v 73 ca . 117R,�fllOft{0OORrlfla '^Mw),��• OPINION — The -Sacramento Bee Locally owned and edged for 138 years JAMES Mcr-LATCHY,editor, 1857-1883 C.K.McCLATCHY,editor.president 1885-1936 ELEANOR McCLATCHY,President 19-1&1978 WALTER P.JONES,edtor, 1936-197d C.K.McCLATCHY,editor, 1974-1989 GREGORY FAVRE,executive editar PETER SCHRAG,editorial page editor FRANK R.J.WH{iii AKER,prasidentand general manager AN OPPOSING VIEW - At the end of the day they go home tired.but proud., They've earned enough to take care of their families.in- Davis—Bacon law'_* ciuding good medical and dental benefits. Using their hands and their heads, they ve built some of the most sophisticated and durable structures in the world. Ev- ery construction worker I've ever known points with e ensures QuahLy special pride as he or she tells the kids,"I built that." hn SOLID,safe infrastructure has served businesa- construction and the 215 lic well. Government construction: aroundd �5 percent of all construction, has also, served the construction industry well. high quality cony-- I struction standards are enforced. Extensive training of I We rebuilt the uorld's busiest freeway in dust 66 apprentices, including growing numbers of minorities days. Part of our success came from contracting with and females. is required. Area wage standards, known. private firms - to speed rebuilding and to keep down as "prevailing wages."are mandatory.-In this way, bids. the cost. are won through efficiency and competence, not wage,- -Gov.Wilson,1995 State of the State Address cutting. Federal law, the Davis-Bacon Act, and state `. ` laws,called"Little Davis-Bacon'laws,have kept carenil Government workers - hand-in-icand with private _ I business - rebuilt Southern California's fractured .,. I control of a dog-ea d industry. dog That's why Pm distressed to see The See expressing freeways in record time and under budget. Support for those poiiticians who want to change all this' -President CZinion,1995 State of the Onion ('prevailing wage reform," April 17). They want to re- Address .., oeal these laws or tinker with the wage protections. I I say to them, lower wages don't mean lower costs. Lower. -; wages mean shoaav work, more accidents, more MJu- - f ries more repairs. What can and should be reformed are the burden= C.C. Ayers is president of C.C. flyers Inc.. a con i some regulations and the enormous paper work r� strucrion company in Rancho Cordova. j quired on government projects. Rather than cutting-- i wages,we should be looking for ways to put more people By C.C.flyers to work. One "endangered species," such as the fairy . T.1/t�DE me proud when the governor of California Shrimp, can stop an entire deveiopment project and the thousands of jobs that go with it.It's time to take a hard.. and the president of the united States publiciv rec- look at laws such as the Endangered Species Act. Its ognized the tremendous success we had in rebuild- time for government to stand up for hard-working, prit w after the Northn in the Santa Monica -Freeway g y � auctive Americans. earthquake.We just repeated that performance rebuild- ing in 21 days the twin bridges wiped out on I-5 after' HAVEN'T built my company and my reputation us- the recent floods. I ing low-wage,unskilled workers. We come in ahead- Like air and water, the highways, dams, bridges,- of schedule and under-budget because we use the schools, sewage systems and other infrastructure ele•, best skilled workers and run the job with a talented ments that make society run are easy to take for grans- management team.'That kind of operation doesn't come ed-until they're missing. When commerce stops, pay-. cheap, but investing in skills is what brings the overall` checks stop,families go hungry. cost of construction down. Those of us in the construction industry know what it' takes to build a first-class infrastructure: government I witness close up every day the core values that have-- commitment, quality contractors and a skilled, hard made America great: hard work and the dignity that.-- working labor force. Each member of my crew, from top. comes with it; individual responsibility combined witli:.. management to the newest apprentice,has a job to do in. teamwork; and most of all, a tierce love of family and a tough, dirty and dangerous occupation. These are country. These are the people who work for me. They skilled,Wigged individuals who take great pride in their work hard, they're treated right, and they get the job work. Manv of them have devoted more than 5.00G- M done.Why would anyone want to screw thas�p. hours to learning their trade. They put in a fu1I day's work and they get a decent day's pay. I vi '� � 'c^� m_.�_ ; .non � � c U?^' �"• � � c o .L �• � �N� '� 3� � m `� v o mum'" 3 o� H o.: �— ci y a+12 C .. -• _ a (n V Oy0 C may. VI i7'O_.t OG ca L1 -• '�7 -� _y� N N �� c > a cs :J_- �� o v_,aj iO �g c � ^ o ms v Jc� � s ,uy � -S0E'— o`� N3 V� > • °y°=D - — n C `�F m u _ y' r� u ° u o > Y p^ 5 1L Vol i i 3 C C G' 3 V z S u r <22 N ZZC _ 7 V— 'p F, N N 1, Q V O s 14 .13 Y r 0 7 u o ' ff � U U �_ J J` V1 I � � W � � � � •� � �Y J �^U V C.L J 14 13 zz- - I ' � _ � � ==" ram— a 3 ; 3 = 3 _ � j� •- _ __ � � � y�•�� V� � p 1 � � ar it r 3 � V � .J u — � �— r O S • OpC y V7 f� T CJ N a7 rJ y — y =—':� — — eft :7•— — r V J= O a r= — i t'ai O y C7 p ? '>J p'^•'J O 7.J O u y ra J - ° y O."S. v=f"" O =. °. 3�v� ' I �/ oD N� Q^J y DO '��r � T� 00 y.p y ,� y y h •" (J ^ C = .7 O= p >� 'u .V N _ •• �... LZ to Q 1= � 13 L 'n.o -, :� �" y .s,nr n >,•= N ` 3 T°• :.7 C y O ,J.,.� .3 j y J y.O = "'J >lz CJ� /( 7 (f �__ •- `J^D �.�. .JJ � ��•c.�9"''._' y Cam. �a O^�.a C"''_+'^� .". J y� � c ._ U �� J c 3 y � Jam, c tj c v ad W c.,, 0 y -� C y .�rG L 03.Z .9 6).�^. > CL n• J v u' mF;C of 9, d /i y ., •� :,n � � n - :� -� 7 w ^ y c y n nZ� nJ id pr 31 -73 \�( t� zd � ,',J,a n 3'� �_Cr � c � �'�=. -r - >,n =•n = � �:� Q � Z .� ;"'-. � �-7 fin`- � � _;•'.� � �._ c„' '� - _ �3 > r n 73 Sfti� Z}C/�YN/� i► Tuesday,June 14,1994 A-15 DAHIE� M. CEN Pending, a law to-end `replacement' of strikers Saoamenro To argue that being"perma- HE WORLD cheered nently replaced"is different than when Polish workers, being fired makes a mockery of the striking illegally, right to strike. brought down an anti- An employer can still hire re- democratic communist placements.they just can't be"per- government. manent."In other words. an em- It was courageous Ukrainian plover can't fire his striking em- and Russian mine workers.again ployees,When a strike is settled. striking illegally,who hastened the win.lose or draw.a worker can go demise of the Soviet empire. back to work-Families aren't de- What a tragic irony that today, strayed;communities aren't npped in 1994, American workers find apart. themselves fighting for them own How cynical that some politi- cians who loudly proclaimed their n,mts.The National Labor Reia- P , lions Act, enacted in 1937, out- support for Polish and Russian lawed the firing of striking work- workers. like Senate Republican 1 era. Going further,it stated that leaders Bob Dole and Orrin Hatch. nothing in the act"shail be con- are just as loudly opposing those strued as to interfere with or im- same rights for American workers. pede in any way the right to Senators Doie and Hatch are strike." out to kilt 555. Since there are That's the way it worked for more than enough votes committed more than 40 years.Worxers didn't to pass 955 in the Senate, they become millionaires, employers have threatened to filibuster the didn't go bankrupt. Living stan- bill. cards rose gradually, and society Person. the I progressed. Everyoodv shared the OR A working ,p (, benefits. But the get-it-while-wu- -fight to withhold one's labor. I can.greedy,me-first decade of the freedom of speech and the 1 1980s changed everything. rights of association and assembly differ- Along with business deregtuia- are what make a free socze ry , tion, hostile takeovers, golden ent from dictatorships•communist parachutes, leveraged buyouts, or otherwise.Just ask the Polish, 1 junk bonds and S&L scandals.em- Russian or Ukrainian workers. plovers abused a iittle-known loop- These associations provide protec- hoie in the law.They realized that tion from oppressive government while it was iilegai to fire striking and a needed balance against em- 4 workers, it was perfectly legal to plover abuse.We call them unions. "5ermanently repiace"them. Its no comcideno that the de- To restore balance to labor- cline in union membership is management relations.Senate bill matched by the decline in real 55 was introduced.S55.known as wages and standard of living. the Workplace Fairness Act and al- Its nw coincidence that the dis- ready approved by the House of integration of family values in Representatives,would make it il- America mirrors the loss of mil- legal for employers to"permanent- lions of middle-class jobs that paid Iv reviace"workers engaged in a decent wages.Twenty percent of lawful economic strike or to give se- Americas full-time employees are nionty to repiacement workers paid less than poverty-levei wages. hired during a lawful strike. The Does anyone nmiv believe that a law will only apply to unionized worker,particruiariy in today's dif- companies, ficult economy,lightly decides to go on strike? Passage of the Workplace Fair- Dmud A Cwtm 13 davetor of the new Act will begin to restore the Caii{+ornia State Council of Carper- balance that went haywire m the te Mos. Prevailing Wage e Laws and the California Economy Michael Reich, Ph.D. Professor Institute of Industrial Relations University of California Berkeley, CA 94720 510-643-7079 mreich@econ.berkeley.edu February 1996 Michael Reich is a Professor of Economics at the University of California at Berkeley, where he has served on the faculty since 1974, as well as Research Director of the National Center for the Workplace. He holds a B.A. in Mathematics from Swarthmore College and a Ph. D. in Economics from HarvardUniversity, where he studied labor economics with John T. Dunlop, former United States Secretary of Labor. Reich has published eight books and monographs and over fifty articles, including studies of racial inequality and of segmented labor markets in the U.S. A 1985 ranking of the best young economists in the U.S. placed him at no. 33. Reich's recent research has focussed on the causes and consequences of high performance workplaces, with a particular interest on the development of worker skills and rewards in Japanese and U.S. employment systems. He has recently completed a book-length manuscript on this topic (co-authored with his Berkeley colleagues Clair Brown and Lloyd Ulman). Reich served from 1986 to 1994 as the Editor of the scholarly journal Industrial Relations and recently finished a term as Associate Director of the Institute of Industrial Relations at UC Berkeley. He is a member of the American Economic Association and the Industrial Relations Research Association. Michael Reich Prevailing Wage Laws and the California Economy Page 1 1. Introduction -•- - of prevailing The Department of Industrial Relations proposes to change the computationp g wages from a modal to a weighted average methodology and to eliminate the payment of predetermined wages (i.e., the so-called "double asterisk"). DIR claims that such a change will reduce labor costs on the state's public works projects by 20 percent, which translates into annual savings of $200 million for the state and state agencies, and DIR contends further that v persons or businesses.' The rim this change will not cause adverse effects to private p primary rationale for the change is to lower labor costs, from union wage levels, to reflect the lower wage rates paid on nonunion private construction projects. A secondary rationale is to reduce the cost of determining prevailing wages by aligning state methodology with Federal methodology and making use of federal surveys.2 DIR cites numerous academic studies, reports of the Legislative Analyst, and its own studies to support its claims. Despite these efforts, the DIR analysis is seriously flawed, drawing upon incorrect data and ignoring important costs that will be generated if the proposal is implemented. I discuss here: errors in DIR's estimates of savings on labor costs, implied but unrecognized increases in survey data costs for the state, reduced tax revenues, disparate impacts upon minorities, effects upon productivity of labor on public works that offset reduced wages, and undiscussed costs upon health and safety, training, benefits and pension fund safety, as well as training and skill formation. I conclude that DIR has not presented a coherent and careful analysis of the effects of the proposed rulemaking change and that it omits many significant costs. At present, the California economy is undergoing a healthy recovery, and the construction sector, which is vital for private and public investment, is recovering as well. The DIR proposal requires considerable 1 a key segment of the economic recovery at risk. further anal analysis because it aces ry y p y g 2. Errors in DIR's cost-saving estimates Problems with DIR's calculations: The DIR cost saving figure of $200 million is based primarily upon an estimate of a 20 percent decline in the prevailing wage, which is in turn based upon data on modal and average wages in employer surveys that DIR conducted in three rural California counties in the late ' DIR Memorandum from Dorothy Vuksich to Director Lloyd Aubry, November 21, 1995, Rulemaking file, Appendix A, page 2. 2 "The amended regulation will make it easier to adopt rates determined for federal public works projects when appropriate. This will reduce duplication of effort in the area of wage investigations and thereby reduce the burden upon the regulated public." DIR, Initial Statement of Reasons, p. 3. Michael Reich Prevailing Wage Laws and the California Economy Page 2 1980s. In reviewing these survey results, I find that the .method used by DIR to compute the percentage wage changes of switching from a modal to a weighted average methodology contains a major error, and produces results that violate elementary-school arithmetic. DIR's error can be seen on pp. 10, 12 and 14 of the rulemaking file, Appendix A, which display survey results for the three counties: Amador, Nevada and Tulare counties respectively. On the Amador County page, the largest percentage difference between mode and mean of any of the crafts is 21.6 percent, for laborers, and only one other craft has a difference in double- digits. Clearly, the overall weighted average for all crafts will be much less than 21.6. Yet the method of calculation used by DIR produces an average 28.79 percent difference! The DIR error is introduced when the weighted average is computed because the weights used in the calculation appear in percentages that add to 100, rather than as decimals that add to 1.00, and no correction is made in the final figure. As a result, the reported weighted average difference is 100 times larger than what would be obtained with a correct arithmetic approach. Eliminating this error reduces the estimated wage change to essentially zero and erases the purported savings claimed by DIR. But this is not the only problem with DIR's calculation. Examination of the survey data indicates that they contain severe shortcomings. My examination of the findings from these surveys indicate internal inconsistencies and lack of representativeness. The inconsistencies include numerous examples in individual crafts and counties in which the mean basic hourly wage rate was higher than the mode.' Yet modal rates are generally thought to be union rates, which generally are higher than nonunion rates. These inconsistencies cast doubt upon the reliability of these data. Whether wage reductions would occur depends upon having reliable and representative survey data. DIR does recognize that the results from these three counties will not be ' representative of more populated areas of the state. Yet it persists in concluding that wage rates will fall overall by 20 percent (DIR, p. 4) to estimate a cost saving of$200 million. It then uses ' In the Amador County August 1987 survey of 23 craft categories, the mean was higher than the mode in 7 crafts and equal to the mode in another 4. The sample size was less than 10 in 8 of the 23 crafts, suggesting high standard errors for the estimates. Although means and modes are displayed, standard error statistics are not provided, so that we cannot know whether mean-mode differences are statistically significant. Inconsistencies between the mode and the mean were also present in cells with large sample sizes (DIR Rulemaking Appendix A, p. 11). In the Nevada County August 1987 survey of 18 craft categories, only 6 have the expected negative wage change and 11 have sample sizes of 10 or less (DIR Rulemaking Appendix A, p. 13). In the Tulare County January 1988 survey of 19 craft categories, only 9 have the expected negative wage change and sample sizes are 10 or less in 7 categories. Among 124 surveyed carpenters, the mean was 49.9 percent higher than the mode. Altogether, over 33 of the 60 wage change estimates in these three surveys do not have the expected sign. Michael Reich Prevailing Wage Laws and the California Economy Page 3 this estimate again, but recognizes that it is unrepresentatively high.4 Because most other estimates are under 10 percent, DIR arbitrarily selects a "midpoint" estimate of 10 percent. DIR then tries another gimmick. Drawing upon a short-run labor demand elasticity of .5, DIR implicitly critiques its own 20 percent savings figure by arguing that a cut in wages of 20 percent would cause construction employers to increase their labor requirements by 10 percent, thereby eliminating half of the purported savings.' DIR then arbitrarily combines an excessive labor cost proportion figure (of 33 percent, see below) to this wage reduction estimate to again a savings compute in s of $160 million, which it suggests provides a lower-bound. DIR goes on to P g congratulate itself for using two very different methods and yet obtaining very similar estimates. This is illogical. A lower-bound estimate of cost savings should be zero, according to the survey estimates discussed above (as well as the careful study by Prus, 1996, which disentangles public and private construction cost figures). A more reasonable short-run direct public cost savings estimate can be derived by using the conventional 22 percent labor cost ratio and an 8 percent wage reduction estimated by Philips (1996; see below). This estimate would come to $40 million. Even this lower estimate has to be modified to account for collateral state tax revenue losses, increased burdens upon public hospitals for those who lose health-care coverage and additional adverse impacts upon the private workforce and businesses.' DIR clearly needs to improve its own data collection and analysis procedures, as well as its cost estimates, so that it can accurately determine the effects of its proposals. Comparisons with Federal prevailing wage rates: DIR's surveys and studies indicate that Federal prevailing wages already are very close to state prevailing wages as determined by the current modal method.' These comparisons 4 The State's Legislative Analyst very recently criticized DIR's estimate of cost savings noting that "while there would be savings resulting from the proposed change, the DIR estimate overstates the magnitude of the potential savings because it extrapolates data from three rural [and far less unionized] counties . . . ." Hill, Analysis of the 1996-97 Budget, Bill. ' Philips et al. (1995) use this .5 estimate to calculate the net impact of prevailing wage laws on total wages. For recent research in this area, see Hamermesh, 1993; Borjas, 1996 for surveys of the literature. ' Longer-run estimates of cost savings will be much smaller than short-run effects as productivity falls because of declines in the supply of skilled construction workers. See below. ' DIR Memorandum from Maria Robbins to Lloyd Aubry and R.W. Stranberg, May 7, 1993, page 635 in the rulemaking file and Attachment A, Comparison of Federal/State Published Rates, pp. 639-58. The arithmetic errors described above are present in the tables here as well, but because most entries are close to or equal to zero, are not consequential. Michael Reich Prevailing Wage Laws and the California Economy Page 4 suggest that wage reductions are likely to be significantly less than the 20 percent figure discussed earlier. Labor costs constitute a declining share of total construction costs: Cost savings from a change in prevailing wage methodology depend not only upon wage rates, but also upon the ratio of labor costs to total costs. Although DIR states that a ratio of 22 percent is standard, the DIR document employs a much higher and arbitrary ratio of 33 percent.' The 22 percent figure is based upon a Federal census of construction in 1987. Projections into future years should recognize that the labor cost proportion has been falling, n partly us partly because annual real wage increases in construction have been negative, and pa t y because labor productivity growth in construction has been positive.' Consequently, DIR's estimates of cost savings are likely to be overstated for future years. How much will construction waves fall? A recent analysis of construction wage rates in different states with alternative policy regimes on prevailing wage laws indicates that the shift to a weighted average methodology is likely to reduce the annual wages of California construction workers by about 8 percent (Philips, 1996). This study uses microdata from the 1990 Census, holds constant many individual characteristics as well as region of the country, and obtains statistically significant results. Remarkably, Philips found that the wage reduction spilled over to all construction P g P — workers in the state, not just those on publicly financed projects. This result indicates both that the cost of lower earnings to private workers exceeds direct cost savings to the state and that revenue losses because of lower sales and income tax collections may be considerable. The revenue implications are discussed further in section 4 below. 3. Additional data collection costs generated by the rulemaking change DIR's approach to new data requirements is internally inconsistent. One the one hand, DIR claims that switching to the weighted average will bring Federal and state methodologies in line with each other and eliminate the state's need to duplicate surveys already carried out by the Federal government. On the other hand, DIR also recognizes that Federal methodology $ DIR cites an estimate of one-third mentioned in testimony by Bernard Anderson, Assistant Secretary of Labor. However, Anderson's estimate is not supported by the Department of Labor's own data. ' See BLS 1981, Chart 6, for evidence of long-term trends of declining onsite labor requirements for new construction, disaggregated for 9 types of construction projects. The value of capital goods per worker in construction grew by 130 percent (in constant dollars) between 1950 and 1983, as reported in Table 900 of the Statistical Abstract of the United States 1986. Michael Reich Prevailing Wage Laws and the California Economy Page 5 actually consists of using union rates from collective bargaining agreements (CBAs) and that the state would need to conduct many expensive surveys to implement its new approach. Although the effects of a switch to a weighted average methodology on prevailing wage determinations remains unresolved by DIR's own data, there is no doubt that the switch will generate substantial new costs to conduct the numerous periodic wage surveys called for by the new method.'o The required surveys need to be representative of wages and benefits for comparably skilled workers in each construction occupation category and local economic area. About 4,000 different determinations must be obtained every six months and sample sizes of actual respondents in each of the state's 58 counties must be large enough so that each survey renders statistically significant results. Household-based surveys are more costly than the current approach of employer-based surveys but may be much more reliable because employers face payroll-tax related incentives to under-report wages, benefits and employment. The annual budget to conduct a large household survey and analyze its findings would easily be many millions of dollars and is beyond the state's current in-house capabilities. An advantage of the current modal system is that it does not require large-scale surveys. Written contracts, when they exist, provide information that is both reliable and much less costly to obtain. Collective bargaining agreements provide much of the database for the current modal methodology. 4. Impact on tax revenues I How much might state tax revenue fall? The major components are changes in sales tax revenue and income tax revenue. A reduction of 20 percent in wage rates for public works (DIR's estimate) is likely to spill over onto the remainder of the construction labor market, causing a reduction of construction workers' average earnings, from about $28,000 in 1995 to 80 percent of that figure, or$21,600. In this range, construction workers would fall into a lower state income tax bracket, in most cases falling by 2 percent. The result is a reduced tax liability of $432 per worker up to $21,600 of earnings, and a reduced tax liability of 6 percent (the average tax rate) on the earnings loss, or $384. Adding these two figures together, we get a 19 5 figure of 512 000 construction jobs in revenue loss of $816. Using the November9J '0 "Should the State modify the definition of prevailing rate from a modal rate to a weighted average, statewide-all-craft wage surveys would be necessary. Adopting the Federal rates would not fulfill the requirements of a weighted average definition... In the absence of a protest, Federal Davis-Bacon rates are set by adopting CBA rates." DIR Memorandum from Maria Robbins to Lloyd Aubry and R.W. Stranberg, May 7, 1993, page 635 in the rulemaking file. Michael Reich Prevailing Wage Laws and the California Economy Page 6 California yields a reduction in revenue of just under $418.million." The reduction in sales tax receipts is likely to be nearly as large.12 A drop in wage rates of 8 percent (Philips' estimate) will cause a smaller reduction in tax revenue, approximately $167 million less in income tax receipts. Against this loss there is a gain in tax revenue created by the earnings from the additional employers' hires of 4 percent (using a demand elasticity of .5 as discussed earlier). This effect restores about $16 million of the lost revenue, leaving a net loss of over $150 million in income tax receipts and again a similar amount in sales and use tax receipts.13 An additional consideration applies to these calculations. The proposed rules changes will adversely affect union pension funds (see the discussion in section 7 below) and consequently investment in state construction bonds by union pension funds will also decrease. Innovative programs to invest union pension fund assets directly in housing and other sectors will also be affected. 5. Impacts upon minorities How might the representation of minorities in construction be affected? Currently, blacks and Hispanics constitute 10.7 and 29.0 percent of registered apprentices, respectively, compared to 7 percent and 26 percent of the state population.14 The black representation has been constant in recent years while the proportion of Hispanics has been increasing steadily. Using a different dataset, Bilginsoy (see section 8 below) finds that minorities comprise 21 percent of new apprentices nationwide. A cross-tabulation of California construction apprentices and race/ethnicity status, taken from the 1990 Census Public Use Microdata Sample (PUMS) and conducted by the present " This estimate must additionally be modified to reflect the reduction in adjusted gross income, which will differ from the reduction in total income. The order of magnitude will remain the same, however. 12 In 1993-94 state sales and use tax revenues amounted to $14.8 billion, while personal income tax revenues (including those on unearned income) amounted to $17.6 billion. Workers receiving construction-level wages do very little saving out of earnings. Putting all these factors together suggests that sales and income tax effects will be broadly similar. Economic Report of the Governor 1995, Table 36. 13 Belman and Voos (1996) calculate sales and income tax revenue losses if Wisconsin were to repeal prevailing wage laws and also find that revenue losses substantially exceed cost savings, creating net adverse effects upon the state budget. 14 Statistical Abstract of California 1995, Table C-16. Michael Reich Prevailing Wage Laws and the California Economy Page 7 author indicates similar findings: blacks constituted 9.5 percent of construction apprentices in 1990, while Hispanics constituted 25.3 percent. (See attached tables.) These findings, together with adverse effects of the rule change upon apprenticeships, indicate substantial adverse employment impacts upon the minority population. The earnings impact will also be adverse, since the earnings of black journeymen in California in 1990 averaged $20,900, compared to $16,500 for all employed blacks; the comparable figures for Hispanics are$17,570 and $13,500. (Computed from California PUMS extracts and the Statistical Abstract). 6. Productivity effects ' documents reco nize that the skill efficiency and quality of the work DIR s internal Y q Y g performed by construction workers play a crucial role in determining actual labor costs on a given construction project.15 Yet DIR makes no effort to incorporate these factors into its analysis. How much could productivity fall? The only careful studies are Allen's examination of union-nonunion productivity differences, based upon data from the early 1970s to the late 1980s. (Allen, 1994; 1988; 1994). Allen finds union-nonunion productivity differentials when prevailing wages do not apply, with productivity differentials of about 20 percent offsetting the 20 percent union-nonunion wage differential in constriction. In other words, paying higher wages does not translate into higher labor costs, as higher-paid union workers are more trained (more on this below) and therefore better skilled, and because employers adjust by better managing their workforce and by utilizing more capital equipment per worker.16 Similar evidence arises from interstate comparisons of highway building costs conducted by this author.States with the highest prevailing wages experience dollar costs per mile that are generally comparable to those in states with no prevailing wage coverage and the lowest wages. These results hold even in a multivariate regression equation that controls for the rate of urbanization in the state and the proportion of bridge or overpass miles in total highway miles.17 Although such results may appear surprising to some, they are predicted by economic theory." Indeed, recent developments in economic theory and evidence have discovered 15 " ..unless other factors such as skill, efficiency and overall quality of construction are held constant, total labor costs may not be reduced." Ibid. 16 Allen (1994) also suggests that union-nonunion productivity differentials narrowed during the 1980s; a prudent estimate today would be about 10 percent. 17 More precisely, the coefficient on the log wage is positive and statistically significant in an equation in which the dependent variable is total miles constructed per worker. " Contemporary economic theory recognizes that a competitive economy can travel on a high-wage, high-productivity path (or equilibrium), or on a low-wage, low-productivity path. Michael Reich Prevailing Wage Laws and the California Economy Page 8 significant complem en tari ties between capital equipment and-skilled labor but not between capital equipment and unskilled labor. Economic theory also predicts that training will not be provided by employers in the absence of long-run employment relations. These findings suggest that institutional mechanisms that provide a long-run supply of higher-paid but also higher-skilled labor provide benefits to the economy as a whole. A simple example would be craft workers whose skills are utilized by employers in other industries when seasonal or cyclical downturns occur in construction. 7. Health care coverage and pension benefits The DIR statement of reasons for changing the method of determining prevailing wages does not discuss how employee benefits -- primarily health care coverage and pension plan contributions -- will be affected. Yet, as Petersen (1996) shows, the impact on this important component of total compensation will be both certain and radical: virtually all workers on prevailing wage projects will lose much of their private health and retirement benefits, and a very large proportion will lose them entirely. The effects upon the state's already-overburdened system of county hospitals and upon future retirees will be dramatic. At present, virtually all unionized construction workers receive health and retirement benefits worth approximately $4.40 per hour, while only 18.5 percent of nonunion construction workers receive any benefits, and those who do receive much less than $4.40. In 1992, collectively-bargained funds paid health and welfare benefits of $1.15 billion and pension benefits of $1.13 billion, compared to $233 million in benefits paid by nonunion construction funds in California." The sharp contrast between the two sectors arises from the short-run character of construction employment as well as the relatively small size of most construction companies. The high burden upon of providing benefits faced by individual employers means that multi employer programs are necessary to collect and pool insurance and retirement funds. In the U.S., such multi-employer plans almost always arise through the process of collective bargaining and are often administered under union auspices. The new rules would eliminate benefits except when unionized workers constitute the overwhelming majority of a locality's construction workforce. But union workers constitute 25 The path that is followed depends to a great extent upon public policy. See, g.Z., Brown and Reich, 1989; Brown, Reich and Stern, 1993; Brown and Reich, 1994; Brown and Reich, 1995. 19 These figures are from Petersen (1996), p. 8 and are based upon Form 5500 data provided to the Pension and Welfare Benefits Administration of the U.S. Department of Labor. Michael Reich Prevailing Wage Laws and the California Economy Page 9 to 30 percent of statewide construction employment.2° Even when unionized workers predominate, a weighted average calculation will probably lower the value of their estimated benefits in excess of$1.00 per hour. As a conservative estimate, assume that unionized workers account for half of the prevailing worker hours in a given year, and that half of those workers lose their benefits entirely while the remainder lose an average of $1.00 per hour. The total would then be approximately $160 million, but at least half and more likely three-fourths of this amount represents health care costs that would be shifted to the public sector. 8. Training and skill formation How much will training fall? Using state data, I calculate that the construction industry accounts for 70.8 percent of all the registered apprentices in California." The total number of apprentices declined from 49,484 in 1990 to 35,542 in 1993, and then increased to 38,768 by the end of 1994 (ibid. . Nearly all of these apprentices are unionized. Bilginsoy (1996) draws upon national data on apprenticeships supplied by the Bureau of Apprenticeship and Training and uses a methodology similar to that of Philips (1996). This study is the first to employ microdata on the characteristics of individual apprentices and classification of state prevailing-wage regimes. Comparing states with modal regimes to those with weighted average regimes, and statistically holding all other variables constant, Bilginsoy finds a reduction of apprenticeships of 13 percent, an increase in dropout rates from apprentice programs of 18 percent, and larger changes for minorities and women. These findings suggest that long-run skill formation will drop significantly. These effects are troubling because they represent future declines in skills of California workers. In the short run, California construction can draw upon the substantial stock of highly 20 Union-provided figures on the number of workers paying into benefit funds, well over 30 percent of construction employment,are substantially larger than the number of union workers found through the Current Population Survey, just under 25 percent. Both numbers may be accurate because they refer to different time periods. The CPS data are based upon monthly averages, whereas benefit fund contributions derive from annual data. Over the course of a year an average individual worker may move in and out of the construction industry many times and on average will not be counted in each of the monthly surveys, but will appear in the funds' data. Recent research in labor economics has discovered surprisingly large flows in and out of jobs, relative to total employment. Equally important, calculations of unionization rates should properly define the boundaries of the industry. Many of the workers in residential construction, which is predominantly nonunion, do not have the skills needed in large-scale public works projects and should be excluded from the calculation. " California Statistical Abstract 1995, Table C-15. Michael Reich Prevailing Wage Laws and the California Economy Page 10 skilled journeymen in the state, few of whom are likely to-emigrate to other areas. Neighboring states that have repealed state Davis-Bacon laws, such as Utah, continue to be able to draw upon the products of California training programs, employing California workers on short-run contracts. This "free riding" has occurred in states in other regions of the U.S. because their neighbors generate a nearby skilled pool of construction labor. No such possibility is available for California, and even experienced and unemployed construction workers in Mexico do not obtain the requisite skills. Many economists have pointed out that real earnings levels have been falling for the 80 percent of the U.S. population that do not graduate from a four-year college program, and that the U.S. economy requires higher levels of work-based training to reverse this trend.22 Although some nonunion employers provide a structured training program, including through multi-employer associations, these do not and will not substitute in quantity or quality for current levels of training. The demise of construction apprentice programs that is likely to occur will make a bad situation even worse. 9. Workplace safety Construction injury rates remain above those for other industries. Using a methodology similar to his colleagues Philips and Bilginsoy, Waitzman (1996) has calculated that the proposed switch in prevailing wage methodology will increase injuries by 8 percent and severe injuries by 5 percent. These increases result from less attention to safety and from reduced enforcement of health and safety laws. Weil (1992), for example, found that unionized worksites are three times more likely to call in safety inspectors than comparable nonunion worksites. The increased injuries, and their attendant human and economic costs (including higher workers compensation premiums) should be considered in making policy changes. 10. Conclusion: Maintaining the economic recovery in California The California economy is currently undergoing a vigorous economic recovery, led in by the expansion of business fixed investment in structures and equipment and by increased demand from the rest of the nation for the products of California businesses.23 The recovery is apparent in the growth of gross state product and total employment. It is less apparent in the state unemployment rate, which continues to be one of the highest in the nation, and in continued declines in real earnings for most workers. The economic recovery in California began later than in the rest of the nation because of the restructuring required by reductions in the defense/aerospace sector, closings of military bases, overbuilding of commercial and industrial structures in the 1980s and the drag of high zz See, for example, the publications of the National Center for the Workplace. zs Council of Economic Advisers 1995; Economic Report of the Governor 1995. Michael Reich Prevailing Wage Laws and the California Economy Page 11 housing prices. But just as the recovery began later than in-the rest of the nation, it is expected to end later as well.24 Since most forecasts suggest that the national economic recovery will not end soon, we can expect that California's will continue as well. The California construction industry is growing along with the rest of the state economy. Construction employment grew 4.4 percent between 1993 and 1994 and was forecast to grow 5.9 percent in 1995 and 6.1 percent in 1996 (Economic Report of the Governor 1995, Table III- November 1995 has been measured as 6.5 percent 4 . Actual growth from November 1994 to g (BNA, Daily Labor Report, February 1, 1996). Employment in heavy construction -- mainly public works and utilities -- grew by 2.6 percent in 1994. The California recovery has generated additional demands to maintain and expand the public infrastructure of the state, which were already strained by increases in seismic upgrading requirements and by increased prison construction. The recent quadrupling in Caltrans' estimates of the cost of retrofitting the bridges of the Bay Area will generate still more fiscal demands and underscores the need for maintaining and replenishing the state's skilled construction workforce. Yet Federal funds for Federal public works or for matching state projects are not expanding, and although the State's fiscal capacity to support public works is growing, competition to,allocate to funds in other areas is also growing. In this environment, it is understandable that the State state g g should seek to reduce costs in every area, and that taxpayers and consumers of state-provided goods would endorse supposed "reform" efforts that promise to save public funds. In so doing, however, public policy should not reduce the capacity of public works spending to provide the infrastructure upon which continued economic growth depends. Reforms also should not generate hidden or indirect costs upon the public or the public purse that would eliminate projected "savings" and injure the health of the state's economy. Public sector construction can continue to support economic growth only if institutional features specific to the construction industry labor market remain in place so that it can perform efficiently. These institutions involve the continuing ready availability of a supply of skilled construction labor and of efficient means to match workers and jobs. Prevailing wage laws, union hiring halls and structured apprenticeship training have constituted critical components of these institutions, and more so in California than in many other states in the nation. I While the dismantling of these institutions in other states has been shown to have moderately harmful effects, the impacts in California could be much larger. The size of the California economy relative to its neighbors within the U.S., the large geographic distances to sources of substitutable labor within the U. S., and the small distance from the Mexican labor market represent features that are not shared by other states and which put California at greater risk. It is not prudent to ignore the economic risks that would be generated by restructuring prevailing wage laws in California in the manner now suggested by DIR. 2' The Economic Report of the Governor forecasts personal income growth in 1996 in the 5-6 percent range, which is 1-2 percent over the forecast for the U.S. Michael Reich Prevailing Wage Laws and the California Economy Page 12 References Allen, Steven G. 1984. "Unionized Production Workers are More Productive," Quarterly Journal of Economics, 99, 2: 251-74. Allen, Steven G. 1988. "Declining Unionization in Construction: the Facts and the Reasons," Industrial and Labor Relations Review, 41, 3: 343-59. Allen, Steven G. 1994. "Developments in Collective Bargaining in Construction in the 1980s and the 1990s." Pp. 411-66 in Paula Voos, ed. Contemporary Collective Bargaining in the Private Sector. Industrial Relations Research Association Series. Belman, Dale and Paula Voos 1996. "Prevailing Wage Laws in Construction: the Costs of Repeal to Wisconsin." Working Paper, The Institute for Wisconsin's Future. Brown, Clair and Michael Reich 1989. "When Does Union-Management Cooperation Work? A Look at NUMMI and GM-Van Nuys." California Management Review, 31,-4:26-44. Brown, Clair, Michael Reich and David Stern 1993. "Becoming a High-Performance Work Organization: the Role of Security, Employee Involvement and Training." International Journal of Human Resource Management, 4, 2:247-75. Brown, Clair, Michael Reich and David Stern 1994. "Training Structures, Skill Formation and Wage Profiles in Japan and the U.S." Industrial Relations Research Association, Proceedings of the 48th Annual Meetings. Brown, Clair and Michael Reich 1995. "Employee Voice and Training in Career Development." Industrial Relations Research Association, Proceedings of the 49th Annual Meetings. Bilginsoy, Cihan 1996. "Apprenticeship Training and Prevailing Wage Laws." Working Paper, Economics Department, University of Utah. Borjas, George 1996. Labor Economics. McGraw-Hill. California Department of Finance 1995. California Statistical Abstract. State of California. Council of Economic Advisers 1995. Economic Report of the President 1995. U.S. Government Printing Office. Economic Report of the Governor 1995. Office of the Governor, State of California. Hamermesh, Daniel 1986. "The Demand for Labor in the Long Run." In Orley Ashenfelter and Richard Layard, eds. The Handbook of Labor Economics. North-Holland. Michael Reich Prevailing Wage Laws and the California Economy Page 13 Hamermesh, Daniel 1993. Labor Demand. Princeton University Press. Krueger, Alan and Lawrence Summers 1987. "Reflections on the Inter-Industry Wage Structure." In K. Lang and J. Leonard, eds. Unemployment and the Structure of Labor Markets. Basil Blackwell. Peterson, Jeffrey 1996. "The Effects for California Construction Workers from Changing the Method of Calculating Prevailing Fringe Benefits." Unpublished paper, School of Public Health, University of California at Berkeley. Philips, Peter 1996. "Results of a Multivariate Regression Analysis of Construction Worker Incomes with a Focus on the Implementation of Prevailing Wage Policies." Working Paper, Economics Department, University of Utah. Philips, Peter et al. 1995. Losing Ground: Lessons from the Repeal of Nine "Little Davis- Bacon" Acts. Working Paper, Economics Department, University of Utah. Prus, Mark 1996. "The Effect of PrevailingWage Laws on Total Construction Costs." Working Paper, Department of Economics, SUNY Cortland. U.S. Bureau of the Census, Census of Population 1990• Public Use Microdata Sample (PUMS). Machine-readable data. U.S. Bureau of Labor Statistics 1991. Productivity and the Economy: a Chartbook, BLS Bulletin 2084. U.S. Government Printing Office. Waitzman, Norman 1996. "Workers Beware: The Relationship Between the Structure of State Prevailing Wage Laws and Injuries in Construction, 1976-91." Working Paper, Economics Department, University of Utah. Weil, David 1992. "Building Safety: the Role of Construction Unions in the Enforcement of OSHA." Journal of Labor Research, 13, 3. January 8, 1996 Percentage Distributions of Construction Journeyman by Race/Ethnicity and Within Age Groups Non Hispanic White Black Asian, Other American Latino Pacific Isl. Indian 18-25 54.54 1.75 1.62 0.14 0.70 41.33 26-30 57.89 3.14 1.41 0.18 0.83 36.56 31-35 63.15 2.98 2.23 0.10 0.77 30.77 36-40 62.75 3.68 4.00 0.16 1.39 28.02 41-45 61.51 4.17 4.05 0.00 1.24- 29.03 46-50 58.42 4.86 5.34 0.19 1.37 29.81 51-55 65.50 5.12 3.45 0.00 1.21 24.72 56-60 65.11 7.30 2.62 0.32 0.58 24.07 61-65 71.84 5.04 2.52 1.29 0.57 18.75 Calculated from the 1990 Census 5% Public Use Microdata Sample. All figures are weighted. Rows sum to 100. Percentage Distributions of Construction Laborers by Race/Ethnicity and Within Age Groups , Non Hispanic White Black Asian, Other American Latino Pacific Isl. Indian 18-25 36.17 2.27 1.54 0.21 0.57 59.24 26-30 40.96 2.98 1.68 0.14 0.57 53.67 31-35 45.73 4.36 2.22 0.00 0.23 47.45 36-40 49.22 4.32 3.00 0.00 0.74 42.73 41-45 50.76 3.41 4.19 0.00 1.85 39.79 46-50 45.50 5.42 2.06 0.00 1.62 45.40 51-55 45.39 7.96 2.88 0.00 2.27 41.49 56-60 52.91 6.24 2.95 0.00 0.52 37.38 61-65 59.70 7.95 1.08 0.00 0.00 31.27 Cross-Tabulations and Percentage Distributions of Employment in the California Construction Trades by Occupation and Race/Ethnicity, 1990 Cross-Tabulations Race/Ethnicity Total Supervisors Journeymen Apprentices Laborers Non-Hispanic White 277,763 45,435 171,392 1$ 41 59,595 Black 15,908 1,546 9,068 199 5,095 Asian, Pac. Isl. 11,260 1,240 6,984 0 3,036 Other - 604 39 429 18 118 American Indian 4,322 633 2,634 0 1,055 Hispanic 179,861 12,505 98,017 527 68,812 Total 489,718 61,398 288,524 2,085 137,711 Percentage Distribution by Race/Ethnicity Race/Ethnicity Total Supervisors Journeymen Apprentices Laborers Non-Hispanic White 56.72 74.00 59.40 64.32 43.28 Black 3.25 2.52 3.14 9.54 3.70 Asian, Pac. Isl. 2.30 2.02 2.42 0.00 2.20 Other 0.12 0.06 0.15 0.86 0.09 American Indian 0.88 1.03 0.91 0.00 0.77 Hispanic 36.73 20.37 33.97 25.28 49.97 Total 100% 100% 100% 100% 100% Calculated from the 1990 Census 5% Public Use Microdata Samples. All figures are weighted. Average Annual earnings for California Construction Workers by Occupation and Race/Ethnicity, 1990 Race/Ethnicity Total Supervisors Journeymen Apprentices Laborers Non-Hispanic White $26,848 37,895 25,106 15,302 23,693 Black 21,277 28,191 20,858 15,694' 20,143 Asian, Pac. Isl. 21,388 35,652 20,250 - 18,180 Other 20,687 54,486' 20,931' 12,040' 9,950' American Indian 22,764 31300 23,149 - 16,680 Hispanic 17,267 27,811 17,570 11,948' 14,960 All Workers 22,979 35,494 22,271 14,464 19,011 Calculated from the 1990 Census Public Use Microdata Sample. All figures are weighted averages. a. Average computed with fewer than 30 observations. y <;:.w �j fi�Sri ttr.r•"�'Y �q y�Qg�.9�f4.fc�* • aYe;<K�i„ �e�a1c'�.r..�'sc a.,r,sr`.3.� -�' _.,�s= .. • • rf .�J' l�;u,. /t''.l3iti,7��'�A.�` 'y� r.-s t t�f at 5•� r.n .+- r �x t • • ' �s�q�� .r.��.s�yiiY�^:�.�/�}."�irr�;�,..G., f, `;''T �1..- _ ..Kts k_.-. _ r.., 4 • •. •• t y.t�,hs.'#r4 y�s`�.,`.'?• 1+a=as+i':t,%+^•,..6'.iY.W st ? lit • •- • • �}�* �,S:�`'rys�4t1�i-{}''4 r.$-i�tilr1��'t^� ���YYW��•�.3�R�yf•,'�• • �k ?��.�'L �"�_r"` z 3 "¢Ti '}xr-i �E µ� � r-i, -� s .�.i��„� ti�...i�yq'�.•F��J./.,�,`'i � - �� sr" „d,��w�'-Si��YS'��i"}�'�t�^�'.��y��•�s" � -ram% YfOG d r� r.'j"t••�•�I'I t�it`�'. f','� �a<�+.r�fLs-^7`.�<�"""'.�}'��"�f,V svY't�*�"1y'S'�',i,._s'✓.r'?,. .��, i.. '�. .-.9 i.tiY."':.J.re_�::?Yacr .SffFIR.�a�'J.R 0 tiII i3F.'_i''i'"Mn'S•�.N,Y�:y MAO• • h The Effect of State Prevailing Wage Laws on Total Construction Costs by Mark J. Prus Department of Economics SUNY, Cortland Cortland,NY 13045 607-753-5758 PRUS@eSNYCORVA.CORTLANDXDU copyright Mark J.Prus January 1996 The Effect of State Prevailing Wage Laws on Total Construction Costs Mark J. Prus Department of Economics SUNY, Cortland Summary Opponents of prevailing wage laws claim that the laws inflate construction costs by at least 25%. Professor Prus's paper tests this by examining the effect of state-level prevailing wage laws on the cost of state public construction. He uses a sample of 7,854 private and public construction projects reported by F.W. Dodge between 1990 and 1994. He includes offices, hospitals, schools, garages, and warehouses - both new construction and alterations and additions. The question he asks is, "Is the presence of a state-level prevailing wage law associated with higher construction costs on state projectsT' Controlling for factors such as project square feet, building type, construction type (new or renovation building materials number ), g r of stories, and the state in which the project was built, Professor Prus finds that state construction was 32% more expensive than private construction in states with prevailing wage laws. However, when he repeats this analysis for states without prevailing wage laws, he finds that state construction was 37% more expensive than comparable P P private construction. When he combines both samples and tests the effect of the presence of a state prevailing wage law on project cost, he finds that it has no effect on the cost of either private or public projects. While public construction is more expensive than private construction, prevailing wage laws have nothing to do with it. Professor Prus's work is the most advanced analysis of the relaxionship of prevailing wage laws to project cost to date. He uses a very large data set with a high number of controls, unlike the earlier research on which prevailing wage law opponents base their claims. His conclusion is that prevailing wage laws have absolutely no effect in boosting the cost of public construction. Abstract The impact of prevailing wage laws on public construction costs remains a puzzle despite the claims that some make regarding the inflationary effects of these laws. Proponents of changing the law argue that substantial savings will be achieved. Estimates of savings based on an examination of wage differentials tend to exaggerate the cost savings, however,by failing to recognize that labor costs are a small percentage of total costs or by assuming that labor productivity is invariant. The only careful study published in a recognized, peer reviewed, academic journal that examines the impact of prevailing wage laws on total construction costs directly finds that the Davis-Bacon Act raised the cost of federally funded projects 26 percent. The authors also find that reporting requirements and decreased competition "do not appear to be major contributors" to the higher costs of federal Davis-Bacon projects. The magnitude of this estimated inflationary impact is difficult to accept given the roughly 20 percent wage differential between workers on public and private projects. Indeed, the authors of this study acknowledge that, due to inadequate controls, they may be confounding cost differences associated with the Davis- Bacon Act with cost differences between public and private projects. Using data published by the F.W. Dodge Company on individual construction projects from 1990 through 1994, the study set forth herein estimates the impact of state prevailing wage laws on public construction project costs through regr.,ssion analysis. By comparing a large number of public and private projects across states both with and without prevailing wage laws, the effects of cost differences between public and private projects can be isolated from the impact of prevailing wage laws on total construction costs. In other words, this study provides the control group that the prior study acknowledged it was missing. The results indicate that public projects are 29.6 percent more expensive than private projects holding other factors such as type of structure, materials, size, and state location constant. At the same time the presence of a prevailing wage law does not have a statistically significant impact on total construction costs. In short, there is no credible evidence showing that public construction done under the regulation of prevailing wage laws is any more expensive than public construction undertaken in the absence of prevailing wage regulations. I Introduction Prevailing wage laws have been a part of public policy in the United States at both the federal and state levels since the early part of this century. These laws require that construction workers be paid the "prevailing rate"when working on publicly funded construction projects. They were initially established to neutralize the government's monopsonistic power as a purchaser of certain types of construction labor and to support the social objective of maintaining a family wage. Currently, the question of whether prevailing wage laws continue to make public policy sense is being debated. One of the most significant issues in this public policy debate is what impact prevailing wage laws have on publicly funded construction costs. Opponents of. prevailing wage laws argue that they raise construction costs and, according to some,that repeal of these laws would result in cost savings on the order of 20 to 30 percent. If such savings were possible,it is argued, school districts could build five schools for the price of four. Claims of cost savings from the repeal of prevailing wage laws,however, are generally based on analysis of the effect of higher wage rates on construction costs. Yet wage differences have a moderate effect on total construction costs. Labor costs are less than a third of total construction costs and have been falling. In 1972,for instance, in an analysis of school construction costs,John Olsen found that onsite wages and salaries excluding benefits were 28.2 percent of total costs. (Monthly Labor Review, 1979,p. 40) According to the Census of Construction,labor costs counting benefits on all types of construction were 30 percent of total costs in 1977 and had fallen to 26 percent by 1987. 2 A second problem leads us to question estimates of the impact of prevailing wage legislation on construction costs based on an analysis of wage differences. These analyses assume implicitly that the same number of hours of each type of labor will continue to be employed and that labor productivity is constant. Neither of these assumptions is necessarily appropriate. The payment of prevailing wages may serve to attract workers who have more experience and training and are, therefore, more productive. Increased labor productivity will result in fewer hours of labor being required thus offsetting the higher wage rate. For instance, Allen has shown that unionized labor in the construction industry is between 17 and 52 percent more productive than nonunion labor. (Allen, 1984) Additionally,higher wage rates may lead to capital or other inputs for labor,mitigating the impact of higher wages contractors to substitute p g g P rates on total construction costs. These possibilities, alone or in combination, make the assumptions underlying the analysis of construction costs based on wage differences inappropriate and cast doubt on the estimates of costs savings. Specifically, it is difficuit to imagine how savings of 20 to 30 percent are possible. To get a true picture of the impact of prevailing wage legislation's impact on total construction costs,one would have to evaluate not only differences in wage rates,but also productivity differences, the incidence of substitution,administrative costs and other ways in which these laws' impact is either mitigated or enhanced. An alternative approach is simply to examine total construction costs directly and compare those costs in the presence and absence of prevailing wage laws, controlling for project differences. 3 Previous Estimate of Impact on Total Costs Only one study has attempted to estimate the impact of prevailing wage legislation based on actual total construction on costs. Fraundorf, et. al. in The Effect of the Davis-Bacon Act on Construction Costs in Rural Areas,"examined 215 new, non-residential construction projects built in 1977 and 1978. (Fraundorf, 1983) Approximately half of these projects were federal P J projects built under the purview of the federal Davis-Bacon Act specifying that prevailing wages be paid. The other half were privately owned projects constructed without the requirement that prevailing wages be paid. Data on total construction costs were then compared using regression analysis to control for the effects of factors other than the presence of prevailing wage requirements. This study controlled for differences in the type of structure, the types of materials used, and project size in an effort to focus on cost differences associated with labor cost differentials resulting from the regulatory regime. It also attempted to control for regional differences in construction costs by grouping projects into four regions; Northeast, North Central, South and West. The dependent variable in the regression analysis was the natural log of the projects bid price deflated to 1977 dollars. The authors of the study found, somewhat surprisingly, that federal construction projects governed by Davis-Bacon were 26 percent more expensive than private construction projects controlling for other cost influencing factors. When they re-estimated their basic model to correct for disproportionate response rates by region and building type, the authors concluded that the impact of Davis-Bacon on total construction costs was as high as 30 percent. While they admitted that their results were high, especially in light of a public-private wage differential estimated at 20 percent,they pointed out that the results were consistent with other aspects of the 4 data. In particular, they did not find evidence of factor substitution which would mitigate the impact of prevailing wage requirements. However,this does not explain why the impact on total costs is greater than the wage differential. They explored other factors that might have contributed to the higher costs of federal Davis-Bacon projects, such as record keeping and reporting; and decreased competition; but neither of those factors appeared to have contributed significantly to costs on federally funded construction. (Fraundorf, et. al, 1983, p. 145) One possible problem with the Fraundorf study is that regional differences in construction costs may have been inadequately controlled for. Given the relatively small sample size, the authors of the study had to group construction projects into relatively broad regions. This creates the potential for comparing a private project in a low cost state, such as Idaho, with a public project in a high cost state such as California. Since both projects are considered to be in the same region, the cost differential is incorrectly attributed to the impact of prevailing wages when, in fact, it is actually due to differences in the cost of living or cost of materials between Idaho and California. Another problem may result from the way in which building types were classified. Each construction project was placed into one of six categories;recreational buildings, storage facilities, industrial buildings, office-commercial, medical and other. These categories were then used to find matches between public and private construction projects. However,these six categories were sufficiently broad to allow rather dissimilar buildings to be considered as comparable. For instance,in the category storage facilities, warehouses were grouped with barns as well as airplane hangars. Likewise office buildings were in the same category as restaurants. Thus, differences in costs between public and private buildings may have resulted 5 from differences in structure type. Rather than from prevailing wage requirements. (Fraundorf, 1982, pp. 14-15) A third and potentially more serious problem with the study is that it failed to isolate adequately the impact of prevailing wage legislation on construction costs. Specifically, Fraundorf compared private projects constructed in the absence of prevailing wage legislation with federal (i.e.,public)projects built using workers paid the prevailing wage. This comparison, while seemingly appropriate, contains the potential for confounding cost differences related to prevailing wage laws with cost differences resulting from other differences between private and public construction projects. The authors recognize this possibility when they point out, "If the government is more exacting than private owners in its quality standards, labor hours (and costs) and possibly material costs would be higher on government projects." (Fraundorf, 1983,p. 145) It may also be that the difference in bidding procedures for public and private contracts, or differences in the time horizon of public and private owners, may contribute to higher costs in the public sector. In other words, the cost differential that Fraundorf attributes to the effect of prevailing wage legislation may, in fact,be due instead to differences between public and private construction. Estimating the Impact of State Prevailing Wage Laws The current study analyzes the impact of prevailing wage legislation on total construction costs using data on nonresidential construction in the United States. The data are from the F. W. Dodge Company, an organization that collects and disseminates data on construction projects to the construction industry. These data provide information on 6 construction costs at the start of the project(i.e., the bid price). They also contain detailed information on structure type, project location, project scale;and technical characteristics of the project such as number of stories and type of frame. The Dodge data also distinguish between public, private and federal projects. One advantage of the Dodge data is that they report on a large number of construction projects allowing for a more appropriate geographical breakdown of projects. In addition, the Dodge data make it possible to compare construction costs on similar projects in the private and public sectors for states both with prevailing wage laws and without. This is essential if one is to isolate the cost differences associated with prevailing wage laws, as opposed to the cost differences associated with public construction.. The Model The model used here to estimate the effect of prevailing wage legislation on construction costs is : C= a + b,S+ b,T+ b3B + b4A + bAST+ b6AItadd+ b,Pubcode + e where C= start cost or bid cost;S= project scale as measured by square footage; T= a vector of dummy variables indicating detailed structure type;B= a vector of building material dummy variables;A = a vector of state dummy variables;ST= the number of stories;Ahadd= a dummy variable indicating whether the project was an alteration or addition as opposed to new construction; and Pubcode=a dummy variable indicating that the project is public. This model is nearly identical to the one used by Fraundorf to estimate the impact of the federal Davis- Bacon Act on total construction costs. It is appropriate for estimating the cost difference be�ween public and private construction projects holding constant other factors such as building 7 type, building materials, state in which the project is undertaken, building size and complexity. In states with prevailing wage laws, the cost difference between public and private projects may be thought of as measuring the inflationary impact of the law. Data The most notable difference between the Fraundorf analysis and this study is the focus here on the impact of state prevailing wage laws, also known as little Davis-Bacon acts, on costs. In the first instance, I selected construction projects in states that had prevailing wage laws in 1990. This focus on state prevailing wage laws' impact on costs, as opposed to Davis- Bacon's impact, is useful for differentiating between costs resulting from the presence of prevailing wage laws and the cost resulting from the construction being public rather than private. Moreover, given that many states are currently considering repeal or modification of their"little"Davis-Bacon acts, the impact of state laws on construction costs is an important question to answer. The projects selected were nonresidential construction categorized as offices, hospitals, elementary schools, middle schools, secondary schools, garages and warehouses. These categories are similar to those used by Fraundorf,et al.,but are more specific and consequently less likely to result in comparisons between dissimilar structures. For example,hospitals is a more detailed class of structures than medical buildings, and warehouses is similarly more specific than storage facilities. Unlike Fraundorf, I included both new construction and additions/alterations. In states with prevailing wage laws, additions/alterations to public structures are covered by the statues. The results of a multiple regression analysis using the natural log of real total project cost 8 as the dependent variable, controlling for other relevant factors, in states that have prevailing wage laws, are reported in Table 1. Public buildings in such states are 31.8 percent more expensive than similar private structures. This result is very similar to the results obtained by Fraundorf in terms of the magnitude of the estimated impact. Given that the prevailing wage laws apply only to public projects, this estimated cost differential can be attributed to the law's existence. This approach to estimating the impact of prevailing wage laws on construction costs, like the earlier study by Fraundorf, is unable to distinguish between cost differences due to ownership differences i.e., public vs. private) and cost differences that result from prevailing wage requirements. By comparing public projects built in states where prevailing wage laws are in effect with private projects in those same states, the impact of the law is confounded with cost differences that are attributable to public versus private ownership To illustrate the problem the model presented earlier was re-estimated using data on construction projects from states without prevailing wage laws. Similar controls were used to insure that public projects were being compared with similar private projects. The results of this regression are reported in the third column of Table 1. Once again,public projects are significantly more expensive (37 percent)than comparable private projects. But because the projects examined were located in states that currently do not have prevailing wage laws, the cost differential can no longer be attributed to the law's impact. This result lends support to the notion that the public sector may be a more exacting owner than the private sector. It also suggests that it is inappropriate to assume that the higher cost of public projects is attributable to the presence of prevailing wage laws. 9 Table 1: Regression Results States w/Laws - States w/o Laws Variable Coefficient Coefficient SCHOOL .586** .769** HGHSCHL .712** .878** HOSP 1.077** 1.229** WARE .016 .425** OFFICE .242** .588** PUBCODE .276** .317** STEELDUM -.045 -.081 WOODFRM .150** .061** CEMENTDM .091** .118** LNSQFEET .562** .581** STOIHES .071** .110** ALTADD -.160** -.066** (Constant) 8.187** 7.569** Adjusted RZ=.625 Adjusted RZ=.671 N=5136 N=2717 F= 210.087 F= 186.245 Dependent Variable is ln(real total costs) where total costs are reported in 1994 dollars ** coefficient is significant at the .01 level the coefficients for the state dummy variables are not reported 10 A more appropriate analysis recognizes that there are two different dimensions to construction cost differentials. On the one hand, comparisons of public projects versus private projects can determine the extent to which the government may be a more exacting owner. The other dimension considers the presence or absence of prevailing wage legislation. Combining these two different dimensions creates four different possibilities: private projects built where no prevailing wage law exists;public projects with no prevailing wage law; private projects where prevailing wage laws exist; and public projects with prevailing wage laws. Only the last category of construction projects is directly affected by the presence of a prevailing wage law. Thus, in order to appropriately assess the impact of prevailing wage legislation on construction costs, this category must be isolated from each of the other possibilities. The model used here to estimate the effect of prevailing wage legislation on construction costs is C = a + b,S + b2T+ b jB + b¢A + b j ST+ b6Altadd+b,PW+ b8Pubcode + belnteract+ e — measured footage; T= a vector where C= start cost or bid cat, S— project..tale as b s m y square� g , of dummy variables indicating detailed structure type;B= a vector of building material dummy variables;ST=the number of stories;A = a vector of state dummy variables;Altadd= a dummy variable indicating whether the project was an alteration or addition; PW=a dummy variable indicating the presence or absence of a prevailing wage law;Pub=a dummy variable indicating ;Interact= (PW x Pubcode). The key components of this model are the variables PW, Pubcode and Interact. In combination these three variables allow us to estimate the effect of prevailing wage laws separately from the effect of public ownership. The Pubcode variable estimates the cost differential that exists between public and private projects, ceteris paribus, 11 regardless of whether a prevailing wage law is in effect. The PW variable estimates the effect of prevailing wage laws on construction projects in states with laws regardless of whether they are public or private. Finally, the Interact variable picks up the direct effect of prevailing wage legislation on public projects since it is equal to one only in those cases where there is a public project in a state with prevailing wages. This model was estimated using the combined data for I� states with and without prevailing wage laws. The results are reported in Table 2. The coefficient on the interaction term(Interact) is positive but statistically insignificant indicating that the direct effect of prevailing wage laws on the cost of public construction projects is negligible. The presence of a prevailing wage law also does not appear to have any significant effect on the costs of construction projects. Public projects in all states, however, are t private projects as indicated b the coefficient more expensive 29.6 percent)than y significantly ( ) P P 1 Pe Pe on the variable Pubcode. Conclusions The results of this analysis of the impact of state prevailing wage laws on total construction costs indicates that, in contrast to earlier academic analyses, as well as some casual statements, there is no measurable cost difference between similar structures as a result of prevailing wage requirements. Consequently,repealing or modifying these laws will not produce the substantial savings promised by proponents of such actions. At the same time there are significant measurable cost differences between public and private projects of a similar nature. Researchers and politicians both should try to determine 1). why this differential exists, 2). whether it should be reduced, and 3). what steps could be taken to achieve that result. 12 Table 2. Regression Results:Determinants of Construction Costs for All States Variable Coefficient SCHOOL •628** HGHSCHL •747** HOSP 1.109** WARE .132** OFFICE 342** PUBCODE .259** LAW .180 INTERACT .050 STEELDUM -.053* WOODFRM .118** CEMENTDM .106** LNSQFEET .569** STORIES .081** ALTADD -.128** (Constant) 7.952** Adjusted R Square = .637 N=7854 F= 224.180 Dependent Variable is ln(real total costs)where total costs are reported in 1994 dollars ** coefficient is significant at the.01 level *coefficient is significant at the.05 level the coefficients for the state dummy variables are not reported I 13 References Steven G. Allen, "Unionized Construction Workers Are More Productive," Quarterly Journal of Economics. (May 1984), pp. 251-274. Martha Fraundorf, John P.Farrell and Robert Mason, "The Effect of the Davis-Bacon Act on Construction Costs in Rural Areas," The Review of Economics and Statistics, 1984,. pp. 142-146. Martha Norby Fraundorf, John P. Farrell, and Robert Mason, "Effect of the Davis-Bacon Act on Construction Costs in Non-Metropolitan Areas of the United States,"Report to the American Farm Bureau Federation, (Corvalis, Oregon: Department of Economics, Oregon State University, 1982). John G. Olsen, "Labor and Material Requirements for New School Construction,"Monthly Labor Review, April, 1979, pp. 38-41. r 14 Biographical Sketch Mark Prus is Associate Professor of Economics at the State University of New York, Cortland. He has published articles on wage and occupational structures in the Cambridge I� Journal of Economics, Journal of Economic History, Quarterly Review of Economics and Finance, Journal of Socio-economics, Journal of Economics Issues, Social Science Journal, and Research in Economic History. A Comparison of Public School Conssiruction Costs In Three Midwestern States that Have Changed Their Prevailing Wage Laws in the 1990s Kentucky, Ohio and Michigan By Peter Philips, Ph.D. Professor of Economics, University of Utah 00 Introduction Proponents and critics of prevailing wage regulations have debated the merits of these regulations for some time. Proponents argue that these regulations promote the development of a slulled labor force in construction, improve rove work P place safety, encourage quality construction, increase apprenticeship training and provide career opportunities in construction for local citizens. Proponents emphasize that prevailing wage regulations also induce contractors to provide health insurance and pension coverage that otherwise would be absent Critics of prevailing wage regulations concede some of the foregoing positions and contest others. But the main argument of critics of prevailing wage regulations is the contention that these laws raise public construction costs. This is a two-sided argument asserting that when prevailing wage regulations are applied they raise public construction costs and when these regulations are eliminated, public construction costs will go down. The magnitude of savings is thought to be substantial ranging anywhere from 10% to 30% or more of total construction costs. This paper focuses on the specific question of whether or not the application of P Pe P� q PP prevailing wage regulations raises costs, and if so, by how much. Those favoring this view theorize that prevailing wage regulations raise wage rates on public construction to higher levels than they otherwise would be. Increased wage rates should lead to increased construction costs that would be passed on to the government and eventually the taxpayer. Proponents of prevailing wage regulations counter that higher wage rates induce contractors to hire or train a more skilled and productive labor force. Higher wage rates also will encourage contractors to better manage their workers and provide them with better and more up-to-date equipment. These responses to higher wage rates may, according to this view, offset some or an of the costs of higher wage rates. Prevailing wage proponents also argue that a more skilled labor force leads to better quality construction that reduces downstream maintenance and repair costs. This paper tests these competing hypotheses regarding the cost-effects of prevailing wage laws. The focus wll be on new public school construction in Kentucky, Ohio and Michigan over the period 1991 to 2000. These states and 2 this time period were chosen because legislative and judicial changes in these states over this period form a natural experiment helpful in isolating the effects of these regulations on costs. In the 1990s, prior to July, 1996, Kentucky did not apply prevailing wage regulations to public school construction. Starting in July of 1996, the state's prevailing wage law regulated public school construction. By itself, this provides some before-and-after information about the effects of applying prevailing wages to school construction costs. Fortuitously, from the standpoint of science, Ohio in the 1990s did almost the opposite of Kentucky. Throughout the 1990s until July of 1997, Ohio applied its prevailing wage law to public school construction. Starting in July 1997, Ohio exempted public school construction from prevailing wage regulations. Thus, almost simultaneously, these neighboring states moved in opposite regulatory directions. The fact that Ohio lifted its law soon after Kentucky applied its law to public school construction hers th:s rat ral '-.-riment isclzte the effects of regulatory policy from other factors that change over time. From an experimenter's perspective, this is nice. But nicer yet, at around the same time, Michigan does yet a third thing with its prevailing wage law. At the end of 1994, a judicial ruling suspended the application of Michigan's prevailing wage law to any public construction. This judicial suspension lasted until July of 1997 when a higher court ruling reapplied Michigan's prevailing wage law to all public construction, including schools. So Michigan suspended its law two-and-one-half years before Ohio, and Michigan reapplied its law in precisely the same month Ohio exempted schools from prevailing wage regulations. Figure 1 shows the variation in prevailing wage policies by state in the 1990s. `- r'`.� Kentucky No Law •�,...1" �Z'`� •�— •:C.f-,.s sJv,�� K''ifv.Ya�..t'1`y Ohio No Law .. .'its•�� �:�: ., '.'C�J f!4► R� .y.r. NichipnLawNO LawLa 1991 19M 1995 1997 1999 2001 .Figure 1:Prevailing Wage Policy by State 1991-2000 3 With these variations in legal policies in hand, we are in a position to assess statistically whether or not changes in prevailing wage policies as they applied to public school construction raised or lowered the cost of building public schools. The Data The FW Dodge Corporation is a private company that provides bidding information.to contractors. In so doing, the Dodge Corporation systematically gathers information on the 'start" cost of construction projects. 'Stan' costs are the agreed-upon bid price of a project at the outset of a bid. The final cost of a project can vary from the start cost based on cost overruns. Proponents of prevailing wage regulations argue that one of the advantages of prevailing wage laws is that they reduce cost overruns. This argument asserts that absent prevailing wage regulations a cutthroat bidding system emerges where low-ball bidders undercut their competitors with unrealistically low bids in the hope and expectation that during the term of the project the contractor can recoup his profits through change orders. This argument asserts that prevailing wage regulations attract a set of bidders that will compete with each other over on-time completion, quality construction and productivity but eschew the strategy of low- ball bidding hooked into profiting from change orders. We cannot test this argument with Dodge data because it does not include cost overruns subsequent to the acceptance of the bid. But with this one limitation in mind, the Dodge data form the single best source on school construction costs across states. In addition to providing the accepted bid price, the dodge data indicate what kind of building project it is; what the total square feet of the project is; where the project will take place; when the bid was accepted; some details on the nature of the project and other useful information. This report uses data from 1991 to September, 2000. We focus on new public school construction only. By eliminating renovations, alterations and additions, we can focus on a relatively homogenous group of buildings—new public schools. We will ask the question--controlling for the size of these public schools, and where they took place, and when they were built, and whether they included a gymnasium-swimming pool facility—did the presence or the absence of prevailing wage regulations affect the total cost of the project? Table 2 describes the new schools used in this study. Characteristic of Schools in Stud Number of New Schools in Study 391 Average Square Foot Size of the School 86,415 Averaoe Total Cost of the Project ear 2000 dollarsl $8,483,937 Percent of All Schools Michigan 38% Ohio 36% Kentucky 26% Percent of School with a Gym-Pod Facility 7% Percent of Urban Schools 32% Percent of Schools Built Under Prevailing Wages 49% j .Table 1:Description d the new xho*used in the study This study involves 391 new public schools built between 1991 and September 2000. Table 1 show that the average size of a school was 86,415 feet and the average total cost was $8,483,937 in year 2000 dollars. The consumer price index urban (CPI-U)was used to update earlier costs into year 2000 dollars. Of the 391 new public schools in the study, Michigan accounted for 38%. Ohio accounted for 36% and Kentucky accounted for 26%. This reflects the relative size of the three states. Dodge data indicated that 7% of the new schools included a swimming pool/gymnasium facility. in our statistical model we expect that even controlling for the size of the school project the inclusion of such a facility is likely to raise the square foot cost of the new school. Thirty-two percent of the new public schools were built in urban areas with the remaining 68% built in rural areas including smaller towns. Almost half, 49% of the projects were built with prevailing wages while the remaining 51 r6 were built without these regulations.ns. Table 2 shows the distribution of when the new public schools were butt. Column c shows that 1991 had unusually few new public schools built, and the year 2000 had unusually many schools built. This reflect the two ends of the business cycle and building cycle. The year 1991 was a recession year and the economies of these states grew progressively from then through the 1990s. One result of this growth has been an expansion of school construction. a Percent of Each Percent of Years New Number Decade's Total Schools Built of New New Schools Built Under Prevailing Year Schools in Each Year Wage Laws a b c d 1991 5 1% 20% 1992 44 11% 61% 1993 28 7% 68% 1994 10 3% 50% 1995 39 10% 33% 1996 49 13% 37% 1997 53 14% 49% 1998 33 8% 58% 19991 56 14% 71% 20001 74 19% 30% Total 391 49% -Table 2:The time distribution of new school construction within the study Table 2 also shows the annual percent of each year's new school construction that was done using prevailing wages. The balance tends to be stable over time. The dip in 1995 and 1996 is due to Michigan suspending its law in those years. (Michigan split construction in 1997 with the first half not using prevailing wages and the second half using this regulation.) In the year 2000, Ohio has a major increase in school construction, ap done absent prevailing wages. This accounts for the somewhat lower percentage of new school built with prevailing wages in , that year. e Median Cost per Square Foot of New Elementary Schools 120 - �.�;Ohio Eliminates L.aw.-. Kentucky Adds law - 100 " 1, --- —♦—Kentucky c o s. .� _ - _ h O b or • .60 c _s__..�........._ Period 11: v Kentucky Has Law covering . 40 Period 1: Ohio Has Law covering - z r 20 1992 1993 1994 1995 1906 1997 1M 1999 2000 •Figure 2:Median Cost of New Public Elementary Schools in Kentucky and Ohio by Year Figure 2 gives a general sense of the experiment we are going to perform. In Figure 2, the median square foot cost to build a new elementary school in Kentuckyand Ohio are compared for the years 1992 to 2000. These data are not , deflated. They are in the actual dollars reported at the time. Consequently, it is not surprising that the median cost rises with time. In our statistical model,we will deflate these prices using the Consumer Price Index and ask the question r whether controlling for Inflation, new school construction costs are still going up. We will also control for a variety of other factors that cannot be controlled for in a simple figure such as the one above. Our additional controls include a control for possible economies of scale associated with larger schools, a control for the differences between urban and rural construction, a control for fancy facilities, a control for tight construction markets possibly pushing up the real cost of construction. With these controls in place, we will look at variations in prevailing wage policies. In Figure 2, changes in prevailing wage policies are shown by two vertical lines one in 1996 representing Kentucky's application of a prevailing wage regulation. And a second vertical line in 1997 representing Ohio's exemption of public schools from prevailing wages. If these laws had an effect on costs, one might expect to see a change in the general relationship between 7 median new public elementary school construction costs in the two states. Visually, this does not seem to appear. We will go beyond visual inspection to examine all 391 new schools and test whether or not, controlling for other factors, changes in prevailing wage regulations have made a difference. Comparison of Mean Square Foot Cost We begin our analysis with a simple comparison of the average or mean inflation- adjusted square foot cost of building a new school. The 391 new schools are broken down into those built in urban areas (126 schools) and those built in rural areas (265 schools). Urban areas include the areas around Cleveland, Columbus, Cincinnati, Dayton, Louisville, Lexington, Detroit, Flint, Grand Rapids and Lansing. Within each group of urban and rural schools, the schools are hmken down into those built under prevailing ►►age regulations and these built without prevailing wages. Table 4 shows the mean, standard deviation' and number of schools in each of four categories: 1) rural schools built without prevailing wages; 2) rural schools built with prevailing wages; 3) urban schools built without prevailing wages; and 4) urban schools built with prevailing wages. New Public Schools Real(Inflation Adjusted)Square Foot Cost a b c I d e f g 1 Rural Schools y Urban Schools 21 Mean Standard Deviation Number Mean Standard Deviation Number. 31 No Law $96 ^',ti- ^ram*':at $2 r t! 161 $114 $36 40 4 Law $98 $24 s:.= 104 $114 $34 86 5 t-test -0.78 0.05 statistically Q-'kfi_„s• ; r,r- SignificantN Z• '�'_ NO 6 Difference? .Table 3:Comparison of the Real(InflatimAdjusted) Square Foot Cost of New Public Schools by Urban and Rural Schools and Butt without or with Prevailing Wages ' The standard deviation is, in essence the wiggle around an average. So for instance,if you had 5 children in a carpool ages 8, 9, 10, 11 and 12, the mean (or average)would be 10 years of age, and the standard deviation or wiggle around the mean would be 1.4 years. a The comparison of mean real square foot costs can be seen in rows 3 and 4, columns b and e. Considering rural schools first, the average or mean real square foot cost of schools built without prevailing wages was $96 per square foot while the mean real square foot cost of schools built with prevailing wages was $98 per square foot There were 161 new rural schools built without prevailing wages in the sample and 104 new rural schools built using prevailing wages in this sample. (See column d, rows 3 and 4). The standard deviation is a statistical measure of the wiggle around each mean. ft is used to construct a statistical test of whether or not the $2 difference in the average cost of construction per square foot is statistically significant The statistical test is called a West Typically, for there to be statistical significance, the t-statistic must be around plus or minus 2. In this r-ncc fr►r rural schcc4s where the difference is $2 per s-1-!?re f_`Ot, the t statistic shown in column b, row 5 is -.75. What this means is, statistically speaking, there is no difference between the average square foot cost found for rural schools built with prevailing wages compared to the average square foot cost of schools built without prevailing wages. Considering schools built in urban areas, in the sample, 126 new schools were built in urban areas with 40 being built without prevailing wages and 86 being built with prevailing wages. The average or mean real (inflation-adjusted) square foot cost of urban schools built with and without prevailing wages was almost equal. Indeed, rounding to whole dollars, they are equal at$114 per square foot (in year 2000 dollars).2 Again, the West indicates that any minor difference in these two means (in this case 34 cents) is not statistically significant Another way of putting this is: the difference in average real square foot construction cost for new public schools is due to random differences and statistically, the averages are equivalent This conclusion holds both when comparing urban schools and when comparing rural schools built with and without prevailing wages. A Statistical Model of New Public School Construction Costs Table 4 presents the resufts of a statistical model of new public school construction costs. The model is called an ordinary least squares linear regression model. This type of statistical model is very commonly used by economists, epidemiologists and others studying social phenomena. The particular model in Table 4 uses the 391 new public schools built in Kentucky, Ohio and Michigan over the 1991 to 2000 period as data to help predict the effects of various factors on total new construction costs. The focus variable in 2 If you do not round to whole dollars, the mean for schools built without prevailing wages was $114.17 and with prevailing wages it was$113.83. 9 the equation is the last variable in the model shown in gray on line 12. But before we get to this issue, let us examine the other aspects of the model. t- Significance Statistically Model Coefficient statistic level Significant? 1 a b c d e f 2 (Constant) 4.45 16.05 0% Yes 3 Size Natural Log of the Total Square Feet of the Project 1.00 41.59 0% Yes 4 Business Boom Time (in years) 2.9% 5.58 0% Yes 5 Location School was built in Ohio -12.6% -3.70 0% Yes 6 School was built in Kentucky -14.6% 4.03 0% Yes 7 School was built in an urban area 10.5% 3.41 0% Yes 8 Special Facilities School had a oym/pool facility 9.2% 1.69 9% Yes 9 Timing School was started Winter quarter -5.6% -1.21 23% No 10 School was started Spring quarter -10.9% -2.75 1% Yes 11 School was started Summer quarter -2.7% -0.63 53% No 121 Law School was built with prevailing wages 1 0.7%1 0.26 - 79%1 No 13 Total Cost of Natural log of real(inflation adjusted)total start cost of each new school(in year 2000 dollars) School 14 Model statistics Adjusted R-s ware a goodness of fit statistic = .85 15 Number of new schools in the sample= 391 •Table 4:A Unear Regression Model of the Total Square Foot Cost of Building New Public Schools in Kentucky,Ohio and Michigan Focusing on the Effect of Prevat7tng Wage Regulations The first key factor in the model is the size of each of the 391 new schools. The coefficient of 1.00 says that as the square foot size of the school increases, the total cost of the school increases proportionately.3 The second variable is simply time measured in years. This variable captures the fact that building costs have been rising faster than inflation in the 1990s. The cost data in the model are inflation adjusted using the Consumer Price Index The time variable indicates that after adjusting for inflation, new public school construction costs in these three states have been rising at 2.9% per year from 1991 to 2000. This result is statistically significant The reason bulding costs have been rising faster than inflation is because the economic boom has led to a very vigorous boom in building leading to heavy demand for construction services. 3 In after shidws,t have found in sortie cases that there were economies of scale in school txorauuction cross. Namely,as school size increased,total cat went up but more slovAy than total size went up. The absence d economies d scale among these 391 schools may be due to the relative homogeneity d the buldngs in the sarripie 10 Rows 5, 6 and 7 in the model present three variables that control for where the school was built. Row 7 indicates whether or not the school was built in an urban area. Urban schools cost 10.5% more than rural schools controlling for other factors such as the size of the school. Rows 5 and 6 indicate whether or not the school was built in Ohio (row 5) or Kentucky (row 6). In the case of urban schools, the reference point was rural schools. In the case of Kentucky and Ohio, the reference point is Michigan. The model indicates that new Ohio schools cost 12.6% less than Michigan schools while new Kentucky public schools cost 14.6% less than Michigan new public schools. These results, too, are based on controlling for other factors such as the size of the school, when the school was constructed, whether or not the school was urban or rural and whether or not the school was built under prevailing wage mandates. Rows 9, 10 and 11 indicate in what season the school was started. Row 9 indicates schools started in the winter quarter(January, February or March). Row 10 indicates schools started in the spring. And row 11 indicates schools started in the summer. In all three of these cases, the reference point is schools started in the fall. These seasonal variables get at the question of whether breaking ground on a new school in the face of winter weather raises the-cost of building that school. The model indicates that schools started in the winter were 5.6% cheaper than schools started in the fall. But this result is not statistically significant The lack of statistical significance means that you cannot be sure there really is any difference in the total cost of schools started in the winter compared to the fall. In the case of schools started in the spring, they were 10.9% cheaper than schools started in the fall (again controlling for other factors such as the size of the school, whether it was an urban or rural school, etc.). In this case, the results of the model are statistically significant That is, this statistical model indicates that you can be confident that breaking ground in the spring will lead to lower cost new school construction compared to breaking ground in the fall. Breaking ground in the summer is estimated to be 2.7°/6 cheaper than breaking ground in the fall, but this is not a statistically significant difference. The point of these results, however, is dear. Don't break ground into the teeth of winter. It will cost you. Indeed, the model shows that by starting in the spring instead of the winter saved enough jmoney for each school to include a gymnasium/pool facility in its specifications. The 7% of all schools that had such facilities paid 9.2% more total costs, controlling for other factors. And compared to starting in the fall, starting in the spring would have offset the cost of a swimming pool. in most cases in the FW Dodge dada,the start date is the date of bid acceptance. Given that there wit be some tag between bid acceptance and ground breaking,fal probably means late fal and spring ply means Late spring Also,given the probable cause of increased costs associated with breaking ground in the late fair is weather,the effect of this seasonal factor is probably stronger in the colder areas within the sample or Michigan,Ohio and Kentucky. 11 Figure 3 shows that for the most part, school boards and contractors know this. Fully 40% of all new public schools in Kentucky, Ohio and Michigan started in the years 1991 to 2000 were started in the spring. Fall starts—the most expensive start time—accounted for only 16% of all starts. However, whatever drove schools boards to start schools in the fall compared to the spring also led them to pay a 10% premium for this choice of when to break ground. Distribution of New School Starts by Season fall winter 16% 17% ®winter PJ - ■spring summer O summer 27% ®fall ' Uiipspring 40% •Figure 3:Percent Distribution of New School Starts by Season for Kentucky,Ohio and Michioan. 1991-2000 The Effect of Prevaifing wages on costs Controlling for seasonality, controlling for differences in rural and urban construction costs, controlling for the size of the project and controlling for the state in which the project was built, the model estimates that using prevailing wage regulations raised school construction costs by 7110t`of 1%. But again,this is not a statistically significant result. In effect, the model says there is no effect on total costs associated with prevailing wages. How can this be? Prevailing wage rates insure that all contractors must pay the wage rates that prevail for an occupation in an area. Without this regulation, ,12 Peter Philips grew up in Compton and Pomona, California. He received his BA from Pomona College in 1970 where he majored in economics and received the Leland Backstrand Graduating Senior Award in Economics. Philips received his MA in economics (1976) and his PhD in applied economics (1980) from Stanford University. Philips is a Professor of Economics at the University of Utah. He is co-editor of Three Wodds of Labor Economics(M.E. Sharpe, 1986)and coauthor of Portable Pensions for Casual Labor Markets:the Central Pension Fund of the Operating Engineers (Quorum Books, 1995). Philips has published widely on the canning and construction industries in journals such as the Journal of Education Finance, Industrial and Labor Relations Review, Industrial Relations, Business History, the Jouma/ of Economic History, Historical Methods, The Journal of Economic Literature, Oxford Encyclopedia of Economic History and the Cambridge Journal of Economics. Philips has been a consultant for the U.S. Labor Department analyzing the supply of cannery labor in California, and he has worked as an expert on the Davis-Bacon Ad for the U.S. Justice Department. The Davis-Bacon Act regulates wage payments to construction workers on federal public works. Philips is a respected expert on prevailing wage laws and on employment, training wages and benefits in the construction industry. He has testified before state legislative committees in Ohio. Indiana, Kansas, Oklahoma, New Mexico and California on their state prevailing wage laws. Along with other researchers at the University of Utah, Philips has analyzed the effects of prevailing wage laws on public construction costs, construction worker incomes, apprenticeship training, worker safety and minority access to construction work Philips is the senior labor economist at the University of Utah. He teaches a wide range of courses in the area of labor economics, econometrics, labor law, collective bargaining and economic history. Philips has received awards for his teaching and community service, including University of Utah Public Service Professorship, the University of Utah Presidential Teaching Scholar Award and the University of Utah, College of Social and Behavior Science Superior Teacher Award. Philips is married with two children. 14 Appendix Simple Comparison of One Year Before and One Year After Legal Change in Kentucky, Ohio and Michigan The following is a simple comparison of the average or mean square foot cost of new public schools one year before compared to one year after a change in the prevailing wage law in the three states—Kentudry, Ohio and Michigan. Kentucky applies prevailing wage regulations to public school construction in July of 1996. Ohio exempts public schools from prevailing wage regulations in July of 1997. A Michigan court suspended the application of prevailing wage regulations in the state in December, 1994. A second court reapplied prevailing wage regulations in July of 1997. These variations allow for four simple 12-month before-and-after comparisons— one for Kentucky, one for Ohio and two for Michigan. In this simple comparison there is no adjustment for inflation and no statistical tests for equality of means. AN that is presented is the raw averages and the number of new public schools that comprise each average. Aver a e Square Foot Cost Before 8 Schools After #Schools I Legal Change Cost Change Kentucky: No-Law to Law $86 17 $86 15 Enactment s0 Ohio:Law to No-Law $77 7 $90 18 Re eal $13 Michigan:Law to No-Law $83 14 $94 5 Suspension $11 Michigan: No-Law to Law $101 40 $108 9 Resumption-f— $6 •A 1:Simple Comparison of Mean Square Foot Cost of New Public Schools for the 12 Months Before and After a Policy Change Table Al shows that for Kentucky, the mean square foot cost of 17 schools built in the 12 months prior to the application of prevailing wage regulation to public school construction equaled the mean square foot cost of 15 new public schools built in the 12 months after the application of prevailing wages. In Ohio, there was a $13 increase in the mean square foot cost of new school construction subsequent to the repeal of the application of prevailing wages to public school construction. In the case of Michigan, when the law was suspended, the mean square foot cost of new public school construction rose in the subsequent 12 months by $11 per square foot When Michigan reapplied prevailing wages to public school construction, the mean square foot cost rose again by $6 per square foot Other factors that might influence these changes include inflation, tightening construction markets, and the mix between urban and rural schools. These factors are controlled for in the econometric model presented in the main body of this report I California's Prevailing Wage: AN INVESTMENT IN QUALITY State Building & Construction Trades Council of California, AFL-CIO 921 11th Street, Suite 400 Sacramento, California 95814 (916) 443-3302 March, 1993 Jerry P.Cremins Richard Zampa President Secretary-Treasurer i Table of Contents 1. An Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2. History of the Prevailing Wage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 3. The Economics of the Prevailing Wage . . . . . . . . . . . . . . . . . . . . . . 5 4. How the Prevailing Wage is Determined . . . . . . . . . . . . . . . . . . . . . 10 5. A Sound Community Investment . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 6. The Impact of Repeal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 7. Is the Prevailing Wage Racist? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 8. Making It Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 9. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 10. Appendix 1 . . . . . . . . . . . . . . . . . . I . . . . . . . . . . . . . . . . . . . . . . . 23 11. Appendix 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 1. An Overview In 1931, with Herbert Hoover still in office, lawmakers in Washington, DC. and here in California took action to ensure reasonable competition in bidding for federal and state construction projects. The federal law was known as the Davis-Bacon Act. The California law is among the so-called "Little Davis-Bacon" Acts passed in 40 states over the years. The basic purpose of the prevailing wage is to restrain government's use of its tremendous economic power to distort the working of the free market in the construction industry. Clearly the standards set by the government's participation in the construction industry--federally-financed or assisted construction currently accounts for about 30% of the industry's activity--have a profound effect. Few communities would be invulnerable to considerable economic and social instability if either the federal or state prevailing wage laws were to be repealed. State and federal prevailing wage laws are comparable to other legislation--an ti- trust, for example--which counteract distortions resulting from the concentration of economic power. Under these laws, government agrees to forego the economic scattershot of market power and to rely on wage rates set in the private sector. A separate but related function of prevailing wage legislation has to do with the nature of public works construction. The competition for federal and state contracts is ordinarily intense. But because specifications in these contracts are frequently quite precise, competition seldom rests on product design. Under such circumstances, where the opportunities to trim costs are limited, irresponsible contractors are naturally inclined to underbid their opponents and increase their profit by cutting wages. The prevailing wage, in other words, guarantees stability. The prevailing wage guarantees that the local contractor, already committed to paying locally prevailing wages, has a fair chance to compete for government jobs. And they help promote efficient, high-quality construction work on government projects by prohibiting wage 1 cutting practices likely to lead only to the employment of less qualified and less skilled workers. Over the last 15 years, state and federal prevailing wage laws have come under regular attack, primarily from non-union contractors such as the Associated Builders and Contractors, and from allied organizations such as the Business Roundtable and the Chamber of Commerce. Unable to win repeal or significant modification of federal law, they have focused primarily on state legislation--in 1979 for example, state chapters of the ABC and the AGC launched repeal actions in 24 states, often with the assistance of the Chamber. In California, efforts to repeal or weaken the prevailing wage law occur regularly. One strategy is to argue for the Department of Industrial Relations' regulatory authority to change the way in which the prevailing wage is calculated, and to take other steps to weaken its enforcement. The basic argument of those who would undermine California's prevailing wage laws is that they artificially increase constriction labor costs, and thereby cost state and federal taxpayers monies which could be better used for other purposes. These charges are based on discredited methodology, and incorrect economic assumptions. They ignore the economic and social benefits resulting from hiring fully qualified and fairly- compensated workers. But they are nonetheless extremely tempting at this particular historical moment for states like California, in desperate need of economic growth and a short-term budget fix. It is therefore increasingly important to document the logic of prevailing wage, to demonstrate its social and economic usefulness, and to expose the inaccuracy of these allegations. This packet of information on the California prevailing wage laws will demonstrate: • prevailing wage laws are a sound economic investment that benefit both the state as a whole as well as local communities. • prevailing wage laws bring stability to the industry and are supported by responsible contractors. • prevailing wage laws are not inflationary, and that attempts to prove otherwise are based on faulty and/or dishonest research. prevailing wage laws should be protected from frivolous tampering and should, in fact, be strengthened. 2 2. A Brief History of the Prevailing Wage "The essence of the thing as I see it is: Is the Government willing for the sake of the lowest bidder to break down all labor standards and have its work done by the cheapest labor that can be secured and shipped from state to state?" Ethelbert Stewart, Commissioner of Labor Statistics under President Herbert Hoover Prevailing wage laws governing state public works construction were first adopted in the late nineteenth century. By the time the Davis-Bacon Act--governing federal construction--was signed by Herbert Hoover in 1931, such laws had been adopted in urban industrial states like New York, New Jersey and Massachusetts and in more rural states like Kansas, Oklahoma, Idaho and Arizona. While a number of states, California among them, adopted prevailing wage laws during the 1930s, others adopted them much later--Pennsylvania in 1961, Delaware in 1962, Michigan in 1965, Wyoming in 1967, Louisiana in 1963, Arkansas in 1969 and Minnesota in 1973. These laws vary widely as well in their provisions -- with respect to the types and sizes of projects requiring payment of the prevailing wages, the formula for determining the prevailing wage and the extent of applicability. Prevailing wage laws, in other words, were not adopted in response to particular historical circumstances, like the Great Depression. They were a solution to basic, underlying and permanent issues of instability and cut-throat competition in the construction industry. Representative Robert L. Bacon (R-NY) first introduced a federal prevailing wage bill in 1927 because of his concern over itinerant contractors who were bidding successfully on federal public works programs and producing work of questionable quality. The bill was enacted four years later, but only after the normally chaotic conditions in the construction industry had been made intolerable by the Great Depression. The 1931 bill was a very Iunited measure. It covered only federal building projects--exempting roads and dams--and left determination of the prevailing wage in the hands of the contractor. The original law also made no real provisions for enforcement. Amendments passed in 1935 extended coverage to all public works projects, empowered the Labor Department to "predetermine" local prevailing wages for specific jobs, and set up stronger enforcement mechanisms. While the Act applies directly only to federal projects, it has been incorporated over the years into scores of statutes involving federal grants, loans and/or guarantees for construction activities carried on by state and local governments and their agencies. California's "Little Davis-Bacon" law was also adopted in 1931. Found in the California Labor Code under various sections, it requires the State to set wage rates for workers in all crafts and classifications on government-funded public construction projects. It applies to all public works jobs on all levels of government. Prevailing wage rates include contributions to pension funds, health benefits and injury compensation, as well as contributions to the maintenance of apprenticeship programs which provide ongoing pools of skilled craft workers for the industry. The law further requires that rules and regulations be adopted by the Department of Industrial Relations to administer and carry out the statute. 3. The Economics of the Prevailing Wage "...A construction job using unskilled labor will result in cost and time overruns. Why? Well, there's the matter of unskilled workers taking longer to do the work. Then there's the problem of the job not being done right in the first place; thus, other contractors, the ones who should have done the work in the first place, will have to come clean up after the "cheaper" contractor. " Philip M. Vermeulen, Executive Director California Association of Sheet Metal and Air Contractors National Association The argument against prevailing wage laws is that they raise the costs of public construction to benefit the construction worker. This criticism, however, is based on "research studies" which are characterized at best by poor methodology and interpretation and at worst by outright dishonesty. For example: In 1992, Assemblyman William Filante introduced legislation to exempt school construction from the California prevailing wage, arguing that such action "could lower total project costs by between 10 and 15W. This "savings", Filante said, "translates into 2,160 new classrooms serving a total of 75,600 new students without raising taxes, incurring additional bond debt or shifting general fund revenues at the expense of other vital programs''. It's a seductive argument, until we begin to look uncler the surface. Filante's figures are based on Prevailing Wage Laws and School Construction Costs, a study by Dr. Armand Thieblot, commissioned by the Associated Builders and Contractors' "Merit Shop Foundation" and published in 1978. Thieblot estimates that repeal of all state prevailing wage laws would save a total of $239 million per year on school construction. While this figure certainly sounds impressive, the methodology on which it was based is so simplistic as to be ludicrous. Thieblot's study is based on a gross comparison of construction cost per classroom during the period 1968-74 between states where prevailing wage laws are fully applicable to school construction and those where prevailing wage laws are only partially applicable, inapplicable or do not exist. Average costs in the "prevailing wage states" turn out to be 13% higher than in the "non-prevailing wage states". On this basis alone, he concludes that school construction would cost 13% less if all state prevailing wage laws were repealed. The approach is obviously severely flawed. It is based on the idea that any difference in construction costs between states must be exclusively due to the presence or absence of prevailing wage laws. Yet many other factors must be considered. A more careful analysis of Thiebiot's own data suggests that school construction costs in states with prevailing wage laws and states without such laws are actually indistinguishable one from the other. For example: material prices vary cozrsiderabiy around the U.S. which will certainly be reflected in school construction costs. CIimate differences also play an important role--the need for air conditioning, heating and insulation will certainly differ among the various regions of the country, as will the costs of these amenities. Urbanization will be another important factor--schools are more expensive to build in cities than out in the countryside. Yet another factor which Thieblot ignores are the considerable regional differences in the quality of school buildurQ�s, reflecting differing local customs, needs and financial resources. It is also necessary to separate the data for elementary and secondary schools, rather than lumping them together as Thieblot has clone. Since secondary schools cost more to build than elementary schools, a state which happened to build a large number of secondary schools and a few elementary schools will appear to have higher costs than a state which happened to build more elementary schools. Additionally it is necessary to adjust for the effects of inflation, which drove school construction costs up by 60% in the seven year period Thieblot studied. Schools in one state might appear to be more expensive than schools in another simply because of the fact that the first state did more of its work toward the end of the period, rather than at the beginning. If, with these adjustments, states are ranked according to the cost per room of school construction, no meaningful correlation between prevailing wage laws and construction costs emerges. In fact the correlation between school construction costs and prevailing wage laws is nearly random. 0 Nine of the twenty states with the triUher per classroom costs have prevailing wage laws which are only partially applicable or not applicable to school construction. 6 • Of the twenty states with the lowest per room average costs, half have prevailing wage laws which are fully applicable. If we want to identify a variable which better explains the state-by-state cost difference, geographic location and climate seem 1111IC11 more influential. Take New England, one of the coldest regions in the country. Even though all six New England states fall within the top eleven in terms of costs, only three of the six have fully applicable prevailing wage laws. The other three are as costly without any wage standards. In 1980, Steven G. Allen and David Reich studied school room construction costs in the same period of time, on a state-by-state basis. They found that among those which paid most for construction of new secondary schools were 12 without prevailing wage laws. Significantly enough, the cost per classroom for the State of California placed it seventeenth from the bottom. The same pattern of fundamentally dishonest research and distorted conclusions reflected in the Thieblot and other studies is repeated in one state after another. In 1988, for example, the Associated Builders and Contractors promised Massachusetts voters a $212 million savings if they were to repeal that state's Little Davis-Bacon law. The figure was based on a study by the far-right Massachusetts Foundation for Economic Research. The Massachusetts Building Trades in turn commissioned the non-political and highly respected Data Resources, Inc./McGraw Hill to analyze ABC's claims of inflation and prepare an assessment of the impact of repeal. DRI found that ABC's estimate of a $212 million savings was inflated by as much as 400%, and was gathered from anecdotal evidence based on a "confidential" estimate provided,by one unidentified contractor of his possible cost savings on the construction of one unspecified small building somewhere in the Boston metropolitan area. More recently, the University of Indiana's Institute for the Study of Labor In Society told the Indiana House-Senate Interim committee in 1987 that "a review of the literature concerning the impact of prevailing wage laws on costs indicated that the Associated Builders and Contractors aSSU111ptIU11 of a 20'% figure for saving in Indiana is wholly unreliable". The flaw tanti-prevailing w Q a1 .v tl a a u 1 es thatpaying 1 e fl 1n he a e � 1_ument Is 1 t It ss 1i ., � � anything more than the lowest possible wage rate will automatically lead to excess cost. In fact, there is strong evidence to suggest that more highly skilled workers attracted by prevailing wages have a significant advantage in productivity which more than makes up for the difference in wage levels. 7 If one individual earns 20% more than another, but can complete 25% more work in the same amount of time, with fewer errors, then employing the more highly-paid person will be more cost effective--it will actually save money. Several statistical studies have confirmed that these circumstances are, in fact, the norm in the public works construction market. This relationship between wages and productivity was captured in a recent study conducted by the International Union of Operating Engineers, verified by an independent Washington-based statistical analysis firm, Ruttenberg, Kilgallon and Associates, Inc. The union's study was based on Federal Highway Administration data on the 10 states with the highest dollar volume of federal aid for 1987-1991. The states represent nearly half of all highway and bridge work in the U.S. The study included four states where the workforce was less than 5% union and six states where the workforce was over 8017c union. The "non-union" states were Texas, Georgia, Florida and Virginia. The "union" states were Illinois, Pennsylvania, New York, Michigan, California and Missouri. The study concluded that: • the average wage cost in the high-wage states to build a mile of highway is 11% lower. • the high-wage states completed the worn: with 5617o less hours. • Florida, Georgia, Texas and Virginia (low-wage states) required 136 million man hours to build 5,109 miles of highway. f California, Illinois, Michigan, Missouri, New York and Pennsylvania (high-wage states) required 77 million manhours to build 5,216 miles of high way. • the high-wage states built 74 more miles of roadbed and 32.8 more miles of bridges for $457 million less with a wage package more than double that of the low-wage states. • the total cost per mile averages $1.35 million for the low-wage states and $1.21 million for the states with higher wages. The difference was time, with higher- paid workers doing their work in 561% of the nnanhours it took lesser paid workers. Those who would repeal prevailing wage laws, in other words, have never made a compelling case that such repeal would in any way ease the budget crises which are so pervasive in virtually every state in tine country. On the contrary, a convincing case can be made that higher wages and better conditions produce more cost-effective construction. "Repeal", according to Representative William Clay (D-MO) would "undoubtedly have the effect of increasing the profits of a few employers, but it does so at the expense of both workers and taxpayers". 9 4. How Prevailing Wages are Determined "The definition of the prevailing rate will not be changed from the modal rate. The existing process has been used to determine the prevailing wage rate for many years and appears to be the most equitable and adequate measure of existing rates". Ron Rinaldi, Director of Industrial Relations under Governor George Deukemejian, 1988 Authority for determination of the prevailing wage in California is vested in the Director of the Department of Industrial Relations, wlio is appointed by the Governor. The actual calculation is performed by the Prevailing Wage Unit, a five-person office within the DIR's Division of Labor Statistics and Research. The Prevailing Wage Unit sets wages according to the "modal" system, established during the administration of Republican Governor Goodwin Knight. The modal rate, as defined in the law, is: "The basic hourly rate being paid to a majority of workers engaged in the particular craft classification or type of \vork within the locality and in the nearest labor market area, if a majority of such workers is paid at a single rate; if there is no single rate being paid to the nla'orlty, then the single rate (modal rate) being paid to the greater number of workers is prevailing". In other words, under the modal system the prevailing wage for any craft in any particular locality is the wage being paid to the largest single group of workers doing comparable work within that area. The basic principle behind the use of the modal rate is that the prevailing wage should be the rate received by the greatest number of workers. This principle is completely in accord with the normal meaning of the term "prevailing", which means most frequent". 10 The alternative is use of an average. The use of a strict average to reach wage determinations is inconsistent with the language and intent of prevailing wage statutes, and its use would create serious practical problems. For example, in a situation where wage rates within an occupation are widely dispersed an average would be purely artificial. Where a substantial number of worl<ers receive a wage which is near the high or the low end of the local wage spectrum, the average wag below the rate received by most workers. In sucli a e rate is likely to be above or case, a determination based on the average would not reflect the wage actually prevailing in the. area. It is clear that where the union rate prevails, the use of the modal rate preserves the union rate.. It is also clear that where a non-union rate prevails, the modal rate preserves that as well. That, ultimately is the purpose of the regulation--to sustain and continue the prevailing wage which is set in the private sector. Indeed, in 1985, after an exhaustive review by awarding agencies, business and labor, the Department of Industrial Relations under Republican Governor George Deukemejian came to the conclusion, as explained by DIR Director Ron Rinaldi, "The definition of the prevailing rate will not be changed from the modal rate. The existing process has been used to determine the prevailing wave rate for many years and appears to be the most equitable and adequate measure of existing rates." 1� rate is tantamount to adopting union Opponents argue that the use of the moc �1 p g wage rates. But in California, from March 1987 to the present, the use of the modal rate has resulted in adoption of the union wage rate in less than 60% of DIR wage survey cases. Does the modal rate of determination in California increase wages disproportionately? A DIR survey of 15 categories of %workers throughout the state demonstrates that state prevailing wage rates, set by the modal system, are in fact extremely close to federal rates. Where state rates slightly exceed federal rates, according to DIR researchers, "In most cases the difference...is a result of the use of superseded rates--scheduled increases in the collective bargaining agreement used as the basis for determination, which were not reflected in the federal rates" State rates set by the modal system are, in fact, sometimes lower than federal rates. For example, the rates for plumbers and carpenters in Northern California, as well as for painters statewide, are substantially below the national norm. State prevailing wage laws also offer a simplified appeal process. California Labor Code Section 1773.4 allows interested parties to challenge wage determinations issued by the Department of Industrial Relations and is used quite frequently. The attack on the modal method of determlrratfon is essentially an attack on the very concept of prevailing wage legislation. It is fueled by the same myths discussed 1.1 earlier, particularly that paying anything in excess of the lowest possible rate drives up costs unnecessarily. Departure from the modal method of determination would depress wages in our state, driving both employers and employees out of the market with dire results for the California economy and threatening increased costs over the long term. The prevailing wage must be paid to workers on projects valued at over $1,000. However, legislation passed in 1989 with the cooperation of the California building and construction unions, gave important flexibility to smaller contractors. First, a $25,000 threshold was created for those public agencies which established a labor compliance program (Labor Code Sec. 1771.5). In addition, a change in force account laws now allows for a threshold of $75,000 in cases where public agencies are in compliance with certain bookkeeping procedures. 12 5. A Sound Community Investment "It wouldn't take very long for (repeal) to be felt by the merchants in those communities--that is by the people who own the shops, the stores and the markets where construction workers spend their income. Ray Marshall, US Secretary of Labor, 1977- 1981 In 1927, when he first proposed the law which later bore his name, Representative John Bacon, a conservative Long Islancl banker, recalled the construction of a veterans hospital in New York which "...was let to a firm from Alabama who brought some thousand noli-union laborers from Alaballla...into Illy congressional district. They were herded onto this job, they were housed in shacks, and were paid a very low wage. "Of course that meant that the labor conditions in that part of New York State where this hospital was to be built were entirely upset." Ancient history? Irl 1982, opponents of prevailing wages persuaded the Kentucky legislature to exempt schools and local construction projects from the coverage of prevailing wage laws. These opponents made grarld promises of lower construction costs and major tax savings. But here's what really happerlecl: Y unemployment in the construction trades rose to a full 60% as fly-by-night contractors brought in untrained help from outside the state. • 1,783 new contractors, many of tliem from out of state--set up shop, further destabilizing an already unstable industry. • corruption became rampant iii the construction business, grand juries convened througliout Kentl.lCky to consider charges of bict riagirlg and conflicts of interest. A mountain of anti-corruption legislation was introduced to undo the damage. • not only were none of the promised savings ever identified, but Kentucky had to implement a $1.3 billion tax increase in 1990. In 1992, facing a $400 million 13 shortfall, the victorious candidate for Governor made a return to prevailing wage a key platform of his campaign. While prevailing wage laws are of obvious benefit to the workers immediately involved, the ripple effect filters through all our communities. Fairly paid construction workers buy the homes, pay the taxes, support the merchants and enhance the quality of life in communities across the state. Eliminating or weakening the prevailing wage law will wipe out high-wage jobs in our state, something no California community can afford. As former Labor Secretary Ray Marshall has pointed out, "It wouldn't take very long for (repeal) to be felt by the merchants in those communities--that is by the people who own the shops, the stores and the markets where construction workers spend their income". It goes well beyond the wages which are paid. The California prevailing wage includes contributions to pension funds, health benefits, vacation and holiday compensation, injury compensation and other fringe benefits as well as contributions to the maintenance of apprenticeship programs which provide new pools of skilled labor. This is particularly important because there is no such thing as "permanent employment" in the construction industry. The construction worker, union or non-union, spends his or her working life on a series of temporary jobs, from one to three days, to as much as six months. He or she faces extended periods of unemployment. Because of the sporadic nature of employment, fringe benefit plans traditionally have eligibility provisions which make it possible to bridge the period of unemployment and prevents the worker from becoming a public charge during those periods. The construction worker who does not receive health and welfare, pension, vacation, holidayor training benefits becomes a net user of public service during periods of unemployment. The construction worker who enjoys these benefits becomes a net contributor. For example: In 1984, according to a study conducted by Lionel Richman and Julius Reich, the City of San Bernardino awarded contracts totaling $8.8 million, generating 158,000 work hours in construction. These hours produced: • $325,000 paid to community hospitals, doctors, dentists, pharmacists and other providers of health benefits. It represents a repayment back into the community of $.83 on each dollar contributed. 0 $553,570 of pension benefits paid directly to pensioners in the community, a repayment of $.94 on each dollar contributed. 14 • $234,300 in vacation benefits paid directly into the community, a repayment of $.96 on the dollar. These payments of $1,113,290 carry a powerful economic multiplier effect. "A comparison of specific studies done for specific regions within the state of California, �- applied to the city of San Bernardino, indicates a multiplier of expenditures of approximately 2.4 in 1981 and 1982... Applying this rationale would produce a direct economic benefit to the City of San Bernardino of S2.7 million from state prevailing wage laws", according to the Richman and Reicli study. Prevailing wage protection is also beneficial to the industry and the community in general because it helps ensure that wages and benefits will be sufficiently high and sufficiently stable and predictable to allow the recruitment, training and retention of a pool of skilled workers able to meet the needs of any contractor who undertakes a project within the area. The transient contractor may not have such a stake. But the local industry does. For this reason there is a substantial community interest in ensuring that there are adequate incentives to keep these skilled \vorkers from drifting away. 15 6. The Impact of Repeal "The only clear result of repealing the Massachusetts prevailing wage law would be lower wages for certain Massachusetts residents. " Data Resources, Inc./McGraw Hill Study of the Economic Impact of Repeal of the Massachusetts Prevailing Wage Law, 1988 Over the last 15 years, state and federal prevailing wage laws have come under sharp attack, primarily from non-union contractors and their associates and allies. Unable to win repeal or significant modification of federal law, they have focussed primarily on state legislation--in 1979 for example, state chapters of the ABC and the AGC launched repeal actions in 24 states, often with the assistance of the Chamber. They have actually prevailed in seven states-- in Florida (1979), Alabama and Utah (1981), Arizona (1984), Colorado, Idaho and New Hampshire (1985). Voters have the right to aslc what the impact of repeal or substantial weakening of the prevailing wage law might be. In 1982, opponents of the prevailing wage persuaded the Kentucky legislature to exempt schools and local construction projects from coverage. As noted in the previous section of this report, the results were disastrous--unemployment increased and none of the promised savings ever materialized. And in 1988, when the Massachusetts prevailing wage law was challenged on the ballot, DRI/McGraw Hill prepared an economic simulation which projected a total wage loss of $196 million, tax and fee savings of S91.3 Illlllloll, or 0.6% of total expected tax revenue of $15.4 billion. The net employment loss was 600. "While it does appear", the research firm conclucled, "that some nominal tax savings can be obtained, the 0.6% reduction in taxes indicated in the "most likely" scenario would take place at the cost of increased instability in the construction labor market; fiercer competition for wort: from out-of-state contractors and workers; and a lower standard of living for Massachusetts workers and their families. 16 "These costs would be incurred even in the event that the tax savings are less than indicated by the quantitative analysis, as increased contractor profits and decreased labor productivity would keep total cost reductions well below the amount indicated by the apparent savings in unit labor costs, while possible increases in state unemployment compensation and other social service expenditures would further offset the apparent initial savings. "The only clear result of repealing the Massachusetts prevailing wage law would be lower wages for certain Massachusetts residents." On November 4, 1988, Measure 2, Nvttich would have repealed the Massachusetts prevailing wage law, was defeated 2-1. 17 7. Is the Prevailing Wage Racist`.' "The Heritage Foundation seeks to justify repeal of (federal prevailing wage laws) by asserting that reducing the wages of minority and female workers is somehow in their interest--rather than the economic interest of the corporate sponsors of the Heritage Foundation. " --Representative Ron Dellums (D-CA) "Enactment of the federal Davis-Brecon Act", according to Robert Cline, "was rooted in prejudice and motivated by racism, to keep lower paid workers from entering the industry. It was to keep black crews from coming up from the south and taking previously white jobs". The only documentation for this conclusion W IS a single comment made during a discussion in the House of Representatives in 1931 by a congressman from the south who stated that an Alabama contractor had hired "cheap colored labor that he transports (and puts) in competition with white labor throughout the country". In the 17 pages of the Congressional Record covering the debate on the Davis-Bacon Act prior to its passage in the House, there is not another remark of that sort. Exploitation of low-wage mobile labor was not particular to any region or to any particular racial group. Testifying in 1931, William Green, President of the American Federation of Labor, placed in the record a summary of the complaints he had received from 16 different areas. His testimony makes it clear that the contractors who presented problems were not merely Southern "invaders of the Norte]". One complaint involved a contractor from Raleigh, NC in connection with a federal project at Fort Bragg. Another involved a Louisville contractor on a post office facility in the same city. Others involved one contractor from El Paso on a project in the same city; a New York City contractor on a Utica, N.Y. project; a Bellingham contractor in Blaine, Washington; and a Pennsylvania contractor on a New Jersey ps oject. Whatever may have been Representative Bacon's original motives, the fact is that by the time Congress was ready to deal in earnest with the prevailing wage proposal, the problem was by no means considered exclusive to itinerant Southern contractors and "imported" labor, black or white. 1s But rather than dwell on empty historical debates, however, we are much more concerned with the effect of prevailing wage laws today in closing the gap between ethnic groups. According to Representative Ron DellUmS (D-CA), "The truth is that minority and female workers have entered construction in increasing numbers over the past decade. Because they are often the newest-members of the industry, these workers are particularly vulnerable to the wage-cutting practices which the Davis-Bacon Act of 1931 is designed to prohibit". a _ e LIP nearly half 46% of the For example, members of minority ,1 Daps mal, i 1 y ( ) population of apprenticeship and training p]'Ogl'S111S ill Callf0l'nla, which are financially supported by contributions mandated under the prevailing wage law. That's why, in Dellums' words, "African-American, Hispanic, Native American and other minority workers, as well as women and young workers especially need this law. For this reason the NAACP, the National Women's Political Caucus, the Navajo Tribal Council and the Mexican American Unity Council have all endorsed the Davis-Bacon Act". 19 S. Making It Work The most effective enforcement tool for any law is voluntary compliance due to the fear of getting caught and paying penalties or going to jail. If you could steal money from your employer without fear of getting caught or going to jail, the government could not hire enough policemen to catch all the thieves. If you are caught stealing money and all you had to do was pay it back, with no further penalties, chances are there would be many more thieves around than there are now. Unfortunately, under many prevailing wage laws, a contractor can steal money from his or her workers with little fear. To demonstrate the lack of deterrent, the National Alliance for Fair Contracting in 1991 sent letters to all SO states requesting a list of debarred contractors. Of the 16 states responding, only two provided a list of debarred contractors. The bottom line is that state budget appropriations are insufficient to properly enforce prevailing wage laws. As a result, organizations are springing up across the country to insure compliance. Fair contracting and work preservation groups are, in effect doing the work of the government by investigating and reporting violations of prevailing wage laws. These groups are funded by both labor and management. There are currently some. 30 of these groups in the U.S. One such group, the Center for Contract Compliance (CCC), covering eleven counties in Southern California: • handled 3,405 worker prevailing wage. complaints .in 1990. • protested 61 bids of which 11 were successful in 1990, protested 93 in 1991, with 27 successes. • investigated 999 job sites in 1990 and 1. 124 in 1991. • had three contractors cebarred in California in 1990 and one in 1991, with two of those serving prison terms for cheatinU workers. • filed successful bid protests of $116 million in 1990 and $161 million in 1991. -)o A similar example is the Foundation for Fair Contracting, covering 56 counties in Northern California, whicli has been jointly funded by labor and management since 1985. From March to August 1992, the Foundation monitored 159 job sites and actively investigated 145. It visited various federal, state and local government agencies 328 times. It filed 69 complaints regarding the improper payment of wages, forcing the assessment of $141,035 in fines. Clearly, enforcement of prevailing wage laws suffers from inadequate state funding. The record of at least two non-governmental agencies discussed above, indicates the degree to which unscrupulous contractors can still evade the law with impunity. 21 9. Conclusion The prevailing rate was formally established sixty years ago under the republican administration of Herbert Hoover toward economic: stabilization of the construction industry. The concept was that construction, which is transient by its nature, was in need of a balanced base in order to avoid the corrosive erosions of fly-by-night contractors and developers intent on high profit at the expense of product quality and worker income. Through the years, under the rubric of prevailing wages, American construction has developed a skilled work force recognized world-wide for its safety and speed of construction record. j This is in part due to the ancillary presence of an apprenticeship program, inutually funded, coordinatecl and run by employer and work force, that has become a symbol of excellence everywhere. Efforts at cloning by organizations such as the Associated Building Contractors have failecl in the absence of training proficiency developed through years of practice. Opponents of prevailing wage argue that the system is inflated. Yet, their argument in all instances is devoid of scientific studies to make their case. The figures 15% and 20% and higher appear again and again, mostly in press statements, but in no instance is research information of consequence supplied. In California, where prevailing rate is establislrecl regionally through surveys, by the Department of Industrial Relations, these figures appear regularly, though the prevailing wages vary widely. The argument also is that the union rate prevails. Yet, again in California where the modal rate pertains, the prevailing rate approximates the union rate in only 60 percent of the cases. In conclusion, research for abandoninent of the prevailing rate makes no legitimate case except for that of economy, which is always usable. However, economy at the singular expense of the construction work force is neither equitable, good economics nor Good policy. APPENDIX 1 Highway Statistics, 1991 DISTRIBUTION OF COSTS ON FEDERAL AID HIGHWAY CONSTRUCTION CONTRACTS OVER $1,000,000 EXCLUDING ALL SECONDARY PROJECTS REPORTED DURING CALENDAR YEAR 1991 AS COMPLETED 1 Otner 'ss. i 22 Equipment ••••••••• : ................. 21.8 j Steel 6.5 VCtef^Cls t C C2^s \ \\ \ \\\ m . Ceent \\\\\\\\\\\\\\ \ 5.2 Acare^Ctes Wcces q 2 .7 1) Gross earnings of contractors' employees in the following classifications: Administrative and Supervisory, Skilled, Intermediate and Unskilled labor. 2) Aggregates consist of sand, gravel, slag, crushed stone, etc., for use in bases, portland cement concrete and bituminous surfaces and portland cement concrete structures. 3) For both roadway and structures. 4) For various types of bituminous surfaces and bases. 5) Includes structural, reinforcing, culvert and miscellaneous steel. 6) Equipment includes fuel and lubricants 2.3%, but excludes operators' and mechanics' wages. Overhead includes contractors' on-site expenses such as mobilization, office rental, taxes, licenses, insurance, etc. SOURCE: Federal Aid Division Office of Engineering, FHWA (202) 366-4661 24 APPENDIX 2 TENT-STATE TOTALS LOW WAGE FAIR WAGE NON-UNION (TX, GA, FL, VA) (IL, PA, NY, MI, CA, MO) Labor Hours 136 Million 77 Million Gross Earnings $1.270 Billion $1,544 Billion 4 Year Total $6.9 Billion $6.3 Billion % U.S. Total 25.5 23.3 Roadbed Miles 116.03 148.8 Bridge Miles 2.27% 2.55% TOTAL MILES 5,108.88 5,216.01 TEN-STATE PROJECT PER MILE AVERAGES LOW WAGE FAIR WAGE (NON-UNION) (UNION) Average Wage $9.31 $19.99 Man-Hours Per Mile 26,651 14,810 Labor Cost per Mile $248,618 $296,077 TOTAL COST PER MILL. $1,348,098 $1,213,569 NOTE: 0 The Fair Wage state total cost per mile is 1017c, lower than the Low Wage state total. • The Fair Wage workers completed the wort: with 56% of the man-hours. • Fewer man-hours also means fewer equipment hours and is reflected in the lower total cost per mile of the Fair `Wage states. 26 O < O O co O O O o CD cD 1 W o o s mn co T N p a G� N Ul ••• -` �• f� Q -64 , o. o c � j O co CD �, W -1 O C!1 , :1 T :1 �N C1 1 V/ , O ffl ffl�� E3 rf , � N f/�1 a 1 0 0 W W O r* SUMMARY The top 10 states represent $13.2 billion or 48.8'( of all highway and bridge work in the United States. These states also reflect a near equal proportion of low-wage non-union to fair wage union dollar volume. Four states - Texas, Georgia, Florida and Virginia - represent the low-wage non- union group (less than 5% union). Six states - Illinois, Pennsylvania, New Yorl:, Nlicliigan, California and Missouri - represent the fair wage union group (more than SO'7 union). At first glance, most of these statistics appear somewhat equal. A careful analysis reveals some startling facts: • The union states built 74.4 more miles of roadbed and 32.8 more miles of bridges for $557 million less with a wage package more than double the non- union states. • The argument that low non-union wanes are cost effective is simply not true. In fact, the opposite is true. • While we seem to holding our market share in most of our traditional union states, a much larger federal highway dollar volume is being shifted with the Population migration to the non-union soutlern states. The chart shows 25.5% of all federal highway dollars were spent in just four southern states. If we project that percentage to the new Surface Transportation Bill, they would receive 25.5% of the $119 billion authorization, or S30.3S billion of authorized highway work over the next six years. t STATE PREVAILING WAGE LAPS AND CONSTRUCTION COSTS : A REPLY TO PROFESSOR THIE3LOT x Losiva Ground: Lessons from the Re-peal of Nine- Little Davis-Bacon" Acts Peter Philips, Garth Manaum Norm W ai tzman, and Anne Yeaalle Working Paper Economics Department University of Utah February 1995 THE ` U RSITY OF LTTAIJ 3 r V sLa=e ors , =_1� nc wage laws represent an imccrt-,-z ceunteroart to t e =ederal Davis-3acon Act. t_ oresent, laws c= s ara cn the bc_�s in 40 states . The -3esL , in Kansas , _] ' _ - J. _s -Ja _J 5=3=2 w- _. - _.. =_"43 cony= and o_=e wa,, their aoolicabili=y to local government units , etc. rr *-=-eVer the ----'--ences , the general RC__ 3t_C. be�-_4.-d all = eselaws is - _y same : _o ensure the eccnc ,i_ o- the CCVer._. .._._t _s not used to d= C_ ��C =:� wa'^' anc la-or stancares of _:a loca, coru:�ur.ity. _he little 0av_s-..accn acts can also ce as _=ws =o t_ea-men_ _.,_ construction contra-=_rs• -_.. ----Se pretra___.g wace =�Cu:re-entS , ccntractorS cot::'- --ed =. -j i i.: - n foal ^a :JC^ICCete r P e t...+ .7 -.- _nC r- C..- a.. e _v •�e a ._ e_s ha - �__...._ rages -o ---e-- e• _ -o_ _ -_ - for 1-u lic --o,ect3 , wh-4 h generally must be a-:arded to =:^e lowest C- --=r• -n __.e aCsance :re'Ja_, ; nq ::a;e -a;JS , COS:��e'___C1 =0= PT1 a coy. g,. .�_.T_.._ cry�e.._s wcu_.. _.._.. __._� - an wages the furthest i n c-�'e- i w_tZ the lcwest bid. s_a; , _� _ to come n ,I e impositicn o: a :lcor under wage levels, ccntr ac-ors are fcrced to compete on the basis C` their s'c:'_1 and e`2i-ie:CY. - :age - _.,section _s also bere:___a_ to =^z n - action i^ 'astr:r and =he cc-r.ur:_=y in gene al 7 - factors must be considered. It is well known that prices vary considerably from place to place in the U.S. , a fact w ich will: certainly be reflected in school construction costs. C1i:�ate differences also play an important role in the pricetage Iof a new school. The need for air conditioning, heating, and insulation will certainly differ among the various regions of the ccuntzv, as wi11 t:^.ei- costs . Urbanization will be another imocc:ant facto: since schools are likely to be much more expensive to build in cit: than out in the countryside. Yet another factor which Th'_biot ignores are the considerable regional differences in the quality and amenities of school buildings , reflecting differing local customs , fin-=racial resources , etc. 2/ While_ i_ is impossible to get an accurate measure of z^e impact o- all these diverse regional factors , it is pcss_cle to make rough adjust.-zents along these lines . Before doing ' s, how- ever, it is important to make two preliminary adjustments to Thiebl figures to remove potential biases resulting frcm his use of highly aggregated data. First, it is necessary to separate the data for elementary and secondary schools, rather than lumping them together as _hieblot An indication of the extent of these regional variations in amenities is provided by data published annually in School !�anageme giving the percentage of schools completed having various spec:aliz acilities (auditorium, cafeteria, gymnasium, language jab, et-::. ) a a region by regior. basis. For example, in 1971, the pe__entace of new secondary schools with auditoriums ranged from 63 . 5i region (N,, NY, PA) to 23. 5i in region 9 (kF, CA, HI, CR, WA), t a percents w_` ^ �=ic=^s ra^��d frcm 9?. 9� in resica 2 to 5Z.7% in rsy-cn 3 -h .y.. '. , Cc, :A, mT, %-T,7, Nm, uT, WY) . these Gaza a:e nc :ed f _ndi' _C-2 u al s t a as , and t::u:3 it 15 10t pCSS:--e to data for t: _3 fag:.._ . A tax Of _..e s p_n-Ja___ng wage laws are t^e single most factc_ wage level s , as C Thieb1Ct c 1 a s . e n11St ex,=Iai_^. «;^.v e -Cs.: _.____3'e Ca�Sd; :/ «a:e StariCa_=S %S . _ , i , .� _ _ _ .-,a _: n:T laws az3 ._ _'Ji. - _Csts , =fie affect �. .giro - _,� m� -V and Ceccnaa_V ^^.I "2_ �; _`c_ bctz e:e. en_ SC.^.ccis . .e= t e figures Show that t e additior.al cost cL elementary SC cols 4n States with li_t_e OaVr.S- aC'Jn is th ee times mc=e _na : t e a 1t_Cnal CCSt fcr Secc-4-ary SC.^.CC15 . It is -`--,tile to s_st_st-cs to e =_ain anoma_v. ,`_ e__.._ ills in see_. eva :I wa-2 -.GINS as t -e c4aC_S-. %e var4 able. ^`-_s can be s:.c,w,; in =ague= way. If' states are ,a..{-d to the cost oe= rocm o` sc ccl ccnst_uction , as in no mea liarful cc=re l at; en with prevailina wage 'laws emerges . act, there is almost a random correlaticn between school =str ctic _osts and prevailing wage laws . NL'ire of the twe.^.t_r states _=Ies- ?er c_.assrocm cgst ::a-re wage laws which are only 'a_t_al_y Cr act 3aD. _=a:la a= all _c SchcCl =nst.uctio'- - the zwe_^-_.I Stains with the 1-- e5t ?e: _cCZ a"e:aae ccs' , h- al- ha-,e prevailing wage laws whams are .fully applicable. := we wart to idantifv a variable which better explains ti.e o�'-sta _ test "=e_ ces cec�ra^: lCCati:.a ar.L "ate- _ _e -an - _C seem to :tea =a- mCre _:, S== is _e 2; T3::e New Tne_d . _e _ t.,o CCt .3Z5t ��'' C.^.a ti.� `.=:e C^.'...._-"•'. L'ie^ =-•CCy~ ?-� �-�' -2w a d states 'ail the _op eleven =e- c= sc.._ __ 15 - ?-e,.ai l i ng Wage Laws and Overall '."-ace Levels A second argument sometimes leveled at the little Davis-Bacon Acts is that their presence somehow causes the overall indus=y-wid; level of wages within particular states to be higher than it othe_w-- would have been . It ' s hard to understand why this would be true, since -nese laws do not i�npeose new wage levels , but only require that workers be paid no less than whatever figure is locally prevailing. Furt'ner state-funded censtruction is a small part of total construction acti and thus wages on state projects should not have ;auch inf_uenca on the average wage level in the industry as a whole. This expectation is confirmed by the results of a statistical study which indicates that the presence or absence of a state prevailing wace law seers to have no effect on interstate wace differentials.1/ This study was done using a standard statistical technique known as multiple regression analysis. Briefly, what was involved was taking a sample of construction workers (drawn from the Census Bureau's Current Population Survey) and setting uo ecuatic which try to explain the variation is earnings from worker to worker on the basis o` various relevant variables -- workers ' age, education sector of the industry, occupation, urban/ncnurban residence, union me-T.bership, etc. The process was then repeated, with an additional variable included indicating whether or not the worker lived in a state with a little Davis-Bacon Act. The measured effect frog this `— '^hese results are based on research in rogress by Dr. S__ven G. Allen, :forth Carolina State University. t f, t r Contents Executive Summary, page iii The Authors, v I. The :history of Prevailing Wage Laws in the United States, 1 Passage of State Prevailing Wage Laws, 2 Passage of The Davis-Bacon Act, 3 Repeals of So=c State Prevailing Wage Laws, 6 E or:s to Repeal Other Prevailing Wage Laws, 7 EZ.;rs to Repeal Davis-Bacon, 8 II. The Economic Effects of Davis-Bacon Repeals, 11 cutthroat Bidding, 11 A Loss of Earnings for All Construction Workers, 16 A Loss of State Tax Revenues, 17 Regression Analysis of the Decline of Construction Worker Earnings, 21 Increased Employment associated with Lower Wages, 24 The Net Effect of Repeals on Government Budgets, 25 Summary, 31 III. lie Effect of State Repeals of ?revailing Wage Laws on Training, Black Unemployment, and Mutortry _ aruc:patron in Training, 33 he Effect of Repeal on Const^scticn Unions and Wages, 34 T'ne Relation between Repeals and Black Unemployment, 37 A Decline in Training, 40 Market Responses: Training, Turnover, and Careers, 42 National Trends•in Registered Appmnticeszip Training, 4Y Scmmary, 55 IV. Construction Safety Put at Risk, 57 'Why Prevailing Wage Law Repeals Lead to Increased Injury Rates, 58 A Comparison of Injury Rates, 60 ?ae Cost of Injuries, 60 Summary, 63 V. Conclusion, 65 T ae Effects of Repeal of P:-vaiiing Wage Laws, 65 ae Goals of Stats Prevailing Wage Laws, 65 The Definition of a Prevailing Wage, 65 T ae Financial Costs of State Repeals, 66 Other Costs of State Repeals, 68 L Estimated Effect of a Davis-Bacon Repeal, 73 End Notes, 76 Rcfcrcnz:s, 32 i r Executive Summary Like the 1931 federal Davis-Bacon Act, legislation in 41 s',a� �1 required that the however, wa-staz-. -e- overnment-f•snded ceastrsc^ca ?role..s. F1 sr a--Zs aid on ,t.... g � in ,,;, ,,.:._s be p 1 (Nine states never had such a law.) The remaining have repealed their prevailing wage laws. (i++ ations in state experience provide useful Information w: have retained prevailing wages- T•nese vari which :o consider ?robabie eft`ecs of additional Sate epeals or .he pr000sed repeal of Davis-Bacon. �. state reoeals of prevailing wage laws had several effecs. T'nis s�.e found that Y�,n' �, mos. lose',;, he to 5udge: `uas ^ct bene^t.�? n n whose experience was e � `_e n za.�-, and from -Yell of '--� , 1 rg W=it 'a-,v. la reed help dive down ,cg �� ' �.'�/�: L s'�ate h� last sudS,::crl rcorte ar, a:d sales Mx revenues. In a Decade before he : repeal -n Utah, ocnst:l:ction �'erker earniIIgs eve rag about 1=j Yerwat of average non-agrcu:r: al earnin=s. v 1993 con :cn worker ea,-niags ad fallen to 103 percent of the average eam�sgsuf for sh B es but als ULaih o because of a subsequent workers. This decline in earnings is because of lower wages, to a less-skilled construction labor force. Second, also n Utah. the sine of total cost ove.:uas on state road construction hs pled in >re dec:de ;fir ce ope:I in ccmparisor. to the previous decade. The shift to a less-skilled t�s labor.`ems have lowerag labor productivity along with wages — and the great,=-, z— ,�-e, - lessened any pessioie savings in public works cons--ruction costs associated with t-be T'aird, loeki g at all the sta .�s, and controlling for a general downward read in '== :onstr'1c icn earnings, variations in s-tt. unemploy:rent rates, and regional dip ere have =st nccs m wages, re yis in earnings, on ave^ (in 1C0 constric'ron in workers in the nine states at least S1,477 per year ni dollar;. The costs may eventually be higher as :he effects of the more recent repeals to :lire, driving wages Lid ti^..ia rg down further. F.;ur-:h, cen-oilirg for a general downward trend in he amount a�ai�Osr'uc-oIIatr�nrP •• a `-,, •1 P vanations in s',at: unemployment :axes, and regional differences n training repeals gave reduced construction training in those states by 40 percent. - Fifth. minority representation in ;casruc'roa Jainm prog".rns .as fallen even =er than have the training pregr-his in repeal states. Until the various state repeals, minority apprennce:;:lip participation mirrored the minority percentage of each state's population. After .eper.i, mincrnes became signincandv under-represented is c cns=c:ion apprenticeship programs. Si.t-h. occspacor al injuries in :,cnss%c:ien -cse by lc ?ecent where state prevailing wage laws `here rice pied. Based oa hose aadings, we coaciude that :f the federal Davis-Bacon AC we.t :tpealed: Federal income tax collection would.ally ar :east SI .ikon ger year in real = :is every - ea' for the forrseeabie mature. iais is because ;,cas-Lu(=ea wage levels would declne acress ail a _.y ,xceren�. of the hire repeal states —COIISa^scaon e^?iovment levels s*.-:--s 3II�. -ase:. Sri LhC well would not rise retie :gh to OF :is ,revenue 'loss. The :figure for lest ar rever_es may «e•. be hig roil If :he experience of the ame states that never had a prevailing wage law is i r The Authors Garth Mangum (PhD, Harvard University, 1960) is Max McGraw Professor of Economics and Management at the University of Utah. He is author or coauthor of 30 books and numerous _ ' monographs and articles on labor and employment, incla Trade Uraon wand Labor Struggles al and Labor in in 1 Copper, nie Operating Engineers: Economic History of the Post office and Union Resilience in Troubled Times: the Story of the Operating Engineers 1960-1993. In addition to teaching at several universities and serving as an arbitrator in more than 600 labor disputes, he has served the federal government as senior research analyst of the t Presidential Railroad Commission, research director of the Senate subcommiree an employment and rr�.._ ower, and executive secretary of the Varional Commission on Technology, Automation . �y and Eccnomic Progress. Peter Philips (PhD, Stanford University, 1980) is a Professor of Economicsauthor at he University of forthcoming t~ _ ee Worlds o Labor Economics and co Utah. He is co-editor of Thr. f Building Retirement Security: the Central Pension Fund of the Operating Engineers. Philips has published widely on the canning and construcrion industries in Journals such as Industrial and Labor Relation Review, Industrial Relations, Business History, the Journal of Economic History, ' and the Cambridge Journal of Economics. Philips has received awards for his teaching, including University of Utah Public Service Professor and Presidential Teaching Schoiar. Norm Waitzman (PhD, Amer can University, 1988) is an assistant Professor of Economics at the University of Utah and the co-author of Business War on the Law: An Analysis of and The Incidence and Costs of Birth eDefec.ss f. deral Health and Safety L.^,v Enforcement 7"ne >alue of Prevention. WaitL-^nan's work on the economics of safety and health appears in t�1 American Journal o Public Health, L'emography, Inquiry, and the Economic Forum. Ize is Am f l and Public Policy. _writing a book, to be called Physician Su., y Anne Yeaaie is a PhD ca ndidate at the University of Utah who is writing her dissertation on the effects of the 1981 repeal of Ut:lz's prevailing wage law. s 3• s: I. The History of Prev::iling Wage Laws in the United States In February 1891, Samuel Gonpers, president of the A,-ner`c zed ratioof lacer" Tan yet.=s Tooeka, K Haas, to spe31{ on what the local newspaper called u gtopic out teg"slarive aims that earlier, t:^e A — at its o`er ci`3n°n Gabor, free ublic schoel ng, compulsory schooling c o rg la% s, elimination of child : P hour work C��, the re the elimination of convict. labor, and prevailing wages on public works. These proposals were based cn a teller fat * .e A."n c-n !abcr market should consist of highly skilled -vo,.-ers earn decent Nags, with time for fami!•�, and with chllareR free to earn an educaticr.. L^. p':rsuit 31 ft '�$' OIILICaI S`at"may '•n Kansas 3�:led .he 1\vpupticr^. �3•':�'. these aims, Go .p... p -he Popui:st GOm�ers�:; 1 '1' ro , the .�.:lid!cs Da.^,, 'i o'er. -o History as U^. ::e .^.1CC::1P.� oL .^ r., 1 za'I of the Star! I=CUSe pa •Ixvit:-drew an aariier ir.�ita^on for ::.... 'o spe .:ti n wor'.<ers had wro -epr----rated 9oo,oGo Reprtse.^.ta,wes, w;..ch the -=I oc:rolled. Gar cd.s, fallen out of favor with the populists, reportedly because of his belie:that :he trade unions should not form a poll ical party with the Alliance.' The Rapubiicans, who controlled the Kansas Senate, invited Gompers to speak there, and he did. Gerrpers was in Kansas to focus on the eight-flour day. Like other Americans, Kansans in i89I typically worked six days per week, ten 'o tvelve hours per dad'. Lri the older and saddle making, where the work pace .vas slow a:d trades and Cr3I'S, Si:Ch as Car-loge making T .tee Re'Ne3r faC ^.�S �roduc'-01 urcer t:a workers' al_ect-:on, the long work day '.seas olerab.e. .il ' . TI the uSba.^. pli�ng-out S%Stems :II ,�eGleWOi:�, shoes, textiles, and :he I1Ke; in the m:n es; an .-* had Made ltS Or:::ie ob;ective a six-cay weeks and tweive-hour days were grueling. The shor:enad work day and work week with as little cut In pay as possible. In his Toceka speec , Gor.:p-ers declared: and on :hat banner Heat is inscribed, "Eight `zcurs pork, Our ;;gazer floats high to :he breeze at :,ours rest and eight hours for mental and moral inir, ovetnert.iz At that time, w;.en there were no income Supplement progr=s for the poor, low-income parents worked and ilad :o send their children to work to make ends Meet. This prat ice was taxer refer-ec to by a 'ier_. Carolina newspaper editor as it the seed corn." Each .;eneratlon of g ew uo as illiterate as &,-,r pear contemned its or!;prirg to poverty because t;^.e c;llid._n ter_ pa e w. ine ;_!valence of c7eap child labor, whic:t accounted for 5 percent of the rn=uractti..; oratorio, e^•ice sec:or workers, kept xages down d ^e .s90 a:.0 a larger n of s... abcr for. :.Z ;;ors wa. ends ;nee m -egularic^ 'o t,p ours to mark! t �o forced ,tr.. per:;ors put in u e long _ • peer o c1?t a crate,, however .alnful In :he srert �' . of a t fc_ce e:r.;,ievers :rd ^.e workers hadne skills :o wage, high-skilled ,owth path where c::ildrzn were In sczol and jusrii:y wages that would allow for a family 'life. C-ompers said, The Federation endorses the total abolition of child iabcr under 14 years of age; an eight hour law for ail labours and mechanics employed by :.he government dimctl through ccna--Ctcrs workmen ,.::gated oa pubic w•^rk, and its rigid cnforccment, pr_tectien of and b _f rsron of s�ii-age as well s _ nor, for Wpio�ed .. :ac:cri:s. s ups and .. .. s-. ... } ea's::i sz• to .votrea....;he m_:sures, not pates. r That in all cases such daily waves shall remain at the minimum rate which was in such cases paid and received prior to the passage of ,he acts T"ae eight-hour bill was one of four Iabor-related bills pending in the legislature. The weekly pay bill, the child-labor bill, and the bill to make the first Monday in September a holiday, which would become known as Labor Day. In addition, that year the Kansas State Federation of Labor approved a resolution calling "for the abolition of convict labor when in competition with free, labor."6 The eight-hour bill, Senate Bill 151, failed in the Kansas senate March 6, 1891, with the Prevailing wage secticn removed. But by March 10, when the prevailing wage section was put a law stated, back in, the bill became law. Thts ._rat prevailing wage g :hat not less than the current rate of per diem wages in the locality where the work is penormed shall be paid to laborers, workmen, mechanics and other persons so employed by or on behalf of the state of Kansas....' At first, however-, the law was not enforced.' Not until 1900, did the Kansas Bureau of Labor and Industry Statistics report enforcement: "there were hundreds of complaints that were attended to by come., ondence, and good results obtained.r9. � Prevailing wage laws in other states. New York was the second state to pass a prevailing wage law. New York's eight-hour law (Chapter 385) was amended in 1894 by Chapter 622 to include a prevailing wage law for those employed on public works. As in Kansas, however, there were many violations."Laws similar to those in Kansas and New ""ork were passed in Oklahoma 1911 Arizona 1912), New Jersey (1913), Massachusetts (1914), and Nebraska (1909), Idaho ( (192-3) (see table 1.1). These laws established a precedent for the creation of the Federal Davis- Bacon prevailing wage law. Passage of The Davis-Bacon Act Three federal laws primarily affect prevailing wages in the United States: the Davis-Bacon Act of 1931 which applies to construction, the Walsh-Healey Public Contracts .act of I936 which covers employers in manufacturing and supply industries, and the Service Contract Act of 1965 (known as the O'Hara-McNamara Service Act), covering suppliers of personal and business services. These laws attempt to neutralize the effeas of government purchases on wage determination in the private sector. The Davis-Bacon Act is the most significant of the three laves. A prevailing wage is intended to prevent the federal government from ar�ectizg local wages and construction conditions; Davis-Bacon disallows the government from pushing down wages in competitive bidding. The government has always been a major purchaser of construction services. As a primary customer of construction services, the government holds : e potential to use its bargaining power to force down wage rates. b 3 Ca%s-Bacot Zereal c:fe= The ! 1�ernment is engage in =uiiding in my district a Veteran's Burcau hospitsl. Bids wear aske' Tar. Se. rai ` Y .on,- ac:or ;,id, had in their bids. ci course, they had to take nto consideration •he high labor standards Pr- in the State of New York...Tae bid, heave•Ter, was ict ro a firm 1;Mm .�iabama who had br.ught somWeore hz tied onto his job,laborers they were Alab=a into Long Island, N.Y.; Into my �c�:.-ict. They housed in shacks, they were paid a rery lc%v wage, and the work proceeded...It saemed :o me that tiie federal vovernment should nat engage in construction work in any state and unde:anine r conditions and :Ic Tabor wages :aid in that State...Tne least the 2ede :.. ral Govement rile iaoo evaiiing in the Iecalir r whe:9 cca cra do is comply wi 11 :e S i stlnda."ds of wages and labor :,- .'•e building onstrucrion :s :c ::xe place• a.- :ai pre• a.:. 7 wa?e law oegan in l92, and cor.::nued :n 1923 and 193), but e ��S :Zr 3 . '93', Bacon's ori�inai proposal, wh ch he :had :ei::xeducsd as .10 ��:. .vas passed. vn liar .. y_ . o 12 r'.R i o i 9, was si?? ed into '.awy P:eSiaent mover. is t Davis-Bacon Act required payment of prevailing wages on fecerslly financed consm;ction projecs. The law essec^:ally ruled out bidding on construction worker wages on federa '.y financed construction. The original lan.;sage was vague, however, and prevailing wages ,v. v I { president Roosevelt sig^ed genera::v were not dete.^aine.. �e or. ;he acceptance of bids. In 93., ls-Bacon Act- Tae - clariivin 3 amend,:.ents to tha act, wiich became the basis or h e cur-,ant Dati of 1935 gave the Secretary of Leber authority to set he prevailing vat:0rai Labor Relators wag=. 1935, RoCSevelt'S Sec:e!=y of Labor, Francis Perkins, established .he o:: - roes for v was said to be he wag= ?ai3 deter=x:,l ng the �3vis-Bacon pr!vain grates. Ttze prevaiiin, x��_ :o &.a majority, I. a m3JorltV ex seed; if not, the 30-percent rule was used. The 30-percent rule :;1er^s 'If 30 per_ent of h e wor'.<e:s in an area are paid the same rate, that ra:a becomes he *-' err-ant rule often resulted in the union wage bei he prevailing prevuI,ng rate -.ere. 1 Ie ��-? - !1'. 30-per cent rule did 1oi 3ooly, because at least 30 �erCe:lt et ,^•e `:'v^rX�rS rn a �!�ate N3c u1d r v r wage ra the average wage rate in .':e :c 31 labor t1:rket did not receive he Same was oa d to work_rs Being :he _ame joo. 1,le prevailing wage was dete^nined his way for 50 years. Li l 85, Pr?side t Reag�^. c:1w^.ged mini u.ation of Davis-Bacon, creating '�`Ie 50-vercent rule. 7-1-.e :e�rised rt;ulation reduces he influence of the negotiated union wage in in areas (see page 9 below). T.:e T-n it .a:r-.rd.,.er: :o .':e Cens-�=,tion re .ncs he ability of he federal ;overnment:o d,�ate .ercact :e-Tls Ter ..e sstec. T.1,1us, work funded errreiy by state or local 3overnmers Is t or ^ -StabllSh Its own p[eValliIlg 'wage t - :td by :V:�- .3CC;.. _�C ,ate, coL`r: city Catl In 1994 :9 ozrcent OF 311 -a=—r-i-/el :e�eral it _-looses :o do so — :.:ou3^ '-.3':siaron. aicen .`:era union ccr.::acts, -t8 percent �ised aye:age Lavis-;;actin orevaiiing wa rates wag' as woes, and he remaining :5 percent of counties used a mix of union and ay..rage depend:ag on t .e ecc::paticn. 5 ✓8'rIS-03 �.. Z:'GCB: -IS x The t":rst bill to repeal the statute was introduced in 1979, only to be vetoed by Democratic Governor Scott Mathheson. In 1991, repeal bills were introduced in 14 states. Only in Utah did repeal succeed that year and it succeeded only ai'er a second veto from Matheson. repeal The bill was approved on almost straight parry lines — Republicans favoring repeal and Democrats opposed. The Salt Lake City Tribune noted that only one Republican representative, who called himself a lifelong Republican and union member,voted against repeal and broke away from part' lines.:0 • "I'm convinced that 'When Matheson vetoed the bill in 1981, he said, at repeal of this law :s not in the best interests of working people in the trades whose skills are essential for a vigorous construction indus-y."Z` Nonetheless, the Senate overrode the veto 21-7 and the repeal took effect Z moms later. Tnose in favor of the repeal maintained that the prevailing wage law was inflationary and pro--union. Republican C. McClain (Mac) Haddow sponsored the 1981 repeal bill. He said, "tze law is outmoded and is preserved only as a tool to extend union control. The law is contrary to u Utah's right-to-work philosophy....» predicted a I0 to I S Roger Evershed, president of the Association of Building Contractors, p percent savings on public works projects with repeal." Ar.'cona. Tnexthe next state to repeal was Arizona in 1984.24 Arizona's statute began as an eight- hour work day in 1912 and, by 1930, became a prevailing wage law. In a court test, the statute was found uncons::tutional in September 1979.3 L-i Nove:ncer 1984, voters repealed the statute in a ballot initiative, Proposition 300. Provisions of the ballot initiative prevented communities from itnciementing local prevailing wage statutes. Ida.�o. Idaho's prevailing wage law was first enacted in 1911 as an eight-hour law. 7ae v aru*e statute was extensively amended until 1965; efforts to repeal it began in 1979. The Iegrsl failed To overode several vetoes but did repeal the law in 1985." At the same time, over:Ime )ay • e r rzor� than eight hours of work were repealed.-s req.ur_r. nts for 85.=� Atte .^.pts for repeal Colorado. Colorado also repealed its prevailing wage law zn I9 it was not until after the governor had vetoed the bill several ti=—as began in he tare 197,01s, but v idden and the repeal passed. Nevertheless, since 1985 at least one em P that tie veto...was o ,o municipally, Pueblo, established its own prevailing wage rate for local constnicuon. Mew Hampshire. New Hampshire joined Colorado and Idaho in 1985 when it, too, repeait-d." Altrough legislators began in 1979 to :ry to repeal the prevailing wage law, they did not s;..coed until 1985. Influenced by reports of inzlated costs on a school construction job, both houses passed repeal without the signature of Governor John Sununu.'' 3ansas and Louisiana. Kansas, the first to have a state prevailing wage law, repealed it in .. 1987." Louisiana followed in 1988 with repeal over the initial veto of the governor.' r• k Efforts to Repeal Other Prevailing Wage Laws 'J The.'Kassachusetts ballot initiative In Massachusetts,in 1988, thousands of union members, already 3c~:e i^ the presidential alecuon, worked wit:`i corimuniry g=pups to help defeat a ba'.:ot inira:i�:e :^at would 1:ave re.z!ed tt'.e s:ate's 1914 prevailing wage 13w. The effort to block repeal :^ , ass-,czusa-s appears also to have sic..ed efforts to repeal other star- prevailing wag= s - Dnvts-taco.^. �c^ea! �;TCCS r • Advocat!s of repeal of Davis-Bacon said: I -iIa act has Inflated construction costs. 2. The act costs the federal government huge amounts of money. 3. The act is poorly administered. 4. Tne act is biased toward union contractors and hurts non-union contractccs. 5. The act has caused wage infl ay.on. o. The act discriminates against rif•.orties, because they are disproportionately represented among :he low-skilled labor ford. 7. %:� ee-tra:xet system is supressed. zhanatd A:_.,ou:Tn the Da��is-Bacon .act was not repealed n I97,th IF, ad+'19 is"dORd the'30 - ' .,:.^.is:raaon in I,85 � feted a _aw ears later. 1ne ad .,, • � finis Y the w�i .n.. yaw is ac�m percent :-uie. Until then, the Department of Labor used the modal — most common — wage to determine the prevailing wage for an occupation in a Iocal labor market, if the modal wage to v accounted for more than 30 percent of all wages for that group." If :he Nodal wa=e the ?e. a � averse wage was declared he f percent of all wages, the me.n ( g ) g r an 3 0 .or fewer than accounted prevailing wage. T^.e ;Zeagan administration raised the threshold to 50 percent before the mode could be -end to be he modal wage and hev :e^c to be above declared .'ae prevailing wage. Union wages t ,, 1 ;ne e�ie::t .-., the Re an ,.a.-n inisu'arve chat Lion. So � the mean_ ar average wage for an occupy of Lowe^ng the prevailing wage in arzas where unions were weak. Some of the compering claims :or and against Davis-Bacon can be tested ag=ins ^e zYperience reoealed state prevailing wage laws, as well as those :hat continue of ihe ;ta,_s — tc`tcse th at `tave to ^ave such laws, and sates hat !nave never legislated a prevailing wage. This s�:dy examn_s the conte:.tions of Davis-Bacon proponen's that prevailing wage laws prevent the disruption of local wage and cors:.ucon labor markets and that prevailing wage laws protect living standards _ and discourage curt--:roat competition. T'nis study examines, as well, the contention of Davis- -^^ent considerable sums of money and '.incriminates Bacon oeoonents ;.hat he law costs joy'-...:.. workers. itte story also raises :-Ne new questions. rirst, against women and minority ccnsr.-,:c=or. �0 32aticn :Il what ate -"he of prevailing wage lays on za:ning and human captt:i Second, wrist execs do these laws have on -he safety and health of corsan:crion work--rs' 3 .ZC Cai :;TCCS II. The Economic Effects of Davis-Bacon Repeals Cutthroat Bidding As soon as the law was repealed, some of these non-union people [contractors] that had been doing small work around town suddenly just took o$; and the union people [contractors] like ourselves, our market share decreased. — President, a union construction company, Salt Lake City, 1993 Our -ompany has consisted of my father and my grandfather and me from about 1963. [We (Our] , . are a double-breasted company.] Company A :s a union [general] contractor that at.=s merit shop companies with no regard to union affihation. Company B is a non-union merit shop company.... Our indusay became very competitive during the mid-eighties, a lot of people are chasing the same type of work. — General contractor, double-breasted company, Salt Lake City, 1993 We'-re been in business for 51 years. Before that my gnat-grandfather ran a construction company and so we've always done construction. Right now we're doing mostly mechanical, and we do utilities, Mountain Fuel, water lines, sewer lines, AT&T jobs. We've built homes. We've built golf courses.We've built apartment buildings. in the last probably about eight years [since :he mid-1930s] there's a lot more small companies —li=le tiny, you Know, dad and his th=e boys. We can't compete against them. We have too much overhead to do that and you get small start-up companies,they`re willing to work for nothing for a while and you know they;l go ouc there for two years and just take these jobs dirt cheap. Sometimes they cant finish. They'll 3o broke in the middle but still, we don't want to work for nothing. We'd just rather to ck the gate and wait. — Office manager, union construction company, Salt Lake City, 1993 When Utah repealed its prevailing wage law in 1981, the structure of the consrucaon industryy us effect was the decline of union membership and union changed cramaticaily. T"ne most obvio contractors. But this was only the most obvious effect. Underlying the decline of union contractors was the rise of the little contractor and increasing turnover of contracting firms in the business. The industrial organization of the industry changed, with an increased reliance on i' subcontr=ors. Comparing the 12 years prior to repeal to 10 years after repeal, the share of total cons-r;uc::on employment accounted for by the typically bigger and more capital-intensive general conactcrs and heavy and highway contractors fell, while the share of total employment tr accounted for by specialty subcontractors rose (fig. 2.1). With the entry into the market of more contractors and smaller contractors, competitive i pressure to win bids heated up. This pushed wages down. An operating engineer familiar with the bidding wars stimulated by Utah's prevailing wage law repeal tells how the bidding affscted labor. tt Da%rs-Bacon Rcpcai �t�ccu �N�iea tnq; re^doled litah's ta•v, a ;..t of companies went out of business because of he cutthrco: ccmpe:aion. .� tot of companies just bought jobs so t:ev oou;d have a Casa flaw to make paym :s on their equipment. The design engineers r would 'stille-0 n Lr ere anyway and the job was Loin; to .:ost a million dollars. The coarrac.o ;ow-ba.11 the bid. Then they would turn around to heir workers and make their wares fit whatever :he;; had to be to fit the low-ball bid. �� the prevailing wage law was ,pealed. f bid sh000ing sr:.,r P T`ac Len ral contractors did a lot o . . The general contractor would get a bid from :.e subconcauor of say S_`0,000 and then he would low-bail the bid. Then, when the 3e^.cral got lire job he weal' go back to the you've got to cut your bid :o S•:0.00O to have subctintrac:or and say .•eah l•:e got the job but ou 3 ;this ;cb :'vc got and he sub wou;d go back to he :vcrxe s and sav JK we've Sot Wis job but ^.ow .. e got :e cut .your wag-s. _ and the Sce -osts o: ..ate^ais and .uppi:cs and equipment were j aLse you of j".:s nob. So the wori:c:s ;lad ;o msl:c tip the asphalt didn't go dcw.• ust :cc 3 - , mcn., off the backs •� A e. of .Heir 'y .,molo d.L:-Kce for all this low-ball bidding. So basicai,y .he y g wages had to fail. of the worker. Whether it was to make money or just to break even,1 engineer, Bountiful, Utah, 1994 — Operating e• g But wages were net the only factor to feel the strain of an dOvor'L c si o no process. Govem ^en: purchasers of construction services we.e now exposed annual cost over-L.^.s for t:.e Utah Depar�nent of bidsover-turi:.ing costs. Average 2 ^� Since -`ie iaLL�s repeal was percent of initial accepted bid (fig. ._ . nce Transvortaticn prior :o • repeal, however, overrun costs ;•lave risen to 7.3 percent of the initial bid. This rise in ng the r costs has come despite the introduction of computers s a tool for contractors in preparing bids. %1e cause of hose increased over-un cons is the pour-repeal tendency for cortrac:ors to '.axe (^a Wien more ;IS&S n the bidding process uncer the pressure Jf Illcreas�d cOr^pal Q�IT 3ro OL �+e prCjeC:'S the save :ills for bids Qr. a project, he state engineer a1repe3v 'iaw'�winning bids averages 91 cost. in are decade prior to the -epeal of Ltah's prevalitng e s es:.rnate. Afte. :h, repeal, winning bids have been, on averag percent of :he state engineer , S9 • percent ��r-the-state enginee r's -situate. Contractors are shaving their bids to �zn state contracs. These rower estimates have not proved to be a windfall for he •ate• t �s-ead, of er L;tah's prevaiiing wage law repeal, final constriction cols have been rttu,-:ing at 95 ^er=e:.t of he sa:e -n;freers iriral estimate. T•ri air:ourts to 5 percentage potnts above r� V ^t of he eng:needs ?:-or :c 7Utan's -epesl, :t^.ai cos—s we e ; ing 93 aerce . esrmate, only RL'0 t^,Ct.^.s zi&,er t~an mina.i accepted bid prices. r d - y ... an at the pre-repe--1 crIsricuon was ultima:eiy cheaper for -- ,...s eS 1c: jecessa. ?--earl viC ric_- and ac:1.131 'QStS w� . .p 'ye^gee.^. acceote� ^ ? the state, out it does •;fern that re reia.lons:.i more ce :gin and that ccatrac:ors promised less before Utah's repeal, but 3eiivere^ more relative tt to the =:-=z enain a ws Cost esimates. V_ not omy c:_ated uncertai ;r in t: Heightened compzction after �tah's repeal has piurrber process, but has also lowered Utah cons-xuction wages across the board. A u.*iion describes ,his: • Ls•.�s-�s:or. Zc. ... �..c�:s r c - r� f . `I 1 C V/o _-.-..-_-•-.•_---_.-_.•---------------------««__...__•_•__•__.. 93% 98% o % --------------- ----- > 94% 0 ----- 92% 950/0 C- COO/ O m ¢` &r% n a 70--81 Attu S2-?4 &V(Eng's Est) ActuaV(Eng's Est) i Figure 2.3 :he ratio of accepted bids and final cost to the Utah state engineer's estimate of -oad construction project cost, before and after repeal of the state's prevailing wage law ;after the Utah repeal of its prevailing wage law, competition among contractors heated up and contractors shaved their bids to win contracts. In the decade before the state repeal, accepted bids averaged 89 percent of the state engineer's estimated project cost on road construction. After repeal, accepted bids fell, on I� average, to 89 percent of the state engineer's estimates. However, this cutthroat bidding did not cut final project costs as a percentage of the state engineer's �.. estimates. In the decade after the repeal, because of a tripling of cost overruns, =. the final project costs averaged 95 percent of the state engineer's estimate. n If one r sumes that rile union dif;erential is 20 percent above :"e non-union :.aze and, he repeal, union wade rails to the non-union wave, both wage :ores .�iIl ha:e to fall en fu:-!her to aft: zn overall 7.5 percent cut in earnings. :�ssumin� tl at tt.e union wave weuie :all to the ,ony.ier :ate arid then they would both tall together, the union wage would have :,) fall by 21 percent and he non-union rate would Nave to Lail by 5 percent to obtain an overall fall of ?.5 Pere_^t.41 In fact, only rareiv does the union rate fall entireiy to :.pie non-uzior. wave. reasonable assumption would be that the union prior -aa repeal was 20 percent above the non-unicn rate and after the repeal fell to 10 percent above the non-union rate. Given a 7.5 percent overall :ail in earnings and a 13 percent union membership :ate, union wages woulc have to `ail 1 oerce.^.t and non-union wages would have to fail 5.3 percent to obta:r. an overt:. fall 7.5 uz ether words, while he union rate would have to .ail :,4:c_ as :1:; as re .. -won of union ra•e ape ion-unicn sector or _onsuuczon workers would have to acsoro much of :he average ^.'rcen:3,Pe wage cut. lie effects of s:3te repeals oI' pre'•a1=1r.g wave :a'.�s a.�e isciated neither to union workers nor to government-t5nanced cons M_ c^cn. They gena:3te across-iatle- board cuts in the earnings of all constriction workers. A Loss of State Tax Revenues The tax -eve.^.ue losses that result from lower consu.ic.ion wage levels are surprisingly large. Whatever ::e source of this eariint s decline among corstr.:c:ion wor'.cers, s•,:tes wZth income taxes have 'ost tax revenues as a result of this decline in taxacie income ar^.org -a workers. Raid, because this lost income means lost purchasing Dower, states Flat izave reYeaiad their prevailing, wage laws have also lost some sales tax revenues. On average, consauc=on workers account for 5 to 6 percent of a state's labor force. In Utah in 1991, lnd:':"duals S20 000 :o 530,%I paid a marginal income tax rate of about 7 percent. Taking the 3 i,5:8 construe::•en �,or'.<er5 employed in Utah in 1991 and anaverage cer :aD[t3 dec::ra :n :ncorr._ of Sl.S,`, '. e tc.ai loss of annual income `.om :he Utah construction industry in Utah in 1991 because Utah's 1931 repeal could be calcuiated as S58 million (51,835 times 31,528). Given a marglnai ax -ate cr 7 percent, 1991 lost ,—.ate income tax revenues might amour.: to S-► in.ilicn (in 1991 �oilars) (table 2.1). Assu...ing a .:iar,inai propensity to consume on sales- axable ;terns frcm ci=ges in income of 80 percent and a sales tax rate of 6.25 percent, lost state sales tax revenues tom 1`lis '.os5 of incotn_ amount :o S2.9 hillier: in 1991.j9 Adding these Zvo losses anti 995 a� n0 the ors•urner rice tnd�s y. s an ear.^at'd loss of Sb.2 -o .. v tleS LSI:.� c P - miii on7:^-�.a:e raves in j:t:h n 1.991 evaiva:ed in 1:�95 doi:a:5. •i' ^.ues -•ay be overestimate for :cur reasons, _ 1::., of 58.2 .::i1.1on in post .Y re•:_ xages fall a.^d :acor Oecor. es -eacer cen:rac:ors might are :ore orKt. however =::s., if - , So we TS :onsider possible incre: zs in tct i ircorre of �crs• -c:on war..- :_suiting om Possible nc:eases :: rota'. construction em�Iovm--t a ter a fall in woes. Second, real wages nave bee: fail:- yin rite U tired States generally, incl•.ding tine cons�c:ior. :r:dust.:y. Sot: a of ti•.e lower wages a:\\ar sm:_ repeals may simply reflect a tong-tzrrn decline in real wages hat woI save taken place anyway. i1 ird, annual earrings in cons�uction are sensinve M u:.e^pioy:::ea. a: d :atl w ure 'i Ploy" increases. Because E=ings nse when unempioyment falls _ 4,317 ---------------------------------------- ............. ----------------- 0 ------- - -------- $2 .- --------•-- ------ 75 � v $1 > ' Eercre Re .J After Repeal States by Groups Figur-e 2 A comparison of construction earnings in nine repeal states only,before and after repeals (in 1991 dollars) In the nine states that repealed their prevailing wage laws between 1979 and 1988, average annual income fell after the repeals (calculated in constant 1991 dollars). This fact does not control for other factors that might have been driving down wages, but it is prima facie evidence that the repeals forced lower earnings not just on pubiic works but across the construction labor market. 0 i` .7 unerzployrrent varies by state and ;ear, some of the difference in earnings mid tt because of 2.4 and 2 5 -_ Last, construction waes vary b" vanaticns in the :ne:nplcy :rent rare (see Ergs. ) region for reasons that are not directly due to the presence or abse.^.ce of preva�irng wage laNs. These regional differences in eamings, unemployment, and long-term trends in wages can ce accounted for by using linear regression analysis. Regression Analysis of the DecA a of Construction Worker Earnings Using linear regression analysis, this sec', On tales U.S. Depa-*-anent of Lubec e:nniovment a:.d e la,�_�l to construcnon : re-?srmate ,�e earrings data for construction wcrkers in states nor earr+ingS OSS resulting _ram Sty:e �'���' o*oreva:ar. wale lays :a= anallysis centreis for le-g- '� ;ne ,ploy-ne!t and va.^.a::or. in wages r�;ion of ,} .erm *rends in wages, vartarons is 2 had a crevaii.ng wage law, coun^r, and then focuses on t.^e !f=Pct of (1) never hav;ng repealing a prevailing wage law, and (3) raising ;he threshold for imple renting a sate prevail:.r.g wage law to contracts worth S500,000 or more. -earnings data rovice detailed information on R and .earning P n U.S. Department of Labor -employment contrac:er.50 annual construction earnings brcktri down by year, state, and type or construc^on For 197 5-91, there are 27,7%3 separate observations. The inclusion in these data or information about prevailing-wage law arus by state and year and translation of—,;I money values into 1�9I allows , to test for (1` e e .ect that e .. `�aving - d r h consumer price index) dollars (usi.. u.e �. .) f a f-�-- .I' e._.,... on indiviL�ai a prevailing wage law has on per capita• consruc:on earnings_^(- n individual earnings of earnings of repealing a state prevailing wage law, and (3) :she _.:e W e law. raising the threshold for apply'lrg a prevailing age In this .est, we conuol for regional diferences in construction earnings, secular trends in S� r' in eariin�s as 3 result of variatic^s in une npioyment, =^.d earnings, cyc.ical varlanons , differences in earnings 5y detailed contractor type.:" c-on T::e data used for this test ir..lude average earnings access all states, yews, and constru trades — 525,6�5 per year in 1991 dollars (table 2. ).53 States that never had a prevailing w=ge ercent of all the observations. States that repealed their laws account :cr law account for 15.6 p .; 10.5 percent of all observations after repeal and 7.3 percent of all observa..ors before -y repealed their laws, for a combined total of 13.3 percent. States's that had and retained m edof ail 9 5 and 1991 account for the :„rnainin, percent thtir prevai:ing wage laws betwe..n gage laws gut �; .�,v d Oklahoma. the ;rates with Yrevaili.^_g obse. adons in area set. lvfa.- land any with .hresnol� levels of pro;ecs costir•g S500.;)CO or more, account for 1 percent of aII yn. p od a•veraged 5.76 percent annual'_y. �-� �.ute:^ luvme^t rates in is ere obser.va*o is. Jtale-bv-�. ,� P c-;::gin -es :its of .tis r,�. scion -nod.I st:,;,ating the erfects of „are repe�°s on c.,nstru »ro s e earnings are statisticaily si�lificant and the o•serail model has a ;pertness of :it of 73 percent, which means that 73 percent of L e overall v- .piton in annual earnings in the data set are explained by the model. Tne results may be read as follows (see tabie 2.3). f annual earnings of S33,005. (This is a star*ang point Begin With a constant amount o calculated by the regression model and is typically cailad the "constant.") Then se!_ct a state and ' .e ' f uctlon. VV'-_• /'are in =•:v Sate .0r ! v ar we lkrlaw :he -a.u's of prevailing wage CC CCnSt:"' , y a (3) U a eva'lin2 via^__3 raw. °JCL, �?�' 199., that use tear.^. S 2r :`C' p:� '1 CoL'T^ tall 1�a y _ L e e, a_. L. 3 S law has repealed.repealed. F'urrh!rmo,-!, !go I wzs i'I ye:."s 3:6er V9�`5"'34C^:: `:'?CSl �Ci•CCIS Table 23 A regression model estimate of the effects of state repeal on constru l trends in earningsction annual ea zas , =�^ controlling for regional differences in earnings and for secular and cyclica Regression Model Examples for 1991 Variables and Coefficients (in 1991 Dollars) Utah Maryland Geareia -71 Starting Point: S33.005 I S33.005 S33.005 S33.005 Regional Control Variables: Alaska S 15.623 Hawaii S7.982 Midwest S4,7 63 Pacific S4.633 :atlantic S4,617 S+.5 New England SI.SyS Corn Belt S1,0I0 Mountain -S79 -S79 South -523 60 T.-and Control Variables: - -5225 -53.89 -».5�9 Secular Trend Unemployment -530-3I -S1?g1 -S1.?84 -51.512 ; Focus on Legal Variables: -�960 Never Had Law S'"960 Repel S1,350 -S1.350 -y Threshold S500.000 Predicted Income: S26266 S30.S36 5='-3"5 .i -:Ile focus =-^�ble in the Tede! is average annual earnings in construction and :.he zypothesis .s that the re:a::::.shiF ber;�een �a.-nings and employment should be negative. As gar Z: ;so dowT. emptoytre:: .:.igct well �o up. iao "ne regression model also includes control {dumt � 5a s � each state and each detaiied industY classiriication (four-digit SIC; such as, p pip'! fitters, SIC :71'.). Thus, the man d ode! predicts construe on erployment in specific f is states, !975 each construct, subclassification, such as ?iambi^g and pipe fitting• the average employment in a four-digit subclassification is 3o emDslo meet workers.ction d there is unemploy-m!::t 'ate, not su'rpr.SiR�tV, negatively at.�c'.s construction small but statistically sigt:irc:�•t u;,ward lend in employ-rent. The etxect ot,prevaii:n; wage :ace repeals eRzpevrtezt is aeg3... ., gut this variable is not statistically si�Tn:Lieart wiZ:eh me:::s ^eats a :mpioyment is zero. the `r•:z ::�;L _. eCt ^,I :Z. �� wet earnin'�s �' o :0 ,e �ect of sta:z repeals on empiovr:lent wo .king ou_ c:o ; .:, :. tare::e • at - I �s ---r::ngs fa' as theoretically expected. is no: ze� �":e =_!ct of earnings oR .mpioymeat is -eialions :ia employment increases and this -estimated effect is stanstically sib i.,can.. From this . we can esimate the indirect etsect of state prevailing wage laws on employment through the repeals' et:ects on earnings. r various levels of earnings decline. In Possible employment erects may be calculated ,or , --suits frcm table �iusrn (!) pres-eants nvpo'.^ercal earnings Mines and, in column (-), the r table 2.4 are used to calculate a predicted increase :n the cons�uc:ion indu-,.., when tt s plumbers and pine firers, SIC 711). anai,/zed .: the detail of s-digit SIC codes (such as . �� „� average a::.ual ccr..�uc:ion earnings :all from a less of S500 to a loss of 5�""O' !rn:• loymen given SIC Indust •y groups rises from 24 new workers :0 1 i8 new workers.`6 Gil;= in average in $ 40 these hypothetical * creases in employment size of a 4-digit-SIC industry group of 3,5 , employment translated in percentage terms to an increase of from 0.7 percent 'Nit=^ -=.'^ings ::II II � -- � 4.0 percent when earnings in con .: cnon :ail by S3,C00. by S500 :o an er.:pioyme t inc, .3 e of The ti Effect of Repels -n Government Budgets I on state e:cpdit clicn The over it er ect of state repeals of prevailing wage S u hare? al and state tax revenues "Will depend on the str:ounts of government cost g and los. :ax revenues :rom a repeal. Government con'tu�o Tust savingss worker produc..end on hvity erl no ques'aen�: now much lower are wage cosy a�.• a •- t o �ernment urc ase? Lost tax a„o w m::c con,-truc::on work does he ao p at low .:��_s, art.. how wor:ters ea:.:ng revenues will depend en (1) iZe -1.3:nal income ::x rate for cor.�:rlc�oR rate, !--) ate marginal properns:ty ro consume S20,OG,' :o S-J NCO ?e: year. (21 -.he sales tax 00 t4) lost n 1 �e� y-ea. . ?er- taxable corn: edi ies :cr co„�:^.:c:ien workers e,-I.n; S-0,�CO to S 0, Zen -etnployme:.t capita co-suuc�on income associated with a repeal, and (5) gained :or.�"_c. va -.��,tis 3 repeal. ine S_O,G00 to 5. O,OGO rom^.;e --compasses npasses ^Zest con •1 -.-repeal on workers.) -^.e associate,, P:2 OI:S es mates of construction cast sarngs asscc.ated W1 a Zypothetl t;ifice fa• cps r federal Datiis-Bacon act range from I to Il percent.' The Congressional Budg an es:ir.:ate of a 1.5 percent cost savings associated w'i'h the wage effect plus a �.3 percent cost savin s bz:.cuss of page-.tiork associated wi ;� Da•ris-Bacon.$a 1 Via•:lRgS ...ay be higher or g lo•Ner. oyment or an ge-s Table 2-5 Effects of construction earnings decline t SIC) of 3,S�10fworkers Yper ee detailed j -1 construction standard industrial Classification (4-d ij Predicted Rise in Percentage Rise Various In I-^lployment . Hypothetical Employment Because of A Fail in Annual Because of a F211 Earnings Const in Earrings f Declines rnctian Earnings (2) (3) (1) -SSCfl 24 0.7�'d •� -S 1.000 47 13% 41,500 71 2.0`b 44 2.7 90 S2.000 -52.500 118 3.3% _ 141 -S3.000 4.0co As repeals force a fall in construction wages and earnings, construction table 2.4 indicates that a SE00 fall in earnings employment rises. The model in results in a 0.7 percent rise in employment. An average annual S ,000 drop in earnings would result in a 4 percent rise in employment. This is 3n "inelastic" !- demand for labor — the percentage that earnings declines is subs=ntiall! higher than the resulting percentage rise in employment (for the group). This means that even though employment rises when wages fall,the rise in employment ;s relatively ges. Consequently, small compared to the fall in wa overall income to construction workers declines after state repeals. r- �i i.. rr- Table 2.6 The relation '3f hypothctical construction-cost savings to tax revenues is y.W 198' 1990 �363 :7336 315_3 34902 Isana (339.397.04+) (536,891.'46) (538,203.'3I) (S+I,Ila:!5) (S46 56�3=11 1(SI,5+5,'93) 1SS.553 at) 3 La" +e .63 498 564 625 'll a Q+ead a78 u: plw+*al .i+O S nea�n. 5:5..06 114.O".�08 SS+.95_3:7 5:a379.IS1 xae. .._: 1'3.:+7,�d6) (SZ:.360•: i;Jb.:93.-08) t;•:_'..J'70) : ±Jet:.aat:•+m*� + '_5.3;.591) '^`:_.`•50") ..25,E _:S�) 1.NS.318) �55._9) .5tS+1) J59. 30) joa:noi:. 7" .S1.i31.35:7 'Sl.'Sd.-ya .Si.:33.+b5) � ! i (S1.;09,513) 9 rx --7"" (St.:++.!30) iS1�36._'.1) (n•t 3.'�'t'-: + IS3 "3.7=.+ 1 S7.Y3J.'301 (Ss;9l.'!61 ' 10 T.oW�ac'aav (1:.�5.J3±� fSJ.Ot��adl I - - I1 Vvua�i Srar.Tr,c�oed Catrel+R>a+ _ I g ip 194.436.6_0 S 8 Ca9.iu"3 ;93.:.5.306 573,661,056 581,518,353 SICa��018 Sl l3.'90�78 t3 Ronda 5:1,117.077 S9,S2+.:'S SI•,1231 S1IX,0.161 =7,577,:80 SI7S�+. 42 .06 ?5.G?S Si 53 S13_51 s,'_0 14 Taal S1 i 5.353.591 91 15 clwo�txuc..�rcia a�C..rscua+Q+C.aRa . I 906-312 ;6 l% SL:S5337 S7r,9.1:5 Si. .:x9 1 51.:51.;•0 "=5,0� SI�__..51 i 52,713.9'36 S3,455.33; 53.�•9.565 S3.T.�54 t7 )% 53.+5o..it1 ;y537,�133: t a s% u.-77.6as + +aa 15.759,:= Sti,:J7,:0S dzc.156 u�395.5.99 S4.531.�h1 � as sa.o6a:= SS�a6.3s1 s9::'3.osa t9 1 '% ;a.c�S.'S9 16.:3J Yaa ::6.7.a1 16.3+4.: ! 520�67,i+J S21.-33.594 11 19.981. 919 18,:36.30920 11 5.:61 r: z1o�99,s37 r.91�m S9 959.+3•+ Si_i71,c.0 S:'•r SS'• ::,�5' �bl .I Vet Gac is:..A '+'u'kc� e ` <, �3-294) 9,410) (=731-Z90) 5:84,134 +(F06736) (3427'69) � S1:'a1,391 SIS5:.176 25 s% szssl.ss. 51�31�2 .^- �61 S:.f05.3J9 21,t5S5t Z6 SU29.w7 u.o.0339 +-9 l a,+6� u,l sa+72 u.195.I3+ s-c.:t o.t-a -% u.46�'2s .� SnZ6:L1I $4.S9r,-2 S6 33E,716 Sj,'31. 7 S6,+i4?93 So.54aS3 :1 c .S '•K 59.+8+S"S /6«i..:+A :9.wi.49.4 S6._`a7.-..• �:1!::i S9:7Lfz! 29_11'1'7 ra 1w rs With an employment level of 6 million construction workers and an average aanual earning of S=7,000, the lost income from lower wages exceeds the gained income from increased employment. This results in differing values of lost income tax revenues depending on the assumed marginal tax rate. With a value for federal construction of $11.5 billion, the hypotheticalsavings in s on construction from a repeal depends on the assumed cost-savings rate- At a marginal income tax rate of 16 percent, net budgetary savings from a repeal occur only with construction cost savings rates above S percent- At a 20 percent marginal tax rate, net budgetary savings from a repeal occur only with construction cost savings rates aboN-e 9 percent. At a 28 percent marginal tax rate, net budgetary savings from a repeal never occur within the range of cost savings between 1 and 11 percent. To short, a repeal of the Davis-Bacon Act will hurt the federal budget deficit. There are approximately 6 million construction workers in the United States."'Table 2.7, rcw 2 shows what would have been trie less in income that these construction workers would aave experienced given the 1994 value (-.S1,477) of our regression estimate of the effect of s`:e repeals en construction income. Row 3 presents an estimate of increased national cons�-uaiol e.:.pioyr..ent associated with lower wages.Row 4 presents average annual intone for constru workers in 1994. Row 5 multiplies gained employment in row 3 times average income in rcw 4 to obtain the increase in total cons::uc-n workers' income associated with a hypoth: :t repeal of:he Davis-�accn Ac:. Row 5 subtrac-s gained workers' income :mom new empioy:;,,fnt from lost interne as a result of lower wages :o yield net lost worker inrame resulting iun a hypot^etical repeal. Rows 8 through 10 present lost income tax revenues due to net lost income at three marginal tax rates of 16, 20 and 28 percent. In fiscal year 1990-91, the government spent 510.491 billion on const:-sc:ion.,C Row 11 presents t.'^.is SUM in 1994 doiT��. Rows i3 i,irou_i 18 present levels of hypot:::-::cal savinggs ?n con a c.10n OOS'�s associated =:-? a repeal of Davis-Bacon. Recall u-►at the Congressional Budget V%Ice es-.:1na:es :c:al the sa"cgs :o be 1.7 percent, but otters i:ave presented savings estimates between 0.5 percent :I percent. Rows 20 through 'IS present the net effete on the federal budget cf ri hypotn.dr:.l a ma- 1 construction cost savings at various projected :aces minus tax revenue Iosses at v ous tax rates. Rows 20 through.25 ":tow that only at very low marginal tax rates and very ii } constr-.:c on cost savings rates does the federal budget bene t from a rep al of Davis-Bacot at a narg:zai tax :ate of 20 perc�zt and a con .. ctien cost savings rate of 3 percent, the fe=er_l annually in 1994 dollars based on the 1991 Ievel of federal govet�L=t budget loses 5838 million a '=f erendi:.ires on construction. Summary In Utah, the repeal of the state prevailing wage law led to an overheated •oidding process !;-iiich :.` added ::.cer;aiay to Line cost of state construction. Ln the decade before .he repeal, cost ov_ i^ e�age- _ pen:: cf accepted bits. --I line decade aft:. on state- nanced roa,. cons:.:. . „n av .. .-_.. . -+ *-u.-a rse :0 7 peace... of&.t act-ctea did. : c:of, repel avtra^ roa,, cons:-::..:to^ cost , yell tnsucc:ion of :he data slowed %hat, aster repeal, corn-actors tended to prtsen: bids at a :o;4 ,5 1 i TII. Tae Effect of State Repeals of Prevailing Wage Laws on Training and Minority Participaron in Training Tais chapter presents a case study of the effects or the repeal in 1991 of Utah's prevailing wage law on unionizal.0n, construction earnings, and training. The Utah repeal accelerated the decline in the union share of the state's construction labor market, drove down average cor_str:c:ion wages in the stars, and decreased union apprenr:ceship training for construction. No public or private source has ofaet he decline i•"• training• In esPense to t,'te decline in union me.:.bership and :raining, contractors have reduced turnover in order to retain skilled workers and to minimizeuaso screening c.-<. rairin costs. In _esporlsa not r.:y to :he dec''ine in consm—,c:.on .tie �s but also to the coincident decline in hea :h and pension -;enef::ts, ^owever, expert:erced construcaon a,��.��s ir. �.er indus=t:s. Taus, wniie corstrsc :0n f,_n r e ear,::g 11e:r trades for V -2c worker's ter. 61 turnover :s on the decline, turnover, in he industry is on.the - e. 1"his chanter examines also whether the ;. taii experence i Sn training can be generalized to their prevailing the eig:�.t other states that have repealed th p g wage laws in construction. The U.S. Department of Labor Bureau of:apprenticeship Training keeps state-by-state records on registered union and ion- inior,. apprenticeship programs in construction. These records suggest that what `tappened in other states after peal of their prevailing hapoentc in Utah is y pical or what has . wage Tne ratio or appremices to Journeymen in construction is higher :n states that retain their ^retailing wage laws compared with sates that never had such a law. TLe -ate of P;r prevailing wage laws was substa.^n=•I;J higher apprenticeship zaining in states :nat repealed pre aiIin� r mains tree even when one ccr.jols for before the repeal compared with after the repeal. T::is e regional differences in training rates, the eriec: of unemployment, and long-term trends in training. There are not many minority workers in 'Utah in con .r c:ion, but nationally ,here are evailing w^ ("tifiroriry" here refers to nonwhites, :Hale and f-_rnale.) Some have argued :hat pr minori workers and lower ;he ai en job op Orrin:iics for =skilled ty e.�s will � P law r.,p F' J unen:pioyment rate of rinori ies, relative to whites. However, there is no evidence to supper this cl aim. Black-white unemployment ratios rose in repeal states after repeals. Biwa e laws unemployment ratios tend to be slightly higher in states that have never had prevailir,� -- compared to sates that have retained :heir laws. `Nhile repealing prevailing wage laws probably °sb t has not caused black-white unemployment ratios :o ;o up. There is no evidenc., to sL^decline. + to decll men. ratios a repeal of he Qa��s-Bacor. .�ct wool' cause black-white ,hemp oy o e - wage law-s has tsoeciaily hurt :he training of ninorizes. There a-- ^o r-p ai 0t ?rey3iIing s- proCc ' onaIeI ^ Cre m:n0r:t:es trat led as ccn5M,:C^.on a ^.remises in states 'sat retain v t!1 :r prevai;in; wage :awT cotrpar=d 'Mli, sates tr:at have never had such laws. In repe?1 stzt.,s, proportion of minerces trait ed in construction apprenticeship programs declines szbstanti:llY after the repeals. T"his remains true after con:re(ling for regional dirrerences in relative training rates, une^piovment, and !ong-term ends in -riror:ty training which are inderendeni of stars repeals of prevailing wage laws. The decline in minorry participation in constriction apprenticeships after repeal is tied to a des:int in i:niori non. 'Union apocenticeship programs tend to be Large. �ppren cesr.:� coorcira ors nose �ppre.^.noes rcm contras:or to con:,actor in or;:er, to broaden T e programs s tie a.^ f the a .e e 1�'^+i�.-ily, I ec:use non-tmion ZCp. �u Snip �CO o per nnc . - lar�� . s --•�'l :;na„-:irm pro,rar s, as occo_z:: to een::aetol, the :.Cn-mi0n prop-arms *end o oe Iavis-Bacon , cpcaE - s f _ Law Repealed in 1981 - 001----'-=!- ' '"------------------------------------------------------- ------ . �- ..__ -- -------..._1------------------------------•-- y 4500- --------...__...._ --�-- � �" 1 r nn ------------------ ------------------ - -----------^--._...-._•-- �VQ 1982 Recession ......................... ........... ...... ................... 3.,00- ---------------------- ry Their _ Union Members Car v 3000~ __ Cards in Their Shoes" ----- -------------- ---------- o 2500 ---------------------------------------------•-- •___•----------.--___.___� ----._._ I � Eventually, Members "Find Z 2000- -------------------------------- a Home in;he Open Shop" 1500 1683 „1.c 1987 1989 „ 1977 1979 1981 Quay` rf�e Y Membership mbershi Total s Figure 3.1 Union membership in construction in Utah, 1977-89 Source: Utah State Building and Construction Trades dues records_ of Union membership began to decline with the prevailing wage law repeal and � onset overalt the 1982 recession. Membership recovered somewhat in 1983 but not as fastas construction employment. With the 1985 downturn in Utah construction employment, union membership began a steady decline to less than half its late-1970s peak. r - on- Typic .:lv, unionized cens.:sc:icn workers receive better health and ?e^�1°e:�sr'co triban dco d:e u,^icc;zed wor'.cers. Lower bene ts, parrcularly health and pen s;on c_ , incr�rse in overall labor turr:over in and out of the con ttrainedusand Ilesstz'cPeriencnd Iabor occ;::,arcnal turnover, wa wnil see, led to a younger, ess for:e. The Relation between Repeals and Black Unemployment Davis-Bacon Act was sassed, in par"., to .restrict southern blacks from It as been .�.r;led that the ., �,'rt n�:r. rar o nc=:tm zor.st^uctien goo ocporr—lines. it s fur:':er claimed --hat':'te cure..t an em iovment rates relative to -white unemployment rates ;s partly due to re ero-er c� jack ur. p t sac prevaiiing wage lawn ;:repose on the ability of unskilled black labor to ,.,rrtvete u... skilled write labor. rrorr. Act :hese beliefs, it :s argued that a^reveal of :he Du vz obsacon for less-skilled lower black unemployment relative to white unemployme.t by opening P J black labor.53 the available evidence. Black unemployment These arguments are not directly supported by rates are separately collected for only five of the A ine�sT3 that zot have large znough bave revealed their lack prevailing wage laws. A,--:Ana, Idaho, -New Hampsh:r_, ar rocu!aaons to generate meaningful unemploymentsatisacs. However, Alabama, Colorado, rlcr.da, :{ansas, and Louis:sna•do have suzncient black poouiaa m cns to test the above ar;ueat. T.e -aao of black-to-white unemployment for five repeal stases can be shown usmg^��, ur.emoiovment rates for white anal blacks and white males and black males (fig-Befo.e he repeal to unemployment. bias'.: •izemployment rates are more than r�nce the rat., of white P Or ;tzte prevailing wade laws, however, the male biack-to-white Punempicyment ra=nos o and the overall black-to-white unempdr loyment ra o were both Less than .h.,:r cor..sponding tress ,—.aces repealed their pre'.ailing wage laws. ri This does not mean that the repeals caused the black-to-white ane�ell93Cs lire rpc�oyment sal��S s were r ina across the country in ` Black. o-white unemployment ratios unemployment ratios si:..?iy reflecs this ti—e and elsewhere. The rise in the black-to-white By corzparirg the states that retain their prevailing wage laws with :'rose states that never '- effect of ar•.e tre^ds in black-to-white ad ?revaiiirg wage laws, we can elimirate the ff�� .� v .,� v -�ent ratio ar:d . male black-to-white �e^ciovc:eni ratios. T::-- bl-16: -o-wi:ite une...t::o .. �d to sates r'V3li'na, wage ,aws zompar_ r 4 ^e::cicyme:.t ratio are hots~: lo��er _°r s.:trs With - v ant r r s acrss States and years from 1 n^ +t arevaiiing wage — 3veragir.; :r_rnv.o. :r .. _ to 1952 (fi;.3.4).� iae male nernpicvtnert raaos ;n figure 3.� are �::=ost the same a^ I radstically they are not different. ese data do not support the preoosition that a repeal of the DaVzs-Bacon Act wool ar eiicrate in any signi:cant way the relative urempiovrzent of blacks :o whites. �s is_3acu: Zc_c�c :Tea 3- ---------- - a2.5 ____- ---•- c , 5 2 _...... _o m M x J Ct5 0.5 Never ;'r ad Law Retained I-aw 0 Males All Figure 3.4 Black-to-white unemployment ratio for states that retained and that never had sure prevailing wage laws Source: US DOL Geographical profile of employment and unemplo ment I974-92. Comparing the black-to-white unemployment ratio in states that retained their state prevailing wage laws throughout the last:5 years -rich the ratio in those states that never had state prevailing wage laws eliminates the effect of a strong time trend that shows up in before-and-after analysis. The male black-to-white unemployment ratio is slightly higher in the states that never bad prevailing wage laws compared with states that retained theirs. The difference is not statistically significant The overall black-tu-white unemployment ratio is significantly greater in the states that never having had a prevailing wage law, but this is because of female unemployment differentials, which are unlikely to be significantly affected by construction employment patterns. yV � O ' y tembersnio °er.{s in 1981 I ~........... ........ ...... .--- --�•`' ----------------------•---12C0 30%............ %11 ,above Norm 1 C00-- -----------•----•-•------- , •------ o/ _- -, .....--- ---- -- -- ................ -- - --------1------\--------------------------------- NCO _j I 10%' ~--------- ~ - ------ ------------- ------- .~_---------- —aG0 'o 0 �' •- - ------------- -----_--_---•_-_ -----------`--- 700 Z poi ..�,r..._----- .J,° , l Beiow Normal :raiai^; Rate 0% , . . . . . . . . . C( i�1 --�- p:ur.;cers-nur, ber —0— Az:Tjrendces-;:e::.ent Fil;sre 3.5 Apprentice plumbers as a gercentz journeymen records.plumbers in Lta}s, I9o1-9I Sour.=: Utah plumbers and pipe fitters locals P The plumbers' union in Utah has historically attempted to train apprentices at a rate of IO e union to :S percent of their journeymen members. As employment boomed in mantic shsp�trainia- could not meet journeyman demand and consequently expanded apprenticeship apprenticeship training Was rapidly. As the numbers of journeymen grew to melt demaad.'epaiiia wage law in I98:, reduced to normal rates. But With the repeal of the state p- g g union membership declined and apprenticesaip training rates -were cut to all-time l0w5- e c; r i the apprentice wage rise to 80 percent of a journeyman's pay. In the union sector, apprentices begin at S7 per hour with an additional S3 in benefits. Their wages rise to S14 per hour plus S3 in bene over five y ears. Non-union apprentices are sponsored by a particular contractor hat oversees on-the-job training, and these apprentices take classwor ata participating is coordinator,community college. Union apprentices work under the direction of an apprenticeship tate among employers for on-the-job training, and tak� classes f�e �on PD enticeunity d complete theeges and n apprenticeship centers. Roughly 90 to 95 pert.. programs and graduate to journeymen status, while only 15 to 20 percent of the non-union apprentices gradu_-te. Given these rates, in four years, out of 846 anti-union apprentices, we r 70 journeymen to be graduated. In five years in the union sector, out of expect 1_5 .0 1 ;; should x� e o to journeymen electrician. Thus, while a graduate Y 123 apprentices, 1l0 to 115 apprentice would grad J the nea-union sector accounts for more than 35 percent of all electrician apprentices, it accounts for about 60 percent of journeymen graduates. are aid less wn.h this pat tern whet herein non-union apprentices p Economic henry is consistentP e and graduate at a lower rate than union apprentices. Economic theory posits that in the absent of mar'.cet•,,+ide institutions or government subsidies, individual workers will have to pay for their own on-the-job training when the skills learned are general to an industry and not specific and 'ties of a particular firm. The worker-learner pays for training by accepting a unique to the ac^:v, marginal product. By working wage t`tat is lower than the value to the firm of that workers mar a less &.an the worker's worth to the employer, the worker pays the employer for on-the-job training.That beginning non-union electrical apprentices earn S6 per hour while union app renzces earn ;;0 per hour (including benefits) is consistent with the theoretical proposition that non-union apprentices pay for their own training by taking a discounted wage below their marginal value to the contractor. g ests that :he es not a much for non-, ion training, the theor su g Because the employer do pay employer has no stake in the worker's training. If the worker leaves, the employer does not 1--se the employer will tolerate high levels of any investment in the workers human capital. So, bs 'AM turnover. Because the worker is receiving less than what the worker can earn in ohe o h current no on-the-job training, the worker may be tempted to exit jobs with traitung when personal 'budget needs become pressing. So, on both the employer side and he worker side, turrov,!r is tolerated in the non-union sector. This view is consistent with among the ohagher mo er ',- rates non-union apprentices, but other factors also contribute .o the gh Y percent differential in non-union to union graduation rates. Because the non-union empieyer prices new hands at discounted wages that shield the employer from investing in the human capital of new workers, the employer does not screen new workers extensively to forestall subsequent turnover. The employer's failure to preselect new workers for aptitudes and attitudes consistent with a long-term attachment to construc tion work adds to the r,:rnover among non-union construction apprentices. In contrast, the joint apprenticeship boards of unions and union contractors do considerable preselection for aptitude and attitude before letting a candidate into an apprenticeship program. This is because the union apprentices' training.': contractors and unions will invest in the union app apparent that Lr In e non-union sector, werk--s may also leave apprenticeships if it becomes the er:.ciover offering training at a discounted wage is not delivering or. :hat tr��ring promise to train. Because employers are able to discount wages of apprentices below their current we:-h to Davis-Bacon Rc--csl E.Tccts 43 ;4°c . ..... ............................................................................................ •....-- ------ --z" \. ..7.-- -- _ ...__. f................ ............` 10 '. o/ 1 ........... ...•.__.._._....-------------------------------'1\ ._-. +x- � _ ` o ......... --------------------------------------------------- t�„`�/.... R...... ------------------------------------- `� -- 40,0 70 75 80 8: °0 BPS --*— All �mpIovmeuc �- C,,naact COnSL --I— Suildina -. heavy Hwy S��ally rpic�ment sts:�ide z:gure 3.6 Turnover in T�tah's construction industry compared with �Il e hJ Source: Utah Job Security, Division of Labor Market Information Aasual Report, table 5. As the number of trained journeymen in i;nion hiring hails declines and the number of non-union journeymen declines, firms respond by reducia g tur nover. o vet. -J 'Ldo i� r^ 0 4 states that never had prevailing wage laws, and 19 states that divided into their prevaii.ng wa,, laws, states ed a sate prevailing wage law throughout tilt period. Tnesne arson (figs can eand ) No A retained " "never-had," and retained-law, for comparison 1987 all �� e r of n e to ones repeal, .. d of the first qu arter a en e c the state had repealed its prevailing wage law by 19 r8. By es had passed their repeals except Louisiana which repealed�n 1988. The data for nine repeal scat Pthat 1987 are for the summer of 1987, after Kansas had rep. ed inthose that retained such a law in the "before" period, states that had prevailing and those that had not yet repealed theirs— typically trained a higher percentage of registered apprentices than the states that never had a prevailing wage gl pen dkntohe statwn aesso that would s, the year aPP o astern. During this pre-repeal 19 7 6 is an exc.,ption to this p the eve repeal their laws had as high or h that kept igher raining Thad fallen fodr all states, ut they had e y .rod. B 1987, training rates their laws throughout the p Yge laws. 1989, the states that had fallen least in states that had retained their parni ii rates law as thesues that never had repealed their prevailing wage laws prevailing wage laws lowers prevailing wage laws. This is clear evidence that repealing state p g formal apprenticeship training. prevailing wage A simple analysis can help isolate the effect on training ° repealing ialnang.pApPrenticeship laws from a general downward trend in covet rcnwa apprenticeship awe in the late 1970s and 1980s are training rates for states that repeal their prevailing g� revailin wage laws presented as a percentage of the training rates of sates that retained their p ga rates that out the I970s, before repeals, the repeal states had training ailing (table 3.2., col. 2). Through _ . r above the average training rates for states that had and tra�ngeratesthatprev fell to as were a. o I990, wage laws. After the repeals in the late 1980s, the repeal states h wage laws. By little as 53 percent of the training rates of slates :liar kept their prevailing g v prevailing the reoeal states had relative training rates that were as low asin of stall sates,he all for rep--,I wage laws. Thus, while training in construe:on vas time trends aside — the been falli ..peal states matched states has been the most precipitous and — settin, i:� rates of the retaining states prior to repeal and fell to the rates of states never Navin; the tr,..in g had prevailing wage laws after the repeal•' le linear regression analysis can Unlike the simple analysis just presented, however, a emplmultoyment regional differences control for other factors, such as differences in state un p is a transformation of the training variable in the analysis v de endento a percentage in t:ai.Zing (table 3.3). The P registered apprentices as P .- rate for each state, where the training rate is cal c•mat Fos technical reasons associated with the in a state and ye at�aral log of the of all construc'son employees assumptions of linear regression analysis, the ac:sal dependent variable is the n rate where the odds ratio is calculated as (the percent trained) divided min odds ratio of the training n the 'red . r with nt trai ) d fo by (one minus the percent rates are controlled In the regression model, regional differences in training regionaltatistics categorizatior;s. regions corresponding to standard Bureau of Labor ear The data are for the years 19 7 5- T ent differences are controlled for by state and y equalling 1 once a state lovcn able C,nemp - 78 and I987-90. The focus variable is REPE:A-, a dummy v� ,� which equals zero for ing w reveals ice prevailing wage law. A second focus variable is N-E - lac„ al te pr--vai states !xcept for taziose nine states that never hgd�a7 obser�atiolns n the data et Calsifornia, For those states, N, EF :� equals 1. There ar, -5 t. Davis-Bacon Reveal Ef`=—Q s 47 i 1 I :..U�O -�CAtOf9 t�608a l------------------------------------------ ....._...._._.......------- ......... ---------------------- -------------------..------ . ------- Fesan law x4.00 o I .' ._. y --------------------................-------------- 3.5.'/o I`:` - � t � ------ Na er-ac w C ..._.. --- - ..-�___:- U _ r- 2.0% ---- i 2.1.. ._ ---- --- - --- - 1.00 ---- r'7.aC e I 1 I / t; i0 V I O a% Repel Mates: States NhIch: States which: Repeal —--9753 arG'967-90 Figure 33 Apprenticeship training rates. by state of Labor Statistics and Apprenticesaip Training. Sonnet U.S. Department of Labor Bureaus States are grouped here into four categories, repeal states before and after their repeals of prevailing wage laws, states that retained their prevailing wage laws, and states that never had prevailing wage laws. This simple pattern shows that repealing or not having prevailing wale laws reduces formal training in constriction. (Par. of this before-and- prevailing ovtrail downward trend in registzred apprenticeship rates in after picture i- due to construction :wer`ime.) Repeals aurt apprenticeship training repeals curt unions. because b 'Jon-union construction contractors do less training and less formal. high quality traiai:.g. a Table 3?Training rates in repeal and never-had states as a percentage of training rates is states that retained their wage laws Repeal S tales States Never Having Had Law -) (3) � 1 ( , 1975 10670 ( 85% 109% 1976 1 I..2,% i 1977 100% I 86% 1978 97% I 81% 1987 87% 74% 1988 68"o 52% It 1989 70% I 70% 1990 63% 60% i L ' Except in 1976, the states that never had prevailing wage laws have training rates which fail from 86 percent of the training rates of states that retain their prevailing wage laws to 60 percent of the training rates of states that such laws. Repeal states mirror the training rates of r_taining states prior to their repeals. After the several repeals of state prevailing wage laws—from 1978 to 1988— the average training rate in repeal states falls to 63 percent of the training rates in states retaining the laws. This is a simple way of L viewing the roughly 40 percent drop in registered construction apprenticeship training caused by state repeals of their prevailing wage laws. f � 1; Li i _ nO� .................................... 3: .i 15.4'/.` .L�rtisOea! i-► o ' � i .00 2.. .� 8.0° ;o 4.0% _ 2.0,o ` 0.0% 'Re�ea1 St as: Stomas whi&,: Sties which: Re J S'M2 Yarn•'Yi S 3 anC'�6;3U F`•,^sre 3.9 Minorities as a percentage of all construction apprentices by state soups, 53 Source': US. Department of Labor Bureaus of Labor Statistics and .kpprenuc:sbip Training. in repeal states, before repeal of their prevailing wage laws in construction, minority participation in registered apprenticeship programs averaged 19.4 percent of all ' apprentices. After the repeals, minority participation fell to 12.5 percent of all apprentices. The n=23 and 0=66 refer to the number of state-year observations in each group. States :bat kept heir prevailing wag? laws and states that never had prevailing wgge taws had rougttiv :he same rate of minority participation throughout 19,754 and 1937-90. On average. however, populations of the states that aever had prevailing wage yaps had much higher proportions of minorities. 1 had prevaiiing wage laws, minority representation rates averassd 83 percent throughout the period. :us, bo& repealing states prior to repeal and "retaining" states throughout the period had pprenticeships that mirrored the state population. In minority participation in construction a contrast, both repealing states after the repeal and states which never had prevailing wage laws had substantially under-represented minority participation in construction apprenticeships. Summary Employment in construction is inherently unstable, because the industry fluctuates cyclically and seasonally — and firms expand and contract their employment as they win and lose job bids. Unions have acted like a flywheel in the industry, creating career workers when there were only casual jobs. Unions did this by facilitating the movement of journeymen from employer to employer and minimizing the employers' transaction and screening costs for the training. Unions also Iowered training turnover by providing a mechanism whereby employers and journeymen could rationally invest in the human capital of apprentices. This raised lo ens to promote the he wages of prentices assage so they would stay with training and induced the unionemployers of apprentices to journeymen in order to preserve their investment. Unions also encouraged the career attachment of trained journeymen by providing relatively high wages and additional wages in the form of health and retirement insurance, which are increasingly attractive to workers as they age. By creating career jobs in a casual labor market, unions created the institutions needed to make human capital investment a rational market activity. capital skills With the decline of unions in Utah, the formation and preservation of human cap have become less-rational. Self-investment by apprentices becomes more precarious as the differe::rial between the apprentices' wage and alternative wages in other industries widens. It simpiy becomes more reasonable for apprentices to leave construction if unforeseen personal budget problems emerge. The high turnover among non-union apprentices represents in tie aggrgore e a considerable loss of'human capital .o the construction industry, even though it is not a loss -he employer or the state pays for direc-dy. With the lowering of construction wages, it becomes reasonable for young construction workers to limit the amount of human capital they invest in themselves. With the worker's lower stake in construction skills and with the disappearance of wages in the form of health and old-age insurance, it becomes more reasonable for journeymen construction workers to abandon the construction field when they start families. This represents an additional loss of built-up human capital. Contractors in Utah have attempted to minimize the effect of this increased skill volatility _ within t1he industry by encouraging firm attachment. Still, despite initiatives, such as profit- sharing, 401K plans, and health insurance, designed to attach key workers to a firm, construction turnover remains we'll above the average for the Utah labor market. In short, union decline has _ meant the decline of the career worker within Utah construction, a diminution in incentives to invest in con �. cror. skills, and an increased loss of accumulated human capital as apprentices d career jobs in this and journeymen leave the trades. Although the loss of human capital an industry does not appear as a private cost on the ledgers of any contractor, the industry and socierf at large pay a price for the loss of financially secure occupations in construction. Not only r lowed to dwindle, but the is qual tv in ?.-he industry put at risk when '-urian capital stocks are al Davis-Bacon Qcpeal E:Yc.:s ' S5 s IV. construction Sj:fety Put at Risk Construct on is dangerous work. In fact, it is the nation's most dangerous industry. Azcording to the U.S. Bureau of Labor Statistics: • Nfore than 900 construction workers are killed each year — 3 to 5 per workday. • S 10,500 work-related injuries and illnesses occur annually — almost 2,000 cases per dyy.'. • 'oa,3oo cases involve l most ,Ycrk days, for a total of 4.6 million days lost fro work per y ear. sever, cannot ore t1 d - �;•aren of the h�'aC"s 3Ssc .fated With con m-ic-Aon .vor'<, he igil substar.^a! variability of accidents and their consequences across lob sites and instirtlticnai d injuries are the product of a compiex interaction berxeen worker v�� nments. Accidents an j e11 this environments.o - orb limited, depending on how w and environment, and injuries will be either rest,.. .d or limp p g . . interaction promotes safety. This chapter focuses on the effect of the repeal of state prevalhag wage laws on injuries in ccris.:uctien. The focus'on safety a eat llneer ss s related lto constriction health, at s juncr:r is strictly a concession to the paucity of . .hole d Wv_ night the repeal of a state prevailing wage law arTect �' e?afey record in coNo er and How does the presence or absence of such a law alter the in ucr-cant interaction of environment? C�rtain parameters are key to tf e incidence of inju.^r. For s'.ance, ronsu� °1 wor. is more dangerous when workers are untrained and inexperenced. ..eases associated a lac.< of job sec_ity, the pace of work, and the possible avenues :or grievance all teed into he en an oo site. cnticWl interaction of work and environment an Y , Na A la�.v tr ini.^.� ceclired as In Utah, following the 1981 repeal of the states prevailing g- d the C ..�.. .CtiOn labor market was ,Grog into recession (see c apter k of training a III . The lack 1 r P o n-'T. c widesrread use of inexperienced workers began to surface as the const^:c:on ecororn rebaurded. One experienced pipe' d ipe fitter recalls of that era: .. Contactors were using inexr.erienced people with ao training. They had no training progra= :o peg n with, fey weer airing peopic off the street with no experience in the rade. j�/aat they ;cold do is eve^one that of him.,:.-' on one project that did not have a history or work 3 red sticictr on their hard i1at. 7^-cr� had :o -xce^.^tee on a ohs action ;ao, :.e� a:ui :a veer s wear that for 30 days. Well zver••whe c you would looked ':err ware red st:ckz.s evcrnvhcr i :st:mate that acout 40 or 5G°'a of the peopie had one on their hat. T'ney called them "hahurgcr kids." — Pipe utter, Salt '�aI<e City, 1994 is Lack of training and inexperience are not the only sources of work injuries. In Utah there was a greater sense of job inse_lrity after the repeal of the states pre•:ailing wave law and the �.. �+� r•� r n '�-tIrIiy ex-',:%ion workers `� :.`1 tralnl.^._ a:,d ralat 7 -ec:ine in ut:l�1 w'crk. Z':th03t uni n se i One ry elves r� :ba races ±ey would not r.a'•: :�:=a prior :o •::_ repeal,. e::—..erie:t�.t round 4&bet:� :..:.. 1= V. union worker who was forced :o take -work in the open s:.op rec :.s. c^ the face of intensified competition and cost overruns. There is a tendency to speed up work and - cut back on safeguards in the face of such pressures. Force turnover. When state prevailing wage laws were repealed, worker turnover Work care...s see ed significantly, as the industry found it harder to retain workers for long-term market, increased gn coon labor may t, uni on share of the construction chapter Li). Repeals resulted in a decline in the ate and decreasing union apprenticeship training in the state driving down average construction wages o contractors n. In response to the decline in union membership and training, or construction. P v o d training f minimize screening an g III attempted to reduce turnover— to retain skilled workers and to minim drove experienced P ' pension benefits xP casts. Still, the decline in wages and in health and construc^on workers from their trades for careers in other industries. Thus, while construction firm turnover is on the decline, turnov er industry in the whole try is on the rise. Those who now work on federally funded Davis-Bacon projects are more likely to be union s. They are likely trained because of the demanding nature of these large, civil engineering eobmore likely to have to know more about new processes and changes in technology, and they ar graduated from certified apprenticeship programs. compare with those that never had such In states that retain their prevailing wage law — cornP training is a law or repealed such a law -- the proportion of construction aiid benefits leads toworkers ga flood of higher and injury rates are lower. A decline in wag toa workers inexperienced workers into the industry as well as a decline in skilled, experienced needed m supervise the recruits and to assure that they work safely. Der?ine in the skill base of the construction labor mark& Experience is a major productivity. Training of skilled construction determinant of safe work performance and o rams, most of which are workers is normally conducted through apprenticeship training programs,g operated by unions and employers through joint trust funds. An integral part of this training is lemming on the job while properly supervised. In that way, workers learn from experience while on a va.-iety of projects. Among other things, apprentices are trained to identify and correct j or release pf ergonomic problems, to detect physical hazards, and to detect the presence or cease.o hazardous chemicals. Knowledge about safety and health hazz ds,�appropriateapprenticeship measures, and hazard communication methods are ail important elements that aPP programs provide. apprenticeship ro arras decline and When little Davis-Bacon acts are repealed, training and app P P to continue incentives er . Without employer (chapter III) _ the skill base of workers erodes � � procedures declines as well. renticeshi pro-rams, knowledge of proper safety and health P trainin , an P P decline in g . P na t com etiro , ..ties e factors — cutthroat p Summary- The combination of c� the safe _related skill and to the industry — affects safety -related an erosion of career attachments to or force. Workers become more injury-prone and low experience base of the construction lab the kinds of risks they are taking. Furthermore, as the workforce becomes less skilled less about to limn safety risks . . _ acid its wages in construction decline, workers are forced to take mod-u often press their competitive speed-up make a living. Furthermore, contractors caught :n the P workers to speed up and take more chances. Workers are put at increased risk in an already hazardous industry. 59 Davis-Bacon Rercal E:Yccts ----- __...._. t5.41% (tii3 ►.� 13.54% n n o M1.30 ' nr'O/ J o/ �o J �- _ 4.00% « z.ca% Se Tore Reoeaf Have Law Never H& Law Ai<er Rer!"2! FiQ:re =.1 Injury rates in construction by status of prevailing wage iaw Source: US. Department of Labor, BLS. Injury rates in construction Were relatively low in the nine repeal states prior :o repeal na average, rose to »I percent_ injury rases, o $ , .- - � u repeals, y -.54 percent), after the var.o sJ r. ' fates that have retained prevailing wage laws, injury rates �.ave been and .,3 ,he 32 s o, w s w In nine states that have never bad state prevailing wa,_ la , �+— r�maia relativel; to „ , in;ury rates Were and remaia reiativ:iv high. The aotaron "n refers to :te _umbers of 't state—gear observations in Cato roup. .For instancet the�e wr.e .�0 stat----year i;. combinations for states that had pre`aiiiag Wa;' tames to .7 3 I. L da or roughly S20 billion. T.nis, for a construction labor force which represents but 5 to 6 percent r" compensation, the of the whole U.S. labor force. In addition to the direct costs of workers' re are numerous industry-related indirect costs connected to work-related injuries or deaths. These s and retraining of workers. include job shutdown ction Industry Institute, "even when theestimates of claims are According to the Constru deleted from cost data, indirect costs still exceed the direct costs. Based on the our regression model of the experience of the nine states that repealed their _ prevailing wage laws, we project that national injury rates" will increase by around 15% if the Davis-Bacon Act is repealed. What this will mean in terms of safety is: • There will be 30,000 new cases of lost-time injuries each year, accounting for 675,000 days lost from work. will increase by about S3 billion per year. • Workers' compensation costs • Because Davis-Bacon construction accounts for approximately 10 percent of ail construction, it is estimated that repeal of the Davis-Bacon Act would increase federal construction toss 1 by 5300 million per year in direct, workers'-compensation-related costs alone, and indirect toss would double this figure. The nurnbers might prove larger, because a Davis-Bacon repeal in the wake of st have 3 ate repeals may lamer impact on the construction industry. Summary The institutional context of work is critical to worker health and safety. State prevailing wage little to do with worker health and safety. But such repeal has laws, on the surface, have _ funda ntalIy altered an institutional context that was more conducive to workplace safety. r: Rareais of state prevailing wage laws, therefore, have had hidden et e Because he onsibili for group, take less ..sC tY ._ ., process becomes overheated; because contrac.ors, as a g p, _ bidair., training and safety; because workers feel less secure on the }ob; and because he workforce becomes less attached to and experienced with construction work; construction becomes more dangeres. Safety in an already relatively dangerous industry is put at risk by the repeal of prevailing wage laws. r Davis-Bazo n Repeal Effects 63 MW V. Conclusion The Erects of the Repeal of Prevailing Wage Laws e:n of government in the j;nited States is sometimes called "democracy's Tne federal sys; workshop." The diverse experiences of the 50 states at 'ord a valuable window for assessin u.� successes and failures of public poiicies. Between 1979 and 1988, nine states repealed their prevailing wage laws regulating the cons icnon of public works. These legislative changes enable us to a Gamine the before-and-after ptc=,ires of he effe,Cs of such repeals. Nine other states never gad prevailing wage ia'•vs go�•ening Yuciic construction, while the remaining hirry-:.vo sates retained prevailing wage taws. T:zese ,:,,ever-had sates and "retaining" sates give :s ; ada.:.cn31 oerspe ^vas on `N.at it ze z s .Q <vep or repeal Jreva�:: wags laws. .agislators are o;yen _'or�ad to act on heorv; *.his is ore instance where they can act on Facts and experience. Tre expo^ence of the last Zo years in :he apoiication and removal of state prevailing wage laws on pubiic construction offers insight a�gaconhe prospective effects of further state ret. peals or the proposed repeal of the Federal The Goals of State Prevailing Wage Laws wage ia•xs were irst anacled at :,e stare level. Kansas passed the �rst prevailing wa t .o Prevail.:; law on public wcrks in 1391 as part of iegis:aticn mend ann; the eight-hour work day. Prev u: :.; wage laws were central to a larger effort to inprove working conditions for American cizecs. r w -hat child labor laws should enable children to be in school and the eight- not ion as The . -hat day should help allow workers time :o spend with hear families. ? mer t ftom us`::; .: J„P.t he o vern. d to re ".e proponents or"prevailing wave Ieg:sla..on wanted p ' -�o � �- a��s of iL s citizens. It was believed that ,overn:..�.:. its pu'chasing power to under-nine he w�� shoo::: set an example, by paying the wages prevailing in a locality for each occupation hired .y gov--mment con: actcrs to build public prcyects. 3efore the Great Depression, Arizona, Idaho, Kansas, Massachusetts, Nebraska, New Jersey, passed prevailing wage laws regulating state building and road Oklahoma aSs g o w York and 0 P P ; ' is New , ccnV-_'c-on. In 1931• Congress passed the Davis-Bacon Act. Soon thereafter, I3 addition states World War II and until 19822, 15 more states passed adopted prev aiiir g wage laws. After Wcr.� revailin 'aws. A11 of hose laws raised the question: what was :nea prevailing wage rtt by a p g wage? The Definition of a Prevailing Wade Wages in local labor markets often have a peculiar dis.:+bu6on. Par,icularly where there a:e e in a local labor -,narxet is oRe-ri the S.cst _- unions, but also in other c:rcumsances, th., hlgrest wag. commonly found wage rate. Even when he highest wage occurs most oaten, however, it will not be '.hie average '�sge simply because the lower wages — however tow or ,tier=y for tt:at occupar.o:� will c^.ng the average wage do x,.. y S r cullinguC'•iiZ D- o •�.aQ� laws are :ctende, to ?'_: .:,� gQv'fnT.ent Jl.'L v`t u�l,. t:St a S J. ♦ .eValli� ae Cavis--=ccn R:?c11 _.Pects k cycle, regional differences in wages an rings and d un-2mployment, and long-term trends in ea.- employment -hat are not associated with repeals of prevailing wage laws. We found ,hat the nine repeais cost construction workers in those states 51,477 (in 1994 dollars) per worker every year since state repeal. This was about a 5 percent drop in construction earnings attributable to each state's repeal of its prevailing wage law on public works. A slight increase in construction employment Proponents of state repeals maintained that the Iowering of wages would be offset by an increase in construction employment. While high- paid, high-skilled workers would be hurt by a repeal, it was believed, low-paid, low-skill-!d workers would have more job opportunities in construction. Re-eal proponents were right that cheaper construction labor would lead to an increase n con c:ion employment. Again using regression analysis, we found that the repeal stases experienced a 1.7 percent increase in eonstruc`.ion employment that would not have occur.ed without these repeals. This was an unfavorable trade-off from the standpoint of workers, howev-tr, as their wages fell by 5 percent overall while their employment rose by less than 2 percent. It turned out to be a tough trade-off for government budget-watchers as well. Lost tax revenues. As a group, construction workers lost income, because their wages their total employment rose by Iess than 2 percent. This caused the dropped by 5 percent and government to lose substantial tax revenues. In recent years, the state of Utah has lost S3 million to S5 million annually in sales tax and income tax revenues because it repealed its prevailing wage law in construction. Increased construction cost overruns. Cost overruns are a hidden cost repealing Increased prevailing wage laws. In Utah, the overruns in Utah resulted from an over bidddingg prrods in which contractors, uncertain about each other's labor costs and confronted with the entry or shaved their bids in a desperate effort many start-up construction companies, to obtain government contracts. After the repeal, winning bids on state jobs came in lower than ever before, but the final job costs were a higner percentage of original es:,=-tes than ever berore (chapter 2, fig. 2.3). Having u: '—"-i�d:obs, contractors and subcont'actors would arrange change orders to get the jobs done or si,npiy walk away from badly underbid jobs and leave he state to pick up the pieces. In Utah, cost overruns on the construction of state roads tripled in the 10 years after repeal, compared with the 10 years before. Tne bottom line for Utals's budget The Congressional Budget Office estimates that, should the federal Davis-Bacon Act be repealed, the federal government might save a total of 1.7 percent on its co:is.;ucrion costs. T'nis savings might even be less." Using an even more conserva:Mve `- figure of 3 percent to estimate what Utah saved in construction cons by repealing its prevailm, wage law, we calculate that the Utah state budget almost — but not quite — broke evcn. Balancing construction cost savings against lost tax revenues, in two of the years since 1987 the ' Utah budget saved more money in construction costs than it lost in tax revenues. In five of the years since 1987, the state lost more in tax revenues than it saved in corstr c-ion costs (fig- 5.1). In either case, the net savings or tosses were small compared with the lost earnings of Utah's citizens (table 2.6, row 3). But construction workers — and the industry — were to lose more than monev when these repeals were enacted. r- 67' b Davis-Bacon Zcrcal '::ccis '� v t Other Costs of State Repeals A less-skilled labor force. Unions and union contractors do the lion's share of worker training in the constr uction industry. Some very large , rge ,1-traine n:.onstn:c Ioncontractorsdworkers and .work s but most non-union contractors hire out-or who have learned a trade on a catch-as-catch-can basis. Most �are afrvaid enough to afford to train and retain their own labor force. Contractors, understanda.bly, thnnir that in :he First slack period, the workers they =ainedllnarket failure byave them an inducirwork orurien compe^tors. Unions historically have compensated for this tf e:'s aooreatices. contra::ors to share the burden of gaining and to share each o . . In Ptah, the repeal of the prevailing wage law Ied to a dramatic decline in ur.:on aporer.,:::eship programs because he repeal led to a �r a was not nclitned to pour money into local Having repealed .ife prev I..,g wage law, to, star., comn unity collates and vocational wining centers *.ea along up the differenc . At if in the trough of a business cfcle repeal :n 1981, the Utah construction economy limp o missed. Non-union contractors hired so the absence of quality training systems was not stron5irty c Jon workers pro4'ded arelatively the older eneramon of co out-of-work union members and g SKIi1ed labor force to the Open shoo of nonunion construction. In the last three years, however, Utah has experienced a nassive constr ivJn n bnoo Jo^-g train-, to, systems were in place to ,;,eet this booth. Utah has ::i1ed he Sac by . g rni which is in a construction slump. Utah has also ;-tied o n cons-unnon workers rrom Ca lifornia,a, a less-skilled labor force. Whether vtah will be able to continue to re:y on Caiifor-Iia w�o,k,.:-s a la's economy picks up, many or the skilled Califoru:a travelers wail �-,� een if Californ lobes reta..:..s , likely return Dome to the increased wages there. „ r was not u.. ., tail's experience with declining svailabiliy of cons:rsctien tra1.^.Ina . Comna.^ng the decade before repeals to the decade after repeals, union and nen-iu.^Ian apprt ::zeship rates in, construct:oa fell by more z an half 1n thenine wage laws didsnot Lose ground in prev in- wage laws. States that retW�*:ed ttleir pr vallln, g a rsnticesnip training and states tl;at never had prevailing wage Laws had relatively low `r=-n>:1g PF rates in construction throughout the period. repeal or prevailing wage laws thus had the indirect effect of reducing trainingrepeal, p •<; hindenng the formation of a skilled labor force. When unions declined in �`Ie wake of only ;rate gOVe^,r�.e^.. could have picked up �e ;,feces. jne cost of expanded state-financed voca^onal ^3I.^.i;.g ;s a hidden cost of repealing prevailing a hidden cost Wage laws. So :a:, :t is =ew ^eai states have been walling to pay. , ills in that __ �-� - gains by r-:inor�ty workm A fa:tenng stock of humaa skills Sowed economic g 'wage Psulted from state repeals of prevailing cons:.uc:ion is not the only nonmone:ary cost that re- laws. Con=uctien used to be one of:he few blue-colla.� occupations ee the averageft, where ococst:sc-,:on a coEt-ge education could ear.. a ::addle-cities incot^.e. r income in 1994 was S�;,�00. Becoming a skilled construction worker was a road out OF pove.:y r,.• for rniroc y workers. Before the nine rate repeals, participation by minority group me nbzrs _ male a:^,� `.-male ICaw2ltCes — In CCns1T':c:,on 3 prentiC�Sl:1pS mfrrCCe'� t!'ie T.t 1C.^•ty Paouiations to ezz 'ate. acCcunt�`r tCr In e repeal states before tilt- re?eel ot" z:. ^te%:a.,in, wag mfno - ._s t ' ....._. ...__...--........._...................-............ 100 ----..__ -------- ---------- - IZ 600 ~ - T 40/o o .._ _ G --------"_ 200 0°I° After RepealBefore Repeal 100%Mincrity App. Reflect State Poo. divided by the percenta; of Figure 3.2 The percentage of minority nority reflection prentices in construction, lntage— for nine reel states minority population in the state _ Source: Table 3.8. In the nine repeal states where separate data were available on minority populations, in the decade before repeal, minority apprentices were slightly over- ` represented relative to their proportion of the state population. The minority I� reflection percentage was 107 percent. In the decade after the repeals, the minority reflection percentage fell to 35 percent, indicating significantly under- represented minorities. 1~ - - 3 b Estimated Effect of a Davis-Bacon Repeal -s that repealed Democracy's workshop has given us an opporunity. Z"noe ni e slices of he repel of the fedr�ai Davis-Bacon Acts offer a chance to estimate the likely Q Davis-Bacon Act. Based on. this study, we project the following. or ail First, construction earnngs would drop if the `most 1 law was S5 billionrper your inolreal tzt?nsf every construction workers, this would .nean a loss of al r lower wages in construction, federal income tax collections would fall by year. As a esu.. of to g- - urchased consmicrion almost rou�riy Sl billion per year. Pro}ected cos savings on tedera' P certain , would be ''ess. (fig. 5.=) �� v ;0 percent. T::e Second, we estimate that ormai training in con ...ic:.cn could fail b, l� e-coi'ar work-�rs xrriing a ^ziddle-c!:sa income 'o indus`t-; would :-love from one of ski 'z: biu P* es �ii :ority access to gc•.d a muca-less-skilled labor force earning substantia:l-v� lowt rates. .vouid be using training likely would fall even farther han overall training _ 'filed workers would be � g Y or them, but the less ski and paying less t ,; re construction workers P Y ,ictien serrices would more projects. Purchasers of constr building less and adding less value to buildingp j not necessarily profit from lower-wage labor if that labor is also less skilled. This is a potential lose-icse situa:ior.. =ter its repeal of construction labor force a_ U-ah was abie to patch together a large-enough censtru s.ems in plats n ;e orrac:ors in Utah rode freely on the , a:n:ng sy ' _.r. s P g w .aw. C I Da%—s-But.. Califcr^..ia. But the country as a whole cannot go on a si ii.ar t. '1P �� reoea :d and construction trazn:ng nationally deci;nes szarpiy, the United St_...s • 'll not be a small state like Utah turning to California for its rescue. Nationally, thare will be no plats o .. is the federal government prepared to spend the money to establish its own acprenticesh:p nir- ment induce or _1, _e ccntr�ctors to ;oin pro�ra_-ns in construction? Alternateiv, will ate ge•:ern into :ooperanve raining progra�:s? If preval.. wag. legistation :s -epeai`d, t is likely ;:at sor.:= additional measures will be needed to ensure cccupational trains^g tar e =onstrsc �n indt_s—y. e would produce 30.000 Last, but not least, we estimate that. Z,, e ccns rut:ion job site addi7onal serious injuries yearly. inese njuries would add a large but srl1-urdeterm:red financial cos to :he uirmate price of repeal. o process that lowers It goes wz-hout saying that the public benefits from a bidding pr to rivers must be kept within certai., c_nds, P rt con=uc-ion COsa. But the b'.dding prec.ss ti �o� hilt and societal costs. consepLences that could lead :o increased — rartitr d.an decreased — ac --^,rs on =i'y• States and =ommvnities erno.oy Comce^tive pressures tempt cornac.ors :o C a co,.•. - ,ca�iy ,:.pions a••z 3ssu-'nod the role bu i:Lng inspectors to assure that gL:ilry is ==1ntaz e:. ^.i'.� , of "building inspector" for safery ar.4 =aiaing in .`fie oors�sc^.on :ndusu-y. r o �y1e Indust-y t: Tie role of unions. Ernploy-nent in cons=uction :s inherently :.^stable beaus. a and less fluc�sa:es cyclically and sersorally. Firms expo-*id and contract etnp?eyment as hey job yids. A worker rarely has a lop;-terra attachment :o one employer in .`:e Indus-try. and :he ► work-reiated instirition ate worker mows. construction union may be Lie on.y stable, o ee eor*t:.nities out of a -.-,. act like a flywhdel in the industry, cr.acing carmer op. m le Cons....c..cn unions - „ -;�,:,;.�z ..c:n e p �•'-• t; by :acii::a^ng •tee ,:ic��ement or?e.. casual labor mrrket. nit..^.s do .his and -een,ng costs. to !MF:oy_r and nini niang ^e erroioyers' traZsa:non c. '3 Davis-Bacon Reveal =,_fees r in construction injuries, a 40 percent decrease in apprenticeship training, and an even further decline in minority apprenticeship training. All this was sacrificed to save an estimated 1.7 percent in state construction costs. Even that savings was squandered by the loss in state tax indeed. revenues from an impoverished construction labor force — a poor bargain • n 75 -'' Davis-Bacon Re-,cal E.Kects i ' :i t W. r. w9n 7 5% Percent Percent • decline in Decline Decline rN' in Union in Non- Wage Uaior. W S 100 0.13 510.00 0.87 510.26 S9.49 S;1.00 0.13 S 10.00 0.87 S 10.13 14.1% 6.3% S I0.31 0.13 59.37 0.87 $9.49 11 The percentage decline in union wages is 14 percent (from S12 per hour to 510.31 per hour), and the percentage decline in non-union wages is 6 per=-nt (from SIO per hour :o 59.37 per hour). 49. Unless described differently, figures are given in 1991 dollar amounts. 50.The data are provided in four-digit detail of the Standard Industrial Cand Ema loyment)Statistiics, Data Analysis s ata are from U'S' Department of Labor. Bureau of Labor Statistics, Office of Earnings P Section, Special Tape XC4057, provided by Darrell E. Carr- 5 1. "Secc:iar" trends refers to trends in earrings that are not due to fluctuations in the business cycle nor due to the state -crzals of prevailing wage laws. nings 52.Cc--trolling for contractor type is a cansezvative procedure Overall construction of s a ning�Vthin trades m''�d of a s rift in the mix of construction worker ripe. We are focusing of any .recline resulting from a shift in the mix of trades. Additional earnings losses may be calculated assde line With a shift to a mix of less skilled workers. This is one reason why the regressioc w mixes timate f it in the states is lo-,,,:r than the simple estimate -vhrch does include the effec, of chars g construction industries. Unemployment rates are for each state for each year. 53. T-ris is an annual earnings average by S;C group. When earnings are weighted by the number of workers in each =pup, earnings fall to slightly below 5Z5,000. 54. ?'Ichnic al derails: This regression model was tested on 27,778 observations. Control (dummy) variables for 26 detailed 4-digit standard industry code (SIC)clsssificatioas were included in the regression model but not reported La t!a table. Each coe:ficient reported in column(2)of the table is statistically significant e xcept the control for for�e mo:stain states region. (This mcans that :he estimated regional effect on annual construction earning tcoutim n region of-579 is small and probably not different from zero.) The unemployment Maryland. �rel9mca for ts s C t3.:.D1C states shown in table 2.3 were'4.9 n°with 'he weight being being thedsqusre root 5.9 for s of the annual average gencralized least-iqusres weighted rcc:_ss empioyment in each SIC industry for that year. The R- is0.73, which means the model is a good tit of the data 55. The model also estimates a negative effect on annual earnings of 51173 associated with raising the threshold for c 1 he model contracts covered by a prevailing wage law to 5500,000 or more. This suggests that at some point r recce rais::.3:lie threshold has a siruiar effect:o re:^esiing the law altogether. However,this result is based on expe from only two states,Maryland with a 5500.000 threshold and Oklahoma with a 5600 ra t e whild. ea could cot find hegstive effects on earnings from lower thresholds in the 51001000 tom tha0.0t1i0r0esa ids which 5500,400 have cauriour about .:is result. A censc-Vati%•e iatc?rctation of this result may _ a ^ini i tri impact on construction camin,s while ':ireshoids above S500,000 have progressively negative effects on 79 Ca:•is-I;zcon Reresl EfTects 19. Net--on,Richard " 'State labor legislation etscted in i981" U.S. Depat':^•Cnt of Labor c:nolov=ent Standards AdmIm rtion Division of State E npiovtne tt Standards,!. 20. Salt „ke City Tribune, Jan. 23, 1981. B2. 21. Salt ;;,ake City Tribune January 23, 1981 32. 2'_. Salt Lake City Tribune Jan. 16. 1981, A6, col 1. 23. Salt _' -ke Ciry Tribune, January 16, 1991 (A6, col I). 24. A::zona laws 1984 S.C.R. N0.1001. *Tattle Davis-Bacon Act is R::icd Unconstitutional,' Sept. 12. 1979, p 'j, T;c ?hoenix Gazette, . A l 26. Thi_-blot, !2. 27. Idaho Session Laws Ch 3 HB7. 28. 7 ieblot, 165. 29. Colorado Revised Statutes 1985 Article 16,8-16-101. 30. Thicblot, 159-o0. 31. New Hampshire Ch 117(SB71). 32. 7:ieblot, 184. 33. K-.=s Ch 186 S.3.11:. 34. Louisiana Act No.18 H.B.2. 35. zr;ich, 4 and 6. 36. Erlich, 7. .� Impact of i 37. Re,c==i iaforr=tion(:roue of Data Resources. "Ex--stive Summ sry of the Study of the Economic Repel; �( e Viassachuse::s ?;eva'ling `7�azc :caw" Le--9toc, XvS.•�. Au..T-st r w _ 38. Inc �tassac�use,.:s Foundation `or Econor:ic Research. The peculiar Prevailing 3'agc Law ;1I1 Cabot t;eet N 4 -,jha=� MA 2 1114), March 1988. !' 39. Massachusetts Foundation for economic Resesrch, 1 �i 40. Data Resources. I. 41.Data Resources, S. 4, ter,- ch, 101, 1 i`. and 112.. Davis-?sccn R-pcsl -t~cots 70. U.S. Bureau of the Census, 1970 '—ensu.r of?opularion. h any data are 71.Figures 3.7 and 3.8 include all sates for'o��hich there are noa Bureau+of Apprentrc ship-:airing dataDt�r the of Coiumba, Hawaii, and P hode island f second period- We exclude these states and the District of Columbia (for the same reason). 1?,We do not know what accounts for the unusually high training rate for 'never-had'states in 19-5.Thus anomaly inin disappears when average tra g rates by decades are compared ry assumptions of ar 73.Tht:s ts ?Ic transforms, into the log of an odds ratio m^pion, with thnorm—ae regression weightedeby-' square root of TJe technique used is 3eneraiized .east ecuares regression, cd) times ("torte employment). (pence n: trained) times one minus perccn, train r Cx sure to toxic mate:iais are test orsibie for as mcouated and thus 74. LaIcnt iLresses resulti:3 �rorn _. the costs of wrich are borne as refaced prodtu::�nry, undete-ained additional my-abcr of injuries and illnesses ruined rives for workers and:heir families,and burdens on worxers'cozaPenS3uoa and other soc:sl security ryste.:s. For a mix of reasons, there are no reliable estimates on the aumber of such illnesses. 75.C.Calver,M•Marshall,and C.Connolly,Construction Accidenu: 992 orkers'Compensation✓ata Base, 1983- coring, 1 1988, Washington. DC, OSi�-A Offica of Construction3m 76. In l st•re 4.I, n refers to the number of observatiocs in each state-la out the ryriodr -ts.cc, wore Z;0 combinations for stasis that had prevsilin3 wage laws throng.. pe state-year • 77.Ia•�c case of lost workdays per injury,the:sported result is of the expected sign,but not staur_=11y sitcant- jt^^.ie Kinze, !redirect Costs oj�onsrr•.rc:ion Accidenu, Seattle: T'ae University of Washi:non, 1992, 14. 7g 79. Because of small numbers, there are no reliable csti=tcs on how rexsl would a `ect 'ea` rtes. T'nus. '�e cannot:siculate the projected increase =fatalities due to repeal. If.however, they we per eoro ore atYccted at the same ma�ir,:le as are injuries, we would expect sre increase of 130 to 150 fatalities `or Payments.. 1970-%4 du+published �z 1483 mC :994 KP° rm • .roccued ym • 80. ::tait DepatCnent of Transportation. ,c oral:stiauus � of total costs_tf you aa'wose labor costs try 10 ` 81• The s Y rep ore so email hecwne labor costs on public works are only roughly-S Pcm� F w have cut foul costa ry only :.S persctst 3?.. Geor3e F. Will.•ss time to mpeai the Dow-3acon Act,, Deseret;lewi. February J. 1995- 7 ce Cocnoily,CORsrrc:icri.4cciden-Is: rc ; �rrers'Compensation .� ^�• � silver Michael Marshall.ird CJnstan a^ 8a _..at.es C eern Data Bare, 1985-1988. Office of Consaucuoa and fis+ gnd • r.. r• Davis-Baczn Repeal r.ffects L _ FINANCIAL LM?ACT ON STATE AND LOCAL GOVERN- ENTS OF THE PROPOSED REPEAT, OF OREGON'S PREVAILING WAGE STATUTE Review of the Proposed Impact Statement I -- ill 4 i � 1 c ae� F. Sheehan Ph.D. Mi h . F , I August 1994 i , MICHA.EL F. SHEEHAN PLDLFC :LN,:„tiCE AN-D GErERAL. ECONOMICS 33126 S.W. CALLS v ROAD •• SC APPOOSE, ORECO� 97056 503-513-7173 FA K: 503-543-7173 FIN;LNCIAL IMPACT ON STATE AND LCCAL GOVEc'L", ENTS OF THZ" PROPOSZD REPEAL OF OREGON ' S PREVAILING WAGZ STATUTE Comments on the proposed Imract Statement Michael F. Sheehan, Ph.D. Scappoose, Oregon After reviewing the two SOLI databases, materials from the several committee meetings this week, and a variety of other materials ( including the study my colleague Professor Deter .isle_ and 1 ccmcie_ed in Iowa in 1935 cn same have reached the =011Cw` ng Conclusicns : Qualty of the Database: The SOLI Projects database was not meant to and is not capable of supporting this sort of serious analysis. The following points t' problems: w_1_i illustrate :.ne p o Lack of Validation There is apparently no ef=ort made to require ff that ' source of funds info nation _e provided correctly, or cr that matter at all There is a Lot of Missing Data Many local entities didn' t put down any funding source: Staff assumed that these were local. There are $30 million in split t funding scarce prc j ects. Many .. of these don' t provide the percentages. Staff asar.nes 50i 50, but who knows. There are a substantial number of duplicate entries. Numerical Errors The database is full of what appear to be numerical entry errors; BOLL has caught some, most notably the $25, 730, 0-00 (p.066 ) listed as $267,330,000, but other remain ( see below) . Siven the obvious shakiness of the data set to begin w4l _.., I spc} checked some of t.._ _a. I identified r .Large de:.:ar value items to check at random yes`arday afternoon. Of the five, tiro didn't call back, one was correct at S31 million of local funds; a:_d two had significant problems : - * A Linn Countv Road Department project listed at $10, 400, 465 turned out to be S1, 040, 000; * Jackson Cour.:v' s S17, 6 0, 524 urban renewal budget is iZ real_t.: an'v S1 .76 .m-" I-'on, w= S3CC, '='G0 Ci that federal CDSG funds, which probabi': mak•_s the i ' )a1CLv.P.ShF71.�.Y res:.was W C.fas:.f 3m.asu "�'--- toga Sc+ss.e.► 0. T`= PAGE e j Other Federal Funds The very limited time allowed for research has cut short my effort to identify the universe cf federal construction funding in Oregon. However, I note that the Higher Ed and highway funds are not all the federal construction - funds expended in Oregon.' For example, there is $14 million in - stcta dispersed federal CDBG funds; and direct-to-City CDBG funds to larger cities constitute another significant amount. In sum, in six telephone calls I was able to identify $32 million in reductions in state prevailing wage-covered projects, and a number of other errors and missing data the value of which could not be quantified in the time available. These limited corrections are reflected in Table 17. TABLE 17 BOLI PROJECTS DATABASE i REVIEW OF THREE ERRORS Item Database Corrected Excess Linn Cty Roads $10, 400, 465 S1, 040, 000 $9,360, 465 ' ( Jackson Cty 17, 610, 824 1, 760, 000 17, 610, 824 Urban Renewal (but CDBG) Higher Ed Sederal Funding 133,319 5,010, 715 5,010, 715 T—ITAL $28, 011, 289 S1, 040,0001 S31, 982, OG4 Note: The Higher Ed adjustment is not offset by the $133, 319 i since we have no way of knowing whether the $113,319 is part of the $4, 555, 196. j Rolling these limited corrections into Table 9 gives Table 9A. As is evident from the table, even the application of a small amount of effort in checking has produced significant reductions in the values we've been working with. i i 1 Table 522 in the Statistical Abstract of the United States ( 1992 edition) shows that federal grants to state and local governments in Oregon in 1992 were in ex=ess of S2 billion. And white much of this is certainly not for construction, some of it ti sure_y is. vtQt�,r-s�t�r - ^soc Pas.r :a.r 9ca.uwiar . PAGE 3 TA3LZ 9A PROJECTS BY FUNDING SOURCE c_' 1994 With Limitad Projects Amount Staff Ccrrecticns TOTAL PROJECTS $769, 089, 657 FEO=RAL 51, 139, 079 � I CC? FCTIONS 235, 1_32 � TOTAL NET OF FEDERAL AN-D CORRECTIONS 717, 663, 446 722,000, 000 712, 654, 731 STATE: Higher Education 46, 149, 088 41, 138, 373 Cther State 61, =8Y, 570 T-5 9, 0 s 0, 6-,9 TC"_'AL STATE 1C7, 1v3, 658 I 105, C00, 000 I 30, 189 2_.. LCCA� 506, 230, 778 617, 000, OOC z31, 693, '10 i I ` Note: Adjustments are to the "Amount" column. Source: BOLL Projects Database Conclusion The one year SOLI Projects database which is the basis for the proposed language is not capable of supporting this analysis. Wage Differentials : There is no way to know, given the limited data, wha� the currant wage differential is, especially given divergent market conditions around the state. Last week' s discussions in Salem were based on a 30-9s figure. My analysis, based on current but limited BOLI data, shows a 10$ differentia! in the Portland area, rising to about 20$ in the Salem-Eugene area, to a little under 29-W in the rural areas of the stag . My estimate of a sate-w_,ie we,gh ted average ::aye { d_f' rantial is 17. 88!�. See my Table 0314. %GCNAE,F.S}CEEPW4 a�.e � :3::1 S.W sz. Iud PAGE 4 sa-v�t- ►.�: . .,_ None of the data is very good, but the data in table DS14 is BCLI Oregon data roughly disaggregated by regions . The weights were derived from the BOLL Projects database based on value of construction by region. The 17.88* is rough, but Oregon-specific and recent. Passthrough Figures: The Staff assumes a range of 0$ to 40-W passthrough. This is supported by our Iowa data. Data for Oregon is entirely lacking, however. Given the significance of local competitive conditions, the condition of the labor market, and Oregon' s substantial geographic variations in these markets, it' s not possible to come up with a concrete single figure for the passthrough, even limiting it to a consideration of excess profit retention by non-union contractors. The data is highly variable. Our Iowa data set for Higher Ed projects in Iowa City, for example, contained 14 projects for which we had high and low bids ( Iowa did not have prevailing wage lagislation at the time) . Of the fourteen projects, in three cases the union firms were the low bidders. Of the others, the varia_�.cn in bidding relati•re to the correspcnding low union bids was dra.;�atic, displaying no fi---n ongoing tendency. Unfortunately, data even of this quality does not appear to be available for Oregon at this time. Productivity: Our Iowa study did not consider productivit-r variations be-ween unionized and non-unionized workers. The range of es-imates in the literature runs from 29* to 51-9s to IOC , as noted by the staff in last week' s meetings. None of these figures are Oregon specific. Without having a good productivity figure iz is i impossible to determine a financial impact figure with confidence. t Loss of State Income Tax Revenue on the Lost Wages: ! I have reviewed ORS 250. 125 and have come to the following conclusion,: It may be that no market-mediated effects, e.g. bidding levels on public contracts, should be counted as "direct" effects. The statute and the legislative history would support such an interpretation. On the other hand, if market-mediated effects not "required" by the repeal, are to be included within the impacts to be Z The author has been a active member of the Oregon Bar since 1983 . ►acA4=t.514EEXnv M S SM.cw"ft lo.d _ Sesm.s'.ae�w 'r16 PAGE 5 ' r _ estimated ( in that hey are %nvolve[d] " ) , then it is difficult to ( see how the first round, i. e. "direct, " "reduction in state revenues" should be excluded. What we did in Iowa and what I am proposing here is that first round, i.e. "direct" , reductions in state income tax payments be reflected in the estimates. Using the Staffs basic analysis, I have calculated the direct tax effects, i.e. the loss of tax revenue from decreased payroll and increased excess contractor profits, depending on scenario, of a repeal on income tax revenues--and the effects are substantial re��tive to the magnitude of the range of possible savings. Using =F abstracting awe from the productivi�y and abs�r Y s-a�� assumptions, 9 problem, adding in —e tax effects shows a re-_ necative impact on t:e state treasury ac the low end of the staff ' s range. While I recognize--as noted above--that the tax effects may be i held to be "indirect, " they are certainly no more spec-zlative than the rest of the analysis, given the quality of the data. I Tables DB12A, 123 and 12C present a financial analysis reflecting various scenarios. Each of these tables is based upon the SOLI Projects data set. I offer them to illustrate the tax impacts of repeal -n the event that the BOLI Project data set is adopted as authoritative, notwithstanding what we know about it. One part of Tables 123 and 12C deserves particular emphasis. ;v•::)tice that once the tax impacts are reflected it' s clear ( given t e weaknesses of the data) that the state treasury could actually lose money in net. Revenue Bonds: Last week' s discussions showed the staff tending toward a decision to ignore the fact that some portion of the BOLI Projects database is financed by revenue sources which are neither taxes or general revenues. Some of the more expensive Higher Ed projects, for example, are financed by F( 1 ) bonds- ; some of the local projects are apparently for proprietary enterprises and financed through the sale of goods or services. If these are to be included, this should perhaps be reflected explicitly, as we did in the Iowa study, sinca cc:^-ercial customers and not taxpayers are t:a prospective _eneficlaries in those cases, and the amounts a The BOLI Projects database shows a little over 5200, GCO in Art. XI F( 1 ) bond projects . The facilities star`, however, reported that F(l ) bonding was a little over 57.7 million in Fy 1994 for the same small sa.mpie disc-ssed earlier. KICiAF.'.l.Sf�Htti� f`noe?era.d:.mr Sm.ma D•r 5.W. --- �+ PAGE 6 xa-s,s-� t.�c:•r tr y involved appear to be significant.' _ Summary _ In my view, Y based on m research on this issue to date, whether there is a financial impact either positive or negative associated with repeal is indeterminate and speculative: I base this on the following factors: ( 1) Too many parameter values are either speculative or missing altogether; ( ) P 2 No comparative bid based database similar ,to th e one we had in Iowa apparently exists for Oregon; (3 ) Even using the Staff's assumptions, there is a large variation in the impact range; (4) The Staff s wage differential figure appears to be either too high cr based on the figure for the outlying areas of the state where a relatively small amount of public construction occurs; ( 5) Adjusting for taxes shows that the state treasury could come out a net loser, depending on the scenario; and, ( 6 ) the likely magnitude of the benefits is so small that they are dominated by the large likelihood of error in the data and/or assumptions. In light of all this, interdeterminancy is the only conclusion that can be reached with confidence. ICI August 3, 1994 Scappcose, Oregon � n I followed up with the Treasury to obtain the volume of - revenue bonds issued by sta-e and local government in FY 1994 is Oregon. This value is $780 Million. This includes, however, some F refunding bonds, which, as of this writing, Treasury had not been E able to sort out. saC:.s.r.staatar 13s3 LW.C.U.A_t o 301Pvq.orpa -: PAGE 7 XM-sa.Zr.rwus,}-;-Z ' TABLE 9A FPROJECTS BY FUNDING SOURCE FY 1994 i With Limited Projects Amount Staff MEMMINVEMO� Corrections TOTAL PROJECTS $769, 089, 657 FEDERAL 51, 189,079 CORRECTIONS 235, 132 TOTAL NET OF FEDERAL AND CORRECTIONS 717, 665, 446 722,000,000 712, 654, 731 STATE: Higher Education 46, 149, 088 41, 138,373 Other State 61,484, 570 59, 050, 849 TOTAL STATE _ 107, 633, 658 105, 000, CCO 100, 189, 222 FOCAL 606, 230, 778 617,000, 000 581, 693, 21C If f Note: Adjustments are to the "Amount" column. Projects Database Source: BOLI Pro � j L � 1 TABLE 9 PROJECTS BY FUNDING SOURCE FY 1994 Projects Amount Staff TOTAL PROJECTS S769 ,089 , 657 FEDERAL 51, 189 , 079 C^.PRECTTONS 235, 132 TOTAL NET OF FEDERAL ?L\M CORRECT=ONS 717, 665, 4.16 722, 000, 000 STATE: Higher Education 46, 149, 088 Other State 61, 484, 5701 TOTAL STATE 107, 633, 6-=8 105, C00, 000 r,OCA;, I 606 , 23C , 778 6!7,CC,C, 000 Source: 30LI Projects Database DDT.~E1� i i A1.G63 %en\db 12a.wk1 :NANC!AL!MPAC T S ON STATE AND LOCAL GOVE�NMENT CF REPEAL OF OREGON'S PREVAILING WAGE STATUTE (Staff Assumptions) Item Amount State Local Project Cost Per Year $722.0 $105.0 $617.0 Constrjction Payroll 18.0% 18.9 111.1 Wage Rate Savings 30.0% 5.7 33.3 ------------------------------ Pass7irough 0.0% 0.0 0.0 Excess C,,,ntractcr Profits @ 0%Passthrough 100.0% 5.7 33.3 State Income Tax Revenue Impact @ 0%: Lost Tax Revenue On Reduced Payroll 9.0% (0.5) (3.0) Additional Tax Revenue From Corporate Income Tax�)n Excess Carp Profits 6.6% 0.4 2.2 Net Impact @ 0%: (0.936) -- ' Pass,hrough 40.0% 2.3 13.3 Excess Contractor Profits @ -�0% Passthrough 60.0% 3.4 20.0 Revenue Impact @4C%: Lost Tax Revenue Cn .Peducad Payroll 9.0% (0.5) (3.0) Adcitional Tax Revenue Frvm Corporate Inccme Tax on Excess Corp Prc6ts 6.6% 0.2 1.3 Net !mpact on State J40%: 0.303 SUMMARY Net !moact 0% (0.936) 0.000 Net 'rrract @ 40% 0.303 13.327 S;urce: Basic zata from CAS Memo cf.'u,y 22, !994. Tax rates trcm :1,,e Department of Revenue. Gar.'c�t 2a.+vk t A1.G63 %en1db12b.wk1 FiNANC1ALIMPACTS ON STATE AND LOCAL GOVERNMENT OF REPEAL OF OREGON'S PREVAILING WAGE STATUTE (Staff Assumptions with Ltd Corrections) Item Amount State Local _ Project Cost Per Year ---- ------ $712.7 $100.2 $581.7 Construction Payroll 18.0% 18.0 104.7 Wage Rate Savings 30.0% 5.4 31.4 Passthrough ------ 0.0°6 0.0 0.0 Excess Contractor Profits @ 0%Passthrough 100.0% 5.4 31.4 State Income Tax Revenue Impact @ 0%: Lost Tax Revenue On Reduced Payroll 9.0% (0.5) (2.8) Additional Tax Revenue From Corporate Income Tax on Excess Corp Profits 6.6% 0.4 2.1 Net Impact @ 0%: (0.884) -------------------------------------------------- Passtrough 40.0% 2.2 12.6 Excess Contractor Profits @ 40%Passthrough 60.0% 3.2 18.8 Revenue Impact @40%: i - 1 Lost Tax Revenue Cn Reduced Payroll 9.0% (0.5) (2.8) i Additional Tax Revenue From Corporate Income Tax on Excess Corp Profits 6.6% 0.2 1.2 Net impact on State @40%: 0.308 SUMMARY Net Impact @ 0% (0.884) 0.000 Net Impact 'D 40% 0.3ca 12.565 Source: cask ata born DAS Memo of July 22, 1994. Tax rates from the Department of Revenue. t d A 1.G63 %enldb 12c.wk 1 FINANC;ALIMPACTS ON S T ATE AND LOCALGOVERNMENT OF RE' OF OREGON'S PREVAILING WAGE STATUTE (Wage Differential 17.38%with Ltd Corrections) Item Amount State Local Project Cost Per Year $712.7 $100.2 $581.7 Construction Payroll 18.0% 18.0 104.7 Wage Rate Savings 17.9% 3.2 18.7 - ough 0.0---- ----------------------------- ------ - 0.0 Excess Contractor Prcfits @ 0% Passthrough 100.0% 3.2 18.7 State Income Tax Revenue Impact @ 0%: Lost Tax Revenue On Reduced Payroll 9.0% (0.3) (1.7) Additional Tax Revenue From Corporate Income Tax on Excess Corp Profrts 6.6% 0.2 1.2 Net Impact @ 0%: (0•5277) Paszzhrough 40.0% 1.3 7.5 Excess Contractor Profits @ 4C% Passthrough 60.0% 1.9 11.2 Revenue Impact @40%: List Tax Revenue On Reduced Payroll 9.0% (0.3) (1.7) Additional -ax Revenue -rom C:.rxrate Income Tax on Excess Corp P-dits 6.6% 0.1 0.7 Net 'mpact on State @4C%: 0.184 SUMMARY Net Impact @ 0% (0.527) 0.0cO Net Impact .@ 4C% 0.184 7.489 Saurce: Basic data from DAS .Merno of.my 22. 199s. Tax rates cm ti,e Dacartmert;:t =avenua. J A REVIEW OF: "EFFECT OF THE DAVIS-BACON ACT ON CONSTRUCTION COSTS IN NON-METROPOLITAN AREAS OF THE UNITED STATES" Martha Norby Fraundorf John P. Farrell Robert Mason OSU 1982 ro The 1982 Fraundorr study (Study) provides no support for the proposition tha t repeal of the Oregon Prevailing Wage law would produce any positive savings to state and local governments in Oregon. If anything, it supports just the opposite proposition, i.e. that repeal would have no impact The weaknesses of the Study relative to this proposition are as follows: The Study Only Deals with the Federal Davis-Bacon Act Projects covered by state Little Davis-Bacorr Acts were deleted from the national data set by design. For Oregon, of the one or two public projects probably surveyed (see below), neither were covered by the Oregon prevailing wage law, since the e Or e on revailino Study explicitly eliminated all Oregon projects coveredby the g p w the survey. .33 bottom right). wage law from y (p * VirtuallyNo Data From Oregon The Study's conclusions were based on 215 b interviews with contractors nationwide. Only 39 successful interviews were from projects from the thirteen states of the Western Census region, including California. There is no information to the study that tells us how many of the 39 interviews were from Oregon. 94% of the population of the Western States region was = in Oregon in 1980 (see below). If we assume that successful interviews were roughly in proportion to the populations of the various western states in the Census Region as of 1980, then only 2-4 of the 39 interviews would have been from Oregon. Since the interviews were "paired" (p.31), this means that the total number of public projects surveyed in Oregon was no more than two. The Study presents no information at all relating specifically to the Oregon projects studies, if, indeed, there were any. _ =` * The Study Found that the Federal Davis-Bacon Did Not Add Significantly to Costs in the Western Census Region Including Oregon The Study found that "the (Davis-Bacon) Act in fact has little effect on casts in rapidly growing regions such as the West and South." (p.25, left column). _ Rs,M;+mow+=mL:a+ nus 1-w. -i...r ti PAGE 1 ,0-,4.,:FAX S&I-n. * The Study is Dated All projects in the Study's database were drawn from the period 1976-9, with 95170 of the projects from 1976-78. * All Projects Were Rural All projects in the Study's database were, by design, _ eon toda y, roughly 72% of all from rural counties (p.31). In contrast, to Oregon y, g Y projects covered by the Oregon prevailing wage law are from the Portland :Metro and Salem and Eugene areas, with only roughly 2817o (in terms of dollar value) from the remainder of the state. BOLI data shows that the union/non— union wage differe:tial is sharply lower in the non—rural areas where the bulk of the value of public projects occurs. �t Buildings The Study was designed to include only projects involving buildings, projects covered b the Oregon Prevailing Wage law whereas a large share of l Y g g P are road and highway projects, which tend to have a lower labor coefficient than do buildings. . * Productivity/Quality/Building Type The Study is explicit in saying that differences in the types of building beings built, other federal non—Davis—Bacon federal requirements (for instance, arfirmative action (p.2=) and a greater insistence on quality work), may produce some or all of the difference in casts right attributed to the Davis—Bacon Act variable (p.23 g ht column). * Peculiarly, the Study Shows Dramatically Higher Costs for All Public Even Those Clearly Project Expense Items, E Y Unrelated to the Davis—Bacon 1 P Act The Study claims that fcr federal projects compared to comparable private projects, materials costs are 54O10 higher, equipment cost 158% higher, "other h direct costs 1.1.E .o h and overhead,and profit n 80% hierr.� (Table 4-12 p.23) none of these items are controlled by the Davis—Bacon Act, and there is no reason to suppose that project costs--if the Study's assumption that the l apples"public and private protects are "apples and , p 1e " but for the Davis—Bacon aspect--should differ by these large levels. These results imply :hat there is some other factor that is driving all costs (including labor) and that Davis— Bacon is not the cuiora. * Not All Variations of the Model as Developed by the Authors Produce a .1„ Davis—Bacon Impact The authors tested various versions of u....r mode?. Not all of these versions produced statistically sipifcant values for the Davis— Bacon variable. (p.22, right column and Table 4-11). Iarzer Firms Mean Highe, Costs The Study did not incorporate a fire size variable into the final model form chosen, even though 'he authors found that sine was statistically sin izicant and tended to reduce the impact ascribed to the Neon ph r -0 Sam.="O z;.i S.w PAC,E 2 r Davis—Bacon variable. (p.24). * The Cost Impact Attributed by the Study to the Davis—Bacon Act is Many Times Larger than the Differential in Wage Rates Also Found in the Study The Study claims an increase in labor costs of 78.2% ('72, p.23), even though the Study also finds the nationwide wage differential is only 20%. (p.22; p.23, 412). ROOM.33cen"N.ML:CM M=A.W.C.O. L.► PAGE 3 ���nz 4 Prevailing Wage Laws Effect on School Construction Independent studies done by Peter Phillips, Ph.D.',University of Utah, Economic Department, Howard Wial, Ph.D., The Keystone Research Center, and Mark J. Prus, Ph.D. Department of Economics SUNY Courtland on Prevailing Wage Laws,; and their effects on the cost of school construction. Complied By Orange County Labor Management Cooperative Trust 309 N. Rampart, Suite M Orange, CA 92868 (714) 939.3146 ASS wm RFR01� OF G CLERK JDAN L FLYW CCtY Prevailing Wage Studies Table of Contents Introduction ■ Do Higher Wages Raise Labor Costs? Or do workforce skills have greatest cost impact? By Robert Gasperow ■ Do Lower Prevailing Wages Reduce Public Construction Costs? By Howard Wial ■ Square Foot Construction Costs for Newly Constructed State and Local Schools, Offices and Warehouses in Nine Southwestern and Intermountain States. Prepared for the Legislative Education Study Committee of the New Mexico State Legislature. By Peter Philips ■ The Effect of State Prevailing Wage Laws on Total Construction Costs By Mark J. Prus ■ Prevailing Laws and School Construction Costs An Analysis of Public and School Construction in Maryland and the Mid Atlantic States. By Mark J. Prus ■ A Comparison of Public School Construction Costs In Three Midwestern States that Have Changed Their Prevailing Wage Laws in the 1990's By Peter Philips Ph.D. ■ Prevailing Wage Regulations and School Construction Costs: Evidence from British Colombia By Cihan Bilginsoy and Peter Phillips Ph.D. ■ Kentucky's Prevailing Wage Law Its History, Purpose and Effect. By Peter Philips, Ph.D. ■ Kansas and Prevailing Wage Legislation on By Peter Philips, Ph.D. ■ Losing Ground Lesson from Repeal of Nine "Little Davis-Bacon Acts Working Paper Economics Department University of Utah. By Peter Philips, Ph.D., Garth Mangum, Norm Waitzman, and Anne Yeagle ■ Prevailing Wage Laws and School Construction Costs Study on Davis Bacon and Virgina. By Peter Philips, Ph.D. Other Studies: (All of these studies have findings consistent with the studies included herein) These additional studies are not included in this folder. • Riech, Michael, Ph.D., UC Berkeley Institute of Industrial Relations, Prevailing Wage Laws and the California Economy, February, 1996. ■ Biiginsoy, Cihan, Department of Economics, University of Utah, Apprenticeship Training and Prevailing Wage Laws, February, 1996. • Carlson, Richard, Spectrum Economics, Inc., Impact of Repealing California's Prevailing Wage Laws on California's Local Economies, February, 1996. ■ Petersen, Jeffrey, Ph.D., School of Public Health, University of California, Berkeley, Health Care and Pension Benefits for Construction Workers: The Role of Prevailing Wage Laws, April, 1997. ■ Petersen, Jeffrey, Ph.D., School of Public Health, University of California, Berkeley, The Effects for California Construction Workers from Changing the Method of Calculating Prevailing Wage Benefits, February, 1996. ■ Phillips, Peter, Ph.D., Economics Department, University of Utah, Results of Multi Variate Regression Analysis of Construction Workers Incomes with a Focus on the Implementation of Prevailing Wage Policies, February, 1996. ■ Waitzman, Norman, Ph.D., Department of Economics, University of Utah, Worker Beware; The Relationship Between the Strength of State Prevailing Wage Laws and Injuries in Construction, 1976-1991, 1996. Prevailing Wages Introduction Scores of studies have examined the impact of the payment of prevailing wages and benefits throughout the United States. ■ The payment of prevailing wages (living wages) does not increase the cost of construction significantly, if at all. ■ In fact, the payment of prevailing wages often reduces costs because of the increase in productivity, the decrease in job site injuries, and savings derived from significant reductions in future maintenance because of the higher quality of the original construction. ■ Where prevailing wages are paid, training programs are more prevalent, including higher participation by minorities and disadvantaged workers. ■ The failure to pay prevailing wages and benefits creates a direct cost to taxpayers because it shifts the cost of health care and pensions from employers to public health systems. • The failure to pay prevailing wages creates direct costs to taxpayers and insurers because of the significant increase in job site injuries created by less-well trained workers. • The failure to pay prevailing wages often increase construction costs, reduces tax revenues, lowers the general economy and forces skilled workers to migrate to other areas, further reducing local tax bases. Do Higher Wages Raise Labor Costs? Or do workforce skills have greatest cost impact? By Robert Gasperow Robert Gasperow Bob Gasperow is the executive director of the Construction Labor Research Council (CLRC) in Washington, D.C. He has headed CLRC since it was established by employer associations in the 1970's to develop data of construction labor costs and related information. He is responsible for maintaining those data bases, as well as producing various reports and analyses of industry performance and trends. Mr. Gasperow also represents the industry on the Business Research Advisory Council of the Bureau of Labor Statistics and makes frequent presentations at industry conferences. He received an M.B.A. in Quantitative Analysis from Columbia University and a B.S. in Economics from Miami University. Summary: A 14-year study of highway construction in the United States from 1980-1993, found that skills and productivity, not differences in wage rates, the critical determiner of bottom line labor costs. The federal study found that the payment of prevailing wages and the use of higher paid higher skilled workers reaped an average of$123,057 per mile in savings. The study found that"There is no basis to the claim that lower wage rates result in lower construction costs. Do Higher Wages Raise Labor Costs? Or do workforce skills have greatest cost impact? Do higher wages raise labor costs? Not according to Bob Gasperow, Director of the Construction Labor Research Council (CLRC). In a review of Federal Highway Administration (FHA) information in 1995, Bob Gasperow analyzed the available data to determine the correlation between wages, man-hours and highway construction expenditures. His study's findings illustrate how skills and productivity- not differences in wage rates - are the critical determiner of bottom line labor costs. Owners, public bodies and local and state legislators tend to believe higher wages add up to higher construction costs, and their reasoning seems to be because prevailing wage opponents constantly promote it. Gasperow's analysis uses data compiled by the Federal Highway Administration (FHA) that shows construction expenditures or cost savings are related to wages and productivity - and not to wages alone. Three other factors make the FHA database ideal for this type of scrutiny. It is 1. objective, 2. comprehensive and 3. neutral (not designed to evaluate labor costs). In addition, the data encompasses 14 years so that exceptions and atypical projects reported in a particular state in a particular year have little or no impact upon the findings. Statistics included in study cover all fifty states over the 14-year period from 1980 though 1993 with the following volumes: All States _Top 26 States Total Construction Dollars $87.1 billion $67.9 billion Roadway Miles 98,454 68,976 Bridge Miles 2,138 1,598 Total Construction Miles 100,591 70,573 Labor Hours 1.5 billion 1.2 billion The total volumes listed above are actual construction expenditures. They do not include engineering costs,purchases of right-of-ways or any other cost that is not directly related to actual construction. The analysis selected a grouping of states that averaged over$100 million annually to eliminate any variables that might occur in lower dollar volume states. Highway Costs in 26 Top $$ Volume States (1980-1993) Low Wage* High Wage** U.S.Average Average Hourly Wage $9.76 $17.65 $12.15 Man-Hours Per Mile $22,837 $13,697 $18,348 Labor Costs Per Mile $216,864 $241,465 $235,603 Total Costs Per Mile $1,141,049 $1,017,992 $1,136,963 *Low Wage States:TX,GA, IA,FL,AL,MN,MS,TN,NC,CO,VA,LA,WV` **High Wage States:OH,IL,WI,PA,MO,MI,WA,CA,NY,IN,AR,OR,NJ 1 These 26 states represent • 78% of the total construction dollars, • 70% of total construction miles and, • 79 % of total labor hours over the 14 years. As the above table clearly shows, the man-hours to complete a mile of highway are 40 % lower in the high wagestates - in spite of an 81% higher wage rate. And total dollar costs per mile between low wage and high wage states are 11 % less expensive in high wage states when compared to an 81 %wage rate differential. The bottom line:The use of higher paid; higher skilled workers reaped an average $123,057 per mile savings in the high wage states.This is despite the fact that rates in these states averaged $17.64 an hour compared to $9.21 per hour in lower wage states. Higher skilled productive workers are the key to a project's cost-effectiveness. This study documents that there is only minimal correlation between the hourly wage rate paid to labor and the cost of a mile of highway. Moreover, the limited correlation which does exists indicates that the relationship is inverse - higher hourly rates tend to equate to lower highway cost per mile. CLRC'sGasperow explains that the amount/cost of any single factor in highway construction - various mixes of equipment, labor, materials and management-reveals little about total cost. Up to a point, factors are substitutes for each other because they may be exchanged. Similarly, within a factor category, there may be substitutes. For example, workers with varying skill levels may be utilized. Although there are higher costs per unit of time for the more highly skilled, these workers require fewer labor inputs. Therefore, if the gain in output per unit of time exceeds the premium paid to the more highly skilled worker, this becomes a more cost-effective alternative. The analysis of FHA data documents the impact on highway costs of utilizing various amounts of labor inputs at varying hourly rates. Gasperow explains: "It substantiates the lack of correlation between labor inputted into a mile of highway and total cost of project. "Using higher skilled, higher hourly cost labor substantially lowers the required labor inputs- often to the extent that cost per mile is lower when paying higher hourly labor rates. Gasperow's conclusion: "There is no basis to the claim that lower wage rates result in lower construction costs." Note:To see a full copy of the information above contact the National Alliance For Fair Contracting and ask for a copy of the publication tided, "Wages,Productivity,and Highway Construction Costs".1 North Old State Capitol Plaza,Suite 525,Springfield,IL 62701 Phone:(866)523-NAFC Fax:(217)522-6588 2 Do Lower Prevailing Wages Reduce Public Construction Costs? Howard Wial Briefing Paper 99/2 - For additional copies of this report contact: KEYSTONE RESEARCH CENTER Keystone Research Center•412 North Third Street• Harrisburg,Pennsylvania 17101 Phone: (717) 255-7181 •Fax: (717)255-7193 • Email: KeystoneRC@aol.com This report was published and printed in-house. • KEYSTONE RESEARCH CENTER Do Lower Prevailing Wages Reduce Public Construction Costs? July 1999 Howard Wial PREVAILING WAGE KFYSTONE RE AK44 CENTER KEYSTONE RESEARCH CENTER The Keystone Research Center(KRC),a non-partisan think tank with offices in Harrisburg and the Philadelphia area, conducts research on the Pennsylvania economy and civic institutions. This research documents current conditions and seeks to develop innovative public policy proposals to expand economic opportunity and ensure that all State residents share in the benefits of economic growth. The Keystone Research Center is a non-profit organization as described in section 501(c)(3)of the Internal Revenue Code. All contributions are tax deductible. ABOUT THE AUTHOR Howard Wial is a Senior Fellow of the Keystone Research Center and adjunct professor of law at Rutgers University, Camden,New Jersey. He holds a Ph.D.in economics from the Massachusetts Institute of Technology and a law degree from Yale University. Dr.Wial is co-author of the 1998 Cornell University Press book New Rules for a New Economy: Employment and Opportunity in Postindustrial America. ACKNOWLEDGMENTS Carol Ramsey designed and did the layout of the document. ©Keystone Research Center, 1999 Do Lower Prevailing Wages Reduce Public Construction Costs? Howard Wial Briefing Paper 99/2 For additional copies of this report contact: KEYSTONE RESEARCH CENTER Keystone Research Center•412 North Third Street• Harrisburg,Pennsylvania 17101 Phone: (717)255-7181 • Fax: (717)255-7193 • Email: KeystoneRC@aol.com This report was published and printed in-house. PREVAILING WAGE KEYSTGEW RESEARCH CENTER Overview and benefit laws and the cost of school construction. Pennsylvania law requires construction contractors working on publicly funded • Lower prevailing minimum wages paid to construction or renovation projects to pay workers have no measurable impact on workers at a minimum the"prevailing wages public construction costs partly because and benefits"in their respective trades in the wage declines lead to offsetting declines in geographical area in which the work is per- productivity. formed. In 1997, the Pennsylvania Depart- ment of Labor and Industry(DLI)imple- • Real savings in public construction costs mented a change in its method of detennining are more likely to come from investments prevailing wage and benefit rates resulting in a in worker training,which can make work- reduction in the legally required prevailing ers more productive,thereby lowering rates in many construction trades in much of costs without cutting wages. the state. (Although the Pennsylvania Com- monwealth Court ruled them illegal on Janu- By threatening the construction ary 11, 1999,the prevailing wage and benefit industry's skill base, lowering prevailing rates DLI implemented were in place for more minimum wages and benefits moves than a year.) Pennsylvania's construction industry in the opposite of the direction in which it needs to Opponents of prevailing wage laws move to reduce construction costs over the often argue that repealing or weakening these long term. laws will lower the cost of public construction projects because such action will reduce the Pennsylvania's Prevailing Wage Law wages and benefits of construction workers employed on public projects. Data on Penn- Since 1961,Pennsylvania has required sylvania public school construction projects construction contractors who work on publicly during the period in which lower prevailing funded projects (such as public schools, rates were in effect do not support this claim. government buildings,and highways)to pay construction workers at least the prevailing • Data on public school construction costs wages and benefits for their respective trades show no strong evidence that in the area in which the work is performed. Pennsylvania's lower prevailing wages and This prevailing minimum wage requirement benefits reduced construction costs exists, in one form or another,in many states charged by contractors performing public and at the federal level as well. The purpose works. of the requirement was to prevent building contractors from using out-of-state or out-of- Studies from other states show no statisti- cally significant relationship between the and wage base. existence of prevailing minimum wage 5 PREVAILING WAGE RESEARCH CENTER Since 1979,nine states have repealed In June 1996, DL1 decided to set their prevailing wage laws.' Opponents of prevailing wage and benefit rates in a different prevailing wage laws believed that the laws way. Instead of relying on submissions from kept construction workers' wages and benefits interested parties as contemplated by the artificially hi h because the prevailing mini- statute DLI contracted with the Pennsylvania Y g p g mum wages set under the laws were often the State Data Center to conduct a statewide hourly wages paid to unionized workers, survey of construction workers' wages. The which are generally higher than nonunion survey covered the period from January 1995 workers' hourly wages. If wage and benefit through May 1996. It was sent to building rates were no longer propped up by law, contractors, although certain contractors and prevailing minimum wage opponents argued, certain types of construction projects were wages would fall to their"natural"market excluded. Using the survey results,DLI levels yielding lower construction costs and would declare the prevailing wage and benefit saving taxpayers money. rate in a particular trade and county to be the wage and benefit rate paid to a majority of Pennsylvania's prevailing wage law workers in that trade and county based upon has not been repealed. However,the Pennsyl- DLI's criteria for the determination of majority vania Department of Labor and Industry(DLI) status. If no single wage and/or benefit rate recently weakened the law's effect by chang- was paid to a majority of workers, DLI would ing its method of determining prevailing rates. declare the prevailing wage to be the average (The prevailing wage statute gives DLI the wage and benefit rate of all survey responses. responsibility for determining the prevailing The new method of setting the prevailing wage and benefit rates in each trade or craft.) wage and benefit rates went into effect Febru- Before February 27, 1997,DLI set the prevail- ary 27, 1997,but was the subject of a legal ing wage and benefit rates by accepting challenge which stayed its implementation or submissions of wage and benefit information use from April 8, 1997,until August 13, 1997, from any interested party, including building when the stay was lifted. contractors and trade unions. It then declared the prevailing rates to be the hourly wage and In many trades and counties where a benefit rates that were most frequently paid to large share of construction workers were workers. While the wage and benefit rates of covered by a collective bargaining agreement, nonunion workers often vary from job to job, the prevailing wage and benefit rates were the person to person, and employer to employer, same under the new method as under the old rates in the unionized sector are generally one. But in trades and counties with a lower unform in an area for each specific trade. For union presence or a low survey response rate this reason,the prevailing wage and benefit for unionized projects,the new method usu- rate was usually the collectively bargained ally resulted in a decline in the prevailing wage and benefit rate. minimum wage and benefit rate because that wage and benefit rate was no longer the collectively bargained rate. 6 PREVAILING WAGE KEYSTOW A Kti CENTER The Effect of Lower Prevailing Wages Table 1 shows the average cost per and Benefits on Construction Costs square foot for school construction projects for which contracts were awarded in the months Opponents of prevailing wage Iaws before and after the wage-determination method would expect Pennsylvania's public construc- was changed.' The table shows the average cost tion costs to have declined after the impact of for eight groups of Pennsylvania counties. Each collective bargaining agreements on prevailing group of counties is characterized by the number wages and benefits was weakened. To find of building trades or crafts (out of ten) in out whether such expectations are warranted, which the new prevailing wage and benefit we can look at evidence on school construc- rates were the collectively.bargained wage. tion costs in Pennsylvania during the two-year (There are no before-and-after data for groups period from October 1996 through October of counties in which the new prevailing wage 1998. If lower prevailing wage and benefit and benefit rates were the collectively bar- rates reduced construction costs to public gained rate in one or six trades, so those bodies,then the average cost per square foot county groups are excluded from the table.) for public school construction projects would be expected to have been tower after August 1997 than it was in the earlier part of the two- year period.' Table 1. Average Cost Per Square Foot* for School Construction Projects in Pennsylvania, October 1996-October 1998 Before and After Prevaiflini r Wage Determination Methh d Was Changed Number of Trades,in Average Cost Per Average Cost Per Percentage County Group,in Square Foot Before Square Foot After Change in Which New Prevailing Prevailing Wage Prevailing Wage Average Cost Wage Was the Determined Method Determination Method Per Square Collectively Bargained Was Changed(October Was Changed Foot Wage" 1996-August 1997) (September 1997- October 1998) 0 $ 61.50 $97.33 +58% 2 37.06 77.31 +109 3 125.43 59.28 - 53 4 110.00 238.00 +116 5 80.25 113.90 +42 7 116.71 160.95 +38 8 113.35 159.40 +41 9 151.19 73.61 -51 10 101.12 108.13 + 7 *Average cost per square foot is the ratio of total costs of all construction projects in a county group to total square feet covered by those projects. "Each county was classified according to the number of trades in which the new prevailing wage was the collectively bargained wage. The counties were then grouped together according to this number. For example,"2"indicates that the data pertain to all counties in which the new prevailing wage was the collectively bargained wage in two trades. 7 PREVAILING WAGE KEYSILME RESFARCH CENTER The table shows that average construc- prevailing wage rate change on construction tion costs per square foot declined only in costs from other influences on those costs. counties where either three or nine trades had Multiple regression analysis of Pennsylvania their new prevailing wage rates set equal to school construction data shows no across-the- the collectively bargained wage. In all other board decline in costs after the prevailing county groups, average costs rose after the wage law's determination methods were new wage-determination method went into changed! effect. Studies of prevailing wage laws in Table 1 also provides data on which other states have also produced no strong county groupings saw the largest declines or evidence that these laws have any impact on smallest increases in average cost per square construction costs. foot. Proponents of weakening prevailing wage laws would expect construction costs to . A study of 15 Great Plains states showed fall most(or increase least) in the counties in that the average cost per square foot of which very few trades had their prevailing building new schools did not differ signifi- wage and benefit rates set by collective bar- cantly between states that had prevailing gaining after the change in the prevailing wage wage laws and states that did not.5 determination method. These proponents would therefore expect the top rows of the last . A comparison of school construction costs column of Table 1 to be more negative(or less between prevailing wage and non-prevail- positive)than the bottom few rows. In fact, ing wage states in the Mountain West and the data shows the opposite of what would be Southwest found that average costs per expected if lowering prevailing wage and square foot were actually lower in states benefit rates reduced construction costs: while with prevailing wage laws.' the trend is erratic,the top rows of the last column of Table 1 tend to be higher than the . A multiple regression study of school bottom rows. construction costs in Maryland and other Of course,the projects included in mid-Atlantic states found that prevailing Table l differ in terms of many other things wage laws have no measurable impact on that may affect their costs,not just whether costs. their contracts were awarded before or after The most cited statistical study that the change in the method of determining the produced a contrary result, a multiple regres- prevailing wage rate. For example,they differ sion analysis of the impact of the federal by the number of square feet in the project,the prevailing wage law(the Davis-Bacon Act)on cost of living in the relevant county,the type construction costs in non-metropolitan areas of school,and the particular month in which of the U.S.,was flawed by its deliberate the contract was awarded. A statistical tech- exclusion of publicly funded projects that nique known as multiple regression makes it were not subject to the requirements of the possible to separate the influence of the federal law.' The study thus compared public 8 PREVAILING WAGE Utz^:rSiCe#, RE&EARCH CENTER i projects subject to the federal prevailing wage Why Do Lower Prevailing Wages and law with private projects. Because many Benefits Have No Measurable publicly funded projects such as highways, Effects on Construction Costs? P Y P J ( public transportation systems, and prisons) have no private counterparts,however, it is Many people,without examining the impossible to hold constant the characteristics evidence,might expect the lowering of of the project when comparing public and Pennsylvania's prevailing minimum wages private projects. Therefore, it is impossible to and benefits to reduce construction attribute public-private cost differences to the workers compensation and,with it, construc- public sector's prevailing wage and benefit tion costs. But this would be true only if rate requirement. construction workers' productivity did not change when workers' wages fell. When The recent experience of two Pennsyl- workers become less productive, construction vania school districts shows that even in- costs rise. If the decline in wages causes creases in legally mandated prevailing wage productivity to drop, the effect of the produc- and benefit rates do not necessarily increase tivity decline(to raise costs)works against the public construction costs. In March 1999, effect of the wage decline(to lower costs). If after two months of legal uncertainty about the two effects cancel each other out,the required prevailing wage levels,DLI began prevailing wage change will have no effect on issuing prevailing wage rates that were higher construction costs. If the productivity decline than the 1998 rates. The Blue Mountain outweighs the wage decline,costs will actu- School District, in Schuylkill County,was ally rise. planning to renovate its high school. In April 1999,the school district's construction man- There are several reasons to expect ager estimated that construction costs would construction industry productivity to decline increase by about$670,000 as a result of the substantially when the prevailing wage rate is higher prevailing wage and benefit rates.9 But lowered. Large wage cuts demoralize workers when bids for the project were opened on May and encourage skilled and experienced work- 6,the low bids,which were expected to be ers to travel to higher wages areas or leave the about$15.1 million came in at only industry about indust altogether. Over a longer period of $13.8 million, almost 9 percent below the time, lower wages threaten the continued anticipated level.10 And in April,bids for a existence of construction apprenticeship middle school construction project in programs(most of which are established Tamaqua,which used the same prevailing through collective bargaining and adminis- wage and benefit rates as the Blue Mountain tered jointly by union and employer represen- bids, also came in under budget estimates.il tatives).12 Apprenticeships,which usually last 9 PREVAILING WAGE WXYSTONE RESEARCH `ESTER four or five years, give workers a broad set of in Great Plains states that repealed their prevail- con- g wage r in e laws or never had such laws. In Great skills within their respective trades. If g struction workers' wages are too low and Plains states that kept their prevailing wage laws, workers are unlikely to spend entire careers in the number of apprentices fell by only 27 percent the construction industry,costly apprentice- during the same period.13 ships no longer make economic sense for workers or employers. As apprenticeships Conclusion decline,the construction industry loses its base of skilled workers. As a result,produc- For the period in which they existed, Pennsylvania's lower prevailing wage and tivity declines. Individual employers are u benefit requirements had no measurable impact unwilling to pick up the slack and train work- on construction costs in the state. If the lower ers themselves because they know that any worker they train can easily go to work fora prevailing wage and benefit*rates were intended competitor. Moreover,nonunion employers to save taxpayers money,they failed to achieve that goal. Real savings in public construction have rarely been able to create or sustain costs are more likely to come from investments apprenticeship programs that are comparable in scope to the collectively bargained pro- in worker training,which can make workers grams. more productive,thereby lowering costs without cutting wages. By threatening the construction In other states,eliminating prevailing industry's skill base,lower prevailing rates move wage laws has driven down wages and reduced Pennsylvania s construction industry in the investment in apprenticeship training. Between opposite of the direction in which it needs to go 1973 and 1990,the num ber of construction to reduce construction costs over the long term. apprentices declined by an average of 53 percent 10 PREVAILING WAGE LA KFfSTCOE K&SARCH C ENT FR FOOTNOTES Peter Philips,Kansas and Prevailing Wage Legislation,study prepared for Kansas Senate Labor and Industries Committee, February 20, 1988,p. 14. The prevailing wage law in one additional state,Oklahoma,was judicially overturned. 'The data analyzed are from the F.W.Dodge Company and include information on the bid price at the start of the project,the structure type,project location,project scale,and technical characteristics such as number of stories and type of frame. The Dodge data also distinguish between public,private,and federal projects. 'The data include six private school construction projects and 70 public school projects. The inclusion of this small number of private schools,which are not covered by the prevailing wage law,makes little difference to the analysis. In addition,the prevailing wage law may have an indirect effect on wages paid to workers on private construction projects if public construc- tion projects account for a large share of all construction industry employment in a trade and locality. 'The multiple regression analysis included all Pennsylvania school construction projects for which contracts were awarded between October 1996 and October 1998,except for those in counties where either one or six trades'new prevailing wages were equal to the collectively bargained wage. (The latter projects were excluded from the regression analysis,as they were excluded from Table 1,because there were no before-and-after data.)The regression related the logarithm of the total cost of each project to a measure of the cost of living in the project's county,the time when the project was built(measured in months after October 1996),the type of project(addition,addition/alternation,alteration,alteration/renovation,new construction, unknown/new,or unknown),the number of square feet,the type of school(public or private,and elementary,middle,or high school),and whether the project was built after August 1997 in each county group. The regression results showed that,after controlling for other influences on construction costs,costs went up in some county groups and down in others after the prevailing-wage change took effect. (In which groups costs increased and in which they decreased depended on the precise statistical specification used.) However,none of these cost increases or decreases was statistically significant at the 5 percent 0 level except,in one specification,for a decline in costs in the nine-collective-agreement county group. s Philips,Kansas and Prevailing Wage Legislation,pp. 18-21. s Peter Philips,Summary of New Mexico Study of School Construction Costs,University of Utah,Department of Economics, no date. 'Mark Prus,Prevailing Wage Laws and School Construction Costs:An Analysis of Public'School Construction in Maryland and the Mid-Atlantic States,study prepared for Prince Georges County(Maryland)Council,January 1999. B Martha Norby Fraundorf,John P.Farrell,and Robert Mason,Effect of the Davis-Bacon Act on Construction Costs in Non- Metropolitan Areas of the United States,Oregon State University,Department of Economics,January 1982. 9 Mary Ellen Maher-Harkins,"Prevailing Wage Debate Hits County;Blue Mountain Sees Project Cost Rise,"Schuylkill Online,April 27,1999,www.pottsville.com/pub/1999/Apr/27/D977592.htm 11 Mary Ellen Maher-Harkins,"Blue Mountain Bids Come In Low;School District May Save More Than$1 Million in Construction Costs,"Schuylkill Online,May 7,1999,www.pottsville.com/pub/1999/May/7/E2374.htm "Maher-Harkins,"Prevailing Wage Debate Hits County." 11 PREVAILING WAGE r RCH CENTER. 12 In Pennsylvania during the 1990-95 period,labor-management apprenticeship programs operated jointly by unions and construction contractor associations accounted for 72 percent of new entrants into apprenticeship programs(and a higher share of those graduating from apprenticeships). U.S.Department of Labor,Bureau of Apprenticeship and Training,data provided by Cihan Bilginsoy of the University of Utah. During this period,unions represented about 22 percent of construc- tion and Earnings Data Book Washington,D.C.: tion workers(Barry T.Hirsch and David A.Macpherson,Union Membershipsg ( Bureau of National Affairs),various years.) Therefore,Pennsylvania unionized construction firms did about 9 times as much apprenticeship training relative to their employment as did non-union firms. 13 Philips,Kansas and Prevailing Wage Legislation,pp.35-36. 12 Harrisburg Office: 412 North Third Street•Harrisburg,Pennsylvania 17101 Phone: (717)255-7181 • Fax: (717)255-7193 • Email: KeystoneRC@aol.com Philadelphia Office: 314 Springton Road•Upper Darby,Pennsylvania 19082 Phone: (610) 352-2798 • Fax: (610)352-2798 •Email:krcphila@igc.org This report was designed and printed in-house. Square Foot Construction Costs for Newly Constructed State and Local Schools, Offices and Warehouses in Nine Southwestern and Intermountain States 1992-1994 Prepared for the Legislative Education Study Committee of the New Mexico State Legislature September 6, 1996 by Peter Philips Professor University of Utah Economics Department Abstract:This study compares p q public square foot construction costs in five southwestern and 0 Intermountain states that have state prevailing wage laws with four states in the same region that do not have state prevailing wage laws. The five"have-law"states are New Mexico,Texas,Oklahoma,Wyoming and Nevada. The four"no-law"states are Arizona,Utah,Idaho and Colorado.' These states are often used by either New Mexico or Utah in making public school teacher salary comparisons. During the time period of the study, 1992-94,elementary schools cost$6 per square foot less in the five-state group with prevailing wage laws. Middle schools cost$11 less per square foot in the states with prevailing wage laws. High schools cost$11 less per square foot in the states with prevailing wage laws. Warehouses cost$35 dollars less per square foot in states with prevailing wage laws and office buildings cost$2 per square foot more in the five state group with prevailing wage laws. When Texas is removed from the"have-law"group, elementary and middle schools still cost less in the remaining four"have-law"states while high schools cost $5 more per square foot in the remaining"have law"states and offices cost$8 more per square foot in the remaining"have law"states. By far,the highest number of observations is for elementary schools and this group generates the most stable and reliable results with an 8%cost advantage favoring the"have-law' group when Texas is included and a 5%cost advantage favoring the"have-law"group when Texas is excluded. New Mexico's 14 newly built elementary schools during the time period of the study, 1992-94, cost$66 per square foot as compared to$72 in Arizona,$72 in Utah and$80 in Colorado,the three surrounding states without state prevailing wage laws. Construction costs are sensitive to regional differences in cost of living as well as prevailing wage law regulations,building design and other factors. Consequently,when the number of observations falls for a specific building type,these results must be treated cautiously. An aggressive conclusion from these data would be that prevailing wage laws promote collective bargaining and apprenticeship training and consequently lower public construction costs. A more conservative conclusion from these data notices that these cost differences found do not provide support for the proposition that the elimination of prevailing wage laws saves on public construction costs. ' Two cities in Colorado,Denver and Peublo retain prevailing wage regulations. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 1 Contents I.About the Author 3 I1.Acknowledgments 3 II1.Introduction 3 IV.Review of the Literature on School Construction Costs 6 Bureau of Labor Statistics Report 6 Work of Steven G.Allen 7 V.Why Union Contractors Train More 8 Training and Utah's Repeal of Its Prevailing Wage Law 8 Nonunion Sector's Response to Utah's Repeal 10 General Pattern of Construction Training in the U.S. 11 VI.Construction Costs in Nine Southwestern and Intermountain States 13 New Public Construction, 1992-94(Table 5) 14 Distribution by Type of Construction(Table 6) 15 Distribution by Type and State(Table 7) 16 Square Foot Cost in No-Law States(Table 8) 17 Square Foot Cost in Have-Law States(Table 9) 18 Comparison of Costs in 5 Have-Law States with 4 No-Law States(Table 10) 19 Comparison of Costs Excluding Largest Have-Law State-Texas(Table 11) 21 VII.Conclusion 21 Appendix A:Construction Data for New Mexico 23 Appendix B:U.S.Census of Construction Data on Labor Costs as Percent of Total Costs 24 U.S.Data 27 New Mexico Data 37 Endnotes to Appendix B 42 Peter Philips,Ph.D.,Professor Economics Department,University of Utah 2 I. About the Author Peter Philips received his B.A.from Pomona College(1970)and his M.A.(1976)and Ph.D.(1980)from Stanford University. He is a Professor of Economics at the University of Utah. He is co-editor of Three Worlds of Labor Economics and co-author of Portable Pensions in Casual labor Markets. Philips has published widely on the canning and construction industries in a variety of academic journals including Industrial and Labor Relations Review,Industrial Relations,Business History Review,the Journal of Economic History,and the Cambridge Journal of Economics. Philips has received awards for his teaching and community service including the University of Utah Presidential Teaching Scholar and the Lowell Bennion University Service Professor. Philips has served as an expert advisor to the U.S.Justice Department on the history and impact of the Davis-Bacon Act. He has testified before legislative committees in California,Indiana,Ohio and Oklahoma regarding aspects of those states'prevailing wage laws. Some of Philips'previous research has been funded by the U.S.Labor Department,by alumni of the University of Utah's economics department,by the AFL-CIO and by construction industry organizations. Readers of this report who have questions,comments,suggestions or criticisms may contact the author at the Economics Department,University of Utah,Salt Lake City,UT 84112;(801)585-6465; PHILIPS@ECON.SBS.UTAH.EDU. The author welcomes and encourages such communication. II. Acknowledgments This report was prepared in cooperation with Sharon Ball,Senior Research Analyst,Legislative Education Study Committee,New Mexico State Legislature and presented to the Legislative Education Study Committee on Friday,September 6, 1996 in Albuquerque,New Mexico. Helpful comments and suggestions were received from Cihan Bilginsoy,visiting Assistant Professor,University of Utah,Hamid Azari,Instructor,Economics Department,Alabama State University at Montgomery and Evalyn Hunemuller,New Mexico School Administrators. These individuals have been most helpful. The content and any errors in this report remain the responsibility of the author. Some of the material on the effect of the repeal of Utah's state prevailing wage law on apprenticeship training in construction in that state was previously published in Hamid Azari,Anne Yeagle and Peter Philips, "The Effects of the Repeal of Utah's Prevailing Wage Law on the Labor Market in Construction"in Sheldon Friedman,et a1.,Restoring the Promise ofAmerican Labor Law, Cornell University ILR Press,Ithaca,New York, 1994,pp.207-22. III. Introduction A primary reason to repeal state prevailing wage laws is to attempt to save on state public construction costs. It is argued that state prevailing wage laws encourage the use of unionized construction contractors. It is well known that union construction workers receive higher wage rates and consequently,it is argued that these higher wage rates lead to unnecessarily higher public construction costs. This report attempts to test the proposition that the elimination of state prevailing wage laws lowers public construction costs by focusing on the states surrounding New Mexico that do and do not have state prevailing wage laws.The design of this study is to select states that New Mexico uses when making comparisons of public teacher salaries. New Mexico is well situated for such a study design because some of the surrounding states have state prevailing wage laws(Texas and Oklahoma)and some do not(Utah,Arizona,Colorado). In addition,because the author is from Utah and has an interest in the effect of Utah's repeal of its prevailing wage law on state expenditures,the states surrounding Utah are also included in this study. 2 The time period for this study is 1992 to 1994. In the fall of 1995,Oklahoma's state supreme court declared its state law unconstitutional. In 1982 Oklahoma's legislature,in order to save administrative costs,dropped its procedure for determining what the prevailing wage rate was and adopted the federal Davis-Bacon rates. In 1995,the state supreme court declared this an unconstitutional delegation of state powers to the federal government. For the purposes of this study,Oklahoma is treated as a state with a prevailing wage law. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 3 These are Nevada and Wyoming which have state prevailing wage laws and Idaho which does not have a prevailing wage law. Also,I understand Nevada school teacher salaries are sometimes used in comparison with New Mexico public teacher salary comparisons. In any case,the data are presented in such a way as to see the effects of the inclusion or exclusion of any one or several of these states. States such as California are excluded from this comparison of public construction costs for the same reason that California would be excluded from a comparison of school teacher salaries. Significant variations in the cost of living alter private and public sector construction costs. One way to control for this effect is to select states with similar costs of living. This study also focuses on a short time period, 1992 to 1994 in order to reduce the confounding effects of inflation and the business cycle on comparison of construction costs. All costs are analyzed in 1994 dollars using the consumer price index urban as a deflator. The major focus of this study is on school construction costs because this type of construction is commonly affected by state prevailing wage laws and because this type of public construction provides many observations allowing for more reliable statistical results. For comparison,this study also includes public offices and warehouses. Not surprisingly,there are considerably fewer public office buildings and warehouses built in the selected region during the selected time period. The results for these types of buildings should be viewed conservatively for both statistical and practical reasons. Statistically,the fewer observations only permit more tentative conclusions. Practically,if one wished to save public construction cost by altering prevailing wage law regulations,one should focus on where most state public construction takes place. That would lead to a relatively sharper focus on school construction compared to other building types. This study is in three main parts plus two appendices. Section IV reviews the relevant literature on school construction. No published studies exist that focus on school construction and prevailing wage laws. However,the U.S.Bureau of Labor Statistics did publish one study of school construction costs in the early 1970s. This study found that despite an almost 50%wage rate differential in school construction between the Northeast and the South in 1972,labor costs as a percent of total construction costs were basically the same. The study attributed this to possible differences in labor productivity and the cost of materials by region. Again studying construction in the 1970s,economist Steven Allen found that union construction labor productivity was 20 percent to 50 percent higher than nonunion construction labor productivity. However, Allen also found that schools built by union contractors were more expensive than schools built by nonunion contractors. Allen's sample was small so that he was forced to combine elementary schools with middle and high schools. Furthermore,Allen's sample contained only 1 I nonunion schools and 57 union- built schools across the entire U.S.. In contrast to Allen,this study presents data for 223 schools built only in nine Intermountain and Southwestern states. This allows me to control for differences in square foot costs associated with differences between high schools and elementary schools as well as eliminate the confounding effects of regional differences in cost of living. Section V presents an analysis of training in construction as it relates to prevailing wage laws and union/nonunion differences in contractor strategies. This section shows that in the case of Utah, apprenticeship training declined dramatically after the repeal of Utah's state prevailing wage law. Neither nonunion contractors nor state vocational education programs significantly made up for the loss in jointly managed union-contractor programs in the union sector of construction. Economic theory identifies two problems or market failures that lead to greater training when there is collective bargaining in construction. These are the free rider problem and the bait-and-switch problem. Basically,because construction workers often move from contractor to contractor as work shifts from construction site to construction site,nonunion contractors are hesitant to train workers who will only end up working for their competitors. In the collective bargaining context,the rule is"who comes around goes around". For example,I train the apprentice who may end up working for you while you train the apprentice who because he or she is a 0 member of the union and I am a union contractor,I know he or she may end up working for me. Thus,the Peter Philips,Ph.D.,Professor Economics Department,University of Utah 4 collectively bargained contract tends to overcome the free rider problem and permit training. U.S.Bureau of Apprenticeship data reflect this by showing that three out of every four construction apprentices enrolls 0 in a joint union-management apprenticeship program even though union contractors employ only 25%to 50%of the labor force. Furthermore,in union-management apprenticeship programs graduation rates are relatively high. This is because there is a jointly hired apprenticeship coordinator who serves as a policeman insuring that the quality of the training is good and apprentices are rotated through a variety of work experiences. On the nonunion side there usually is no policeman insuring that training takes place. Thus,some nonunion contractors are tempted into a bait-and-switch strategy of promising training but only providing helper experience. Consequently,the graduation rate on the nonunion side of apprenticeship training is much lower. The combined consequence of free-ride and bait-and-switch problems means that almost nine out of every ten journeyman construction workers graduating from formal,monitored, structured and planned apprenticeship programs come from the unionized sector of construction. This accounts for the higher productivity of unionized construction contractors that tends to pay for the higher union wage rates. Section VI presents the results of comparing the square foot construction costs in five types of public buildings(elementary schools,middle schools,high schools,offices and warehouses)in five states with prevailing wage laws(New Mexico,Texas,Oklahoma,Wyoming and Nevada)with four states without prevailing wage law regulations(Utah,Arizona,Colorado and Idaho). These results show that if anything, square foot construction costs are lower in the states with state prevailing wage laws compared to those without these laws. Comparing New Mexico to the four states without prevailing wage laws yields the same result. Excluding Texas from the group of states with prevailing wage laws(because Texas has a majority of observations in this group)also yields the same result. While one cannot conclude that prevailing wage laws definitely lower state construction costs,one can conclude that in this region of the country and these types of structures,especially schools and especially elementary schools(where we have the largest number of observations)there is no evidence to support the proposition that state prevailing wage laws raise public construction costs. Appendix A lists the F.W.Dodge data for New Mexico used in this study. It should be noted that all cost data in this study are accepted bid prices or start costs. These costs do not include change orders or other sources of cost over-runs. F.W.Dodge does not collect final cost data. An attempt was made to collect final cost data but this proved to be impractical. A large and well financed study might succeed in collecting final cost data but there is no central source for these kind of statistics. In Appendix B to this report,this study presents data from the U.S. Census of Construction showing that labor costs as a percent of total costs in construction runs around 25%to 30%. This is helpful data in setting the limits on plausible estimates of how much money one might save by eliminating state prevailing wage laws. Clearly any estimate asserting that one might save 25%to 30%on total public construction costs by eliminating state prevailing wage laws is unrealistic given that labor costs including benefits are only 25%to 30%of total costs. Peter Philips,Ph.D. Professor Economics Department,University of Utah 5 p , P IV. Review of the Relevant Literature. Summary of a U.S. Bureau of Labor Statistics Study of Wages and School Construction Costs In 1979,the U.S.Bureau of Labor Statistics published a study of school construction costs by region in the United States. In contrast to the material I will present below,the BLS study did not break out schools into elementary,middle and high schools. Nor did it break down construction by state. Rather,the BLS study aggregated school types and presented data on four regions,northeast,midwest,south and west. The relevant data for our purposes is presented below. Table 1:Hourly Wage Rates and Total Costs as a%of Total Construction Costs Elementary • Secondary School • • 1972 Hourly Wage Rate Wages as a Percent of Total Cost Northeast $73 27.9% North Central 29.3% South $ 27.3% West $7. ... 29.0% Source: U.S.Bureau of Labor Statistics,John G.Olsen,"Labor and Material Requirements for New School Construction,"Monthly Labor Review,April 1979,Vol. 102,Number 4,p.41. The key point to be derived from Table I is to note that hourly wage rates varied considerably between the Northeast region and the South($7.75 versus$5.22 in 1972). In contrast, wage costs as a percent of total costs were almost the same in the two regions(27.9%versus 27.3%). The analyst,John Olsen,commented on these facts as follows: Average hourly earnings also varied by region. Hourly earnings for all construction workers averaged$6.78,ranging from$5.22 in the South to$7.75 in the Northeast. Wages as a percent of contract costs varied from just above 27 percent in the South to slightly above 29 percent in the North Central. Although average hourly wage rates in the Northeast were higher than those in the North Central region,wage costs as a percent of total contract costs were lower. Among other factors,this irregular trend could result from regional differences in productivity rates and in relative material costs. (pp.40-41) Why could differences in labor productivity account for the fact that an almost 50%difference in wage rates between the South and the Northeast did not result in any difference in labor costs as a percent of total costs? The answer is simple. If labor productivity in the Northeast was 50%greater than labor productivity in the South,either because of better training,better equipment or both,then the higher wage rates of the more productive workers would be offset by that greater labor productivity. Is it reasonable to believe that construction workers in the Northeast in 1972 were roughly 50%more productive than Southern construction workers at the time? Peter Philips,Ph.D.,Professor Economics Department,University of Utah 6 Summary of Steven G. Men's Work In 1984,an economist at North Carolina State university,Steven G.Allen,published in The Quarterly Journal of Economics an article entitled"Unionized Construction Workers Are More Productive".3 Professor Allen summarizes his study as follows: Apprenticeship training and hiring halls probably raise union productivity[compared to nonunion workers]while jurisdictional disputes and restrictive work rules lower it. Using Brown and Medoff s methodology,union productivity measured by value added per employee is 44 to 53 percent higher than nonunion. The estimate declines to 17 to 22 percent when estimates of interarea construction price differences are used to deflate value added.(p.251) In other words,prior to adjusting for differences in regional cost-of-living and differences in regional construction material costs,union construction labor in the 1970s,the period of Allen's study,was roughly 50%more productive than nonunion labor. The wage rates and material costs in the BLS study were not altered to factor in the effect of differences in regional cost-of-living. Thus,the BLS study is quite consistent with Allen's work and their conclusions are similar. Wage rate differences of 50%across regions with differences in productivity and cost-of-living may not alter labor costs as a percent of total costs. Within a region such as New Mexico or the Intermountain west,where the cost-of-living and the material costs of construction is similar,20 percent differences in wage rates in construction can be offset by differences in productivity between union and nonunion labor. In a subsequent paper also published in the Quarterly Journal of Economics,AIlen comes up with slightly different conclusions. In"Can Union Labor Ever Cost Less?"Allen concludes: ...union contractors have greater economies of scale. This gives them a cost advantage in large commercial office buildings,but in school and hospital construction,nonunion contractors have lower costs at all output levels. Despite the cost differences,profits for nonunion contractors in school and hospital construction are no higher than those for union contractors because the burden of higher union costs is shifted to the buyer.4 In other words,based on a study of 57 union built elementary and secondary schools and 11 nonunion built elementary and secondary schools,Allen concludes in his second article that the union built schools cost more and that the union contractor did not absorb those added costs. Rather those added higher costs were passed on to the school boards and tax payers who paid for that construction. I emphasize Allen's result because it is not consistent with the data I have analyzed for New Mexico and eight other Intermountain or Southwestern states from the early 1990s. The differences in my results which I present below compared to Allen's may be due to 1)differences in time,Allen's data are from the early 1970s,2)differences in region,Allen does not indicate what region of the country his school data are drawn from while my results are specific to nine states,3)effects of mixing elementary,middle and high schools together--Allen does not break out school types while I do;or 4)differences in sample size. Allen is comparing 57 union built schools with 11 nonunion built schools while I compare 223 schools built in prevailing wage law states with 109 schools built in non-prevailing wage law states. As I will point out below,the results of the this study are that square foot cost of school construction in the selected states are,if anything,lower in the states with prevailing wage laws compared to those without prevailing wage laws. Obviously,not all schools in states with prevailing wage laws in this region are built 3 Quarterly Journal of Economics,May, 1984,pp.251 to 274. The Quarterly Journal of Economics is the second oldest economics journal in the United States and is published by Harvard University. It is considered one of the premier journals in economics. 4 Steven G.Allen,"Can Union Labor Ever Cost Less?"Quarterly Journal of Economics,May, 1987,pp. 347 to 373. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 7 by union contractors and not all schools in states without prevailing wage laws are built by nonunion contractors. Allen does not break his analysis down along the dimension of legal regulations. For the purpose of analyzing the advisability and usefulness of prevailing wage law regulations,it is more useful to break construction projects out by whether they were built under these regulations or not rather than whether they were built union or nonunion. Nonetheless,it is appropriate to address the question of differences between union and nonunion contractors as this issue underlies much of the debate surrounding prevailing wage laws. V. Why Union Contractors Train More The Case of Utah's Prevailing Wage Law Repeat When Utah repealed its state prevailing wage law in 1981, training for construction, both in union apprenticeships and through vocational schools declined in Utah. Union apprenticeships are tied to the availability of union jobs. For example,unionized plumbers and pipe filters in Utah, United Association of Journeymen and Apprentices of the Plumbing and Pipe Fitting Industry of the United States and Canada, historically have attempted to maintain apprenticeship rates at between 10 and 15 percent of the number of union journeymen plumbers in the state(fig. 14.3). As employment boomed in the 1970s,however,the union could not meet the demand for journeymen from the unionized contractors. Consequently,the union increased apprenticeship rates to a peak of 25 percent in 1975. The boom persisted,but the backlog had been remedied. So the union lowered its apprenticeship rate back to normal ranges by 1978. Employment during the construction boom peaked in 1979 and membership in the plumbers and pipe fitters'union peaked in 1981. With the repeal of the Utah prevailing wage law, the union dropped its apprenticeship rate to 10 percent, a historical low. Union membership fell slightly in 1982 and then began a steeper decline in 1983. Faced with these sustained declines in membership, the union cut its apprenticeship rate to historical lows in 1986 and thereafter. Unions hit harder by declines in membership have scaled back their apprenticeship programs further. The carpenters'union,Utah locals 194 and 1499 of the United Brotherhood of Carpenters and Joiners of America,which graduated seventy in a class in 1977, graduated five in 1992 (Hansen, Paul. Weber Basin Job Corps Carpenters Union Instructor,personal interview, Sept. 2, 1993). The Utah International Union of Bricklayers and Allied Craftsmen suspended its apprenticeship program altogether. The decline in union apprenticeship training in Utah has not been offset by a rise in other sources of training. Because the repeal of Utah's prevailing wage law was motivated by a desire to limit state expenditures, state legislators were not eager to raise funding for state-sponsored vocational training. Delmar Stevens, who has taught building trades construction at Salt Lake Community College since 1971, chronicles the decline in enrollment. I started at the college in 1971. We went about two or three years and then the enrollment just really started to grow. [When I started we had]probably about 45 students with the first year and second year programs. Then about 1973 or 74 it really started to grow. It jumped up to 200 students and then the crunch came in the last part of the 70s and early 80s it just dropped back to 30 or 40 students. Right now we let in anybody who can walk or crawl. Well probably graduate 6 or 7 this year(Personal interview, September 14,1993). Although the number of vocational graduates in construction grew in the 1970s,the construction labor force grew more rapidly. Thus, while the 1970s was the heyday of vocational training at Salt Lake Community College,vocational graduates as a percentage of the construction labor force had already begun to decline. 5 The next two subsections of this report are drawn from Hamid Azari,Anne Yeagle and Peter Philips,"The Effects of the Repeal of Utah's Prevailing Wage Law on the Construction Labor Market"in Sheldon Friedman,et at.,Restoring the Promise ofAmerican Labor Law,Cornell University,ILR Press, 1994,pp. 207-22. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 8 Stevens argues that the decline in enrollment is driven by a lack of demand for training. "The staff is hanging on by their fingernails because there's not enough students that want to get into the program." He adds that many of his graduates do not stay in construction: "They get out,and they find out that they don't want to work in the cold in the winter and they want to get into something more secure, something that's got benefits." n cold but the loss of benefits and security has occurred Construction has always required workers to work t the ty since 1980 along with stagnating employment,declining rate of unionization,and falling relative wages. Stevens also argues that the decline in construction training is a function of shifting government priorities in education: I can't speak for the school—I don't want to get in trouble that way—but we do have more general ed classes feeding the University[than in the past]and the vocational programs are really suffering. You look at Weber State. They used to have vocational programs and with academic drift, they don't have any vocational programs anymore. Tom Lewis, director of the plumber and pipe fitters'union apprenticeship program, agrees with Stevens that there is an institutional tendency to move away from vocational education. We used to have a pretty good relationship with the community colleges. The reason we bought our own building and moved out here[away from Salt Lake Community College]is that you have an administrator of a vocational school and they don't want to remain the administrator of a vocational school. They move to an applied technology center,then they're a community college,and then they're a university. I started my apprenticeship at Weber Vocational Center which is now Weber State University. Vocational training seems to get set aside as this evolution happens and we eventually just get moved out the back door(Personal interview,Sept3,1993)_ The steady decline in vocational training as a percentage of the construction labor force through good times and bad supports the notion that the state has simply tried to get out of the business of vocational training in construction. The fall in union membership and wages has made construction a less attractive career. At the same time,unions are less able to train construction workers. As unions are weakened and schools drift toward academic offerings, the capacity to respond smoothly to an upsurge in construction jobs is undercut. And federally sponsored Job Corps vocational training is not in a position to fill in the gap. Federal revenues pay for Job Corps training in Utah at both the Weber Basin and Clearfield centers. Federal funding in real terns for these centers has not expanded,but the Weber Basin Job Corps Center,which draws predominately from the Utah population, has significantly contracted its construction worker training throughout the 1980s. This center committed itself to changing from an all-male student population in 1980 to i for traditional) male occupations such as 50 percent female by 1990. To accommodate this switch,training y up construction have been scaled back to accommodate new offerings in traditionally female occupations,such as office management and clerical work. Cement masonry and heavy-equipment training have been eliminated, and instruction in carpentry,painting,and brick laying has been cut in half. The Clearfield Center has graduated approximately one hundred construction trainees per year since the early 1970s. Fewer Clearfield graduates go into the Utah labor market compared with Weber Basin because most of Clearfield's students are from out of state. On the whole,perhaps 10 percent of Clearfield's graduates go into the Utah labor market,but this percentage rises during periods of local labor shortage. It is estimated,however, that at most only 25 percent of Clearfield's graduates will stay in Utah (Hill, Merle. Vocational Education Manager,Clearfield Job Corps,personal interview,Sept.9, 1993). Even without union pressure, a skill shortage in Utah construction may raise wages and induce a new generation of young people to enter vocational training. When that happens, however, high-quality training h t time to create may not be in lace to meet that demand,which will add an additional lag which take y p programs, to the natural time it takes to train a skilled laborer. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 9 Why the Nonunion Sector Did Not Make Up for the Loss of Union Training The market has not successfully made up for the decline in union and state-sponsored training. At the national level,the nonunion Association of Building Contractors(ABC)has attempted to replicate the union system of bargaining for hourly contributions to a training fund. It is difficult, however, to induce ABC's member contractors to include general training costs in their bids. Each contractor fears his competitors will not include training costs. Thus,in an attempt to be low-cost bidder,ABC contractors often refrain from including training costs despite the ABC's initiative. Consequently,very little ABC training has occurred in Utah. Nonunion apprenticeship programs operate,however, in the licensed trades of electricians and plumbers. In 1992, there were 846 nonunion licensed apprentice electricians in Utah and 2,068 nonunion journeymen. Thus, there are 4 apprentices for every 10 journeymen in the nonunion sector. In contrast, there were 123 apprentices and 607 journeymen in the union sector in 1992 or 2 apprentices for every 10 journeymen. In the nonunion sector, apprentices begin at around$6 per hour with no benefits. Over a four-year period,the state mandates that their wage rise to 80 percent of a journeyman's pay. In the union sector,apprentices begin at$7 per hour with an additional$3 in benefits. Their wages rise to$14 per hour plus$3 in benefits over a five-year period(Leroy,Julie.Assistant Business Manager,IBEW Local 354,Sept.25, 1993). Nonunion apprentices are sponsored by a particular contractor that oversees their on-the-job training, and these apprentices also take classwork at a participating community college. Union apprentices work under the direction of an apprenticeship coordinator, rotate among employers for on-the-job training, and take classes at community colleges and union apprenticeship centers. Roughly 90 to 95 percent of the union apprentices complete their programs and graduate to journeymen status,while only 15 to 20 percent of the nonunion apprentices graduate (Dean,Frank.Home Builders Institute Electrical Instructor,Clearfield Job Corps, Sept. 9, 1993). Given these rates,in four years,out of 846 nonunion apprentices,we should expect 125 to 170 journeymen to be graduated. In five years in the union sector,out of 123 apprentices, 110 to 115 apprentices would graduate to journeymen electrician. Thus,while the nonunion sector accounts for more than 85 percent of all electrician apprentices,it accounts for about 60 percent of journeymen graduates. Economic theory is consistent with this pattern wherein nonunion apprentices are paid less and graduate at a lower rate than union apprentices. Economic theory posits that in the absence of marketwide institutions or government subsidies, individual workers will have to pay for their own on-the-job training when the skills learned are general to an industry and not specific and unique to the activities of a particular firm. The worker-learner pays for training by accepting a wage that is lower than the value to the firm of that worker's marginal product. By working for less than what the worker is worth to the employer, the worker pays the employer for on-the-job training. That beginning nonunion electrical apprentices earn$6 per hour while union apprentices earn$10 per hour(including benefits) is consistent with the theoretical proposition that nonunion apprentices pay for their own training by taking a discounted wage below their marginal value to the contractor. Because the employer does not pay for nonunion training,the theory suggests that the employer has no stake in the worker's training. Consequently, if the worker leaves, the employer does not lose any investment in the worker's human capital. Thus, the employer will tolerate high levels of turnover. Because the worker is receiving less than what the worker can earn in other jobs with no on-the-job training, the worker may be tempted to exit jobs with training when current personal budget needs become pressing. So, on both the employer side and the worker side,turnover is tolerated in the nonunion sector. This is consistent with the higher turnover rates among nonunion apprentices, but other factors also contribute to the 15 to 95 percent differential in nonunion to union graduation rates. Because the nonunion employer prices new hands at discounted wages that shield the employer from investing in the human capital of the new workers, the employer does not screen new workers extensively to forestall subsequent turnover. Failure to preselect new workers for aptitudes and attitudes consistent with a long-term attachment to construction work adds to the turnover among nonunion construction apprentices. In contrast, the joint apprenticeship boards of unions and union contractors do considerable preselection for both aptitude and attitude before letting a candidate into an apprenticeship program. This is because both the union Peter Philips,Ph.D.,Professor Economics Department,University of Utah 10 contractors and unions will invest in the union apprentices' training. Not wanting to lose their up-front investment,they seek to eliminate exit once the apprenticeship is begun. In the nonunion sector, workers may also leave apprenticeships if it becomes apparent that the employer offering training at a discounted wage is not delivering on that training promise to train. Because employers are able to discount wages of apprentices below their current worth to the employer,it is tempting to engage in bait-and-switch tactics whereby training is promised but not delivered. By saving on training costs, the employer can earn an additional profit from employing green hands at discounted wages. In the union sector, because employers and union journeymen invest in the training of the apprentices,bait-and-switch tactics are less attractive. Because the apprentices'wage is not discounted as much below what they could earn elsewhere, the apprentices are not as tempted to leave. Thus, economic theory predicts the observed pattern whereby the nonunion sector must begin training five apprentices to graduate one journeyman while the ratio in the union sector is close to one to one. The General Pattern of Training in Construction In basic terms,nonunion contractors have difficulty training because one,the relationship between the contractor and the construction worker is often brief. This leads to a free-rider problem. Why should 1 train you when you are likely to go down the road and work for my competitor. I would just be helping him out and not myself. Two,without an apprenticeship coordinator there is no one policing the training to insure that on-the-job training takes place and is of decent quality. Thus,some contractors are tempted into bait- and-switch strategies where they promise an apprenticeship experience but deliver only the job experiences of a helper. The apprentice spends his or her time holding the dumb end of a measuring tape. The result of these two market failures,the free-rider and bait-and-switch phenomena lead to lower numbers of apprentices in the nonunion sector and lower rates of graduation into journeyman status. Data from the U.S. Bureau of Apprenticeship Training illustrate these effects. Table 2 shows data for the entering class of construction apprentices in 1989. The data exclude California, New York,Delaware and Hawaii. These states currently do not report into the U.S.Bureau of Apprenticeship Training data base. For the remaining states,basically three out of every four construction apprentices enrolled into a jointly managed union-contractor apprenticeship programs despite the fact that union contractors account for less than half of all construction workers. Peter Philips,Ph.D.,Professor Economics Department,University of Utah I 1 Table 2:Apprentices Enrolled in 1989 by Union and Nonunion Sectors and Craft Apprenticeship • • The Class of ' : ' 7,245 58.4% 41.6% 4,752 88.8% 11.2% 4,467 65.6% 34.4% 1,856 74.0% 26.0% 1,234 97.4% 2.6% 1,162 88.0% 12.1% 972 90.6% 9.4% 830 90.0% 10.0% 729 87.9% 12.1% 3,675 78.5% 21.5% All.trades 26 922 T'4.E°0 5 2°14 Table 3 shows that the graduation rates are very different in the union and nonunion sectors of construction. By 1995,six years after enrollment,55.8%of the apprentices in joint union-management programs had turned out as journeymen. In contrast,among the smaller number of nonunion apprentices only 28.7% turned out as journeymen by 1995. The lower number of apprentices in the nonunion sector despite the fact that the nonunion sector employs more workers is an example of the problem of free-rider market failures. The lower graduation rate among the nonunion apprentices reflects the problem of bait-and-switch. Together these problems feed on each other. Table 3:Graduation Rates for the Entering Class of 1989 by Sector and Craft Apprenticeship in Construction The Class of 69.0% 29.5% 44.8% 29.5% 61.5% 26.0% 67.4% 43.6% 56.2% 9.4% 29.2% 17.9% 39.9% 18.7% 49.9% 48.2% 53.8% 6.8% 57.3% 25.7% Atl trades, 5's_.8•�0. .: �$°T°1a Peter Philips,Ph.D.,Professor Economics Department,University of Utah 12 Table 4 shows the combined effects of free rider and bait-and-switch problems. The higher ratio of apprentices to journeymen in the union sector combines with the higher graduation rate in the union sector resulting in union contractors accounting for 85%of all the new journeymen coming from the entering class of 1989. Table 4:The Share of All New Journeymen Coming from the Entering Class of 1989 Apprenticeship • • The Class of ' : ' 76.7% 23.3% 92.3% 7.7% 81.8% 18.2% 81.5% 18.5% 99.6% 0.4% 92.3% 7.7% 95.4% 4.6% 90.3% 9.7% 98.3% 1.7% 89 1% 10 9% Not all good construction workers learned their trade through formal apprenticeship. Not all union construction workers come from formal apprenticeship. Not all union apprenticeship graduates stay within the union sector of construction. However,the fact that almost nine out of every ten journeymen coming out of formal,monitored,planned and supervised apprenticeships come out of union apprenticeships means that union construction contractors are more likely to have a core of well trained craft workers. This allows these contractors to confidently invest in larger or more specialized machinery that require greater skill. This allows those contractors to employ capital intensive and human capital intensive technologies where wage rates can be higher without necessarily raising unit costs. VI. An Analysis of Square Foot Public Construction Costs in Nine Southwestern and Intermountain States The F.W.Dodge Corporation collects data on the accepted bid price or start cost of construction projects across all states and in both the public and private sector. Dodge does this as an information service to contractors bidding on construction jobs. From these data,I have selected five types of new structures built for non-federal public entities for analysis. These structures are elementary schools,middle schools,high schools,offices and warehouses. I limited my analysis to new construction in order to bring my Peter Philips,Ph.D.,Professor Economics Department,University of Utah 13 comparisons as closely as possible to similar construction projects.6 Renovations have more widely varying characteristics. (For instance a new boiler put into an elementary school will have a much higher square foot cost than would the installation of a portable class room.) I looked at construction that began sometime between July 1, 1992 to July 1, 1994. I translated all accepted bid prices into 1994 dollars using the consumer price index. I selected construction in nine western states. These states include five states with state prevailing wage laws regulating wage rates on public construction and four states free from these regulations on state and local public construction. The states selected were Utah which does not have a state prevailing wage law and New Mexico which does have such a law. Plus I selected all surrounding states that either border New Mexico or border Utah or both. Based on this criterion,the five selected states with prevailing wage laws are 1)New Mexico,2)Texas,3)Oklahoma 7,4)Wyoming and Nevada. The four states in which contractors were free from state prevailing wage regulations in my selected group are Arizona,Utah, Colorado and Idaho. With the exception of Nevada where the two main cities,Las Vegas and Reno tend to have cost-of-living characteristics similar to California,the states in this selected group tend to have similar cost-of-living characteristics_ Table 5: The Distribution of Construction by Type e within the Selected Nine States • t 43 10.' VV REFtOUSES 20 5.1 ° s ELEMENTARY SCH04LS 175 44.340" _ MIDDLE SCHOOL 104 26.3Q° rllt SCH007 LS, 53 13 t- In the nine selected states,during the years 1992 to 1994,elementary schools accounted for the largest single type of new non-federal public construction. Table 5 shows the distribution of construction projects across the five selected types. Elementary schools are the most common while warehouses are the Ieast common. In order to make statistically meaningful comparisons between the average square foot public construction costs in states with and without state prevailing wage laws it is necessary to have a sufficient number observations. Thus,results for elementary schools are most reliable and results for warehouses are least reliable. 61 excluded from my sample any project that reported square foot costs less than$20 or over$500. In unreported statistical analyses,I included these outliers. The conclusions of this report are not substantially altered by the exclusion of these cases. 7 In October, 1995,Oklahoma's state supreme court declared Oklahoma's law unconstitutional. It is no longer in effect. However,during the period under study,Oklahoma had a prevailing wage law regulating state public works. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 14 Table 6 shows the breakdown of construction sites by type of structure grouped by states with and without state prevailing wage laws. The distribution of structure types is basically the same in the five states with prevailing wage laws compared to the four states without such laws. Elementary schools are most common and public warehouses are least common. More construction in the selected structure types took place in the five states with prevailing wage laws compared to the four states without such laws. In general,there are sufficient numbers of sites by type in each group of states to make meaningful comparisons with the possible exception of warehouses where there were only 12 built in states with prevailing wage laws and 8 built in states without prevailing wage laws. In contrast, 116 elementary schools were built in states with prevailing wage laws and 59 in states without prevailing wage laws. This is clearly enough observations to overcome the random effects of special aspects on any one particular job site that might raise or lower square foot construction costs. Table 6: Distribution of New Construction by Type and Prevailing Wage Law Status 01FICES 23 8. % 14.60%° WAREHOUSES 12 4:709/0 8 5.80% ELEMENTARY SCHOOLS 116 4540% 59'' 43:10%. !DOLE SCHOOLS 76 201 20.4a1 fGH SCF © .LS` More new public construction took place in the five states with state prevailing wage laws compared to the four states without prevailing wages laws during the period under study. However, the distribution of types of structures was similar in the two groups of states. Table 7 shows Texas(a state with a prevailing wage law)accounts for the most construction projects of any of the nine selected states. Arizona(a state without a prevailing wage law)is second while Wyoming provides the least number of construction sites to analyze. New Mexico built 24 new public structures of the types under analysis during the time period under analysis. Of these 24, 14 were elementary schools. This is a sufficient number to make comparisons using New Mexico alone compared to other states. (These structures are listed in Appendix A.) Peter Philips,Ph.D.,Professor Economics Department,University of Utah 15 Table 7: Distribution of New Construction by Type and State Numberof New State and Local Construction Proiectsbv State 01` 11= 64 5 4 2 1 13 5 3 WAREHOUSES 1 6 1 0 0 0 12 0 0 ELEMENTARY 24 14 8 14 10 12 79 13 1 MIDDLE SCHOOLS 9 6 7 4 5 8 58 6 1 'HtGH SCHOOLS 10 2 5 2 5 3 21 5 0 50 1 2 .2 _ 2�_ 224 .� 9 Texas accounts for the largest amount of new construction projects. Arizona is a distant second. Wyoming accounts for the fewest projects. Table 8 presents data on average square foot cost of construction in 1994 dollars broken down by structure type for the selected four states that do not apply prevailing wage regulations to state and local public works. In addition to average(or mean)square foot costs,Table 4 shows the number of new structures built,the minimum square foot cost on the least costly of these structures and the maximum square foot cost on the most costly of these structures. In each state,elementary schools were the most common structure built. Average square foot costs ranged from a low of$60 per square foot for 8 elementary schools in Idaho to$80 per square foot for 14 elementary schools in Colorado. In terms of individual schools,one elementary school in Arizona and one in Idaho were built for$46 per square foot while at the other extreme,one elementary school in Arizona was built at the cost of$155 per square foot. These cost variations in elementary school construction may have to do with urban versus rural settings,differences in installed equipment,the size of the school or other factors. In comparing square foot construction costs across states with and without prevailing wage laws, my analysis assumes that with increased sample size the effects of these special factors driving costs up or down will offset each other in state averages. This is more likely to be true as the sample size rises. In Table 9 the same square foot construction costs are presented for states with prevailing wage laws. Again,elementary schools are the most commonly found type of new public structure in each state. The average square foot cost ranges from a low of$49 per square foot for 8 elementary schools in Oklahoma to a high of$96 per square foot for 10 elementary schools in Nevada. The 14 elementary schools in New Mexico averaged$66 per square foot with the cheapest coming in at$35 per square foot and the most expensive at$87 8 Oklahoma had the single cheapest elementary school built at$27 per square foot while Texas had one elementary school cost$368 per square foot. The difference in these very cheap and very expensive projects are probably due primarily to distinct differences in specifications of the buildings. However,when the expensive Texas school is averaged in with the 78 other elementary schools built in that state,its effect on the statewide average is minimal. s The cheapest New Mexico elementary school project involved portable metal classroom buildings and the most expensive was the building of Chaparral Elementary School in Santa Fe. In comparing construction costs,one tries to compare apples to apples. Elementary schools as a building type are relatively homogeneous. However,as the example of portable classrooms suggests,even within the category of elementary schools/new construction there can be considerable differences in building design driving square foot cost differences. Similarly,it may be that Santa Fe is a more expensive region within New Mexico. This is why more observations lead to better comparisons of averages as the atypical project or the atypical area gets swamped by a larger number of more typical projects and more representative areas within a state. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 16 Table 8: Square Foot Construction Costs in States without Prevailing Wage Laws Arizona Count Mean Minimum Maximum OFFICES 6 $100° $64.. $i79 WAREHOUSES 1 $65 $65 $65 ELEMENTARY 24 $72 $46 $155> MIDDLE SCHOOLS 9 $77 $53 $100 HIGH.SCH000LS 10 $66- $69: $104 Colorado Count Mean Minimum Maximum OFFICES 4 $108 $62 $168- WAREHOUSES 6 $106 $21 $297 ELEMENTARY 14 $80 $61 $95 MIDDLE SCHOOLS 6 $,79 $41 11330 HIGH SC OOLS 2 "$' -7 Idaho Count Mean Minimum Maximum OFFICES $96 $44 $174 ELI=IVIEN7 �RY 8 $I �:. $7 : M1DaLE:SCH€ tLS 7 $ t $78 HIGfSCIdCLS 5 $ $ .., $1# Utah Count Mean Minimum Maximum OFFICES 5 $71 $34" 11=21 - ELEMENTARY 13 S72 $ a MIDDLE SCHOOLS 6 $94 $59 $2I5""" HIGH Scl oLs .. 6 rt. Average square foot cost of construction by structure type in four states without state prevailing wages laws by structure type. Also lowest and highest cost projects for each structure type by state. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 17 Table 9: Square Foot Construction Costs in States with Prevailing Wage Laws New Mexico Count Mean Minimum Maximum ©FF+GES � $s WAREHOUSES - - ELEMENTARY SCHOOLS 14r. $66 $35 $87 ' MIDDLE SCHOOLS 4- $71, $58 $85`, Hi'GH SCHOOLS $117 Nevada Count Mean Minimum Maximu 9 - 3 $177 WARFI OUSES - -, ELE ±lE4 ARY"S- cHpo- emu; 'ICf $ 6. . $ q $1 6' ?A OLE S}C-HOOLS 5 r $75 $141 Oklahoma Count Mean Minimum Maximum . - t?FEfCE 1 119 $ 1 " MID"1 L!<51 11f7f7LS $ $- $71 H1Gfl SCHOOLS $ 9 Texas Count Mean Minimum Maximu fJEF1GES =1 WAF2EI1�t15ES ELE1V ENTARY SCkit? L . Wyoming Count Mean Minimum Maxi f, _ 77 ¥ �+ 3 E E t1El �lkR'Y"SCH€30LS � M 10bLE to- 1... $39 $39 H1H 0CH0OLS _ Average square foot cost of construction by structure type in five states with state prevailing wages laws by structure type. Also lowest and highest cost projects for each structure type by state. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 18 Table 10: Comparing Average Square Foot Cost of New Public Construction by Type and Prevailing Wage Law Status OFFICES 23 95 20 $93 WAREHOUSES 12 $61 $9 LEM`ENTAR'Y SCHOOLS LL t -, O 5� 11aDLE SCFirULSa 6 , ' 77 HJH SGH3LS 31 = $7( e Square foot construction costs on public projects are not higher in states with prevailing wage laws compared to states without prevailing wage laws. Table 10 presents the key results from the F.W.Dodge data. Let us focus on elementary schools first because this is the largest group. Average square foot new construction costs are$67 in the five Intermountain and southwestern states with prevailing wage laws and$73 per square foot in the four Intermountain and southwestern states without prevailing wage laws. The sample sizes are large and statistical tests show that this difference of$6 per square foot is real.9 Referring back to Tables 4 and 5 one can compare New Mexico to the group of states without prevailing wage laws. The average square foot cost for building 14 elementary schools in New Mexico was$66 compared to an average of $73 for 59 schools in the group of states without prevailing wage laws. Comparing New Mexico with Colorado,Utah and Arizona,the three non-prevailing wage law states bordering New Mexico yields similar results. In Utah, 13 elementary schools were built at an average square foot cost of$72. In Arizona,24 elementary schools were newly built at an average square foot cost of $72,and in Colorado, 14 elementary schools were built at a cost of$80 per square foot. These compare with New Mexico's 14 schools built at a cost of $67 per square foot. Of the four states without prevailing wage laws,only Idaho had a lower average than New Mexico. Idaho built 8 elementary schools at a cost of$60 per square foot. These data do not support the proposition that eliminating prevailing wage laws are likely to lower public elementary school construction costs in a measurable way. Because the sample size for elementary schools is the largest,this is the single most reliable conclusion from the Dodge data. Table 10 also shows that the square foot construction cost of middle schools and high schools are lower in the five states with prevailing wage laws compared to the four states without these laws. The 76 middle schools built in prevailing wage law states cost an average of$66 per square foot while the 28 middle schools built in the four states without prevailing wage laws cost an average of$77 per square foot.In the case of high schools,the contrast is similar. In states with prevailing wage laws,square foot costs for 31 schools were$70 while in states without prevailing wage laws it was$81 for 22 schools. However,this last result disappears when Texas is removed from the group of states that have prevailing wage laws. The sample of high schools in the remaining four states with prevailing wage laws falls from 31 to 10 and the square foot costs rises to$86. Removing Texas lowers the sample sizes for elementary schools in the remaining prevailing wage law states from 116 to 37 and middle schools from 76 to 18. However,the basic results in these two cases are unaltered. Square foot construction costs are still,on average,lower in the remaining four states with prevailing wage laws compared to the four states without prevailing wage laws. 9 A two-tailed t-test shows that these are statistically different numbers at the 5%level of significance. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 19 The sample sizes for middle schools and high schools in New Mexico are low. New Mexico built only four new middle schools and two new high schools during the period under study. The four New Mexican middle schools cost on average$71 per square foot. This compares with square foot costs for middle schools in the four states without prevailing wage laws of $77 in Arizona(9 schools),$79 in Colorado(6 schools),$62 in Idaho(7 schools)and$91 in Utah(6 schools). The two New Mexican high schools cost$105 per square foot(one at$92 and the other at$117). This compares with square foot costs of high schools in the four states without prevailing wage laws of$86 in schools)and 65 in Utah 5 schools). Arizona(10 schools),$107 in Colorado(2 schools),$77 in Idaho(sch ) $ ( ) Thus,there is some suggestion that in contrast to elementary schools and middle schools,high school construction in New Mexico might cost more than in surrounding states with state prevailing wage laws. But this conclusion would be premature. The New Mexico sample is small(two schools)and is in line with the more expensive schools built in each of the states without prevailing wage laws. The high-end costs were$104 per square foot in Arizona,$118 in Colorado,$106 in Idaho and$86 in Utah. The small New Mexican sample may simply only include two more richly designed high schools. Basically,as sample sizes fall,conclusions drawn from comparing averages must become more tentative. The cost of office buildings(Table 6)were basically the same in states with and without prevailing wage laws. The cost of constructing 23 new state office buildings in states with prevailing wage laws was$95 per square foot while the 20 new office buildings in states without prevailing wage laws cost$93 per square foot. When Texas is eliminated from the sample of five states with prevailing wage laws,the sample size falls to 10 and the average square foot cost rises to$101. However,from a statistical standpoint,there is no significant difference between the averages of$101 and$93.10 New Mexico built 4 public office buildings during this period with a square foot cost of$76. Contrasting this$76 square foot cost to the four states without state prevailing wage laws—Arizona built 6 office buildings at an average cost of$100 per square foot;Colorado built 4 public office buildings as an average square foot cost of$108;Idaho built 5 public office buildings at a square foot cost of$96 and Utah built 5 public office buildings at a square foot cost of$71. Texas was the only prevailing wage law state to build public warehouses during the time period under study. Texas built 12 warehouses at an average square foot cost of$61 while in the group of states without prevailing wage laws,Arizona,Colorado and Idaho together built 8 warehouses at a square foot cost of$96. (Utah did not build a public warehouse during this period.) In summary,start cost data from the F.W.Dodge Corporation do not support the notion that square foot construction costs on state and local construction will be cheaper in the absence of state prevailing wage laws. In the largest and perhaps the most homogeneous sample of public construction,elementary schools, if anything,square foot construction costs are cheaper in Intermountain and southwestern states with prevailing wage laws compared to those without these laws. This conclusion holds when comparing New Mexico to the surrounding states without prevailing wage laws. Even though sample sizes decrease,the general conclusion that construction costs for middle schools and high schools are,if anything,lower in states with prevailing wage laws compared to those without prevailing wage laws continues to hold. This is also true for New Mexico compared to surrounding states without prevailing wage laws in the case of middle schools. However,in the case of two high schools built in New Mexico,their costs were on the high range of square foot costs in surrounding states with no state prevailing wage laws. This may simply be due to the specific design of these two schools. Sample sizes for office buildings and warehouses are smaller than those for schools,particularly if Texas is excluded from the group of states with prevailing wage laws. In these small sample cases,one cannot conclude that construction costs are any different in Intermountain and southwestern states with or without prevailing wage laws. 10 A t-test indicates that these means are not significantly different at a 5%level. What this means is that given the small sample sizes and the wide variation of square foot cost within each sample,one cannot conclude that the average differences are the result of anything other than randomness. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 20 Table 11: Excluding Texas and Comparing Average Square Foot Cost of New Public Construction by Type and Prevailing Wage Law Status MP ♦ • • " OFFICES 10 01 20 $ V1IAREHOUSES - 8 $96` ELEMKENTAFZY SCi=f�E3LS 37 5-014 1{- DLE SCI-OO 1$ 8 $ 7 T HIGH SCHOOLS 10 22 $ t Excluding Texas from the group of states with prevailing wage laws does not alter the general conclusion that public construction costs are not higher in states with prevailing wage laws compared to states without prevailing wage laws. VII. Conclusions There are no previous published studies that analyze the relationship between state prevailing wage laws and public school construction costs. Steven Allen's study of union and nonunion contractors in the early 1970s found that while union workers were 20%to 50%more productive than nonunion workers,nonunion contractors were the low-cost contractors in public school construction. Allen's study was limited to 57 union-built and 11 nonunion-built schools. Consequently,he had to collapse together elementary and secondary school buildings and include various regions of the country. The U.S.Bureau of Labor Statistics found that in the early 1970s labor costs as a percent of total costs in school construction did not vary widely by region despite wage rate variations of 50%. The BLS attributed this to differences in regional labor productivity and construction material costs. Azari,Yeagle and Philips show that in the case of Utah,the repeal of the state prevailing wage law in 1981 corresponded to a rapid decline in apprenticeship training in that state,a decline that was not compensated by any increase in job corps or community college training. Bureau of Apprenticeship Training data for the 1990s show that 85%of all apprenticeship-trained construction journeymen come out of jointly sponsored union-management apprenticeship programs. The reasons that nonunion construction contractors are less likely to train are associated with market failures tied to the problems of free-riding contractors waiting for others to train and bait-and-switch contractors offering training but only providing helper experiences on- the-job. Collectively bargained contracts calling for jointly managed apprenticeship programs provides the policing mechanism to overcome these market failures. Thus,regulations that discourage collective bargaining in construction also discourage formal apprenticeship training in construction. The resulting Iower productivity helps account for the fact that lower wage rates in construction do not necessarily lead to lower construction costs. In the case of the nine southwestern and Intermountain states selected for this cost study,Table 10 shows the basic result. The average square foot construction costs for 116 elementary schools built in five states with prevailing wage laws was$67 while for 59 elementary schools built in four states without prevailing wage laws the cost was$73 per square foot. For 76 middle schools built in the states with prevailing wage laws,the average square foot cost was$66 while in the states without prevailing wage laws the cost was$77 per square foot. For 31 high schools built in the five states with prevailing wage laws,the square foot cost Peter Philips,R.D.,Professor Economics Department,University of Utah 21 was$70 while in the four states without state prevailing wage laws the cost was$91. The difference in all of these averages was statistically significant. When New Mexico's 14 newly built elementary schools was compared to elementary schools in the four surrounding states without prevailing wage laws,the average New Mexico square foot cost of$66 was lower than the square foot elementary school cost in Arizona($72),Colorado($80)and Utah($72)but higher than in Idaho($60). During the time period selected for the study(fiscal years 1992-94)New Mexico built on 4 new middle schools and 2 new high schools. Thus,the average costs for these schools is more sensitive to the effect of small numbers. Nonetheless,the average New Mexico square foot cost for middle schools($71)was lower than that for Arizona($77),Colorado($79)and Utah($91)while higher than Idaho($62). In the case of two high schools in New Mexico,the square foot construction cost was $105. This compare with$96 for 10 high schools in Arizona,$107 for 2 high schools in Colorado,$77 for 5 high schools in Idaho and$65 for 5 high schools in Utah. The smaller number of high schools built compared to middle schools and elementary schools make this comparison more sensitive to the effects of small numbers on averages. Similar small number problems exist for offices and warehouses. During the time period under study,New Mexico built four public office buildings and no warehouses. The average square foot cost for public office buildings in New Mexico was$76. This compares to an average of$93 for 20 public office buildings built in the four surrounding states without prevailing wage laws. The basic conclusion of this study is that in the case of New Mexico there is no strong evidence to suggest that the repeal of the state's prevailing wage law would save substantial costs in the construction of public schools. This is especially true in the long run. The reason higher wage rates for construction workers do not necessarily lead to higher construction costs is because those higher wage rate appear to be offset by higher labor productivity. In the short run,lower wage rates might not lead to lower productivity simply because trained construction workers might be forced to accept those lower wage rates. However,in the long run,a migration of trained workers out of construction and a decline in the training of new construction workers would lead to lower productivity canceling out any savings from lower wage rates. Peter Philips,Ph.D.,Professor Economics Department,University of Utah 22 Appendix A; New Mexico Data Used in Study City Project Description ttructure: Project Square Sq.Ft. Sq.Ft. Type Value Feet Cost Cast I"", ANITA FE SCHOOL ADMINISTRATION OFFICE(ADD/REMOD/DEMOLITION) Office $1,290,600 "141000 $92 $46 ALBUQUERQUE OFFICE BUILDING Office $4,000,000 65,000 $62 $641. LBUQUERQUE STATE BAR CENTER PROJECT(OFFICE BUILDING) Office $1,800,000 28,000 $64 $66° SOCORRO EMRTC COMPLEX(OFFICE/LABORATORY/SHOP)(3 BLDGS)(PRE ENGR) Office $3,405,596 431000 $79 $79 ALBUQUERQUE TOMASITA ELEMENTARY SCHOOL Elemental $2,899,900 4%000 $60 $64 SANTA FE CHAPARRAL ELEMENTARY SCHOOL 12 Elerrlentpryl $2,117,922 26;000 $81 $8,7 PORTALES PORTALES ELEMENTARY SCHOOL(9212) Elementary $3,698,921 54,000 $63 $67 ZUNI A:SHIWI ELEMENTARY SCHOOL(PH Elementary $1,177,080 17,450 $67 $72 ALBUQUERQUE' PAJARITO ELEMENTARY SCHOOL CORE FACILITY Elementary $1,287,426 21,000 $61 $65 TOME TOME ELEMENTARY SCHOOL(PH 2)( Eternerrtary $800,000 11,000 $73 $77; ARTESIA YESO ELEMENTARY SCHOOL(24 CLASSROOM)(944) Elerhontao. $3,924,500 5.1,500 $76 $76• LOS LUNAS " LOS LUNAS ELEMENTARY SCHOOL E1el0etit $2,311,000 '42,500 $54 $56 RUIDOSO RUIDOSO ELEMENTARY SCHOOL Sefnehfary_ $1,879,700 :32,000 $59 $61 ALBUQUERQUE 1993 PORTABLE METAL CLASSROOM BLDGS(120 UNITS)(93112) Elementary: $3,611,080 lLf8;000 $33 $35 LOVING LOVING ELEMENTARY SCHOOL(NEW)(PH 1)(A9315) E"ritafy" $2,000,000 UAW $61 $61`i LAS CRUCES ELEMENTARY SCHOOL(139325) Elementary` $3,257,687 ''53;768 $61 $61" MORIARTY MORIARTY ELEMENTARY SCHOOL(PH 2)(23 CLSRMS) Elementary; $1,741,351 27;000 $64 $67 GALLUP EAST ELEMENTARY SCHOOL(PH 1) E1emeYitary $2,410,000 29;700 $81 $81 SANTA FE NEW MID SCHOOL Middle $7,696,771 94,000 $82 $85' BELEN BELEN JUNIOR HIGH SCHOOL(PH 2 Middle $1,504,850 27,116 $56 $59 FARMINGTON FARMINGTON JR HIGH SCHOOL Middle $5,655,423 100,000 $57 $58' TEC KOOGLER MIDDLE SCHOOL(8 CLASS Middle $829,616 10,940 $76 $81 LAS CRUCES ONATE HIGH SCHOOL(1391.04) High $19,768,874 380;000 $110 $117 ALBUQUERQUE,LA CUEVA HIGH SCHOOL PH 3 Hr.h $2 606 500 -_30 000. $87 2m Peter Philips,Ph.D.,Professor Economics Department,University of Utah 23 The Effect of State Prevailing Wage Laws on Total Construction Costs by Mark J. Prus Department of Economics SUNY, Cortland Cortland,NY 13045 607-753-5758 PRUS@SNYCORV A.CORTLAND.EDU C copyright Mark J. Prus January 1996 Introduction c policy in the United States at both the been a art of public o Prevailingwage laws havep p p Y g federal and state levels since the early part of this century. These laws require that construction e aid the "prevailing rate"when working on publicly funded construction projects. workers b g P P They were initially established to neutralize the government's monopsonistic power as a purchaser of certain types of construction labor and to support the social objective of maintaining a family wage. Currently,the question of whether prevailing wage laws continue to make public policy sense is being debated. One of the most significant issues in this public policy debate is what impact prevailing wage laws have on publicly funded construction costs. Opponents of prevailing wage laws argue that they raise construction costs. Repeal of these laws would result in cost savings on the order of 20 to 30 percent according to some. If such savings were possible, it is argued, school districts could build five schools for the price of four. Claims of cost savings from the repeal of prevailing wage laws,however, are generally based on analysis of the effect of higher wage rates on construction costs.Yet wage differences have a moderate effect on total construction costs. Labor costs are less than a third of total construction costs and have been falling. In 1972, for instance, in an analysis of school construction costs,John Olsen found that onsite wages and salaries excluding benefits were 28.2 percent of total costs. (Monthly Labor Review, 1979,p.40) According to the Census of Construction, labor costs counting benefits on all types of construction were 30 percent of total costs in 1977 and had fallen to 26 percent by 1987. A second problem leads us to question estimates of the impact of prevailing wage 2 legislation on construction costs based on an analysis of wage differences. Because they assume implicitly that the same number of hours of each type of labor will continue to be employed and that labor is of invariant productivity the impact on costs is driven by the wage differential. Neither of these assumptions are necessarily appropriate. The payment of prevailing wages may serve to attract workers with more experience and training. Increased labor productivity may result in fewer hours of labor being required thus offsetting the higher wage rate. For instance, Allen has shown that unionized labor in the construction industry is between 17 and 52 percent more productive than nonunion labor. (Allen, 1984) Additionally,higher wage rates may lead contractors to substitute capital or other inputs for labor,mitigating the impact of higher wages rates on total construction costs. These possibilities, alone or in combination,make the assumptions underlying the analysis of construction costs based on wage differences inappropriate and cast doubt on the estimates of costs savings. Specifically, it is difficult to imagine how savings of 20 to 30 percent are g .possible To et a true picture of the impact of prevailing wage legislation's impact on total P construction costs, one could evaluate not only differences in wage rates,but also productivity differences,the incidence of substitution, administrative costs and other ways in which these laws's impact is either mitigated or enhanced. An alternative approach is to simply examine total construction costs directly and compare costs in the presence and absence of prevailing wage laws controlling for project differences. 3 Previous Estimates of Impact on Total Costs Only one studies have attempted to estimate the impact of prevailing wage legislation based on actual total construction costs. Fraundorf, et. al., in"The Effect of the Davis-Bacon Act on Construction Costs in Rural Areas," examined 215 new,non-residential construction projects built in 1977 and 1978. (Fraundorf, 1983) Approximately half of these projects were federal projects built under the purview of the federal Davis Bacon Act specifying that prevailing wages be paid. The other half were privately owned projects constructed without the requirement that prevailing wages be paid. Data on total construction costs were then compared using multivariate regression analysis to control for the effects of factors other than the presence of prevailing wage requirements. This study controlled for differences in the type of structure,the types of materials used, and project size in an effort to focus on cost di fferences associated with labor cost differentials resulting from the dichotomy in regulatory regimes. It also attempted to control for regional differences in construction costs by grouping projects into four regions; Northeast,North Central, South and West. The dependent variable in their regression analysis was the natural log of the project's bid price deflated to 1977 dollars. The authors of the study found, somewhat surprisingly,that federal construction projects governed by Davis Bacon were 23 percent more expensive than private construction projects controlling for other cost influencing factors. When they re-estimate their basic model to correct for disproportionate response rates by region and building type, Fraundorf finds that the impact of Davis Bacon on total construction costs is as high as 30 percent. While they admit that these results are high, especially in light of a public-private wage differential estimated at 20 percent,they point out that 4 the results are consistent with other aspects of the data. In particular they do not find evidence of factor substitution which would mitigate the impact of prevailing wage requirements. However, this does not explain why the impact on total costs is greater than the wage differential. They explore other factors that might contribute to the higher costs of federal Davis Bacon projects such as record keeping and reporting,and decreased competition. Neither of these factors appear to significantly contribute to costs on federally funded construction. (Fraundorf,et. al, 1983,p. 145) One possible problem with the Fraundorf study is that regional differences in construction costs may have been inadequately controlled for. Given the relatively small sample size,the authors of this study had to group construction projects into relatively broad regions. This creates the potential for comparing a private project in a low cost state such as Idaho with a public project in a high cost state such as California. Since both projects are considered to be in the same region the cost differential is incorrectly attributed to the impact of prevailing wages when in fact it is due to differences in the cost of living or cost of materials between Idaho and California. Another problem may result from the way in which building types were classified. Each construction project was placed into one of six categories; recreational buildings,storage facilities,industrial buildings, office-commercial,medical and other. These categories were then used to find matches between public and private construction projects. However,these six categories were sufficiently broad to allow rather dissimilar buildings to be considered comparable. For instance, in the category storage facilities,warehouses were grouped with barns as well as airplane hangars. Likewise office buildings were in the same category as restaurants. Differences in costs between public and private buildings may have resulted from differences in 5 structure type and not from prevailing wage requirements. (Fraundorf, 1982) his stud is that it fails to problem with t 1 more serious rob y A second and potentially p adequately isolate the impact of prevailing wage legislation on construction costs. Specifically, Fraundorf compares private projects constructed in the absence of prevailing wage legislation with federal(i.e.,public)projects built using workers paid the prevailing wage. This comparison,while seemingly appropriate, contains the potential for confounding cost differences related to prevailing wage laws with cost differences resulting from other differences between private and public construction projects. The authors recognize this possibility when they point out,"If the government is more exacting than private owners in its quality standards, labor hours (and costs)and possibly material costs would be higher on government projects. (Fraundorf, 1983,p. 145) It may also be that the difference in bidding procedures for public and private contracts or differences in the time horizon of public and private owners may contribute to higher costs in the public sector. In other words,the cost differential that Fraundorf attributes to the effect of prevailing wage legislation may in fact be due to differences between public and private construction. Estimating the Impact of State Prevailing Wage Laws This study analyzes the impact of prevailing wage legislation on total construction costs using data on nonresidential construction in the United States. The data are from the F.W. j Dodge Company,an organization that collects and disseminates data on construction projects to the construction industry. These data give information on construction costs at the start of the project,or bid price. They also contain information on detailed structure type,project location, 6 project scale, and technical characteristics of the project such as number of stories and type of frame. The Dodge data also distinguish between public,private and federal projects. One advantage of the Dodge data is that they report on a large number of construction projects allowing for a more appropriate geographical breakdown of projects. In addition, the Dodge data make it possible to compare construction costs on similar projects in the private and public sectors for states both with prevailing wage laws and without. This is essential if one is to sort out the cost differences associated with public construction from the cost differences associated with prevailing wage laws. The Model The model used here to estimate the effect of prevailing wage legislation on construction costs is C=a + bIS+ b2T+ b3B + NA + b5ST+ b6Altadd+ b7Pubcode + e where C= start cost or bid cost;S= project scale as measured by square footage; T= a vector of dummy variables indicating detailed structure type;B= a vector of building material dummy variables;A= a vector of state dummy variables;ST= the number of stories;Altadd= a dummy variable indicating whether the project was an alteration or addition as opposed to new construction; and Pubcode=a dummy variable indicating that the project is public. This model is nearly identical to the one used by Fraundorf to estimate the impact of the federal Davis Bacon Act on total construction costs. It is appropriate for estimating the cost difference between public and private construction projects holding other factors such as building type,building materials,state in which the project is undertaken,building size and complexity constant. In 7 states with prevailing wage laws,the cost difference between public and private projects may be thought of as measuring the inflationary impact of the law. Data The most notable difference between the Fraundorf analysis and this study is the focus here on the impact of state prevailing wage laws, also known as little Davis Bacon acts, on costs. In the first instance,I selected construction projects in states that had prevailing wage laws in 1990. This focus on state prevailing wage laws's impact on costs, as opposed to Davis Bacon's impact, is useful for sorting out differences in costs associated with prevailing wage laws from cost differences between public and private buildings. Moreover, given that many states are currently considering reforms or repeal of their little Davis Bacon acts;the impact of state laws on construction costs is an important question to answer. The projects selected were nonresidential construction categorized as offices,hospitals,elementary schools,middle schools, secondary schools, garages and warehouses. These categories are similar to those used by Fraundorf, et al.,but are more specific and consequently less likely to result in comparisons between dissimilar structures. For example,hospitals is a more detailed class of structures than medical buildings, and warehouses is similarly more specific than storage facilities. Unlike Fraundorf, I included both new construction and additions/alterations. In states with prevailing wage laws additions/alterations to public structures and roads are also covered. The results of a multiple regression analysis using the natural log of real total project cost as the dependent variable indicate that,controlling for other relevant factors, in states that have in Table 1. Public projects in states having prevailing wage laws public buildings are reported p 8 prevailing wage laws are 27.6 percent more expensive than private structures. These results are very similar to the results obtained by Fraundorf. in terms of the magnitude of the estimated impact. Given that the prevailing wage laws in these states apply to public projects,this estimated cost differential can be attributed to the law's existence. This approach to estimating the impact of prevailing wage laws on construction costs, to distinguish like the earlier stud by Fraundorf is unableguish between cost differences due to Y ownership differences and cost differences that result from prevailing wage requirements. By comparing public projects built in states where prevailing wage laws are in effect with private projects,the impact of the law is confounded with cost differences between public and private projects. To illustrate the problem the model presented earlier was re-estimated using data on construction projects from states without prevailing wage laws. Similar controls were used to insure that public projects were being compared with similar private projects. The results of this regression are reported in the third column of Table 1. Once again,public projects are significantly more expensive(31.7 percent)than comparable private projects. But because the projects examined were located in states that currently do not have prevailing wage laws,the cost differential can no longer be attributed to the law's impact. This result lends support to the notion that the public may be a more exacting owner than the private sector. It also suggests that it is inappropriate to assume that the higher costs of public projects are attributable to the presence of prevailing wage laws. A more appropriate analysis recognizes that there are two different dimensions to construction cost differentials. On the one hand,comparisons of public projects versus private projects can determine the extent to which the government may be a more 9 exacting owner. The other dimension considers the presence or absence of prevailing wage legislation. Combining these two different dimensions creates four different possibilities;private projects built where no prevailing wage law exists,public projects with no prevailing wage law, private projects where prevailing wage laws exist, and public projects with prevailing wage laws. Only the last category of construction projects is directly affected by the presence of a prevailing wage law. 10 Table 1: Regression Results States w/Laws States w/o Laws Variable Coefficient Coefficient SCHOOL .586013** .769542** HGHSCHL .712087** .878536** HOSP 1.077311** 1.229640** WARE .016312 .425241** OFFICE .242328** .588873** PUBCODE .276200** .317366** STEELDUM -:045994 -.081977 WOODFRM .150933** .061925** CEMENTDM .091849** .118236** LNSQFEET .562699** .581263** STORIES .071699** .110672** ALTADD -.160042** -.066969** (Constant) 8.187471** 7.569623** Adjusted R2= .62534 Adjusted R2= .67164 N=5136 N=2717 F= 210.08747 F= 186.24560 Dependent Variable is ln(real total costs)where total costs are reported in 1994 dollars ** coefficient is significant at the .01 level the coefficients for the state dummy variables are not reported 11 Thus,in order to appropriately assess the impact of prevailing wage legislation on construction costs,this category must be isolated from each of the other possibilities. The model used here to estimate the effect of prevailing wage legislation on construction costs is C =a +b1S+ b2T+ b3B + b,49 + b5 ST+ b6Altadd+b,PW+ bBPubcode + b8Interact+ e where C= start cost or bid cost; S= project scale as measured by square footage; T= a vector of dummy variables indicating detailed structure type;B= a vector of building material dummy variables; ST=the number of stories;A = a vector of state dummy variables;Altadd= a dummy variable indicating whether the project was an alteration or addition;PW=a dummy variable indicating the presence or absence of a prevailing wage law;Pub=a dummy variable indicating ; Interact=(PW x Pubcode). The key components of this model are the variables PW,Pubcode and Interact. In combination these three variables allow us to estimate the effect of prevailing wage laws separately from the effect of public ownership. The Pubcode variable estimates the cost differential that exists between public and private projects,ceteris paribus,regardless of whether a prevailing wage law is in effect. The PW variable estimates the effect of prevailing wage laws on construction projects in states with laws regardless of whether they are public or private. Finally,the Interact variable picks up the direct effect of prevailing wage legislation on public projects since it is equal to one only in those cases where there is a public project in a state with prevailing wages. This model was estimated using the combined data for states with and without prevailing wage laws. The results are reported in Table 2. The coefficient on the interaction term(Interact) is positive but statistically insignificant indicating that the direct effect of prevailing wage laws on the cost of public construction projects is negligible. The presence of 12 a prevailing wage law also does not appear to have any significant effect on the costs of construction projects. Public projects in all states,however, are significantly more expensive (25.9 percent)than private projects as indicated by the coefficient on the variable Pubcode. Conclusions The results of this multivariate analysis of the impact of state prevailing wage laws on total construction costs indicates that, in contrast to earlier academic analyses as well as some casual statements,there is no measurable cost difference between similar structures as a result of prevailing wage requirements. Consequently,reforming or repealing these laws will not lead to the kinds of substantial savings promised by proponents of repeal. At the same time there are significant measurable cost differences between public and private projects of a similar nature. Researchers and politicians both should try to determine 1). why this differential exists and how and 2). what steps could be taken to lessen the difference. 13 Table 2. Regression Results: Determinants of Construction Costs for All States Variable Coefficient SCHOOL .628581** HGHSCHL .747973** HOSP 1.109748** WARE .132805** OFFICE .342607** PUBCODE .259601** LAW .180579 INTERACT .050779 STEELDUM -.053093 * 40 W ODFRM .118633** i O CEMENTDM .106521** LNSQFEET .569458** STORIES .081728** ALTADD -.128803** (Constant) 7.952916** Adjusted R Square = .63792 N=7854 F= 224.18097 Dependent Variable is In(real total costs)where total costs are reported in 1994 dollars **coefficient is significant at the.01 level the coefficients for the state dummy variables are not reported 14 References Steven G. Allen, "Unionized Construction Workers Are More Productive," Quart-rly Journal of Economics,(May 1984),pp. 251-274. Martha Fraundorf, John P. Farrell and Robert Mason,"The Effect of the Davis-Bacon Act on Construction Costs in Rural Areas,"The Review of Economics and Statistics, 1993, pp. 142- 146. Martha Norby Fraundorf, John P. Farrell, and Robert Mason, "Effect of the Davis-Bacon Act on Construction Costs in Non-Metropolitan Areas of the United States,"Report to the American Farm Bureau Federation, (Corvalis, Oregon: Department of Economics, Oregon State University, 1982). John G. Olsen,"Labor and Material Requirements for New School Construction,"Monthl April 1979 . 38-41. Labor Review,Ap ,pp IS Biographical Sketch Mark Prus is Associate Professor of Economics at the State University of New York, Cortland. He has published articles on wage and occupational structures in the Cambridge Journal of Economics, Journal of Economic History, Quarterly Review of Economics and Finance, Journal of Socio-economics, Journal of Economics Issues, Social Science Journal,and Research in Economic History. P Hing Wage Laws and School Co ion Costs An Analysis of Public School Construction in Maryland and the Mid Atlantic States By Mark J. Prus, Ph.D. Associate Professor Department of Economics SUNY Cortland Prepared for the Prince George's County Council, Maryland January 1999 Table of Contents List of Tables and Figures About the Author Acknowledgements Executive Summary Chapter One: Prevailing Wage Regulations in Maryland and the U.S. Chapter Two: The Impact of Prevailing Wage Laws on School Construction in the Mid- Atlantic Region Chapter Three: The Impact of Prevailing Wage Laws on School Construction in Maryland Chapter Four: Prevailing Wage Laws and Competition between Contractors Chapter Five: The Effect of Prevailing Wage Laws on Construction Industry Wages List of Tables and Figures Table 1: Prevailing Wage Laws, by State, Year Passed and Repealed Figure 1: Median Square Foot School Construction Costs for Mid-Atlantic States With and Without Prevailing Wage Laws Table 2: Square Foot School Construction Cost in States with and without State Prevailing Wage Laws Table 3: Square Foot School Construction Costs by Public and Private Projects Table 4: Square Foot Cost of New School Construction Broken Down by State WITH and WITHOUT State Prevailing Wage Laws and then Broken Down by Public and Private Schools Table 5: A"Here-There" Cross-State Linear Regression Model Predicting Total Construction Costs for New Schools Table 6: Data on School Construction Projects Used in the 1989 Department of Fiscal Services Report on Maryland's Prevailing Wage Law Table 7: Regression Results for 20 Public School Projects Used in the Department of Fiscal Services Report on Prevailing Wage Laws, 1989 Table 8: Distribution of School Construction Projects by County and Type of Construction within Maryland Table 9: Regression Results Estimating School Construction Costs within Maryland Table 10: Distribution of Contractors Working Inside and Outside "Local" Area Table 11: Measurement of Distance Traveled by Differences in Zip Code Table 12: Regression Results Predicting the Probability of a Contractor Being Awarded a School Construction Project Table 13: Earnings of All Construction Industry Workers Relative to All Non-Agricultural Employees by Prevailing Wage and No Prevailing Wage Law States Table 14: Relative Earnings of Construction Workers in Maryland Table 15: Relative Earnings of Construction Workers to All Non-Agricultural Workers, by School Prevailing Wage Law Jurisdiction, 1988-1996 About the Author Mark Prus grew up in Northeastern Ohio and Arizona. He lived in Cumberland, Maryland briefly. He received his B. A. in economics from the University of Notre Dame in 1979. Prus received his Ph.D. (1985) from the University of Utah where he was the Marriner S. Eccles Fellow in Political Economy. Prus is an Associate Professor of Economics, and Chairman of the Economics Department at the State University of New York at Cortland. Prus has published widely on labor market issues in journals such as the Journal of Economic History, The Journal of Economic Issues, the Cambridge Journal of Economics, Research in Economic History, and the Quarterly Review of Economics and Finance. Prus is a respected expert on prevailing wage laws and their effects on construction costs. He has served as a consultant to the Construction Labour Relations Association of British Columbia and testified before the Department of Industrial Relations in California on their state prevailing wage law. He has presented the results of his research on the construction industry at the Western Economics Association Annual International Conference and the Economics Research Network. Acknowledgements A number of people gave generously of their time in the preparation of this report. Dr. Yale Stenzler, of the Interagency Committee on School Construction, described at length the the process of public school construction funding in the state of Maryland and also provided me with copies of the reports on Prevailing Wages in Maryland produced by the Department of Fiscal Services. Dr. Stenzler also gave me the names of facility planners in each county, allowing me to check on the coverage, or lack thereof, of prevailing wage laws. Tammy Woolford and Richard Avalone at the Prevailing Wage Unit of the Maryland Department of Labor helped me identify school projects in Maryland built with prevailing wage requirements. Jim Correll of the Baltimore Building and Construction Trades and Sheldon Sugarman of the Baltimore City Wage Rate Commission helped me understand the application of Baltimore City's local prevailing wage law, and identify schools in Baltimore City built with prevailing wage requirements. Dr. David Miller generated the map of the Mid Atlantic states and Dr. Lisi Krall patiently read the entire report and offered helpful comments. All of these people, and many others patiently answered my questions, even when I asked them more than once. Without their help my chances of getting it right were much smaller. Or course I alone am responsible for any remaining errors. Executive Summary At the request of the Prince George's County Council, I conducted this analysis of the impact of prevailing wages on public school construction projects. The county council commissioned this study to guide them as they consider the adoption of a bill to require adherence to state prevailing wage rates in the construction of public schools. While the state of Maryland has a prevailing wage law, it mandates the payment of prevailing wages percent or more of the funding. which the state provides 75 erc 9 onlyfor those school projects for p p . . Geor e's County is embarking on a six year capital program for the construction and Prince ty g Y 9 /or renovation of eighteen schools in the county. I have attempted to address the following four concerns of the county council: • Compare school construction costs in states with prevailing wage laws to those in states without prevailing wage laws in the mid Atlantic region. • Compare school construction costs within Maryland for those local jurisdictions that pay prevailing wages to costs in those areas where prevailing wage rates are not required. • Analyze the extent to which local contractors have been harmed by unfair competition from outside contractors due to the absence of prevailing wage requirements on school construction projects. • Examine the extent to which the absence of prevailing wage rates in school construction impacts construction wages across the construction industry. The analysis presented here provides answers to each of these questions primarily through the statistical manipulation of data on individual school construction projects provided by the F. W. Dodge Corporation. The analysis exploits variations across the mid Atlantic region, or within the state of Maryland, in the application of prevailing wage requirements. Within the mid Atlantic region, Delaware, Pennsylvania and West Virginia have prevailing wage laws that apply generally to public school construction. North Carolina and Virginia do not have state prevailing wage laws, and while Maryland has a state prevailing wage law, it only applies to public school construction projects if the state government provides 75 percent or more of the funding for the project. The Maryland state law does, however, allow for voluntary adherence to prevailing wages, and two local jurisdictions (Allegany County and Baltimore City) have elected to do so. i brief history of prevailing wage laws and their impact on Chapter 1 resents a ry p 9 9 p p construction costs. In Chapter 2 of this study, I examine the actual square foot construction costs of schools built in mid Atlantic states with and without prevailing wage laws over the period 1991 to 1997. This study includes data on school construction projects for Delaware, Pennsylvania, West Virginia, Maryland, North Carolina and Virginia. The first three states have prevailing wage laws that apply to school construction; the latter three do not. While square foot costs for new school construction are generally higher in prevailing wage law states, this may be due to regional differences in the cost of living. Public school construction costs also tend to be higher than private school construction costs, but this is true irrespective of whether or not a state has a prevailing wage law. A formal linear regression statistical model capable of controlling for these and other factors confirms the hypothesis that there is no measurable or statistically significant increase in construction costs associated with prevailing wage regulations. Chapter 3 examines school construction costs within the state of Maryland. A previous study by the Department of Fiscal Services, which found that prevailing wage requirements within the state raised school construction costs by 15 percent, is examined. The results of that study are shown to be very sensitive to the inclusion or exclusion of particular variables. At the time of that previous study most schools were built in Maryland with prevailing wage requirements. Since that time changes in the allocation of state funds have led to the vast majority of schools being built without prevailing wages. Some local jurisdictions, however, have already adopted prevailing wage requirements of the kind being considered by Prince George's County. These include Allegany County and Baltimore City. A comparison of school construction costs between those jurisdictions with prevailing wages and those without is conducted. Controlling for other influences on costs, prevailing wage requirements do not measurably affect the costs of school construction. Chapter 4 examines the relationship between prevailing wage laws and the extent of competition between contractors. One of the original intents of prevailing wage laws was to limit competition from itinerant contractors who would import low wage labor from outside the local labor market. Little empirical work has been done in the past to determine the extent to which this original intent has been fulfilled. Regression analysis is used to determine the nature of the correlation between prevailing wage laws and whether contractors from outside the local construction labor market are successful in bidding for school construction projects. Prevailing wage regulations appear to restrict the ability of urban contractors to win rural construction contracts but encourages rural contractors winning of urban jobs. Chapter 5 looks at the impact that prevailing wage laws have on construction workers'wages. Using data from the U.S. Census Bureau's County Business Patterns, the wages of construction workers relative to all employees' wages are compared for prevailing wage and non-prevailing wage jurisdictions. Construction workers earn a premium in prevailing wage jurisdictions over other nonagricultural workers. This premium ranges from 7 to 16 percent. Prevailing Wage Laws in Maryland and the U. S. : Their History, Intent And Impact Nationally, prevailing wage laws date back to the Republican Congress of 1868 that passed a National Eight-Hour Law that provided for an eight-hour day on public construction. Congressional debate made it clear that when the working day was shortened from 12 or 10 hours per day to eight, workers were still to be paid the prevailing daily wage. Ulysses Grant was the first President to call for the enforcement of prevailing wage regulations. Kansas was the first state (1891) to pass a prevailing wage law. In upholding the constitutionality of this law, Supreme Court Justice John Marshall Harlan stated that the purpose of the law was to raise labor standards not only in construction but by example for all blue collar workers. Seven states passed prevailing wage laws between 1891 (Kansas) and 1923 (Nebraska). Sixteen states passed prevailing wage laws between 1931 and 1937. Maryland's prevailing wage law was originally passed in 1945. Eventually all but nine states would pass prevailing wage laws. (See Table 1.) At the national level, the Davis-Bacon Act, passed in 1931, required payment of prevailing wages on federally financed construction projects. However, the original language of the law was vague, and prevailing wages generally were not determined before the acceptance of bids. In 1935, President Roosevelt signed clarifying amendments to the act, which became the basis of the current Davis-Bacon Act. Prevailing wage laws emerged from a concern that cutthroat competition over wages in construction would lead the industry down a low-wage, low-skill development path. This was said to put the quality of construction at risk and lead to an itinerant, footloose, low-wage construction labor force. Poor construction workers would make poor neighbors and potential burdens on the community. Reasonably paid construction workers, on the other hand, held out the possibility of being solid neighbors, good citizens and productive members of the community. Government, by the operation of prevailing wage laws, was supposed to get out of the business of cutting government costs by cutting the wages of its citizens. Whatever labor standards had been established, whatever wages prevailed in a local community; that is what the law said government should pay on public works. i Maryland's prevailing wage law establishes Like other prevailing wage laws, ry p 9 9 minimum wage levels to be paid on public construction projects. Maryland's current law enacted in 1969, requires that prevailing wages be paid to workers on any public construction project which receives 50 percent or more in state funding and is valued at $500,00 or more. An earlier prevailing wage law, adopted in 1945, applied to highway projects in certain parts of the state. Public school construction is subject to prevailing wage requirements in Maryland if 75 percent or more of the funding comes from the state. While this threshold appears to have been met in the majority of school construction projects in the 1980s, changes in the method of allocating state funding for school construction in 1989 reversed this situation so that most schools in Maryland are built without prevailing wage requirements. Allegany County and Baltimore City have enacted local prevailing wage laws that typically apply to school construction projects. Baltimore City's Ordinance, approved in 1973, required the payment of minimum wages on "each and every contract in excess of five thousand dollars made by the Mayor and City Council of Baltimore." The minimum wages required in these contracts would be set by the Board of Estimates at least once every year to conform to area prevailing wage rates. Since the late 1970s, prevailing wage laws at both the national and state level have come under pressure. Calls for repeal or reform of prevailing wage laws have been motivated by the suspicion that they increase public construction costs and hinder competition. Opponents of prevailing wage laws argue that they raise construction costs. Repeal of these laws would result in cost savings on the order of 20 to 30 percent according to some. If such savings were possible, it is argued, school districts could build five schools for the price of four. But while a number of states have moved to repeal their prevailing wage laws, Prince George's county is not alone in considering the adoption of stronger prevailing wage statutes. Kentucky, for example, recently re-extended its state prevailing wage law to cover school construction costs. Outside of the United States, the provincial government in British Columbia adopted a prevailing wage regulation through its Skill Development and Fair Wage Act in 1992. Estimating the Impact of Prevailing Wage Laws on Construction Costs In all cases, policy makers are concerned with the costs associated with prevailing wage laws. Claims of the added cost associated with prevailing wage laws and of cost savings from repeal have not been adequately supported by empirical evidence. Some efforts to estimate the impact of prevailing wages are construction costs have used differences in wage rates between union and nonunion construction workers. Yet wage differences have a moderate effect on total construction costs. cos ts and may have been failing. d total construction9 Labor costs are less than a third ofY In 1972, for instance, in an analysis of school construction costs, John Olsen found that 2 onsite wages and salaries excluding benefits were 28.2 percent of total costs. (Monthly Labor Review, 1979, p. 40) According to the Census of Construction, labor costs, counting benefits, on all types of construction were 30 percent of total costs in 1977 and had fallen to 26 percent by 1987. A second problem leads us to question estimates of the impact of prevailing wage legislation on construction costs based on an analysis of wage differences. Because they assume implicitly that the same number of hours of each type of labor will continue to be employed and that labor is of invariant productivity the impact on costs is driven by the wage differential. Neither of these assumptions is necessarily appropriate. The payment of prevailing wages may serve to attract workers with more experience and training. Increased labor productivity may result in fewer hours of labor being required thus offsetting the higher wage rate. For instance, Allen has shown that unionized labor in the construction industry rY is between 17 and 52 percent more productive than nonunion labor. (Allen, 1984) Additionally, higher wage rates may lead contractors to substitute capital or other inputs for labor, mitigating the impact of higher wages rates on total construction costs. These possibilities, alone or in combination, make the assumptions underlying the analysis of construction costs based on wage differences inappropriate and cast doubt on the estimates of cost savings. Specifically, it is difficult to imagine how savings of 20 to 30 percent are possible. To get a true picture of the impact of prevailing wage legislation's impact on total construction costs, one could evaluate not only differences in wage rates, but also productivity differences, the incidence of substitution, administrative costs and other ways in which the impact of these laws is either mitigated or enhanced. An alternative approach is to simply examine total construction costs directly and compare costs in the presence and absence of prevailing wage laws controlling for project differences. Few studies have attempted to estimate the impact of prevailing wage legislation based on actual total construction costs. Fraundorf, et. al., in "The Effect of the Davis-Bacon Act on Construction Costs in Rural Areas," examined 215 new, non- residential construction projects built in 1977 and 1978. (Fraundorf, 1983) Approximately half of these projects were federal projects built under the purview of the federal Davis Bacon Act specifying that prevailing wages be paid. The other half were privately owned projects constructed without the requirement that prevailing wages be paid. Data on total construction costs were then compared using multivariate regression analysis to control for the effects of factors other than the presence of prevailing wage requirements. This study controlled for differences in the type of structure, the types of materials used, and project size in an effort to focus on cost differences associated with labor cost differentials resulting from the dichotomy in regulatory regimes. It also attempted to control for regional differences in construction costs by grouping projects into four regions; Northeast, North Central, South and West. The dependent variable in their regression analysis was the natural log of the project's bid price deflated to 1977 dollars. The authors of the study found, somewhat 3 construction projects governed b Davis Bacon were 23 surprisingly, that federal cons p ) g Y percent more expensive than private construction projects controlling for other cost influencing factors. When they re-estimate their basic model to correct for disproportionate response rates by region and building type, Fraundorf finds that the impact of Davis Bacon on total construction costs is as high as 30 percent. While they admit that these results are high, especially in light of a public-private wage differential estimated at 20 percent, they point out that the results are consistent with other aspects of the data. In particular they do not find evidence of factor substitution which would mitigate the impact of prevailing wage requirements. However, this does not explain why the impact on total costs is greater than the wage differential. They explore other factors that might contribute to the higher costs of federal Davis Bacon projects such as record keeping and reporting, and decreased competition. Neither of these factors appear to significantly contribute to costs on federally funded construction. (Fraundorf, et. al, 1983, p. 145) One possible problem with the Fraundorf study is that regional differences in construction costs may have been inadequately controlled for. Given the relatively small sample size, the authors of this study had to group construction projects into relatively broad regions. This creates the potential for comparing a private project in a low cost state such as Idaho with a public project in a high cost state such as California. Since both projects are considered to be in the same region the cost differential is incorrectly attributed to the impact of prevailing wages when in fact it is due to differences in the cost of living or cost of materials between Idaho and California. Another problem may result from the way in which building types were classified. Each construction project was placed into one of six categories; recreational buildings, storage facilities, industrial buildings, office-commercial, medical and other. These categories were then used to find matches between public and private construction projects. However, these six categories were sufficiently broad to allow rather dissimilar buildings to be considered comparable. For instance, in the category storage facilities, warehouses were grouped with barns as well as airplane hangars. Likewise office buildings were in the same category as restaurants. Differences in costs between public and private buildings may have resulted from differences in structure type and not from prevailing wage requirements. (Fraundorf, 1982, pp. 14-15) A second and potentially more serious problem with this study is that it fails to adequately isolate the impact of prevailing wage legislation on construction costs. Specifically, Fraundorf compares private projects constructed in the absence of prevailing wage legislation with federal (i.e., public) projects built using workers paid the prevailing wage. This comparison, while seemingly appropriate, contains the potential for confounding cost differences related to prevailing wage laws with cost differences resulting from other differences between private and public construction projects. The authors recognize this possibility when they point out, "If the government is more exacting than private owners in its quality standards, labor hours (and costs) and 4 possibly material costs would be higher on government projects." (Fraundorf, 1983, p. 145) It may also be that the difference in bidding procedures for public and private contracts or differences in the time horizon of public and private owners may contribute to higher costs in the public sector. In other words, the cost differential that Fraundorf attributes to the effect of prevailing wage legislation may in fact be due to differences between public and private construction. While the Fraundorf study suffered from certain problems in the specification of the model, it opened the way for the use of regression analysis for studying the impact of prevailing wage laws on public construction costs. I have used a regression model patterned after the Fraundorf study to analyze total construction costs and prevailing wage laws in the U.S. and in British Columbia. In my analysis of state prevailing wage laws, I found that while public projects were significantly more expensive than similar private projects, this was true in both prevailing wage law states and non-prevailing wage law states. Consequently, the higher costs of public projects could not be attributed to the presence of prevailing wage laws. In fact, the estimated effect of prevailing wage laws, controlling for other factors, including differences in the type of ownership, was not statistically different from zero. 5 Table 1: Prevailing Wage Laws by State, Year Passed and Repealed States having Year prevailing wage laws passed States without prevailing wage laws Alaska 1931 Georgia Arkansas 1955 Iowa California 1931 Mississippi Connecticut 1935 North Carolina Delaware 1933 North Dakota District of Columbia 1931 South Carolina Hawaii 1955 South Dakota Illinois 1931 Vermont Indiana 1935 Virginia Kentucky 1940 Maine 1933 Maryland 1945 Massachusetts 1914 Michigan 1965 Minnesota 1973 Missouri 1957 Montana 1931 Nebraska 1923 Nevada 1937 New Jersey 1913 New Mexico 1937 New York 1894 Ohio 1931 Oklahoma* 1909 Oregon 1959 Pennsylvania 1961 Rhode Island 1935 Tennessee 1953 Texas 1933 Washington 1945 West Virginia 1933 Wisconsin 1931 Wyoming 1967 States that repealed Year Year of prevailing wage laws passed repeal Alabama 1941 1980 Arizona 1912 1984 Colorado 1933 1985 Florida 1933 197 9 Idaho 1911 1985 Kansas 1891 1987 Louisiana 1968 1988 New Hampshire 1941 1985 Utah 1933 1981 *The enforcement of Oklahoma's law was judicially suspended in 1995. 6 The Impact of Prevailing Wage Laws on School Construction Costs A Case Study of the School Construction Costs in Mid Atlantic States with and without Prevailing Wage Laws One way of estimating the impact of prevailing wage laws on school construction projects that avoids the problems associated with the necessarily restrictive assumptions involved in wage rate comparisons, is to compare actual total school construction costs in the presence and absence of prevailing wage regulations. Maryland, while it has a state prevailing wage law, constructs most schools in the 1990s without requiring prevailing wages. In contrast, a number of surrounding states in the Mid Atlantic region have prevailing wage laws that P generallyapply to school projects. These include Pennsylvania, Delaware and West Virginia. Other states, namely, Virginia and North Carolina do not have a state prevailing wage law. This regional variation creates the opportunity for a "here and there" analysis of school construction costs using a multiple regression model similar to both the Fraundorf model and the model I have used previously to estimate the impact of state prevailing wage laws on public construction costs. Before that model is presented, I develop a more intuitive comparison of square foot costs for school construction. A Comparison of School Construction Costs As a first exercise in comparing school construction costs, the median square foot costs of new public schools built from 1991 to 1997 in each of the six states in the Mid Atlantic regionare presented in Figure 1. As can be seen, square foot costs are highest in Pennsylvania and Delaware and lowest in North Carolina, Virginia and West Virginia. While Pennsylvania and Delaware are two states with prevailing wage laws, and North Carolina and Virginia are states without prevailing wage laws,it would not necessarily be appropriate to conclude that prevailing wage laws raise school construction costs. costsFigure 1: Median square foot school construction • • and without Prevailing wage 1.00 No P.W. Law fi+ � Partiai coverage Tables 2 and 3 show the median square foot construction costs for school construction projects in the mid Atlantic region over the period 1991 to July, 1997. The accepted bid price of the schools were inflated to 1997 dollars using the consumer price index-housing. This allows for a direct comparison of square foot costs for school construction projects in different years.' For all schools included in Table 2 the cost of schools in states without prevailing wage laws is $75.57 per square foot, while the square foot cost of schools in prevailing wage law states is $94.07. Schools in prevailing wage law states appear to cost 24.5 percent more. When school construction projects are disaggregated into elementary, middle and high school projects, we see that this difference is attributable to substantial differences in middle school and high school construction projects. The cost of elementary school projects in prevailing wage law states is actually lower. Table 2: Square Foot School Construction Costs by States with and without State Prevailing Wage Laws (notice caution in footnote) Square • .t School Constructionby with • without State Prevailing W Laws age LEGAL ELEMENTARY SCHOOLS Number MIDDLE SCHOOLS Number HIGH SCHOOLS Number STATUS Median of Obs. Median of Obs. Median of Obs. Note: Data are for 1991 to 1997 Inflated to 1997 Dollars Using the Consumer Price Index-Housing Source: F.W.Dodge Corporation Start Cost Data Caution:these data are for private schools only--no public schools are included However, Table 2 presents data for private schools only. These schools were built without prevailing wage regulations regardless of whether or not they were in a state with a prevailing wage law. Prevailing wage laws cover public projects only. The fact that the median square foot cost was higher for private schools in states with prevailing wage laws simply reflects the fact that the states without prevailing wage laws in the Mid Atlantic region are generally further south and may have dramatically different costs of living and costs of construction. Square foot construction costs are generally lower in these regions for private as well as public projects for a variety of reasons. Thus, in assessing school construction costs in states with and without state prevailing wage laws, we will need to take into consideration overall differences in construction costs in these groups of states. 'The data are from the F. W. Dodge Corporation,the standard service provider of project information in the construction industry. Alternative price indices were tried to examine whether results were dependent on the price index chosen. Results were basically the same regardless of the price index used to translate information into constant 1997 dollars. 9 Table 3 divides school construction into publicly and privately built r 'ects public schools cost $84.09 ion o hoot construct , schools. Combining all school projects, P per square foot compared to $75.57 for private schools. Public schools cost 11.3% more per square foot than private schools. When school projects are disaggregated by type, public elementary schools cost 3 percent more than private schools P 9 and public high schools cost more than 26% more per square o foot to build than private high schools. These data, however, refer only to public and private schools built in states that do not have a state prevailing wage law. Thus, the public-private cost differential cannot be laid at the foot of prevailing wage regulations. This reminds us that in assessing the effects of prevailing wage regulations on building costs, we must keep in mind that similar public and private buildings, such as elementary schools or high schools, may differ in the quality and nature of their construction. Table 3: Square Foot School Construction Costs by Public and Private Projects Square Foot Public and Private School OWNERSHIP Number ELEMENTARY SCHOOLS Number MIDDLE SCHOOLS Number HIGH SCHOOLS Number OF PROJECT of Obs. Median Sq.Ft.Cost of Obs. Median Sq.Ft.Cost of Obs. Median Sq.Ft.Cost of Obs. Note:Data are for 1991 to 1997 Inflated to 1997 Dollars Using the Consumer Price Index-Housing Note:Public Projects exclude Federal projects. Source: F.W.Dodge Corporation Start Cost Data Table 4 presents a more appropriate comparison, though it points to the problem of small subsample sizes at the same time. The square foot cost of new construction for elementary, middle and high schools is presented. These data state prevailing wage laws and states that are first broken down into states with s p g 9 do not have state prevailing wage laws. Then the data are broken down a second time into public schools and private schools. Finally, for both states with laws and states without laws, a comparison is made. How much more or less expensive is it to build a public school? Table 4 compares 76 public elementary schools in states with a prevailing wage law to the construction of 4 privately built elementary schools in those states. The public elementary schools cost 80% more per square foot than the private elementary schools. Perhaps this implies that prevailing wage laws raise elementary school construction costs by about 80%, though the magnitude of this cost differential exceeds virtually all estimates of the impact of prevailing requirements on costs. wage Public middle schools actually cost 2.8% less than private middle schools in prevailing wage law states—compared to the .8% difference in states without 10 prevailing wage laws. Public high schools in non-prevailing wage law states cost 26.7% more than private high schools in those states. In stark contrast, in prevailing wage law states, public high schools cost 32.7% less than private high schools, though there is only one private high school (and an exceptionally costly one at that) in prevailing wage law states. While these small subsamples make any inference drawn from these comparisons shaky, the data in Table 4 do not provide strong support for the contention that prevailing wage laws raise school construction costs. Table 4: Square Foot Cost of New School Construction Broken Down by State WITH and WITHOUT State Prevailing Wage Laws and then Broken Down by Public and Private Schools LEGAL OWNERSHIP ELEMENTARY SCHOOLS Number MIDDLE SCHOOLS Number HIGH SCHOOLS Number STATUS OF PROJECT Median Sq.Ft.Cost of Obs. Median Sq.Ft.Cost of Obs. Median Sq.Ft.Cost of Obs. NO LAW STATE Private Project $59.97 N=4 $112.09 N=4 $124.35 N=1 LAW Public Project $107.96 N=76 $108.99 N=22 $83.65 N=22 STATE %Increase in Public Cost 80.0% -2.81 -32.70 Note:Data are for 1991 to 1997 Inflated to 1997 Dollars Using the Consumer Price Index-Housing Note:Public Projects exclude Federal projects. Source: F.W.Dodge Corporation Start Cost Data Using a Linear Regression Model to Measure the Effect of Prevailing Wage Laws on School Construction Costs. regression is a standard called linear re In economics, a statistical technique g method for measuring the effect one factor has upon another controlling for other things. For instance, we can develop a model designed to predict the cost of building a school based on • Whether it is an elementary, middle or high school • how many square feet are in the project • how many stories the building is • what kind of building materials are used in construction • whether or not the school is public or private, and Controlling for these factors, we can then ask the question: if the school is being publicly built in a state with a prevailing wage law, will it cost more? This statistical technique is the same used by Fraundorf, et. al. to examine the impact of the Davis Bacon Act on public construction costs. �1 Table 5 presents a linear regression model that predicts total school construction costs (excluding land acquisition, architect fees and construction management fees) based on the size of the building, the number of stories, the type of building materials used, whether or not the school is an elementary, middle or high school, whether the school is public or private and whether the school was built under a prevailing wage law. Table 5: A "Here-There" Cross State Linear Regression Model Predicting Total Construction Costs for New Elementary, Middle and High Schools, 1991-1997 Constant -9.866 Yes 0% Log of Total Square Feet 0.872 Yes 0% Log of the Number of Stories 0.066 Yes 10% Marker for Wall Board Framing -0.016 No 60% Marker for Wood Framing 0.055 Yes 8% Marker for Steel Framing -0.239 No 31% Marker for Cement -0.051 No 49% Middle School 0.007 No 84% High School 0.046 No 26% Public Project 0.264 Yes 0% Log of Regional CPI 3.334 Yes 0% Effect of Prevailing Wage Law 0.038 No 26% Dependent Variable:Log of the Total Project Value in 1997 Dollars Deflating with the CPI-Housing Adjusted R Square=.887 Number of Observations=358 This linear regression model covers the construction of schools in Delaware, Maryland, North Carolina, Pennsylvania, Virginia and West Virginia. In order to control for cost of living differences across states a cost of living index was included as a control variable. The model is a good fit of the data (as indicated by an adjusted R-square statistic of 89%). The model indicates that as the size of a school goes up, the total cost of the school rises. But it also indicates that there are economies of scale associated with larger schools so that while the total cost goes up with increasing size, the square foot cost goes down. This is shown in the estimated effect (or coefficient) for the variable--the log of total square feet for the school being built. This coefficient is .87. This means that if you doubled the size of a school from (say) 50,000 square feet to 100,000 square feet, the size would go up by 100% but the cost would only go up by 87%. This indicates that as schools get larger, the total cost goes up, but the square foot cost goes down. The model in Table 5 controls for a variety of technical factors--total square feet, number of stories, type of building materials--and the model also controls for whether or not a middle school costs more than an elementary school or whether a high school costs more than an elementary school. For a middle school of exactly the same size as an elementary school, built of the same material, having the same number of stories, the model estimates that a 12 middle school will cost .7% more. A high school of identical size, using the same materials will cost 4.6% more. Neither of these results, however, are statistically significant. There are three standard levels of statistical significance--1%, 5% and 10%. In simple terms, 10% means statistically I have a 1-in-10 chance of being wrong, and 1% means I have a 1-in-100 chance of being wrong. Rarely do economists accept higher levels of probability in this test as statistically significant. This means that for all practical purposes, an elementary school, a middle school and a high school of the same size will cost the same amount. Now the model asks the question whether or not public schools cost more than private schools controlling for other factors such as size. The model estimates that public schools (in states with and without prevailing wage laws) cost 26.4% more than private schools. The estimated cost difference associated with public school buildings is statistically significant at all standard levels of probability. This cost difference may be due to design differences or other features typically found in public schools compared to private schools. Public school buildings may have a longer life span than private school buildings, or other factors may account for this cost difference. But this cost difference exists in both states with and without prevailing wage laws. This is not a cost differential that can be attributed to prevailing wage laws simply because this cost differential is found where there are no prevailing wage regulations. Finally, the model estimates the cost effect of prevailing wage laws. The model estimates that controlling for other factors, building a public school in a prevailing wage law state will cost 3.8% more than building the same public school in a state without a prevailing wage law. However, this is not a statistically significant estimate. For all practical purposes there is no statistical difference between building a public school in a state with or without a prevailing wage law. How can the model say there is no difference in the cost of public school construction in states with prevailing wage laws compared to states without prevailing wage laws when Table 4 suggests that on average square foot public school construction costs are higher in states with prevailing wage laws? Once again, the answer is that, on average, private school construction costs are also higher in states with prevailing wage laws. Once these cross-state differences in construction costs are accounted for, there is no statistically measurable effect on total construction costs associated with prevailing wage regulations. Conclusion A "here-and-there" linear regression model was developed to estimate the effect of prevailing wage regulations on total construction costs for schools, controlling for other factors. This model controlled for the type of school, the size of the project, and building characteristics. It also controlled for general differences in construction costs between states with and without prevailing wage 13 � laws and general differences between the cost of public and private construction n (whether or not done under prevailing wage regulations). Controlling for these factors, this model could find no statistically significant impact on total construction costs due to prevailing wage requirements. is 14 0 The Impact of Prevailing Wage Laws on School Construction Costs in Maryland Introduction In Chapter 2 we compared school construction costs in mid Atlantic states that have prevailing wage laws with costs in those states without prevailing wage laws in the mid Atlantic region. In this chapter, we exploit local variation in the application of prevailing wage laws to compare school construction costs within the state of Maryland. Unlike the previous comparison where the opportunity for dramatic differences in the cost of living could contaminate the results, comparing school construction costs in those jurisdictions with prevailing wage requirements in Maryland to costs of school construction for those jurisdictions where prevailing wages are not paid provides a clearer focus on the law's impact. A Reconstruction of the Department of Fiscal Services' Study This is not the first time the impact of prevailing wage laws on school construction costs in Maryland has been examined. A report prepared by the Department of Fiscal Services (January 1989) ten years earlier evaluated the impact of Maryland's prevailing wage with a special focus on public school construction. The study examined 20 school projects funded by the Interagency Committee for Public School Construction in 1987 and 1988. Of these projects, 14 were built under the guise of prevailing wages while the remaining 6 were not. A multiple regression analysis was performed using square foot cost as the dependent variable and proxies for building design, location, and the applicability of prevailing wage regulations as independent variables. The model estimated that prevailing wage requirements raised school construction costs by approximately$11 per square foot, or roughly 15 percent. The report by the Department of Fiscal Services includes a detailed discussion of the methodology and data used in coming to this conclusion. These data allow me to replicate the earlier study and discuss some problems 15 associated with the methodology and findings. I have reproduced the data from the Appendix 4 of the Department of Fiscal Services report in Table 6. Table 6: Data on School Construction Projects Used in the 1989 Department of Fiscal Services Report on Maryland's Prevailing Wage Law 0 0 School Name County Year /o Sq Ft /o Est Site /o Special Sq Ft Cost Prevailing Renovation Preparation Conditions Wage? S. Middle A.Arundel 1986 53 9 0 76 No G. Fox A.Arundel 1987 88 8 0 44 No G.Washington B.City 1987 0 11 0 88 Yes Garrison B. City 1987 91 10 0 73 Yes Dundalk B. County 1986 100 0 0 50 Yes Sunderland Calvert 1987 0 10 0 88 Yes Appeal Calvert 1987 40 9 0 90 Yes Manchester Carroll 1987 35 12 0 76 Yes Voc/Tech Carroll 1986 0 11 0 104 Yes Jenifer Charles 1986 0 11 0 82 Yes Hillcrest Frederick 1987 0 11 0 77 Yes Dublin Harfors 1986 92 3 1 38 Yes F. Douglas P.George's 1987 79 5 0 70 Yes Flow Hill Montgomery 1984 0 11 0 79 No Oak View Montgomery 1984 79 5 1 50 No G.Lake Montgomery 1986 0 7 11 106 No B. Hills Montgomery 1983 55 5 0 52 No 8th District St. Mary's 1988 0 12 0 92 Yes Bester Washington 1987 99 5 0 75 Yes Pinehurst Wicomico 1988 82 6 2 70 Yes Source:Department of Fiscal Services,"Maryland's Prevailing Wage Law:A Study of Costs and Effects Two potentially serious, and related, problems with the 1989 study are the limited number of observations and the intermingling of new schools with renovation projects. Of the 20 projects listed, 8 are new schools and the remaining 12 involve at least some renovation work. If we look only at the 8 new schools included in the study, there are 2 projects built in the absence of prevailing wage requirements and 6 schools constructed with prevailing wages. Comparing the mean square foot costs of construction, we find that schools built with prevailing wages cost $88.50 per square foot while schools built without prevailing wages averaged $92.50 per square foot. Prevailing wage schools were, on average, 4.5 percent less expensive. Given the variation even amongst new school costs, this difference is not statistically significant. In fact a careful examination of the raw data indicates that the most expensive new school in the sample was built without prevailing wage requirements. This observation is responsible for inflating the average cost of new schools in the absence of prevailing wages. In order to overcome the problem of small sample size, school projects that included renovation work were added to the sample. When all 20 school 16 construction projects are compared, the average square foot costs of projects not requiring prevailing wages is $67.83 compared to $76.64 for those projects paying prevailing wages. This difference of 13 percent is not, however, statistically significant given the variation across both groups. A large part of the variation in school construction costs can be attributed to the amount of renovation work included in the project. Renovation work itself is variable and affects square foot costs differently. For example, renovation work such as installing a new boiler would increase square foot costs significantly both because of the expense of materials and the small square footage involved. Painting, on the other hand, may be far less expensive overall, but also involve a large area, substantially lowering the square foot costs. Overall, square foot costs appear to decline as the percentage of square foot renovation work increases. In order to illustrate some of the problems with the Department of Fiscal Services study I reconstruct their regression analysis. To begin with I use the 8 projects and regress square foot costs against a observations on new school g q 9 P 1 dummy variable for the prevailing wage. Dummy variables are variables that indicate the presence or absence of a quality or characteristic, in this case, the absence of prevailing wage requirements. irements. The results of this presence or a q p P 9 9 regression are presented in Table 7. In the first model the estimated effect of prevailing wages on square foot costs is negative though not statistically significant. In addition, overall, the regression equation does not produce statistically significant results pointing to the need to increase the number of sta P 9 Y 9 observations. In the next model, renovation projects are included with new school projects. In addition to the prevailing wage variable, the percent of renovation work in the project is included as an explanatory variable. As expected, the percent of renovation work is negatively correlated with square foot costs as indicated by the minus sign on the coefficient. The effect of prevailing wage requirements, controlling for renovation work, is now positive though not si gn of the estimated coefficient i statistically significant. In other words, while the g is positive, we cannot conclude that the effect is significantly different from zero. Thus prevailing wage requirements do not appear to have a measurable impact on costs! The Department of Fiscal Services study also included information on other factors that could affect costs. These included the percent of total costs attributable to site preparation, and a control for region. Including site preparation in our regression equation, we find that site preparation is positively correlated with square foot costs but not significantly so. The percent renovation work remains significantly negatively correlated to square foot costs. Most importantly, while the estimated coefficient on the prevailing wage variable is still positive it is not statistically significant. Including the regional control, which identifies school projects in the Baltimore area, leaves the prevailing wage variable still not statistically different from zero. Table 7: Regression Results for 20 Public School Projects Used in the Department of Fiscal Services Report on Prevailing Wage Laws, 1989 Variable Model Model Model Model Model Prevailing Wage -4 8.222 5.987 7.468 12.3 (-.426) (-1.351) (0.82) (1.11) (2.05) • Sq Ft Renovation - -0.348 -0.294 -0.295 -0.177 (-5.006) (-2.734) (-2.636) (-1.684) % Site Preparation - - 0.87 0.872 2.238 (0.66) (0.64) (1.77) Regional Control - - - 0.733 0.906 (0.08) (0.12) % Special Conds. - - - - 3.215 (2.62) Constant 92.5 83.76 74.784 74.825 52.731 (11.38) (13.95) (5.01) (4.85) (3.39) No. of Observations 8 20 20 20 20 Adjusted R Square -0.132 0.57 0.555 0.525 0.659 Note: t-statistics are in parentheses. Only when the control for special conditions costs is included in the regression model, as I do in the last column in Table 7, does the effect of prevailing wage requirements become significant. But this result, in which the prevailing wage variable is significantly positive for the first time, appears to hinge on the inclusion of one variable. This exercise in replication, if nothing else, points out the fragility of the results reported in the earlier study and calls for further efforts to estimate the effect of prevailing wage laws on school costs. A Regression Analysis of School Construction Costs in Maryland In order to overcome the primary limitation of the earlier analysis of school costs in the presence of prevailing wages, I use Dodge data from 1991 to 1997 to estimate a regression model like that presented in Chapter 2. In this case the observations used are limited to the state of Maryland. In contrast to the period used in the Department of Fiscal Services report, when a majority of schools were built with prevailing wages, changes in the way school projects were funded have reversed the situation so that the vast majority of school construction projects do not require the payment of prevailing wages. Consequently finding a sufficiently large number of prevailing wage projects for comparison was a challenge. Allegany County and Baltimore City have enacted local prevailing wage requirements for school projects, providing the bulk of observations. In addition, the Maryland State Department of Labor has identified a limited number of other school projects for which prevailing wage determinations were requested. These were projects for which the Maryland state law (requiring 75 percent state funding) applied. 18 s The distribution of school projects by county and by type of construction within Maryland is presented in Table 8. Of the 186 school construction projects from 1991 to 1997, seventy one projects (38%) were new schools. The vast majority of projects were concentrated in Anne Arundel, Baltimore, Montgomery, and Prince George's counties. Eleven projects (6%) were identified as prevailing wage projects. Nearly 60 percent of the projects were elementary schools, with the others almost evenly divided between middle schools and high schools. Table 8: Distribution of School Projects by County and type of Construction County Elementary Middle High School Allegany 2 0 1 Anne Arundel 13 3 5 Baltimore 15 5 4 Calvert 3 1 1 Caroline 1 1 0 Carroll 4 4 0 Cecil 0 2 0 Charles 3 1 1 Dorchester 3 0 0 Frederick 3 5 0 Garrett 1 0 1 Harford 7 1 0 Kent 0 1 0 Montgomery 23 10 5 Prince George's 12 1 4 Queen Anne's 1 0 0 St Mary's 2 0 1 Somerset 0 1 1 Talbot 2 0 1 Washington 4 0 5 W icomico 0 0 1 Worcester 1 0 0 Source: F.W. Dodge Corporation Table 9 presents the results of a linear regression model identical to that used in the previous chapter that predicts total school construction costs (excluding land acquisition, architect fees and construction management fees) based on the size of the building, the number of stories, the type of building 19 materials used, whether or not the school is an elementary, middle or high school, whether the school is public or private and whether the school was built under a prevailing wage law. Table 9: Regression Results Estimating School Construction Costs within Maryland Constant 7.383 Yes 0% Log of Total Square Feet 0.678 Yes 0% Log of the Number of Stories -0.02 No 84% Marker for Wall Board Framing -0.017 No 86% Marker for Wood Framing 0.162 Yes 10% Marker for Cement -0.022 No 90% Marker for Steel Framing -0.167 No 57% Middle School -0.019 No 85% High School 0.329 Yes 0% Renovation -0.252 Yes 0% Public School 0.406 Yes 0% Effect of Prevailing Wage Law 0.018 No 90% Dependent Variable:Log of the Total Project Value in 1997 Dollars Deflating with the CPI-Housing Adjusted R Square=.817 Number of Observations=124 Just as in the case of the Cross state regression model presented in Chapter 2 the model used here to analyze school construction costs within Maryland is a good fit of the data (as indicated by an adjusted R-square statistic of 82%). In other words, 82 percent of the variation in total school construction costs is accounted for by the explanatory variables listed in Table 9. Again, the model indicates that as the size of a school goes up, the total cost of the school rises. But it also indicates that there are economies of scale associated with larger schools so that while the total cost goes up with increasing size, the square foot cost goes down. The estimated effect (or coefficient) for the variable--the log of total square feet for the school being built-- is .68 indicating that as the square feet of a school project doubles, the cost increases by only 68 percent. The model in Table 9 controls for a variety of technical factors--total square feet, number of stories, type of building materials--and the model also controls for whether or not a middle school costs more than an elementary school or whether a high school costs more than an elementary school. For a middle school of exactly the same size as an elementary school, built of the same material, having the same number of stories, the model estimates that a middle school will cost 2 percent less. This result, however, is not statistically significant. On the other hand, a high school of identical size, using the same materials will cost 33% more. 20 Now the model asks the question whether or not public schools cost more than private schools controlling for other factors such as size. The model estimates that public schools (in areas with and without prevailing wage laws) cost 40.6% more than private schools. The estimated cost difference associated with public school buildings is statistically significant at all standard levels of probability. This result is consistent with the result for the cross state comparison in Chapter 2 and, again, is likely due to design differences or other features typically found in public schools compared to private schools. it is important to recognize that this cost difference exists in both areas with and without prevailing wage laws, and, as such, is not a cost differential that can be attributed to prevailing wage laws simply because this cost differential is found where there are no prevailing wage regulations. Finally, the model estimates the cost effect of prevailing wage laws. The model estimates that controlling for other factors, building a public school in a prevailing wage law jurisdiction within Maryland will cost 1.9% more than building the same public school without prevailing wage requirements. However, this is not a statistically significant estimate. For all practical purposes there is no statistical difference between building a public school with prevailing wages and building a public school without prevailing wages. 21 Prevailing Wage Laws And Competition Between Contractors Date Limitations One of the original rationales for prevailing wage laws was the proposition that such regulations would discourage outside contractors from bringing into an area low wage labor that undercut local wage and working conditions. No academic studies have been done to test whether or not this is indeed one of the effects of these regulations. This report is able to shed some light on this question using F.W. Dodge reports of public school contracts awarded to general contractors in the six states of Maryland, Pennsylvania, Delaware, West Virginia, Virginia and North Carolina over the period October 1996 to September 1998. The data are limited to information regarding the origin of general contractors. No information regarding the origin of subcontractors is available. The data also do not provide information regarding the workers employed by the contractor. It is possible for an outside general contractor to come into an area and hire many, if not all, workers locally. In other words, the data do not outside contractor bringing n in cheap distinguish between an outs g g p labor and an outside contractor hiring local labor at wage rates consistent with local labor market conditions. Definition of Outside Contractor With these data limitations in mind, the following analysis looks at three questions. First, within states, do prevailing wage laws influence the movement of urban contractors from major metropolitan areas to suburban and rural areas? Second, within states, do prevailing wage laws influence the movement of suburban and rural contractors into major metropolitan areas? Three, across states, do prevailing wage laws influence the movement of out-of-state contractors into other states influenced? 22 Analytical Framework Obviously factors other than prevailing wage regulations can influence whether or not outside contractors bid on distant jobs. In this analysis we are able to control for two of these factors. First, the size of the project will influence the universe of contractors willing to bid on that project. Presumably there are economies of scale that will make distant projects that are larger more attractive to potential bidders. Thus, all other things being equal, the larger the project the more likely it is that outside contractors will bid on the job and the more likely it is that one of those outside contractors will win the bid. Thus, we predict that the dollar value of a school project will be positively correlated with the probability that the general contractor is an outside contractor. Second, population density will make it more likely that an "outside" contractor will be close at hand. Heavily urbanized areas, such as around Philadelphia or Baltimore, are more likely to be exposed to and provide outside contractors compared to rural areas of West Virginia or Eastern North Carolina. Using the percent urban of a state's population as a proxy for the density of economic activity and the proximity of outside contractors, we predict that the more urban an area, the more likely will the contractor winning a school project be an outside contractor. With these two controls in place, we examine whether or not prevailing wage regulations influence the probability that the winning contractor on a public school job is an outside contractor. General Patterns in the Data Urban-Suburban-Rural Patterns within State Within the six states--Maryland, Delaware, Pennsylvania, West Virginia, Virginia and North Carolina, the data provide 601 cases where a public school project was awarded to a general contractor. Of these awards, 368 were awarded without the influence of prevailing wage regulations while 233 were awarded under prevailing wage regulations. In order to examine the flow of contractors between urban, suburban and rural areas, these contract awards and contractors were divided into urban and non-urban areas based on the zip codes of the school owner and the zip codes of the general contractor. Within the six states under analysis, Baltimore, Philadelphia, Wilmington, Pittsburgh, Richmond, Norfolk-Virginia Beach-Newport News, Greensboro-Winston Salem- High Point and Raleigh-Durham-Chapel Hill were determined to be major metropolitan areas. No area in West Virginia was classified as a major metropolitan area. 23 Whenever the zip code of a general contractor or owner fell within the urban boundaries of one of these major metropolitan areas, that owner or contractor was labeled an urban owner or contractor. All remaining owners and contractors within these six states were put in one suburban-or-rural group. This labeling allows us to analyze the movement of contractors within these six states across this line drawn between major metropolitan centers and other areas within each state. Table 10: Distribution of Contractors Working Inside and Outside "Local'Area Law No Law City Contractor Working in Suburban or Rural Area 11% 13% Suburban or Rural Contractor Working in City 10% 5%0 Cross-Boundary Work Subtotal 21% 18% City Contractor Working in City 8% 31% Suburban or Rural Contractor Working in Suburbs or Rural Area 71% 51% Within Boundary Work Subtotal 79% 82% Table 10 excludes out-of-state contractors in order to focus on the movement of general contractors within each state working on public school projects. Table 10 shows that within these six states most bids are won by general contractors coming from within the sector from which the bid is offered. This is true both for jurisdictions that enforce prevailing wage regulations and jurisdictions that do not have these laws. For instance, 79% of all bids under prevailing wage regulations were won by contractors coming from within the sector from which the bid originated. In comparison, 82% of the bids offered without prevailing wage regulations were won by contractors coming from within the sector from which the bid originated. Thus, regardless of legal regimes, roughly four-out-of-five in-state general contractors working on public schools came from the same sector as the school owner. Movement of Contractors within the Suburban-Rural Sector In this analysis, the suburban-rural sector in each state is geographically large. There may be considerable geographical movement of contractors within each state within the suburban-rural sector. The potential distance traveled, of course, depends upon the size of the state. Table 11: Measurement of Distance Traveled by Differences in Zip Code Average Zip Code Difference Zi Code of Owner Average Difference as Between Owner and General Contractor Minimum Maximum Range a Percent of the Range WV Law 347 24740 26807 2067 17% MD No Law 14 21629 21740 111 13% VA No Law 294 22030 24588 2558 12% DE Law 21 19711 19901 190 11% PA Law 412 15037 19609 4572 9% NC No Law, 147 27016 28906 1890 8% Table 11 measures the distance traveled by suburban and rural general contractors to build public schools by using zip codes. This table only looks at suburban and rural contractors within a state who received a public school construction contractor within the suburban-rural sector of that state. In column three the average difference between the zip code of the school owner and the general contractor is reported.2 Columns four, five and six report the minimum and maximum zip codes of school owners in each state and the range or difference between the top zip code and the bottom zip code. The range of zip codes is correlated to the size of suburban-rural sector of the state. The range for Maryland is smaller than the range for Delaware simply because the school projects put out to bid in Delaware during the period of this study were geographically more disperse than the suburban and rural jobs let out to bid in Maryland. Thus, the zip code range measures the size of the public school market within the suburban-rural sector of each state rather than the geographical size of this sector. Only when every corner of the state has jobs open to bid will the zip code range be perfectly correlated with the geographical range of the state. The focus column in Table 11 is the last column which shows the average difference between the zip code of the owner and the zip code of the general contractor as a percent of the range of zip codes for owners. Obviously in-state contractors in Delaware cannot move as far to jobs as instate contractors in Pennsylvania or Virginia. This percentage standardizes for the different sizes of each state. In absolute terms, Delaware contractors do not move nearly as far as Pennsylvania contractors. In relative (or percentage) terms, Delaware and Pennsylvania contractors move similar distances. In either absolute or relative terms, there is no clear pattern of movement correlated to the presence or absence of prevailing wage laws. In absolute terms, Pennsylvania and West Virginia suburban and rural contractors move the most while Delaware contractors move second to the least. These are the three states where prevailing wage laws govern suburban and rural public school construction. In relative terms, West Virginia contractors move the most followed by Virginia and Maryland. There is no obvious stacking of these states' contractor mobility patterns in relative terms by legal policy. It may be that patterns would emerge if the suburban-rural sector was broken apart into two sectors and/or additional states were added to the analysis. However, these research extensions were not possible within the time frame of this report. 2 In technical terms,the absolute value of the difference is reported. Using absolute values avoids the problems created by two equally distant contractors but where one has a lower zip code than the owner and the other has a higher zip code. Using absolute values treats these two hypothetical contractors as equally distant from the owner. 25 An Analysis of the Influence of Prevailing Wage Laws on the Movement of Contractors across the Major Metropolitan Boundary Controlling for Other Factors The basic conclusion of the foregoing analysis is that most general contractors building schools are local. However, at the margin, distant contractors do come into local areas. And their impact on local labor markets and construction markets may be disproportional to their numbers. Consequently, it is reasonable to pursue the question--what determines the crossing of the line? What determines the movement of an urban contractor into the suburban-rural sector of public school construction. What determines the movement of a suburban or rural contractor into the urban sector? What determines the movement of an out-of-state contractor into a state? To answer these questions we again use the statistical technique of multivariate regression analysis. The specific form of regression analysis used to answer the above questions is called logistic regression analysis. This is the appropriate tool when asking a yes-no question such as is the contractor an outside contractor? What a logistic regression model does is ask a question such as--will the contractor be an urban contractor working in a rural area (yes- no)--and then pose a set of variables designed to predict whether or not under certain circumstances the contractor in the rural area will have come from a major metropolitan area. What we want to do is design such a model and insert into it the presence or absence of prevailing wage regulations to focus on the question do prevailing wage laws influence the presence or absence of outside contractors. What other variables might influence whether or not a contractor will be a local or outside contractor? We propose three factors or variables that might help decide whether or not a local contract is won by an outside general contractor. First. the size of the contract. Our reasoning here is as follows. It costs money to bid on contracts. The farther away the project, the more information the contractor has to gather, the more costs and risks the contractor undertakes. Larger jobs are more likely to cover these risks and costs. So distant contractors are more likely to bid on and therefore more likely to win big jobs compared to smaller jobs. Our data set includes both new school construction, and additions, renovations and repairs. We expect outside contractors to be more common when these jobs are large. We will measure size by the total dollar value of the project. Second, economies can be densely populated or sparsely populated. In heavily urbanized areas the physical distance between a major metropolitan area and suburban and rural areas will be closer than in less urbanized, more sparsely populated areas. The cost of crossing the line between the suburbs and the city or even between states will be lower the extent to which states are more urban. 26 So we hypothesize that the states within our sample that are more urbanized are more likely to experience the use of outside contractors. Third, the wider the difference between the wages paid between a city and the surrounding suburban and rural areas the more likely it will be that suburban and rural contractors will move into the urban area and that urban contractors will not move out into suburban and rural areas. Furthermore, we expect that the higher one state's wages are relative to another, the more likely will out-of-state contractors come into that state. These are the expectations that we had in building what turned out to be four logistic regression models. One asks what is the probability that a urban contractor would win a suburban or rural school construction project? The second asks what it the probability that a suburban or rural contractor would win and urban school project? And the last two ask the question what is the probability that an out-of-state contractor would win an in- state school job? Table 12 shows the results of this statistical modeling. Table 12: Regression Results Predicting the Probability of a Contractor Being Awarded a School Construction Project Variables in Model: Urban Contractor Rural Contractor Out of State into State to Rural Owner to Urban Owner Model 1 Model 2 a b O d e h Constant 8.14 14% -21.56 0% -5.15 2% 36.56 0%u Total Cost of Project 5.67E-09 1% -9.90E-08 5% 7.73E-08 0% 2.87E-08 32% Urbanization of State 0.02 12% 0.05 4% 0.04 2% 0.02 35% City Wage Relative to State Wage -12.26 2% 14.79 0% -41.55 0% Prevailinq Wage Regulation -0.41 12% 1.02 2% -0.82 1% -0.51 28% Model Statistics: Chi Square Test of Fit 15.46 0% 21.08 0% 25.69 0% 204.00 0% Total Number of Contractors 572 572 634 633 Number of Contractors Working Across Line 1 57 39 61 61 Numbers in Bold Are Statistically Significant In columns a and b of Table 12, the results of modeling the question what is the probability that an urban contractor would win a suburban or rural school project are presented. There is a constant and four variables in the model. There are 572 observations, 57 of which are urban contractors who won suburban-rural public school projects. The Chi Square statistic indicates that the model does a reasonable job of predicting the observed outcomes. In all these models the constant is in for technical reasons. Thus for purposes of general exposition this variable can be ignored. Now to the interesting stuff. As we expected, the sign on the coefficient in column (a) for total cost of project is positive indicating that the larger the rural project--all other things being 27 equal--the more likely is an urban contractor to bid on that project and therefore the more likely is the prospect that an urban contractor will win that project. The model also reports that the more urbanized a state, the more likely will urban contractors venture across the line into suburban and rural school construction. However, this result is not statistically significant. The higher the urban wage relative to suburban and rural wages--the wider is this wage gap--the less likely is the urban contractor to win jobs in suburban and rural areas. This result is statistically significant. Finally, controlling for these factors, the presence of a prevailing wage law discourages urban contractors from working in rural areas but again this is not statistically significant. In short, we can say that the bigger the project the more likely an urban contractor will venture forth into outlying areas but the wider the wage gap between higher paying urban areas and lower paying outlying areas, the less likely will the urban contractor move out into these areas. Prevailing wage laws do not appear to strongly affect this dynamic. The dynamic determining whether a rural contractor comes into the city is somewhat different. Not surprisingly, the wider the wage gap between the rural and urban area, the more likely will the rural/suburban contractor win jobs in the urban area. The more densely packed the economy as measured by urbanization, the more likely will the suburban/rural contractor cross the line into urban school construction. These are both expected and statistically significant results. Now for some surprises. First of all, rural contractors are more likely to take on urban school construction to the extent that the urban project is small (not large). This result is statistically significant and is contrary to the expectations we had when we built the model. Why would it be that urban contractors go after larger projects in rural areas but that suburban/rural contractors go after smaller projects in urban areas? Our data include both new school construction, additions, renovations and repairs as long as the total value of the project exceeded $750,000. The average value of a project was over $4 million. We suspect that suburban/rural general contractors competing in urban areas are at a disadvantage if the project is large or technical. They may have a less skilled work force or support staff. Consequently, they aim at smaller, less demanding projects. Conversely, the city contractors may be best positioned in suburban/rural markets when the project is large and technically demanding. Thus, the result we obtained may not be as surprising as it seems. But prevailing wage laws (all other things being equal) encourage the use of suburban/rural contractors in urban areas. This is told from the positive sign to the coefficient for prevailing wage laws in column (c) and this result is statistically significant. Now this is a surprise. Urban areas have higher wage rates. In Virginia and North Carolina there are no regulations requiring that a suburban/rural contractor pay those higher wage rates. But in Maryland, a suburban/rural contractor working in Baltimore must pay those higher rates. This is also true in Pennsylvania and Delaware. (West Virginia does not have a major metropolitan area.) Why would forcing suburban/rural contractors into 28 paying urban wage rates facilitate suburban/rural contractors winning urban jobs? We can think of two possible explanations. First, this result may be an artifact of the age and structure of cities in the North versus the South. The Southern cities in our sample are conglomerate cities--Greensboro, Winston Salem, High Point--Raleigh, Durham, Chapel Hill. Our Northern cities (Baltimore, Pittsburgh, Philadelphia) are not sprawling conglomerates. Furthermore, Table 10 shows that a greater proportion of all school construction in the South is occurring in the major metropolitan areas compared to the metropolitan areas in the North. Perhaps'in the Northern city contractors have moved their offices to the suburbs where more work is occurring and yet these contractors maintain their city operations as well. Perhaps in the South the institutional gap between city and suburb-rural areas has not been bridged by the movement of city contractors outside the metropolitan areas. A related explanation involves unionization. Construction is more unionized in the North. Union contractors move from area to area but hire locally and pre-determined locally bargained wages. Collectively bargained contracts and the provision of local labor from hiring halls reduces the uncertainty of ramping up a job in a distant area. Perhaps the positive effect of prevailing wage regulations on the probability of a suburban/rural contractor obtaining an urban school project is capturing the effect of collective bargaining reducing the cost of bidding and working at a distance. Further research is required to test these and possibly other explanations for this result. Models 1 and 2 presented in columns (e) through (h) examine the factors that influence the hiring of out-of-state contractors. Model 1 includes all the explanatory variables except the wage gap between construction workers within the state compared to construction workers from the state of the out-of-state contractor.3 The cross-state wage is eliminated from model 1 simply because its effect shown in model 2 is so strong that it is helpful to see what the model shows without including this variable. Model 1 shows what we expected. The larger the project, the more likely contractor. In more economically dense urbanized it is to go to an out-of-state _ . I is the 'ob to o to an out-of-state contractor. This is states, the more likely � g capturing the effect of the urban corridor running from New York City to Washington DC on the use of out-of-state contractors. Finally, prevailing wage laws discourage the use of out-of-state contractors. All of these are statistically significant results. But the statistical significance of these results disappear when we add the wage of the state compared to the wage of the state from which the I outside contractor comes from. This variable swamps the measurable effects of 3 Data on state wages come from the 1992 Census of Construction. Data on metropolitan wages are from the 1994 BLS Occupational Wage Survey. By normalizing on wage rates in each year relative to the US average for that year and then comparing them,the inflationary differences between the years is eliminated. 29 all other variables, and more surprisingly it says that the closer the wage is between the two states, the more likely will an outside contractor be used. What is going on here? We think this captures a proximity effect. Namely, New York contractors are working in Pennsylvania and Eastern Tennessee contractors are working in Western North Carolina. For the most part, contractors cross state lines where those lines are close at hand and consequently the wage differences will be minimal. In model 2, the signs of the other variables have not changed compared to model 1, but their statistical significance has fallen away. We believe that the tentative conclusion should be that prevailing wages do discourage the use of out-of-state contractors but to be confident about that conclusion more research is necessary. Conclusion Most construction work is done by local contractors. Less than 10% of all school projects valued above $750,000 in the six states under study are done by general contractors from outside those states. Within states, 80% of all school projects are done by contractors that come from the urban or suburban/rural sector within which the job was let. Within the suburban/rural sector of each state, general contractors move farthest in states that are larger. With only six states under study, prevailing wage regulations do not appear to effect the distances over which suburban/rural contractors look for work. Do prevailing wage laws discourage the use of outside contractors? Multivariate logistic regression analysis provides only tentative answers to this question. Across state lines we believe the answer is yes--prevailing wage laws discourage out-of-state general contractors. But more research is required to be confident in this answer. Prevailing wage laws may discourage urban contractors from working in rural areas and conversely, may encourage suburban/rural contractors working in urban areas. But these results also are fragile and require further research. Our data were limited to information on general contractors. The relationship of prevailing wage laws to the movement of specialty contractors may differ to some extent. 30 The Impact of Prevailing Wage Laws on Wages in the Construction Industry Introduction In the preceding portions of this report I have shown that prevailing wage requirements do not have the dire negative consequences on either school construction costs or the competitive environment that many of the laws opponents espouse. The question to which I now turn considers one possible positive consequence of prevailing wage laws, namely, the extent to which they serve their intention of promoting the path of high wage, high skill development within the construction industry. In this chapter, I analyze construction wages in prevailing wage and non-prevailing wage environments to determine whether they are significantly higher in prevailing wage jurisdictions compared to non- prevailing wage jurisdictions. In thinking about this question, economists tend to employ a simple labor market model in which prevailing wage regulations impose a minimum wage above the equilibrium wage that would exist if supply and demand were left unfettered. Consequently, prevailing wages are assumed to always be above the competitive equilibrium wage in a local labor market. This conception of the issue, however, ignores the institutional detail of how prevailing wages are determined. The idea underlying the administration of nearly all prevailing wage laws is to protect local labor market standards. In other words, the determination of what contractors must pay as the prevailing wage is based on existing local labor market conditions. Of course, the devil is in the details. Workers within a single trade may not always receive exactly the same wage; wage rates will different depending on a number of factors. These include unionization, seniority, and differences in certification or training, urbanization and possibly others. In the state of Maryland, prevailing wages are determined by the Prevailing Wage Unit of the Department of Labor according to the following 31 formula. If 50 percent or more of all workers in a trade are paid exactly the same rate, that rate is considered the prevailing wage. If less than 50 percent are paid the same rate, then the prevailing wage is the wage paid to 40 percent or more of the workers within a trade. If less than 40 percent receive the same rate, then the prevailing wage is determined as the weighted average of the wage rates received by workers within a trade. Prevailing wage determinations are made for each county within Maryland, and Baltimore City. Especially in cases where the weighted average method for determining the prevailing wage is used, the prevailing wage may not differ significantly from what economists imagine to be the "equilibrium" wage. Consequently, prevailing wage laws, far from increasing wage rates, may simply reinforce existing labor market conditions. On the other hand, in areas with concentrations of unionized construction workers high enough for the 40 percent rule to kick in, there can be a significant difference between the prevailing wage and the nonunion wage. In the final analysis, whether prevailing wage laws inflate wage rates is an open question subject to empirical verification. Empirical Analysis of Construction Industry Wages Data for such an analysis of construction industry earnings is available from the U.S. Bureau of the Census through their County Business Patterns series. These data report number of employees and annual payroll by two digit Standard Industrial Classification (SIC) code. County Business Patterns reports on the construction industry in general and also disaggregates the industry into general contractors, heavy and highway construction, and specialty contractors. A limitation of these data is that workers in school construction cannot be distinguished from workers in other market segments. Consequently, it is not possible to draw any direct inference about the impact that the inclusion or exclusion of school construction from prevailing wage requirements might have on construction workers' wages. Similar data are also available through the Census Bureau on a statewide level. These data can be used to construct an index of relative earnings in the construction industry for prevailing wage and non-prevailing wage areas. In Table 13, 1 present results of a statewide comparison of relative earnings levels for the years 1993 through 1996, the most recent year for which data are available. The states included are the Mid-Atlantic states used throughout this study. For the purposes of this analysis Maryland was excluded from the analysis. The prevailing wage states used here include Delaware, Pennsylvania and West Virginia. The "no law" states include North Carolina and Virginia. The relative earnings index was calculated by dividing the average annual earnings per employee in the construction industry by the average annual earnings per employee in all nonagricultural industries. A comparison of the relative earnings indexes indicates that for the construction industry as a whole 32 (SIC 15), the relative earnings of employees is higher in prevailing wage law states than in non-prevailing wage law states. Earnings in non prevailing wage law states were approximately equal to average nonagricultural earnings. In prevailing wage law states, construction workers earned a 9 to 15 percent premium over average nonagricultural employees. These results are consistent with the proposition that prevailing wage laws tend to raise the wages of workers in the construction industry. Table 13: Earnings of All Construction Industry Workers Relative to All Non- Agricultural Employees by Prevailing Wage and No Prevailing Wage Law States Year Prevailing Wage Law No Prevailing Wage Law 1993 1.12 1.01 1994 1.15 1.02 1995 1.09 1.00 1996 1.11 1.01 Source: U.S. Bureau of the Census The Census Bureau's County Business Patterns reports number of employees and payroll by SIC code for the years 1988 to 1996. In Table 14, 1 use these data to calculate the relative earnings of construction workers to all nonagricultural workers with Maryland. For the state as a whole, construction workers earn premiums similar to the premiums earned in other prevailing wage law states, namely from 7 to 16 percent more than nonagricultural employees. Table 14: Relative Earnings of Construction Workers in Maryland Year Relative Earnings Index 1988 1.16 1989 1.14 1990 1.14 1991 1.14 1992 1.11 1993 1.11 1994 1.13 1995 1.07 1996 1.08 Source: County Business Patterns, 1988-1996 33 isDisaggregating the data for Maryland by county allows us to examine the impact f school construction b of the differential coverage o Y prevailing wage laws. Counties were characterized as having a prevailing wage law or not depending on the existence of a local statute covering school construction for that jurisdiction. This is an imperfect test because, as mentioned earlier, it is not possible to identify only those workers employed in school construction. Consequently, the impact of prevailing wage laws covering school construction could be overwhelmed by Maryland's general prevailing wage law covering public construction. Nevertheless, it appears that in most years, those jurisdictions having prevailing wage laws covering schools have higher relative earnings for construction workers, as shown in Table 15. Table 15: Relative Earnings of Construction Workers to All Non-Agricultural Workers, by School Prevailing Wage Law Jurisdiction, 1988-1996 Year Prevailing Wage e Law No Prevailing Wage Law 1988 1.38 1.14 1989 1.08 1.14 1990 1.22 1.13 1991 1.33 1.12 1992 1.17 1.11 1993 1.19 1.11 1994 1.12 1.13 1995 1.06 1.08 1996 1.15 1.07 Source: County Business Patterns, 1988-1996 A statistical test of the difference between the average relative earnings index in prevailing wage law jurisdictions and non-prevailing wage law jurisdictions indicates that, overall, earnings are higher in prevailing wage law jurisdictions. One might be tempted to conclude that, since Baltimore City is a prevailing wage jurisdiction that wages high es in this urban center are pulling the average up. But 9 this would not be an appropriate inference because it ignores the fact that construction worker wages in Baltimore are being compared to the earnings of other workers in Baltimore. What this test, and all of the numbers presented throughout this chapter, indicates is that construction industry earnings do appear to be higher in areas with prevailing wage laws. This conclusion is consistent with one of the original intents of prevailing wage laws, namely, to promote a path of high wage economic development. 34 A l on Of 1c cam����� costs In Three Midwestern States that Have Changed Theirg Prevailing Wage Laws in the 1990s Kentucky, Ohio and Michigan BY p Peter Philips,i s Ph.D. Professor of Economics, University of Utah 00 0 Introduction Proponents and critics of prevailing wage regulations have debated the merits of these regulations for some time. Proponents argue that these regulations promote the development of a skilled labor force in construction, improve work place safety, encourage quality construction, increase apprenticeship training and provide career opportunities in construction for local citizens. Proponents emphasize that prevailing wage regulations also induce contractors to provide health insurance and pension coverage that otherwise would be absent. Critics of prevailing wage regulations concede some of the foregoing positions and contest others. But the main argument of critics of prevailing wage regulations is the contention that these laws raise public construction costs. This is a two-sided argument asserting that when prevailing wage regulations are applied they raise public construction costs and when these regulations are eliminated, public construction costs will go down. The magnitude of savings is thought to be substantial ranging anywhere from 10% to 30% or more of total construction costs. This paper focuses on the specific question of whether or not the application of prevailing wage regulations raises costs, and if so, by how much. Those favoring this view theorize that prevailing wage regulations raise wage rates on public construction to higher levels than they otherwise would be. Increased wage rates should lead to increased construction costs that would be passed on to the government and eventually the taxpayer. Proponents of prevailing wage regulations counter that higher wage rates induce contractors to hire or train a more skilled and productive labor force. Higher wage rates also will encourage contractors to better manage their v�orkers and provide them with better and more up-to-date equipment. These responses to higher wage rates may, according to this view, offset some or all of the costs of higher wage rates. Prevailing wage proponents also argue that a more skilled labor force leads to better quality construction that reduces downstream maintenance and repair costs. This paper tests these competing hypotheses regarding the cost-effects of prevailing wage laws. The focus will be on new public school construction in Kentucky, Ohio and Michigan over the period 1991 to 2000. These states and 2 this time period were chosen because legislative and judicial changes in these states over this period form a natural experiment helpful in isolating the effects of these regulations on costs. In the 1990s, prior to July1996, Kentucky did not apply prevailing wage regulations to public school construction. Starting in July of 1996, the state's prevailing wage law regulated public school construction. By itself, this provides some before-and-after information about the effects of applying prevailing wages to school construction costs. Fortuitously, from the standpoint of science, Ohio in the 1990s did almost the opposite of Kentucky. Throughout the 1990s until July of 1997, Ohio applied its prevailing wage law to public school construction. Starting in July 1997, Ohio exempted public school construction from prevailing wage regulations. Thus, almost simultaneously, these neighboring states moved in opposite regulatory directions. The fact that Ohio lifted its law soon after Kentucky applied its law to public school construction helps this natural experiment isolate the effects of regulatory policy from other factors that change over time. From an experimenter's perspective, this is nice. But nicer yet, at around the same time, Michigan does yet a third thing with its prevailing wage law. At the end of 1994, a judicial ruling suspended the application of Michigan's prevailing wage law to any public construction. This judicial suspension lasted until July of 1997 when a higher court ruling reapplied Michigan's prevailing wage law to all public construction, including schools. So Michigan suspended its law two-and-one-half years before Ohio, and Michigan reapplied its law in precisely the same month Ohio exempted schools from prevailing wage regulations. Figure 1 shows the variation r g a � on m prevailing wage policies by state in the 1990s. Kentucky I�t€3 ate Ohio No Law Michigan NEi; Lave 1991 1993 1995 1997 1999 2001 -Figure 1: Prevailing Wage Policy by State 1991-2000 3 With these variations in legal policies in hand, we are in a position to assess statistically whether or not changes in prevailing wage policies as they applied to public school construction raised or lowered the cost of building public schools. The Data The FW Dodge Corporation is a private company that provides bidding information to contractors. In so doing, the Dodge Corporation systematically gathers information on the "start" cost of construction projects. "Start" costs are the agreed-upon bid price of a project at the outset of a bid. The final cost of a project can vary from the start cost based on cost overruns. Proponents of prevailing wage regulations argue that one of the advantages of prevailing wage laws is that they reduce cost overruns. This argument asserts that absent prevailing wage regulations a cutthroat bidding system emerges where low-ball bidders undercut their competitors with unrealistically low bids in the hope and expectation that during the term of the project the contractor can recoup his profits through change orders. This argument asserts that prevailing wage regulations attract a set of bidders that will compete with each other over on-time completion, quality construction and productivity but eschew the strategy of low- ball bidding hooked into profiting from change orders. We cannot test this argument with Dodge data because it does not include cost overruns subsequent to the acceptance of the bid. But with this one limitation in mind, the Dodge data form the single best source on school construction costs across states. In addition to providing the accepted bid price, the dodge data indicate what kind of building project it is; what the total square feet of the project is; where the project will take place; when the bid was accepted; some details on the nature of the project, and other useful information. This report uses data from 1991 t. We focus on new public school construction only. By eliminating renovation alterations and additions, we can focus on a relatively homogenous group of buildings—new public schools. We will ask the question—controlling for the size of these public schools, and where they took place, and when they were built, and whether they included a gymnasium- swimming pool facility—did the presence or the absence of prevailing wage regulations affect the total cost of the project? Table 2 describes the new schools used in this study. 4 Characteristic of Schools in Stud Number of New Schools in Study 391 Average Square Foot Size of the School 86,415 Average Total Cost of the Project Year 2000 dollars $8 483 937 Percent of All Schools Michigan 38% Ohio 36% Kentucky 26%0 Percent of School with a Gym-Pool Facility 7% Percent of Urban Schools 32% Percent of Schools Built Under Prevailing Wages 49% •Table 1:Description of the new schools used in the study This study involves 391 new public schools built between 1991 and September 2000. Table 1 show that the average size of a school was 86,415 feet and the average total cost was $8,483,937 in year 2000 dollars. The consumer price index urban (CPl-U)was used to update earlier costs into year 2000 dollars. Of the 391 new public schools in the study, Michigan accounted for 38%. Ohio accounted for 36% and Kentucky accounted for 26%. This reflects the relative size of the three states. Dodge data indicated that 7% of the new schools included a swimming pool/gymnasium facility. In our statistical model we expect that even controlling for the size of the school project, the inclusion of such a facility is likely to raise the square foot cost of the new school. Thirty-two percent of the new public schools were built in urban areas with the remaining 68% built in rural areas including smaller towns. Almost If, °ha 49/° of theprojects wer uil e built with prevailing wages while the remaining 51%were built without these regulations. Table 2 shows the distribution of when the new public schools were built. Column c shows that 1991 had unusually few new public schools built, and the year 2000 had unusually many schools built. This reflects the two ends of the business cycle and building cycle. The year 1991 was a recession year and the economies of these states grew progressively from then through the 1990s. One result of this growth has been an expansion of school construction. 5 Percent of Each Percent of Year's New Number Decade's Total Schools Built of New New Schools Built Under Prevailing Year Schools in Each Year Wage Laws a b c d 1991 5 1% 20% 1992 44 11% 61% 1993 28 7% 68% 1994 10 3% 50% 1995 39 10% 33% 1996 49 13% 37% 1997 53 14% 49% 1998 33 8% 58% 1999 56 14% 71% 2000 74 9%1 30% Total 391 1 49% •Table 2:The time distribution of new school construction within the study Table 2 also shows the annual percent of each year's new school construction that was done using prevailing wages. The balance tends to be stable over time. The dip in 1995 and 1996 is due to Michigan suspending its law in those years. (Michigan split construction in 1997 with the first half not using prevailing wages and the second half using this regulation.) In the year 2000, Ohio has a major increase in school construction, all done absent prevailing wages. This accounts for the somewhat lower percentage of new school built with prevailing wages in that year. 6 Median Cost per Square Foot of New Elementary Schools 120 100 80 ntuck 0 C Ohio LL 5 N 60 w `Period IL v Kentucky Has Law Covering 40 Period I: Ohio Has Law Covering 20 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 •Figure 2: Median Cost of New Public Elementary Schools in Kentucky and Ohio by Year Figure 2 gives a general sense of the experiment we are going to perform. In Figure 2, the median square foot cost to build a new elementary school in Kentucky and Ohio are compared for the years 1992 to 2000. These data are not deflated. They are in the actual dollars reported at the time. Consequently, it is not surprising that the median cost rises with time. In our statistical model,we will deflate these prices using the Consumer Price Index and ask the question whether controlling for inflation, new school construction costs are still going up. We will also control for a variety of other factors that cannot be controlled for in a simple figure such as the one above. Our additional controls include a control for possible economies of scale associated with larger schools, a control for the differences between urban and rural construction, a control for fancy facilities, a control for tight construction markets possibly pushing up the real cost of construction. With these controls in place, we will look at variations in prevailing wage policies. In Figure 2, changes in prevailing wage policies are shown by two vertical lines, one in 1996 representing Kentucky's application of a prevailing wage regulation. And a second vertical line in 1997 representing Ohio's exemption of public schools from prevailing wages. If these laws had an effect on costs, one might expect to see a change in the general relationship between 7 median new public elementary school construction costs in the two states. Visually, this does not seem to appear. We will go beyond visual inspection to examine all 391 new schools and test whether or not, controlling for other factors, changes in prevailing wage regulations have made a difference. Comparison of Mean Square Foot Cost We begin our analysis with a simple comparison of the average or mean inflation- adjusted square foot cost of building a new school. The 391 new schools are broken down into those built in urban areas (126 schools) and those built in rural areas (265 schools). Urban areas include the areas around Cleveland, Columbus, Cincinnati, Dayton, Louisville, Lexington, Detroit, Flint, Grand Rapids and Lansing. Within each group of urban and rural schools,the schools are broken down into those built under prevailing wage regulations and those built without prevailing wages. Table 4 shows the mean, standard deviation' and number of schools in each of four categories: 1) rural schools built without prevailing wages; 2) rural schools built with prevailing wages; 3) urban schools built without prevailing wages; and 4) urban schools built with prevailing wages. New Public Schools Real Inflation Adjusted Square Foot Cost a b c d e 1 1 Urban Schools 2 Mean Standard Deviation Number 3 No Law $114 36 40 4 Law $114 34 86 5 West 0.05 Statistically lh Significant NO g Difference? -Table 3: Comparison of the Real(Inflation-Adjusted)Square Foot Cost of New Public Schools by Urban and Rural Schools and Built without or with Prevailing Wages ' The standard deviation is, in essence the wiggle around an average. So for instance, if you had 5 children in a carpool ages 8, 9, 10, 11 and 12,the mean (or average)would be 10 years of age, and the standard deviation or wiggle around the mean would be 1.4 years. 8 The comparison of mean real square foot costs can be seen in rows 3 and 4, columns b and e. Considering rural schools first, the average or mean real square foot cost of schools built without prevailing wages was$96 per square foot while the mean real square foot cost of schools built with prevailing wages was $98 per square foot. There were 161 new rural schools built without prevailing wages in the sample and 104 new rural schools built using prevailing wages in this sample. (See column d, rows 3 and 4). The standard deviation is a statistical measure of the wiggle around each mean. It is used to construct a statistical test of whether or not the $2 difference in the average cost of construction per square foot is statistically significant. The statistical test is called a Mest. Typically, for there to be statistical significance, the t statistic must be around plus or minus 2. In this case for rural schools where the difference is $2 per square foot, the t- statistic shown in column b, row 5 is -.75. What this means is, statistically speaking, there is no difference between the average square foot cost found for rural schools built with prevailing wages compared to the average square foot cost of schools built without prevailing wages. Considering schools built in urban areas, in the sample, 126 new schools were built in urban areas with 40 being built without prevailing wages and 86 being built with prevailing wages. The average or mean real (inflation-adjusted) square foot cost of urban schools built with and without prevailing wages was almost equal. Indeed, rounding to whole dollars, they are equal at$114 per square foot (in year 2000 dollars)2 Again, the Rest indicates that any minor difference in these two means(in this case 34 cents) is not statistically significant. Another way of putting this is: the difference in average real square foot construction cost for new public schools is due to random differences and statistically, the averages are equivalent. This conclusion holds both when comparing urban schools and when comparing rural schools built with and without prevailing wages. A Statistical Model of New Public School Construction Costs Table 4 presents the results of a statistical model of new public school construction costs. The model is called an ordinary least squares linear regression model. This type of statistical model is very commonly used by economists, epidemiologists and others studying social phenomena. The particular model in Table 4 uses the 391 new public schools built in Kentucky, Ohio and Michigan over the 1991 to 2000 period as data to help predict the effects of various factors on total new construction costs. The focus variable in 2 If you do not round to whole dollars, the mean for schools built without prevailing wages was$114.17 and with prevailing wages it was$113.83. 9 the equation is the last variable in the model shown in gray on line 12. But before we get to this issue, let us examine the other aspects of the model. t- Significance Statistically Model Coefficient statistic level Significant? 1 a b c d e f 2 (Constant) 4.45 16.05 0% Yes 3 Size Natural Log of the Total Square Feet of the Project 1.00 41.59 0% Yes 4 Business Boom Time(in years) 2.9% 5.58 0% Yes 5 Location School was built in Ohio -12.6% -3.70 0% Yes 6 School was built in Kentucky -14.6% -4.03 0% Yes 7 School was built in an urban area 10.5% 3.41 0% Yes 8 Special Facilities School had a gym/pool facility 9.2% 1.69 9% Yes 9 Timing School was started Winter quarter -5.6% -1.21 1%. No 10 School was started Spring quarter -10.9% -2.75 Yes 11 School was started Summer quarter -2.7% -0.63 No ,art. "S 13 Total Cost of Natural log of real (inflation adjusted)total start cost of each new school (in year 2000 dollars) School 14 Model statistics Adjusted R-s uare a goodness of fit statistic = .85 15 Number of new schools in the sample=391 .Table 4:A Linear Regression Model of the Total Square Foot Cost of Building New Public Schools in Kentucky, Ohio and Michigan Focusing on the Effect of Prevailing Wage Regulations The first key factor in the model is the size of each of the 391 new schools. The coefficient of 1.00 says that as the square foot size of the school increases, the total cost of the school increases proportionately.3 The second variable is simply time measured in years. This variable captures the fact that building costs have been rising faster than inflation in the 1990s. The cost data in the model are inflation adjusted using the Consumer Price Index. The time variable indicates that after adjusting for inflation, new public school construction costs in these three states have been rising at 2.9% per year from 1991 to 2000. This result is statistically significant. The reason building costs have been rising faster than inflation is because the economic boom has led to a very vigorous boom in building leading to heavy demand for construction services. 3 In other studies, I have found in some cases that there were economies of scale in school construction costs. Namely,as school size increased,total cost went up but more slowly than total size went up. The absence of economies of scale among these 391 schools may be due to the relative homogeneity of the buildings in the sample 10 Rows 5, 6 and 7 in the model present three variables that control for where the school was built. Row 7 indicates whether or not the school was built in an urban area. Urban schools cost 10.5% more than rural schools controlling for other factors such as the size of the school. Rows 5 and 6 indicate whether or not the school was built in Ohio (row 5) or Kentucky (row 6). In the case of urban schools, the reference point was rural schools. In the case of Kentucky and Ohio, the reference point is Michigan. The model indicates that new Ohio schools cost 12.6% less than Michigan schools while new Kentucky public schools cost 14.6% less than Michigan new public schools. These results, too, are based on controlling for other factors such as the size of the school, when the school was constructed, whether or not the school was urban or rural and whether or not the school was built under prevailing wage mandates. Rows 9, 10 and 11 indicate in what season the school was started. Row 9 indicates schools started in the winter quarter(January, February or March). Row 10 indicates schools started in the spring. And row 11 indicates schools started in the summer. In all three of these cases, the reference point is schools started in the fall. These seasonal variables get at the question of whether breaking ground on a new school in the face of winter weather raises the cost of building that school. The model indicates that schools started in the winter were 5.6% cheaper than schools started in the fall. But this result is not statistically significant. The lack of statistical significance means that you cannot be sure there really is any difference in the total cost of schools started in the winter compared to the fall. In the case of schools started in the spring, they were 10.9% cheaper than schools started in the fall (again controlling for other factors such as the size of the school, whether it was an urban or rural school, etc.). In this case, the results of the model are statistically significant. That is, this statistical model indicates that you can be confident that breaking ground in the spring will lead to lower cost new school construction compared to breaking ground in the fall 4 Breaking ground in the summer is estimated to be 2.7% cheaper than breaking ground in the fall, but this is not a statistically significant difference. The point of these results, however, is clear. Don't break ground into the teeth of winter. It will cost you. Indeed, the model shows that by starting in the spring instead of the winter saved enough money for each school to include a gymnasium/pool facility in its specifications. The 7% of all schools that had such facilities paid 9.2% more total costs, controlling for other factors. And compared to starting in the fall, starting in the spring would have offset the cost of a swimming pool. 4 In most cases in the FW Dodge data,the start date is the date of bid acceptance. Given that there will be some lag between bid acceptance and ground breaking,fall probably means late fall and spring probably means late spring. Also,given the probable cause of increased costs associated with breaking ground in the late fall is weather,the effect of this seasonal factor is probably stronger in the colder areas within the sample of Michigan, Ohio and Kentucky. 11 Figure 3 shows that for the most part, school boards and contractors know this. Fully 40% of all new public schools in Kentucky, Ohio and Michigan started in the years 1991 to 2000 were started in the spring. Fall starts--the most expensive start time—accounted for only 16% of all starts. However, whatever drove schools boards to start schools in the fall compared to the spring also led them to pay a 10% premium for this choice of when to break ground. Distribution of New School Starts by Season fall winter 16% 17% winter ®spring El summer ummer 27% fall spring 40% •Figure 3: Percent Distribution of New School Starts by Season for Kentucky, Ohio and Michigan, 1991-2000 The Effect of Prevailing Wages on Costs Controlling for seasonality, controlling for differences in rural and urban construction costs, controlling for the size of the project and controlling for the state in which the project was built, the model estimates that using prevailing wage regulations raised school construction costs by 7/101'of 1%. But again,this is not a statistically significant result. In effect, the model says there is no effect on total costs associated with prevailing wages. How can this be? Prevailing wage rates insure that all contractors must pay the wage rates that prevail for an occupation in an area. Without this regulation, 12 contractors are free to pay whatever they want(or the market will allow). The fact is that prevailing wage regulations induce contractors to hire a more skilled labor force and equip them with better, more up-to-date, tools, materials and equipment. It also induces management to compete over better management strategies and techniques. Thus, the higher wage rates are offset to a large extent by higher skilled, better equipped, and better managed workers. It may be, however, with more observations, we would find that the 0.7% higher cost associated with prevailing wage regulations would turn out to be statistically significant. And, in a market where the government is obliged to accept the lowest bidder regardless of the reputation or history of the contractor, that 0.7% difference could lead to an entire changeover in the contractors doing business in building schools. But we must remember that this model is based on start costs—accepted bid price. The ultimate cost of a new school includes cost overruns and the downstream cost of maintenance. The potential 0.7% savings may be an offer for school boards to become penny wise and pound foolish. Conclusion A simple comparison of the mean (or average) inflation-adjusted square foot cost of building 391 new public schools in Kentucky, Ohio and Michigan broken down by urban and rural schools finds no statistically significant difference between those public schools built with prevailing wages and those public schools built without this regulation. A more complex statistical model that estimates new public school construction costs based on the size of the project, whether it was an urban or rural school, which state built the school, and at what time of the year the school was built again finds no statistically significant effect on total new school construction costs associated with whether or not the school was built with prevailing wages. While net effect of prevailing wage regulations is apparent, school boards can save 10% on new school construction costs by starting in the spring and not breaking ground in the face of winter weather. While 40% of all new schools do start in the spring, 16% of new schools had to pay this 10% penalty by starting as winter approached. The data used in this study come from FW Dodge reports and show the start cost—or accepted bid price—of the new school. Final cost of new public school include cost overruns and downstream maintenance costs. The hgher wage rates required by prevailing wage regulations insure that all contractors bidding on the job will use skilled labor when building the school. If you have to pay for the high-priced spread, you might as well buy it. Thus, prevailing wage regulations offer school boards some assurance that the project will be skillfully built and workers on the job will be carefully managed. Consequently, prevailing wage regulations provide some assurance against cost overruns and downstream maintenance costs. 13 Peter Philips grew up in Compton and Pomona, California. He received his B.A.from Pomona College in 1970 where he majored in economics and received the Leland Backstrand Graduating Senior Award in Economics. Philips received his M.A. in economics (1976) and his PhD in applied economics (1980) from Stanford University. Philips is a Professor of Economics at the University of Utah. He is co-editor of Three Worlds of Labor Economics (M.E. Sharpe, 1986) and coauthor of Portable Pensions for Casual Labor Markets:the Central Pension Fund of the Operating Engineers (Quorum Books, 1995). Philips has published widely on the canning and construction industries in journals such as the Journal of Education Finance, Industrial and Labor Relations Review, Industrial Relations, Business History, the Journal of Economic History, Historical Methods, The Journal of Economic Literature, Oxford Encyclopedia of Economic History and the Cambridge Journal of Economics. Philips has been a consultant for the U.S. Labor Department analyzing the supply of cannery labor in California, and he has worked as an expert on the Davis-Bacon Act for the U.S.Justice Department. The Davis-Bacon Act regulates wage payments to construction workers on federal public works. Philips is a respected expert on prevailing wage laws and on employment, training wages and benefits in the construction industry. He has testified before state legislative committees in Ohio, Indiana, Kansas, Oklahoma, New Mexico and California on their state prevailing wage laws. Along with other researchers at the University of Utah, Philips has analyzed the effects of prevailing wage laws on public construction costs, construction worker incomes, apprenticeship training,worker safety and minority access to construction work. Philips is the senior labor economist at the University of Utah. He teaches a wide range of courses in the area of labor economics, econometrics, labor law, collective bargaining and economic history. Philips has received awards for his teaching and community service, including University of Utah Public Service Professorship, the University of Utah Presidential Teaching Scholar Award and the University of Utah, College of Social and Behavior Science Superior Teacher Award. Philips is married with two children. 14 Appendoc Simple Comparison of One Year Before and One Year After Legal Change in Kentucky, Ohio and Michigan The following is a simple comparison of the average or mean square foot cost of new public schools one year before compared to one year after a change in the prevailing wage law in the three states—Kentucky, Ohio and Michigan. Kentucky applies prevailing wage regulations to public school construction in July of 1996. Ohio exempts public schools from prevailing wage regulations in July of 1997. A Michigan court suspended the application of prevailing wage regulations in the state in December, 1994. A second court reapplied prevailing wage regulations in July of 1997. These variations allow for four simple 12-month before-and-after comparisons-- one for Kentucky, one for Ohio and two for Michigan. In this simple comparison there is no adjustment for inflation and no statistical tests for equality of means. All that is presented is the raw averages and the number of new public schools that comprise each average. Average Square Foot Cost Before #Schools Legal Change Cost Change Kentucky:No-Law to Law $86 17 Enactment $0 Ohio: Law to No-Law 77 7 Repeal 13 Michigan: Law to No-Law $83 14 Suspension $11 Michigan: No-Law to Law $101 40 Resum tion $6 •A 1: Simple Comparison of Mean Square Foot Cost of New Public Schools for the 12 Months Before and After a Policy Change 15 Table Al shows that for Kentucky, the mean square foot cost of 17 schools built in the 12 months prior to the application of prevailing wage regulation to public school construction equaled the mean square foot cost of 15 new public schools built in the 12 months after the application of prevailing wages. In Ohio, there was a $13 increase in the mean square foot cost of new school construction subsequent to the repeal of the application of prevailing wages to public school construction. In the case of Michigan, when the law was suspended, the mean square foot cost of new public school construction rose in the subsequent 12 months by $11 per square foot. When Michigan reapplied prevailing wages to public school construction, the mean square foot cost rose again by $6 per square foot. Other factors that might influence these changes include inflation, tightening construction markets, and the mix between urban and rural schools. These factors are controlled for in the econometric model presented in the main body of this report. 16 Journal of Education Finance Vol.24(Winter 2000)pp.415432 The Journal of Education Finance is published by the Association of School Business Officials International, 11401 North Shore Drive,Reston VA 20190-4232 (Posted with permission of the Association of School Business Officials Int.) PREVAILING WAGE REGULATIONS AND SCHOOL CONSTRUCTION COSTS: EVIDENCE FROM BRITISH COLUMBIA Cihan Bilginsoy Assistant Professor (801) 581-7691 bilginso@econ.sbs.utah.edu and Peter Philips Professor (801) 585-6465 philips@econ.sbs.utah.edu University of Utah Department of Economics 1645 E Central Campus Dr Front Salt Lake City UT 84112-9300 I PREVAILING WAGE REGULATIONS AND SCHOOL CONSTRUCTION COSTS: EVIDENCE FROM BRITISH COLUMBIA The stock of public school buildings constructed during the baby boom is aging along with that generation of Americans. Soon much of this building stock will have to be replaced.' The financing of this rebuilding of America's schools is an emergent political issue of considerable importance. Given these pressures on school construction financing,any proposal that promises to substantially lower the price tag for this reconstruction garners considerable public interest and potential support. One such proposal is the elimination of prevailing wage regulations that in 31 states and the District of Columbia govern the payment of wage rates on public school construction.2 Prevailing wage laws require that state mandated wage rates be paid on public road and building construction. The purpose of these laws is to encourage collective bargaining in construction and to discourage the payment on public works of wages below those prevailing locally. Critics of prevailing wage laws assert that these laws raise public construction costs and discourage nonunion contractors from bidding on public works. Some politicians and school officials believe that the elimination of prevailing wage regulations could cut public school construction costs by 10 to 40 percent. For instance,Gary Johnson,the Governor of New Mexico,in his 1997 address to the New Mexico state legislature foretold a 33 percent reduction in total state school construction costs from the elimination of the state's prevailing wage law.3 In Ohio in 1997 the state legislature held hearings on the possible elimination of prevailing wage regulations for school construction. School officials foresaw substantial cost savings from such an exemption. For instance,Rick Savors, Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415432 2 Deputy Director of Legislative Networks for the Ohio School Boards Association(OSBA) testified that The OSBA believes that the state's current prevailing wage law adds significant costs to a project and limits the number of contractors willing to offer bids on school contracts...The cost savings for districts,taxpayers and the state can be significant...those savings could reach perhaps as high as 30-40 percent in some instances.4 In the same hearings,Richard Maxwell,Deputy Executive Director for Government Relations for the Buckeye Association of School Administrators(BASA)testified: BASA supports the exclusion of the prevailing wage provision for school construction and facility improvements. This exclusion will benefit Ohio's schools by extending the scarce taxp aY er's dollars to improve facilities. As a superintendent of schools for 23 years in districts I managed many school buildings built in 1912 through 1964. We had an extensive program of renovation and improvement of these buildings...BASA members tell us they believe they can do 10-15 percent more construction and renovation if the prevailing wage provisions did not apply to school construction and renovations.5 The primary basis for asserting that the elimination of prevailing wage regulations on school construction will substantially lower total school construction costs is the proposition that absent these regulations,wage rates on school construction would be substantially lower. In consideration of Ohio Senate Bill 102,Jim Shirey,Legislative Liaison for the Athens Ohio City Schools Board of Education made this argument explicit. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415432 3 "Prevailing Wage"is nothing less than price fixing perpetrated by state government. n scale,regardless if the the rice of construction labor at uruo s Prevailing Wage sets , g g g p local labor market. One of the contractors who testified on this bill before the Senate Finance Committee did a study of actual construction costs and he showed very clearly that Prevailing Wage can inflate the cost of construction labor by as much as 60 percent.6 There are no academic studies on the relationship between prevailing wage regulations and school construction costs. This paper attempts to inform the ongoing debate over the impact of the prevailing wage laws on school construction costs by using a unique dataset from Canada.Wages including benefits constitute about 30 percent of the total construction cost, excluding land acquisition and architectural costs.' The province of British Columbia(B.C.) established the Skill and Fair Wage Development PolicY ( W SDFP)on March 30, 1992. It p g mandated payment of the prevailing wage on public construction projects and determined"fair" wage schedules for each occupation within the building trades.The fair wage was set at about 90 percent of the collectively bargained wage rate for each construction occupation in B.C. The B.C.experiment with prevailing wages provides a new opportunity to make an empirical assessment of the debated question of whether or not prevailing wage regulations measurably increase school construction costs.The objective of this paper is to compare final construction costs before and after the SDFWP was implemented in order to determine whether legislated I wages resulted in higher school construction cost in British Columbia. For this purpose we will g use final cost data from new elementary and secondary public school construction projects from six school districts in B.C.tendered between 1989 and 1995. The paper is organized as follows. Section 1 presents a brief review of the empirical studies on the impact of the prevailing wage laws on public construction,and states the contribution of this paper to that literature. Section 2 describes the data. Section 3 presents an empirical model to test for the impact of the SDFWP.The model is estimated and results are discussed in section 4. Section 5 concludes. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 4 1. The Literature In this section we will first outline the general features of the debate over the prevailing wage laws.In the last part of the section we will outline the specific findings concerning school construction. The literature on the construction cost impact of prevailing wage laws focuses almost exclusively on the U.S. federal Davis Bacon Act. With one exception,the Davis-Bacon literature estimates cost savings from the elimination of the federal regulation to range from I to 3 percent of total federal construction costs. The one exception estimates a 26 percent savings in rural building construction from the elimination of federal prevailing wage regulations. Because the Davis Bacon Act was passed in 1931,there is no possibility of comparing the cost of construction prior to and subsequent to the passage of this act. This leaves three avenues of research. The first estimates what wage rates would be paid on federal construction absent prevailing wage regulations and then tries to compute what total costs would be under two wage rate regimes. The second exploits a moment of time in 1971 when the Davis Bacon act was suspended for 34 days. The last compares the cost of construction on public works with comparable private construction projects. The first and most common method in the literature tries to assess what wage rates would be paid on federal construction absent Davis-Bacon. With this calculation in hand,these studies then attempt to estimate theoretically what construction costs would be under these hypothetical wage rates. Finally,these studies then attempt to compare hypothetical construction costs under hypothetical wage payments.As expected,this line of research is fraught with debates over the methodology. Most studies of this type conclude that the Davis-Bacon wage rate is biased upwards toward the union wage, and consequently predict that Davis Bacon raises the cost of construction.8 The estimated cost inflation attributable to this wage differential is in the order of 1.5 to 3 percent of public construction expenditure(Gujarati, 1967;GAO, 1979, 1983; is Goldfarb and Morrall, 1978, 1981; Gould, 1971;Gould and Bittlingmayer, 1980;GAO, 1979, Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415432 5 1983;Thieblot, 1986). On the other hand,Bourdon and Levitt(1980) found no such bias,and Allen(1983)argued that the Davis-Bacon effect is modest,raising construction costs merely by 0.3 to 0.4 percent. Thieblot(1975)adopted a second,more direct approach by taking advantage of President Nixon's temporary suspension of the Davis-Bacon in 1971.Thieblot compared bid prices of projects tendered but not contracted in this period with their rebid prices in the following period.He concluded that in the absence of Davis-Bacon the lowest bid were lower by 0.63 percent. Thieblot's re-examination of the data leads him to the conclusion that savings, including administration and wage costs,would be as much as 4.74 percent if the Davis-Bacon were repealed (Thieblot, 1986, 105-106). Fraundorf et al. (1983),followed a third methodology.They utilized multivariate methods to determine the impact of prevailing wage laws on construction costs directly. Fraundorf et al. compared 215 federal and private non-residential construction projects controlling for non-labor cost factors(such as the type of structure,technical characteristics, size,and geographic location).They found that federal projects were 26 percent more expensive than the private projects,and attributed this difference to the Federal Davis-Bacon Act. The Fraundorf result is substantially higher than previous studies and may be the academic basis for claims by school officials that total construction costs can be cut by 10 to 40 percent through the elimination of prevailing wage regulations. However,the Fraundorf study has two weaknesses that the present study overcomes. First,the Fraundorf study is not a study of school construction costs. It may be that school construction is distinct from rural building construction. Second,the Fraundorf study derives its projected cost savings from a comparison of public building costs to private building costs. If private buildings differ from public buildings in ways that are not adequately controlled for,this may conflate cost differences derived from public-private building differences with cost differences derived from prevailing wage regulations. Our study attempts to overcome these potentially confounding factors by focusing only on public school construction. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 6 Because state and provincial prevailing wage regulations differ in their specifications and implementations,and school construction is regulated at the state and provincial level, a variety of case studies are needed. Our study is limited to assessing the effect of one prevailing wage law in one province at one particular period of time. 2. The Data The recent experience of B.C.with the SDFWP permits an assessment of the impact of wage regulation on us the school construction cost by controlling for many factors that confounded the results of the studies mentioned above.The empirical analysis of this paper is based on tender data compiled by the Construction Labour Relations Association of B.C. from six school districts(out of 18)for a study of the impact of the SDFWP on construction costs (Government of Canada, 1997).These school districts are all located in the southern Lower mainland education region of B.C. and are geographically in close proximity.The data cover 54 new school construction projects tendered and completed between 1989 and 1995. Of the 54 tenders,25 were received before March 31, 1992 and 29 were received afterwards, allowing the comparison of costs before and after the establishment of the SDFWP. All projects cost more than$250,0009 and were therefore covered by the Fair Wage Policy during the period it was in effect.In addition to the final cost,the raw data also provide information on the list of bidders and bid prices for all projects,the contract price,the type of school(elementary or secondary),and structural characteristics of projects,including gross area,construction type, foundations,and special features(e.g.remote location,difficult site).For many projects estimated and/or actual durations of construction are available.Thus,while the number of observations is relatively small,the data have the advantage of being appropriately detailed.In addition,we also collected information on the size of the general contractors to whom the contract was awarded. Table 1 summarizes the distribution of school types across the two periods.The number and the type of schools are distributed approximately evenly across the two periods. Gross Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 7 sizes of elementary schools range from 2,017 to 3,950 square-meters,while secondary school gross sizes range from 6,033 to 15,787 square-meters. Table 2 reports summary statistics for the unit(square-meter)final cost(in 1989 prices). The average unit cost is$1,423 but there is a significant difference between the pre- and post- SDFWP periods.After the establishment of the Policy,the average unit final cost increases by $207,from$1,308 to$1,515.This 16 percent difference,which is statistically significant at the one percent level suggests that the SDFWP constitutes a serious burden on the public purse. In the next section,we will address this question in a multi variate context controlling for other potentially relevant factors that may impinge on the unit cost of construction. 3. The Empirical Model The cost of construction,by definition,is equal to the sum of the bid price and change orders. The unit bid price,in turn,is equal to the sum of the estimated cost of production and the profit margin.Let FP stand for the unit final price,m be the mark-up rate,and CO be the total change orders cost,the latter including the mark-up on change orders.10 Estimated total cost of construction,in turn,can be written as the sum of labor and non-labor costs.Letting w and n stand for the prices of labor and non-labor inputs, and L and N stand for the quantities of labor and non-labor inputs,the final cost can be written as: FP=(1+m)(wL+nN)+CO. (1) This equation is the basis of the empirical model to be used in predicting the cost of construction.Four sets of factors that potentially affect the independent variables are the physical features of the project,state of the construction business cycle and the degree of competition among contractors,regulatory environment in the labor market,and characteristics of the contractor. Physical features of the project(including the size,the type of foundation and frame, location,number of stories,type of school)influence the quantity and the type of inputs. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415432 8 Fraundorf et al.,controlled for technical characteristics such as foundation,frame,exterior walls,floor finishing,and frame type.In our sample there is little,if any,variation in technical characteristics for which information is available. Structural specifications of all projects were reported in the cost estimation forms prepared by the architect prior to bidding.Almost all schools are steel frame structures with slab on grade foundation.The type of school -- elementary vs. secondary -- summarizes most of the other technical differences. Secondary schools are larger than elementary and require approximately twice as long to build. If there exist economies of scale to size or longer construction period,second schools are expected g p �Y to have lower unit costs.Finally,secondary schools are more likely to include higher unit cost structures(e.g. laboratories)on the one hand,lower cost areas such as sports fields,on the other.'1 These considerations make it difficult to form an expectation on how the school type would affect the unit cost. In order to control for these technical features we introduce a dummy variable that takes the value of one for secondary schools and zero for the elementary.We also utilize dummies for school districts to control for any potential location effects. Input prices vary with the state of the economy,reflecting relative abundance of the labor, capital and intermediate goods. Hence the unit cost is expected to be pro-cyclical.The impact of the cycle on the mark-up is indeterminate. During the expansion of the economic activity,for instance,the firm may raise the profit margins taking advantage of rising demand.It is also conceivable that the firm shaves the mark-up in order to protect its market share from potential entrants attracted to the industry. In order to capture the construction industry business cycle effects we fitted a linear annual trend to the values of non-residential building permits(in constant prices)over the 1989-1995 period by ordinary least squares. We then measured the cyclical fluctuation in the construction sector activity as the percentage deviations of actual values of permits from the trend values.It should be bore in mind that this measurement of the cycle based on the annual experience of only seven year is crude. Quarterly data would probably yield a more precise profile of the cycles,but they are not yet available. The immediate effect of the SDFWP on the construction cost is via the wage rate.If the legislated wage is higher than the market wage,then labor costs increase. This anticipation is the basis of most of the empirical studies on the cost effect of the prevailing wage laws. It should be Silginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 9 kept in mind that this cost effect would be tempered by substitution between different qualities of labor as well as between labor and capital.Total cost is not expected to rise proportionately because contractors facing higher labor costs would alter techniques of production,employing more capital and skilled labor intensive techniques. In addition,the SDFWP may also affect the final cost by changing the type of uncertainty facing the bidders.Assuming that there is a monotonic relationship between the accepted bid price and the final cost of construction,the final cost would be influenced by the degree of competition because each contractor takes into consideration the number of bidders its competing against in determining the mark-up.The auction theory predicts that if the "independent private values"model applies then bidders lower their mark-ups in the face of higher competition in order to maximize their chances of winning the contract.'z This "competition effect"implies that the mark-up and the bid price are inversely related to the number of bidders.When the"common values"model applies,however,bidders would be subject to the so-called"winner's curse." The winner's curse refers to the fact that in a first- price sealed-bid auction,although expected value of each bidder's bid is unbiased,the winning bid will be biased downward,or that the winner is the one who underestimated the construction cost the most.Consequently,the winner is likely to make less than anticipated profits or may even lose money(Thaler, 1988). Since rational bidders do not make consistent errors,the optimum bidding requires experienced bidders protect themselves from the winner's curse by adding a surcharge to the estimated cost. Optimal behavior under the common values model requires this surcharge to rise with the number of competing bidders;offsetting the competition effect. To the extent that the SDFWP reduces uncertainty over labor costs,the relationship between the bid price and the degree of competition measured in terms of the number of bidders would change from more to less positive or from less to more negative following the implementation of the Law. In the extreme case where introduction of the law results in a switch from common to private values model,the degree of competition would influence bid prices positively before the passage of the Law and negatively afterwards.Bilginsoy(1999)uses the B.0 school tender data of the projects considered in this paper to provide evidence for such a switch in the response of bid prices to the degree of competition. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415432 10 The cost of construction also depends on the capital stock and capacity utilization of the contractor. If there are economies of scale associated with larger size,final cost should decline with the size of the firm. If the contractor faces diminishing returns as it approaches full capacity, final cost would increase.Ideally,one should be able to control for the capital and the level of activity separately, but the data do not allow this. Instead,it is possible to build a dataset for the "size"of the contractor,defined as the average annual volume of sales. The obvious shortcoming of this variable is that it cannot distinguish between the capital stock and capacity utilization,and therefore its impact on the final cost is indeterminate. In the absence of data that can make finer distinctions,we used this as a proxy to control for the f m-i characteristics. We constructed data series on size from the Canadata database and the Journal of Commerce's "Substantial Performers"and"Leaders"reports of the leading contractors for 1992, 1995 and 1996.'3 Theory suggests little about determinants of change orders.There is some evidence indicating that cost ovemms may be directly related to the size of the project,and that the cost overrun is more likely if the bid is below the owner's estimate(Jahren and Ashe, 1990). Although it is possible to control for the area of the school,data on owner's pre-tender estimate are far from complete in our sample and therefore not included as an explanatory variable. There may be a direct link between change orders and the SDFWP,however.Dyer and Kagel (1996,p.1470)identify negotiations over the change orders one method whereby a contractor who has underbid may recoup his/her losses.If we follow this argument,change orders are expected to be higher in the SDFWP period when contractors are more likely to be victims of the winner's curse.We calculated the change orders for the projects in the data set as a percentage of the accepted bid price. For all the projects in the dataset,the average change order was 2.08 percent of the lowest bid price.For the pre- and post-SDFWP periods, however,the mean values of change orders were 2.54 and 1.68 percent,respectively.The difference in means is statistically significant at the 10 percent level,providing some evidence for the hypothesis that change orders declined after the establishment of the Policy. Hence it may be hypothesized that the SDFWP affects the final cost directly by lowering the final unit cost. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 I1 Finally,the final cost of construction may change gradually over time for reasons other than those listed above. Such factors may include technological changes which raise productivity and lower cost over time,improving methods of cost estimation, gradual specification changes in construction,14 or adjustment process of contractors to the new legal regime. In order to capture the effect of these gradual changes we add a monthly time index starting in January 1989 to the explanatory variables. Since it is impossible to distinguish between the separate effects of these secular factors on this monthly time trend we do not have any priors on the direction of its impact on the final cost. The project size was excluded from the final regression equation because it was highly correlated with the school type and therefore introduced collinearity problems.The estimated regression equation is then: ln(Final Cost) = 6,, + g Secondary school+,8,Construction cycle +A(1 I Number of bidders)+Ai Contractor size; +ATime+ 8,jDistrict i +rl. (2) Final Cost is the square-meter final cost of construction deflated by the non-residential construction price index,Number of bidders is the number of bidders per tender, Contractor size is a vector of dummy variables for contractor size categories(indexed by i), Time is the monthly time trend starting in January 1989,and District is a vector of dummy variables for the six school districts(indexed by j). r)is the error term. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415432 12 4. Estimation We estimated regression equation(2)for the whole period,as well as the pre- and post-SDFWP periods by the ordinary least squares method.This method not only shows how the SDFWP affected the final cost controlling for these variables,but also if and how the response of the final cost to these variables were altered after the establishment of the Policy.In estimation,contractor size turned out to be statistically insignificant in all regressions and,given degree of freedom constraints,we decided to exclude these from the final regressions results reported on Table 3. Table 3 reports the parameter estimates of equation(2)for two specifications of the model.In the first specification,reported by columns(1)to(3)the time index is omitted.In column(1)the regression is run over the complete sample.According to the adjusted R2 of column(1)the equation explains 33 percent of the total variation of the variation in unit final costs for the whole sample Columns(2)and(3)apply the equation separately to the pre- and post-SDFWP sub-periods. We tested the hypothesis that the coefficients of explanatory variables are equal across the two sub-periods. This test yielded the F-value of 5.73,which is statistically significant at the one percent level.Rejection of the hypothesis of stability of coefficients across the periods indicates that it is inappropriate to pool the pre- and post- SDFWP periods in estimating the final cost regression equation. Indeed the explanatory power of the regression increases significantly once the sample is divided on the basis of the legal regime.Adjusted coefficients of determination reported in columns(2)and(3)are 0.55 and 0.48,substantially higher than what is found in the case of the full sample.According to the second column of Table 3,before the establishment of the Policy, secondary school unit cost was on average 14.4 percent higher than that of the elementary schools.The negative coefficient of Construction cycle indicates countercyclical behavior for final cost but the parameter estimate is not statistically significantly different from zero.The final cost increases with the number of bidders for the project and this result is statistically highly significant(p=0.003).All else being constant,an increase in the number of bidders(at the mean) Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 13 raises the square-meter final cost by 3.3 percent.This is consistent with the optimum bidding behavior of contractors who realize that they are susceptible to the"winner's curse." Contractors face collective uncertainty over the cost of the project prior to the SDFWP and, as economic theory suggests,they increase their bid prices in response to an increase in the number of bidders. The third column shows that after the SDFWP secondary school unit cost was still higher than that of the elementary but now only by 5.6 percent and this result is not statistically significant.Final cost is pro-cyclical and,again,the result is statistically significant(p=0.03): a one percent increase above the trend growth raised the square-meter cost by approximately 0.7 percent.The most striking effect of the SDFWP concerns the impact of the number of bidders. The sign of the coefficient is now positive(p--0.03): a unit increase in the number of bidders(at the mean)now lowers the unit final cost by 1.8 percent. In comparison with the pre-SDFWP period,this finding suggests a change in the type of uncertainty facing the bidders.It is now private uncertainty,rather than collective,that is more relevant in the bidding process and rising competition induces contractors to lower their mark-ups(and consequently bids and final cost) in order to increase the probability of winning the contract.15 In order to compare average unit costs of the pre and post-SDFWP periods we predicted the construction cost of an elementary school in school district 6.We assumed that the building permits grow at the trend rate and that eight contractors make bids for the project. Under these assumptions,the predicted average construction cost is$1,238 before the SDFWP and$1,313 afterwards.This cost differential is not statistically significantly different from zero.The 95 percent confidence interval for the pre-SDFWP prediction($1,097-$1,397) includes the post-SDFWP figure. The remaining three columns of Table 3 report estimation results of the specification including the time index.Estimation of this specification yields results that are consistent with those reported earlier.Again the hypothesis of structural stability of coefficients across the two periods is rejected at the one percent level(F-value=5.20).In the pre-SDFWP period secondary schools are more expensive to build by 12.5 percent(p=0.02).The cost of construction is varies directly with the number of bidders.An increase in the number of bidders Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 14 (at the mean)is estimated to raise the final cost by 2.0 percent(p=0.09).Unit construction cost is also estimated to rise at 0.6 percent per month(p=0.08).In the post-SDFWP period,the type of school no longer makes a difference on the cost of construction. Similar to the earlier finding reported on column(3),final cost is inversely related to the number of bidders.It declines by 2.27 percent in response to a unit rise in the number of bidders(at the mean). The final price also declines by 0.05 percent per month in the post-SDFWP period. Final cost is not responsive to the business cycle in either period.While the theory did not predict unambiguous sign for the coefficient of this variable,it should still be remembered that quarterly data on business cycles would have yielded more reliable assessment of the cyclical behavior of the final school construction costs. We again predicted the average cost of construction under the assumptions stated above.These predictions do not indicate a significant change in the cost of construction following the establishment of the SDFWP.The predicted cost increased gradually between January 1989 and March 1992,reaching$1,347 just before the establishment of the SDFWP. At the end of March 1992,the SDFWP was established.On April 1992,the predicted price jumped to$1,474, and then it declined gradually.It took about 20 months for the price to decline to its March 1992 level. 5. Conclusion A number of politicians and school officials proposed the elimination of prevailing wage regulations as a means to lower public school construction costs.The B.C. data,at first sight, seem to support this view. The brvariate"before and after"comparison of final unit price indicate that the SDFWP caused the construction cost to increase by 16 percent,even higher than some of the figures estimated by the critiques of the U.S.prevailing wage law.This difference is also found to be statistically highly significant.However,when the same experiment is carried out controlling for other factors including the construction business cycle,number of competitors,type of school,district dummies,and the time trend,this result is no longer holds. There is indeed a structural break in the determination of the final cost after the SDFWP,but Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415432 15 there is no evidence of significant changes in the unit costs before and after the establishment of the SDFWP. Instead,what is observed is a 6.1 percent increase in price,if the impact of the time trend is ignored.No statistical significance can be attached to this figure.If time trend is also included in the analysis,the price rises by 9.4 percent at the time of introduction of the Law followed by a steady decline afterwards. This steady decline,over time,offsets the immediate inflationary cost impact. These findings indicate that the appeal for the repeal of prevailing wage laws to reduce the school construction costs and the burden on the public budget is misguided. Regression results also indicate that the prevailing wage affects construction costs through a variety of subtle channels that are overlooked in the literature.Present findings suggest two factors playing important roles in cost determination.The first is the impact of competition on the final price. The final price is directly(inversely)related to the number of bidders before (after)the SDFWP.This finding supports the hypothesis that the SDFWP altered the nature of uncertainty facing the bidders and made bid surcharges to avoid the winner's curse unnecessary. The second factor is the trend change.We do not know exactly which factors lie behind the rising trend costs in the pre-SDFWP and declining trend of the post-SDFWP periods. Possible answers to this question may he in the areas of technological and non-wage regulatory changes, as well as the learning behavior of contractors and their adjustment process to the changing regulatory environment. SDFWP may also affect costs by lowering the cost of change orders. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415432 16 TABLES Table 1: School Construction Projects in Six B.C.Districts: 1989-1995 Pre-SDFWP Post-SDFWP Whole period Elementary School 18 21 39 Secondary School 7 8 15 All Schools 25 29 54 Table 2: Final Square-meter School Construction Cost(in 1989 prices—Can.$) _ No. of Obs. Mean Median Stan.Dev. Range All Tenders 54 $1,423 $1,440 $217 $968-1,811 Pre-SDFWP 25 1,308- 1,248 189 968-1,690 Post-SDFWP 29 1,515a 1,508 195 1,194-1,811 Note:We tested the hypotheses that means and standard deviations of the post-SDFWP periods are equal to their pre-SDFWP counterparts(one-tailed tests). Superscript a indicates that the hypothesis is rejected at the one percent level. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415432 17 Table 3: Regression Model for Final Cost (Dependent variable:natural log of square-meter cost) Specification 1 Specification 2 Whole Pre- Post- Whole Pre-SDFWP Post- Period SDFWP SDFWP Period SDFWP (1) (2) (3) (4) (5) (6) Constant 7.272 7.379 7.007 7.118 7.115 7.260 (88.30)*** (81.84)*** (70.39)*** (76.11)*** (42.97)*** (49.28)*** School type 0.090 0.144 0.056 0.094 0.125 0.040 (=1 if secondary) (1.94)* (3.03)*** (1.26) (2.56)** (2.75)** (0.97) Construction 0.007 -0.008 0.007 0.008 -0.009 -0.002 Cycle (1.94)** (1.24) (2.27)** (2.47)** (1.50) (0.45) 1/(No.of bidders) -0.646 -2.063 1.390 -0.380 -1.268 1.757 (131) (3.50)*** (2.36)** (0.81) (1.82)* (3.11)*** Time 0.002 0.006 0.005 (2.85)*** (1.85)* (2.18)** District dummies Included Included Included Included Included Included w 0.43 0.70 0.63 0.52 0.76 0.71 Adj.RZ 0.33 0.55 0.48 0.42 0.61 0.57 F 4.19*** 4.74*** 4.28*** 5.21*** 5.23*** 5.05*** Observations 54 25 29 54 25 29 NOTE:t-statistics in parentheses. ***,**and*indicate statistical significance at the 1 5,and 10 percent levels,respectively. 0 Bilginsoy&Philips,Journai of Education Finance, Vol.24(Winter 2000)pp.415432 18 REFERENCES Allen, Steven G., 1983,"Much Ado about Davis-Bacon:A Critical Review and New Evidence,"Journal of Law and Economics,Vol. 25, October, 707-736. Bilginsoy,Cihan, 1999,"Labor Market Regulation and the Winner's Curse,"Economic Inquiry(forthcoming). Bourdon,Clinton C. and Raymond E.Levitt, 1980, Union and Open Shop Construction, Lexington,MA: Lexington Books. Dyer,Douglas and John H.Kagel. 1996.`Bidding in Common Value Auctions:How the Commercial Construction Industry Corrects for the Winner's Curse,"Management Science, October,Vol. 42,No. 10, 1463-1475. Fraundorf,M.,Farrell,J.P.,Mason,Robert, 1983,"The Effect of the Davis-Bacon Act on Construction Costs in Rural Areas," The Review of Economics and Statistics,Vol.66, February, 142-146. Gilley,Otis W.and Gordon V.Karels, 1981,"The Competitive Effect of Bonus Bidding:New Evidence," The Bell Journal of Economics,Vol. 12,Autumn,637-648. Goldfarb,Robert S. and John F.Morrall, 1981,"The Davis-Bacon Act:An Appraisal of Recent Studies,"Industrial and Labor Relations Review,Vol. 34,No. 2, 191-206. Goldfarb,Robert S.and John F.Morrall, 1978,"Cost Implications of Changing Davis-Bacon Administration,"Policy Analysis,Vol.4,No.4,439-453. Gould John P. 1971 Davis-Bacon Act: The Economics o P > revailin a of Prevailing Wage Laws, Washington,D.C.: American Enterprise Institute for Public Policy Research. d John P.and Bittlin a er George, 1980 The Economics o Gould, gm y g , Davis-Bacon Act:An.f Analysis of Prevailing Wage Laws,Washington,D.C.:American Enterprise Institute for Public Policy Research. GAO, 1979, The Davis-Bacon Act Should Be Repealed,Washington,D.C. GAO, 1981,Modifying the Davis-Bacon Act:Implications for the Labor Market and the Federal Budget,Washington,D.C. Government of Canada, 1997,"The Impact of Skills Development and fair Wage Policy on Construction Costs in British Columbia: An Empirical Analysis of Some Key Issues,"May. Bilginsoy&Philips,Journal ojEducation Finance, Vol.24(Winter 2000)pp.415-432 19 Gujarati,D.N., 1967, "The Economics of the Davis-Bacon Act,"Journal of Business,Vol. 40,July,303-316. Hendricks,Kenneth and Robert H.Porter, 1988,"An Empirical Study of an Auction with Asymmetric Information,"American Economic Review,Vol. 78,No.5,December, 865-883. Jahren,Charles T. and Andrew M. Ashe, 1990,"Predictors of Cost-Overrun Rates,"Journal of Construction Engineering and Management, September,Vol. 116,No. 3, 548-552. Milgrom,Paul, 1989,"Auctions and Bidding:A Primer, " The Journal of Economic Perspectives, Summer,Vol. 3,No. 3, 3-21. Thompson,David C.,R. Craig Wood and David S.Honeyman, 1994,Fiscal Leadership for Schools,New York,Longman. Thaler, R.H., 1988,"The Winner's Curse,"Journal of Economic Perspectives,Vol.2,No.1, Winter, 191-201. Thieblot,Armand J.,Jr., 1975,The Davis-Bacon Act,Philadelphia: Industrial Research Unit, The Wharton School,University of Pennsylvania. Thieblot,Armand J.,Jr., 1986,Prevailing Wage Legislation: The Davis Bacon Act, State "Little Davis-Bacon"Acts, The Walsh-Healey Act, and The Service Contract Act, Philadelphia: Industrial Research Unit,The Wharton School,University of Pennsylvania. U.S.Bureau of the Census, 1997, Census of Construction, 1992,GPO,Washington,D.C. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 20 ENDNOTES ' For instance,Thompson, et al. (1994,p. 553)state"...it is truly startling to recognize that almost half of all school buildings in the nation have only marginal future utility,and another 20%-30%are candidates for reconstruction or abandonment because they are more than 50 years old." 2 The first state prevailing wage law was passed in Kansas in 1891. Eventually 41 states would adopt such legislation. Between 1979 and 1988,nine states repealed their prevailing wage laws. In 1995,Oklahoma's law was judicially overturned and in 1997 Ohio exempted school construction from the purview of its prevailing wage law. The federal Davis Bacon Act was passed in 1931. Because most public school construction does not include federal funds in the U.S.,public school construction is regulated by state laws. 3 "...without the constraint of the Little Davis-Bacon Act,we could build four schools instead of three for the same amount of money."New Mexico Governor Gary Johnson,State of the State Address,January 16, 1996. 4 Rick Savors,Ohio School Boards Association Deputy Director of Legislative Networks, Testimony on Ohio Senate Bill 102,May 8, 1997. 5 Richard Maxwell,Testimony on Ohio Senate Bill 102,May 8, 1997. 6 Jim Shirey,Testimony on Ohio Senate Bill 102,May 8, 1997(emphasis in the original). This result is from the U.S. Census of the Construction Industry, 1992, Industry Series, United States Summary, CC92-I-27 Table 3,page 27-8. There is no comparable Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 21 construction census for Canada. The data are for contractors with payroll. The numerator of the ratio of labor costs to total costs includes total wages paid to construction workers plus the value of both legally mandated benefits(such as social security)and voluntary benefits(such as a private pension)paid for by the employer. This benefit figure includes not only benefits paid to construction workers but also to all other clerical and support workers employed by construction contractors. Thus,this is an overestimate of the wages and benefits paid to workers covered by prevailing wage regulations. The denominator is the net value of construction work done. This figure avoids double counting associated with subcontracting. Anecdotally construction contractors sometimes report higher labor costs as a percent of total costs. While labor costs as a percent of total costs vary among contractors and any one contractor is unlikely to reflect the average,anecdotal reports suffer from a second handicap. Often when contractors calculate their labor costs as a percent of total costs,they are thinking of labor costs to them. Consequently they exclude their charge for capital depreciation and profit to the purchaser of their services. The U.S. Census of Construction figure includes the contractor's markup in the total costs paid by the owner or purchaser of the construction. This is the relevant figure when trying to assess the effect of a savings on labor costs to school board construction costs. 8 Davis Bacon periodically surveys counties. Davis Bacon sets the"prevailing"wage rate at the modal rate found in the survey for specific occupations if the mode accounts for 50 percent plus one wage rates found.Otherwise,the average wage rate is said to prevail. 9 All prices are in Canadian$. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 22 10 The mark-up on change order is likely to be different from m because the change order price is negotiated separately. Dyer and Kagel(1996)argue on the basis of U.S.field data that contractors who underbid may use these negotiations to recover their losses. " The data provide only the gross area size. 12 Private values model implies that the bidders' estimation errors of cost of construction are independent.In the alternative case of common values model,estimation errors are interdependent.The latter would be the case if all bidders face some common uncertainty concerning the cost of construction which may be attributable to factors such as weather conditions or the state of the general labor market(Milgrom, 1999). 13 We identified six size categories in terms of average volume of sales:less than$15 million, g g $15 to 30 million,$30 to 60 million, $60 to 100 million,$100 to 200 million, and above$200 million. 14 In B.C.,school construction specifications,including building,mechanical and electrical codes,were changed in the late 1980s and early 1990s in order to make structures more earthquake resilient. 15 The assumption of exogeneity of the number of bidders deserves further comment.The number of bidders may be exogenously determined if the bid/no-bid decision is influenced by the desirability of the project.If so,the number of bidders would no longer be an appropriate measure of the degree of competition,and the estimated parameters would be biased and inefficient. In the economics of auctions literature it is suggested that the decision to bid would be affected by factors including the variance of the estimated value of the project and Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 23 i asymmetric distribution of information across the bidders. These factors were shown to be signif cant in the outer continental shelf(OCS)hydrocarbon lease auctions(Gilley and Karels, 1981;Hendricks and Porter, 1988). In addition,in school construction some contractors may chose not to bid in school districts that are difficult to work with. These problems are not serious in our sample because school construction is a far more standard item to bid in comparison with the OCS leases and the scope of asymmetric information across contractors,if any exists,is likely to be very narrow.The variance of the submitted bids for any project suggests a much smaller variation in cost estimates.Inspection of"special features"in the architect's cost estimation forms indicates little variation across projects and between the pre- and post-SDFWP periods, and supports these contentions.Finally,the spread of the average number of bidders across districts is quite narrow,ranging from 7.3 to 9.3,and do not suggest that contractors deem any district to be less desirable than the others. Bilginsoy&Philips,Journal of Education Finance, Vol.24(Winter 2000)pp.415-432 lin Wa Kentucky'sPff—w-Valg ge fts Hisl�, Purpose and EfF By Peter Philips, Ph.D. Professor of Economics, University of Utah October, 1999 1 Table of Contents About the Author Executive Summary Chapter One: History of Prevailing Wage Regulations in Kentucky and the U.S. Appendix: The Current Kentucky Prevailing Wage Law Chapter Two: Are Prevailing Wage Laws Jim Crow Laws? ChVter Three: An Analysis of the Impact of Prevailing Wage Laws on Construction Costs Chapter Four: How Does Kentucky's Prevailing Wage Law Effect Apprenticeship Training? Chapter Five: The Rise in Serious Injuries in Kentucky Construction After Schools and Municipalities Were Exempted from the state's Prevailing Wage Law in 1982 Chapter Six: Prevailing Wage Regulations, Family Health Insurance and Old Age Security 2 About the Author Peter Philips grew up in Compton and Pomona, California. He received his B.A. from Pomona College in 1970 where received the Leland Backstrand Graduating Senior Award in Economics. Philips received his MA. (1976) and his Ph.D. (1980) from Stanford University. Philips is a Professor of Economics at the University of Utah. He is co-editor of Three Worlds of Labor Economics (M.E. Sharpe, 1986) and co-author of Portable Pensions for Casual Labor Markets (Quorum Books, 1995). Philips has published widely on the canning and construction industries in journals such as Industrial and Labor Relations Review, Industrial Relations, Business History, the Journal of Economic History, The Journal of Economic Literature and the Cambridge Journal of Economics. Philips has been a consultant for the U.S. Labor Department analyzing the supply of cannery labor in California, and he has worked as an expert on the Davis-Bacon Act for the U.S. Justice Department. The Davis-Bacon Act regulates wage payments to construction workers on federal public works. Philips is a respected expert on prevailing wage laws and on employment, training wages and benefits in the construction industry. He has testified before state legislative committees in Ohio, Indiana, Oklahoma, New Mexico and California on their state prevailing wage laws. Along with other researchers at the University of Utah, Philips has analyzed the effects of prevailing wage laws on public construction costs, construction worker incomes, apprenticeship training, worker safety and minority access to construction work. Philips is currently working with Anthony Suruda, M.D., M.P.H., of the University of Utah Department of Family and Preventative Medicine on a National Institute of Occupational Safety and Health funded study of child labor fatalities in construction. Philips has received awards for his teaching and community service, including the University of Utah Lowell Bennion Public Service Professorship, the University of Utah Presidential Teaching Scholar Award and the University of Utah, College of Social and Behavior Science Superior Teacher Award.Philips is married with two children. 3 Executive Summary-The Main Points HIS —the first federal prevailing wage law was passed by the Republican Congress of 1868 and enforced by President Ulysses Grant. The first state prevailing wage law was enacted in 1891 by the Republican and Populist legislature of Kansas. The current federal law—the Davis Bacon Act—was passed by the Republican Congress of 1931 and signed by Herbert Hoover. The Kentucky law dates to 1940. In 1982 Kentucky schools and some municipal work were exempted from the law's coverage but in 1996 the law was re-applied to school and municipal construction. Currently, collectively bargained wage rates are required to be paid on public works when a majority of workers of a particular trade in a local labor market are receiving those wages. When less than half of an occupation is receiving the collectively bargained rate, then the prevailing wage is the weighted average of all wage rates for the trade in the area. The PU'p06e of P11�lkig W to LANs--Prevailing wage laws were originally intended to encourage the development of a high-skill, high-wage growth path for the labor market in general, and the construction labor market in particular. Where prevailing wage regulations are applied, union and nonunion contractors win public works jobs based on having the most productive, best equipped and best managed workforce. Under prevailing wage regulations, wages established in the market through collective bargaining are not undercut through the power of government. Critics of prevailing wage regulations argue that the government should use its bargaining power to cut local wage rates. They contend that local wage rates could be cut by as much as 50%. And they contend that such a race to the bottom can cut public construction costs substantially. But you get what you pay for. When wages are cut substantially, worker skills, experience and motivation also fall off. Contractors no longer compete on the basis of who can best train, best equip and best manage a construction crew. Rather they compete on who can find the cheapest workers 4 either locally or through importing labor from elsewhere. This puts the quality of construction at risk. It may lead to increased cost overruns. It may lead to downstream increases in building and road maintenance costs. And it does lead to an increase in construction injuries and a decline in the health and pension coverage of construction workers. This puts pressure on worker compensation costs. And it puts pressure on social services—as family health needs go unmet and retired workers cannot make ends meet. The original prevailing wage laws were passed at the same time and often in tandem with compulsory schooling and child labor laws. All these laws share a similar philosophy. In the short run, some employers can turn a profit based on cheaper labor. Indeed, that was the claim of those who hired child labor and opposed compulsory public schooling. But in the long run, society is better off supporting regulations that encourage skill formation and competition based on increased worker productivity. Prevailing wage regulations, by not undercutting collective bargaining, support the largest privately financed system of higher education in the country— apprenticeship training. Three, four and five year apprenticeship training in construction is the foundation of a productive construction workforce and quality workmanship. When apprenticeship and journey-worker training is cut back, productivity falls. That is why you can cut wages 50% and show little or no effect in overall labor costs. The philosophy underlying prevailing wage regulations is that a community is better off in the long run by encouraging competition that builds skills builds productivity, pays decent wages and provides for the health p Y� p Y 9 and old age needs of its citizens. That is the philosophy and that is what prevailing wage regulations do. • Race—Some claim that prevailing wage laws were originally Jim Crow laws designed to keep blacks out of construction. There is no evidence to support this in the case of Kentucky. Nor does it square with what we know about those who supported the early federal and state laws. The first federal law in 1868 was passed by the same Republican Congress that enacted the 13th, 14th and 15th Amendments to the Constitution—the legal bases for equal rights in this country. The first state law was declared constitutional by U.S. Supreme Court Justice John Marshall Harlan—the leading judicial critic of Jim Crow laws in his day. The current federal law was vigorously supported by NY Republican Congressman Fiorello LaGuardia who subsequently, as mayor of New York, played a strong supportive role in bringing Jackie Robinson to the Brooklyn Dodgers. Critics nonetheless claim that prevailing wage laws keep blacks and other minorities off public construction. They claim that blacks are less skilled and prevailing wage rates require that contractors staff public construction with their best workers. So, because blacks are second rate, they are kept off the job. But the claim that black construction workers are less capable is unsupported. An 5 econometric analysis of the relationship between prevailing wage regulations and the presence of blacks in construction fails to find support for this claim. Furthermore, in states with prevailing wage laws, the proportion of minorities in construction apprenticeship programs reflect the percentage of minorities in those states' populations. In contrast, in states that do not have prevailing wage laws, minorities are under-represented in construction apprenticeships. In states without prevailing wage regulations there are 18% to 19%fewer minorities in construction apprenticeship than one would expect from those states' minority populations. The probable reason for this divergence in representation is the size of apprenticeship programs. Under collective bargaining, apprenticeship programs are typically multi-employer programs of considerable size. Many merit shop apprenticeship programs are single-employer programs with only two or three apprentices. Affirmative action regulations in apprenticeship only apply to programs of five or more apprentices. So union apprenticeship programs almost always fall under regulations prohibiting discrimination,while many open shop apprenticeship programs do not. Finally, when it comes to graduating minorities from construction apprenticeship programs, the graduation rates are much higher under collective bargaining. Onl 18% of minorities in merit shop gY apprenticeship programs graduate from those programs. Thus, minority apprentices are heavily under-represented in those states that do not have prevailing wage regulations. And if those minorities are in programs organized by merit shop contractors, the odds of those few ever turning out as journeymen and women are grim. • COStSr—Some critics of prevailing wage regulations erroneously claim that you could build four schools for the cost of three if prevailing wage regulations were lifted. But labor costs in construction are only about 30% of total costs. To build four schools for the cost of three, everyone on the job site would have to work for free. Real--inflation adjusted—costs of school construction in Kentucky have risen in the 1990s. But this rise occurred prior to the application of prevailing wage regulations for schools. Since 1997, real median square foot cost of public school construction has been stable. Chapter 3 presents a study of over 6,000 school construction projects in the U.S. in the 1990s. This study does not find a statistically significant relationship between prevailing wage regulations and school construction costs. However,the study does find that school authorities can save as much as 4% on total construction costs if they begin their projects after the coldest weather has subsided. Late winter and spring start-ups run 4% less compared to similar projects begun in late summer or fall. School authorities can also save if they do not build into a tight construction market. A two percent drop in the state's unemployment rate leads to a 4% rise in total construction 6 costs. Counter-cyclical construction spending on the part of public schools can save taxpayer dollars and also soften the blow of slow times on the local community. • $loll S11o1S and Tlalil —The long expansion of the 1990s has led to severe skill shortages in construction. This is almost exclusively a problem for contractors that do not use collective bargaining. The Business Roundtable reported the results of a survey of its members in late 1996: "Over 60 percent of the survey respondents indicated that they had encountered a shortage of skilled craft workers, and 75 percent reported the trend had increased over the past five years...The union sector has always excelled in craft training through the joint labor/management apprenticeship programs....the open shop, as a whole, has not supported formal craft training to the extent necessary." Prevailing wage regulations promote training directly by requiring all contractors on public works to pay prevailing wage rates that include apprenticeship costs when collectively bargained rates apply. Prevailing wage regulations support training indirectly by not undercutting collective bargaining. In construction, 66% to 99% of all graduating apprentices, depending on trade, come out of joint labor-management programs established through collective bargaining. • IrjL.7E1!S—When prevailing wage regulations are eliminated, worker wages and benefits go down. Indeed, that is the stated aim of deregulation—the cutting of wages and benefits in the hopes of cutting costs. Proponents of deregulation often claim wages will be cut in half. But when this does occur, experienced and trained workers leave the industry. A younger, less experienced and less trained set of workers shoulder the job. This not only puts at risk the quality of work, it puts at risk the workers themselves. In Kentucky, in the nine years after schools were exempted from prevailing wage regulations in 1982, serious injuries among Kentucky construction workers rose 11% compared to the six years prior to exemption. Serious injuries as a percent of all injuries rose by 8%. And the days lost per serious injury rose 16%to an average of 21 lost days of work per injured worker. • Heaft Insurance and Old Age Seculity—When collective bargaining is absent in construction, most workers are not paid health insurance or pensions. In Kentucky,the average contractor that has collectively bargained 7 with his or her workers pays $1,240 per year in employer contributions to each worker's pension, and $2,567 in employer contributions to each worker's health insurance. The typical contractor that does not collectively bargain with his workers pays, on average, $51 per worker into pensions and $242 into health insurance per year. These are averages. Actually, Kentucky's open shop contractors pay their key workers more than $51 in pensions and the rest of their workers get no employer-sponsored pensions. Ninety-six percent of Kentucky construction workers who do not have the protection of collective bargaining are not receiving pensions. Over half are not receiving health insurance. Construction workers form roughly 5% of the labor force. When the construction industry and ultimately its customers fail to pay for the family health costs and the old-age costs of 5%of the community, then eventually social services are forced to pick up some of that tab. t Wr a . 8 The History of Prevailing Wage Regulations with a Special Focus on Kentucky The First Federal Prevailing Wage Law(1868). Ulysses S. Grant was the first President to seek enforcement of a federal prevailing wage law. The first federal eight hour day law was enacted on June 25, 1868. It also was the first federal prevailing wage law. The country had just passed through a Civil War that among other things had kick-started massive industrialization across the north and west of the country. The next thirty years would see the emergence of a new class of wealth and power in the country. Men such as J.P. Morgan, John D. Rockefeller, and Andrew Carnegie were using the rapid growth stimulated by the Civil War as a foundation for accumulating economic power never before seen in the country. At the same time, the lives of working people were in flux. Hours of labor had always been long but they had moved to the pace and the rhythms of the farm. Shoe factories in New England, meat packing plants in Chicago and woolen mills in California changed all that. Work was being harnessed to the time clock and the production line supervisor. People were being ground down by the pace of machinery, the demands of the supervisor and the strain of 12 hour days and six day weeks. In 1868 Congress addressed this issue with the National Eight Hour Day law. The idea was to set labor standards, to guide the labor market, to nudge it away from the stretching out of the workday towards competitive behavior that emphasized increased productivity within a limited set of hours. It was felt that the market could not get there by itself. Short run competitive pressures would continually push for the longer day. But by regulating the market, it could be forced to find its own best self-interest, competition over productivity rather than competition over sweating labor. The legal doctrine of individual contract prevented Congress from directly regulating the market, but Congress could regulate its own contracts. Thus, public works was targeted as a way of indirectly trying to regulate all labor markets. Republican Senator Conness of California captured most of these ideas in one line of argument: 9 [The Eight-Hour Law] is but a very small boon that the working men of America ask from the Congress of the United States, namely: that the example be set by the Government of reducing the number of hours of labor. I know that the passage of this bill cannot control in the labor of the country; but the example to be set by the Government, by the passage of this bill, is due to the laboring men of the country, in my opinion. I know that labor in the main, like every other commodity, must depend upon the demand and supply. But, sir, I for one will be glad, a thousand times glad, when the industry of the country shall become accommodated to a reduced number of hours in the performance of labor. After forty or fifty }ears of such advance in the production of the world's fabrics by the great improvements that have been made by inventions, and the application of steam as a power, by which the capital of the world has been aggregated and increase many fold, I think that it is time that the bones and muscles of the country were promised a small percentage of cessation and rest from labor, as a consequence of that great increase in the productive industries of the country. 1 Prevailing wage regulations were an integral part of the first national eight-hour law. For Congress said that when hours on public works were cut from 12 to 8, the daily wage should not be cut from (say) $1.20 to 80 cents. In those days, construction workers were paid by the day. Congress said that when hours were cut, the contractor on public works still had to pay the daily wage that was current in the locale in which the work was being done. Enforcement of the current wage provision proved difficult. Twice Republican President Grant had to issue proclamations directing contractors and government agents to respect the current wage provision of the eight-hour day law.2 Thus, the principle of a prevailing wage law at the federal level predates the Davis-Bacon Act by fifty years. The purpose of the federal law was to set labor standards regarding hours and wage rates in the public sector presumably with the hope that these standards night spread to the private sector. That the purpose was thwarted in enforcement is indicated by Grants need to make the same proclamation twice. It was also thwarted by legal decisions emphasizing the rights of individuals to contract without government interference. Frustrated by problems of implementation and court rulings, the American Federation of Labor, in its first convention in 1881 stated what it thought the purpose of the law was and complained that it was not being enforced: 1 Congressional Globe,op cit. 2 On May 19, 1869,President Grant issued the following proclamation: that,from and after this date no reduction shall be made in the wages paid by the Government by the day to such laborers,workmen and mechanics on account of any such reduction of hours of labor. On May 11, 1872 Grant reiterated with greater detail and emphasis in a second proclamation that per diem wages should not be cut with the required shorter hours: ...I,Ulysses S.Grant,President of the United States,do hereby again call attention to the act of Congress aforesaid,and direct all officers of the executive department of the government having charge of the employment and payment of laborers,workmen and mechanics employed by or on behalf of the government of the United States to make no reduction in the labor wages paid by the day to such laborers, workmen and mechanics on account of the reduction of the hours of labor. The Statutes at Large and Proclamations of the United States of America from March 1871 to March 1873,Vol.XVII,Boston,1873,pp.955-56. The Statutes at Lar e and Proclamations of the United States of America from December 1869 to March 1871,Vol.XVI,Boston,1871,p. 1127. 10 Resolved...that the National Eight Hour law is one intended to benefit labor and to relieve it partly of its heavy burdens, that the evasion of its true spirit and intent is contrary to the best interest of the Nation; we therefore demand the enforcement of said law in the spirit of its designers.' The next year the AFL convention went on to argue "that the system of letting out Government work by contract tends to intensify the competition between workmen, and we demand the speedy abolishment of the same." Further by focusing on enforcing the federal law, "the enforcement of the National Eight-hour law will secure adoption of similar provisions in nearly all the States of the Union."4 Thus, the AFL wanted to get the government out of the business of pushing wages down and into the business of pushing hours of work down. Public works were targeted for regulation not so much because construction unions were a particularly powerful interest group but because under legal theories of the time, general governmental regulation of the labor market was viewed as a violation of the individual right to freely make contracts. However, he government was a party to contracts for public construction. Therefore, the government, like any party to a contract, could set conditions under which it was willing to contract for construction services. Proponents of hours and wage regulations on government works hoped these conditions would serve as a model and standard for private work in and out of construction. British(1890) and Canadian(1900)Laws. The country has no interest in keeping down the price of labour ; on the contrary, the country is interested in the advancement of the labour market....It is better for the workingman, for high wages enable him to supply himself with more of the necessaries, more of the comforts, more of the luxuries of life. This is better for the country also, as it stimulates the consumption of manufactured goods of all kinds. Higher wages benefit not only him who receives but him who gives, and they benefit not only the parties directly concerned,but the whole community. Canadian Postmaster General 1900 Workmen's Wages on Government Contracts Debate In England in 1890, the House of Lords issued the Report of the Sweating Commission. Sweatshop labor conditions had become a scandal. Construction was seen as one of the sweatshop industries. The system of contracting and subcontracting and lowest bidder acceptance led to a form of competition that was deleterious. To obtain a contract in the short run, contractors would ignore long term costs of the industry, such as training. Having shaved on a bid to win a government contract, contractors were trying to offset their costs through shoddy workmanship. Contractors who won a job would shop it around to laborers, seeing who would take the biggest pay cut to get a job. In response to these practices, Parliament enacted a prevailing wage law as part of a larger set of reforms designed to reign in the prevalence of sweatshop competitive practices. Canada followed the English example in 1900. The Canadian Parliament was persuaded that there was a high-wage, high-skilled growth path and a low-wage, low-skilled growth path 3 Federation of the Organized Trades and Labor Unions of the United States and Canada, 1881, "Declaration of Principles"in Proceedings of the American Federation of Labor, 1881 to 1888,Reprinted 1905,Pantograph Printing,Bloomington,Illinois(hereinafter Proceedings), 1881 p.3. This organization changed its name to the American Federation of Labor in 1886. 4 Proceedings, 1882,pp.4 and 18. 11 opening up before Canada. The high-wage path was seen as preferable because it promoted solid skills and good workmanship on public works, it created middle class citizens and it stimulated demand for local manufactured goods. The First State Prevailing Wage Lave-Kansas(1891). In February 1891, Samuel Gompers, president of the American Federation of Labor, visited Topeka, Kansas, to speak on what the local newspaper called "the great topic of labor." Ten years earlier, the AFL — at its own creation — had laid out legislative aims that included the eight-hour work day, the elimination of child labor, free public schooling, compulsory schooling laws, the elimination of convict labor, and prevailing wages on public works. These proposals were based on a belief that the American labor market should consist of highly skilled workers earning decent wages, with time for family, and with children free to earn an education. In pursuit of these aims,Gompers'political strategy in Kansas allied him with the Republican Party. On the morning of Gompers' arrival, the Alliance Party, known to history as the Populist Party, withdrew an earlier invitation for him to speak in the hall of the state House of Representatives, which the party controlled. Gompers, who represented 900,000 workers, had fallen out of favor with the Populists, reportedly because of his belief that the trade unions should not form a political party with the Alliance. Gompers and the AFL bok the position that unions should be nonpartisan. Rather than form a labor party, Gompers advocated that unions support those of any party who would support the needs of working men and women. In Kansas in 1891, this made Samuel Gompers an ally of the Republican Party. The Republicans, who controlled the Kansas Senate, invited Gompers to speak there, and he did. Gompers was in Kansas to focus on the eight-hour day. Like other Americans, Kansans in 1891 typically worked six days per week, ten to twelve hours per day. In the older trades and crafts, such as carriage making and saddle making, where the work pace was slow and under the workers' direction, the long work-day was tolerable. In the newer factories producing shoes, textiles, and the like; in the mines; and in the urban putting-out systems in needlework, six-day weeks and twelve-hour days were grueling. The AFL had made its prime objective a shortened work-day and work week with as little cut in pay as possible. In his Topeka speech, Gompers declared: Our banner floats high to the breeze and on that banner float is inscribed, "Eight hours work, eight hours rest and eight hours for mental and moral improvement.'b At that time, when there were no income supplement programs for the poor, low-income parents worked and had to send their children to work to make ends meet. This practice was later referred to by a North Carolina newspaper editor as "eating the seed corn." Each generation of poor condemned its offspring to poverty because the children grew up as illiterate as their parents. The prevalence of cheap child labor, which accounted for 5 percent of the manufacturing labor force in 1890 and a larger proportion of service sector workers, kept wages down and forced adult workers to put in the long hours to make ends meet. Gompers wanted regulation to force employers and the poor to adopt a strategy, however painful in the short run, of a high- 5.Topeka State Journal,February 24, 1891,co1.4,p.4. 6 Topeka Daily Capital,February 25, 1891,p.l. 12 wage, high-skilled growth path where children were in school and workers had the skills to justify wages that would allow for a family life.Gompers said, The Federation endorses the total abolition of child labor under 14 years of age; an eight hour law for all laborers and mechanics employed by the government directly through contractors engaged on public work, and its rigid enforcement; protection of life and limb of workmen employed in factories, shops and mines; ...the extension of suffrage as well as equal work for equal pay to women....The Federation favors measures, not parties. Gompers also pleaded for workers to be paid the "current" daily wage so they could afford the reduced work time. Government was being asked to set a good example for the private sector, to show that a refreshed labor force could produce in eight hours what a fatigued and bedraggled labor force turned out in ten or twelve hours. The prevailing wage law in its infancy was an attempt to obtain shorter working hours for all labor. The AFL paid attention to public works, however, because government at all levels was a major purchaser of construction. The AFL said government should not try to save money by eroding the wages of its citizens. With similar logic, the AFL called for an end to convict labor. Many states employed convicts to pay for their keep. Convicts built roads on chain gangs, operated farms, made textiles, and sewed garments. Convict-made goods were sold, forcing down prices and the wages of working free citizens. In February 1891, the Second Annual Convention of the Kansas State Federation of Labor, in Topeka, approved a bill concerning state-paid wages. That month, the bill, which included the prevailing wage section, called "for an Eight Hour Law" and was brought forth by Mr.Avery of the Typographical Union No.121,Topeka. The bill stated, That in no case shall any officer, board, or commission, doing or performing any service or furnishing any supplies to the State of Kansas under the provisions of the act be allowed to reduce the daily wages paid to employees engaged with him (or them) in performing such service or f imishing such supplies, on account of the reduction of hours provided for in the act. That in all cases such daily wages shall remain at the minimum rate which was in such cases paid and received prior to the passage of the act.8 The eight-hour bill was one of four labor-related bills pending in the legislature: the weekly pay bill, the child-labor bill, and the bill to make the first Monday in September a holiday, which would become known as Labor Day. In addition, that year the Kansas State Federation of Labor approved a resolution calling "for the abolition of convict labor when in competition with free labor."9 7.Topeka State Journal,February 25, 1891,col.3-4,p.1. g.Sixth Annual,215. 9.Sixth Annual, 124. 13 The eight-hour bill, Senate Bill 151, failed in the Kansas senate March 6, 1891, with the prevailing wage section removed But by March 10, when the prevailing wage section was put back in,the bill became law.This first prevailing wage law stated: That not less than the current rate of per diem wages in the locality where the work is performed shall be paid to laborers, workmen, mechanics and other persons so employed by or on behalf of the state of Kansas....1 o We do not know the immediate impact of the Kansas prevailing wage law. But a report from the Oklahoma labor commissioner in 1910 may well have applied to Kansas. The Oklahoma law was patterned after the Kansas act. It was passed in 1908. It was reported to have had the intended effect of setting wage and hour standards not only on public works but in related labor markets. The Oklahoma Commissioner of Labor stated in 1910: The eight hour law has been of inestimable value to the laboring men of this state....The common laborer, who was heretofore employed ten and twelve hours per day, is now, under the provisions of this bill, allowed to work but eight hours....The law has not only affected the laborers and those who are dependent upon this class of work for a living, but it has gone further, and in many localities has gradually forced railroad companies, private contractors [i.e. private construction] and people of that class to pay a high rate of wages for unskilled labor.t 1 The Federal Davis-Bacon Act(1931) > For four years before the 1931 passage of the Davis-Bacon Act, 14 bills were introduced in Congress to establish prevailing wages in construction.Republican Representative Robert L. Bacon(NY)in 1927 introduced the first bill proposing a prevailing wage for construction,H.R. 17069.This member of Congress justified his measure as follows: The Government is engaged in building in my district a Veteran's Bureau hospital. Bids were asked for. Several New York contractors bid,and in their bids,of course,they had to take into consideration the high labor standards prevailing in the State of New York...The bid,however,was let to a firm from Alabama who had brought some thousand non-union laborers from Alabama into Long Island, 10.L. 1891 Ch. 114 p.192-193. 11 Chas.L Daugherty,Labor Commissioner, Oklahoma Department of Labor,Third Annual Report, Oklahoma City,OK,1910,p.327.The primary concern in both Kansas and Oklahoma was to use public works hours and wage policies to set and improve local labor standards. A typical enforcement case in Oklahoma as reported by the Labor Commissioner follows: [Anadarko.May 10. 19081 We were advised that the O'Neill Construction Company had cut the wages on public works at Anadarko from twenty-five cents to seventeen and one-half cents per hour....[C]ontract was taken with the understanding that twenty-five cents per hour should be paid. The work was not progressing as rapidly as necessary to the cost within the estimate,hence the contractors tried to take advantage of the situation by reducing pay. After thoroughly discussing the matter before the[city] council and contractor,the wages were restored to twenty-five cents. (p.320) Second Annual Report Oklahoma Labor Commissioner Chas.L Daugherty,Oklahoma City,OK,August 7, 1909. 14 N.Y.;into my district.They were herded onto this job,they were housed in shacks,they were paid a very low wage,and the work proceeded...It seemed to me that the federal Government should not engage in construction work in any state and undermine the labor conditions and the labor wages paid in that State...The least the federal Government can do is comply with the local standards of wages and labor prevailing in the locality where the building construction is to take place.t z Hearings for a federal prevailing wage law began in 1927 and continued in 1928 and 1930,but no bill was passed. On March 3, 1931,Bacon's original proposal,which he had reintroduced as H.R. 16619,was signed into law by Republican President Herbert Hoover.13 The Davis-Bacon Act required payment of prevailing wages on federally financed construction projects.However,the original language of the law was vague,and prevailing wages generally were not determined before the acceptance of bids.In 1935,Democratic President Roosevelt signed clarifying amendments to the act,which became the basis of the current Davis-Bacon Act. In 1935,Roosevelt's Secretary of Labor,Francis Perkins,established the original rules for determining the Davis-Bacon prevailing rates.The prevailing wage was said to be the wage paid to the majority,if a majority existed; if not,the 30-percent rule was used. The 30-percent rule means if 30 percent of the workers in an area are paid the same rate,that rate becomes the prevailing rate there. The 30-percent rule often resulted in the union wage being the prevailing wage.If the 30-percent rule did not apply,because at least 30 percent of the workers in a given occupation in the local labor market did not receive the same wage rate,the average wage rate was paid to workers doing the same job.The prevailing wage was determined this way for 50 years. In 1985,Republican President Reagan changed administration of Davis-Bacon,creating the 50-percent rule.This rule holds that if 50%plus one wage rates for an occupation in a local labor market are the same to the penny,then that wage rate is said to prevail. If no one wage rate accounts for more than 50%of all wage rates for an occupation in a local labor market,then the average wage rate for that occupation prevails. Under the old rules,if union wage rates accounted for more than 30%of all wage rates for an occupation,then the union wage rate prevailed. Under the new rules,union wage rates must represent more than 50%of all wage rates in an occupation before union wage scales prevail under Davis-Bacon. 3 Republican President Herbert Hoover supported the passage of the Davis Bacon Act. The Act was named after Republican Representative Robert Bacon and Republican Senator James Davis. Hoover signed the Act in 1931. 12 U.S. House of Representatives,Hearings before the Committee on Labor on HR 17069,Efl`Congress,2' Session,p.2,February 18, 1927. 13 Hearings Before the Committee on Labor,House of Representatives-Seventy-First Congress.January31, 1931.Bacons proposal was re-introduced in 1930 as H.R.9232 by Congressman Elliot W.Sproul from Illinois,while Bacon proposed a complementary bill. 15 Kentucky(1940). Seven states passed prevailing wage laws between 1891 (Kansas)and 1923 (Nebraska). Seventeen states,including Kentucky(1940)passed prevailing wage laws between 1931 and 1940. Eventually all but nine states would pass prevailing wage laws. (See Table 2.) The first record I have found regarding the history and implementation of the Kentucky law comes from a visit by U.S. Labor Department officials to he Kentucky Labor Commissioner in 1951. A memorandum of that visit summarized the state of the law 11 years after its passage. The Kentucky law (See Baldwin's Kentucky Revised Statutes, Sections 337.150- 337.550 inclusive; and Section 337.990)provides for the following: 1. The contracting agency before advertising for bids, is required to ascertain prevailing rates for labors [sic.] workmen, mechanics, etc., in the locality where the work is to be performed. The schedule must be made part of the specification and part of any contract resulting therefrom. 2. The wages paid must not be less than those indicated by the schedule and if more than 90 days relapse between the establishment of the schedule and the confirmation of the contract the appropriate public authority must make a redetermination. 3. Contractors and subcontractors are charged with strict compliance and must pay the wages in legal tender without deductions....Contractors are required to keep accurate payroll records and to post copies of the prevailing wage rate. 4. Work schedules are limited to 8 hours a day and 40 hours a week except in cases of emergency. (A footnote stated: "During the war an amendment was passed relaxing hours of work in connection with the construction of highways. Designation by the commissioner of highways is necessary, for the duration of the National emergency." This amendment was passed in 1942 and repealed in 1950. (Repealed 1950 Ky. Acts ch. 176,sec. 1. —Created 1942 Ky. Acts ch. 136,secs. 1 and 3.)] 5. The department (presumably the contracting department) is empowered to assist workers in collecting the full amount of wages due. The worker is authorized to bring a civil action on claims for unpaid wages. 6. Various penalties are provided for violations of wage and hour requirements. The U.S. Labor Department officials went on to note: It would seem that possibly the best solutions to [Kentucky Labor Commissioner] Mr. Willis's [sic.] problem would be the enactment of a good wage payment and wage collection law. Kentucky has laws relating to the 16 payment of wages...but there is no authority given to the commissioner of labor to take assignments from workers and to bring suit to collect.14 In the 1940s and 1950s, contracting agencies determined the prevailing wage. The role of the Labor Commissioner in enforcing the prevailing wage law was unclear. The state Attorney opined in 1945 that it was the du of the Labor Commissioner to assist public works General h' . p laborers in the collection of prevailing wages and the Commissioner should also institute criminal proceedings under the law.15 But apparently,this was not happening. We also get some idea of what a prevailing wage was interpreted to be in the 1940s from the Attorney General's opinions. In August of 1946, the Kentucky Attorney General states that the wage rate for the majority of workers in a locality is the proper rate to apply as the prevailing wage where there are insufficient employees in the area working under a collective bargaining t 6 vagueness in agreement. This was an effort on the part of the Attorney General to tackle the 40s and 19 S0s the law stated: the law regarding what was a prevailing wage. In the 19 , g g P g The prevailing wages shall be the wages paid in the locality under collective agreements between unions and employers at the time of the contract, if such agreements apply to a sufficient number of workers in the locality." Because contracting agencies determined the prevailing wage, each agency was free to decide what was "sufficient". The Attorney General was attempting to define and regularize the alternative procedure if a contracting agency found union contracts to be insufficient. The Attorney General's a q ' alternative required the use of what statisticians call the mode, or the most Ge effect the likely, thrs would be the union wage rate. So in , nl found wage rate. Most g commonly g Y� Attorney General was telling contracting agencies to go back to the collectively bargained rate whether or not the agency thought it covered a"sufficient"number of workers. This ambiguity got the law into trouble. In 1960, a Prevailing Wage Board was established to centralize the determination of prevailing wage rates. The amendment to the law required the Board to make a determination on prevailing wages based upon union scale in an area where most workers in the category under consideration were covered by union contracts. However, a case arose in a county in which the overwhelming majority of laborers available to work on a project were not covered by a union contract. The Kentucky Court of Appeals stated: The law is completely silent with respect to the scope of the Board's authority or duty in the establishment of prevailing wages when union rates are not applicable. This deficiency falls within the scope of the rule that where the intention of the legislature is so obscure as to defy a rational meaning,the law cannot be given effect.18 This led Kentucky to adopt the federal Davis-Bacon prevailing wage law formula to this problem. Through an amendment in 1962, the law defined prevailing wages as the wages that 14 Memo from John B Kneipple dated January 18, 1951 in the U.S.Labor Department,Prevailing Wage Division files. 15 Attorney General opinion,August 27, 1945. 16 Attorney General opinion,August 15, 1946. 17 Kentucky Revised Statues,1955. 18 Labor law Reports Summary,No.54,June 19, 1961. 17 are or have been paid to the majority of laborers...employed in each classification of construction upon reasonably comparable construction in the locality where the work is to be performed....If a majority does not receive the same rate, then the rate paid to the greater number—provided that it is at least 30 percent of those employe"etemnines the prevailing wage; or if less than 30 percent are receiving the same rate, then the average rate is prevailing. The 1962 amendment also gave guidance on what information was to be considered in determining the prevailing wage. These included 1. wage rates paid on recent previous public works in the same locality, 2. wage rates paid on reasonably comparable recent private construction projects in the locality,and 3. recent collective bargaining agreements between bona fide labor organizations and their contractors whenever applicable to a locality.19 The definition of a prevailing "wage' in the original act and in its 1962 amended form did not address the issue of fringe benefits. Starting in the early 1950s, collective bargaining agreements began to establish health benefits for construction workers. By the early 1960s, pension benefits were also being provided to construction workers through collective bargaining. At first these benefits were modest, maybe a nickel for pensions. But by the mid-1960s, these were increasingly important aspects of the construction labor market. They allowed construction workers to provide for their family's health and old age security for themselves and their spouses. Consequently, in 1968, the Kentucky legislature addressed this issue by breaking down the prevailing wage as the sum of"the basic hourly rate and an additional amount per hour equal to the hourly rate of contribution irrevocably made to a trustee or third person pursuant to an enforceable commitment to carry out a financially responsible plan for medical, pension, death or injury benefits 20 This language was similar to the Davis Bacon Act. So a law that began in 1940 as a prevailing "wage" law based on "sufficient" numbers of union workers in a locality as determined by the contracting agency, had by 1968 become a prevailing "total compensation" law. The determination of this prevailing wage was based on a switching rule that used collectively bargained rates where 30% of an occupation in a locality was covered by a collectively bargained contact. Otherwise, the average total compensation for the locality was used. By 1968, the Commissioner of Labor was responsible for making these determinations and along with individual rights of redress, the Commissioner was responsible for enforcing the law. In 1982, political winds were blowing against prevailing wage laws. Florida had repealed its law in 1979. U.S. Senator Orin Hatch held hearings on the federal Davis Bacon prevailing wage law aimed at its repeal. Alabama repealed its law in 1980 and Utah repealed its law in 1981. Kentucky did not repeal its law, but it exempted all public institutions of learning from the laws coverage. Kentucky also exempted city, county- and urban-county governments from the laws coverage if the construction was financed with less than 50% state funds. It also established a threshold for coverage of any project of 9250,000. This threshold was to rise with the consumer price index. Also, following regulations instituted by the Reagan Administration regarding the Davis-Bacon Act, the switching rule between most commonly found wage rates and the average wage rate was changed from 30% to 50%. With these changes in mind, a 19 U.S.Labor Department,Prevailing Wage Division,Annual State Report,Kentucky, 1963-64. 20 Kentucky,Regular Session, 1968 New Laws Page 509,Senate Bill No. 123. 18 comparison of the federal Davis Bacon Act and the Kentucky law in 1983 was made by the U.S. Labor Department. That comparison is presented in Table 1. In 1996,Kentucky reversed course and re-applied the state prevailing wage regulation to public schools and to municipal and county construction not using a majority of state funds. It also clarified the threshold provision,preventing projects from being broken into sub-projects below the threshold. Also,the threshold is no longer tied to the consumer price index. Furthermore,the eight-hour provision that had been in the law from the beginning was modified to permit a four-tens work week. With these changes in hand,the Kentucky prevailing wage law can be said to be similar to the current federal Davis-Bacon Act. "A' I E11- 29 19 Table 1:A Comparison of the Federal and Kentucky Prevailing Wage Laws Circa 1983 Federal entuc y IN luepartMent 01 Lauer KentUCKY luepartment of Lauer Contracts to which the United States or the All buildings, roads. streets, alleys, District of Columbia is a party. for construc- sewers,ditches, sewage disposal plants, tion,alteration,and/or repair including paint- waterworks and all other structures or work ing and decorating, of public buildings or except for buildings constructed as institu- public works of the United States or the tions of learning constructed under contract District of Columbia within their geographical with any public authority and involving limits, and which requires or involves the laborers,workmen, and mechanics. However, employment of mechanics and/or laborers. the law does not apply to construction con- ducted by a city, county, urban-county government or school district unless such construction is financed with 50 percent or more of State funds. $2,000.00 $250,000, with an adjustment each July 1 based on change in the CPI. asic hourly rate ot pay plus fringe benetits. Basic hourly rate o pay p us Hoge ene us. IMM Rate of wages paid in the area in which the Rate of wages paid in the locality where work is to be performed, to the majority of the work is to be performed. to the majority those employed in that classification in similar. of those employed in each classification of construction in the area. construction. If there is not a majority paid at the same If there is not a majority paid at the rate, then the rate paid to the greater number: same rate, then the basic hourly rate of pay Provided, such number constitutes 30 percent of shall be the average basic hourly rate. those employed- (The elimination of the 30 (The commissioner shall conduct a public percent rule is one of a number of proposed hearing in the locality, after notice has changes in the Federal regulations. These been given, for the purpose of making initial proposed regulations were the subject of a determinations or revisions of a prevailing U.S. District Court decision which has been wage schedule for the construction of public appealed. Since the 30 percent rule was not works. The locality may be one or more enjoined by this decision, steps to implement counties, but may not be less than an entire the change are currently under consideration.) county.) If less than 30 percent of those so employed If both Federal and State rates are appli- receive the same rate, then the average rate. cable, those wages in each classification which are higher shall prevail. Statements showing wage rates paid on projects. Wage rates paid on previous public works Signed collective bargaining agreements. constructed in the localities. Wage rates determined for public construction Wage rates previously paid on reasonably by State and local officials pursuant to pre- icomparable private construction projects vailing wage legislation. constructed in the localities. Information furnished by Federal and State Collective bargaining agreements agencies. Any other information pertinent to the deter- mination of prevailing wage rates. Wage surveys conducted by the U.S. Department of Labor, Federal agencies, or other interested parties. 20 federal entuc y Payroll records are to be submitted weekly. Payroll records need not be submitted. They must be open to inspection by the Commissioner of Labor, and may not be destroyed or.removed from the State for one year following completion of the improvement in connection with which they are made. If the contracting officer finds that any Any laborer, workman, or mechanic employed on laborer or mechanic employed by the contractor public works may file a complaint of any or any subcontractor directly on the site of violation with the department. The department the work covered by the contract has been or is shall assist in the collection of claims of being paid a rate of wages less than that re- wages due. The commissioner shall investigate quired by the contract. the Government may, by and enforce the prevailing wage provisions to written notice to the contractor, terminate his the fullest and shall bring all actions to right to proceed with the work, or the part of collect wages due and shall take action the work for which there has been a wage viola- against any contractor or subcontractor to tion. and to complete the work by contract or restrain violations of the law. Such otherwise, and the contractor and his sureties contractor or subcontractor may be held shall be liable to the Government for any excess ineligible to bid on public works until costs incurred. determined to be in substantial compliance. The Comptroller General of the U.S. is author- A laborer, workman, or mechanic may by civil ized to pay directly to laborers and mechanics action recover any sum due as the result of any wages found to be due from any accrued pay- the failure of the employer to comply with ments withheld under the terms of the contract. the prevailing wage law. The commissioner may also bring any legal action necessary to collect claims on behalf of any or all laborers, workmen, or mechanics The Comptroller General of the U.S. is authorized Any contractor or subcontractor who violates and directed to distribute a list to all depart- any wage or work hours provision in any ments of the Government giving the names of contract under the prevailing wage law shall persons or firms found to have violated the law. be fined not more than $100 for each offense No contract is to be awarded to the persons or and be required to make full restitution to firms appearing on the list until three years all employees to whom legally indebted as a have elapsed from the date of publication of result of the violation. The prime contractor the list. shall be jointly and severally liable with a subcontractor for wages due an employee of the subcontractor.For a flagrant or re- peated violation the offending contractor or subcontractor shall be barred from bidding on or working on, any and all public works con- tracts for a period of two years from the date of the last offense. Each day of violation constitutes a separate offense, and the violation as it effects each individual worker shall constitute a separate offense. Source: U.S. DEPARTMENT OF LABOR Employment Standards Administration Office of State Liaison and Legislative Analysis Division of State Employment Standards Programs 21 Conclusions. Prevailing wage laws emerged from a concern that cutthroat competition over wages in construction would lead the industry down a low-wage, low-skill development path. This was said to put the quality of construction at risk and lead to an itinerant, footloose low-wage construction labor force. Poor construction workers would make poor neighbors and potential burdens on the community. Reasonably paid construction workers, on the other hand, held out the possibility of being solid neighbors, good citizens and productive members of the community. Government, by the operation of prevailing wage laws, was supposed to get out of the business of cutting government costs by cutting the wages of its citizens. Whatever labor standards had been established, whatever wages prevailed in a local community, that is what the law said government should pay on public works. The bidding process on government works differs significantly from the private sector. In the private sector, the owner of the construction project can overlook the lowest bidder for a higher bid that promises better quality or performance. In the public sector, the lowest bonded bid must be accepted. Contractors have an incentive to shave costs on the initial bid and hope to make up those costs in change orders or in a favorable interpretation of the jobs specifications. Owner dissatisfaction cannot lead to debarment from subsequent public jobs as long as the letter of the law and specifications are adhered to. This bidding structure puts an added downward pressure on wages and an upward pressure on hidden costs. But if these laws are successful in focusing competition on the factors that raise the productivity of construction in the long run and help justify a better-paid construction labor force, then three things must be true. First, where prevailing wage laws exist, training in construction must be more common and of a higher quality. Second, where prevailing wage laws exist, the income and benefits of construction workers must be higher. And third, despite higher wages, income and benefits, where prevailing wage laws exist, construction costs must be roughly equivalent to construction costs where prevailing wage laws are absent. We will soon turn to the task of investigating these three questions. But first, we examine a unique criticism of prevailing wage laws. Some critics allege that prevailing wage laws in general and the Davis- Bacon Act in particular are racist laws. This criticism alleges that these laws were originally passed to exclude blacks and other minorities from construction. And furthermore, this criticism argues that the current effect of prevailing wage laws is to exclude blacks and other minority workers. 22 Table 2: Prevailing Wage Laws by State,Year Passed and Repealed States having Year prevailing wage laws passed States without prevailing wage laws Alaska 1931 Georgia Arkansas 1955 Iowa California 1931 Mississippi Connecticut 1935 North Carolina Kentucky 1933 North Dakota District of Columbia 1931 South Carolina Hawaii 1955 South Dakota Illinois 1931 Vermont Indiana 1935 Virginia Kentucky 1940 Maine 1933 Maryland 1945 Massachusetts 1914 Michigan 1965 Minnesota 1973 Missouri 1957 Montana 1931 Nebraska 1923 Nevada 1937 New Jersey 1913 New Mexico 1937 New York 1894 Ohio 1931 Oklahoma* 1909 Oregon 1959 Pennsylvania 196161 Rhode Island 1935 Tennessee 1953 Texas 1933 Washington 1945 West Virginia 1933 Wisconsin 1931 Wyoming 1967 States that repealed Year Year of prevailing wage laws passed repeal Alabama 1941 1980 Arizona 1912 1984 Colorado 1933 1985 Florida 1933 1979 Idaho 1911 1985 Kansas 1891 1987 Louisiana 1968 1988 New Hampshire 1941 1985 Utah 1933 1981 *The enforcement of Oklahoma's law was judicially suspended in 1995. 23 Appendix: the Current Kentucky Prevailing Wage Law 337.010 Definitions for chapter and specific ranges in chapter. (1)As used in this chapter,unless the context requires otherwise: (a) "Commissioner" means commissioner of the Department of Workplace Standards under the direction and supervision of the secretary of the Labor Cabinet; (b) DeP P "Department"means Department of Workplace Standards in the Labor Cabinet; " includes an compensation due to an employee b reason of his employment, including (c) Wages include y pY overtime a severance or dismissal pay, earned bonuses, salaries, commissions, vested vacation pay, o pay, p Y and any other similar advantages agreed upon by the employer and the employee or provided to employees as an established policy. The wages shall be payable in legal tender of the United States or checks on banks convertible into cash on demand at full face value,subject to the allowances made in this chapter; (d) "Employer" is any person, either individual, corporation, partnership, agency, or firm who employs an employee and includes any person, either individual, corporation, partnership, agency, or firm acting directly or indirectly in the interest of an employer in relation to an employee;and (e)"Employee"is any person employed by or suffered or permitted to work for an employer. (2) As used in KRS 337.275 to 337.325, 337.345, and KRS 337.385 to 337.405, unless the context requires otherwise: (a) "Employee" is any person employed by or suffered or permitted to work for an employer,but shall not include: 1.Any individual employed in agriculture; 2. Any individual employed in a bona fide executive, administrative, supervisory, or professional capacity, or in the capacity of outside salesman, or as an outside collector as the terms are defined by administrative regulations of the commissioner; 3.Any individual employed by the United States; 4. Any individual employed in domestic service in or about a private home.The provisions of this section shall include individuals employed in domestic service in or about the home of an employer where there is more than one(1)domestic servant regularly employed; 5. Any individual classified and given a certificate by the commissioner showing a status of learner, apprentice, worker with a disability, sheltered workshop employee, and student under administrative procedures and administrative regulations prescribed and promulgated by the commissioner. This certificate shall authorize employment at the wages, less than the established fixed minimum fair wage rates, and for the period of time fixed by the commissioner and stated in the certificate issued to the person; 6. Employees of retail stores, service industries, hotels, motels, and restaurant operations whose average annual gross volume of sales made for business done is less than ninety-five thousand dollars($95,000) for the five (5) preceding years exclusive of excise taxes at the retail level or if the employee is the parent, spouse,child,or other member of his employer's immediate family; 7. Any individual employed as a baby-sitter in an employer's home, or an individual employed as a 1 person or the persons immediate family,to care for that companion b a sick convalescing, or elder y p p Y g YP sick,convalescing,or elderly person and whose principal duties do not include housekeeping; 8.Any individual engaged in the delivery of newspapers to the consumer; 9. Any individual subject to the provisions of KRS Chapters 7, 16, 27A, 30A,and 18A provided that the secretary of the Personnel Cabinet shall have the authority to prescribe by administrative regulation those emergency employees, or others, who shall receive overtime pay rates necessary for the efficient operation of government and the protection of affected employees; 10. Any employee employed by an establishment which is an organized nonprofit camp, religious, or nonprofit educational conference center, if it does not operate for more than seven (7) months in any calendar year;or -four 24 hour residential care on the employer's 11. An employee whose function is to provide twenty ( ) p Y dr who are primarily dependent, neglected, and abused and who are premises in a parental role to children p y p g in the care of private, nonprofit childcaring facilities licensed by the Cabinet for Families and Children under KRS 199.640 to 199.670. 24 (b) "Agriculture" means farming in all its branches, including cultivation and tillage of the soil; dairying; production, cultivation, growing, and harvesting of any agricultural or horticultural commodity; raising of livestock, bees, furbearing animals, or poultry; and any practice, including any forestry or lumbering operations, performed on a farm in conjunction with farming operations, including preparation and delivery of produce to storage,to market,or to carriers for transportation to market; (c) "Gratuity" means voluntary monetary contribution received by an employee from a guest, patron, or customer for services rendered; (d) "Tipped employee" means any employee engaged in an occupation in which he customarily and regularly receives more than thirty dollars($30)per month in tips;and (e)"U.S.C."means the United States Code. (3)As used in KRS 337.505 to 337.550,unless the context requires otherwise: (a) Construction includes construction, reconstruction, improvement, enlargement, alteration, or repair of any public works project by contract fairly estimated to cost more than two hundred fifty thousand dollars ($250,000). No public works project, if procured under a single contract and subject to the requirements of this section, may be divided into multiple contracts of lesser value to avoid compliance with the provisions of this section; (b) "Contractor" and "subcontractor" include any superintendent, foreman, or other authorized agent of any contractor or subcontractor who is in charge of the construction of the public works or who is in charge of the employment or payment of the employees of the contractor or subcontractor who are employed in performing the work to be done or being done by the contractor or subcontractor under the particular contract with any public authority; (c) 1. "Locality" shall be determined by the commissioner. The commissioner may designate more than one (1) county as a single locality,but if more than one (1) county is designated,the multicounty locality shall not extend beyond the boundaries of a state Senatorial district. The commissioner shall not designate less than an entire county as a locality. If there is not available in the locality a sufficient number of competent, skilled laborers, workmen, and mechanics to efficiently and properly construct the public works, locality shall include any other locality near est the one in which the work of construction is to be performed and from which such available skilled laborers, workmen, and mechanics may be obtained in sufficient number to perform the work;and 2. "Locality" with respect to contracts advertised or awarded by the Transportation Cabinet of this state shall be determined by the secretary of the Transportation Cabinet. The secretary may designate any number of counties as constituting a single locality. The secretary may also designate all counties of the Commonwealth as a single locality,but he shall not designate less than an entire county as a locality; (d) "Public authority" means any officer,board,or commission of this state,or any political subdivision or department thereof in the state, or any institution supported in whole or in part by public funds, including publicly owned or controlled corporations, authorized by law to enter into any contract for the construction of public works and any nonprofit corporation funded to act as an agency and instrumentality of the government agency in connection with the construction of public works, and any "private provider", as defined in KRS 197.500, which enters into any contract for the construction of an "adult correctional facility",as defined in KRS 197.500;and (e) "Public works" includes all buildings, roads, streets, alleys, sewers, ditches, sewage disposal plants, waterworks, and all other structures or work, including "adult correctional facilities", as defined in KRS 197.500,constructed under contract with any public authority. (4) If the federal government or any of its agencies furnishes by loans or grants any part of the funds used in constructing public works, and if the federal government or its agencies prescribe predetermined prevailing minimum wages to be paid to mechanics, workmen, and laborers employed in the construction of the public works, and if KRS 337.505 to 337.550 is also applicable, those wages in each classification which are higher shall prevail. Effective:July 15,1998 History: Amended 1998 Ky. Acts ch. 154,sec. 92,effective July 15,1998;ch.426,sec. 558,effective July 15, 1998; and ch. 606, sec. 113, effective July 15, 1998. -Amended 1996 Ky. Acts ch. 48, sec.1, effective July 15,1996;ch. 100, sec. 1,effective July 15,1996; and ch.115,sec. 1,effective July 15,1996.-Amended 1994 Ky. Acts ch. 405, sec. 85, effective July 15, 1994; and ch. 492, sec. 1, effective July 15, 1994. -Amended 1986 Ky. Acts ch. 208, sec. 2, effective July 15, 1986. -Amended 1984 Ky. Acts ch. 414, sec. 12, effective July 13, 1984. - Amended 1982 Ky. Acts ch. 54, sec. 1, effective July 15, 1982. --Amended 1978 Ky. Acts ch.141,sec.1,effective June 17,1978;and ch.340,sec. 1,effective June 17,1978.--Amended 1976 Ky.Acts ch. 223,sec. 1. --Amended 1974 Ky. Acts ch. 341,sec.1;and ch.391,sec.1. --Amended 1970 Ky.Acts ch. 25 33,sec.1. —Amended 1968 Ky. Acts ch.100,sec. 6. —Amended 1966 Ky. Acts ch.158,sec. 1.—Recodified 1942 Ky. Acts ch. 208, sec. 1, effective October 1, 1942, from Ky. Stat. secs. 1599c-4, 1599c-39, 2290c-1, 2290c-2,4767a-1,4767a-17. Legislative Research Commission Note (7/15/98).This section was amended by 1998 Ky. Acts chs.154, 426,and 606 which do not appear to be in conflict and have been codified together. 337.505 Definition of"prevailing wage,"fringe benefits included For the purpose of KRS 337.505 to 337.550, the term prevailing wage for each classification of laborers, workmen, and mechanics engaged in the construction of public works within the Commonwealth of Kentucky,means the sum of: (1) The basic hourly rate paid or being paid subsequent to the labor commissioner's most recent wage determination to the majority of laborers, workmen, and mechanics employed in each classification of construction upon reasonably comparable construction in the locality where the work is to be performed; such rate shall be determined by the commissioner in accordance with paragraphs (a), (b), and (c) of subsection(3) of KRS 337.520; in the event that there is not a majority paid at the same rate,then the basic hourly rate of pay shall be the average basic hourly rate which shall be determined by adding the basic hourly rates paid to all workers in the classification and dividing by the total number of such workers, and (2) An additional amount per hour equal to the hourly rate of contribution irrevocably made or to be made by an employer on behalf of employees within each classification of construction to a trustee or to a third person pursuant to an enforceable commitment to carry out a financially responsible plan or program, which was communicated in writing to the employees affected, for the following fringe benefits: medical or hospital care, pensions on retirement, death compensation for injuries or illness resulting from occupational activity or insurance to provide any of the foregoing, unemployment benefits, life insurance, disability and sickness insurance, accident insurance, vacation and holiday pay, defraying costs of apprenticeship or other similar programs, or other bona fide fringe benefits,but only where the employer is not required by other federal, state or local law to provide any of such benefits: provided, said additional amount may, at the discretion of the employer, be paid either in cash to the employee or by contributions br fringe benefits, or partly in cash and partly by such contributions, it being the intention of this subsection to recognize fringe benefits as a part of the prevailing wage rate where made in accordance with this subsection. Effective:July 15,1982 History: Amended 1982 Ky. Acts ch. 54, sec. 2,effective July 15, 1982. —Amended 1974 Ky. Acts ch.341, sec.2.—Amended 1968 Ky.Acts ch.33,sec.1.--Created 1962 Ky.Acts ch.173,sec.1. 337.510 Public authority's duties as to inclusion of prevailing wage in proposals and contracts. (1) Before advertising for bids or entering into any contract for construction of public works,every public authority shall notify the department in writing of the specific public work to be constructed, and shall ascertain from the department the prevailing rates of wages for each classification of laborers, workmen, and mechanics for the class of work called for in the construction of such public works in the locality in rate of wages shall include a med. This schedule of the prevailing g where the work is to be perfor p statement that it has been determined in accordance with the provisions of KRS 337.505 to 337.550 and shall be attached to and made part of the specifications for the work and shall be printed on the bidding blanks and made a part of every contract for the construction of public works. (2)The public authority advertising and awarding the contract shall cause to be inserted in the proposal and contract a stipulation to the effect that not less than the prevailing hourly rate of wages as determined by the commissioner shall be paid to all laborers, workmen, and mechanics performing work under the contract. It shall also require in all the contractor's bonds that the contractor include such provisions as will guarantee the faithful performance of the prevailing hourly wage clause as provided by contract. It shall be the duty of the public authority awarding the contract, and its agents and officers, to take cognizance of all complaints of all violations of the provisions of KRS 337.505 to 337.550 committed in the course of the execution of the contract,and when making payments to the contractor becoming due under the contract, to withhold,and retain therefrom all sums and amounts due and owing as a result of any violation thereof. It shall be lawful for any contractor to withhold from any subcontractor under him sufficient sums to cover any penalties withheld from him by the awarding authority, on account of the 26 subcontractor's failure to comply with the terms thereof and if payment has already been made to him, the contractor may recover from him the amount of the penalty in a suit at law. History:Amended 1974 Ky.Acts ch.341,sec.3. —Amended 1970 Ky. Acts ch. 33,sec. 3. —Amended 1960 Ky. Acts ch. 56, sec. 1,effective June 16,1960. --Recodified 1942 Ky.Acts ch.208,sec.1,effective October 1,1942,from Ky.Stat.sec.2290c-2. 337.512 Duties of individual officers with respect to prevailing wage law. (1) No public official, authorized to contract for or construct public works shall fail,before advertising for bids or undertaking such construction, to ascertain from the commissioner the prevailing rates of wages as provided in KRS 337.505 to 337.550. (2) No member of a public authority authorized to contract for or construct public works shall vote for the award of any contract for the construction of such public works, or vote for the disbursement of any funds on account of the construction of such public works, unless such public authority has first ascertained from the commissioner the prevailing rates of wages of laborers,workmen, and mechanics for the classes of work called for by such public works in the locality where the work is to be performed and the determination of prevailing wages has been made a part of the proposal specifications and contract for such public works. History:Amended 1974 Ky.Acts ch.341,sec.4.—Created 1970 Ky.Acts ch.33,sec.4. 337.520 Determination of prevailing wages—Administrative regulations—Filing wage contract. (1) The commissioner shall make initial determinations and current revisions of schedules of rates of prevailing wages,of the amount of fringe benefits included as defined in KRS 337.505,and the number of hours applicable. The commissioner may promulgate administrative regulations to carry out the provisions and purposes of KRS 337.505 to 337.550 and to prevent their circumvention or evasion. The administrative regulations shall not include a provision that each contractor and subcontractor furnish a sworn affidavit with respect to the wages paid each employee. No administrative regulation shall be issued by the commissioner except upon reasonable notice to, and opportunity to be heard by, any interested person. (2) The commissioner shall require the filing of all wage contracts of all laborers, workmen, and mechanics in this state which have been agreed to between bona fide organizations of labor and an employer or associations of employers. The contracts shall be filed within ten (10) days after they are signed. (3)The commissioner shall have the authority to determine schedules and current revisions of the rates of prevailing wages as defined in KRS 337.505,but in no case shall the commissioner determine wages to be paid for a legal day's work to laborers, workmen, and mechanics engaged in the construction of public works at less than the prevailing wages paid in the localities. The commissioner, in determining what rates of wages prevail,shall consider the following criteria: (a) Wage rates paid on previous public works constructed in the localities. In considering the rates, the commissioner shall ascertain,insofar as practicable,the names and addresses of the contractors,including subcontractors,the locations, approximate costs, dates of construction and types of projects, the number of workers employed on each project, and the respective wage rates paid each worker who was engaged in the construction of these projects. (b)Wage rates previously paid on reasonably comparable private construction projects constructed in the localities. In considering the rates the commissioner shall ascertain, insofar as practicable,the names and addresses of the contractors, including subcontractors, the locations, approximate costs, dates of construction and types of projects, the number of workers employed on each project, and the respective wage rates paid each worker who was engaged in the construction of these projects. (c) Collective bargaining agreements or understandings between bona fide organizations of labor and their employers located in the Commonwealth of Kentucky which agreements apply or pertain to the localities in which the public works are to be constructed. (4) The wage rates to be used by the public authority in a contract for the construction of public works shall be the prevailing wage as of the date the public works project is advertised and offered for bid. If contracts are not awarded within ninety (90) days after the date of offering for bid, the public authority shall ascertain the prevailing rate of wages from the department before the contract is awarded. The schedule or scale of prevailing wages shall be incorporated in and made a part of each contract. (5) The 27 commissioner may promulgate administrative regulations authorizing the employment of apprentices and trainees in skilled trades at wages lower than the applicable prevailing wage. Effective:July 15,1996 History. Amended 1996 Ky. Acts ch. 48, sec. 2, effective July 15, 1996. --Amended 1982 Ky. Acts ch. 54, sec. 3, effective July 15, 1982. —Amended 1974 Ky. Acts ch. 341, sec. 5. —Amended 1970 Ky. Acts ch. 33, sec. 6. — Amended 1968 Ky. Acts ch. 33, sec. 2. — Amended 1962 Ky. Acts ch. 173, sec. 2(1) to (4). — from Recodified 1942 Ky.Acts ch.208,sec.1,effective October 1,1942,f m Ky.Stat.sec.2290c-3. 337.530 Contractor to pay prevailing wages and post rates—Payroll records— On-site inspections. (1) Where a prevailing rate of wages has been determined and prescribed, the contract executed between a public authority and the successful bidder or contractor shall contain a provision requiring the successful bidder and all of his subcontractors to pay not less than the rate of wages so established.The successful bidder or contractor and all subcontractors shall strictly comply with these provisions of the contract. subcontractors required b KRS 337.505 to 337.550 and by contracts with any All contractors and subco y 2 �1 public authority to pay not less than the prevailing rate of wages, shall pay such wages in legal tender without any deductions. These provisions shall not apply where the employer and employee enter into an agreement in writing at the beginning of or during any term of employment covering deductions for food, sleeping accommodations or any similar item if this agreement is submitted by the employer to the department and is approved by the department as fair and reasonable. All contractors and subcontractors affected by the terms of KRS 337.505 to 337.550 shall keep full and accurate payroll records covering all disbursements of wages to their employees to whom they are required to pay not less than the prevailing rate of wages. Such records shall indicate the hours worked each day by each employee in each classification of work and the amount aid each employee for his work in each classification.They shall cla P be open to the inspection and transcript of the commissioner or his authorized representative at any reasonable time, and shall be in compliance with all regulations issued by the commissioner. These payroll records shall not be destroyed or removed from this state for one (1) year following the completion of the improvement in connection with which they are made. (3) Each contractor and subcontractor subject to the provisions of KRS 337.505 to 337.550 shall post and keep posted in a conspicuous place or places at the site of the construction work a copy or copies of t with the pub lic authority, i e contract h'- revailin rates of wages and working hours as prescribed in p prevailing g g showing the rates of wages prescribed and the working hours for each class of laborers, workmen, and mechanics employed by him in the work of constructing the public works provided for in the contract with the public authority. (4) Every employer shall permit the commissioner or his authorized agents to question any of his employees at the site of the public work and during work hours in respect to the wages paid, hours worked and duties of such employee or other employees. Effective:July 15,1982 History: Amended 1982 Ky. Acts ch. 54, sec. 4, effective July 15,1982. —Amended 1974 Ky. Acts ch. 341, sec.7. --Amended 1970 Ky. Acts ch.33,sec. 8. —Amended 1962 Ky. Acts ch. 173,sec.4. --Amended 1960 Ky. Acts ch. 56, sec. 3. —Recodified 1942 Ky. Acts ch. 208, sec. 1,effective October 1,1942,from Ky.Stat. sec.2290c-4. 337.540 Limitation of working hours—Exceptions—Overtime. (1) Every public authority, before advertising for bids, shall include with the schedule of wages a provision that no laborer, workman, or mechanic shall be permitted to work more than eight(8)hours in one (1) calendar day, which shall constitute a legal day's work; nor more than forty (40) hours in one (1) week, which shall constitute a legal workweek, except in cases of emergency caused by fire, flood, or damage to life or property.This limitation of work hours shall be made a part of the specifications for the work and printed on bid blanks where the work is done by contract and shall be incorporated as a part of each contract. This shall not prohibit any laborer, workman, or mechanic from working more than eight (8) hours in one (1) calendar day, but not more than ten (10) hours in one (1) calendar day where the employee and employer enter into an agreement in writing prior to the working of any one (1) day in excess of eight(8)hours,or where provided for in a collective bargaining agreement. 28 (2) No laborer, workman,or mechanic shall be permitted to work more than eight(8)hours in any one (1) calendar day,nor more than forty (40)hours in any one (1)week,except in cases of emergency caused by fire,flood, or damage to life or property, on the construction of public works which is being constructed under contract with any public authority.This shall not prohibit any laborer,workman,or mechanic from working more than eight (8) hours in one (1) calendar day, but not more than ten (10) hours in one (1) calendar day where the employee and employer enter into an agreement in writing prior to the working of any one(1)day in excess of eight(8)hours,or where provided for in a collective bargaining agreement. (3) Any laborer,workman,or mechanic worked in excess of eight(8) hours per day or forty (40)hours per week, except in cases of emergency shall be paid not less than one and one-half (1-1/2) times the basic hourly rate of pay as defined and fixed under this chapter for all overtime worked, and each contract with any public authority for the construction of public works shall so provide. In any case where a laborer, workman, or mechanic works it excess of eight (8) hours per day, but not more than ten (10) hours per day in accordance with subsection(2) of this section,it will not be a violation of this subsection provided the laborer, workman, or mechanic who works in excess of ten (10) hours in any one (1) calendar day shall be paid not less than one and one-half(1-1/2)times the basic hourly rate of pay. (4)The determination of exception provided in this section of when an emergency exists shall be made by the public authority letting the contract. Effective:July 15,1994 History:Amended 1994 Ky.Acts ch.258,sec.1,effective July 15,1994.—Amended 1974 Ky.Acts ch.341,sec.8.—Amended 1968 Ky.Acts ch.33,sec.3.— Recodified 1942 Ky.Acts ch.208,sec.1,effective October 1,1942,from Ky.Stat. sec.2290c-5. 337.548 Injunction of violation of prevailing wage law. If it is found that a public authority has not complied with KRS 317,101 to 317,110, the commissioner shall give notice thereof in writing to such public authority. Sufficient time may be allowed for compliance therewith as the commissioner deems necessary. After the expiration of the time prescribed in the notice,the department shall at the earliest possible time bring suit in the Circuit Court of the county in which such public body is located to enjoin the award of such contract for a public works or any further work or payments thereunder if the contract has been awarded until the requirements of such notice are complied with.The court may issue a temporary restraining order without notice to the defendant in such action. Upon final hearing thereof, if the court is satisfied that the requirements of the notice by the department to the defendant were not unreasonable or arbitrary, it shall issue an order enjoining the defendant from awarding such contract for a public works or any further work or payments thereunder if the contract has been awarded until the notice is complied with. Such injunction shall continue operative until the court is satisfied that the requirements of the notice have been complied with and the court shall have and exercise with respect to the enforcement of such injunctions all the power invested in it in other similar cases. Both the plaintiff and the defendant in such action have the same rights of appeal as are provided by law in other injunction actions. History.Created 1970 Ky.Acts ch.33,sec.9. 337.550 Department to aid in enforcement—Remedies of laborer. (1) Any laborer, workman, or mechanic employed on public works may file a complaint of any violation of any provision of KRS 337.505 to 337.550 with the department. The department shall assist him in the collection of claims of wages due him and shall also assist to the fullest extent in the administration and enforcement of KRS 337.505 to 337.550. The commissioner shall investigate and enforce the provisions of KRS 337.505 to 337.550 to the fullest and shall bring all actions to collect wages due any laborer, workman,or mechanic and shall take action against any contractor or subcontractor to restrain violations of KRS 337.505 to 337.550. If any contractor or subcontractor is found to be in violation of any provisions of KRS 337.505 to 337.550, then the commissioner shall inform the secretary for finance and administration of the Commonwealth of Kentucky,and the secretary for finance and administration shall hold such contractor or subcontractor ineligible to bid on public works,until such time as that contractor or subcontractor is in substantial compliance as determined by the commissioner. (2) A laborer, workman, or mechanic may by civil action recover any sum due him as the result of the failure of his employer to comply with the terms of KRS 337.505 to 337.550. The commissioner may also bring any legal action necessary to collect claims on behalf of any or all laborers,workmen,or mechanics. No employer shall take any punitive measure or action against an employee because such employee has 29 made a charge, testified, assisted or participated in any manner in an investigation,proceeding or hearing under KRS 337.505 to 337.550.The commissioner shall not be required to pay the filing fee,or other costs, in connection with such action. Effective:June 17,1978 History:Amended 1978 Ky. Acts ch.340,sec.5,effective June 17,1978. —Amended 1974 Ky.Acts ch.341, sec. 9. — Amended 1970 Ky. Acts ch. 33, sec. 10. — Recodified 1942 Ky. Acts ch. 208, sec. 1, effective October 1,1942,from Ky.Stat.sec.2290c-6. 337.990 Penalties. The following civil penalties shall be imposed, in accordance with the provisions in KRS 336.985, for violations of the provisions of this chapter: (1) Any firm, individual, partnership, or corporation that violates KRS 337.020 shall be assessed a civil penalty of not less than one hundred dollars ($100) nor more than one thousand dollars ($1,000)for each offense. Each failure to pay an employee the wages when due him under KRS 337.020 shall constitute a separate offense. (2) Any employer who violates KRS 337.050 shall be assessed a civil penalty of not less than one hundred dollars($100)nor more than one thousand dollars($1,000). (3)Any employer who violates KRS 337.055 shall be assessed a civil penalty of not less than one hundred dollars ($100) nor more than one thousand dollars ($1,000) for each offense and shall make full payment to the employee by reason of the violation. Each failure to pay an employee the wages as required by KRS 337.055 shall constitute a separate offense. (4) Any employer who violates KRS 337.060 shall be assessed a civil penalty of not less than one hundred dollars ($100) nor more than one thousand dollars ($1,000) and shall also be liable to the affected employee for the amount withheld,plus interest at the rate of ten percent(10%)per annum. (5) Any employer who violates the provisions of KRS 337.065 shall be assessed a civil penalty of not less than one hundred dollars ($100) nor more than one thousand dollars ($1,000) for each offense and shall make full payment to the employee by reason of the violation. (6)Any person who fails to comply with KRS 337.070 shall be assessed a civil penalty of not less than one hundred dollars ($100) nor more than one thousand dollars ($1,000) for each offense and each day that the failure continues shall be deemed a separate offense. (7) Any employer who violates any provision of KRS 337.275 to 337.325,KRS 337.345, and KRS 337.385 to 337.405, or willfully hinders or delays the commissioner or his authorized representative in the performance of his duties under KRS 337.295, or fails to keep and preserve any records as required under KRS 337.320 and 337.325, or falsifies any record, or refuses to make any record or transcription thereof accessible to the commissioner or his authorized representative shall be assessed a civil penalty of not less than one hundred dollars($100)nor more than one thousand dollars($1,000). (8) Any employer who pays or agrees to pay wages at a rate less than the rate applicable under KRS 337.275 and 337.285, or any wage order issued pursuant thereto shall be assessed a civil penalty of not less than one hundred dollars($100)nor more than one thousand dollars($1,000). (9) Any employer who discharges or in any other manner discriminates against any employee because the employee has made any complaint to his employer, to the commissioner, or to his authorized representative that he has not been paid wages in accordance with KRS 337.275 and 337.285 or regulations issued thereunder, or because the employee has caused to be instituted or is about to cause to be instituted any proceeding under or related to KRS 337.385, or because the employee has testified or is about to testify in any such proceeding, shall be deemed in violation of KRS 337.275 to 337.325, KRS 337.345, and KRS 337.385 to 337.405 and shall be assessed a civil penalty of not less than one hundred dollars($100)nor more than one thousand dollars($1,000). (10)Any employer who violates KRS 337.365 shall be assessed a civil penalty of not less than one hundred dollars($100)nor more than one thousand dollars($1,000). (11)Any person who violates KRS 337.530 shall be assessed a civil penalty of not less than one hundred dollars($100)nor more than one thousand dollars($1,000). (12) Any contractor or subcontractor who violates any wage or work hours provision in any contract under KRS 337.505 to 337.550 shall be assessed a civil penalty of not less than one hundred dollars($100) nor more than one thousand dollars ($1,000) for each offense, and the contractor or subcontractor shall make full restitution to all employees to whom he is legally indebted by reason of said violation. The prime contractor shall be jointly and severally liable with a subcontractor for wages due an employee of 30 the subcontractor. For a flagrant or repeated violation the offending contractor or subcontractor shall be barred from bidding on, or working on, any and all public works contracts, either in his name or in the name of any other company, firm, or other entity in which he might be interested for a period of two (2) years from the date of the last offense. Each day of violation shall constitute a separate offense, and the violation as affects each individual worker shall constitute a separate offense. (13)Any public authority, public official,or member of a public authority who willfully fails to comply or to require compliance with KRS 337.505 to 337.550 shall be assessed a civil penalty of not less than one hundred dollars ($100) nor more than one thousand dollars ($1,000) for each offense. Each day of violation shall constitute a separate offense. If a public authority, public official or member of a public authority willfully or negligently fails to comply with KRS 337.505 to 337.550 and the failure results in damages, injury or loss to any person, the public authority, public official, or member of a public authority may be held liable in a civil action. (14) A person shall be assessed a civil penalty of not less than one hundred dollars ($100) nor more than one thousand dollars ($1,000) when that person discharges or in any other manner discriminates against an employee because the employee has: (a)Made any complaint to his employer,the commissioner,or any other person; or (b)Instituted,or caused to be instituted,any proceeding under or related to KRS 337.420 to 337.433;or (c)Testified,or is about to testify,in any such proceedings. Effective:July 13,1990 History: Amended 1990 Ky. Acts ch. 42, sec. 3, effective July 13, 1990. —Amended 1980 Ky. Acts ch. 188, sec. 262, effective July 15, 1980. — Amended 1976 Ky. Acts ch. 222, sec. 2. —Amended 1974 Ky. Acts ch. 391, sec. 13. — Amended 1970 Ky. Acts ch. 33, sec. 11. —Amended 1960 Ky. Acts ch. 56, sec. 4, effective June 16, 1960. —Amended 1944 Ky. Acts ch. 63,sec. 2. —Recodified 1942 Ky. Actsch. 208, sec.1,effective October 1,1942,from Ky.Stat.secs.576a-2,1350,1599a-19,1599a-20,2290c-4,2290c-7,4767a-16,4866b-7. Legislative Research Commission Note (10/23/90).Through an apparent clerical or typographical error, the reference to KRS 337.505 to 337.550 in the first sentence of what is now subsection(13) of this statute was transformed into "KRS 337.505 or 337.550." Compare 1970 Ky.Acts ch.33,sec. 11,with 1974 Ky. Acts ch.391,sec.13.Pursuant to KRS 7.136(1),446.270,and 446.280,the prior wording has been restored. 31 Are Prevailing Wage Laws Jim Crow Laws? Do These Lams Purposely for Inadvertently) Exclude Minorities from Public Construction WorkV (This chapter was written with Dale Belman,Professor of Economics,University of Wisconsin at Milwaukee.) The Davis-Bacon Act, which requires that federal construction contractors pay their workers `prevailing wages"was passed by Congress in 1931 with the intent offavoring white workers who belonged to white-only unions over non-unionized black workers. The act continues to have discriminatory effects today. David Bernstein,"The Davis-Bacon Act: Let's Bring Jim Crow to an End"2 1 Introduction. Until the mid-1970's debate over prevailing wage laws in construction was limited to its effect on project costs, taxpayer expenses, the benefits of collective bargaining and apprenticeship training. In 1975 Armond Thieblot introduced a new argument, that the Davis Bacon Act was, at least in part, motivated by racial bigotry. Meblot noted that the issue of race was mentioned explicitly only once during the House debate on Davis Bacon by a Southern Congressman, but Thieblot asserted that thinly veiled allusions to race could be found in other speeches including those of Congressman Bacon.22 In recent years, Thieblot's initial assertion has been refined and advanced by some Washington think tanks, notably the CATO foundation and the Institute for Justice. 3 Their basic argument is that prevailing wage statutes discriminated against African-American workers because the higher wages on public projects inclined contractors to pass over lesser skilled workers, such as African Americans. These think tanks also allege that such discrimination was not an unintended by product of the law, but reflected the purpose of the supporters of the Davis- Bacon Act. This interpretation of prevailing wage laws in general and the Davis Bacon Act in 21. David Bernstein,"The Davis-Bacon Act:Let's Bring Jim Crow to an End," Cato Briefing-Pape No. 17,Cato Institute, 1000 Massachusetts Avenue,N.W.,Washington,D.C.,January 18, 1993,Executive Summary. 22. Armond Thieblot The Davis Bacon Act,Industrial Research Unit,Report No. 10, Wharton School,University of Pennsylvania,Philadelphia, 1975,p.9. 23. Institute for Justice lawyers presented arguments on behalf of plaintiffs in seeking the constitutional overturning of the Davis Bacon Act as a racially discriminatory law. Brazier Construction Co.,Inc.,et al.,Plaintiffs,v.Robert Reich,Secretary of Labor,et al.,Defendants,Civil Action No.93-2318 WBB. 32 particular has received favorable attention from the media and in congress.24 But these claims are not based on a careful review of the legislative histories of prevailing wage laws or analysis of the effects of prevailing wage laws on minority employment. Focusing on the issue of race, this chapter provides both an evaluation of the legislative history of prevailing wage laws in construction, and new empirical work estimating the effect of state prevailing wages on minority employment. Were the Laws Leading Up to the Davis Bacon Act Racist? Competition in post Civil-War construction labor markets segmented along racial lines. Blacks outside the South tended to compete with immigrant labor for unskilled work and tended to be excluded from the skilled trades. 25 White unions in construction reinforced this pattern of racially segmented competition.26 These same construction unions were important supporters of the National Eight Hour Day Law of 1868. Given that unions supporting this law also engaged in racially exclusionary membership practices, can it be said that this law was intended primarily or substantially as a barrier against black employment on public works? The Congressional debate surrounding the National Eight Hour Day Law in 1868 was fought over class lines and not racial lines. For instance, the Abolitionist Republican Senator from Massachusetts argued in favor of the Eight Hour Law by explicitly favoring the rights of labor over capital: In this matter of manual labor I look only to the rights and interests of labor. In this country and in this age...capital needs no champion;...whatever tends to dignify manual labor or to lighten its burdens, to increase its rewards or enlarge its knowledge should receive our supp ort.27 Opponents of the Eight Hour Law felt the market !could be allowed to regulate the teams of employment and that the law violated the freedom of individuals to make contracts as they pleased. For instance, Abolitionist Maine Republican Senator Fessenden in opposing the law argued: Let men make contracts as they please; let this matter be regulated by the great regulator, demand and supply; and so long as it continues to be, those who are smart, capable, and 24. Scott Alan Hodge,"Davis Bacon:Racist Then,Racist Now,"guest editorial by Heritage Foundation analyst in The Wall Street Journal,June 25, 1990,p.A14;George Will,"It Is Time to Repeal the Davis-bacon Act," syndicated column in many papers,February 5, 1995;Tony Brown,Black Lies,White Lies:the Truth According to Tony Brown,William Morrow and Company,Inc.,New York, 1995,pp.304-310. 25. Only 14 African American masons and bricklayers were reported working in New York City in 1870. Blacks in construction tended to be construction laborers. "The longshoremen and common laborers[in New York City] are outnumbered by foreign competition;but as a general thing,their services as good honest laborers are preferred, and to a certain extent when business is brisk,get their share of employment." The Elevator,March 18, 1870,Vol. V,No.50,p.4,col. 1. 26. For instance,in 1869,the National Labor Convention of Colored Men complained:"the exclusion of colored men and apprentices from the right to labor in any department of industry or workshops in any of the States and Territories of the Unties States by what is known as"Trades Unions"is an insult to God and an injury to us." The [Washington] Evening Star,Wednesday,December 8, 1869 Vol.34 No.5224,p.4 Col.4. 27. Senate Debates,The Congressional Globe,June 24, 1868,40th Congress,2nd session,pp. 3424-3429. 33 intelligent, who make themselves skilled workmen, will receive the rewards of their labor, and those who have less capacity and less industry will not be on a level with them, but will receive an adequate reward for their labor[op.cite.] In his evaluation of this debate historian David Montgomery concluded that the National Eight Hour Law was passed primarily with the support of Radical Republicans, the same political group that pushed the passage of the 13t , 14th and 15th Amendments to the Constitution.28 There is no evidence in the Congressional debates that the first prevailing wage law in the United States was primarily or substantially aimed at limiting the labor market options of racial or ethnic minorities.29 U.S. Supreme Court Justice John Marshall - Harlan, an outspoken legal opponent of Jim Crow laws in the 1890s, upheld the constitutionality of prevailing wage laws. He said the purpose of the Kansas law was to ` . shorten the working day without decreasing the - prevailing daily wage. Had prevailing wage laws been Jim Crow laws in intent or effect, Justice Harlan would have objected to their constitutionality. Nor can the first three state prevailing wage laws, Kansas in 1891, New York in 1894 and Oklahoma in 1908, be construed as racially motivated laws. The Kansas and Oklahoma laws were similar to the National Eight Hour Law. It mandated eight hours to be the legal working day on public construction and it required that contractors pay the common daily wage. The law intended that when the working day was shortened from 12 or 10 hours to 8, the daily wage would not be correspondingly reduced. In summarizing the purpose of the Kansas Act, in Ashby v. Kansas, the court case which found the act constitutional, Justice John Harlan, of the U.S. Supreme Court wrote: When the eight hour law was passed the legislature had under consideration the general subject of the length of a day's labor, without specific reference to the purpose or occasion of their employment. The leading idea clearly was to limit the hours of toil of laborers, workmen, mechanics and other persons in like employment to eight hours, without reduction in compensation for the days service 30 28. David Montgomery,Beyond Equality.Labor and the Radical Republicans,1862-1872,Alfred A.Knopf,New York, 1967,Chapter 8. 29. In California in the mid-1860s,labor unions had two main legislative goals--the exclusion of Chinese immigrants and the eight-hour day. California Republican U.S.Senator John Conness distanced himself from Chinese exclusion but led the fight for the National Eight Hour law. Montgomery,Ibid. p.315. 30. Quoted in:Oklahoma,Department of Labor,Second Annual Report,Oklahoma City,OK, 1909,p.327. 34 If the Kansas law had racial animus as a central motivation, that motivation escaped the notice of the most eminent dissenter in Plessey vs. Fergeson, the 1890 case that established the legality of segregation based on the principle of separate but equal public accommodations and services. The primary concern of New York's prevailing wage law was the deleterious effects of cheap, itinerant, foreign and non-local labor on local labor standards. 31 Those who were U.S. citizens were said to have a prior right to jobs. Foreign labor was described as itinerant, spending little and remitting most earnings back home. To the extent race was a consideration, union supporters of New York's prevailing wage law and its citizenship corollary had only one race in mind--whites. Unions complained of cheap, itinerant "birds of passage" from England, Canada, Sweden and Denmark.32 Of course, cheap foreign labor was not always truly foreign or non-U.S. dtizens. Cheap domestic labor also threatened local labor standards. 33 But that cheap domestic labor (at least in 31. The link between cheap foreign labor and cheap domestic labor from other regions of the country was made by a writer from the New York City Bricklayers and Masons'International Union Local No.47: "The only difficulty the bricklayers have is the influx of members of their craft from other States and countries to this city which is almost impossible to overcome." -New York,Bureau of Statistics of Labor,Thirteenth Annual Report,p.387. George D.Gaillard of New York District Council of the United Brotherhood of Carpenters similarly stated: "I think there should be something done about foreigners coming here in the spring and working during the summer and then again returning to Europe in the fall. They come over here and work for less money than the native American,thus depriving him of work." -Thirteenth Annual Report, p.388. 32. For example,Edward F.O'Brien,the Secretary of the Bricklayers and Masons'International Union No.32 said: ...when business in our trade is brisk,the crowd of masons that come here to work from England is awful. They work during the summer here,live poorly,bank all they get,fill our positions and take all they earn back to England,to come again next summer. (p.387) A Brooklyn writer for the Brotherhood of Carpenters and Joiners No.258 said: We recommend restriction of immigration,for our trade suffers greatly from foreigners coming here and undermining the American citizens by working for whatever they can get. At the present time you will find that most of the carpenters out of work are citizens of the United States;while those employed are foreigners,especially Swedes...(pp.387-88). The Secretary of the Buffalo Brotherhood of Carpenters and Joiners No.355 said: We expect that you will do something for us here at Buffalo to prevent the importation of foreign labor,such as Canadians and labor from other States,to take all the employment away from us here at Buffalo. (p.388) A Brooklyn writer for the Brotherhood of Painters and Decorators No. 110 commented: Our wages have been brought down to$2.75 per day and less,by the amount of foreign labor in the market,mostly Swedes and Danes. (p.388) -Thirteenth,Annual Report 33. Mervyn Pratt of the United Tin and Sheet Iron Workers'Association of New York City emphasized that 35 the 1890s in New York state) was white, not persons of color. For instance a writer for the New York City Bricklayers and Masons International Union No. 34 noted in 1899: For some years what we term birds of passage' came over from Europe in the spring, worked here until fall, and then returned to the old country, but on account of the hard times they haven't been coming over lately. We are now affected by the flood of westerners, and there is an overplus of bricklayers in the city...34 Was the Davis Bacon Act of 1931 Primarily Motivated by Racial Animus? David Bernstein argues that racist comments found in the Congressional debate over the Davis- Bacon Act (1931) and preceding related acts (1930 and 1927) prove the Davis-Bacon Act was motivated by racial animus. His interpretation of the Congressional record divides into two parts-- a limited number of statements that directly referred to race and a larger number of statements that he believes are coded references to race. Bernstein states: The comments of various congressmen reveal the racial animus that motivated the sponsors and supported of the bill. In 1930, Representative John I Cochran of Missouri stated that he had "received numerous complaints in recent months about southern contractors employing low-paid colored mechanics getting work and bringing the employees from the South." [Alabama] Representative Clayton Allgood, supporting Davis-Bacon on the floor of the House, complained of "cheap colored labor" that "is in competition with white labor throughout the country." Other congressmen were more circumspect in their references to black labor. They railed against `cheap labor,' `cheap imported labor,' men `lured from distant places to work on this new hospital' `transient labor,' and `unattached migratory workmen.' While the congressmen were not referring exclusively to black labor, it is quite clear that despite their `thinly veiled' references, they had black labor primarily in mind.3 s In fact, direct reference to race in the debate over Davis-Bacon was rare. Of the 31 Senators and Representatives who spoke in favor of the Davis-Bacon Act in 1931, Alabama Representative Allgood is the only one to have explicitly mentioned the issue of race. Furthermore, only one of the thirteen witnesses who spoke at Senate and House hearings in that year mentioned the issue of race. Thus, the view that Congressional debate demonstrates that the Davis-Bacon Act was motivated by racial animus relies primarily on the view that proponents of the Act hid their animus with racial code words. In this view, when proponents of the Davis- Bacon Act complained of cheap, itinerant, foreign, non-local labor undercutting local labor standards,these proponents were using these adjectives as code words for African Americans. There should be some law which would prevent foreign contractors--I mean contractors from other States--from coming here and taking contracts,as it brings on many troubles. The wages in other places than New York city are certain to be lower than here and they only want the wages and work the hours of the places from which they came. -Thirteenth Annual Report,p.535. 34. 16th Annual Report of the New York Bureau of Labor Statistics for 1898,Albany 1899,p. 1042. 35. Bernstein ibid.,p.3. 36 One weakness with the code word hypothesis is that racial and ethnic discrimination was widely accepted at the time and people, including political representatives, were unlikely to use code words when speaking openly of the `problem' was so acceptable. Another weakness is that these same adjectives were explicitly applied to white Europeans in the debate over New York state prevailing wage law. A racial animus interpretation of prevailing wage laws would require that these initiatives and their code words be used primarily or only emerge when the cheap labor is a racial minority. A third issue with the code word hypothesis as applied to Davis-Bacon is that most, cheap, itinerant labor coming into high wage states in the North was not from the South and, even among itinerant southern construction labor coming north, most were white. Table 3 shows the proportion of all construction activity in each of the high wage states accounted for by contractors either from seven north-west-central states or eight southern states. As a group, contractors the north-west- central states accounted for 8% of the construction activity in these en high wage states in 1929 while contractors from the southern states accounted for 1% of the construction activity in these 10 northern states. The pattern of activity is partly determined by regional proximity. Illinois which is close to the northwestern states has the highest involvement of contractors from Western states; Pennsylvania which is closest to the South has the highest southern contractor involvement. Table 3: Importing Average Percent of All Construction Activity i Importing State Construction State Accounted for by Contractors from: Income North Wes 8 All Out of Region Central Southern NY $2,254 7% 0% 28% IL $2,113 29% 1% 66% NJ $2,036 1% 1% 8% MI $1,921 12% 1% 55% MA $1,874 1% 0% 16% CN $1,842 0% 0% 2% OH $1,786 15% 3% 61% RI $1,774 0% 0% 7% PA $1,755 1% 5% 50% IN $1,581 10% 2% 35% 10 States 8% 1% 32% Source: 1930 U.S. Census of Population,Construction Massachusetts, Connecticut and Rhode Island had little involvement from contractors of either the South or the West. Ohio and New York, states equally distant from the South and the plains states had significantly greater involvement from plains state contractors compared to southern contractors 36 36. The two regional grouping,eight southern states and seven plains states,had similarly sized construction industries at the time. The labor force in the southern states earned on average around$1,200 per year and were 37 Furthermore, even when a southern general contractor came north with a work crew, that crew was likely to be composed of both white and African American workers. Construction occupations were racially segregated in the South A crew, which would require the craftsmen from a variety of construction occupations, would necessarily include workers of both races. Thus, the general contractor would likely bring black laborers and hod carriers and perhaps brick masons. But the same contractor would probably bring white carpenters. 7 If a southern contractor came north with an integrated crew at the proportions typical of the racial composition of the southern construction labor force, then the majority of southern workers coming mrth would be white. (See Table 4). Table 4:Black Construction Workers as a Percent of All Construction Workers in Southern States,1930 State African American as Percent of All Construction Workers Alabama 25% Florida 17% Georgia 31% Louisiana 28% Mississippi 30% North Carolina 24% South Carolina 39% Virginia 15% Source: 1930 Census of Population,Occupations Proponents of the Jim Crow interpretation of the Davis-Bacon Act point to an example of an Alabama contractor who came into Representative Robert Bacon's Long Island district around 1926 and built a veterans hospital. This example was mentioned several times in the Davis- Bacon debate and in discussions of earlier related laws. As early as 1927, Representative Bacon complained that this Alabama contractor undercut local labor standards by using cheaper outside labor. The Jim Crow view assumes that this contractor brought a primarily black labor force with him. Bernstein argues this was a coded complaint against the employment of black workers in Bacons district.38 Bernstein relies on a memorandum written by U.S. Commissioner of Labor Ethelbert Stewart in 1928 that characterized the Alabama contractors crew as primarily or essentially black.39 However, in hearings for a predecessor bill, Bacon indicated that the roughly 20%African American. The plains states workers earned roughly$1,450 per year and few were African American. Workers in the northern states selected earned around$1,800 per year and emp toyed few African Americans. 37. For instance,data for Virginia in the 1920s indicate that virtually all construction contractors in that state operated with racially integrated construction crews even though in most cases occupations were racially segregated. 38. Bernstein,op.cit.,p.3. 39. Steward wrote: 38 Alabama contractor had brought an integrated crew and that the issue was not race, in any case, but rather the undercutting of local labor standards 40 In the 1931 debate over the Davis-Bacon bill itself, Representative Fiorello LaGuardia from New York City described his memory of the Alabama contractor that came to Long Island some five years earlier: A contractor from Alabama was awarded the contract for the Northport Hospital, a Veterans' Bureau hospital. I saw with my own eyes the labor fiat he imported there from the South and the conditions under which they were working. These unfortunate men were huddled in shacks living under most wretched conditions and being paid wages far below the standard. These unfortunate men were being exploited by the contractor. Local skilled and unskilled labor were not employed. The workmanship of the cheap imported labor was of course very inferior....all that this bill does, gentlemen, is to protect the Government, as well as the workers, in carrying out the policy of paying decent American wages to workers on Government contracts. [Applause.] 41 New York Republican Representative Fiorello LaGuardia strongly defended the Davis Bacon Act. He decried the exploitation of Southern `�- workers--both black and white—and claimed that the Davis Bacon Act's purpose was to ensure that decent wages were paid on Government projects. [the Alabama contractor]brought with him an entire outfit of negro laborers from the South,housed them in barracks and box cars,permitting no one to see them,that he employed no local labor whatsoever. (Quoted in Bernstein,p.4) 40. Bacon stated: ...the contractor has also brought in skilled nonunion labor from the South to do this work,some of them negroes and some of them white,but all of them are being paid very much less than the wage scale prevailing in New York State... If this contractor hired no local labor,then the skilled labor would very likely have been white southerners. In any case,Bacon explicitly stated that the issue was not whether the outside labor was black but rather whether the outside labor undercut local union wages and working conditions. When Georgia congressperson Upshaw suggested that the problem was created by the presence of black labor,Bacon responded: the same thing would be true if you should bring in a lot of Mexican laborers or if you brought in any nonunion laborers from any state. This response is consistent with the debate around the New York state prevailing wage law thirty years before that sought to reduce the employment prospects of European whites and cheaper labor from western states. Sixty-Ninth Congress,Second Session,House of Representatives,Hearings Before the Committee on Labor,H.R. 17069,Washington,GPO, 1927,pp.24. 41. U.S.,Seventy-First Congress,Third Session,Congressional Record-House,February 28, 1931,p.6510. 39 Although a small number of racial references can be found in the Davis-Bacon debate and previous, related debates, the principle issue of the debate is the protection of labor standards 42 There is no question but that between 1868 and 1931 most construction unions were racially exclusionary institutions, and these unions were supporters of prevailing wage laws. But race was not he primary or an essential concern of prevailing wage laws. The line of support for prevailing wage laws drawn from Radical Republican and Abolitionist Senator Henry Wilson in 1868 to anti-Jim Crow Justice John Harlan at the turn of the Century to Progressive Republican Fiorello LaGuardia in 1931 is inconsistent with the Jim Crow interpretation of prevailing wage laws. Racial Employment Effects of the Davis-Bacon Act. Whatever the intent of supporters of prevailing wage laws, could it be that these laws nevertheless act to exclude Afiican-American employees from the construction industry? Critics of the law suggest that African-American workers are disadvantaged both by the higher wage required by prevailing wage laws and by the lack of low wage entry occupations other than apprentice. These critics argue that higher wage rates make less skilled and less productive employees unattractive to contractors because the wage level cannot be adjusted to conform to the productivity of such employees. Contractors will prefer higher skilled workers, workers who are overwhelmingly white due to hiring and training practices, and will avoid hiring the lower skilled African-American workers. In addition the only type of employee who can be paid at less than the journeyman rates under current administrative practice is an apprentice. The lack of alternative lower wage positions, such as 'helper' or 'trainee', precludes less skilled workers from being hired onto jobs where they could develop the skills needed to qualify as a journeyman. This restriction on the ports of entry for lower skilled workers acts to exclude African-Americans in particular. Both arguments premise that African-Americans in the building trades and related fields have lesser skills than other workers in construction occupations. This might be due to discrimination in entry to apprenticeship programs, in hiring into jobs for which there is union representation,or a lack of family background in the building trades.43 Representing this criticism,Richard Vedder and David Gallaway argue: Representative Bacon was partly successful in his efforts to maintain a predominantly white labor force in construction. Despite a reduction in racially prejudicial conduct by employers over time, blacks continue to be under-represented in construction employment, more so than in other comparable occupations not subject to the strictures imposed by Davis-Bacon. While minimum wage laws such as Davis-Bacon increase unemployment for all groups and raise costs of production, the negative impact of this 42. Further evidence on the racial intent of supporters and opponents of the Davis-Bacon Act might be found voting in the voting pattern of proclaimed segregationists in the House and Senate but the vote was taken by voice and there is no record of who voted for or against the bill. Only one House member spoke against the bill and his dissent was because of the pro-union provisions of the Act. 43. As noted previously in this paper,there is historic evidence of discrimination in acceptance into apprenticeship programs. More recent work by Bilginsoy suggests that such practices have largely been ended(1998).Further, apprenticeship programs provide only half of the trained journeymen in the industry. Other important sources include training in the military,community colleges and on the job training. 40 legislation has fallen disproportionately on individuals subject to discrimination.44 Vedder and Galloway support their conclusions with two measures of the change in the labor market situation of African-American employees in construction between 1930 and 1980: 1) the shift in their unemployment rate relative to white employees, and 2) the change in African- American representation in construction relative to the change in their representation in comparable non construction occupations. First, that African American unemployment rates in construction would increase after passage of prevailing wage legislation is a straightforward extension of the thesis that prevailing wage laws reduce employment opportunity for minority construction employees. Recognizing that unemployment rates among African-American employment rose more than white rates from 1930 to 1980, Vedder and Galloway further suggest that the general trend is due to the federal minimum wage law, and is "aggravated within construction by the existence of Davis-Bacon.i45 The difference in the unemployment rate of African-American and white workers in construction should then have widened over the intervening 50 years and should have widened more than the difference in unemployment rates between African-Americans and white workers in the balance of the labor force. Minority unemployment rates in construction did rise relative to white rates from 1930 to 1980. The gap between white and African-American unemployment rates in construction increased from 1.2 percentage points to 3.9 percentage points between 1930 and 1980 (African- American unemployment construction rose from 13.5% to 16.4%, for white workers it rose from 12.2% to 12.5%). Although the authors argue that the ".. differential more than tripled", the increase in the gap in construction was smaller than the increase in the gap for the full labor force (see Table 5) . Data from another article by Vedder and Galloway (1994) indicates that the AfricanAmerican/white unemployment gap for the entire national work force increased eleven times, from -.52 percentage points to +5.52 percentage points between 1930 and 1980. Expressed less dramatically, the ratio of African-American to white unemployment in construction rose from 1.10 to 1.31 between 1930 and 1980, while it rose from .92 to 2.07 for the labor force as a whole. Once appropriately benchmarked, national unemployment data does not support the view that Davis-Bacon increased unemployment among African-American construction workers.46 44. Richard Vedder and David Gallaway,Cracked Foundation:Repealing the Davis-Bacon Act,Center for the Study of American Business,Policy Study Number 127,November, 1995,p.23. 45. Vedder and Gallaway,Cracked Foundation,Fn.39,p.28. 46. Given the vast changes in a African-American employment structures over the period in question,shifts from a predominantly agricultural labor force with concealed unemployment and underemployment,to an urban work force more readily counted and measured,the most appropriate conclusion might be that the data is inadequate to the purpose. 41 Table 5:Unemployment Rates for Construction and the General Economy,1930-1980 Male Unemployment Rates Construction Non-Construction Whites Blacks Whites Blacks 1930 12 .22% 13 .48% 6 . 59% 6 .07% 1980 12 .49% 16 .36% 10.70% The second argument considers the change in the labor force involvement of African- Americans in six construction occupations (carpenters, electricians, plumbers, painters bricklayers/stonemasons and plasterers) and compares this to the change seven other blue collar occupations (typesetters, compositors and printing-press operators; tool and die makers; cabinet makers; butchers; cotton mill operatives; and machinists) for the period 1900- 1990. African- ' American participation in an occupation is indexed by the ratio %of the African-American male labor force in Occupation %of male labor force in Occupation When this index is greater than 1, African-American male construction workers in the occupation are over-represented in the selected occupation. Conversely, they are under-represented when the index is below 1. Rather than make this comparison occupation by occupation, the two sets of occupations are aggregated into a construction and non-construction indices using employment weights. Data for the indices is taken from the decennial census. Prior to Davis-Bacon, African-Americans were more severely under-represented in the non- construction group, the construction index was 31.5% against 13% for the comparison. By 1990, this had reversed with the index for the comparison group reaching 125% against a construction index of 70.4%47 Vedder and Galloway argue that black participation in construction should have risen as fast as it did in the comparison occupations. Its failure to do so was because construction was regulated by Davis-Bacon while the comparison occupations were not so regulated. Although bench marking the construction index to other occupations provides partial control for factors other than Davis-Bacon which have affected the racial composition of occupations, the comparison is ultimately unconvincing. Sensible changes in the occupations used to construct the indices alter the results. Addition of construction laborers to the construction index, unskilled workers comparable to textile mill operatives, lifts that index from 70.4% to almost 100%. The effects of economic and social changes over the period under consideration -- changes such as the shift of the cotton and f imiture industries to the South, the shift of consumer preferences from beef (a Northern and Western industry) to chicken (a Southern industry), and the growth of construction in the Sunbelt would have to be sorted out before these indices could be used to measure the effect of Davis-Bacon on the racial composition of the construction trades. 47. Vedder and Gallaway,Cracked Foundations,Figure 2,p. 15. 42 Contemporary Effects of'Little' Davis-Bacon Laws on Minority Employment: The issue addressed in this empirical research, an issue at the core of the controversy about prevailing wage statutes, is whether such statutes reduce African-American representation in the construction labor force. Our strategy for investigating this issue is to first use descriptive statistics to illustrate the inter-relationship between statutes and the racial composition of the labor force. We then supplement this with estimates of five progressively more complete multi variate models of the racial composition of the construction labor force. These models include factors such as union membership, individual characteristics and occupation which may influence employment in construction. The descriptive statistics illuminate the central features of the relationship of interest; the mul&variate models assure that the effects of statutes have been isolated from those of correlated factors as well as provide statistical tests of the relationship between statutes and racial composition. Our research finds no relationship between prevailing wage statutes and the racial composition of the construction labor force. There is a simple negative correlation between prevailing wage laws and the probability of observing an Afiican-American in the blue collar construction labor force. Although this is consistent with the views of the critics of prevailing wage laws, it neglects the role of the racial composition of labor supply on the characteristics of the construction labor force. Many of the states which lack prevailing wage laws are in the South and have a large proportion of African-Americans in their labor force. Once we allow for differences in the labor supply between states, there is no evidence of a relationship between state prevailing wage laws and the proportion of African-Americans in construction. This pattern is apparent in the descriptive statistics and across all specifications of the multi variate models. This analysis focuses on the effect of state prevailing wage laws on the racial composition of the construction labor force. Analysis of the Federal Davis Bacon Act is difficult as there is little cross sectional or inter-temporal variation in provisions and application of the Act. In contrast, there is considerable variation between states with respect to both the presence and provisions of state prevailing wage statutes. In 1994, thirty-three states (including the District of Columbia) had prevailing wage statutes which applied to construction, eighteen did not. Among the 33 states with laws there are differences in construction projects subject to the laws and the formula used to deternune the prevailing wage. Thieblot, a proponent of the notion that prevailing wage regulations discourage black employment, uses a classification scheme of states according to whether their prevailing wage law is 'strong', 'average' or 'weak. The following analysis uses Thieblot's classification and yet directly refutes his conclusions as Data for this analysis is taken from the 1994 Outgoing Rotation File (ORG) provided by the BLS. These files include individuals who are in the last month of their CPS rotation and who are asked questions about their wages, hours of work, and union membership. We include all individuals who report being employed as a 'precision production' (craft), operative, transportation operative or laborer in the construction industry from the 1994 ORG files of the BLS. There are 5,886 observations in the data set, 5.96°/a of the employees self report as African American. 48. Thieblot,"Impact of Prevailing Rates on Black Employment in the Construction Industry,"expert report submitted on behalf of plaintiffs in Brazier Construction Company,et al.V.Robert Reich,pp.cit. Thieblot repeats this analysis in a brief note in the Journal of Labor Research,Spring, 1999.Thieblot uses a two to 17 point system in an earlier work State Prevailine Wage Laws An Assessment at the Start of 1995,Associated Builders and Contractors,Inc.,Rosslyn,VA, 1995. 43 No one argues that prevailing wage laws and the racial make up of the labor force are the only factors affecting African-American representation in construction. Other factors may influence employment in construction and may, if not controlled for, mask the true effect of prevailing wage statutes. We address this estimating a four multrvariate models, working from a simple model which only allows for the influence of prevailing wage statutes, to models which better reflect the complexity of the employment decision. The initial model includes only the three prevailing wage indicators: strong law, average law, and weak law, as explanatory variables (Model I) . Individuals are assigned values for the prevailing wage variables according to their reported state of residence. The next model adds a control for the percent of African-Americans in the non construction labor force of the state. (Model II) The third model includes two variables related to unionization, union membership and union density by state, for the construction industry (Model III). Some construction .unions have historically acted to exclude Afiican-Amencans from membership and from their trade. Although such practices have been determined to be illegal by the courts, unions may still engage in practices which de facto serve to exclude African-Americans from employment in construction. These two measures of unionization control for and measure the effect of construction unionization on African Americans employment independent of the effects of state prevailing wage laws. Individual characteristics, such as age, education, place of residence and gender may influence the suitability of individuals for employment in construction. The fourth model follows the work of Heywood and Peoples (op.cite.) in adding controls for demographic characteristics and educational attainment (Model M. The final model, model V, provides controls for three digit occupation. The argument for racial hiring consequences of prevailing wage laws suggests that such laws systematically favor more skilled, and hence more productive, workers. African American workers are on the lower end of the skill distribution, so prevailing wage laws act to exclude them from the industry. But skills in construction are, for the most part, specific to occupations. Those excluded by prevailing laws are excluded because they are on the lower end of the skill distribution for their occupation. To this point coefficients have been estimated without regard to occupation and, as such, combine `within' occupation and `between' occupation effects. This could veil the racial effects of prevailing wage laws if such effects occur entirely within occupations. The addition of controls for occupation resolves this as `between' occupation effects are accounted for by the occupational controls and the non-occupational coefficients capture only `within' occupation effects. Although most econometric research control for occupation at the level of major occupation (23 categories) or, less frequently, detailed occupational (45 classifications) controls, this research uses three digit occupational controls to better delineate the craft structure of the industry. The models are estimated using probit, but as the error term has both individual and state error components, consistent estimation is more complex than the typical probit is. The two component error structure, an implication of inclusion of state level variables in the model, results in an n.i.i.d. error which is correlated across individuals within states. If this were a linear model, OLS estimates would be consistent but inefficient 49 The implications for estimation of a maximum likelihood model are more serious, coefficient estimates are not consistent. This can be corrected with a model which allows for a random state error component. There are several 49. The standard errors obtained from the OLS routine in a typical software package would,however,be wrong,as they are calculated under the assumption of independence of error terms. The correct OLS errors can be obtained by methods typically referred to as robust or White-Huber corrections. 44 methods of estimating such a model, we utilize Butler & Moffit's (1982) approach.50 We illustrate the issue of random components by estimating Model I with a conventional probit and the random effects corrected estimator used throughout the balance of the paper. Estimates of the derivatives of the likelihood functions, the non-linear counterpart of regression coefficients, are provided in Table 6.51 Estimates of model I derived from a conventional probit are found in the first column in Table 6, estimates for a model which allows the state error term are in the second column. The derivatives of the coefficients from the conventional probit are similar across the three classifications of prevailing wage laws. The presence of a law reduces the likelihood of observing an African-American employee by approximately 3.5% without regard to the strength of the law. The coefficients are statistically significant at conventional levels, but the level of significance varies widely, from significant in a 1% one tailed test for strong laws, to 5% in a one tailed test for weak laws, to 10% for average laws. Despite such differences -- but parallel with the descriptive statistics -- the conventional probit estimates of Model I may be taken as supporting the racial exclusion theory. These results are, however, misleading both with regard to coefficient estimates and statistical significance. Correction for random state effects (column two) has little impact on the estimated effect of the coefficients on weak and average laws, African-American employment is reduced between 3% and 3.8% in the presence of such laws. The standard errors for these two variables are substantially smaller than those in the conventional probit, both coefficients are significance at better than a 1% level. The more striking change is the decline in the estimated effect of strong laws, to one third the level indicated by the conventional probit, and its consequent loss of statistical significance in any conventional test 52 The result for the strong law coefficient is at variance with the racial exclusion theory as strong laws should have a more marked exclusionary effects than average or weak laws. The random effects estimates might be interpreted as providing partial support for the racial exclusion theory, but it more clearly illustrates the need to use an appropriate estimator. Model R adds a variable for the proportion of African-American's in the state's non- construction labor force and the estimates suggest this is a critical determinant of the proportion of such workers in construction. The derivative of the coefficient on the proportion of African- American in the port-construction labor force is .5002 with a tstatistic of 13.9; a state with ten percentage points more African-American workers in its labor force will have a five percentage points higher level of African-American employment in its construction labor force. As with the descriptive statistics, inclusion of this variable in the model eliminates the relationship between prevailing wage laws and African-American employment in construction. The coefficients on the prevailing wage variables become smaller in magnitude, the point estimates of the derivatives range between -.003 and +.0094; this decline in magnitude causes the coefficients to become non- significant. This result also carries through models III - V, prevailing wage coefficients are never 50. Estimation with this procedure can be sensitive to the procedures used for estimation,such as the number of quadratures used,but estimates with this data were stable across variations on the routine. 51. The complete estimates are available from the authors. 52. The divergence in the effect of strong laws from that of other laws can be tested by comparing this model to one in which the strong,average and weak coefficients are constrained to be the equal. The hypothesis of equality between the coefficients on the three prevailing wage variables can be rejected in a 1%Wald test. 45 significant in models which include the proportion of African Americans. Addition of controls for union membership and union density, model III, do not alter any the estimates. The coefficient on the proportion of AfricanAmencans in the state labor force remains large and statistically significant, those on the prevailing wage law variables continue to have small, non-significant coefficients and the coefficients on union membership and union density by state are also small and nor-significant. This outcome, which is maintained in all further estimates, is unexpected given the historic and legal record of some building trades unions with regard to employment of African-American workers. It may reflect the success of legal and institutional efforts to end discriminatory practices. Whatever the source, this research suggests that construction employees who are union members are no less likely to be African-American than those who are not African American. Further, that the increased bargaining power provided by greater union organization of construction labor markets is not being used to exclude African Americans from employment in constriction. Model IV, which controls for factors such as age, education and residence, which might t alter the 'duals for employment in construction,influence the suitability of individuals n, does no relationship between prevailing wage laws and minority anployment. The effect of the proportion of African Americans in the state labor force remains large, the effects of prevailing wage laws and of union membership and density remain small in magnitude and non significant. Other important determinants of African-American employment are age and its square, metropolitan residence, marital status and holding a college degree. Older employees are more likely to be African- American, although the relationship is convex. Considering the effect of age alone, a twenty year old has an 8% probability of being African-American, a thirty year old has a 10.61/o probability, a forty the probability is 12.3%, at fifty it is 13.1%. The probability begins to decline between fifty and sixty and at sixty it is 12.9%. One possible source of this pattern are the recent shifts in minority employment in construction, with Hispanics increasingly competing with African- American workers over the last twenty years. Older African-American construction workers, who have ties to the industry, would have remained employed at relatively high rates. But fewer young African-Americans would find employment in construction as Hispanics have moved into the industry (Belman and Bilginsoi, 1997). In addition to age, residence in a metropolitan area increases the probability of an employee being African American by 1.7%. Being married and holding a college degree both decrease the likelihood of observing an AfricanAmerican, by 2.3% ° respectively. Educational attainment other than a college degree has little effect on the and 4.2/o respect y g racial composition of the construction labor force, a result in keeping with the importance of occupation specific rather than general skill training in the industry ( Belman and Bilginsoi, op.cite.). Model V, the final model in this series, differs from prior estimates in controlling for a fixed effect b three digit occupation. Again, b removing the effects of inter-occupational factors Y g p g Y including skill related factors, this model should eliminate any masking of the effects of prevailing wage laws by occupational factors. The thrust of the prior results remains. The cardinal explanatory variable is the proportion of AfricanAmericans in the state labor force, the effects of prevailing wage laws and unionization are small in magnitude and nonsignificant. Model V suggests varied patterns of racial employment b trade. There are thirty-four distinct trades in this gg Y data set including three grades of mechanic, carpet layers, iron workers, electricians, apprentices, and bricklayers. There is evidence that African-Americans are significantly less likely to be observed in occupations such as construction supervisor, heating-ventiliation-air-conditioning, carpenter, electrician, painter, plumber, ironworker, sheetmetal, welder, operating engineer or material moving operative. Although no simple pattern is apparent in this set of occupations, it 46 appears that African-Americans are less likely to be employed in licensed occupations (such as plumbing and electrical) and occupations which require formal training (such as operating engineer, electrician and plumber). But, carpenters and welders, occupations which are often self taught or learned on the job, are also less likely to be African-Americans. The estimates also indicate that apprentices are no less likely to be African-American than other construction workers. This cuts against the argument that such positions do not provide ports of entry to construction for African- Americans and is consistent with Bilginsoy's research on apprenticeships. The small size of the sample argues against putting too much credence in this result.53 Conclusion: A prominent criticism of prevailing wage laws has been that they reduced the employment of African-Americans in the construction industry. This premise has been supported by evidence from legislative records and theoretic arguments about administered wages role as a bar to the employment of the lesser skilled African-American worker. The argument was further buttressed with evidence on discrimination against African-American employees by building trades unions. This chapter has addressed both of these issues, providing an overview of an extensive review of the historic record of prevailing wage laws and a statistical analysis of the current relationship between 'little Davis-Bacon' acts and minority employment. We find that, although those involved in passage of prevailing wage laws did have exclusionary intent, the intent was towards low wage, transient workers including, at various times, white Northern Europeans and migrants from the Northern Great Plains. The argument for anti African-American bias in the legislative history of Davis-Bacon itself is based on overplaying one Senator's comments, misreading of the record and misinterpretation of the historic circumstances at the time of the passage. Our empirical research moves away from discussion of intent to one of measurable consequences. Utilizing a conventional data source and a procedure incorporating a state and individual error component, we find a moderate negative simple correlation between state prevailing wage laws and minority employment in blue collar construction. This correlation is, however, the product of the lack of such laws in the South, the region with the largest proportion of African Americans in its labor force. Once adjusted, the association between prevailing wage laws and minority employment disappears. The debate surrounding the Davis Bacon Act will continue on other grounds. How the Act effects the cost of public construction, the quality of work done, the amount of training that takes place in construction, the extent to which the law promotes labor standards and encourages collective bargaining, all these issues remain and will be addressed in subsequent chapters. However, the proposition that the Davis Bacon Act was primarily or substantially intended to restrict African American access to federal construction work is not supported by the historical record, and the idea that the Davis Bacon Act currently restricts minority access to construction work is not consistent with current racial patterns of employment. 53. The helper classification is of interest as opponents of prevailing wage legislation suggest that the helper category is utilized by African-Americans as a point of entry to the construction labor force. The relationship between employment as a helper and racial status could not be tested as there were few helpers in the data set and since none were African-American it could not be included in the model. 47 . . . � f% : 4 $ §§ / kU) 0 cm a) Co CD CL 70 \ & £ a)_ ® §_ / © ƒ -00.2 . 2Q & \ %0 / > 7� 0 ƒ 22co 0ƒ o� c� r-2 w� a) % } m 2 . R\ k/ \� \� \� \\ � ) /\ k EI ; i � — ee � t 2f k k\ 11 0 ƒ / � ° . ! 0 22 ° 27 \ , U / Co 0 — �7 M7 a « e @ �CO : . 5IP R� CO § Gm ± C ks 2 ° c1 ] ; R§ o § oo ¥2 09 C § x k U) ; o ® . c - o . — 0M\ \ «kz . k\ [ cc k3� _ � $ \ ¥ 4 0 §f = # _� F - R 2 to � -0 \ , CN % Cq � o §« $= oo 4) @ c = o o � o o � h ow of o ® 2m $ » � 23 f3 §3 RZ, C! % � / ) ` . 2 � = oe > , m e J}\ C 2 0a, \ CD° # c . ¥ E � 7 i 2 c i , c o ¥ ? q _ . \ — 0 2 .t - y 2 f / e � 0 \ f\ § G0 G S \ \m ° / \ 0 \ § 2 § 03 f \ - R 00 EI e3 Q = & U) a) \ ] % EJ ` - y a) ƒ § ) 2en \ <a� 2 iF \ \ m \7 0) \ c0 " / \� / � < � C �o W\ �+ a)a2 . oa)oT ' ! 2 2 2 | & C: � \\ \ \ \ \0 C � . c 5 0 \ �\ 0 $ f £\ � e . i% 0 \k \ �\ g |\ e ° { \ \ � ® 00 | @ / gm § c2 » ; ; u 2 c ' o _ &= e c a ® 0 i2 a / » 2t^ c . § e \ % § ' � E 0 � I � e� f / / \ \ CL J \ \ 13 § 0 An Analysis of the Impact of Prevailing y P 9 Wage Laws on Construction Costs u th a special Focus on School Construction Costs Introduction. Few people would object to a law that had as its purpose the promotion of decent wages and benefits for the citizens of a state. Fewer still would object to a law that promoted training and the creation of skills that would justify the payment of decent wages. Not many would object to a law that promoted the availability of health insurance for working class families. Few would be the opponents to widespread private pension programs that helped provide old age security for construction workers and their spouses. The other chapters of this report show that the intent and effect of prevailing wage laws are, in fact, ones that few people would object to. The purpose of these laws was and is to promote productive skills, decent wages, solid health insurance, and widespread pensions among construction workers both on public and private construction. Construction workers make up around 5% of the total labor market. Historically, construction has been a place where working class families could make middle class incomes. The carpenter next door to you could afford his or her house. Economic security led to social security. Construction workers made good neighbors. With the aid of collective bargaining and prevailing wage regulations, the construction labor market has been a major American success story and a minor economic miracle. Left unorganized and unregulated, the construction labor market has all the makings of a secondary labor market of low skilled, itinerant, foot-loose workers. With perhaps the exception of the harvest labor market in agriculture, construction is the most unstable area of work; it is seasonal; and it suffers from wide swings of booms and busts. Workers move from contractor to contractor. Outside the framework of collective bargaining and prevailing wage regulations, contractors have few incentives to train most of their workers. Worker retention to the industry is also problematic. These are the classic signs of what economists call a casual, low-wage, secondary labor market. The American success and the minor miracle is the fact that through collective bargaining and prevailing wage regulations, many construction workers have been able to build highly skilled, craft-based careers out of the flux and uncertainty that characterizes almost all 49 I aspects of construction. This benefits these workers directly and it benefits the communities within which they live. If construction develops along a high-skill, high- wage growth path, then the 5% of the workforce in construction and their families become community assets. If construction develops along a low-skill, low wage path, then the construction workers become a community liability. Given the choice between a � high skill growth path that also insures quality workmanship and safe construction—and h and safe at risk w a low skill growth path that puts workmanship in to question safety k, Y � would anyone oppose prevailing wage laws? Some will oppose prevailing wage laws simply because they have a vested interest in the low-wage development path. These contractors seek a comparative advantage in using low-wage, low-skilled labor. They hope the race will become a race to the bottomcompetition, Y and in such a com etitio the think they have an advantage. Prevailing major comparative advantage present serious problems for contractors whose � p g wage laws pres is that they do not pay health insurance, they do not provide pensions, and they do not train apprentices. But few argue to change prevailing wage regulations simply to expand the opportunities of low-wage contractors. The major rationale presented to justify a repeal of prevailing wage regulations is raise government construction costs. This the allegation that prevailing wage laws g g argument is appealing for two reasons. First, if government could save significant sums of money on its construction costs without sacrificing the quality of construction, then several worthwhile constituencies might well gain from these savings. For example, perhaps we could build more schools for the same amount of money; or cut taxes; or both, if the savings were truly substantial. Second, it seems intuitively plausible that if are wages saved. This is where these critics of prevailing wage laws cut,>; money wi ll be often go wrong. Sure, if you cut wage rates, and nothing else changes either now or in the future, then you will save money on construction costs. But in what walk of life can you cut wage rates by 30% to 50% and eliminate benefits and have no effect on worker morale,labor productivity or business strategies? You can build a dam wi th buckets and shovels. Your workers will need few skills and your wage rates will be low. Or you could build a dam with heavy earth moving equipment and you will need to hire high-skilled, high-wage equipment operators. But just because the bucket-and-shovel approach allows for low wage rates does not mean that your dam will be either cheaper or quicker to build The belief that cutting wage 1 rates cuts labor costs or construction costs needs to be careful y examined. Building Four Schools for the Cost of Three. Critics of prevailing wage regulations contend that public agencies can save substantially t of construction if prevailing wage regulation were eliminated. A not on the cos p g g d be cut b 25% or more. For construction costs cool uncommon proposition is that Y instance, Gary Johnson, Governor of New Mexico asserted in his state-of-the-state address in 1996 that 50 "...without the constraint of the Little Davis-Bacon Act, we could build four schools instead of three for the same amount of money." State of the State Address,January 16, 1996 • If labor costs are almost half of total costs,(50%) • And labor costs fall by about half(50)% • Then,by eliminating prevailing wage regulations--overall costs must fall by 25%. This calculation requires one key assumption. It explicitly assumes that when wages and benefits fall by 50%, labor productivity remains the same. If, at a lower wage rate, contractors hire less experienced or less skilled workers, then those lower wages do not necessarily translate into cost savings for the state. Either the contractor will have to hire more workers to offset their lower productivity, or work less productive workers longer, or tolerate lower quality results. Any of these factors could partially or completely wipe out hypothetical savings from lower wage rates. This type of analysis is hypothetical. It is not examining the cost of public construction under prevailing wage regulations with the cost of public construction absent those regulations. Rather, it is developing a hypothetical scenario of what might happen. Like all hypothetical scenarios, this analysis is only as reliable as the assumptions made. As it turns out, standard sources on costs in the construction industry do not support the notion that labor costs are more than 30% of total costs. Indeed, in Kentucky labor costs nun around 26% of total construction costs. A drop in wages of 50% with no change in productivity or the type of equipment used or the amount of training done, would yield a 15% savings assuming wages dropped in half. If wages fell by 25%, the savings would fall to 7.5%. And if productivity fell off, if training fell off, if experienced workers went elsewhere, that hypothetical 7.5% could fall substantially. Indeed, if cost over-runs increased, if the cost of maintaining poorly constructed buildings and roads increased,the hypothetical savings could even fall into the red. Labor Costs According to the United States Census of Construction. There is a standard source on labor costs in the construction industry. The U.S. Census of Construction surveys construction contractors in every state every five years. The results of the most recent survey, taken in 1997, have yet to be released. However, we have data on labor costs as a percent of total costs in construction for the United States as a whole and for Kentucky in 1992. These data are for thousands of contractors and they are not gathered for the purpose of any specific study. The Census of Construction is systematically relied upon by researchers and analysts of the construction industry. We will see that labor costs as a percent of total costs are much lower than is required to calculate substantial savings. For all construction in Kentucky, labor costs-- including wages, benefits and payroll taxes—run around 26 % of total construction costs. Figure 1 shows labor costs as a percent of total costs broken down into wages and benefits for Kentucky and the U.S. in 1992. Benefit costs in this graph are somewhat overstated because they include not only benefits going to construction workers, but also 51 benefits paid to non-construction workers employed by construction contractors—office workers,estimators,engineers and architects. All Construction 30% 6 5 20% 10% Benefits as a%of Total Costs EfflWage Cost as a%of E 0°l0 Total Cost � US KY Source: 1992 U.S. Census of Construction Benefits overstated by including office&other non-construction workers Figure 1:Labor Costs as a Percent of Total Costs for All Construction in US Kentucky,1992 The Census of Construction for Kentucky also breaks labor costs and total costs down by contractor type. In this breakdown only wage costs are included. Figure 2 shows, for Kentucky, wage costs as a percent of total costs for specific types of general and heavy-highway contractors.54 For light commercial contractors (office buildings, schools, churches, etc.), wage costs account for 20% of the total cost of construction. Estimated benefits—including payroll taxes--probably account for an additional 4% of total costs. Thus, labor costs account for about 24% of the net value of the work done by light commercial contractors including school builders. I say net value because the Census of Construction quite rightly excludes from commercial contractors the value of work done by subcontractors. To calculate overall labor costs on a commercial job or a school project, one must consider not only the costs of the general contractor but also the subcontractors. We will do this momentarily. However, first, let us look at heavy and highway contractors. 54 School construction contractors are aggregated with office builders,church and other non-resdenlialgeneal building contractors. In the industry,this is usually referred to as light commercial general contractors. A school project will also include some specialty subcontractors. 52 General & Highway Contractors Single-family housg 1 Other residential 0 M Indusl bldgiwarehse V L Offices,Schools etc in � Highway&street con c Bridge,tunnel const 0 Water,sewer,utility IZ Heavy constrn,nec 0 V 0% Mean Wage Cost as a %o of Total Cost Source: 1992 U.S. Census of Construction Benefits data not available for industry breakdown Figure 2:Labor Costs as a Percent of Total Costs for Kentucky General Contractors,and Heavy and Highway Contractors For highway and street contractors, the use of non-heavy and highway subcontractors is limited. Thus, the 18% wage-cost-as-a-percent-of-total-cost reported by highway and street contractors is an accurate reflection of wage costs in this type of construction. Labor costs are typically lower in heavy and highway construction because the use of heavy equipment increases labor productivity substantially. The use of labor augmenting equipment that raises labor productivity permits the payment of higher wage rates, while at the same time cutting labor costs as a percent of total costs. You could build Hoover dam with buckets and shovels; wage rates would be low but labor costs as a percent of total costs would be high. And,the dam would probably cost more to build. Labor costs as a percent of total costs are typically higher for specialty subcontractors compared to general contractors and heavy/highway contractors. This is because typically the general contractors bear a larger share of material costs and heavy/highway contractors have heavier equipment to augment the productivity of their workers. Figure 3 shows labor costs as a percent of total costs for specialty contractors in Kentucky. 53 General & Highway Contractors Painting,paper hang Electrical work L Masonry,stonework 0) Plastering,insulatn 0 Terrazzo,tile,marble 0 U Carpentry work Floor laying Roofing,sheet metal '0 Concrete work c C Structural steel ere Glass and glazing n Excavation work L in Installing equip,nec 0 Special trade,nec Mops: 10% 20% 36% 40% Mean Wage Cost as a % of Total Cost Source: 1992 U.S. Census of Construction Benefits data not available for industry breakdown Figure 3:Labor Costs as a Percent of Total Costs for Kentucky Specialty Contractors One of the first things to notice in these figures is that labor costs as a percent of total costs do not fluctuate directly with wage rates. For instance, in Figure 3, labor costs for high paid electricians account for 27% of total costs. In contrast, lower paid painters' wages 29%o of the total value of work done by painting contractors. The highest wage costs as a percent of total cost in Kentucky in 1992 were masonry and stone contractors. Typically brick masons earn less than electricians. These differences are partly due to the differing cost of materials paid for by a masonry contractor, a painting contractor and an electrical contractor. But also the higher labor productivity of a well trained electrician can offset his or her higher wage rates. High wage rates, if they induce higher labor productivity, can actually reduce labor costs as a percent of total costs. Low wage rates, if they mean a loss of skills, can in some cases result in higher labor costs as a percent of total costs. 54 Second, while the Census of Construction does not break out school construction contractors as a separate category, a U.S. Department of Labor study has done this. In 1979, the U.S. Bureau of Labor Statistics published a study of school construction costs by region in the United States. The BLS study aggregated school types and presented data on four regions, Northeast, Midwest, South and West. The relevant data for our purposes is presented below. Table 7:Hourly Wage Rates and Total Costs as a%of Total Construction Costs Elementary • Secondary School • • 1972 Wacies as a Percent of Total Cost Northeast 11775 27.9% North Central $7. 3 29.3% South $ ': 27.3% West a _. 29.0% Source: U.S. Bureau of Labor Statistics, John G. Olsen, "Labor and Material Requirements for New School Construction,"Monthly Labor Review,April 1979,Vol. 102,Number 4,p.41. These are old statistics but their age make them more instructive. In 1972, prevailing wage laws were widely enforced in the North (including Kentucky). If prevailing wage laws bloat relative labor costs now, they should have bloated those costs then. But, in fact wage costs as a percent of total costs were 27.9% in the Northeast. (This does not include benefits.) Adding another 6 % for benefit and payroll taxes would bring labor costs up to about 34% of total costs in 1972. If prevailing wage laws did not bloat labor costs then,there is little reason to believe they are doing so now. Wage Rates and Labor Costs. An interesting point to be derived from Table 7 is that hourly wage rates varied considerably between the Northeast region and the South ($7.75 versus $5.22 in 1972). In contrast, wage costs as a percent of total costs were almost the same in the two regions (27.9%versus 27.3%). The analyst,John Olsen,commented on these facts as follows: Average hourly earnings also varied by region. Hourly earnings for all construction workers averaged $6.78, ranging from $5.22 in the South to $7.75 in the Northeast. Wages as a percent of contract costs varied from just above 27 percent in the South to slightly above 29 percent in the North Central. Although average hourly wage rates in the Northeast were higher than those in the North Central region, wage costs as a percent of total contract costs were lower. Among other factors, this irregular trend could result from regional differences in productivity rates and in relative material costs. (pp.40-41) Could it be that as wage rates are cut experienced workers leave for better paying jobs elsewhere? Could it be that as wage rates rise, contractors find it worth their while 55 to spend the money needed to better train their workers and provide them with new, better equipment? A Direct Look at School Construction Costs Table 8: Annual Median Real Square Foot Construction Costs for Kentucky Public Schools Kentucky Year Median Projects 1991 $80 7 1992 $79 18 1993 $78 14 1994 $85 11 1995 $89 17 1996 $93 22 1997 $91 17 1998 $88 17 1999 $91 10 Restrictions: School Size between 20,000 & 120,000 Sq Foot Value$20 to$180 Deflated Using ENR-B Source: FW Dodge Start Cost Tables 8 shows the median square foot construction costs for school building projects in Kentucky over the period July, 1991 to July, 1999. The accepted bid price of the schools were inflated to 1999 dollars using Engineering News Record's index of building construction costs. This allows for a direct comparison of square foot construction costs for school built in different years.55 The data in Table 8 are graphically presented in Figure 4. 55 The data are from the F.W.Dodge Corporation,the standard service provider of project information in the construction industry. Mernative price indices were tied to examine whether results were dependent on the price index chosen. Results were basically the same regardless of the price index used to translate information into constant 1999 dollars. 56 Annual Median Square Foot Cast of Kentucky Public School Construction Proms(inn 1M Dollars) sus Regtiar�9a StfK3cl� S8i1 5 1991 1992 1 MI, 19N 1996 1996 1997 1%8 1999 Figure 4:Annual Median Square Foot Public School Construction Cost in Kentucky,1991 to 1999 As one can see, real public school construction costs in Kentucky rose substantially in the 1990s. But these costs began rising prior to the implementation of Kentucky's prevailing wage law to public school construction. Subsequent to the implementation of the law, public school construction costs have not been rising (once you adjust for inflation). The peak year in real terms is 1996 and the vast majority of projects built in 1996 came before prevailing wage regulations. By itself, this is strong evidence that the rise in school construction costs in Kentucky cannot be primarily or substantially attributed to prevailing wage regulations. A primary contributor to the increase in school construction costs has been the extended economic expansion of the 1990s. This has led not only to the tightening of labor markets but also to a sellers market for construction contractors. This has occurred not only in states with prevailing wage regulations but also in states without prevailing wage laws. One way to see this is through the use of an econometric model of the determinants of school construction costs. The beauty of econometric models is that they allow one to statistically isolate the independent effects of different factors or variables that influence the cost of building a school. Table 9 presents five related econometric models of school construction costs. These econometric models were tested against data on school construction costs for all states over the years 1991 to 1999 which come from FW Dodge information on the accepted bid or start costs of several thousand school construction projects. These school building projects can be identified by a) state, b) he year and month in which construction started, c) the square foot size of the project, d) the type of project (that is, new construction, addition, alteration or additions and alterations), e) the number of stories involved, f) the type of school (elementary, middle or high school), and 57 g) whether or not the owner was public or private. And, of course, in each case we know what the accepted bid or start cost was. To complete the data set, the Engineering News Record building construction cost index was added to obtain an inflation-adjusted measure of start costs. Also, for each year, the state unemployment rate for the entire labor market (not just construction) was added as a measure of the stage of the local business cycle. To understand how these models work, let us focus on Model I. Eventually, we will focus on the possible effects of prevailing wage regulations of school construction costs. But we will do this in the context of controlling for the other factors that influence school construction costs. To understand these controls, we start with model I. Model I looks only at school construction in states and time periods where no prevailing wage laws are on the books or in force. The "dependent variable" or the phenomenon Model I seeks to explain is total start cost of a school building project (denominated in 1999 dollars using the ERN building cost index). To meet the technical requirement of a normally distributed dependent variable, the natural log of the start cost of each project is the actual dependent variable.56 This transformation will also help us in interpreting some of the "independent variables" or explanitory factors in the model.57 The first explanitory variable in the model is not a variable at all, really. It is called the "constant" and is sort of the starting point in estimating the effects of the true explanitory variables in the model. So for the moment, we can ignore the constant. A major factor explaining differences in the cost of different school building projects is the size of the project. All other things being equal, typically larger projects cost more than smaller projects. (This is not always true if a smaller project has a lot of equipmetn installation--the putting a new boiler into a school.) But it might also be that one project that is twice the size of another might cost more, but not twice as much as the smaller project. In other words, as the size of a project goes up, the total cost might go up at a slowing rate. Thus, two variables are put into the model to capture the possiblity that there might be economies of scale in building schools. The first variable is the square foot size of the project. The second variable is the square foot size of the project times itself (or square foot squared). The econometric models presented in Table 9 are called "ordinary least squared linear regression models." This is the most commonly used type of econometric model. But it is a linear model. The phenomenon of economies of scale is a nonlinear phenomenon. Using a variable and its square to capture nonlinear relationships between a cause and effect is a conventionaly device in this type of model. "If economies of scale in school construction exist we expect the following. The square foot size of the project will be positively related to the total cost of the project. The square of the square foot size of the project will be negatively 56 This is a requirement of ordinary least squares linear regression models,the type of model used in these tests. Total school construction costs are not normally distributed. Rather there is a longer tail in the distribution towards the high end of costs. Taking the log of total cost normalizes this distribution by balancing the two tails of the distribution. 57 Also construction projects were limited to those that were between 20,000 square feet and 120,000 square feet with a square foot cost that fell between$20 and$180 in 1999 dollars. Also the total value of the projects all exceeded$750,000. 58 related to the total cost of the project. When projects are small, the size of a project and the square of the size of the project will be numbers that are relatively close together. As the project gets larger, the size and its square will get progressively farther apart. If the size is positively related to total costs and the square of the size is negatively related to total costs, and if the magnitude of the relationships line up, then total costs will be pushed up by size and pulled down by size squared. The larger the project gets, the stronger will be the downward force of size squared. Thus, the model has the potential of capturing economies of scale in school construction if those economies in fact exist. And indeed, there does appear to be economies of scale in the construction of schools in states without prevailing wage laws. To see this, look at the independent variables in rows 2 and 3 under column a in Table 9. There you will find square feet and square feet squared In column b for rows 2 and 3 you will find the estimated coefficents for these two variables. Estimated coefficients are the model's estimate of the effect of these variables on school construction costs. The square feet of a project is estimated to positively effect total costs and the square feet squared negatively effects total costs. Furthermore, the magnitude of the effect of square feet is larger. The coefficient is larger. w' 59 c (6 O elfr M 0 o N u O M 4 O 6 r e- of N N � N N o'i � �;`•�_ �O C N C N O Cl ca O O ['E' a NO Oz o M M r r 0 C r M 0 to w ch r N O1 O N t0 O O co v co cno Cl) — O O O r r d' In f� '� � N � It N O IA OM M `� CO � N � N — O O r V. ca U) M O O O O 00 Q O O O O o O e o o e o o e e o 0 0 o N 0 O r O o 0 0\0\' o 0 o e \ \ \ \ \ o e o o o co t 1 M O V T.- t- K m 0 r N N �p r� v �- ul r t0 M O h co co O LO O O N c0 O N et N N O O w st N O r h Ln O 4.4 O 0) — O O CJ C13 O Cl OO ca O J U M O O o \ o 0 0\0\0 \ o o\o o\° o e o e \° o\o \ o\o o co (� � '0 O 00 tD O N 00 r O co M 00 O M p0 e- �- t0 d' co (0 — Q CD N 1(� N In •- �- !� f- 00 m w c'i .- O 00 4� 3 M °o � � rA (aO o p ^C O o c Z c o O d 3 -� }y V = J b ca .a d ♦+( U -O.N � a) (v c > V 7 O O Q' •j O M N > _ d CT L a1 m c d O ( 0 �U+ Q � O O LC Lo > Wta) (D >, o6 .a .o _ C LL U- o L o i0 -0 NO L. R y in o Uj ° o r> v �n co ao m a W Z m . o y Z c 7 7 0 0 -0 L '6 'O Z rn rn rn rn rn m rn m Z a > N d '►- ° o Q 3 L � o' v v c a S o 'o � CZ O �? cn UO)- _ ¢ ¢ Q � 000 � [� o � � 0 cna0a ¢ O �I r NMd' � t01� 00.01 O rN Md' tOWr- 000Or NMI O (D r 0N r r r r r T.- r r r T.- N N N N N N N N N �T Projected Total Project Casts and Cost per Square Foot by Size of School Project-Florida High Schaal W,000A00 $140,00 $10,000,000 MOM $100.00 W000,000 $00.00 *6,000,0007 r $4,000,000 $40.00 $�,000000 $20.00 $0.00 Figure 5: Model I Projection of Economies of Scale for a Public High School Built in Florida in 1999 Figure 5 gives a graphical presentation of the results of Model I in Table 9 regarding economies of scale. The Figure presents the model's projection for total cost and cost per square foot of a new public high school built in Florida in 1999. The costs vary by the size of the project. At 20,000 the total cost would be small, just over $2 million, but the cost per square foot is high—over $120 per square foot. Read the total cost off the left-hand vertical axis and the square foot cost off the right-hand vertical axis. The line with circles represents the model's projection of total costs and the line-with- squares represents the model's projection of square foot cost. 61 Econometric models present estimates of the effect of all the variables included in the model. Sometimes those effects are small but nonetheless, statistically significant. The coefficient estimated for square foot squared is a small number but statistical tests indicated that this number is still statistically significantly greater than zero. A zero would indicate no effect or no relationship between the independent variable and the dependent variable-in this case school construction costs. Econometric models typically do not generate zero coefficients. Rather, they generate a coefficient but statistical tests indicate that the coefficient is not statistically significantly different from zero. In Model I the square foot and square foot squared variable coefficients are both statistically significant. This is indicated by the bolding of these two numbers in rows 2 and 3 under the column for Model I(column b). In rows 4 and 5 under column a are two additional variables controlling for the size and shape of the project. These are marker variables, (often called "dummy" variables) that turn on if the school project is a two-story project (in row 4) or a three-or- more story project (in row 5). It turns out, that controlling for the square foot size of the project, going up as opposed to building out adds to the total cost of the school. Total cost in these models are construction costs, not land acquisition costs or architectuaral a costs. In terms of buildingcosts a two-story school project adds 4.7/o to the cost of IY p J construction and a multi-storied project adds 13.8% to total costs. Because many mullti story schools are built in heavily urban areas, some of this effect may be due to the costs of urban construction rather than multistory construction per se. Also, because multi- story construction is sometimes used in colder climates, some of this multistory effect may be due to the costs of cold-weather construction design requirements. In states that do not have prevailing wage laws (that is, in Model D a middle school that is the same size and style as an elementary school will nonetheless cost 2.6% more. A high school of the same size will cost 5% more than an elementary school. These results can be seen in rows 6 and 7 under column b in Table 9. Both results are statistically significant. Relative to new school construction of the same size, involving the same number of stories and at the same type of school, additions are estimated to cost 2.2% less. However, this result is not statistically significant. That is, while the computer's best guess is that additions are cheaper than new construction of the same size, the computer is uncertain about this. Statistical tests suggest that there is no statistically significant difference between this —2.2% estimate and zero. So there may be no difference between constructing an addition than there is in new construction of the same size. Once the project includes additions and alterations, the model estimates that costs go up by 11.8% over new construction of the same size. This estimate is statistically significant. Alterations by themselves raise costs over similar-sized, new construction by 5.1%. This,too, is statistically significant. Now we come to the issue of time. The reference point is 1991-1992. The question is, controlling for other factors like the type and size of the school and the nature 62 of the project, and realizing that we have already translated all the cost data into 1999 dollars—have the real costs of school construction increased since 1991-92? We test this question with a series of marker or dummy variables. The variable D 1993 tums on if the school project was started in 1993 and turns off otherwise. The variables D 1994 through D1999 similarly turn on and off when their year is indicated by the start date of a project. The coefficients for these marker variables in column b indicate whether or not in that year, controlling for the nature and size of the project, the total cost of the project has gone up over the reference period 1991-92. In 1993, the model estimates that real costs went down by 1% and in 1994, the model estimates that real costs went up 1.6% over 1991-92. But in both these cases the estimates are not statistically significant That means that for all practical purposes, the real cost of school construction in states without prevailing wage laws did not change in either 1993 or 1994 over what they were in 1991- 92. However, by 1995, statistically significant real increases in school construction costs can be observed. From 1995 on, real costs show an upward trend over 1991-92, from a 7.3% increase n 1995 to a 16.8% increase in 1999. These real increases in the total cost of building a similar project are obviously not due to prevailing wage regulations because Model I restricts observations to 2,686 school construction projects that took place without any prevailing wage regulations. This is consistent with Kentucky's experience of seeing real median square foot costs of public school construction rising in the period prior to the implementation of prevailing wage regulations on public school construction. We not only know what year these various school construction projects were started, we know in what month each started. In cold-climate areas, starting a project in the teeth of cold weather can conceivably effect total construction costs. We test this with the variables winter, spring and summer in rows 19, 20 and 21. Again these are marker or dummy variables that turn on if a school is started in the winter (January through March), spring (April through June) or summer (July through September) compared to fall (October through December). The start date in the data is the date a bid has been accepted. Thus, actual groundbreaking will lag behind the start date somewhat. Keeping this in mind, a bid-acceptance date in "winter" probably means a groundbreaking date in winter-spring, and so on. So the reference period, "fall" refers to a groundbreaking period of probably November-December-January. Model I estimates that a bid-acceptance date in winter or spring lowers costs compared to a bid acceptance date in the summer or fall. Starting coming out of winter is 1% to 2% cheaper than starting going into winter. But these estimates are not statistically significantly different from zero. This should not be too surprising. Half the states that do not have prevailing wage laws are sunbelt states—eight southern states and Arizona. The snowbelt states without prevailing wage laws tend to be smaller—the Dakotas, New Hampshire, Vermont, Idaho and Utah. The largest snowbelt states without prevailing wage laws are Iowa and Colorado. In the time period under study, 1991 to 1999, Oklahoma, Michigan, Ohio and Kentucky flipped around between regulation and deregulation. So the effect of building into the colder months will be muted by the dominance of sunbelt states in this group. Thus, while a small savings is found for school groundbreaking that avoids the winter season,this effect is not statistically significant 10 63 Row 22 presents an important variable—the cost of building a public school compared to a private school. In states without prevailing wage laws, for school construction projects of similar size and type, public schools cost 8.4% more. This has nothing to do with prevailing wage regulations. Again, we are only considering construction projects done in jurisdictions that do not have prevailing wage regulations. What this probably has to do with—is differing architectual design and materials. While there are certainly some upscale and expensive private schools, on balance, public schools are designed more expensively. This result is statistically significant and practically significant. We will find that public schools in states with prevailing wage laws cost more than private schools in those states. But this cannot be easily or primarily attributed to prevailing wage regulations because this phenomenon is found in the absence of those regulations. In addition to all these variables reported in Table 9, Model I includes marker or dummy varibles for each of the states except a reference state, in this case Florida. These marker variables capture the differences in construction costs among states such as Florida and Vermont or Florida and Idaho that are associated with differences in building codes, seismic and geographical patterns, cost-of-living and other factors that effect differences in state building costs. The coefficient estimates for these state dummy variables are not included in Table 9. (Variables in rows 23 and 24 are not used until Model III,so these cells are blank in Model I.) Finally, Model I presents two statistics in rows 25 and 26. The adjusted R square statistic in row 25 is a measure of the goodness-of-fit of the overall model to the data. This statistic shows a good fit 58 And row 26 reports the number of projects studied in this model--2,686. Having worked somewhat carefully through Model I, Model II can be discussed more rapidly. Model II looks only at schools (public and private) built in jurisdictions where prevailing wage laws hold sway. The results are similar to Model I—the no-law states. Total costs rise with total square feet but there are economies of scale (rows 2 and 3 column c). Two story and multistory schools cost 5.1% and 9.7% more respectively. The result for two-story schools is very similar to that in Model 1. The multistory result is somewhat lower than in Model I. This may be due to a different mix of three, four and more story projects in the two samples. All these results are statistically significant. In the case of states with prevailing wage laws, middle schools of the same size as elementary schools are not more expensive. The estimate is only 0.7% higher and it is not statistically significant. High schools are more expensive, but only 2.91/o more compared to 5% more in states without prevailing wage laws. Additions were less expensive compared to new construction (6.9% less for similar sized projects), additions and alterations were 4.1% more expensive than new construction, and additions and alterations together were were less expensive than new construction (10.2%). All these results were statistically significant 58 The F-statistic is 234 yielding a second confirmation of goodness-of-fit,ADlheo#w models report sirialyhigh F-statistics. 64 The business cycle differed in states with prevailing wage laws compared to states without these regulations in the 1990s. Compared to 1991-92, real, inflation-adjusted, costs of school construction went down in 1993 and 1994, and the result was statistically significant in 1994. Statistically significant increases over 1991-92 in the real cost of school construction in this group only come about late in the cycle. By 1998, real school construction costs had risen 9.5% over 1991-92 and in 1999 real costs were 15.1%above the 1991-92 bench mark. The difference between states with and without prevailing wage regulations in the pace and pattern of real cost increases can be tied to the fact that jurisdictions without prevailing wage laws are concentrated in the South and the Rocky Mountain-High Plains West. The U.S. business cycle has always varied geographically. The fact that real costs rose first and somewhat faster in states without prevailing wage regulations cannot be tied to the lack of those regulations. Deregulation did not cause school construction costs to rise. But we need to keep in mind the effects of the business cycle if we want to try to measure any possible effect of legal policies on school construction costs. The timing of the start of school construction matters more to states with prevailing wage laws. A bid acceptance date in winter or spring will result in about a 4% decrease in school construction costs. This result is statistically significant. And it is due to the fact that many prevailing wage law states are in the snowbelt. If in real estate, location is everything, in the snowbelt timing is something to consider. In prevailing wage law states, public schools are 10.7% more expensive than private schools. This is a statistically significant result. Given that public schools in states without prevailing wage laws were 8.4% more expensive than similar-sized private schools we cannot say that the more expensive public schools in states with p Y p prevailing wage laws are 10.7% more expensive because they were built under these regulations. But can we say the difference between 8.4% in states without regulations and 10.7% in states with regulations can be tied to the regulations?59 Model 111 exanines this question. In Model III, we consider all states—both those with and without prevailing wage regulations. This increases our observations to 6,568 school construction projects. In combining our data, we must add two variables. Not only must we include the marker or dummy variable for public school (compared to private school), we must also add a variable for whether or not the state has a prevailing wage law. This variable is in row 23 and turns on if a project was built in a state with a prevailing wage regulation. But the project could be a private school in a state with a prevailing wage regulation. So an additional marker variable is added in row 24 turning on when the project is a public project in a state with a prevailing wage law. In Model III these two variables in rows 23 and 24 are our focus variables. Controlling for all other factors, when we add them together, we get the model's estimate of the effect of prevailing wage regulations on public school construction costs. 59 In order to push on to this more interesting question,I will only footnote the fact that Model II had 3,382 observations,and the goodness of fit statistics were acceptable. 65 By themselves, neither of the two new variables are statistically significant. Added together, the model estimates that prevailing wage regulations raise public school construction costs by 2.4%. But this result, shown in row 27 column d is not statistically Significant either.60 proponents of the hypothesis that prevailing wage regulations raise school construction costs may take some comfort in the model's estimate that they do— by 2.41/o. But the fact that this result is statistically insignificant means that this is an unreliable result. This putative savings is uncertain and unsure. There is something else to consider. 'These data are accepted-bid data. They do not include change orders, cost over-runs or downstream maintenance costs. The bidding structure in public construction is unique. Reputation of the contractor cannot be considered in most instances. Thus, there is room in most jurisdictions for some contractors to shave his or her bid in order to win the public job with the hope that change orders will restore the contractor's profit margin. If this occurs, or if the contractor attempts to cut costs through shoddy workmanship, then the front end 21/o savings from the low bidder may not result eventually in a 2%savings in final costs. An uncertain and potentially costly 2.41/o savings on public school construction costs may nonetheless be an attractive goal given that public budgets are always tight. Subsequent parts of this report argue that going after this uncertain 2.4% creates known costs. Eliminating prevailing wage regulations will result in lower apprenticeship training rates, higher injury rates and greater pressure on public health and old-age assistance programs. Are there alternatives that may be less costly ways of saving on school construction costs? Models II and TV say that there are. In Model II, where only states with prevailing wage laws are considered, a statistically significant 4% savings on total cost can be achieved simply by timing school construction to begin in the winter or spring. Recall that these start dates are the date of bid-acceptance. Bid-acceptances in the winter or spring yield 4% lower overall bids for similar projects compared to accepting bids in the fall. Figure 6 shows that school authorities do concentrate bid-acceptances in the spring. But fully one-fifth of bid-acceptances are in the fall with construction beginning in the teeth of winter. An alternative to prevailing wage deregulation would be a more optimal seasonal timing of construction. Model IV includes all states and is similar to Model III with one exception. The various marker variables for years have been replaced by a single variable, the unemployment rate in each state for each year. This is a measure of the overall business conditions in each state. The estimated coefficient for this variable found in row 18, column a shows that a one percentage point increase in a state unemployment rate yields a 2.1% decline in total school construction cost. Building into a builder's market costs money. Building into a slack market saves money. And such a strategy yields social dividends as well. By spending counter-cyclically, school authorities can soften the effects on their community of slowdowns in the economy. 60 The t-test in this case is(bl+b2)/sgrt(var bl +var b2+2'cov bl b2) 66 Distribution of Start Dates by Season in States with Prevailing Wage Law 35% 339� 30% 25% 110% 1'9re 2U°k 15% 10% 5% 0% Winter Sprite Summer Fall Figure 7:In states with Prevailing Wage Laws Moving Bid-Acceptances Back to the Winter and Spring Would Save 4%on Construction Costs Distribution of Bid-Acceptance Dates for Schools in Kentucky by Season 40% . 35% 38% 30% - 25% 24% 20% 21 15% 17% 10% 5% 0% Winter Spring Summer Fall Figure 6: Kentucky Concentrates Bid-Acceptances in the Spring But Could Save Money by Moving More Late Summer and Fall Acceptances into the Winter-Spring Period 67 Conclusion Simple, back-of-the-envelope calculations that claim you can build four schools for the cost of three if you eliminate prevailing wage regulations are wrong. They miscalculate what labor costs are as a percent of total costs and they inaccurately assume that draconian cuts in construction wage rates will not effect worker skills and morale, contractor purchases of equipment and crew management or industry training. Real public school construction costs in Kentucky rose in the 1990s. But that real increase in costs occurred before the state applied prevailing wage regulations to school construction. Econometric models of school construction costs based on over 6,000 projects built in the 1990s do find that one might save 2.4% on total school construction costs if prevailing wage regulations were lifted. But this result is statistically doubtful. Technically, this estimate is not statistically significantly different from no savings whatsoever. However, proponents of prevailing wage deregulation may wish to believe the result nonetheless. School authorities mindful of tight construction budgets may grasp at this straw for lack of alternatives. But such policy changes out of bias or desperation are unwise. The bidding structure in the public arena is ill-suited to fend off bid-shaving strategies that cut initial bid offers in the expectation of manipulating change orders down the road to restore profit margins. Lowest-bonded-bidder procedures in the public sector do not consider contractor reputation and crew workmanship as these factors are considered in the private sector. This weakness raises the specter of higher downstream maintenance costs offsetting any small, putative upfront savings associated with deregulation. And there are alternatives which offer a higher and statistically more certain prospect for saving on school construction costs. The models of school construction indicate that time means money. In this case, tinting the start of construction to avoid cold weather conditions and tight construction markets have the potential of saving two to four times as much as any deregulatory scheme. And not only is the magnitude greater, but statistically,these sorts of savings are more certain. 68 i How Does Kentucky►'s Prevailing Wage Law Effect Apprenticeship Training? WM a Special Focus on the Current Slaps Crisis in Construction. 9 r. C f Roofers in the 1920s 69 The Engineering News Record (ENR) is a significant industry trade paper in construction. The Business Roundtable is a major organization of owners who purchase construction services. Cockshaw's Construction Labor News+Opinion(Cockshaw's Report) is the most widely read independent newsletter on construction labor issues. All have recently done surveys attesting to the presence of significant skill shortages in the U.S. construction industry. Our recent extended economic prosperity has expanded the demand for construction services. This expanded demand has precipitated the skill crisis in construction. But the underlying cause of this crisis has been the slow shift over the last twenty years towards open shop construction. On average, open shop contractors train less. They employ a less skilled labor force. These facts help account for the fact that union contractors can pay measurably higher wage rates to a more trained labor force and not have significantly higher labor costs. As this chapter will show, the open shop trains only one out of every four construction apprentices. Because apprentices in open shop programs have lower graduation rates, less than one out of every five apprentices graduating to journeyman status in construction come from the open shop. In some trades, such as operating engineers and iron workers, only 2% of the graduating apprentices come from open shop programs. In collectively bargained apprenticeship programs, women and minority apprentices are less likely to graduate than white male apprentices. In open shop programs this is also true. However, a vwman or a minority apprentice in a collectively bargained program is more likely to graduate than a white male apprentice is likely to graduate in an open shop program. Thus, a change in public policy discouraging collective bargaining in construction would have the dual effect of lowering the amount of apprenticeship training in the face of a skills shortage and lowering even further the opportunities for women and minorities to obtain skilled careers through apprenticeships. We begin a discussion of these issues by considering some of what the Engineering News Record and the Business Roundtable have said about the scope and causes of the current skills and training crisis in construction. The industry has known for much of the past decade that it was headed for manpower trouble when the business cycle turned up.—Nonunion contractors working in bustling areas appear to have the biggest manpower problems, according to the survey results. For example, 56% of the union crafts in the West reportedly have no labor shortages while only 10% of the open shop crafts have no problem. "I would guess that some of the labor shortage exists because the open shop has pirated all the available, qualified union workers, and now suffers the lack of training programs of their own to produce open-shop crafts people," says Donald A. McKay,chairman of union mechanical and sheet metal contractor Tougher Industries, Albany, N.Y. "It's frustrating to hear them whine to the owners for help with their educational programs, while spending a pittance on training." McKay notes that the Alliance of Mechanical, Electrical and Sheet Metal Contractors spends about$100 million a year to train union workers in those trades.... "Craft Shortages Creeping In,"Engineering New Record (ENR), December 25, 1995, Vol. 235, No. 26, pp. 34-5. 70 Although the economic and market picture for the open shop appears very bright, the [America's Open Shop Committee) survey revealed storm clouds on the horizon. One dampening note was the continuing, and worrisome, shortage of skilled tradesmen. This was revealed in a survey question which asked respondent to categorize the availability of 13 craft classifications as `scarce," "adequate," or `plentiful." Eleven AGC chapters advised that 11 or more crafts were scarce in their markets. Moreover, there were a number of regions where all 13 trades were reported in scarce supply. Cockshaw's Report, September 1999 Companies are currently experiencing significant problems in staffing construction projects, resulting in escalating costs and costly schedule delays....In late 1996, The Business Roundtable surveyed its member companies...Over 60 percent of the survey respondents indicated they had encountered a shortage of skilled craft workers,and 75 percent reported the trend had increased over the past five years.—The union sector has always excelled in craft training through the joint labor/management apprenticeship programs open shop, as a whole has not supported formal craft n p p, pp training to the extent necessary. They have succeeded by attracting skilled workers from the union sector as market share shifted and recruiting skilled workers from competitors as individual workload changed. As the well begins to dry up, the ability to use these methods decreases....Through the years, the subject of funding for training has come up repeatedly. All of the discussion has been on the open shop side. Training on the union side has always been required and paid for by the owner. A trained work force was expected and guaranteed by the contractors with costs passed on to the owner as part of the collective bargaining labor rate. It has been a different story on the open shop side. "Confronting the Skilled Construction Work Force Shortage," The Business Roundtable, Construction cost Effectiveness Task Force, October, 1997,pp. 2, 5, 6, 8 & 14. Prevailing wage laws promote collective bargaining. Collective bargaining, in turn, promotes higher wage rates in construction. Yet, as the previous chapter has shown, there is no measurable correlation between prevailing wage laws and substantially higher construction costs. How can this be? How can contractors pay higher wage rates and yet not have measurably higher construction costs? The answer lies in training. Apprenticeship Training Under Collective Bargaining and the Open Shop Contractors that participate in collective bargaining do the lions share of apprenticeship training in construction. The following Figure 8 shows the number of newly enrolled construction apprentices in the United States for each year since 1989 61 These data are 61 These data are from the U.S.Department of Labor,Bureau of Apprenticeship Training. They represent approximately 70 percent of all construction apprenticeship programs in the United States. Similar results are found using Kentucky state apprenticeship data in the study by William J.Londrigan,M.P.A.and Joseph B.Wise,III,M.B.A, Apprentice Training in Kentucky,A Comparison of Union and Nonunion Programs in the Building Trades,Building Trades Apprenticeship Coordinators/Directors Association of 71 broken down by the apprentices employed by union and nonunion contractors. While the number of new apprentices entering construction varies with the construction business cycle, the proportion trained under collective bargaining remains roughly the same. Approximately three out of every four new apprentices enroll in programs created by collective bargaining. New Apprentices by Year 35,000 30,000 s,7az 2 25,000 6,684 7,asa 7,516 7s85 � I m 6,857 6,993 G 20,000 Q 0 Non-union 15,000 ■Uninn d za,ass 20,730 21,510 20,518 20,755 Z 10,000 17,737 16saa 5,000 1989 1990 1991 1992 1993 1994 1995 Figure 8:The Number of Newly Enrolled Construction Apprentices by Year and Union and Nonunion Programs,1989 to 1995 Collective bargaining requires that contractors contribute a specified amount of money for every hour of work into an apprenticeship training fund. This fund is used to hire instructors, to buy tools, equipment and materials and to pay for instructional facilities. In effect, all apprentices who enter programs maintained by collective bargaining are on scholarships provided by their employers. This means contractors who have signed collectively bargained agreements have a vested interest in seeing that their apprentices get trained and successfully graduate. In open shop apprenticeship programs, the apprentice typically must pay a larger share of his or her own training costs. This may come in the form of tuition payments, lower wages, or both. As a result, the contractors have less of a vested interest in assuring that enrolled apprentices successfully graduate. Consequently, not only are three out of every four new apprentices enrolled in collectively bargained programs, but also, once enrolled in a collectively bargained program, the apprentice is almost twice as Kentucky,Inc,March, 1997. 72 likely to graduate. The following Figure 9 shows that for the entering classes of 1989 and 1990, by the end of 1995, 37% of the apprentices in collectively bargained programs had dropped out. In contrast, over half--54%--in the open shop programs had dropped out. In the collectively bargained programs, 41% of the classes entering in 1989 and 1990 had graduated to journeyman status while only 25% of the apprentices in the open shop had graduated to journeyman status. In both types of programs, 21% of the classes entering in 1989-90 were still enrolled apprentices by the end of 1995. Thus, not only does collective bargaining encourage training,it encourages the completion of training. Graduation Rates for Apprentices Under Collective Bargaining and in Non-Union Programs 100% 37% 80% 54% 60% �1Cancellation 21%' ®Active 'Z ■Completion 40% 21% 41% 20% 25% 0% Union Non-union Figure 9:Graduation Rates by 1998 for Apprentices Enrolled in 1989-90 Under Collective Bargaining and the Open Shop ® A A ! ! • ! s ! o • � • ! , • o a s 73 The rubber meets the road in construction apprenticeship training when the apprentice turns out as a journeyperson. As the following Figure 10 indicates, overall, collectively bargained programs turn out 82% of all construction journeymen and women trained through apprenticeship in the construction industry. In some crafts, open shop apprenticeship programs account for only one or two percent of all the apprentices graduating to journeyman status. For instance, only 1% to 2% of the apprentices graduating to journeymen status among operating engineers or structural steel workers (iron workers) come from open shop apprentice programs. Only 9% of the graduating bricklayers and 8% of the graduating carpenters come from open shop apprenticeship programs. Even among plumbers where the open shop has its largest share of graduating apprentices, two-thirds of all plumber-apprentices graduate from collectively bargained programs.62 Relative Contributions of Collectively Bargained and Open Shop Programs to Graduating Journeyworkers (Classes of'89, '90 and '91) Bricklayer 91% Carpenter 92% Electrician 70 Oper.Eng. 99% Painter 96% �Union Pipefitter 88% 0 Non-union Plumber 66% Roofer 92%® Sheet Metal 86% Struc.Steel 98% Other 844 i All occupations 82% 0% 20% 40% 60% 80% 100% Figure 10:Share of Journeyworkers Graduating from Apprenticeship Programs Broken Down by Occupation and Collectively Bargained versus Open Shop Program 62 These are graduation rates at the end of 1995 for apprenticeship classes having entered in 1989,1990 or 1991. 74 Prevailing Wage Regulations,Minorities and Apprenticeship Training Historically, minorities have been excluded from good construction work. During the 1950s and 1960s, minority leaders fought hard to break down discriminatory barriers in construction. As a consequence, construction apprenticeship programs have been closely regulated in the last three decades to insure fair admissions procedures. These regulations apply to both joint labor-management programs organized under collective bargaining and programs operated by merit shop employers. However, anti- discriminatory regulations and oversight apply only to apprenticeship programs with five or more apprentices. This covers virtually all joint labor-management programs under collective bargaining because these are multi-employer programs. Programs with many contractors participating tend to be well over 5 apprentices. For instance, in Kentucky, on average, the heavy and highway contractors working with the International Union of Operating Engineers have on average 63 apprentices enrolled at any time. The Kentucky Chapter of the National Electrical Contractors Association in cooperation with Kentucky locals of the International Brotherhood of Electrical Workers have over 1,000 apprentices at any one time. The outside linemen have a separate apprenticeship program with around 34 apprentices enrolled annually. The plumbers and fitters in conjunction with the Kentucky Mechanical Contractors Association have around 220 apprentices enrolled per year. The carpenters have around 250 enrolled per year. All these programs must meet strict guidelines in their application and enrollment procedures. But open shop, single contractor programs with fewer than five apprentices at any one time need not concern themselves with issues of affirmative action. The consequence is that in states that do not have prevailing wage regulations there are fewer programs overall, fewer apprenticeship programs organized under collective bargaining and a higher proportion of small programs operated by single, merit shop contractors. Thus, it is not surprising that in states without prevailing wage laws, minorities are under-represented in apprenticeship programs while in states with prevailing wage regulations, the representation of minorities in apprenticeship programs reflects the overall proportion of minorities in the state population. Figure 11 shows the minority representation ratio for states by prevailing wage regulation status. The data are from the U.S. Bureau of Apprenticeship and Training for the years 1987 to 1989. They show that in states with prevailing wage laws, the minority apprenticeship representation ratio is 96%. At 100%, the representation of minorities in apprenticeships would look just like the representation of minorities in these states' populations. So a ratio of 96% means that minorities are slightly under-represented in construction apprentices in states with prevailing wage laws. However, in states that never had prevailing wage laws, the minority representation ratio is 82% and in states that repealed their laws, the minority representation rate 81%. In short, minorities have a much more difficult time getting into apprenticeship programs where prevailing wage regulations are absent. This is due to the fact that deregulation encourages single- employer apprenticeship programs and takes apprenticeship out from under affinnative action regulations. 75 _ ( onsmaction Industry) 109%means the percentair or ninari.4 apprertces uz rhts the percengge otminPrmes in poPulafian. S x a � c�tiroe:US Cy l94T,�5Al�,r�cyut 5 E?ate xe for{ �47 19EQ f „ s Figure 11:Minority Representation Ratio for Apprentices by Prevailing Wage Regulation Status It is one thing to get into an apprenticeship program, it is another thing to get out. Graduation rates of merit shop apprenticeship programs are substantially lower than those for apprenticeship programs organized under collective bargaining. This should not be a surprise. Under collective bargaining, contractors are obliged by the collectively bargained contract to contribute so much per hour for every hour worked by all workers into a training fund. This fund is used to invest in new apprentices. Contractors, looking for a return to their investment, try to bring in the most promising new apprentices possible. Once enrolled, signatory contractors have every incentive to see to it that these apprentices get solid training that leads to graduation as skilled journeyman and women. Some merit shop contractors are equally committed to apprenticeship training. But when the upfront investment is not there or is less, the incentive to see to it that apprentices turn out as journeyworkers is less. Consequently, graduation rates of apprentices in the open shop is less. 76 Relative Performance of White and Minority Apprentices in the Union and Nonunion Programs 60% 62% 62% 60% ❑Cancellation 40% � 20% Adve 2f% go 'on 28% a7 1 Union Union Cron-union Non-union White Minority � "e Mina* Figure 12: Status of the Apprenticeship Class of 1989-90 at the End of 1995 Figure 12 shows the status of the entering class of construction apprentices in 1989-90 at the end of 1995 broken down by union and nonunion programs and then broken down by white and minority apprentices. (These data are from the U.S. Bureau of Apprenticeship and Training. Minorities are black, Asian, American Indian and Hispanic.) For white apprentices entering into joint labor-management apprenticeship programs in 1989-90, one third (34%) had dropped out of construction apprenticeship training by 1995. Forty-five percent had graduated into joumeyworker status in their trade and 21% were still enrolled in the apprenticeship program. In contrast, slightly over half of the 1989-90 entering minority apprentices in joint labor-management programs had dropped out by the end of 1995. Twenty percent were still enrolled and 28% had graduated. Clearly, minority apprentices in joint labor-management programs do not fare as well as white apprentices. But, minority apprentices in union programs do as well or slightly better than white apprentices in merit shop programs. The real casualties are minority apprentices in merit shop apprenticeship programs. Not only are there fewer minority apprentices, but few of these few graduate. Only 18% of the minority entering class of 1989-90 in merit shop programs had graduated by 1995. Fully, 62% had dropped out. Thus, minority apprentices are heavily under-represented in apprenticeship programs in states without prevailing wage laws. And if those programs are merit shop 77 programs, the odds of those few minorities ever turning out as journeymen and women are grim. The Use of Skilled and Unskilled Labor Under Collective Bargaining and the Open Shop Union contractors and nonunion contractors approach the job site with very different work crews. The average nonunion contractors use a labor strategy that relies upon a limited number of skilled workers leading a larger number of unskilled workers through the construction process. For example, a skilled nonunion electrician may oversee a set on unskilled wire-pullers. The skilled nonunion electrician will color code studs and the wire-pullers will pull wire through those studs based on color coding. In a contrasting example, the skilled union electrician may pull the wires himself or have an apprentice do this. The nonunion contractor may have a cost advantage with his low- wage, unskilled approach to wiring a building as long as everything is simple and expected. But the union contractor may have the cost advantage if unexpected problems crop up that only a skilled and knowledgeable worker can handle. The nonunion contractor relies on his limited number of skilled workers to handle the unexpected. However, an unskilled worker can make a mistake without even knowing that a mistake was made. The union contractor with a more skilled labor force seeks a cost advantage by knowing that each activity is done by someone who knows the ins and outs of that activity. So problems are nipped in the bud rather than done over when caught. In short, the union contractor pays a higher wage rate but tries to offset that higher wage rate with fewer,more skilled and productive workers. Figure 13 illustrates the difference between the labor force strategies of the average union and nonunion contractors. These are data from Pennsylvania in 1995-96 but they likely reflect trends nationally including Kentucky. They divide building construction contractors into those working under collectively bargained contracts and those working without a collectively bargained contract. The top 5 most commonly found occupations in building construction are reported in Figure 13. These occupations account for 68% of all union workers and 62% of all nonunion workers with the exception of apprentices which were excluded from the survey.63 Union contractors employed a higher percentage of carpenters, electricians, sheetmetal workers and plumbers-pipefitters compared to nonunion contractors. In contrast, a higher percentage of the nonunion contractors' crew mix consisted of unskilled laborers. The foregoing apprenticeship data and these employment data show that on average, open shop contractors train less and use fewer skilled workers compared to union contractors. s I ,Ili III !< + i / • # A! ! I i i i' ese a a come om a ennsyvan+a a e a en er reva++ng age urvey comm+ss+one y e Pennsylvania State Labor Department,data file'Working1txt". 78 Crew Mix of Union and Open Shop Contractors in Pennsylvania 1995-96 L 0 25% LL L. O .0 J 20% O 15% M Nonunion E Union 10% r" C �b Q V 5% r.. O d V 0% L 41 a carpenter electrician sheetmetal plumber, laborer worker pipefitter Figure 13:Relative Use of Skilled and Unskilled Labor by Union and Nonunion Contractors, Pennsylvania,1995-96 t 8 79 Table 10:Nonunion Contractor Use of Skilled and Unskilled Labor Compared to Union Contractors,Pennsylvania 1995-96 Occuoation Nonunion Union carpenter 18% 23% electrician 13% 17% sheetmetal worker 8% 11% plumber,pipefitter 6% 9% laborer 17% 8% 111 To see more clearly the relative use of skilled and unskilled labor under collective bargaining and in the open shop, Table 10 repeats the foregoing data in Figure 13 on carpenters, electricians, sheetmetal workers, plumbers-fitters and laborers. By taking the ratio of nonunion to union use of these occupations in their crew mix, we see that on average, nonunion contractors in Pennsylvania in 1995-96 use 79% as many carpenters, 76% as many electricians, 73% as many sheetmetal workers, 67% as many plumbers and pipefitters but over twice as many laborers. The employment of relatively fewer skilled workers by open shop contractors may help explain, in part, why open shop contractors train fewer apprentices. But the fact that open shop contractors train fewer apprentices may also help explain why they end up using relatively fewer skilled workers. Possible Reasons Why the Open Shop Trains Less and Uses Fewer Skilled Workers Why do [open shop] contractors not train? Many reasons are given. It is cost prohibitive. Investment is lost when a trained worker moves to a competitor. Many fail to recognize the need or appreciate the productivity effects. ---"Confronting the Skilled Construction Work Force Shortage," The Business Roundtable,Construction cost Effectiveness Task Force,October, 1997,p. 6. Training in the construction industry is a classic case of what economists call a market failure. Construction is a boom bust industry in many respects. Not only does the construction business cycle swing much more widely than does the economy as a whole, but also, specific contractors have to gear up and slow down their operations based on their own particular fortunes at winning construction bids. Along with this boonrbust, ramp-up/shut-down structure that is fairly unique to construction, the industry is organized along a complicated structure of subcontracting. Subcontracting is a way for a contractor to allow a more expert subcontractor to handle a particularly difficult or specialized part of a project. It is also a way to export headaches. When in doubt, it is sometimes better to contract out. Labor skill shortages can be just the kind of headache worth contracting out. 80 The boom-bust, ramp-up/ramp-down, subcontract-out headaches structure of construction makes most contractors focus on the short-run. In the short-nun, the available supply of trained construction workers is fixed. If you have a shortage, all you can do is bid craftsmen away from someone else. It takes four to five years to turn an electrician, plumbing, fitter or sheet metal apprentice into a skilled journeyman. By the time you train someone for the job,the job is gone. Anyway, if you train someone, you might just be subsidizing your competitor. The worker you train in all likelihood will be down the road and working for your competitor in the not too distant future. If you undergo training costs and your competitor does not, then your competitor can have his cake and eat it too. He can win that job today, since he has lower costs today because he does not train. And, he has just as much chance as you do of having skilled labor tomorrow because skilled labor moves around. You, the honest contractor that diligently trains for the future--you're a chump in the cutthroat competition that is the construction industry. The historical solution to the market's failure to train in construction has been collective bargaining. A collectively bargained contract between a union representing construction workers and an association representing contractors has traditionally resolved the Problem of meeting long-term training needs in a market that rewards only the short run calculations of contractors. If you and I as contractors are signatories to a collectively bargained contract, that contract will not allow me to be injured by you, my competition. Together, you and I and the other signatory contractors have agreed that for the good of the industry in the long-run, so much per hour (say 50 cents) will be put into an apprenticeship training fund. That means for every hour any of my workers are on a job, 50 cents goes for training apprentices. When I write up my bid, I know I have this cost. But what is more, I know you have this cost as well. I know that you might win the bid over me, but it won't be because I kept in mind the future training needs of the industry and you didn't. We both have to put the collectively bargained training costs into our bid. No pirating is possible because in the future I may hire the worker you trained but I shared in the cost of that worker's training. Thus, with collective bargaining in place, the contract serves as a mechanism for the market to provide training. s r s 81 Conclusion. Public policies that reduce construction apprenticeship training in the face of a critical skills shortage are wrong-headed and self-defeating. Repealing Kentucky's prevailing wage law would seriously threaten the ability of the state's construction industry to provide a well-trained workforce to meet the public and private needs for high-quality construction services. The fact is that over 8 out of every 10 graduating construction apprentices in the United States come from programs financed and managed by collective bargaining. In some crafts such as operating engineers and structural steel workers (also known as iron workers) collective bargaining accounts for virtually all the apprentices trained in these skills. Construction apprenticeship programs jointly managed by contractors and unions are one of the best examples of the labor-management cooperation that many observers believe is crucial for world-class competitive success in the next century. Public policy should encourage this kind of cooperation. And that is just precisely what prevailing wage laws encourage. By tolerating and promoting the practice of collective bargaining in construction, prevailing wage laws promote labor- management cooperation, apprenticeship training, continued training of the journeyworker and the development of a high-skilled labor force. Solid, high-quality, registered and monitored apprenticeship training also makes for a safer labor force. Safety training is a central focus of each and every collectively bargained apprenticeship program in Kentucky. It stands to reason that a more carefully trained labor force is a safer labor force. As the next chapter will show, the exemption of Kentucky schools and municipalities from Kentucky's prevailing wage law in 1982 led to higher serious injury rates in Kentucky construction. Higher injury rates hurt the worker and the worker's family. They also increase workers compensation costs and interfere with construction schedules. Construction is an inherently dangerous industry. Prevailing wage laws keep down the costs of injuries the right way by encouraging the training that reduces the risk of injury in the first place. 82 The Rise in Serious Injuries in Kentucky Construction After Schools and Municipalities Were Exempted from the State's Prevailing Wage Law The general recipe for safety in construction is simple: larger, more experienced contractors working with well-trained and experienced crews are safer than smaller, less- experienced contractors working with less experienced and less trained workers. 64 The suspension of Kentucky's prevailing wage law for schools and municipalities in the 1980s set in motion a train of events that led to the proliferation of less experienced contractors teamingu with less trained and less experienced workers. This led to more p p and longer serious injuries for Kentucky construction workers. The General Relationship Between Prevailing Wage Regulations and Safety Cutthroat competitiveness in contracting. The repeal of the state prevailing wage laws often will lead to a burgeoning of start-up contractors with limited track records. These new entrants join existing contractors in a heated bidding process that can put safety at risk. Because of their relative inexperience, new firms tend to face greater on-site coordination problems than firms with longer track records. Such problems can add to costs, but also directly endanger safety. Problems in coordination, perhaps related to delivery of materials and equipment, or in scheduling work with subcontractors, lead to greater uncertainty with respect to the construction schedule. Uncertainty is a breeder of safety risk, as workers can less easily anticipate and plan for the daily contingencies of work. 64 C.Culver,M.Marshall,and C.Connolly,Construction Accidents: The Workers'Compensation Data Base, 1985-1988, Washington,DC,OSHA Office of Construction Engineering, 1992. 83 i New entrants in the industry also are generally smaller in size than established firms. Smaller firms have worse safety records than larger firms, in part because of greater laxity of enforcement of safety rules and the relative absence of formal safety programs. Of greatest importance, however, is the firm's reaction to increased pressure to cut costs in the face of intensified competition and cost overruns. There is a tendency to speed up work and cut back on safeguards in the face of such pressures. Workforce turnover. When state prevailing wage laws were repealed, worker turnover increased significantly, as the industry found it harder to retain workers for long-term careers (see Chapter Three). Repeals resulted in a decline in the union share of the construction labor market, driving down average construction wages in the state and decreasing union apprenticeship training for construction. The decline in wages and in health and pension benefits (see Chapter Five) drove experienced construction workers from their trades to careers in other industries. In states that retain their prevailing wage law — compared with those that never had such a law or repealed such a law — the proportion of construction workers receiving training is higher and injury rates are lower. A decline in wages and benefits leads to a flood of inexperienced workers into the industry as well as a decline in skilled, experienced workers needed to supervise the recruits and to assure that they work safely. Decline in the skill base of the construction labor market. Experience is a major determinant of safe work performance — and productivity. Training of skilled construction workers is normally conducted through apprenticeship training programs, most of which are operated by unions and employers through joint trust funds. An integral part of this training is learning on the job while properly supervised. In that way, workers learn from experience while on a variety of projects. Among other things, apprentices are trained to identify and correct ergonomic problems, to detect physical hazards, and to detect the presence or release of hazardous chemicals. Knowledge about safety and health hazards, appropriate protective measures, and hazard communication methods are all important elements that apprenticeship programs provide. When prevailing wage acts are repealed, training and apprenticeship programs decline and the skill base of workers erodes. Without employer incentives to continue apprenticeship programs, knowledge of proper safety and health procedures declines as well. Summary. The combination of these factors — cutthroat competition, a decline in training, and an erosion of career attachments to the industry — affects the safety-related skill and experience base of the construction labor force. Workers become more injury- prone and less knowledgeable about the kinds of risks they are taking. Furthermore, as the workforce becomes less skilled and its wages in construction decline, workers are forced to take more safety risks to simply make a living. Furthermore, contractors caught in the competitive speed-up often press their workers to hurry and take more chances. Workers are put at increased risk in an already hazardous industry. 84 The Rise in Serious Injuries in Kentucky After Schools Were Exempted in 1982 Annually, the various state departments of labor in cooperation with the U.S. Department of Labor, Bureau of Labor Statistics, conduct an occupational injury and illness survey. This survey reports for a variety of industries, including construction. The survey reports the number of workers employed in each industry category, the number of injury cases and the number of injury cases that result in lost days from work. I have gathered these data for Kentucky for the period 1976 to 1991. This allows us to examine injury rates in Kentucky construction prior to and after schools were exempted from the state prevailing wage law in July of 1982. The Bureau of Labor Statistics reports the number of injury cases, the number of serious injury cases, the days lost from work from serious construction injuries. These data are reported annually along with the average number of construction workers employed for each year. From these data we can calculate three indices of the prevalence of serious injuries in construction. In Table 11, we compare serious injury cases as a percent of all construction workers average for the six years prior to suspension (1976 to 1981) and for the nine years after suspension(1983 to 1991).65 Table 11: Comparing Annual Average Serious Injury Rates in Kentucky for the Six Years Prior to and the Nine Years After the 1982 Exemption of Schools and Municipalities from the State P P Prevailing Wage Law Serious Injuries Lost Days per Seious Injuries as a per Worker Serious Injury Percent of All Injuries Before School & Municipal Exemptions 4.9% 18.6 45.3% After School & Municipal Exemptions 5.4% 21.6 48.9% Increase 0.5% 3.0 3.6% t statistic 2.4 2.6 3.2 Statistical) significant? Yes Yes Yes Percent Increase 11% 16% 8% ss 1982 is omitted because until July 15,1982,the law was applied to schools and municipalities and after July 15 the law was suspended. These annual data cannot break down 1982s4sies into before and allerrrarthswithin the year. 85 Serious injuries are defined by the BLS as work injuries that result in lost days from work. In the period 1976 to 1981, on average 45.3% of all injuries were serious enough to result in lost days from work. In this pre-exemption period, on average 4.9% of all construction workers were annually seriously injured enough on-the-job to lose days from work. On average,these seriously injured workers lost 18.6 days of work. In the nine years after schools and municipalities were exempted from prevailing wage regulations, 48.9% of all injuries annually were serious injuries. Annually, 5.4% of all construction workers were injured on-the-job seriously enough to miss days of work. On average, these seriously injured workers missed 21.6 days of work. In each case, the average index of serious injuries rose after exemption compared to before exemption. And in each case this increase was statistically significant. (The t statistics are a measure of statistical significance and in each case statistical significance is found at the 5% level or lower). These differences are statistically significant. Are they practically significant? Increase in Kentucky Serious Injury Mates After Prevailing Wage Exemptions in 1882 18% 16°fo 1b°b 14% 12% 1 � 8�0 8% 4% 2% C Serious Injuries per Wofiser Lost Days per Serious injury Siericaas Injuries as a Percent of As I*Hes Figure 14:The Increase in the Measures of Serious Injuries After the Exemption of Kentucky School and Municipalities from the State Prevailing Wage Regulations Figure 14 shows the relative magnitude of the increases in serious injury after prevailing wage regulations were lifted. Serious injuries per worker rose by 11 percent. Serious injuries as a share of all injuries rose by 8 percent and lost work days per serious 86 injury rose by 16%. Serious injuries create costs to the worker and to construction. The worker is partially compensated for his injury through workers compensation. To the extent the worker is compensated, the industry and eventually the consumers of construction services pay for this increase in serious injuries. But the costs do not stop there. To the extent the worker is not made whole or cannot be made whole, the worker and his or her family pay for the injury in lowered earning capacity and lost quality of life. Everyone wants to reduce workplace injuries. Open shop and union shop contractors abhor workplace injuries and seek to promote workplace safety. Prevailing wage regulations are one means of doing this. By encouraging skill formation, nurturing and preserving experience, prevailing wage regulations cut the hidden but very real and sometimes tragic costs associated with workplace accidents and injuries. 87 Prevailing Wage Regulations Support and Promote Family Health Insurance and Old Age Security Pension and health benefits play two crucial roles in the construction industry. First, by providing needed income security in old age and needed health coverage today, these benefits permit adults with families to participate in the industry while knowing that their families' basic needs are insured. Second, pension and health benefits help create and preserve needed skills within the industry. People willing and capable of acquiring the skills needed for solid, high quality construction are also people capable of acquiring the skills needed by many industries. If the construction industry cannot provide the basic benefits needed by families, the construction industry will steadily lose its better and more experienced workers to other industries that will provide these benefits. Merit shop contractors have difficulty paying their workers pension benefits or health insurance. This difficulty is rooted in the same market failure that prevents training on the open shop side of the industry. Construction workers move from job to job. They must move simply because today's building gets built and today's road gets paved. So eventually, the construction worker has to move on. In doing so, the worker often changes employers. Merit shop contractors find it both awkward, and not worth their while, to insure the health and old age of workers who will be with them a limited amount of time. So merit shop contractors develop insurance programs for their key workers who do stay for years. But the merit shop contractors find little reason and much difficulty in providing these same insurance benefits to the transient worker. Collective bargaining provides a mechanism for allowing and inducing contractors to provide health insurance and pensions. Construction projects still come to an end. Construction workers still move on to new employers. But the new employer like the old is a signatory to the collective bargaining agreement. That agreement requires that each employer contribute so much per hour on the worker's behalf into a pension fund and into health insurance. thus, when a union construction worker's child gets sick, the child is covered by health insurance. And when a union construction worker retires, he or she has something more than Social Security to look forward to. This is not only good for the construction worker and his or her family, it is good for the community as well. Construction represents around 5% of the labor market. Thus, in round terms, construction workers and their families represent 5% of our neighbors--neighbors one would hope could afford a doctor when a child is ill and neighbors one would hope could take care of themselves when they are old. Such 88 neighbors are less of a burden on the community as a whole and better neighbors to live next to. Annual Average Employer Contribution per Worker to ERISA Regulated Construction Pension and Health.Programs In Kentucky 1982.1992 $3,OCQ $2.58T gt boaM collective Bargaking $1,24t7 12Merit Sfro i,Q00 ASCU Pension IiaaRh ®calective B a' ' $1,2441 $2,587 > Mari Sha $51 $24 Figure 15:Comparison of Employer Contributions to Worker Pension and Health Plans in Kentuckyunder Collective ollective Bargaining and Within the Open Shop Figure 15 shows the average per worker employer contributions to pension and health plans in Kentucky that are regulated by EFJSA.66 In Kentucky,over the decade 1982 to 1992 k3' , the annual average employer contribution per worker to pensions was 1 24 p $ 0 under collective p echve bargaining and$51 among merit shop contractors. This underscores the problems merit shop contractors have in paying worker benefits. This do es es not mean that no merit shop contractor pays worker benefits. In fact, while in the minority, three types of merit shop contractors do pay benefits. The fast is the merit shop contractor that is trying to slip-stream behind the collectively bargained contract in his or her area and trade. This merit shop contractor follows a high-end strategy of using highly trained workers through paying top dollar wages and benefits. This contractor tries to beat his competitors by avoiding the administrative costs of collective bargaining (contractor organization dues, union dues, grievance-arbitration costs and sometimes apprenticeship costs). This high-end merit shop contractor has to pay benefits to attract and keep an experienced, trained workforce. He or she often tries to hire workers trained under collective ss Employee Retirement Insurance Security Act(ERISA)is a federal law that regulates all worker benefits that are held in trust. These data come from the ERISA required form 5500 and cover the years1982to1992 Alperdon benefits will be regulated by ERISA but some forms of health insurance will not. 89 bargaining. A second merit shop contractor who pays benefits is a small contractor that tries to avoid the ups and downs of the construction business cycle and provide his small crews with steady work. This contractor has to stay small so as not to get caught up in the booms and busts of construction employment. Often this contractor will focus on maintenance work and sometime he (or she) will platoon his workers through the unemployment system when work gets slow. This employer will pay benefits to hang on to an experienced and trusted workforce through the years. But this merit shop contractor is and has to be exceptional. By its nature constriction booms and busts. All contractors,indeed most contractors cannot avoid this instability. Only by staying small and out-of-the-way, can the exceptional contractor in either the union or open shop avoid the ups and downs of the employment cycle. The last and most common merit shop contractor that pays benefits plays the business cycle. This contractor expands in the boom to take on new and profitable work and necessarily hunkers down when things slow down. This contractor cannot keep all the workers he or she hired in the boom nor can the larger contractor platoon large numbers of workers through the unemployment office. So he lets them go. He does, however, try to keep his key experienced workers. These he tries to keep steadily busy even when things slow down. And he pays them benefits. Not all merit shop workers want benefits anyway. Merit shop workers are younger, and younger people are not as worried about old age or health problems. This is a chicken and egg phenomenon. Not only are the younger and less experienced merit shop workers less concerned about old-age and family benefits,but also because the merit shop contractor is not offering those benefits to most workers, the merit shop contractor attracts a younger and less experienced workforce. So while the lack of benefits may not be a problem for some open shop workers, it becomes a problem for the construction industry and society. When construction work is staffed by a younger and less experienced workforce, construction becomes more dangerous. Younger, less experienced workers are more prone to accident. This puts pressure on workers compensation costs and third-party wrongful death and injury costs. Less experienced workers are also more prone to shoddy work. This puts pressure on downstream building and road maintenance costs. Inexperienced workers can also be one cause of cost-overruns. But societal costs rise as well. Construction workers form around 5% of the labor market. When these workers go year after year without pension contributions, eventually society pays higher costs in old-age services. Even in the short run, the failure to provide health insurance for construction workers and their families puts pressure on public health services. We can estimate how many open shop workers in Kentucky construction are not covered by pensions and health insurance. Virtually all Kentucky construction workers working under collective bargaining receive employer contributions into a pension plan. The average contribution is $1,240. If we assume that the merit shop contractor who is paying pension contributions pays the same amount per worker as the union shop contractor, then we can calculate how many merit shop workers are being covered by pensions. On average, the merit shop contractor in Kentucky pays$51 per worker in pension contributions. If each merit shop worker who actually receives a pension contribution receives $1,240 per year, then 4.1% are getting $1,240 and 95.9% are getting $0 per year. (This breakdown will yield an average of$51). Who are these chosen few—this 41/o? They are the key workers in many open shop 90 companies, and all the workers among the exceptional merit shop contractor that either slip- streams behind the union shop or works the quiet non- segments of construction. A similar calculation can be made for health insurance. Based on Form 5500 data and assuming that the merit shop contractor pays the same health insurance premium for his (or her) worker as the signatory contractor pays for his slightly less than 10% of Kentucky's open shop workers are covered by health insurance. This is undoubtedly an underestimate. Unlike pensions, not all health programs are regulated by ERISA. But even if we assume that there are four times as many norrERISA health insurance programs in the open shop, still half of all merit shop workers would go without health insurance in Kentucky. Do prevailing wage regulations encourage the payment cf benefits? Yes. They do this in two ways. First, prevailing wage laws encourage collective bargaining in construction. And collective bargaining makes the payment of benefits possible even when you are working in the main stream of construction demand. Second, it encourages merit shop contractors who do work on public construction to establish qualified pension and health plans of their own. We can look at this phenomenon through the lens of another data source—the U.S.Census of Construction. Average Employer-Contribute Voluntary Benefits per Worker, 199 $3,500 (in 1099 Dollars) . . _ __ �.� $3,0002,$$Q $2,500 $2.000 $1,500 1,415 $1,000 $500 w $0 Never Had(0 ) Repealed(9 tes) Law(,32 States) Figure 16: Comparing Average Employer Contributions to Health,Pension,& Apprenticeship Benefits by States—Those with Prevailing Wage Laws and Those Without in 1992 Source: 1992 Census of Construction 91 i Figure 16 shows employer contributions to benefits not by union and nonunion contractors but by states with and without prevailing wage laws. So in states with prevailing wage laws, the contributions are an average of union and nonunion contractors in those states. Similarly, in states that never had prevailing wage laws and in states that repealed their prevailing wage laws, Figure 16 shows the average voluntary employer benefit contributions to health insurance, old-age pensions, and apprenticeship training. In states with prevailing wage laws, union and nonunion contractors together set aside on average $2,880 for health, pensions and training. In states that never had prevailing wage laws, union and nonunion contractors together set aside about half that amount,$1,415. In states that repealed their laws, union and nonunion contractors together set aside $1,639 for health, pensions and apprenticeship training. In short, the health of construction workers' families were better taken care of in states with prevailing wage laws. The old age of construction workers and their spouses were better provided for in states with prevailing wage laws. The training of young construction workers was better assured in states with prevailing wage laws. Solid communities need solid health and old age insurance. People who cannot take care of themselves when they are ill or when they are old become burdens on their families and burdens on the community. We saw in Chapter 3 that the alleged cost savings from prevailing wage repeal do not exist, or, at best are insubstantial. In this chapter we find that the policies that discourage collective bargaining in construction or the payment of prevailing wages on public works pose a real and measurable threat of lost health insurance and a less secure old age for many of Kentucky's citizens. Prevailing wage laws encourage nonunion contractors to pay benefits when doing public construction. These merit shop contractors can chose to put legally mandated fringe benefit contributions in the worker's pocket But they can, and some do put those contributions into pension and health plans. Prevailing wage laws also encourage collective bargaining in construction. Collective bargaining, in tum, ensures the payment of health and pension benefits not simply to a handful of key construction workers who move with the contractor from job to job,but also to the majority of construction J J � J h' workers who move from contractor to contractor. Collective bargaining, in short, privatizes the cost of health and old age. Without prevailing wage regulations and collective bargaining in construction, Kentucky risks pushing onto the taxpayer the health and retirement costs of caring for Kentucky construction workers and their families. 92 Kansas and Pruvailing V"Wav- 1110,1MM islation . By Peter Philips, Ph.D. Professor y Economics, University¥Utah Prepared for the Kansas Senate Labor and Industries Committee, Fbu92, 198 \« � \ / �\ « � . . . . . a . . Table of Contents List of Tables and Figures About the Author Executive Summary Chapter One: History of Prevailing Wage Regulations in Kansas and the U.S. Chapter Two: A Case Study of the Effect of Prevailing Wage Repeal on Costs: the Cost of School Construction in Kansas and Surrounding Great Plains States Chapter Three: Loss of Construction Worker Income Due to Repeal Chapter Four: Loss of Apprenticeship Training in Kansas Due to Repeal Chapter Five: Increase in Job-Site Injuries in Kansas Due to Repeal Chapter Six: Loss of Pension and Health Insurance Coverage in Kansas Due to Repeal Chapter Seven: Summary and Conclusion 1 List of Maps, Tables and Figures Map: Legal Status of Prevailing Wage Law in Kansas and 14 Great Plains Comparison States Table 1: Square Foot Construction Costs of Public and Private Elementary Schools by 15 States and Legal Status Table 2: A Comparison of Average Square Foot Cost of A New Elementary School in Great Plains States with and without a Prevailing Wage Law Table 3: Square Foot Construction Costs of Public and Private Middle Schools by 15 States and Legal Status Table 4: A Comparison of Average Square Foot Cost of A New Middle School in Great Plains States with and without a Prevailing Wage Law Table 5: Square Foot Construction Costs of Public and Private High Schools by 15 States and Legal Status Table 6: A Comparison of Average Square Foot Cost of A New High School in Great Plains States with and without a Prevailing Wage Law Table 7: Wages as a Percent of School Construction Cost and Average Wage Rate in School Construction by Region of the U.S. Table 8: Distribution of Apprenticeship Training Under Collective Bargaining and in the Merit Shop, in Kansas, Total and by Craft Table 9: Number of Construction Apprentices by State and Year for Kansas and 14 Great Plains Comparison States by Legal Status of Prevailing Wage Law Table 10: Number of Minority Construction Apprentices by State and Year for Kansas and 14 Great Plains Comparison States by Legal Status of Prevailing Wage Law 2 Table 11: Number of Female Construction Apprentices by State and Year for Kansas and 14 Great Plains Comparison States by Legal Status of Prevailing Wage Law Table 12:Average Annual Injury Rates in Construction in Kansas Before and After the Repeal of Kansas' Prevailing Wage Law Table 13: Average Annual Injury Rates in Construction in Kansas and 14 Comparison Great Plains by Legal Status of Prevailing Wage Law Table 14: Annual Average Per Worker Employer Contributions to Pensions and Health Insurance in Kansas and the Percent of Workers Insured by Merit Shop Contractors Figure 1: Labor Costs Including Benefits as a Percent of Total Costs in Kansas, 1977 to 1992 Figure 2: Annual Wage Income of Construction Workers in Five Great Plains States without Prevailing Wage Laws Compared to Ten Great Plains States with Prevailing Wage Laws Figure 3: Wage Costs as a Percent of Total Costs in Five Great Plains States without Prevailing Wage Laws Compared to Ten Great Plains States with Prevailing Wage Laws Figure 4: Average Wage Income of Kansas Construction Workers Compared to the Wages of Construction Workers in Four Great Plains States with No Prevailing Wage Law and Ten Great Plains States with Prevailing Wage Laws Before and After the Kansas Repeal Figure 5: Annual Number of Injuries per Construction Worker in Kansas Before and After the Repeal of Kansas' Prevailing Wage Law Figure 6: Annual Number of Serious Injuries per Construction Worker in Kansas Before and After the Repeal of Kansas' Prevailing Wage Law Figure 7: Total Annual Average Employer Contribution to Pension and Health Insurance in Kansas Prior to and After Repeal of Kansas' Prevailing Wage Law 3 Abou t the Author Peter Philips grew up in Compton and Pomona, California. He received his B.A. from Pomona College in 1970 where received the Leland Backstrand Graduating Senior Award in Economics. Philips received his MA. (1976) and his Ph.D. (1980) from Stanford University. Philips is a Professor of Economics at the University of Utah. He is co-editor of Three Worlds of Labor Economics(M.E. Sharpe, 1986) and co-author of Portable Pensions for Casual Labor Markets (Quorum Books, 1995). Philips has published widely on the canning and construction industries in journals such as Industrial and Labor Relations Review, Industrial Relations, Business History, the Journal of Economic History, The Journal of Economic Literature and the Cambridge Journal of Economics. Philips has been a consultant for the U.S. Labor Department analyzing the supply of cannery labor in California, and he has worked as an expert on the Davis-Bacon Act for the U.S. Justice Department. The Davis-Bacon Act regulates wage payments to construction workers on federal public works. Philips is a respected expert on prevailing wage laws and on employment, training wages and benefits in the construction industry. He has testified before state legislative committees in Ohio, Indiana, Oklahoma, New Mexico and California on their state prevailing wage laws. Along with other researchers at the University of Utah, Philips has analyzed the effects of prevailing wage laws on public construction costs, construction worker incomes, apprenticeship training, worker safety and minority access to construction work. Philips has received awards for his teaching and community service, including the University of Utah Lowell Bennion Public Service Professorship, the University of Utah Presidential Teaching Scholar Award and the University of Utah, College of Social and Behavior Science Superior Teacher Award. Philips is married with two children. 4 12 0 Executive Summary n p } Legal Status of Prevailing Wage Law Repealed 1985&87 (2) Judicially Annulled 1995(1) Never Had Law (3) ❑Has Law (9) A 15 Great Plains state comparison shows that after Kansas repealed its prevailing wage law in 1987 • Wage incomes in Kansas construction fell by 10% not just on public works but across all construction. • Employer pension and health insurance contributions fell by 17%. • While almost all construction workers covered by collective bargaining in Kansas receive health insurance and employer pension contributions, only 10% of the workers in 5 the open (or merit) shop receive pension coverage and only 4% receive health insurance from their employer. • Apprenticeship training in Kansas construction fell by 38% after repeal. Minority apprenticeship training in Kansas fell by 54%. • This was due to a shift away from collective bargaining towards open shop (or merit) shop construction. Merit shop contractors account for only 12% of all apprentices being trained in Kansas. As the merit shop share of the market grew after repeal, apprenticeship training fell substantially. • With lower wages and benefits and less training, a new,. younger, less-skilled, less-experienced work force entered Kansas construction. Serious-injury rates in Kansas construction rose by 21 % after repeal of the state prevailing wage law. • pain of repeal is real and measurable the While the , p p projected gain from repeal--a 6% to 17% savings on state construction costs--failed to materialize. • Elementary school, middle school and high school new construction costs are virtually identical between Great Plains states with and without prevailing wage laws. s -- -- was Overview. Kansas prevailing wage law the first the country --was in 1891 to help prod the Kansas labor market in general and the construction labor market in particular down a high-skilled, high-wage growth path. Confronted with falling wage rates and longer working days, the Republican government of Kansas embraced a series of reforms including child labor laws, compulsory schooling, convict labor laws, the eight-hour day and prevailing wages. All of these reforms were aimed at the same goal. The Kansas labor market was to be regulated so that young people were in school, apprenticeships would be encouraged, the working day would be limited, and competition would be built upon a system of skill-formation that generated and justified rising wages and incomes. Kansas legislators did not want businesses to prove profitable simply because people were working longer for less, and younger with less skills. Almost 100 years after its original passage,Kansas' prevailing wage law 0 Kansas taxpayers was repealed on the promise that would save from 6/o to 17% on total construction costs depending on the project---and in some cases the savings would be even higher. To obtain these gains, workers wages on public works would have to be cut. If there were Epill-over effects on wages outside public works, that would be the additional cost of saving money on public construction. The immediate effect of the repeal of Kansas' prevailing wage law was that construction wages were cut--not only on public construction—but across the entire Kansas construction labor market. Adjusted for inflation, Kansas construction workers wage incomes fell by 11% from 1987, the year of the repeal to 1991. This amounted to a drop in average wages from $25,573 to $22,807. In the nine Great Plains states surrounding Kansas that retained their prevailing wage laws, wage income fell--but only by 2%. So the predicted pain of prevailing wage repeal had been achieved. Was there a corresponding gain for that pain? Were state construction costs cut by from 6% to 17% or even higher? A case-study comparison of new school construction costs in Kansas compared to surrounding Great Plains states that have retained their prevailing wage laws finds no difference in square foot construction costs. The average square foot construction cost of building 365 elementary schools in nine Great Plains states with prevailing wage laws was$76.86. The average square foot construction costs of building 81 new elementary schools in six Great Plains states--including Kansas--that do not have prevailing wage laws was$76.23. Comparison of the square foot costs of middle schools and high schools yielded similar results. There is no statistically significant difference in school construction costs between comparable states with and without prevailing wage laws. 7 Why could wages be cut substantially and yet, no construction savings were forthcoming? The answer is—training and productivity fell with wage rates. Apprenticeship training in Kansas fell by 38% after the state repealed its prevailing wage law. Minority apprentices fell even more by 56% after the repeal of the Kansas law. The balance of construction shifted away from collective bargaining towards the open shop. Currently, open shop contractors account for only 12% of all enrolled apprentices in Kansas. Thus, as the unions declined, the open shop did not take up the slack in apprenticeship training. Rather, in the short-run, merit shop contractors hired union-trained journeymen at substantially lower wage rates and markedly reduced pension and health programs. Total employer contributions to pension and health insurance in Kansas fell by 17%after the state repealed its prevailing wage law. This was a drop from an annual average of$20 million per year to $16.6 million. This drop was due to a shift from collective bargaining to the merit shop. Almost all union contractors in Kansas provide pension coverage and health insurance. Currently, only 10% of merit shop workers in Kansas are covered by a company pension and only 4% receive company health insurance. With lower wages and benefits, experienced and skilled workers eventually migrated out of the industry or retired. With a 38% fall-off in apprenticeship training, skilled and experienced older workers were replaced by younger, less-experienced, less trained workers. Thus, the promised construction savings were based on a false premise--that wage rates could be cut without effecting productivity, and collective bargaining could be terminated without effecting training. Both these premises proved false. In place of lower construction costs, Kansas reaped a costly, higher injury rate in construction. Less trained, younger, inexperienced and poorly paid workers got hurt on the job much more dten. In the five years after repeal, serious-injury rates in Kansas construction rose by 21% compared to prior to repeal. A comparison of Great Plains states with prevailing wage laws compared to those like Kansas shows that states without prevailing wage laws have a 26% higher injury rate in construction. The History of Prevailing Wage Regulations in Kansas and the U.S. In February 1891, Samuel Gompers, president of the American Federation of Labor, visited Topeka, Kansas, to speak on what the local newspaper called "the great topic of labor"Ten years earlier, the AFL— at its own creation—had laid out legislative aims that included the eight- hour work day, the elimination of child labor,free public schooling, compulsory schooling laws, the elimination of convict labor, and prevailing wages on public works. These proposals were based on a belief that the American labor market should consist of highly skilled workers earning decent wages,with time for family, and with children free to earn an education. In pursuit of these aims, Gompers' political strategy in Kansas allied him with the Republican Party. On the morning of Gompers'arrival,the Alliance Party, known to history as the Populist Party, withdrew an earlier invitation for him to speak in the hall of the state House of Representatives,which the party controlled. Gompers, who represented 900,000 workers, had fallen out of favor with the Populists, reportedly because of his belief that the trade unions should not form a political party with the Alliance.' Gompers and the AFL took the position that unions should be nonpartisan. Rather than form a labor party, Gompers advocated that unions support those of any party who would support the needs of working men and women. In Kansas in 1891, this made Samuel Gompers an ally of the Republican Party. The Republicans, who con trolled the Kansas Senate, invited Gompers to speak there, and he did. Gompers was in Kansas to focus on the eight-hour day. Like other Americans, Kansans in 1891 typically worked six days per week, ten to twelve hours per Y day. In the older trades and crafts, such as carriage 1 Topeka State Journal,February 24, 1891,col.4,p.4. 9 making and saddle making, where the work pace was slow and under the workers'direction, the long work-day was tolerable. In the newer factories producing shoes, textiles, and the like; in the mines; and in the urban putting-out systems in needlework, six-day weeks and twelve-hour days . objective a shorten ed work- were grueling. The AFL had made its prime day and work week with as little cut in pay as possible. In his Topeka speech, Gompers declared: Our banner floats high to the breeze and on that banner float is inscribed, "Eight hours work, eight hours rest and eight hours for mental and moral improvement.,2 At that time, when there were no income supplement programs for the poor, low-income parents worked and had to send their children to work to make ends meet. This practice was later referred to by a North Carolina newspaper editor as "eating the seed corn." Each generation of poor condemned its offspring to poverty because the children grew up as illiterate as their parents. The prevalence of cheap child labor, which accounted for 5 percent of the manufacturing labor force in 1890 and a larger proportion of service sector workers, kept wages down and forced adult workers to put in the long hours to make ends meet. Gompers wanted regulation to force employers and the poor to adopt a strategy, however painful in the short run, of a high-wage, high-skilled growth path where children were in school and workers had the skills to justify wages that would allow for a family life. Gompers said, The Federation endorses the total abolition of child labor under 14 years of age; an eight hour law for all laborers and mechanics employed by the government directly through contractors engaged on public work, and its rigid enforcement; protection of life and limb of workmen employed in factories, shops and mines; ...the extension of suffrage as well as equal work for equal pay to women....The Federation favors measures, not parties.3 Gompers also pleaded for workers to be paid the "current" daily wage so they could afford the reduced work time. Government was being asked to set a good example for the private sector, to show that a refreshed labor force could produce in eight hours what a fatigued and bedraggled labor force turned out in ten or twelve hours.The prevailing wage law in its 2. Topeka Daily Capital,February 25, 1891,p.l. 3 . Topeka State Journal,February 25, 1891,col.34,p.1. 10 infancy was an attempt to obtain shorter working hours for all labor. The AFL paid attention to public works, however, because government at all levels was a major purchaser of construction. The AFL said government should not try to save money by eroding the wages of its citizens. With similar logic,the AFL called for an end to convict labor. Many states employed convicts to pay for their keep. Convicts built roads on chain gangs, operated farms, made textiles, and sewed garments. Convict- made goods were sold,forcing down prices and the wages of working free citizens. In February 1891,the Second Annual Convention of the Kansas State Federation of Labor, in Topeka, approved a bill concerning state-paid wages.That month, the bill,which included the prevailing wage section, called "for an Eight Hour Law"and was brought forth by Mr.Avery of the Typographical Union No.121, Topeka.The bill stated, That in no case shall any officer,board,or commission,doing or performing any service or furnishing any supplies to the State of Kansas under the provisions of the act be allowed to reduce the daily wages paid to employees engaged with him (or them) in performing such service or furnishing such supplies, on account of the reduction of hours provided for in the act.That in all cases such daily wages shall remain at the minimum rate which was in such cases paid and received prior to the passage of the act.4 The eight-hour bill was one of four labor-related bills pending in the legislature: the weekly pay bill,the child-labor bill, and the bill to make the first Monday in September a holiday,which would become known as Labor Day. In addition, that year the Kansas State Federation of Labor approved a resolution calling "for the abolition of convict labor when in competition with free labor."5 The eight-hour bill, Senate Bill 151,failed in the Kansas senate March 6, 1891, with the prevailing wage section removed. But by March 10, when the prevailing wage section was put back in, the bill became law. This first prevailing wage law stated: That not less than the current rate of per diem wages in the locality where the work is performed shall be paid to laborers, 4 .Sixth Annual,215. 5 Sixth Annual, 124. 11 mechanics and other persons so employed b or on workmen, mec pY behalf of the state of Kansas....s We do not know the immediate impact of the Kansas prevailing wage law. But a report from the Oklahoma labor commissioner in 1910 may well have applied to Kansas. The Oklahoma law which was patterned after the Kansas act. It was passed in 1908. It was reported to have had the intended effect of setting wage and hour standards not only on public works but in related labor markets. The Oklahoma Commissioner of Labor stated in 1910: The eight hour law has been of inestimable value to the laboring men of this state....The common laborer, who was heretofore employed ten and twelve hours per day, is now, under the provisions of this bill, allowed to work but eight hours....The law has not only affected the laborers and those who are dependent upon this class of work for a living, but it has gone further, and in many localities has gradually force railroad companies, private contractors [i.e. private construction] and people of that class to pay a high rate of wages for unskilled labor. Some argued have that the historic reason prevailing wage laws were people passed was to exclude African Americans from construction job sites. Prevailing wage laws have been described by some as Jim Crow laws. This is a difficult case to make for Kansas. The Kansas law was examined by the U.S. Supreme Court.in Ashby v. Kansas. The Supr eme Court Justice who wrote the deciding opinion upholding the constitutionality of the Kansas prevailing wage law was Justice John Marshall Harlan. Harlan wrote: When the eight hour law was passed the legislature had under consideration the general subject of the length of a day's labor, 6 L. 1891 Ch. 114 p.192-193. z Chas. L Daugherty, Labor Commissioner, Oklahoma Department of Labor, Third Annual Report, Oklahoma City, OK, 1910,p. 327. The primary concern in both Kansas and Oklahoma was to use public works hours and wage policies to set and improve local labor standards. A typical enforcement case in Oklahoma as reported by the Labor Commissioner follows: [Anadarko. May 10. 19081 We were advised that the O'Neill Construction Company had cut the wages on public works at Anadarko from twenty-five cents to seventeen and one-half cents per hour....[C]ontract was taken with the understanding that twenty-five cents per hour should be paid. The work was not progressing as rapidly as necessary to the cost within the estimate,hence the contractors tried to take advantage of the situation by reducing pay. After thoroughly discussing the matter before the[city]council and contractor,the wages were restored to twenty-five cents. (p.320) Second Annual Report Oklahoma Labor Commissioner Chas.L Daugherty,Oklahoma City,OK,August 7, 1909. 12 without specific reference to the purpose or occasion of their employment. The leading idea clearly was to limit the hours of toil of laborers, workmen, mechanics and other persons in like employment to eight hours, without reduction in compensation for the day's service a John Marshall Harlan,Supreme Court Justice Harlan's opinion about the purpose of Kansas' law is especially interesting telling in light of the largely unsupported proposition that these laws were Jim Crow laws. Justice Harlan is known to history as the single Supreme Court Justice who spoke out against Jim Crow. In his famous dissent against the separate but equal doctrine that legitimized racial segregation in the case of? in 189?, Harlan argue vigorously for equal treatment of the races. if the Kansas law had been a Jim Crow law in intent or effect, Justice Harlan would have been the first to declare it so and argue against its existence. Those who have argued that prevailing wage laws are Jim Crow laws typically point to one incident associated with the passage of the federal prevailing wage law in 1931, the Davis Bacon Act. Republican Representative Robert Bacon complained of an Alabama contractor who came to his New York district in 1926 to build a federal veterans hospital. Rep. Bacon complained that the Alabama contractor was undercutting local wages and hours of work by importing cheaper southern labor. Critics of the Davis-Bacon Act have assumed that Rep. Bacon was aiming his complaint at black labor. But in fact Rep. Bacon had indicated that the Alabama contractor had brought up a mixed crew of both black and white workers. Indeed, at the time,two-thirds of all Alabama construction workers were white. While the hod carriers and laborers were likely to have been blacks from Alabama,the brick masons and carpenters were likely to have been white. The notion that Rep. Bacon was aiming his legislation as a Jim Crow attack on southern blacks is thinly supported speculation. Republican Representative Fiorelo LaGuardia was familiar with this particular Alabama contractor. He mentioned this issue as he argued for the passage of the Davis Bacon Act in 1931. He argued on the floor of the House: A contractor from Alabama was awarded the contract for the Northport Hospital, a Veterans' Bureau hospital. I saw with my own eyes the labor that he imported there from the South and the conditions under which they were working. These unfortunate men were huddled in shacks living under most wretched conditions and being paid wages far below the standard. These unfortunate men were being exploited s Quoted in: Oklahoma,Department of Labor, Second Annual Report,Oklahoma City,OK, 1909,p.327. 13 by the contractor. Local skilled and unskilled labor were not employed. The workmanship of the cheap imported labor was of course very inferior....all that this bill does,gentlemen, is to protect the Government, as well as the workers, in carrying out the policy of paying decent American wages to workers on Government contracts. [Applause.]9 Prevailing wage laws were Republican legislation. The Davis Bacon Act was named after a Republican representative from New York and a Republican Senator from Pennsylvania. The Davis Bacon Act was signed by Republican President Herbert Hoover. The rationale for prevailing wage laws is rooted in a philosophy of economic growth. Prevailing wage laws support higher wage rates and greater unionization in construction. The absence of prevailing wage laws permits the spread of lower wage rates and the growth of nonunion construction. As will be seen in later chapters of this report,states with and without prevailing wage laws have very different construction industries. The ones with prevailing wage laws have more apprenticeship training taking place, their workplace is safer, more construction workers have pensions and health insurance and construction workers are more productive and earn higher incomes. Despite these advantages associated with prevailing wage policies,beginning in 1979,there was a widespread effort to repeal existing prevailing wage laws. Between 1979 and 1988, nine states repealed their state prevailing wage laws. In 1995, the Oklahoma law was judicially overturned based on the notion that the state's prevailing wage survey was unconstitutionally over- reliant on the federal survey. The major reason state laws were repealed is that proponents of repeal promised substantial savings on public construction costs. As the next chapter demonstrates, there is no evidence that Kansas has saved a significant amount of money because it repealed its state prevailing wage law. if little was gained by repealing Kansas' law, it is time to consider what was lost. That topic will be taken up in subsequent chapters. 9 U.S.,Seventy-First Congress,Third.Session,Congressional Record-House,Febn mry28,1931,p. 6510. 14 0 The Cost of School Construction in Kansas and Surrounding Great Plains States A Case Study of the Effect of Prevailing Wage Repeal on State Construction Costs by Looking at School Construction in States with and without Prevailing Wage Laws The effect of inflated wages and deflated productivity combines to a net increase in cost to the Kansas Tax Payer for state construction of from 6% to 17% depending on the project and in some cases higher. Carl Conrod, Associated Builders and Contractors Testimony before the Senate Labor and Industry Committee on Bill 112, February 16, 198710 Kansas taxpayers were promised a 6% to 17% reduction in their state construction costs with the repeal of the state prevailing wage law. Sometimes the saving would be higher. This chapter looks for those savings by looking at the cost of school construction—broken down separately into new elementary schools, new middle schools and new high schools--built in Kansas and surrounding states, from July 1991 to June 1997. If Kansas taxpayers have saved 17%or even more on school construction costs, then the cost of building schools in Kansas should be substantially cheaper than the cost of building schools in surrounding states that have retained their prevailing wage laws. Even if Kansas has saved only 6% on its school construction costs,with enough observations this sort of savings should be clear. Kansas is surrounded by a set of states, some of which have prevailing wage laws governing school construction and some of which do not have these ns B comparing the square foot cost of new school construction in regulations. y p g q 9 these differing states, we can estimate the effect of prevailing wage laws on public construction costs. In this chapter we examine separately the mean and median square foot cost of building new public schools. Schools are broken down into three types— elementary schools, middle schools and high schools.. Square foot costs are the total cost of construction excluding land acquisition, architect fees or construction management fees divided by the total square feet of the project. ")George Barbee,Executive Director of Kansas Consulting Engineers in testimony the samedayonSB 112 characterized the general estimate at the hearings of cost savings from repeal as being 20%. 15 The data are from the start of construction as reported by the F.W. Dodge Corporation, the standard bid reporting service for the construction industry. The time period of the analysis is July 1991 to June 1997.11 Earlier construction costs are brought to constant 1997 dollars by using the consumer price index for housing costs. The map below shows the 15 states in the cost comparison. Nine states have state prevailing wage laws, five including Kansas do not. One state— Oklahoma—switched from having a law to not having a law during the time period of the comparison. t� y Legal Status of Prevailing Wage Law Repealed 1985&87 (2) Judicially Annulled 1995(1) Never Had Law (3) ❑ HasLaw (9) Table 1 presents both the mean and the median square foot cost of construction for elementary schools. The mean is the numerical average "F.W.Dodge Corp."Dodge Reports"Start Cost for New Construction. (Earlier data are not available.) 16 while the median is the midpoint cost between the cheapest and most expensive elementary school built in that state. 12 Table 1:Square Foot Construction Costs of Public and Private Elementary Schools by State and Legal Status 1991-97(in constant 1997 dollars) ELEMENTARY SCHOOLS Ne Law Rtafa PW I-W..+tat- gtate Public or Private Moan Median Number of Seho_ols __ Mean Median Ntimber of Schools AR Public Owner $53 $52 N=17 CO Public Owner $82 $85 N=40 IA Public Owner $72 $70 N=8 KS Public Owner $83 $75 N=18 24 MN Public Owner $66 $76 N=N=5 MO Private Schooi(s) $ $ Public Owner $68 $70 N=30 ND Public Owner $56 $56 N=2 • NE Private School(s) $ $34 N=1 Public Owner $80 $84 N=11 NM Public Owner $83 $82 N=27 OK Public Owner $49 $48 N=8 $55 $55 N=14 SD Private School(s) $109 $109 N=1 Public Owner $67 $66 N=5 TX No Public Owner $81 $71 =19 Public Owner $81 $71 N=2 WI Private School(s) $71 $6 N=2 Public Owner $83 $83 N= 2 WY Public Owner $83 $83 N=2 2 Along with public elementary schools,a handful of private elementary schools were built in some of the states. While the small number of private schools makes statistical cost comparisons difficult,these are nonetheless interesting observations simply because private construction is not governed bypmvalirg wage Lim.There are some data in Table 1 that a critic of prevailing wage laws might take as evidence that these laws raise construction costs. Missouri has a prevailing wage law governing public school construction. Private elementary schools in Missouri on average,cost slightly less per square foot than public schools. Perhaps this is due to Missouri's prevailing wage regulation. But when one looks at the median square foot cost,private schools are more expensive to build in Missouri. The average or mean is more sensitive to outliers—anexhNnelyexpensiveor extremely cheap new school. Setting aside the effect of one or two exceptions,the median actually suggests that private construction is more expensive than prevailing wage construction of elementary schools in Missouri. However,one should not rush to this conclusion because the number of private schools built(5)is small. In Texas with a similar number of private schools(4),both the median and mean square foot construction cost for private elementary schools is higher than the public schools built under prevailing wage regulations. In Nebraska,the one private elementary school built since 1991 was built quite cheaply at$33 per square foot. On the whole,a comparison of private and public elementary school construction yields ambiguous results. Sometimes private elementary schools are more expensive. Sometimes they are less. This is precisely the result you would expect if prevailing wage laws had little effect on construction costs. 17 Table 2 shows the mean or average square foot construction cost of elementary schools broken down into those built in states with prevailing wage laws and those states without such regulations. The Table shows that the average square foot cost for 365 new elementary schools built in states with prevailing wage laws is $76.86. The average square foot construction cost of elementary schools in states that do not apply prevailing wage regulations is $76.23. Applying a standard statistical test comparing the values of sample means, we can say emphatically that there is no statistically significant difference between these two numbers.'3 The cost of elementary school construction is basically the same whether or not the state applies prevailing wage regulations. Table 2:A Comparison of the Average or Mean Square Foot Cost of Building a New Elementary School in States with and without Prevailinq Wage Laws Legal Std. Std.Error Status N Mean Deviation Mean Square Foot Cost in1997 No Law 81 $76.2309 $21.3523 $2.3725 Dollars Using CPI-Housing State Deflator PW Law 365 $76.8644 $54.5442 $2.8550 State Table 1 has one more item that a critic of prevailing wage laws might seize upon to demonstrate how costly these laws are. Oklahoma's law was overturned by judicial decision in 1995. Square foot construction costs of elementary schools in Oklahoma were lower after the law was eliminated. Surely this is evidence of the laws costly impact. If such an analysis were true, then one would expect that the cost of new middle school construction in Oklahoma would decline after the law was eliminated. Table 3 repeats the calculations in Table 1 for middle schools. In Oklahoma, average middle school construction costs rose after the termination of the state prevailing wage law. 13 Formal results of tests for statistical significance are presented in the Appendix to this chapter. 18 Table 3: Square Foot Cost of New Public and Private Middle School Construction by State and Legal Status MIDDLE SCHOOLS State Public or Private No Law State PW Law State Mean Median Number of Schools Mean Median Number of Schools AR Private School(s) $45 $45 N=2 Publi c Owner $47 $46 N=13 CO Public Owner $84 $82 N=10 IA Public Owner $67 $67 N=3 KS Public Owner $69 $68 N=12 MN Private School(s) $126 $126 N=1 Public Owner $80 $80 N=14 MO Public Owner $75 $69 N=26 MT Public Owner $59 $59 N=4 ND Public Owner $65 $65 N=2 NE Public Owner $71 $71 N NM Public Owner $90 $90 N=9 OK Public Owner $54 $54 N=2 $51 $49 N=11 SD Private School(s) $70 $70 N=1 Public Owner $72 $72 N=1 TX Private School(s) $70 $ N=3 $70 68 N=131 Public Owner $70 N=1 WI Private School(s) $7 $6 3 N=20 Public Owner $ $ WY Public Owner $6655 $6655 N=2 After the elimination of Oklahoma's law, average (mean) square foot costs of middle school construction (controlling for inflation) rose 6%. But given the number of school s built (2) this re sult has little statistical significance. An examination of public versus private construction yields few result also. In Arkansas, a state with a prevailing wage law, private middle school come in about $2 cheaper per square foot. But in South Dakota, a state without a prevailing wage law, private middle schools also come in about $2 cheaper per square foot. In Wyoming, a law state, one private middle school was built slightly below the public average. But in Minnesota, another law state, the fiddle school built during one private m 9 the period was substantially more expensive. Generally, the number of private schools is small and comparison of averages is consequently statistically unreliable. A comparison of average square foot costs for new public middle schools broken down by states with and without prevailing wage laws yields the same result as with elementary schools. There is no statistically significant e cost of th e two groups of schools. In states with avers c p difference in the 9 . 9 no state prevailing wage law, the average square foot new construction cost was $72.35. In states with prevailing wage laws, the average was $70.02. The lower cost of construction in states with prevailing wage laws is not statistically significant. t4 See Appendix to this chapter for formal test results. 19 Table 4:Average(Mean)Square Foot Construction Costs of Middle Schools by States with and without Prevailing Wage Laws Group Statistics State Std. Std.Error With PW N Mean Deviation Mean Square Foot Cost in1997 No Law 30 $72.3547 $19.7813 $3.6116 Dollars Using State CPI-Housing Deflator PW Law 238 $70.025 $ $2 23.7157 1.5373 State square foot construction costs broken down by public and private high schools and then broken down by states with and without prevailing wage laws. Private high schools in Minnesota and Texas (both states with laws) were cheaper to build than public high schools. But in Kansas this was also true even though Kansas does not have a prevailing wage law. In Wisconsin there was little difference in the cost of building a high school privately or publicly even though the public school were built under prevailing wage regulations. Table 5: Square Foot Cost of New Private and Public High School Construction by State and Legal Status HIGH SCHOOLS States No Law State PW Law State Mean Median Number of Schools Mean Median Number of Schools AR Public Owner $60 $55 N=13 CO Public Owner $81 $82 N=12 IA Public Owner $70 $70 N=6 KS Private School(s) $24 $24 N=1 Public Owner $66 $69 N=9 MN Private School(s) $64 $64 N=1 Public Owner $81 $83 N=23 MO Public Owner $62 $63 N=20 MT Public Owner $65 $68 N=3 ND Public Owner $102 $102 N=1 NE Public Owner $83 $88 N=3 NM Public Owner $97 $96 N=5 OK Public Owner $53 $50 N=5 $53 $53 N=4 SD Public Owner $62 $62 N=2 TX Private School(s) $65 $59 N=7 Public Owner $76 $71 N=86 WI Private School(s) $69 $69 N=2 Public Owner $69 $70 N=25 WY Public Owner $65 $57 N=5 20 When we compare average public high school square foot construction costs by states with and without prevailing wage laws (Table 6), the results are similar to what we found for elementary and middle schools. High schools in states with prevailing wage laws cost, on average, $72.87 per square foot while high schools in states without prevailing wage laws cost $70.72 per square foot. This$2 difference was not statistically significant.15 Once again, there is no measurable difference among these 15 states in school prevailing presence or absence of reva g construction costs associated with the p P wage laws. Table 6: Average Square Foot Construction Costs of New High Schools by Great Plains States with and without Previaling Wage Laws Group statistics State Std. Std.Error With PW N Mean Deviation Mean Square Foot Cost in1997 No Law 35 $70.7255 $20.7515 $3.5076 Dollars Using State CPI-Housing Deflator PW Law 187 $72.8742 $37.7920 $2.7636 State How can this be when wage rates on prevailing wage projects are usually substantially higher than the wage rates on private jobs done by nonunion contractors? Wage Rates and Labor Costs as a Percent of Total Costs When Kansas repealed its prevailing wage law, Kansans were promised anywhere from a 6% to a 17% savings on public construction costs. How were such estimates calculated? The answer is the estimates were hypothetical calculations. The calculation typically went like this. • Assume that labor costs are 50% of total construction costs. • Assume wage rates fall by 12% to 40% with the repeal of Kansas' prevailing wage law. 15 See Appendix to this chapter for formal test results. 21 • Assume labor productivity does not fall when wage rates fall by 12%to 40%. • With these three assumptions in hand, the hypothetical calculation is simple. If 50%of total costs fall by 40%,then 100% of total costs will fall by 20%. If 50% of total costs fall by 12%, then total costs fall by 6%. There you have it.A savings of 6% to 20% on total construction costs. Kansans can now build five schools for the cost of four(a 20% savings) by repealing the state's prevailing wage law. The only problem with this hypothetical calculation is that all its assumptions are wrong. Labor costs are not 50% of total costs. They are around 30% in building construction and less on street and highway construction. Furthermore, labor productivity is not constant when wage rates fall. Skilled and experienced workers leave for better jobs elsewhere. Training falls off. Consequently, productivity falls--offsetting in part, or in full, the fall in wage rates. The key source on information for the construction industry is the U.S. Census of Construction. This Census comes out every five years. The results for 1997 are not yet released. In 1992, for all construction in Kansas, labor costs--wages, benefits, payroll taxes of construction workers--as a percent of total construction costs were 25%. Total cost here does not include land acquisition, architect fees or construction management fees. It also adjusts for possible over-counting by netting out of each contractor's value of construction the cost charged to that contractor by subcontractors. So total cost is the net value of construction built by each contractor and subcontractor. Figure 1 shows for Kansas, labor costs as a percent of total costs for each census year, 1977 to 1992. Kansas repeated its state All Construction 40% 0 U m p 30% 5 f- 0 4 5 5 a � 20% 51 7 in �.n Jaene0ls as a%of 0 10% '- � 3" To al Costs U age Cost as a%of % - L Total Cost 77 82 87 92 Year Source:1992 U.S.Census of Construction Benefits overstated by including office 8.other non-construction workers Figure 1: Labor Costs as a Percent of Total Costs in Ail Kansas Construction, 1977-1992 22 prevailing wage law in 1987. Labor costs as a percent of total costs subsequently fell. But this cannot be laid at the feet of the law's repeal. Labor costs have been failing at least from 1977 onward at a fairly steady rate. This decline has more to do with increased labor productivity and the use of prefabricated material in construction than it has to do with repealing prevailing wage regulations. The Census of Construction does not break out school construction contractors as a separate category. However, a U.S. Department of Labor study has done this. In 1979,the U.S. Bureau of Labor Statistics published a study of school construction costs by region in the United States. The BLS study aggregated school types and presented data on four regions, Northeast, Midwest, South and West. The relevant data for our purposes is presented below. Table 7: Wage Costs as a Percent of Total Costs in School Construction by Regions of the U.S. Elementary • Secondary School • • 1972 Hourly Wa2e Rate wages as a Percent of Total Cost Northeast $7.75 27.9% North Central $7.4 29.3% South $5.22 27.3% West 7:22 29.0% Source: U.S. Bureau of Labor Statistics, John G. Olsen, "Labor and Material Requirements for New School Construction,"Monthly Labor Review,April 1979,Vol.102, Number 4, p.41. These are old data but their age make them more instructive. In 1972, prevailing wage laws were widely enforced on school construction outside the South.16 If prevailing wage laws bloat relative labor costs now, they should have bloated those costs then. But, in fact wage costs as a percent of total costs were 27.9% in the Northeast Compared to 27.3% in the South. Table 7 shows that the U.S. Bureau of Labor Statistics found that in school construction, hourly wage rates varied considerably. For instance, hourly wage rates were 50%higher in the Northeast region compared to the South in 1972 ($7.75 versus $5.22 in 1972). In contrast,wage costs as a percent of total costs were almost the same in the two regions (27.9% versus 27.3%). The analyst, John Olsen, commented on these facts as follows: 16 The only non-southern states without prevailing wage laws in 1972 were North and South Dakota,Iowa and Vermont. Virginia,North Carolina,South Carolina,Georgia and Mississippi also did not have prevailing wage laws in 1972. 23 Average hourly earnings also varied by region. Hourly earnings for all construction workers averaged$6.78, ranging from $5.22 in the South to$7.75 in the Northeast. Wages as a percent of contract costs varied from just above 27 percent in the South to slightly above 29 percent in the North Central. Although average hourly wage rates in the Northeast were higher than those in the North Central region,wage costs as a percent of total contract costs were lower. Among other factors, this irregular trend could result from regional differences in productivity rates and in relative material costs.' Could it be that as wage rates are cut experienced workers leave for better paying jobs elsewhere? Could it be that as wage rates rise, contractors find it worth their while to spend the money to better train their workers and provide them with new, better equipment? Could it be, in other words, that it is wrong to assume that a major wage cut would not effect, whatsoever, in the short run or in the long run, labor productivity in construction? In sum, can wage rates go up without increasing labor costs as a percent of total costs? Annual income of Construction Workers 5 No-Law States (Incl. KS)v. 10 PW States $22,400. $22.20 = 22203 $22,00 $21,80 E $21,60 o $21,40 ` _ m 2I367 =� � Q $21,200 ' No Law Has P Law Legal Status Source:1992 U.S.Census of Construction Figure 2: Average Wages of Construction Workers in 5 States with No Prevailing Wage Law(Including Kansas) Compared to Surrounding States with Prevailing Wage Laws Figure 2 shows that on average, for the ten states around Kansas that do have state prevailing wage laws,the average wage income of construction workers was $22,203 in 1992. In contrast, in the five states including 17 U.S. Bureau of Labor Statistics, John G. Olsen,"Labor and Material Requirements for New School Construction,"Monthly Labor Review,April 1979,Vol.102,Number 4,pp.40- 41. 24 Kansas without state prevailing wage laws, the average construction worker annual wages was $21,367--4% less than in the surrounding states with prevailing wage laws. Did these lower wages result in lower wage costs as a percent of total construction costs? No. Actually, as Figure 3 shows, wage costs as a percent of total construction costs were slightly higher in the lower wage states. Wage Costs as a Percent of Total Costs 5 No-Law States (Incl. KS) v. 10 PW States 19.50 U p 19.40 19 44�MCI � 0 19.30 � (0 19.20 e o ._ a) 19.10 rn ca 19.00 a 18.90 � 18.96 No Law Has PW Law Legal Status Source: 1992 U.S. Census of Construction Figure 3: Wage Costs as a Percent of Total Costs in Five No-Law States (Including Kansas) Compared to 10 Surroundinq States with State So the result found by the U.S. Bureau of Labor Statistics in 1972--that higher wage rates do not necessarily mean higher wage costs as a percent of total costs--still holds true today. Prevailing wage regulations support higher wages but not necessarily higher costs. How can this be? The answer lies in the incentives prevailing wage regulations put in place to encourage training, the retention of skilled workers, and the use of modern equipment. We now turn to the issue of training. 25 Appendix to Chapter 2 Statistical Output from Test Results Comparing Means of Square Foot Construction Costs Elementary Schools Independent Samples Test Levene's Test for Equality of Variances West for Equality of Means 95%Confidence Sig. Mean Std.Error Interval of the Mean F Sig. t df (2-tailed) Difference Difference Lower Upper Square Foot Equal Cost in1997 variances .972 .325 -.103 444 .918 -$.6335 $6.1671 -$12.7538 $11.4868 Dollars Using assumed CPI-Housing Equal Deflator variances 171 328.197 .865 $.6335 $3.7121 $7.9360 $6.6690 not assumed Middles Schools Independent Samples Test Levene's Test for Equality of Variances West for Equality of Means 95%Confidence Sig. Mean Sid.Error Interval of the Mean F Sig. t df (24ailed) Difference Difference Lower I Upper Square Foot Equal Cost in1997 variances .153 .696 .516 266 .606 $2.3322 $4.5178 -$6.5630 $11.2274 Dollars Using assumed CPI-Housing Equal Deflator variances not 594 40.298 .556 $2.3322 $3.9251 $5.5989 $10.2633 assumed High Schools Independent Samples Test Levene's Test for Equality of Variances West for Equality of Means 95%Confidence Sig. Mean Std.Error Interval of the Mean F Sig. t df (2-tailed) Difference Difference Lower Upper Square Foot Equal Cost in1997 variances .179 .672 -.327 220 .744 -$2.1487 $6.5738 -$15.1044 $10.8069 Dollars Using assumed CPI-Housing Equal Deflator variances -.481 83.436 .632 -$2.1487 $4.4656 -$11.0299 $6.7324 not 26 The Loss of construction Worker Income Associated With the Repeal of Prevailing Wage Laws With a Focus on the Effect of Kansas' Repeal High school coaches are fond of advising their players that there is no gain without pain. Such was the philosophy of prevailing wage repeal in Kansas. The gain was alleged to be savings on public construction costs. The pain was that workers would have to endure wage cuts. We saw in Chapter Two that the gain was not there. There are no measurable savings in public construction costs that can be attributed to Kansas' repeal of its prevailing wage law. But while the gain was not real, the pain was. There is one fact upon which all analysts of prevailing wage law repeals agree. These repeals have cut the wages and incomes of construction ur of workers. After all, the precise pose prevailing p p 9 wage law repeals is to cut worker wages--in the hopes that this will save on public construction costs. We saw in Chapter Three that construction cost savings were so minimal that they did not register on standard statistical tests. In fact, we cannot say that there were any savings at all. Is this because construction workers'wages did not decline substantially? No. All analysts agree that construction workers' wages and income have declined due to the elimination of prevailing wage regulations. And the negative effect of repeals have not been limited to the wages of construction workers on public projects. Repeals have lowered construction workers wages across-the-boards in states that have repealed their prevailing wage laws. Before looking at the general effect of prevailing wage repeals on wages, let us examine what happened in Kansas. zi Average Construction Wage by Legal Status, 1986-1991 $27,000 $26.000 $25,000 $24.000 No Law Have Law $23,000 $22,000 $21,000 $20,000 1986 1987 1988 1989 1990 1991 Figure 4: Average Inflation-Adjusted Wage Income of Kansas Construction Workers Compared to Four Surrounding States with No Prevailing Wage Law and Ten Surrounding States with Prevailing Wage Laws, 1986 to 1991. Source: U.S. Bureau of Labor Statistics Figure 4 shows the average wages of Kansas construction workers from 1986,just prior to the repeal of Kansas'state prevailing wage law to 1991. These wages are adjusted for inflation by presenting all years in 1991 dollars. Kansas' wages were slightly higher than the average for ten surrounding states that also had prevailing wage laws at the time.18 The average construction wage in Kansas was substantially higher than the average for four surrounding states that did not have prevailing wage laws in 1986.19 With the repeal of Kansas' state prevailing wage law in 1987, these wage relationships began to change. Over the next five years, the average wages for ten states with prevailing wage laws fluctuated but remained basically the same in inflation-adjusted dollars. In 1987 the 18 These ten states,from north to south,were Montana,Minnesota,Wisconsin,Wyoming,Nebraska,Missouri, Arkansas,Oklahoma,New Mexico and Texas. 19 These four states were North Dakota,South Dakota,Iowa and Colorado. Two cities in Colorado did have city prevailing wage regulations,Denver and Pueblo. 28 waavera9 a construction a in these ten prevailing wage law states was g $25,692. In 1991,these inflation-adjusted wages averaged$25,216. This was a drop in real wages and real consumer power of 2%. Having your wages fall by 2% over five years is no fun. But compared to what happened in the states without prevailing wage laws,a real drop in income of 2% looks good. The real, inflation-adjusted wages of construction workers in the four states surrounding Kansas that did not have prevailing wage laws fell by 11%. That means, adjusting for the cost of living, construction workers in these four states found their annual wages cut, on average, from $24,204 in 1987 to $21,609 in 1991. Having repealed the state prevailing wage law in 1987, Kansas construction workers shared in the fate of surrounding states that did not have prevailing wage laws.Between 1987 and 1991,average construction wages in Kansas--adjusted for inflation--fell from$25,573 to$22,807.This was a drop in real consumer power of 11%. The pain was real. However, other factors may have contributed to the decline in construction worker wages in Kansas after the state repealed its prevailing wage law. Although it is difficult to identify what that might be. General unemployment in Kansas was at 5.4% in 1986 and fell steadily to 4.4% in 1991. In inflation adjusted terms, all Americans'wages were falling during this period--although only by a small percentage, not by 11%. The 2% decline in the real wages of construction workers in surrounding states with state prevailing wage laws reflects the general downward trend in real wages. Would re-establishing prevailing wage regulations in Kansas restore construction worker wages to where they were prior to repeal? Probably not entirely and certainly not right away. The damage of repeal goes deep. Apprenticeship training has fallen substantially. The provision of health insurance and pension contributions has fallen by 25%. The Kansas construction work force needs rebuilding. This needs time. But re-instituting Kansas prevailing wage regulations is part of the solution to moving this industry back towards a high-skill, high-wage growth path. LI Prevailing Wage Regulations and Apprenticeship Training The construction industry is in a training and skills crisis. A January, 1996 report"Gulf Coast Staffing/Retention" commissioned by Brown and Root, Fluor Daniel, and H.B. Zachry--three of the largest nonunion contractors in the country—described the problem. The report wrote: Magnitude of the Problem While the overall availability of construction manpower is declining, the quality of the workforce or crews of highly skilled craftsmen is the real issue....The project execution problems associated with this issue (schedule slippage, work quality, tumover/absenteeism) haven't changed but their magnitude is greater....Failure to address this issue may create an interesting paradox--large contractors shifting more of their work from self-perform to subcontract status and small contractors become even less capable of dealing with the problem due to lack of resources and capital. Driving Forces....Wage erosion has become increasingly worse over the past decade and is causing substandard living conditions. Clients have created a"playing field"which forces contractors to undercut one another to obtain work. Owners do not understand the impact their decisions have on field activities. The accounting/procurement mentality is driving them, thus the industry. Combined with the fact that craftsmen are treated as expendable commodities, woefully inadequate training opportunities over the years, and alternative service sector jobs which are now available at competitive wage rates with superior benefits, it is easy to understand why large numbers of people aren't knocking at the industry's door. Key Issues--Results/Consequences There will be no total system collapse, but the end result of inaction will be a higher cost of doing business. Both clients and customers will pay for the industry's inability/unwillingness to creatively address the problem. The intensity 30 to increase with "high skill" labor shortages will continue9 of regional g craft areas being the worst impacted20 The crisis in the training and retention of skilled construction workers has been a long time in coming. The 1980s were a period of de-regulation, de- unionization training in man pa rts of the in a g Y unionization and the breakdown apprenticeship United States. To understand what has happened, the industry trade magazine—Engineering News Record (ENR)—surveyed the "Top 400 U.S. general contractors and the Top 600 specialty contractors." This is what ENR found: The industry has known for much of the past decade that it was headed for manpower trouble when the business cycle turned up.... Nonunion contractors working in bustling areas appear to have the biggest manpower problems, according to the survey results. For example, 56% of the union crafts in the West reportedly have no labor shortages while only 10% of the open shop crafts have no problem. Only 10% of the union crafts have a severe craft shortage problem while 29% of the nonunion crafts are severely short. "I would guess that some of the labor shortage exists because the open shop has pirated all the available, qualified union workers, and now suffers the lack of training programs of their own to produce open- shop crafts people," says Donald A. McKay, chairman of union mechanical and sheet metal contractor Tougher Industries, Albany, N.Y. "Its frustrating to hear them whine to the owners for help with their educational programs, while spending a pittance on training." McKay notes that the Alliance of Mechanical, Electrical and Sheet Metal Contractors spends about $100 million a year to train union workers in those trades.... Some of the journeymen "pirated" by the open shop may be returning to union construction. "Union contractor backlogs are such that some guys that had been working nonunion are coming back." Says G. Scot Haines,director of business development for union electrical contractor L.E. Meyers, Co, Rolling Meadows, III.... But the battle for the hearts, minds and wallets of skilled worker s knows no bounds of union and nonunion loyalty. In Phoenix, nonunion Haci Mechanical Contractors Inc. reports severe shortages of sheet metal workers and pipefitters. The open shop has an active local training program, but the union sector has been stealing journeymen as soon as they are trained, complains Vice President Tim King. "We pay about$12 an hour and the union pays$18," he notes.... 20"Gulf Coast Staffing/Retention-Cause and Mitigation"by Maxim,Inc.commissioned byBrum and Root Fluor Daniel and H.B.Zachry,January 17, 199& 31 The spreading craft labor shortage problem is underscored by the results of an open-shop survey....Of the 2,437 [open shop contractor] responses, 1,808 or 74% reported shortages in their areas for 14 crafts.21 Why has the industry known for almost a decade that it would face skilled labor shortages once the business cycle picked up? Why would that labor shortage affect nonunion contractors most? Why would nonunion contractors rely upon the ability to hire away union-trained craftsmen? Once the business cycle picked up, why would some union-trained skilled craftsmen return to union shops? And why would nonunion-trained journeymen migrate over to the union sector? Finally, and most important, is this kind of labor shortage good for the industry,for owners and for the community? Training in the construction industry is a classic case of what economists call a market failure. Construction is a boom-bust industry in many respects. Not only does the construction business cycle swing much more widely than does the economy as a whole, but also specific contractors have to gear up and slow down their operations based on their own particular fortunes at winning construction bids. Along with this boom-bust, ramp-up/shut-down structure that is fairly unique to construction, the industry is organized along a complicated structure of sub-contracting. Subcontracting is a way for a contractor to allow a more expert subcontractor to handle a particularly difficult or specialized part of a project. It is also a way to export headaches. When in doubt it is sometimes better to contract out. Labor skill shortages can be just the kind of headache worth contracting out. For example, as the Brown and Root, et al., report quoted above states: Failure to address this issue [i.e. skilled labor shortages], may create an interesting paradox--large contractors shifting more of their work from self-perform to subcontract status and small contractors become even less capable of dealing with the problem due to lack of resources and capital. 22 The boom-bust, ramp-up/ramp-down, subcontract-out headaches structure of construction makes most contractors focus on the short-run. In the short-run, the available supply of trained construction workers is fixed. If you have a shortage, all you can do is bid craftsmen away from someone else. It takes four to five years to turn an electrician, plumbing, fitter or sheet metal apprentice into a skilled journeyman. By the time you train someone for the job, the job is gone. Anyway, if you train someone, you might just be subsidizing your competitor. With the exception of harvest labor in agriculture, there is no major industry with as high a labor turnover rate as in construction. The worker you train in Z'"Craft Shortages Creeping In;'Engineering New Record(ENR),December 25,1995,Vol.235,No.26,pp.34-5. 0 22"Gulf Coast Staffing and Retention"op.cit. 32 all likelihood will be down the road and working for your competitor in the not too distant future. If you undergo training costs and your competitor does not, eat it too. He can win that j ' r can have his cake and then your competitor Job today because he has lower costs today because he does not train. And he has just as much chance as you of having skilled labor tomorrow because skilled labor moves around. You, the honest contractor that diligently tr ains for the future--you're a chump in the cutthroat competition that is the construction industry. Some of the very largest contractors might be able to get around these problems. They may be big enough to always have new jobs on-line when old jobs go away. They might just be able to train internally like in many other g ) industries, have on-going jobs available and have the internal incentives to retain skilled workers. But the smaller contractor cannot follow this strategy except for in the case of a few key workers. Even the largest nonunion companies have difficulty training and retaining skilled workers in the face of industry competitive pressures. Again from the Brown and Root et al. study: Clients have created a "playing-field" which forces contractors to undercut one another to obtain work. Owners do not understand the impact their decisions have on field activities. Combined with the fact that craftsmen are treated as expendable commodities, woefully inadequate training opportunities over the years, and alternative service sector jobs which are now available at competitive wage rates with superior benefits, it is easy to understand why large numbers of le aren't knocking on the industryes door.23 people g The historical solution to the market failing to train in construction has been collective bargaining. A collectively bargained contract between a union representing construction workers and an association representing contractors has traditionally resolved the problem of meeting long-term training needs in a market that rewards only the short run calculations of contractors. If you and I as contractors are signatories to a collectively bargained contract, that contract will not allow me to get screwed by you. contractors have agreed tha t for you and I and the other signatory contra g Together, 9 Tog , y the good of the industry in the long-run, so much per hour (say 50 cents) will be put into an apprenticeship training fund. That means for every hour any of my workers are on a job,50 cents goes for training apprentices. When I write up my bid, I know I have this cost. But what is more, I know you have this cost as well. I know that you might win the bid over me, but it won't be because I kept in mind the future training needs of the industry and you didn't. We both have to put the collectively bargained training costs into our bid. No pirating is possible because in the future I may hire the worker you trained but I shared in the cost of that worker's training. Thus, with collective bargaining sa{bid. 33 in place, the contract serves as a mechanism for the market to provide training. Who provides construction apprenticeship training in Kansas today? Table 8: Distribution of Apprenticeship Training by Craft and Program Type, Kansas 1989-95 Distribution of Aonrentices TVDe of Proararn Trade Nonunion conectiveiv Baraained Bricklayers 0% 100% Carpenters 4% 96% Electricians 28% 72% Ironworkers 0% 100% Painters 0% 100% Pipefitters 21% 79% Plumbers 9% 91% Roofers 10% 90% Sheetmetal Workers 18% 82% Other 1% 99% Total 12% 88% Source: U.S. Bureau of Apprenticeship Training Table 8 shows the distribution of construction apprentices in Kansas over the period 1989 to 1995. These data do not include Kansans serving in apprenticeship programs headquartered in Kansas City, Missouri. However, for registered apprenticeship programs in Kansas, Table 8 shows that overall 88% of all apprentices are trained in collectively bargained apprenticeship programs. This may understate the number of apprentices trained by nonunion contractors by not measuring programs that are less formal and unregistered. Almost always, collectively bargained apprenticeship programs are registered and entail formal training procedures. Some informal, nonunion programs may exist but go unrecorded by the U.S. Bureau of Apprenticeship Training. Nonetheless, the overall pattern is clear. Apprenticeship training in Kansas construction takes place primarily under the auspices of collective bargaining. This fits with what we know about the market dynamics of construction. The problems is, with the repeal of Kansas' prevailing wage law, collective bargaining in construction has declined. With it, apprenticeship training in construction has also declined. 34 Table 9. Construction Apprentices in both Union and Nonunion Program s by State, 1973-1990 1973 863 1949 604 3543 3276 813 824 1135 1378, 7870 3005 348 1974 1019 2548 849 3600 3464 981 961 1213 1851 8761 3687 435 1975 1184 2415 900 3621 3619 1153 981 1252 2092 10514 3358 671 1976 1053 2061 854 3004 3299 1020 873 1335 2046 10365 3030 554 1977 1117 1702 846 2919 3100 1081 844 1236 2070 10144 3010 569 1978 1131 1644 950 3101 3596 1079 788 1291 1907 9989 3495 613 1979 980 4024 4609 1134 887 1491 2370 10852 3832 682 1987 869 2656 5536 295 424 993 1253 5939. 143 1988 1468 2858 5285 279 378 1013 1222 5253 2719 155 1989 782 6309 2837 641 310 1033, 1182 5079. 143 1973 1979 1050 3402 3566 1037 880 1279 1959 9785 3345 539 1987-1990 976 3627 4526 378 366 1065 1245 5294 3170 745 0 Average Percent Change Law States -27% Indicates No Prevailing Wage Law Repeal and No Law States -53° Table 9 shows construction apprentices in training by year for Kansas and for the fourteen states we have been comparing with Kansas. The highlighted numbers refer to the states during the years in which--in that state—there was no prevailing wage law. These data are from the U.S. Bureau of Apprenticeship Training. Data in their records were not available for 1980 to 1986. The first thing to notice in Table 9 is that on average for the 1970s, 861 apprentices were in construction programs in Kansas each year. In the first four years after Kansas repeals its state prevailing wage law, the number of apprentices fell to an annual average of 530 24 This is a decline of 38%. But can we attribute this decline to the elimination of prevailing wage regulations? Apprenticeship training has been on the decline for other reasons, most notably a decline in collective bargaining independent of prevailing wage regulations. In some stronger union states such as Minnesota and Missouri, apprenticeship training did not decline. In others such as Wisconsin, the decline was small. But in Oklahoma and Texas, union decline independent of prevailing wage regulations led to declines in apprenticeship training equal to that in Kansas. However, if we take all these states as a group, we can tease out an independent effect of prevailing wage regulations on the decline in apprenticeship training. In the five states, including Kansas, where 24 In 1995,the number was less than half that average--248apprentices.While this undoubtedly rellectsafurther decline in apprenticeship training,it may also reflect a movement of training to Kansas City as programs shrank in Kansas. 35 prevailing wage laws were absent or repealed, apprenticeship training declined on average-53%from the 1970s to the late 1980s. In the states with prevailing wage laws, apprenticeship training declined, on average, - 27%. Thus, repealing prevailing wage regulations acted like rubbing salt into a wound. Training was on the decline anyway, and the elimination of prevailing wage regulations made this trend worse. And what was bad for construction was even worse for minority construction workers. i Table 10: Minority Participation in Construction Apprenticeship Programs by State, 1973-1990 AR CO IA KS MN MO MT NF NM No OK So TX WI WY 1973 148 595 107 120 566 47 78 622 332 1977 231 48 1974 175 544 114 154 618 83 107 722 375 2413 237 49 1975 146 498 109 106 623 70 111 737 405 2530 176 47 1976 182 420 110 99 568 56 118 721 399 2374 155 50 1977 181 382 133 103 610 55 127 711 391 2443 171 52 1978 174 385 134 118 741 52 138 777 547_ 2674 186 56 1979 148 446 134 149 772 62 122 852 527 2934 175 60 1987 87 148 495 34 55 627 227 1452. 26 1988 408 158 417 38 53 656�29S 1 1309 114 33 1989 57 , 803 150 114 32 7012- 1233. 31 90 54� 9 43 54 7 821 _ 1336 174 39 1973-1979 165 121 643 61 114 735 425- 2478 190 52 1987-1990 151'. - - 319 373 60 44 701 231 1333 144 32 Percent "Chan a 8° 16 -42° -1°0 -61° _goo, _480� -4 -24° -38% Average Percent Change Law States -11% __° Indicates No Prevailing Wage Law lRepealand N • Source-U.S.Bureau of Apprenticeship Trainina Table 10 shows minority participation in construction apprenticeship programs in Kansas and the fourteen comparison states.. Comparing the 1970s to the late 1980s, we see that in Kansas minority participation dropped by -54%. This was typical of states with no prevailing wage law. The average drop in minority participation in the five states that never had or repealed their prevailing wage law was -50%. In contrast minority participation in the ten states with prevailing wage laws fell by only -11%, much less than the overall drop in apprentices in those states. This was in part due to a big jump in minority participation in Minnesota. However, even excluding Minnesota, the drop in minority participation is only -27% compared to-50% in states without prevailing wage laws. 36 Table 11: Female Construction Apprenticeship Participation by State, 1973 to 1990 1973 1 11 1 1 1 5 0 1 1 a 2 0 1974 1 7 1 5 1 32 1 0 6 32 6 1 1975 1 12 2 2 4 50 3 5 17 54 7 0 1976 2 13 3 1 16 59 2 13 21 99 10 4 1977 5 34 4 8 21 79 7 16^ 25 104 11 7 1978 15 58 25 23 47 82 15 36 81 188 34 16 1979 17 117 29 63 147 101 22 90 a 113'. 434 73 32 1987 17 105 159 16 7 34 21 a 210. 7 1988 65: 104 145 17 5 41 24r° 152 63 5 283 93 32 6 42_ 19 133. 4 1989 8 1973-1979 61M 15 34 58 7 23 38 131 20 9 0JMMEMEM 0 0 -ffi0 0 -1. -37 9A 955o _ o Average Percent Change Law States 171% Indicates No Prevailing Wage Law o hi Table 11 shows female participation in construction apprenticeship programs in Kansas and the fourteen comparison states. There are more minority apprentices than females in construction generally including Kansas. In 1979, there were only 29 female construction apprentices in Kansas compared to 134 minority apprentices. By 1990 there were only 9 female construction apprentices in Kansas compared to 57 minority apprentices. Looking at these numbers alone,the female participation fell by two-thirds while minority participation fell by around one-half. But there were so few female apprentices in Kansas in the early 1970s that the average for that decade was substantially less than its peak of 29 in 1979. Consequently, comparing the drop in female participation in the 1970s as a whole compared to the late 1980s shows a drop of only -14%. But with some exceptions, female participation has been on the rise elsewhere. Particularly in Minnesota, female apprentices rose between the 1970s and late 1980s by over 900%from 15 to 149. In Missouri,they rose from 34 to 134. On average in the states without prevailing wage laws including Kansas and Colorado25 that repealed their laws, women construction apprentice participation rose by 68%from the 1970 to the late 25 Two cities in Colorado retain city prevailing wages,Denver and Pueblo. 37 1980s. However, female participation rose much faster, on average 171%, in the states that retained their prevailing wage laws. Why? Collectively bargained apprenticeship programs involve many contractors. Consequently, on average, they are larger than non-collectively-bargained programs that usually involve a single contractor. Affirmative action regulations do not apply to apprenticeship programs of less than 5 apprentices. Consequently, when training on the collectively bargained side of the industry declines, the programs most likely to fall, under affirmative action regulations become a smaller percentage of all apprenticeship training. There is an interrelationship between prevailing wage regulations and affirmative action regulations in the construction labor market. The repeal of prevailing wage regulations brought with it an diminution of legal pressure to enroll women and minorities into construction apprenticeships. Not only will there be fewer trained construction workers in Kansas due to the repeal the states prevailing wage law, but of those that remain,fewer will be skilled minority or women craft workers. �i p � 38 The Increase in Injuries in Kansas Construction After the State Repealed Its Prevailing Wage Law And a Comparison with Surrounding States The General Relationship Between Prevailing Wage Regulations and Safety The general recipe for safety in construction is simple: larger, more experienced contractors working with well-trained and experienced crews are safer than smaller, less-experienced contractors working with less experienced and less trained workers.26 Repeals of state prevailing wage laws set in motion a train of events that lead to the proliferation of less experienced contractors teaming up with less trained and less experienced workers. This leads to more injuries. Cutthroat competitiveness in contracting. The repeal of the state prevailing wage laws often lead to a burgeoning of start-up contractors with limited track records.These new entrants join existing contractors in a heated bidding process that can put safety at risk. Because of their relative inexperience, new firms tend to face greater on- site coordination problems than firms with longer track records. Such problems can add to costs, but also directly endanger safety. Problems in coordination, perhaps related to delivery of materials and equipment, or in scheduling work with subcontractors, lead to greater uncertainty with 26.C.Culver,M.Marshall,and C.Connolly,Construction Accidents: The Workers'Compensation Data Base, 1985-1988, Washington,DC,OSHA Office of Construction Engineering, 1992. 39 respect to the construction schedule. Uncertainty is a breeder of safety risk, as workers can less easily anticipate and plan for the daily contingencies of work. New entrants in the industry also are generally smaller in size than established firms. Smaller firms have worse safety records than larger firms, in part because of greater laxity of enforcement of safety rules and the relative absence of formal safety programs. Of greatest importance, however, is the firm's reaction to increased pressure to cut costs in the face of intensified competition and cost overruns.There is a tendency to speed up work and cut back on safeguards in the face of such pressures. Workforce turnover, When state prevailing wage laws were repealed, worker turnover increased significantly, as the industry found it harder to n -term careers see Chapter Three). Repeals retain workers for log ( p ) resulted in a decline in the union share of the construction labor market, driving down average construction wages in the state and decreasing union apprenticeship training for construction. In response to the decline i n union membership and training, contractors attempted to reduce turnover —to retain skilled workers and to minimize screening and training costs. Still, the decline in wages and in health and pension benefits drove experienced construction workers from their trades for careers in other industries. In states that retain their prevailing wage law—compared with those that never had such a law or repealed such a law—the proportion of construction workers receiving training is higher and injury rates are lower. A decline in wages and benefits leads to a flood of inexperienced workers into the industry as well as a decline in skilled, experienced workers needed to supervise the recruits and to assure that they work safely. Decline in the skill base of the construction labormarket. Experience is a major determinant of safe work performance—and productivity. Training of skilled construction workers is normally conducted through apprenticeship training programs, most of which are operated by unions and employers through joint trust funds.An integral part of this training is learning on the job while properly supervised. In that way, workers learn from experience while on a variety of projects. Among other things, apprentices are trained to identify and correct ergonomic problems, to detect physical hazards, and to detect the presence or release of hazardous chemicals. Knowledge about safety and health hazards, appropriate protective measures, and hazard communication methods are all important elements that apprenticeship programs provide. 40 When prevailing wage acts are repealed, training and apprenticeship programs decline and the skill base of workers erodes.Without employer incentives to continue apprenticeship programs, knowledge of proper safety and health procedures declines as well. Summary.The combination of these factors—cutthroat competition, a decline in training, and an erosion of career attachments to the industry— affects the.safety-related skill and experience base of the construction labor force. Workers become more injury-prone and know less about the kinds of risks they are taking.Furthermore,as the workforce becomes less skilled and its wages in construction decline, workers are forced to take more safety risks to simply make a living. Furthermore,contractors caught in the competitive speed-up often press their workers to speed up and take more chances. Workers are put at increased risk in an already hazardous industry. The Rise in Injuries in Kansas Annually, the various state departments of labor in cooperation with the U.S. Department of Labor, Bureau of Labor Statistics, conduct an occupational injury and illness survey. This survey reports for a variety of industries, including construction. In Table 5 of the survey, the survey reports the number of workers employed in each industry category, the number of injury cases and the number of injury cases that result in lost days from work. I have gathered these surveys for the period 1976 to 1991. For this period, Figure 5 shows the number of injury cases per worker. Kansas repealed its prevailing wage in 1987. The number of injuries per worker in construction immediately jumps from an annual average of.11 to above .13. That is,injury cases rose after repeal from an annual average of 11 injury cases per 100 construction workers to more than 13 annual injury cases per construction worker. This is a 19% increase in injuries annually after repeal. 41 Number of Injuries per Worker in Kansas All Cases .15 .14 .13 .12 lA -- �j .10 . 09 76 77 78 80 81 82 83 84 85 86 87 88 89 90 91 YEAR Figure 5: Number of Injury Cases per Construction Worker in Kansas, 1976 to 1991. Source: U.S. Bureau of Labor Statistics Serious injuries that resulted in several lost days of work rose from 4.4 serious cases per 100 construction workers to 5.3 serious cases per 100 Kansas construction workers. Thus, there was a 21.5% increase in serious injuries after Kansas repealed its state prevailing wage law. Figure 6 shows these data. 42 Number of Injuries per Worker in Kansas Resulting in Lost Workdays by Year .07 .06 iv n a i .05 'c .04 ca � .03 76 77 78 80 81 82 83 84 85 86 87 88 89 90 91 YEAR Figure 6: Serious Injuries per Worker(Resulting in Lost Days from Work) in Kansas Construction, 1976 to 1991. Source: U.S. Bureau of Labor A statistical test of whether or not these increases in injury rates are significant yields the answer Yes. (See Appendix.) Table 12 shows the basic data for this test. At all standard levels of statistical significance, we can say that both injury rates and serious injury rates rose in Kansas construction after the repeal of Kansas' state prevailing wage law. Table 12: Mean and Standard Deviation of Injury Rates in Kansas Construction Before and After the Repeal of Kansas' State Prevailing Wage Law Group Statistics Legal Std. Std.Error Status N Mean Deviation Mean Lost Day Has PW 10 4.382E-02 3.901E-03 1.234E-03 Cases per Law Worker No Law 5 5.324E-02 4.521E-03 2.022E-03 Cases per Has PW 10 .1107 7.325E-03 2.316E-03 Worker Law No Law 5 .1320 9.857E-03 4.408E-03 43 In the second Chapter of this report, school construction costs were compared in states around Kansas that do not have a prevailing wage law to states around Kansas that do have this regulation in construction. A similar comparison of injury rates in construction can be made. This allows us to check the results of our analysis of Kansas by itself. Table 13 shows injury rates per worker and serious injury rates per worker in construction for the years 1976 to 1991 broken down by injuries that occurred in states with prevailing wage laws and in states without prevailing wage laws. The states in the analysis are the fifteen states used in Chapter Two of this report. These include Iowa, North Dakota and South Dakota—states that never had a prevailing wage law. It also includes Colorado and Kansas--states that had a prevailing wage law during the first part of the period but later repealed their law. However, Colorado data are only available for two years, both during the period in which it had a prevailing wage law. The data also include Montana, Wyoming, Minnesota,Wisconsin, Missouri,Arkansas, New Mexico,Texas and Nebraska--all states with prevailing wage laws. Oklahoma is also included. During the time period under consideration, 1976 to 1991, Oklahoma public construction was regulated by a prevailing wage law. Table 13: Injury Rates and Serious-Injury Rates in Construction for 15 States Broken Down by Having or Not Having a Prevailing Wage Law, 1976-1991 Source: U.S. Bureau of Labor Statistics Group Statistics Legal Std. Sid.Error Status N Mean Deviation Mean Lost Day Has PW Cases per Law 106 4.654E-02 7.689E-03 7.469E-04 Worker No Law 20 5.301 E-02 8.329E-03 1.863E-03 Cases per Has PW 106 .1134 1.737E-02 1.687E-03 Worker Law No Law 20 .1425 6.431 E-02 1.438E-02 While the number of observations has risen substantially in Table 13 compared to Table 12, the basic result is the same. Total-Injury rates and serious-injury rates are higher where prevailing wage laws are absent. Injuries per worker rise from 11 per 100 construction workers to 14 per 100 construction workers. Serious injuries resulting in lost work days rise from 4.7 per 100 workers to 5.3 per 100 workers. These are increases in 44 injury rates of 26%and 14%respectively. And these differences are statistically significant. (See Technical Appendix to this Chapter for results of statistical significance tests.) Prevailing wage laws regulate the construction industry in a way that promotes safety. The absence of prevailing wage laws leads to a less safe work place with all the explicit and hidden costs injuries create for the worker, the industry and the community. p nu � w } i 45 Chapter 5 Technical Appendix Tests of Statistical Significance for Table 12: Injury Rates in Kansas Construction Before and After Repeal Independent Samples Test Levene's Test for Equality of Variances t-test for Equahtv,of Means 95%Confidence Sig. Mean Std.Error Interval of the Mean F Sin. F df 2-tailed )ifference ifference Lower U er Lost Day Equal Casesper variances .033 .858 13 .001 -9.42E-03 .247E-03 -1.43E-02 4.57E-03 Worker assumed Equal variances -3.979 7.095 .005 -9.42E-03 .369E-03 -1.50E-02 -3.84E-03 not assumed Cases per Equal Worker variances .773 .395 -4.750 13 .000 -2.13E-02 .485E-03 -3.10E-02 -1.16E-02 assumed Equal variances -4.278 6.300 .005 -2.13E-02 .980E-03 -3.33E-02 -9.26E-03 not assumed Tests of Statistical Significance for Table 12: Injury Rates in Kansas Construction Before and After Repeal Independent Samples Test Levene's Test for Equality of Variances t-test for Equality of Means 95%Confidence Sig. Mean Std.Error rinterval the Mean F Sig. t df (2-tailed) Difference DifferenceUpper Lost Day EqualCases per variances .005 .946 -3.409 124 .001 -6.47E-03 1.899E-03 -2.71E-03 Worker assumedEqual variances -3.226 25.482 .003 -6.47E-03 2.007E-03 -2.35E-03 not assumed Cases per Equal Worker variances 6.140 .015 -3.998 124 .000 -2.91E-02 7.269E-03 -4.35E-02 -1.47E-02 assumed Equal variances -2.007 19.526 .059 -2.91E-02 1.448E-02 -5.93E-02 .185E-03 not assumed 46 Pension and Health Benefits in Construction Before and After the Repeal of Kansas' Prevailing Wage Law Pension and health benefits play two crucial roles in the construction industry. First, by providing needed income security in old age and needed health coverage today, these benefits permit adults with families to participate in the industry while knowing that their families' basic needs are insured. Second, pension and health benefits help create and preserve needed skills within the industry. People willing and capable of acquiring the skills needed for solid, high quality construction are also people capable of acquiring the skills needed by many industries. if the construction industry cannot provide the basic benefits needed by families, the construction industry will steadily lose its better and more experienced workers to other industries that will provide these benefits. Annual Average Employer Contribution to Pension and Health Insurance in Kansas Construction in 1996 Dollars by Before and After Repeal $22,000,000 $20,032,695 ,� - $20.000,000 - $18,000,000 $16,583,342 $16,000,000 $14,000,000 - - $12,000,000 $10,000,000 Before Repeal(1982-86) After Repeal (1987-92) Figure 7 : Total Annual Employer Contributions to Pension and Health Insurance in Kansas Before and After Repeal. Source: U.S. Labor Department Form 5500 47 As Figure 7 shows, total annual average employer contributions towards o pensions and health insurance in Kansas construction fell by 17/o after the 1987 repeal of the state's prevailing wage law. Why? The simple answer is that the repeal helped shift Kansas construction work away from collective bargaining towards the merit or open shop. Merit shop contractors have difficulty paying their workers pension benefits or health insurance. This difficulty is rooted in the same market failure that prevents training on the open shop side of the industry. Construction workers move from job to Pb. They have to simply because today's building gets built and today's road gets paved. So eventually, the construction worker has to move on. In doing so, the worker often changes employers. Merit shop contractors find it both awkward and not worth Their while to insure the health and old age of workers that will be with them a limited amount of time. So merit shop contractors develop insurance programs for their key workers who do stay for years. But the merit shop contractors find little reason and much difficulty in providing these same insurance benefits to the transient worker. Collective bargaining provides a mechanism for allowing and inducing contractors to provide health insurance and pensions. Construction projects still come to an end. Construction workers still move on to new employers. But the new employer like the old is a signatory to the collective bargaining agreement. That agreement requires that each employer contribute so much per hour on the worker's behalf into a pension fund and into health insurance. Thus, when a union construction worker's child gets sick, the child is covered by health insurance. And when a union construction worker retires, he or she has something more than Social Security to look forward to. This is rot only good for the construction worker and his or her family. It is good for the community as well. Construction represents around 5% of the labor marker. Thus, in round terms, construction workers and their families represent 5% of our neighbors. Neighbors that can afford a doctor when a child is ill-- neighbors who can take care of themselves when they are old—these are neighbors that are less a burden on the community as a whole. Table 14 shows the average employer contribution per worker in Kansas construction on an annual basis from 1982 to 1992. The figures are inflation adjusted so that earlier years can be directly compared to later years. 48 Table 14:Annual Average Kansas Employer Contributions per Workerto Pensions and Health Insurance in Kansas Construction 1982 to 1992 And the Percentage of Merit(or Open) Shop Workers Covered by Insurance in Kansas 1982 •$3,228 $2,700 $99 $31 1983 $2,429 $2,653 $88 $43 1984 $3,011 $3,104 $79 $58 1985 $2,637 $3,345 $90 $74 1986 $2,771 $3,429 $89 $69 1987 $2,793 $3.070 $94 $54 , 1988 $2,415 $3,306 $143 $119 `. 1989 $2,332 $3,048 $165 $138 1990 $2,338 $3,122 $219 $248 1991 $3,211 $4,252 $223 $292 � 1992 $2,890 4 897 287 190`= Source:U.S.Labor Department Form 5500 In constant(or inflation adjusted)1996 dollars Looking at union employers first, Table 15 shows that over the ten years-- 1982 to 1992—in Kansas, union contractors have contributed around $3,000 per year to pension programs for their workers. In inflation- adjusted dollars, this contribution has been fairly steady over the time period. In contrast, union employer contributions to health insurance almost doubled over the period,from $2,700 per worker to almost$5,000 per worker. The reason for this is clear. Health costs rose dramatically over the period. Union contractors attempting to preserve their workers' health benefits found they had to pay an increasing premium for health coverage. Nonunion contractors in Kansas also increased their health premium per worker over the period 1982 to 1992. However,the average premium per worker was low to begin with ($31 per worker) and low at the end ($190). This is not because merit contractors could find cheap health insurance that would give coverage for $190 per worker per year. Rather it is because most of the merit shop workers simply were not covered. Interestingly, Kansas merit shop contractors pay more per worker in pension contributions that they do in health contributions. Under collective bargaining, union employers pay more in health premiums. The reason for this is the advent of 401 k plans. This has allowed employers to contribute to pensions that can move with the worker. What percentage of merit shop workers are covered by health insurance from their employers? An estimate can be made from looking at the average health premiums of a merit shop and a union shop worker. 49 contractor does not provide substandard health Assuming the meat shop insurance for the worker who is covered, then the cost of insurance for a construction worker should be roughly the same on the union and merit side of the industry. Thus, if the merit shop contractor pays the same for health insurance as the union shop contractor, and the average premium on the merit shop side of the market is only 4% of the contribution per worker on the union side of the market, then only 4% of the merit contractor's workers are being covered by health insurance. If more than 4% are being covered, then it is because the merit shop contractor is buying less health coverage. A similar analysis can be made for pensions. If all merit shop workers are covered by a pension, then merit contractors in Kansas are paying only $287 per year to help out their workers in retirement. Alternatively, if the merit contractor contributes almost $3,000 per year towards his workers' retirement, then only 10% of his workers are being covered by pensions. In sum, the repeal of Kansas' prevailing wage law helped shift the state's construction away from collective bargaining. On the merit shop side of Kansas construction, only 10% of the workers are covered by pensions and only 4% are covered by health insurance. So quite naturally, total contributions into pension and health insurance fell after the repeal. We saw in Chapter Three that construction worker wage incomes across the entire state fell by around 10% after the repeal of the state prevailing wage law. Now we see that pension and health insurance contributions fell by even more--17%. This is a problem for construction workers in Kansas. But it is also a problem for Kansas. Solid communities need solid health and old age insurance. People who cannot take care of themselves when they are ill or when they are old become burdens on their families and burdens on the community. We saw in Chapter two that the alleged gain from prevailing wage repeal does not exist. In this chapter we find that the pain of lost health insurance and a less secure old age is real and measurable. 50 Summary and Condusions Kansas was the first state to pass a state prevailing wage law regulating the payment of wage rates on public works. The Republican legislators who wrote this law, embedded it in larger legislation seeking to reduce the working day in Kansas from 10 or 12 hours to 8 hours per day. Kansas' prevailing wage law came within a broader legislative initiative to impose factory safety inspections, to limit the use of child labor and prison labor, and to make schooling compulsory. The general purpose of all these laws—including Kansas' prevailing wage law—was to encourage the Kansas labor market to develop up a high-skill, high-wage growth path. Competition was to focus upon which employer could train and equip a skilled labor force to do a quality job. Kansas Republican legislators specifically wanted to avoid competition over which employer could stretch out the day longer, employ more children, employ more prison workers, sacrifice safety to the bottom line and/or dodge long-term training costs for short-term market victories. Eventually, 41 states and the Federal Government followed Kansas' example and passed prevailing wage laws of their own. Between 1979 and 1988, nine states repealed their prevailing wage laws. Repeals came with the promise that by cutting wage rates and benefits on public works, taxpayers could save substantial sums of money on public construction costs. In Kansas, merit shop contractors predicted that Kansas would save from 6% to 17% on state construction costs, and in some cases even more money would be saved. The pain of lower wages and fewer benefits for Kansas construction workers would be more than offset by the gain to the taxpayer. In this study of the effects of the repeal of Kansas' prevailing wage law,we looked for the construction costs savings that proponents of repeal predicted. Kansas new school construction costs from 1991 to 1997 were compared to new construction costs in 14 Great Plains states. Five of those states (including Kansas) did not have prevailing wage laws. Nine 51 of these states retained their prevailing wage law. And Oklahoma's prevailing wage law was judicially annulled in 1995, in the middle of the study years. Schools provide a useful example of the effects of prevailing wage laws on public construction costs for three reasons. First, schools are a major part of state and local public construction expenditures. Second, when broken down into elementary, middle and high schools,these building types make for a good apples-to-apples comparison. Third, many public schools are built—more than any other single type of government building. So enough observations are available in the case of schools to make meaningful statistical comparisons. The results of this study are clear. There were no statistically significant differences in the construction costs of new schools in Great Plains states with prevailing wage laws compared to those states without prevailing wage laws. Furthermore, Kansas fits into this pattern precisely. On average, Kansas does not build schools any cheaper than surrounding states that have prevailing wage laws regulating the construction of their schools. For example, in the case of elementary schools, we have the most observations, and the structure types are the most similar from one school to the next. The average square foot construction costs of new elementary schools in Great Plains states with no prevailing wage law was $76.23. The average square foot construction costs on new elementary schools in Great Plains states with prevailing wage laws was$76.86. This difference of 66 cents per square foot was not statistically significant. Kansas' average square bot cost on 18 new elementary schools from 1991 to 1997 was $83 per square foot.27 This higher average cost, however, was not statistically significantly different from the overall average for all Great Plains states. For new construction of elementary, middle and high schools,there were no statistically significant,measurable cost differences between states with prevailing wage laws compared to states without prevailing wage laws. The predicted substantial gains from the repeal of Kansas' prevailing wage law are simply not there. But the predicted pain from the repeal of Kansas' prevailing wage law did arrive. Wage income for all construction workers in Kansas—not just on public works but on all construction sites--fell by 11% after the repeal. In contrast, the wage incomes of construction workers in surrounding states that retained their prevailing wage law fell by 2%. Roughly speaking, Kansas construction workers wages are 10% lower due to the repeal of the state prevailing wage law. Construction contractor contributions to pensions and health insurance in Kansas fell by an even larger amount after repeal. In inflation adjusted 27 All cost comparisons are adjusted for inflation and in 1996 dollars. 52 dollars annual average employer contributions to pension and health 0 insurance fell by $3.5 million per year. This represents a 17% drop in contributions to pension and health coverage for Kansas construction workers. The reason for the fall in pension and health coverage in construction is not hard to find. While the major rationale for repeal was to cut public construction costs, advocates of repeal also argued that the elimination of Kansas prevailing wage law would open up business opportunities for merit shop contractors. Union contractors collectively bargain over wages and benefits and sign a contract binding all signatories of that contract to specific hourly contributions into apprenticeship training, pensions and health coverage. Merit shop contractors do not have collectively bargained contracts. They are free to pay each of their workers on each individual's own merit. As a consequence of this freedom from collective bargaining, only 10% of merit shop workers in Kansas are covered by pensions contributed to by their employers. Only 4% of merit shop workers are covered by health insurance provided by their employers. Only 12%of all registered apprentices in Kansas are trained by merit shop contractors. This is not because merit shop contractors do not want to train their workers or provide good benefits and health coverage. It is because outside the workings of collective bargaining in construction, it is difficult to provide for these long-term needs of the industry. Under collective bargaining, the contractors as-a-group agree with the workers as-a-group to provide so much per hour for apprenticeship training, so much per hour for health insurance and so much per hour for old-age and disability pensions. Each contractor must--by the rules of the contract—include these costs in each and every bid they submit. Consequently, the contract forces long-term industry needs and costs into the short-run bid considerations of each signatory contractor. On the merit shop side of the industry, no collective contract governs bidding. A contractor may wish to include training costs and health insurance. But that contractor—in a cutthroat bidding environment—always must face the prospect that his competitor will skip those long-term costs to get this job in the short-run. So apprenticeship funds go wanting. Pension and health benefits are shaved. The problem is exacerbated by the fact that in construction, workers go from contractor to contractor as jobs ramp-up and then shut down. Under collective bargaining, each contractor agrees to pay for the training of not only his own apprentices but also those of his signatory competitors. The contract requires it. And it makes sense. Your competitor's apprentice may one day soon be your journeyman. 53 But in the merit shopsector of the industry, each contractor has an ry incentive not to train. If I train at my cost an apprentice that later goes to work for my competitor, I am simply cutting my own throat by subsidizing my competitor. Consequently it is not surprise that merit shop contractors in Kansas account for only 12% of all registered apprentices. Many merit shop contractors try to avoid apprenticeship training. If they train, they train informally--only for the immediate skills needed on this job, and only as a last resort if they cannot find the needed skills out in the market. The first result of the repeal of Kansas'prevailing wage law on training was that apprenticeship training fell by 38%. But the long-run effect was the creation of a labor force with not only fewer skills but a narrower base of skills. Registered apprenticeship training seeks to train workers in the general skills of their craft not the narrow skills of one specific job. Consequently, the shift away from formal apprenticeships to informal, problem-at-hand training has proven to be a shift towards thinly skilled workers with limited commitment to construction as a craft or career. While this study documents the effect of Kansas'repeal of prevailing wage regulations on the skill and manpower crisis in construction,the problem is wider than simply in Kansas. In a story "Craft Shortages Creeping In," The Engineering NewRecord surveyed the top 400 general contractors and top 600 specialty contractors around the country. ENR stated: "The industry has known for the past decade that it was headed for manpower trouble...Nonunion contractors vwrking in bustling areas appear to have the biggest manpower problems. For example,56%of the union crafts in the West reportedly have no labor shortages while only 10% of the open shop crafts have no problem."28 ENR stated that the South had the greatest craft labor shortages. (The Deep South is the one area in the country where no state has a prevailing wage law.) But the problem really is tied to the movement away from the disciplines given to the industry by prevailing wage regulations and a reasonable amount of collective bargaining. Nonunion contractors, themselves, recognize the problem. In a report commissioned by three major merit shop contractors, the writers state: Clients [i.e. owners purchasing construction services] have created a `playing field' which forces contractors to undercut one another to obtain work. Combined with the fact that craftsmen are treated as expendable commodities, woefully inadequate training opportunities over the years, and alternative service sector jobs which are row available at competitive wage rates and superior benefits, it is easy to understand why large numbers of people aren't knocking on the industry's door. 28 ENR,December 25,1995,p.34. 54 in wage laws have joined the group of States that have repealed their prevailing g J 9 construction industry"clients"that have created a cutthroat playing field where under-bidding today is the only rule of business. As a result craftsmen become "expendable commodities". As a result, training opportunities becoming "woefully inadequate". As a result, the service sector provides "competitive wages and superior benefits"compared to construction. Kansas construction has seen a loss of its experienced workers to other industries and retirement. These skilled workers are being replaced by a younger cohort of less trained, less skilled, less experienced and less career- committed workers. Consequently the industry has become less safe. Serious-injury rates in Kansas construction rose by 21% after the repeal of the state prevailing wage law. A comparison of Great P lains states with prevailing wage laws to those—including Kansas—without this regulation finds that injury rates are 26% higher in the states without prevailing wage regulations. These are not simply injuries on public works. These are injuries across all of Kansas construction. Prevailing wage repeal contributes to these injuries by cutting out the support for apprenticeship training that makes the worker more knowledgeable of job site hazards. Prevailing wage repeal contributes to higher injuries by cutting out support for the payment of pensions and health insurance. Experienced, middle age workers are safer workers. But the absence of pension and health benefits in construction encourages construction workers to leave the industry once they start forming families. This leaves the playing field to younger, less experienced, less trained workers who are more injury-prone. Prevailing wage repeal contributes to a higher injury rate by helping further erode construction wages. Whe n construction wages es become secondary wages, people no longer see construction as a place to develop a career. The loss of career workers creates a more dangerous workplace for those who remain. Inexperienced workers are a danger to those who work around them as well as themselves. On a job site replete with career workers, the inexperienced worker receives guidance. An inexperienced workforce--left to its own devices--measurably increases the risks and costs of injuries. The original purpose of prevailing wage laws was to avoid the costs of an unskilled and inexperienced work force. These costs are social as well as economic. A construction worker who has health benefits and can look forward to a pension is less likely to become a burden on his or her family and community. The promise of repeal was lower public construction costs. But that promise went unfulfilled. The cost of prevailing wage repeal in Kansas has been substantial. 55 Construction workers,themselves have lost income and benefits—but that was the predicted by supporters of repeal. Construction in Kansas has become more dangerous. That was an unforeseen consequence. Skilled workers have left the industry. That too was unforeseen. Training has declined substantially. Again this was not predicted by repeal proponents. Now,and in the future, Kansas as a community will face the problems of an uninsured construction labor force. The health and old age problems of Kansas construction workers may simply go unmet, or the cost of these peoples' health and old age may be shifted to Kansas taxpayers. This too was an unforeseen cost of repeal. Because the benefits of repeal in terms of cost savings on public construction are minimal at best--and more likely simply not there—now may be the time to revisit Kansas' repeal of the first prevailing wage law in the country. Because the costs of repeal are significant, measurable and on-going, now may be the best time to re-enact Kansas' prevailing wage law. 56 Losing Ground : Lessons from the Repeal of Nine "Little Davis-Bacon" Ads Peter Philips, Garth Mangum Norm Waitzman, and Anne Yeagle Working Paper Economics Department University of Utah February 1995 THE. UNIVERSITY OFUTAH Summary: A major study of nine states (Alabama, Arizona, Colorado, Florida, Idaho, Kansas,_Louisiana, New Hampshire, and Utah) that had repealed prevailing wages found that the repeals had negative impacts on all state budgets. The loss of construction earnings and sales tax revenues had an adverse impact, and cost overruns on road construction also increased costs. In Utah, for example, these cost overruns tripled after the repeal. Training was reduced by 40%,"minority representation was reduced in training programs and injuries increased by 15%. "The study concluded that if the federal Davis-Bacon Act was repealed that federal tax revenues would drop by$1 billion per year, and that there would be 76,000 additional workplace injuries in construction annually, with more than 675,000 work days lost each year. These increases would be felt in increased workers compensation costs and costs placed on public health systems by workers without health and pension benefits. Losing Ground: Lessons from the Repeal of Nine "Little Davis-Bacon" Acts Peter Philips, Garth Mangum Norm Waitzman, and Anne Yeagle Working Paper Economics Department University of Utah February1995 S YO 6 THE UNIVERSITY OFUTAH Acknowledgements The authors wish to thank Hamid Azari-Rad,Randy Brown,Van Hemeyer,Matt Hotchkiss, Gary Ray,and Scott Smith-all students at the University of Utah-for help gathering information for this study. We also wish to thank the many government officials at the state and federal levels who helpfully provided data for our analysis. This study was originally funded in 1992 by a gift from Local 3 of the International Union of Operating Engineers. Subsequent funding came from the United Association of Plumbers and Pipe fitters of Utah and from the AFL-CIO. Some material in this volume originally appeared in Hamid Azari-Rad,Peter Philips,and Anne Yeagle,"The Effects of the Repeal of Utah's Prevailing Wage Law on the Labor Market in Construction," in Sheldon Friedman,et al.,eds.,Restoring the Promise ofAmerican LaborLaw(Ithaca:Cornell University ILR Press, 1994),207-22. This is a working paper of the Economics Department of the University of Utah.Inquiries, comments,criticisms and suggestions should be directed to: Peter Philips Professor Economics Department University of Utah Salt Lake City,Utah 84112 TEL 801 585-6465 FAX 801 585-5649 INTERNET PHILIPS@ECON.SBS.UTAH.EDU The authors welcome your comments. This working paper will be periodically updated as our research progresses. Updated electronic versions of this working paper are available from the anonymous ftp site at the Economics Department of the University of Utah. Internet users may use ftp(file transfer protocol)commands to obtain an updated version. The file is in Wordperfect 5.1 with extended postscript attachments. It may be printed out in Wordperfect 5.1 or 5.2 on any postscript laser printer. To obtain an electronic copy do the following: (1)log on to an internet terminal;(2)at the prompt type ftp keynes.econ.utah.edu;(3)at the request for user name type anonymous;(4)at the password request type you e-mail internet address;(5)once you have entered the University of Utah Economics Department ftp site type cd tmp [then return] then type binary; (6) then type is to obtain a listing of the files in the ftp site. The file you are looking for is called DAVISBAC. (7) Type get davisbac c:\davisbac.my where the second phrase c:\davisbac.my refers to the drive you want the file to go to and the name you wish the file to have. The analysis and opinions expressed in this paper are those of the authors and do not necessarily reflect the opinion of the University of Utah. Copyright 1995. All rights reserved. Peter Philips Garth Mangum Norm Waitzman Anne Yeagle Davis-Bacon Repeal Effects 3 Contents Executive Summary,page iii The Authors,v I. The History of Prevailing Wage Laws in the United States, 1 Passage of State Prevailing Wage Laws,2 Passage of The Davis-Bacon Act,3 Repeals of Some State Prevailing Wage Laws,6 Efforts to Repeal Other Prevailing Wage Laws,7 Efforts to Repeal Davis-Bacon,8 II. The Economic Effects of Davis-Bacon Repeals, 11 Cutthroat Bidding, 11 A Loss of Earnings for All Construction Workers, 16 A Loss of State Tax Revenues, 17 Regression Analysis of the Decline of Construction Worker Earnings,21 Increased Employment Associated with Lower Wages,24 The Net Effect of Repeals on Government Budgets,25 Summary,31 III. The Effect of State Repeals of Prevailing Wage Laws on Training,Black Unemployment,and Minority Participation in Training,33 The Effect of Repeal on Construction Unions and Wages,34 The Relation between Repeals and Black Unemployment,37 A Decline in Training,40 Market Responses: Training,Turnover,and Careers,43 National Trends in Registered Apprenticeship Training,45 Summary,56 IV. Construction Safety Put at Risk, 59 Why Prevailing Wage Law Repeals Lead to Increased Injury Rates,60 A Comparison of Injury Rates,62 The Cost of Injuries,62 Summary,65 V. Conclusion,67 The Effects of Repeal of Prevailing Wage Laws,67 The Goals of State Prevailing Wage Laws, 67 The Definition of a Prevailing Wage,67 The Financial Costs of State Repeals, 68 Other Costs of State Repeals, 71 Estimated Effect of a Davis-Bacon Repeal,75 End Notes,77 References, 83 Figures 2.1 The mix of construction employment in Utah,by contractor type, before and after the repeal of the state's prevailing wage law, 12 2.2 Average cost overruns as a percentage of accepted bids on Utah road construction,before and after the repeal of the state's prevailing wage law, 14 2.3 The ratio of accepted bids and final cost to the Utah state engineer's estimate of road construction project cost,before and after repeal of the state's prevailing wage law, 15 2.4 A comparison of annual construction earnings,by status of prevailing wage law, 18 2.5 A comparison of construction earnings in nine repeal states only, before and after repeals (in 1991 dollars), 19 3.1 Union membership in construction in Utah, 1977-89,35 3.2 Wages and employment in construction in Utah relative to wages,36 3.3 The ratio of black to white unemployment in five repeal states,before and after repeals,39 3.4 Black,white unemployment ratio for states that retained and never had state prevailing wage laws,40 3.5 Apprentice plumbers as a percentage of journeymen plumbers in Utah, 1961-91,42 3.6 Turnover in Utah's construction industry compared with all employment statewide,46 3.7 Apprenticeship training rates,by state groups,before and after repeals,49 3.8 Apprenticeship training rates,by state groups,50 3.9 Minorities as a percentage of all construction apprentices by state groups,54 3.10 Ratio of the percentage of minorities in construction as a ratio of the percentage of minorities in the state population,by state groups, 55 4.1 Injury rates in construction by status of prevailing wage law,63 5.1 Annual income-tax revenue loss, construction cost savings, and resulting effect of repeal on Utah budget, 1987-93, in 1994 dollars,70 5.2 The percentage of minority pp ri apprentices in construction, divided by the percentage of minority nine repeal states 73 population in the state-the minority reflection percentage-forp PoP ty 5.3 Estimated effect of a repeal of the Davis-Bacon Act on income-tax revenues,construction cost, and total budget(billions of dollars),74 Tables 1.1 Prevailing wage laws,by state,4 2.1 A simple estimate of Utah tax revenues lost in 1991 as a result of the 1981 state prevailing wage law repeal,20 2.2 A description of the data used in regression model of construction earnings decline,22 t f the effects of state repeal on construction annual earnings,controlling ion model estimate o 2.3 Are regression P g for regional differences in earnings,and for secular and cyclical trends in earnings,23 2.4 Effects of wages on employment,controlling for state differences in employment,differences in the size of SIC groupings,the direct effects of repeals,and secular and cyclical trends,26 2.5 Effects of construction earnings decline on employment for an average-size detailed construction standard industrial classification(4-digit SIC)of 3,540 workers per state,27 2.6 The relation of hypothetical construction-cost savings to tax revenues,29 2.7 Projected effects of a repeal of Davis-Bacon on the federal budget,30 3.1 Linear regression model of turnover rate in construction in Utah, 1956-91,47 3.2 Training rates in repeal and never-had states as a percentage of training rates in states that retained their wage laws,52 3.3 Training rates in repeal and never-had states as a percentage of training rates in states that retained their wage laws, 1975-78 and 1987-90,53 4.1 Regression model of the effect of state repeals on injury rates for plumbers and pipe fitters,64 0 Executive Summary Like the 1931 federal Davis-Bacon Act,legislation in 41 states has required that the"prevailing"wage be paid on state-government-funded construction projects. Betwee 1979 to 1988,however,nine states repealed their prevailing wage laws.(Nine states never had such a law.)The remaining 32 states have e p g . P retained prevailing wages.These variations in state experience provide useful information with which to consider probable effects of additional state repeals or the proposed repeal of Davis-Bacon.This study found that state repeals of prevailing wage laws had several effects. First, in Utah,whose experience was examined most closely,the state budget has not benefited from repeal of the prevailing wage law.The repeal helped drive down construction earnings and as a result,the state has lost substantial income tax and sales tax revenues. In the decade before the 1981 repeal in Utah,construction worker earnings averaged about 125 percent of average non-agricultural earnings. By 1993, construction worker earnings had fallen to 103 percent of the average earnings for Utah workers.This decline in earnings is a result of both lower wages and a subsequent shift to a less- skilled construction labor force. Second, also in Utah,the size of total cost overruns on state road construction has tripled in the decade since repeal in comparison to the previous decade. The shift to a less-skilled labor force — lowering labor productivity along with wages — and the greater frequency of cost overruns have lessened any possible savings in public works construction costs associated with the repeal. general downward trend in real construction for en the states and controlling g looking at all g Third, g � earnings,variations in state unemployment rates, es and regional t onal differences in wages,repeals eals have cost construction workers in the nine states at least$1,477 per year in earnings,on average(in 1994 dollars). The costs may eventually be higher as the effects of the more recent repeals mature,driving wages and training down further. Fourth,controlling for a general downward trend in the amount of construction training,variations in state unemployment rates,and regional differences in training availability,the nine state repeals have reduced construction training in those states by 40 percent. Fifth,minority representation in construction training programs has fallen even faster than have the training programs in repeal states.Until the various state repeals,minority apprenticeship participation each state's population.After repeal,minorities became significantly ri percentage p ed the wino P mirror minority p g under-represented in construction apprenticeship programs. Sixth,occupational injuries in construction rose by 15 percent where state prevailing wage laws, were repealed. Based on these findings,we conclude that,if the federal Davis-Bacon Act were repealed: • Federal income tax collections would fall by at least$1 billion per year in real terms every year for the foreseeable future.This is because construction wage levels would decline across all states and-based on the experience of the nine repeal states-construction employment levels would not revenues may well be higher. If the n e loss.The figure for lost tax Y rise enough to offset this revenue to g experience of the nine states that never had a prevailing wage law is 6 Davis-Bacon Repeal Effects • indicative,lost tax revenues from a repeal of Davis-Bacon could rise to$2 billion per year.Whether the losses are$1 billion or$2 billion,the government cannot count on making them up with its cost savings as a purchaser of construction. The government will not break even. 0 There would be 76,000 additional workplace injuries in construction annually,with 30,000 of them serious and thus requiring time offfrom work to recover.As a result,more than 675,000 work days would be lost each year in construction.This could lead to additional workers'compensation costs of about$3 billion per year,of which$300 million would be passed on to the federal government as increased costs on public works. • Utah'sexperience suggests that repeal of Davis-Bacon would generate a period of significant cost overruns and the increased use of expensive change orders.Although we cannot measure the exact costs of such practices,it is generally accepted that change orders add substantially to construction costs. iv Davis-Bacon Repeal Effects The Authors Peter Philips (PhD, Stanford University, 1980) is a Professor of Economics at the University of Utah.He is co-editor of Three Worlds of Labor Economics and coauthor of the forthcoming Building Retirement Security: the Central Pension Fund of the Operating Engineers. Philips has published widely on the canning and construction industries in journals such as Industrial and Labor Relations Review,Industrial Relations,Business History,the Journal ofEconomic History,and the Cambridge Journal of Economics. Philips has received awards for his teaching, including University of Utah Public Service Professor and Presidential Teaching Scholar. Garth Mangum (PhD, Harvard University, 1960) is Max McGraw Professor of Economics and Management at the University of Utah. He is author or coauthor of more than 30 books and numerous monographs and articles on labor and employment, including Capital and Labor in American Copper, The Operating Engineers: Economic History of a Trade Union, and Labor Struggles in the Post Office and Union Resilience in Troubled Times: the Story of the Operating .ff Engineers 1960-1993.In addition to teaching at several universities and serving as an arbitrator in more than 600 labor disputes,he has served the federal government as senior research analyst of the Presidential Railroad Commission,Research Director of the Senate Subcommittee on Employment and Manpower,and Executive Secretary of the National Commission on Technology,Automation and Economic Progress. Norm Waitzman(PhD,American University, 1988)is an Assistant Professor of Economics at the University of Utah and the co-author of Business War on the Law: An Analysis of the Benefits of Federal Health and Safety Law Enforcement and The Incidence and Costs of Birth Defects: The Value of Prevention. Waitzman's work on the economics of safety and health appears in the writing American Journal of Public Health,Demography,Inquiry, and the Economic Forum. He is itin g a book,to be called Physician Supply and Public Policy. Anne Yeagle is a PhD candidate at the University of Utah who is writing her dissertation on the effects of the 1981 repeal of Utah's prevailing wage law. I. The History of Prevailing Wage Laws in the United States In February 1891,Samuel Gompers,president of the American Federation of Labor,visited Topeka, Kansas,to speak on what the local newspaper called"the great topic of labor."Ten years earlier,the AFL—at its own creation—had laid out legislative aims that included the eight-hour work day,the elimination of child labor, free public schooling, compulsory schooling laws, the elimination of convict labor, and prevailing wages on public works. These proposals were based on a belief that the American labor market should consist of highly skilled workers earning decent wages,with time for family,and with children free to earn an education.In pursuit of these aims,Gompers'political strategy in Kansas allied him with the Republican Party. On the morning p of Gom ers's arrival,the Alliance Party,known to history as the Populist Party, withdrew an earlier invitation for him to speak in the hall of the state House of Representatives, represented 900 000 workers,had fallen out of favor with controlled.Gompers,who re which the arty p res p p the populists,reportedly because of his belief that the trade unions should not form a political party with the Alliance.' The Republicans,who controlled the Kansas Senate,invited Gompers to speak there, and he did. Gompers was in Kansas to focus on the eight-hour day.Like other Americans,Kansans in 1891 typically worked six days per week,ten to twelve hours per day.In the older trades and crafts,such as carriage making and saddle making, where the work pace was slow and under the workers' direction,the long work day was tolerable.In the newer factories producing shoes,textiles,and the like; in the mines; and in the urban putting-out systems in needlework, six-day weeks and twelve- hour days were grueling. The AFL had made its prime objective a shortened work day and work week with as little cut in pay as possible. In his Topeka speech, Gompers declared: Our banner floats high to the breeze and on that banner float is inscribed,'Bight hours work,eight hours rest and eight hours for mental and moral improvement.t2 At that time,when there were no income supplement programs for the poor,low-income parents worked and had to send their children to work to make ends meet. This practice was later referred to by a North Carolina newspaper editor as "eating the seed corn." Each generation of poor condemned its offspring to poverty because the children grew up as illiterate as their parents. The prevalence of cheap child labor,which accounted for 5 percent of the manufacturing labor force in 1890 and a larger proportion of service sector workers,kept wages down and forced adult workers to put in the long hours to make ends meet.Gompers wanted regulation to force employers and the poor to adopt a strategy,however painful in the short run,of a high-wage,high-skilled growth path where children were in school and workers had the skills to justify wages that would allow for a family life. Gompers said, The Federation endorses the total abolition of child labor under 14 years of age;an eight hour law for all laborers and mechanics employed by the government directly through contractors engaged on public work, and its rigid enforcement;protection of life and limb of workmen employed in factories, shops and mines; ...the extension of suffrage as well as equal work for equal pay to women....The Federation favors measures,not parties.3 re Although it was not clear at the time whether government could require private sector employers to honor the eight-hour day,government could set an example,Gompers believed.In state after state,he pleaded for the eight-hour day for government workers and private sector workers employed on public works. Gompers also pleaded for workers to be paid the "current" daily wage so they could afford the reduced work time.Government was being asked to set a good example for the private sector,to show that a refreshed labor force could produce in eight hours what a fatigued and bedraggled labor force turned out in ten or twelve hours.The prevailing wage law in its infancy was an attempt to obtain shorter working hours for all labor.The AFL paid attention to public works, however, because government at all levels was a major purchaser of construction. The AFL said government should not try to save money by eroding the wages of its citizens. With similar logic,the AFL called for an end to convict labor.Many states employed convicts to pay for their keep.Convicts built roads on chain gangs,operated farms,made textiles,and sewed garments. Convict-made goods were sold, forcing down prices and the wages,of working free citizens. Thus, prevailing wage law legislation, at its birth, was embedded in an overarching intent to shorten the grueling working day for all labor, to compel the working poor to make ends meet in some fashion other than by sending their children into the factories,to compel children into schools so that they might become better workers and better citizens, to compel employers to adopt techniques that profited on the employment of skilled adult workers rather than unskilled child labor, to present government as an exemplar of good management by establishing the eight-hour day in government employment and on public works,and to abolish the practice of government saving tax dollars by grinding down wages on public works or through convict labor. I t i s n o t surprising,then, that the first prevailing wage law passed in the United States—in Kansas—was part of an eight-hour-day law. Passage of State Prevailing Wage Laws The Kansas Eight-Hour law. Kansas established the first prevailing wage law in 1891. In January 1890, the Kansas Bureau of Labor and Industrial Statistics, in preparation for its Sixth Annual Report,distributed a questionnaire to each trade union and the Knights of Labor Assembly. In response to a question about needed legislation,the Molder's Union of Parsons,Kansas,replied that he wanted"a law...against the letting of contracts for State work to unfair employers.,4 This plea for the state to let out contracts fairly appears to be one of the first reports leading up to the enactment of a prevailing wage law. In February 1891,the Second Annual Convention of the Kansas State Federation of Labor,in Topeka, approved a bill concerning state-paid wages. That month, the bill, which included the prevailing wage section,called"for an Eight Hour Law"and was brought forth by Mr.Avery of the Typographical Union No.121,Topeka. The bill stated, That in no case shall any officer, board, or commission, doing or performing any service or furnishing any supplies to the State of Kansas under the provisions of the act be allowed to reduce the daily wages paid to employees engaged with him (or them) in performing such service or furnishing such supplies,on account of the reduction of hours provided for in the act.That in all cases such daily wages shall remain at the minimum rate which was in such cases paid and 2 Davis-Bacon Repeal Effects received prior to the passage of the act.' The eight-hour bill was one of four labor-related bills pending in the legislature:the weekly pay bill,the child-labor bill,and the bill to make the first Monday in September a holiday,which would become known as Labor Day.In addition,that year the Kansas State Federation of Labor approved a resolution calling "for the abolition of convict labor when in competition with free labor."6 The eight-hour bill, Senate Bill 151, failed in the Kansas senate March 6, 1891, with the prevailing wage section removed.But by March 10,when the prevailing wage section was put back in,the bill became law.This first prevailing wage law stated, That not less than the current rate of per diem wages in the locality where the work is performed shall be paid to laborers,workmen,mechanics and other persons so employed by or on behalf of the state of Kansas....' At first,however,the law was not enforced!Not until 1900,did the Kansas Bureau of Labor and Industrial Statistics report enforcement: "there were hundreds of complaints that were attended to by correspondence, and good results obtained."'. Prevailing wage laws in other states.New York was the second state to pass a prevailing wage law.New York's eight-hour law(Chapter385) was amended in 1894 by Chapter 622 to include a prevailing wage law for those employed on public works.As in Kansas,however,there were many violations.1'Laws similar to those in Kansas and New York were passed in Oklahoma(1909),Idaho (1911),Arizona(1912),New Jersey(1913),Massachusetts(1914),and Nebraska(1923)(see table 1.1).These laws established a precedent for the creation of the federal Davis-Bacon prevailing wage law. Passage of The Davis-Bacon Act Three federal laws primarily affect prevailing wages in the United States: the Davis-Bacon Act of 1931 which applies to construction,the Walsh-Healey Public Contracts Act of 1936 which covers employers in manufacturing and supply industries,and the Service Contract Act of 1965(known as the OTIara-McNamara Service Act), covering suppliers of personal and business services. These laws attempt to neutralize the effects of government purchases on wage determination in the private sector. The Davis-Bacon Act is the most significant of the three laws. Its objective is to prevent the federal government from affecting local wages and construction conditions; Davis-Bacon disallows the government from pushing down wages in competitive bidding.The government has always been a major purchaser of construction services. As such,the government holds the potential to use its bargaining power to force down wage rates. Davis-Bacon Repeal Effects 3 Table 1.1 Prevailing Wage Laws,by State States having Year prevailing wage laws passed States without prevailing wage laws Alaska 1931 Georgia Arkansas 1955 Iowa California 1931 Mississippi Connecticut 1935 North Carolina Delaware 1962 North Dakota District of Columbia 1931 South Carolina Hawaii 1955 South Dakota Illinois 1931 Vermont Indiana 1935 Virginia Kentucky 1940 Maine 1933 Maryland 1945 Year Year of Massachusetts 1914 States that repealed Michigan 1965 passed repeal g prevailing wage laws Minnesota 1973 Missouri 1957 1941 1980 Montana 1931 Alabama 23 Arizona 1912 1984 Nebraska 19 Nevada 1923 Colorado 1933 1985 Florida 1933 1979 New Jersey 1913 N Idaho 1911 1985 New Mexico 1937 New York 1894 Kansas 1891 1987 Ohio 1931 Louisiana 1968 1988 Oklahoma 1909 New Hampshire 1941 1985 Oregon 1959 Utah 1933 1981 Pennsylvania 1961 Rhode Island 1935 Tennessee 1953 Texas 1933 Washington 1945 West Virginia 1933 Wisconsin 1931 Wyoming 1967 Note:The District of Columbia is listed here,but not included in the count of states. Source:State laws and corrected version of Armand J.Thieblot,Jr.,Prevailing Wage Legislation:The Davis-Bacon Act, State 'Little Davis-Bacon Acts," The Walsh-Healey Act, and The Service Contract Act. Philadelphia: The Wharton School, 1986,p.140. For four years before the 1931 passage of the Davis-Bacon Act, 14 bills were introduced in Congress to establish prevailing wages in construction.Robert L.Bacon in 1927 introduced the first bill proposing a prevailing wage for construction, H.R. 17069. The member of Congress justified his measure as follows: 4 Davis-Bacon Repeal Effects The Government is engaged in building in my district a Veteran's Bureau hospital. Bids were asked for. Several New York contractors bid,and in their bids,of course,they had to take into consideration the high labor standards prevailing in the State of New York...The bid,however,was let to a firm from Alabama who had brought some thousand non-union laborers from Alabama into Long Island,N.Y.; into my district. They were herded onto this job, they were housed in shacks,they were paid a very low wage,and the work proceeded...It seemed to me that the federal Government should not engage in construction work in any state and undermine the labor conditions and the labor wages paid in that State...The least the federal Government can do is comply with the local standards of wages and labor prevailing in the locality where the building construction is to take place." Hearings for a federal prevailing wage law began in 1927 and continued in 1928 and 1930,but no bill was passed. On March 3, 1931, Bacon's original proposal,which he had reintroduced as H.R. 16619,was signed into law by President Hoover.12 The Davis-Bacon Act required payment of prevailing wages on federally financed construction projects.The law essentially ruled out bidding on construction worker wages on federally financed construction. The original language was vague,however, and prevailing wages generally were not determined before the acceptance of bids. In 1935, President Roosevelt signed clarifying amendments to the act,which became the basis of the current Davis-Bacon Act.The National Labor Relations Act of 1935 gave the Secretary of Labor authority to set the prevailing wage. In 1935, Roosevelt's Secretary of Labor, Francis Perkins, established the original rules for determining the Davis-Bacon prevailing rates.The prevailing wage was said to be the wage paid to the majority, if a majority existed; if not,the 30-percent rule was used. The 30-percent rule means if 30 percent of the workers in an area are paid the same rate,that rate becomes the prevailing rate there. The 30-percent rule often resulted in the union wage being the prevailing wage. If the 30- percent rule did not apply, because at least 30 percent of the workers in a given occupation in the local labor market did not receive the same wage rate,the average wage rate was paid to workers doing the same job. The prevailing wage was determined this way for 50 years. In 1985,President Reagan changed administration of Davis-Bacon,creating the 50-percent rule. The revised regulation reduces the influence of the negotiated union wage in most areas(see page 9,below). The Tenth Amendment to the Constitution restricts the ability of the federal government to dictate contract terms for the states.Thus,work funded entirely by state or local governments is not covered by Davis-Bacon. Each state, county, or city can establish its own prevailing wage—if it chooses to do so—through legislation. In 1994,29 percent of all county-level federal Davis-Bacon prevailing wage rates were taken from union contracts, 48 percent used average wages, and the remaining 23 percent of counties used a mix of union and average wages, depending on the occupation. Davis-Bacon Repeal Effects 5 Repeals of Some State Prevailing Wage Laws Kansas had passed the first prevailing wage law in 1891 and,by 1969,41 states and the District of Columbia had prevailing wage laws.Several cities also passed local prevailing wage laws affecting construction.However,state governments began experiencing fiscal crises in the late 1970s.In 1978, California voters passed Proposition 13,restricting state expenditures,and the Labor Law Reform Bill failed in Congress. In this political context, many state legislatures believed that, to save tax dollars,government should use its bargaining power to lower construction costs,even if the probable effect of this action would be the lowering of construction wage rates and a possible effect might be the lowering of quality in the construction industry. More than 51 bills have been introduced in 23 state legislatures to repeal or curtail so-called little Davis-Bacon legislation.13 Alabama, Arizona, Colorado, Florida, Idaho, New Hampshire, Kansas,Louisiana, and Utah have repealed their prevailing wage laws. Florida.Florida,which passed its prevailing wage law in 1933,was the first state to repeal.The statute was repealed over the veto of the governor in 1979.14 One of the most populous counties, Broward,established its own local prevailing wage law and several cities in Broward passed similar laws.15 Alabama. Alabama was the next state to repeal,in 1980.16 After Alabama's repeal,the entire South from Virginia to Mississippi, except Tennessee, was without state prevailing wage law. Unsuccessful attempts were made in 1983 and 1984 to reinstate the 1968 Alabama laws.However, prevailing wage laws exist at the local level,such as one in Mobile for city-sponsored construction." Utah. Utah's prevailing wage law had been passed in 1933. Eventually,prevailing rates were set by hearings held in three districts that were created for this purpose. In addition to covering construction,the Utah statute established prevailing rates for piece work. The first indications of intent to repeal the Utah law were heard from the local chapter of the national Associated Builders and Contractors (ABC) in 1978. (The ABC, nationally and in Utah, sought to represent the interests of non-union contractors.)The Utah ABC outlined its strategy in a letter to other state ABC chapters in 1978: It is our hope that the major argument in favor of repeal would be based on tax savings and unnecessary government spending,rather than a union versus non-union argument.18 The ABC lobbying effort became public during the Utah legislative session in 1979. The sponsor of the Utah repeal, Republican Representative S. Garth Jones wrote in the Deseret News: The prevailing wage rate is substantially the union pay scale. In 1933 the law was designed to place money into a depressed economy,to increase wages to get the economy moving.The law does the same thing today.But today,the economy is not depressed;inflation is the problem and the cost of government is too high. Repealing the prevailing wage will allow the free enterprise system to establish the wages of tradesmen at a substantial savings to the taxpayers.The prevailing wage law is inflationary.Additionally,the prevailing wage rate discourages non-union contractors from bidding public contracts.It encourages union contractors to bid public contracts.The effect is to force people looking for work to go to union contractors.The law is inconsistent with Utah's Right to Work law.(Feb. 23, 1979) 6 Davis-Bacon Repeal Effects The first bill to repeal the statute was introduced in 1979, only to be vetoed by Democratic Governor Scott Matheson.In 1981,repeal bills were introduced in 14 states.Only in Utah did repeal succeed that year and it succeeded only after a second veto from Matheson.19 The bill was approved on almost straight party lines — Republicans favoring repeal and Democrats opposed.The Salt Lake City Tribune noted that only one Republican representative,who called himself a lifelong Republican and union member,voted against repeal and broke away from party lines.20 When Matheson vetoed the bill in 1981,he said, "I'm convinced that repeal of this law is not in the best interests of working people in the trades whose skills are essential for a vigorous construction industry.0' Nonetheless,the Senate overrode the veto 21-7 and the repeal took effect 2 months later. Those in favor of the repeal maintained that the prevailing wage law was inflationary and pro- union. Republican C. McClain(Mac)Haddow sponsored the 1981 repeal bill. He said, "the law is outmoded and is preserved only as a tool to extend union control.The law is contrary to Utah's right- to-work philosophy...... 22 Roger Evershed, president of the Associated Builders and Contractors, predicted a 10 to 15 percent savings on public works projects with repeal.21 Arizona. The next state to repeal was Arizona in 1984.24 Arizona's statute began as an eight- hour work day in 1912 and,by 1930,became a prevailing wage law.In a court test,the statute was found unconstitutional in September 1979.25 In November 1984, voters repealed the statute in a ballot initiative, Proposition 300. Provisions of the ballot initiative prevented communities from implementing local prevailing wage statutes.26 Idaho. Idaho's prevailing wage law was first enacted in 1911 as an eight-hour law.The statute was extensively amended until 1965; efforts to repeal it began in 1979. The legislature failed to override several vetoes but did repeal the law in 1985.2'At the same time,overtime pay requirements for more than eight hours of work were repealed.zs Colorado. Colorado also repealed its prevailing wage law in 1985.29 Attempts for repeal began in the late 1970s,but it was not until after the governor had vetoed the bill several times that the veto was overridden and the repeal passed. Nevertheless, since 1985 at least one municipality,Pueblo, established its own prevailing wage rate for local construction." New Hampshire. New Hampshire joined Colorado and Idaho in 1985 when it,too,repealed.31 Although legislators began in 1979 to try to repeal the prevailing wage law, they did not succeed until 1985. Influenced by reports of inflated costs on a school construction job,both houses passed repeal without the signature of Governor John Sununu.32 Kansas and Louisiana. Kansas, the first to have a state prevailing wage law, repealed it in 1987.33 Louisiana followed in 1988 with repeal over the initial veto of the governor" Efforts to Repeal Other Prevailing Wage Laws The Massachusetts ballot initiative. In Massachusetts,in 1988,thousands of union members, already active in the presidential election, worked with community groups to help defeat a ballot Davis-Bacon Repeal Effects 7 initiative that would have repealed the states 1914 prevailing wage law.The effort to block repeal in Massachusetts appears also to have slowed efforts to repeal other state prevailing wage laws until the midterm elections of 1994.Massachusetts Question 2,the repeal initiative and the hottest issue on the ballot that year,was defeated 58 to 42 percent on November 8" The Massachusetts law requires contractors to pay employees on state-financed projects a predetermined wage. Prevailing wage rates are most often based on collective bargaining agreements, which vary by trade and geographical jurisdiction.16 In 1988, the Associated Builders and Contractors (ABC) and Citizens for Limited Taxation formed a coalition that spearheaded the repeal effort,with a signature drive run by the "Fair Wage Committee." In March, a report by the Massachusetts Foundation for Economic Research, The Peculiar Prevailing Wage Law, presented the public rationale for a repeal of the state law.37 The report stated that the many attempts to modify the prevailing wage law were defeated before reaching the governor's desk 38 Using confidential data collected from a construction contractor,the authors estimated that the prevailing wage law increased construction costs by 14 percent through higher wage costs.The report concluded that,"in 1987,the prevailing wage law cost Massachusetts at least$212 million dollars."" In August, in response to the report by the Foundation for Economic Research,the Regional Information Group of Data Resources Inc. presented a contrasting view. Data Resources said the earlier report had used insufficient data and oversimplified analyses 40 Data Resources maintained that a repeal in 1990 would result in a "total wage loss of$196 million and a net employment loss of 600."Data Resources concluded that although there would be nominal tax savings with a repeal, the overall impact would be to increase unemployment and lower living standards 41 By the end of a hard-fought campaign, community support included the Catholic Church; the Jewish Labor Committee; the Massachusetts Nurses Association; the National Women's Political Caucus; and the National Organization for Women.az A similar effort in 1994 to repeal by initiative failed on the Oregon ballot.The battleground has shifted back to state legislatures and the U.S. Congress. Efforts to Repeal Davis-Bacon The onset of state efforts to repeal prevailing wage laws coincided with U.S.Senate hearings in 1979 to repeal Davis-Bacon. During the first hearings, Davis-Bacon proponents defended the law with these points: 1. The act prevents the disruption of local wage and construction market conditions by the introduction of federally financed construction. 2. The act protects the prevailing living standards of construction workers by discouraging cutthroat competition by construction contractors. 3. The act provides equality of opportunity for contractors who are free to bid on the basis of skill, efficiency, and knowledge, rather than on their ability to slash labor standards. 8 Davis-Bacon Repeal Effects 4.The act helps maintain the high quality of the construction labor force and equal employment opportunity in the construction trades by encouraging use of bona fide training programs on federally funded construction.13 Advocates of repeal of Davis-Bacon said: 1. The act has inflated construction costs. 2.The act costs the federal government huge amounts of money. 3. The act is poorly administered. 4.The act is biased toward union contractors and hurts non-union contractors. 5. The act has caused wage inflation. 6. The act discriminates against minorities, because they are disproportionately represented among the low-skilled labor force. 7. The free-market system is suppressed. Although the Davis-Bacon Act was not repealed in 1979, the Reagan administration changed the way the law is administered a few years later.The administration in 1985 altered the 30 percent rule.Until then,the Department of Labor used the modal —most common—wage to determine the prevailing wage for an occupation in a local labor market,if the modal wage to the penny accounted for more than 30 percent of all wages for that group.`If the modal wage accounted for fewer than 30 percent of all wages,the mean(average)wage was declared the prevailing wage. The Reagan administration raised the threshold to 50 percent before the modal could be declared the prevailing wage.Union wages tend to be the modal wage and they tend to be above the mean or average wage for an occupation. So the Reagan administrative change had the effect of lowering the prevailing wage in areas where unions were weak. Given construction unionization rates have fallen from around 80 percent of the construction labor force in the 1940s to around 60 percent in the 1960s to around 25 percent in the 1980s, the impact of the Reagan administrative changes were substantial.45 Some of the competing claims for and against Davis-Bacon can be tested against the experience of the states—those that have repealed state prevailing wage laws, as well as those that continue to have such laws, and states that have never legislated a prevailing wage. This study examines the contentions of Davis-Bacon proponents that prevailing wage laws prevent the disruption of local wage and construction labor markets and that prevailing wage laws protect living standards and discourage cutthroat competition. This study examines, as well, the contention of Davis-Bacon opponents that the law costs government considerable sums of money and discriminates against women and minority construction workers.The study also raises two new questions.First,what are the effects of prevailing wage laws on training and human capital formation in construction?Second, what effects do these laws have on the safety and health of construction workers? Davis-Bacon Repeal Effects 9 10 Davis-Bacon Repeal Effects II. The Economic Effects of Davis-Bacon Repeals Cutthroat Bidding As soon as the law was repealed,some of these non-union people[contractors]that had been doing small work around town suddenly just took off,and the union people[contractors]like ourselves, our market share decreased. -President,a union construction company, Salt Lake City, 1993 [Our]company has consisted of my father and my grandfather and me from about 1963.[We are a double-breasted company.] Company A is a union [general] contractor that hires merit shop companies with no regard to union affiliation.Company B is a non-union merit shop company.... Our industry became very competitive during the mid-eighties, a lot of people are chasing the same type of work. -General contractor,double-breasted company,Salt Lake City, 1993 We'vebeen in business for 51 years.Before that my great-grandfather ran a construction company and so we've always done construction. Right now weYe doing mostly mechanical, and we do utilities,Mountain Fuel,water lines,sewer lines,AT&T jobs.We've built homes.We've built golf courses.Were built apartment buildings. In the last probably about eight years[since the mid- 1980s]there's a lot more small companies-little tiny,you know,dad and his three boys.We can't compete against them. We have too much overhead to do that and you get small start-up companies,they're willing to work for nothing for a while and you know they'll go out there for two years and just take these jobs dirt cheap.Sometimes they can't finish.They'll go broke in the middle but still,we don't want to work for nothing. We'd just rather lock the gate and wait. -Office manager,union construction company,Salt Lake City, 1993 When Utah repealed its prevailing wage law in 1981, the structure of the construction industry changed dramatically. The most obvious effect was the decline of union membership and union contractors.But this was only the most obvious effect.Underlying the decline of union contractors was the rise of the small contractor and increasing turnover of contracting firms in the business.The industrial organization of the industry changed, with an increased reliance on subcontractors. Comparing the 12 years prior to repeal to 10 years after repeal, the share of total construction employment accounted for by the typically bigger and more capital-intensive general contractors and heavy and highway contractors fell,while the share of total employment accounted for by specialty subcontractors rose (fig. 2.1). With the entry into the market of more contractors and smaller contractors,competitive pressure to win bids heated up. This pushed wages down. An operating engineer familiar with the bidding wars stimulated by Utah's prevailing wage law repeal tells how the bidding affected labor. Davis-Bacon Repeal Effects 11 60%- ---------------------------- 55%-//------------------------------------------------------- ------ ------------------------------------------------------ ------------------------------------------------------ ...... c: 450/6- 0 t ------------- ------ z' 40%- ---------------------------------------- 0 ------------------- C) 35 01c ----------------------------------- 45 30%- -------------------------------- ...... 25%- 20%-/ Before 70-81 V After 82-91 15% General Heavy Specialty Figure 2.1 The mix of construction employment in Utah by contractor type,before and after the repeal of the state's prevailing wage law. Source:Utah LMI Annual Report,Table 5. After the repeal of the state's prevail*m,9 wage el taw,the di tributi na-*�f employment- shifted among types of contractors intahr. IM TWal a hig.her,pereentage of construction emplpy es wor ked for gep6r-41,conttac kors and- heavy,and-highway contractors. i here are icai y ar r = AAWA)w I higher,perceniage of ernployb c4ntra which be worked smaller°fikms. 12 Davis-Bacon Repeal Effects When they repealed Utah's law, a lot of companies went out of business because of the cutthroat competition.A lot of companies just bought jobs so they could have a cash flow to make payments on their equipment.The design engineers would tell the contractor that let's say the job was going to cost a million dollars.The contractor would still go in there anyway and low-ball the bid.Then they would turn around to their workers and make their wages fit whatever they had to be to fit the low-ball bid. The general contractors did a lot of bid shopping after the prevailing wage law was repealed. The general contractor would get a bid from the subcontractor of say$50,000 and then he would low-ball the bid.Then,when the general got the job he would go back to the subcontractor and say yeah I've got the job but you've got to cut your bid to$40,000 to have this job I've got and the sub would go back to the workers and say OK we've got this job but now I've got to cut your wages. See costs of materials and supplies and equipment were stable. The price of bricks and the asphalt didnl go down just because you got this job.So the workers had to make up the difference for all this low-ball bidding. So basically the employer got their money off the backs of the worker.Whether it was to make money or just to break even,wages had to fall. -Operating engineer,Bountiful,Utah, 1994 But wages were not the only factor to feel the strain of an overheated bidding process. Government purchasers of construction services were now exposed to practices of low-balling bids and over-running costs. Average annual cost overruns for the Utah Department of Transportation prior to the law's repeal was 2 percent of initial accepted bid(fig. 2.2). Since the repeal,however, overrun costs have risen to 7.3 percent of the initial bid.This rise in overrun costs has come despite the introduction of computers as a tool for contractors in preparing their bids. The cause of these increased overrun costs is the post-repeal tendency for contractors to take more risks in the bidding process under the pressure of increased competition(fig. 2.3). When the state calls for bids on a project,the.state engineer prepares an initial estimate of the project's cost. In the decade prior to the repeal of Utah's prevailing wage law,winning bids averaged 91 percent of the state engineer's estimate. After the repeal,winning bids have been,on average, 89 percent of the state engineer's estimate. Although contractors are apparently shaving their bids to win state contracts, these lower estimates have not proved to be a windfall for the state. Instead,after Utah's prevailing wage law repeal, final construction costs have been running at 95 percent of the state engineer's initial estimate. This amounts to 6 percentage points above the accepted bids.Prior to Utah's repeal,final costs were running 93 percent of the engineer's estimate, only two points higher than initial accepted bid prices. This does not necessarily mean that the pre-repeal construction was ultimately cheaper for the state,but it does mean that the relationship between accepted bid price and actual costs was more certain and that contractors promised less before Utah's repeal, but delivered more relative to the state engineer's cost estimates6 Davis-Bacon Repeal Effects 13 8.00% 7.00% --------------------------------------------------- L 6.00 o O 5.00% U 4.00% Q 3.00% ------------------------------------------- 2.00% 1.00% 0.00% Before(1970-81) After(1982-94) So°rce:Utah,Dept of Transportation,"Final Estimates Processed for Payments",1%5 and 1994 Figure 2.2 Average cost overruns as a percentage of accepted bids on Utah road construction, before and after repeal of the state prevailing wage law Cost overruns:crn the.construction o€Utah ro#d ..averaged 2,percent.over#ccepted bids in the.ciecase be€ t-e Vta h's repeal:of zts pa wage iw. in the decade after repeal,average eQt ovens ro&ete 1fj over the aoscept Change orders associated w!th.,eost overruns axe one of the more expensive components of-vanstruction-costs. 14 Davis-Bacon Repeal Effects ------------------------ -------------------------------------------------------- 100 --------------------------------------------------------------------- 93% 9$% E 96% N ------------ 91% ----------------- --------- > 94% :.. ....... --------- v 92% 0) -------- IL 90% a 88% Q f 86% 89% Q 84% 82% Before 70-81 80% After 82-94 Bid/(Eng's Est) Actual/(Eng's Est) Figure 2.3 The ratio of accepted bids and final cost to the Utah state engineer's estimate of road construction project cost,before and after repeal of the state's prevailing wage law After the Ta#ah repeal of*, g wage eompet'106 a apin g e-ontraetors heated up and contractors shaved their bids to win contract. - the,decade before the state repeal,accepted bids averaged 91 percent of the state engineer's estimated project cost on road construction. After repeal;accepted bids-fell,on average,to 89 percent of-the state engineer's estimates. However,this cutthroat bidding did not cut final project costs as a percentage of the state ogineer's estimates. In the decade prior to the repeal of Utah's prevailing wage law,final costs averaged 93 percent of the state engineer's project cost estimate.In the decade after the repeal,because of a tripling of cost overruns,the final project costs averaged"95 percent of the state epgineerls estiinat . Davis-Bacon Repeal Effects 15 A Loss of Earnings for All Construction Workers Heightened competition after Utah's repeal has not only created uncertainty in the bidding process, but has also lowered Utah construction wages across the board. A union plumber describes this: After Utah repealed its little Davis-Bacon law I was working on a job as a union plumber. The electricians on the job were non-union. At that time there was terrific pressure on wages and, as I remember, the IBEW [International Brotherhood of Electrical Workers] took a big wage cut-something like$3-from$16 to$13.Anyway,the day after the union electricians took that cut, the contractor came on the job and told these non-union guys they would have to take a $3 cut too. There was a lot of animosity around that but they took the cut anyway. They had to. Our union held off two years before we had to do the same thing the electricians did,and when we took our cut the non-union plumbers'wages fell right along with ours. -Union plumber, Salt lake City, 1994 Utah repealed its prevailing wage law just as the economy was falling into the 1982 recession. Thus,the effects of the repeal initially were tangled up with the effects of the recession. However, some of the nine states that have repealed their prevailing wage laws did so in good times and some in bad times. A comparison across states can somewhat disentangle effects of the business cycle from effects of a repeal. Whatever a government might save in construction expenses from the repeal of a prevailing wage law, the saving has to be balanced against the loss of other revenues. The lower wages paid on government-financed construction have a ripple effect, lowering wages throughout the local construction industry.Construction workers in states that have a prevailing wage law have a higher average annual income than construction workers in states that have repealed a law; and those workers,in turn,earn more,on average,than do construction workers in states that have never had a prevailing wage law (fig. 2.4). That pattern may be explainable, however, for more than one reason.States that have different prevailing wage law policies may have higher or lower construction earnings for reasons unrelated to the wage law. For instance,repeal states might also be low-wage states in general. It may thus be more useful to isolate earnings data for repeal states only-before and after(fig. 2.5). Average annual construction-worker earnings in the nine states that repealed their prevailing wage laws from 1979 through 1988 show a drop of$1,835 from $24,317, or about 7.5 percent in wages,adjusted for inflation and denominated in 1991 dollars,or$2,016 in 1994 dollars.The nine states are not heavily unionized and a fall of this magnitude cannot be accounted for simply by a fall of union wages to the non-union level. In recent years,the average construction unionization rate in the nine states that repealed their state prevailing wage laws has been around 13 percent of the construction labor force."With this level of union coverage,for a fall in the union wage to account for all of the fall in the average wage, at the outset of the repeal,union workers would have had to have been earning 60 percent more than non-union workers.41 Union wage differentials typically are around 10 to 20 percent above non-union wages. Because union wages are not sufficiently high and union coverage not sufficiently wide to account for all the fall in construction wages in these repeal states,we know that non-union workers have had to absorb some share of this average earnings decline. 16 Davis-Bacon Repeal Effects If one assumes that the union differential is 20 percent above the non-union wage g and,after the repeal,the union wage falls to the non-union wage,both wage rates will have to fall even further to attain an overall 7.5 percent cut in earnings. Assuming that the union wage would fall to the non- union rate and then they would both fall together,the union wage would have to fall by 21 percent and the non-union rate would have to fall by 5 percent to obtain an overall fall of 7.5 percent 49 In fact, only rarely does the union rate fall entirely to the non-union wage. A reasonable assumption would be that the union rate prior to a repeal was 20 percent above the non-union rate and after the repeal fell to 10 percent above the non-union rate. Given a 7.5 percent overall fall in earnings and a 13 percent union membership rate, union wages would have to fall 14 percent and non-union wages would have to fall 6.3 percent to obtain an overall fall of 7.5 percent.In other words,while the union rate would have to fall twice as much as the non-union rate, the non-union sector of construction workers would have to absorb much of the average percentage wage cut. The effects of state repeals of prevailing wage laws are isolated neither to union workers nor to government- financed construction.50 They generate across-the-board cuts in the earnings of all construction workers. A Loss of State Tax Revenues The tax revenue losses that result from lower construction wage levels are surprisingly large. Whatever the source of this earnings decline among construction workers,states with income taxes have lost tax revenues as a result of this decline in taxable income among construction workers.And, because this lost income means lost purchasing power, states that have repealed their prevailing wage laws have also lost some sales tax revenues. On average,construction workers account for 5 to 6 percent of a state's labor force. In Utah in 1991,individuals earning$20,000 to$30,000 paid a marginal state income tax rate of about 7 percent.Taking the 31,528 construction workers employed in Utah in 1991 and an average per capita decline in income of$1,835, the total loss of annual income from the Utah construction industry in Utah in 1991 because Utah's 1981 repeal could be calculated as $58 million($1,835 times 31,528). Given a marginal tax rate of 7 percent, 1991 lost state income tax revenues might amount to $4 million (in 1991 dollars) (table 2.1). Assuming a marginal propensity to consume on sales-taxable items from changes in income of 80 percent and a sales tax rate of 6.25 percent,lost state sales tax revenues from this loss of income amount to$2.9 million in 1991.51 Adding these two losses and bringing them to 1995 values using the consumer price index yields an estimated loss of$8.2 million in state taxes in Utah in 1991 evaluated in 1995 dollars. The figure of $8.2 million in lost tax revenues may be an overestimate for four reasons, however. First, if wages fall and labor becomes cheaper,contractors might hire more workers. So we must consider possible increases in total income of construction workers resulting from possible increases in total construction employment after a fall in wages. Second, real wages have been falling in the United States generally,including the construction industry.Some of the lower wages after state repeals may simply reflect a long-term decline in real wages that would have taken place anyway.Third,annual earnings in construction are sensitive to unemployment.Earnings rise when unemployment falls and fall when unemployment increases. Because Davis-Bacon Repeal Effects 17 i $30 $26,842 --------------------------------------- --------------------- $25 $22,482 $21,125 o i 0 - --------- ---------- ....... $20 c -- m -------- ----- 8 (n $15 o 6 h- ------ ------ c $10 m -------- -------- ----- - - Q $5 $0 Have Law After Repeal Never Had Law States by Groups Figure 2.4 A comparison of annual construction earnings,by status of prevailing wage law and Earnings,Source:US DOL Employmentg , 1975-91. Figure 2;4;groups states into thi e+e caaegor=ies(from deft to right}. The first.bar,an titer.shows average annual income i4 f99. dollars for construction workers in All states and years whore a state prevailing wage law was en forced. This includes repeal states.prior to repeal. The-second-bar shows the average annual earnings of construction workeis=in repeal, tes=after repeal! The thrd.bar represents a er ge-annual earnings for.c tr o arl r .tl ro giic 't 4`9to M,in all states,that never hacl a prcva ng-gage _ T,# data prove isiitia vi fence that repealing or,never-having.a p",W[ug wfage law lowers cau eruct on income nstruction industr not only oh public works but across the-eh#ire state-coy. 18 Davis-Bacon Repeal Effects $30 $24,317 --------------I--- ------------------------------------------------------------------------ L $25 sl $22,482 0 C) r- - -------------- -------- rn $20 cc: ---- -------------- --------- $15 o ---- ----------- --------- Q $10 a) CU $5 01— Before Repeal After Repeal Source:DOL Employment and Eamings States by Groups Data for Years 197$!1991 Figure 2.5 A comparison of construction earnings in nine repeal states only,before and after repeals(in 1991 dollars) In the,nine states that. pealed weir rt' g i s het- een"I979 I 84 average anal ncQe f�,after the repels{�cate instant'i99 -dolknrsj• 77 This fact does not en xral" nr Oka.IA-xors Ilia , i�ht l�a�e�c�n(riving.dov�n wages,but it is pirima facie evidence that the repeals forced lower earnings not just on public wer*s--hat across the construction labor market. Davis-Bacon Repeal Effects 19 Table 2.1 A simpl e estimate of Utah tax revenues lost in 1991 as a result of the 1981 state prevailing wage law repeal Individual construction income prior to repeal(1991 dollars) $24,317 Individual construction income after repeal $22,482 (1991 dollars) Lost income due to repeal(1991 dollars) $1,835 1991 Utah construction employment 31,528 Total lost income in construction(1991 dollars) $57,853,880 Lost Utah income tax $4,049,772 Lost Utah sales tax $2,776,986 Total lost tax revenues 6 8$ 26,758 Total lost tax revenues in 1995 dollars $8,192,109 The averse annual constructivxtw eating ?En 19 ticla' fr_nie repel.stt to the years after 1975 Aad=before each stat'e3s��l was$2 ,3°17. #lie years after each repeal up to 1991,the average constructiot'earnings fell to$22,482. "Utah construction employment in 1991 was 31,528 workers and multiplying these by an annual loss of income of$1,835 yields a total lost income fn Utah.construction of 857 8 million. BAS0�n i3tttfi's l -l e tax atp sl y�rver 7 pe cent.-and a sales tax rate of sltltLy over 6 percent:anl a marl rsity to consume taxable items of-SO:V scent,total lost4Ok talc'reymues,were$6:8 million. 10 1995 dollars,is 198:2 miio . 20 Davis-Bacon Repeal Effects unemployment varies by state and year,some of the difference in earnings might be because of variations in the unemployment rate. Last, construction wages vary by region for reasons that are not directly due to the presence or absence of prevailing wage laws. These regional differences in earnings, unemployment, and long-term trends in wages can be accounted for by using linear regression analysis. Regression Analysis of the Decline of Construction Worker Earnings Using linear regression analysis, this section uses U.S. Department of Labor employment and earnings data for construction workers broken down by states for 1975-91 to re-estimate the construction earnings loss resulting from state repeals of prevailing wage laws.The analysis controls for long-term trends in wages,variations in unemployment,and variation in wages by region of the country,and then focuses on the effect of(1)never having had a prevailing wage law,(2)repealing a prevailing wage law, and(3)raising the threshold for implementing a state prevailing wage law to contracts worth$500,000 or more. U.S. Department of Labor employment and earnings data provide detailed information on annual construction earnings broken down by year,state,and type of construction contractor.S2 For 1975-91, there are 27,778 separate observations. The inclusion in these data of information about prevailing-wage law status by state and year and translation of all money values into 1991 dollars (using the consumer price index)allows us to test for(1)the effect that never having had a prevailing wage law has on per capita construction earnings,(2)the effect on individual earnings of repealing a state prevailing wage law, and (3) the effect on individual earnings of raising the threshold for applying a prevailing wage law. In this test, we control for regional differences in construction earnings, secular trends in earnings,13 cyclical variations in earnings as a result of variations in unemployment,and differences in earnings by detailed contractor type.sa The data used for this test include average earnings across all states, years, and construction trades—$26,645 per year in 1991 dollars(table 2.2).55 States that never had a prevailing wage law account for 15.6 percent of all the observations. States that repealed their laws account for 10.5 percent of all observations after repeal and 7.8 percent of all observations before they repealed their laws, for a combined total of 18.3 percent. States that had and retained their prevailing wage laws between 1975 and 1991 account for the remaining 66.1 percent of all observations in the data set. Maryland and Oklahoma,the states with prevailing wage laws but with threshold levels of projects costing$500,000 or more,account for 4 percent of all observations. State-by-state unemployment rates in this period averaged 6.76 percent annually. The results of this regression model estimating the effects of state repeals on construction earnings are statistically significant and the overall model has a goodness of fit of 73 percent,which means that 73 percent of the overall variation in annual earnings in the data set are explained by the model. The results may be read as follows (see table 2.3). Begin with a constant amount of annual earnings of$33,005.(This is a starting point calculated by the regression model and is typically called the"constant.")Then select a state and a year.In any state for any year we know the status of prevailing wage laws for construction. We use Utah as an example in column (3). Utah once had a prevailing wage law, but, by 1991, that law had been repealed. Furthermore, 1991 was 17 years after the beginning of the data set Davis-Bacon Repeal Effects 21 Source: U.S. Bureau of Labor Statistics,Office of Earnings and Employment Statistics. Table 2.2 A description of the data used in regression model of earnings decline Observations 27,778 Average Earnings $26,645 Variables Percentage of data Percent of all observations by region South 29.9% Midwest 13.9% Atlantic 10.5% Mountain 10.5% Corn Belt 10.3% Pacific 8.8% New England 8.2% Hawaii 1.1% Alaska 0.4% Other control variables Average state unemployment 6.76% rate Percentage of states with 4.3% threshold for applying state law of more than$500,000 Legal variables Percentage of all states that 18.3% are repeal states Percentage of all states that 15.6% never had state law 22 Davis- eal Effects Bacon Rep Table 2.3 A regression estimate of the effects of state repeals on construction annual earnings, controlling for regional differences in earnings,and secular and cyclical trends in earnings Regression Model Examples for 1991 Variables and Coefficients in 1991 Dollars Utah Maryland Geor is 1 2 3 4 5 Starting Point: $33,005 $33,005 $33,005 $33,005 Regional Control Variables: Alaska $15,628 Hawaii $7,982 Midwest $4,768 Pacific $4,638 Atlantic $4,617 $4,617 New England $1,545 Corn Belt $1,010 Mountain -$79 -$79 South -$2,360 42,360 Trend Control Variables: Secular Trend -$225 -$3,829 43,829 -$3,829 Unemployment -$30,231 $1,481 -$1,784 -$1,512 Focus on Legal Variables: Never Had Law $2,960 -$2,960 Repeal $1,350 -$1,350 Threshold$500,000 -$1,174 -$1,174 Predicted Income: $26,266 $30,836 $22,345 Table 2.3- GoutrQliiag'for re giu ai 4"Mreuces in construction anuoat earnings, and secular trends and,cycii a variatious4n earnings,repeals in 9 states lowered construction earnings by$1,350 annually in 19"dollars. Having a threshold of $500 f000 for applying themstate f v_had.Amast the same effect as a repeal but this is based-on the expeiencef o�riy two states. ever hay .lad state prealing wage laws has al €�s :dauh t e. egai effect rn-ea i s}eo pa edto having; recently repealed the law ,.�' S gg��ts �t tf uegae�efl�ects� f°repea`<s o� earnings tnaynot have.fnlly matured by 1991,the end`of our data series: Davis-Bacon Repeal Effects 23 and thus,the time variable is set at 17 and Utah is in the mountain states region.Set all other regional variables to zero and multiply the mountain states control coefficient by 1.Multiply the secular trend control variable by 17 because this is the seventeenth year of the data set. Multiply P Y the unemployment control by 4.9%because that was the unemployment rate in Utah in 1991. Set the "never had law"variable to zero because Utah did have a prevailing wage law up to 1981 and set the threshold variable to zero,because in 1991 Utah did not have a prevailing wage law(and even when it did,the threshold was below$500,000).Now,set the repeal variable to 1 and multiply it times the repeal coefficient. Thus, the model now predicts Utah's 1991 construction income to be$26,266. That is$33,005 (the starting point)minus $79 (lower wages in the mountain states)minus$3,829 (secular down trend in real wages) minus $1,481 (associated with unemployment) minus $1,350 (because of Utah's prevailing-wage law repeal). The same exercise yields a predicted income of $30,836 for Maryland in 1991 and$22,345 for Georgia in 1991.Change the year and/or the state and the model predictions change.The W statistic of 73 percent indicates that the model fits the data well and that the predicted values are close to the actual earnings in the various states for the various years.S6 Controlling for all these variables,the model estimates that the effect of the repeal of the nine state prevailing wage laws was a negative$1,350 annual hit on construction earnings.Given average annual earnings of$26,645,this means a decline in earnings of 5.1 percent. This is a low estimate of a repeal's effect on earnings. The effect of a repeal may accumulate with time. The states that never had prevailing wage laws in construction have lower construction wages-after controlling for regional differences in wages and differences in unemployment rates.The model estimates that, in the nine repeal states, construction earnings are$2,960 less than in other states, controlling for other factors.This is an 11 percent reduction in construction earnings associated with never having had a prevailing wage law. The simple procedure in the previous section which compares construction earnings in repeal states before and after repeals estimates the repeal effect to have a 7.5 percent negative effect on earnings. Thus, the range of estimated effects varies from 5.1 percent to 7.5 percent to an 11 percent decline in construction earnings associated with the repeal or absence of prevailing wage laws.57 Increased Employment Associated with Lower Wages As construction labor becomes cheaper, contractors may alter their crew mix to use more workers who are unskilled. Have the nine state repeals of prevailing wage laws generated higher levels of employment? Construction employment varies markedly with seasonal and cyclical trends in the economy. These employment swin s can hide the effect of more jobs generated by falling wages. g For instance,Utah repealed its prevailing wage law just as the construction economy was going into recession.On the surface,it looked like the repeal and wage cuts did not generate more construction employment.Multivariate linear regression analysis can control for these variations and pick out the g potentially hidden effect of a repeal,controlling for other factors. Table 2.4 presents the results of a generalized least-squares regression test of the hypothesis that, as construction earnings fall,all other things being equal,construction employment will rise. The model controls for variations in unemployment,secular trends in employment construction,and any nonwage effect on employment associated with the repeal of a state prevailing wage law. The focus variable in the model is average annual earnings in construction and the hypothesis is that the 24 Davis-Bacon Repeal Effects negative.relationship between earnings and employment should be g ative. As earnings g go down employment might well go up. The regression model also includes control(dummy)variables for each state and each detailed industry classification(four-digit SIC;such as,plumbers and pipe fitters, SIC 1711). Thus, the model predicts construction employment in specific states, years, and each construction subclassification, such as plumbing and pipe fitting. In the data set for 1975-91, the average employment in a four-digit subclassification is 3,540 construction workers. The unemployment rate,not surprisingly,negatively affects construction employment and there is a small but statistically significant upward trend in employment.The effect of prevailing wage rate repeals on employment is negative, but this variable is not statistically significant which means the true direct effect of repeals on employment is zero. However,the indirect effect of state repeals on employment working through lower earnings is not zero. The effect of earnings on employment is as theoretically expected. As earnings fall, employment increases and this estimated effect is statistically significant.From this relationship,we can estimate the indirect effect of state prevailing wage laws on employment through the repeals' effects on earnings. Possible employment effects may be calculated for various levels of earnings decline.In table 2.5 column(1)presents hypothetical earnings declines and,in column(2),the results from table 2.4 are used to calculate a predicted increase in the construction industry when it is analyzed at the detail of 4-digit SIC codes(such as plumbers and pipe fitters, SIC 1711).As average annual construction earnings fall from a loss of$500 to a loss of$3,000,employment in given SIC industry groups rises from 24 new workers to 118 new workers.58 Given an average employment size of a 4-digit-SIC industry group of 3,540,these hypothetical increases in employment translated in percentage terms to an increase of from 0.7 percent when earnings fall by $500 to an employment increase of 4.0 percent when earnings in construction fall b 3 P g Y$ ,000. The Net Effect of Repeals on Government Budgets The overall effect of state repeals of prevailing wage laws on state expenditures in construction and state tax revenues will depend on the amounts of government cost savings from such a repeal and lost tax revenues from a repeal. Government construction cost savings will depend on three questions: how much lower are wage costs after a repeal,how much lower is worker productivity at lower wages,and how much construction work does the government purchase?Lost tax revenues will depend on(1)the marginal income tax rate for construction workers earning$20,000 to$40,000 per year, (2) the sales tax rate, (3) the marginal propensity to consume taxable commodities for construction workers earning$20,000 to$40,000 per year,(4)lost per-capita construction income associated with a repeal, and (5) gained construction employment associated with a repeal. (The $20,000 to$40,000 range encompasses most construction workers.) Previous estimates of construction cost savings associated with a hypothetical repeal of the federal Davis-Bacon Act range from 1 to 11 percent.59 The Congressional Budget Office favors an estimate of a 1.5 percent cost savings associated with the wage effect plus a 0.2 percent cost savings because of paperwork associated with Davis-Bacon.60 The savings may be higher or lower. Davis-Bacon Repeal Effects 25 Table 2.4 Effects of wages on employment,controlling for state differences in employment, differences in the size of SIC groupings,the direct effects of repeals,and secular and cyclical trends 1 2 Starting point: 5,500 workers in each SIC group in state Unemployment rate Subtract 211 workers for each percentage- point rise in unemployment Secular trend Add 31 workers for each additional year Repeal Subtract 170 workers(not statistically significant) Average Earnings Subtract 47.1 workers per$1,000 increase in earnings Number of Observations 27,778 Avg.Employment in SIC Group 3,540 by State and Year Note:An example of a four-digit SIC(Standard Industrial Classification)group is plumbers and pipe fitters, SIC 1711. Controlling for state differences in construction employment,differences in the size of four-digit-SfC'groups(such as plumbing versus electrical),secular trends, and cyclical variatious in employment in,each,state-and the-direct effect of repeals on employment-a fall in earnings resuloug,frclm a.fall is wages braises rou _of 3 54fl workers,a employment iu co SIC nstruction, For an average-sue g p 26 Davis-Bacon Repeal Effects Table 2.5 Effects of construction earnings decline on employment for an average-sized detailed standard industrial classification of 3,540 workers per state Various Predicted Rise in Percentage Rise Hypothetical Employment Because of In Employment Earnings A Fall in Annual Because of a Fall Declines Construction Earnings in Earnings 1 2 3 -$500 24 0.7% -$1,000 47 1.3% -$1,500 71 2.0% $2,000 94 2.7% -$2,500 118 3.3% s repeals force a fall in u► �? Y©n._W ges anal earni gs, nstruc on employment rises. The;m©del.Yn table 2:4``indicates that a Q fall 4iu eirn nes; results.in a 0.7 percent rise in_employment; An average ann op ii► earniggs would result in a 4,p an. "inelastic" demand for labor- ,,fie c�� thA1t e�r���{�s�� e��s� rs�antially higher t}ian therresuiting perc±en�*.rt a in M- (4,,irient(for the 4 t SIC gro�xp). This means tltat:evengl ;hgln? nt rsebB. uage fall,the:rise in employment is relatively small compiretl io the fall in ruagesA Consequently, Davis-Bacon Repeal Effects 27 n w will simply accept all ranges of hypothetical or estimated The effect an Utah.In this section, e p y p g yp savings rates from 1 to 11 percent in order to examine our model of lost tax revenues as it applies to Utah(see table 2.6). Rows 1 through 10 of table 2.6 provide half of the information needed to calculate the net effect on Utah's budget balances associated with the repeal of Utah's prevailing wage law in construction. Row 2 shows the level of employment in construction in Utah for 1987 to 1993. Taking from our regression model the value of lost income associated with a repeal of a state prevailing wage law(- $1,350)and translating that into 1994 dollars,using the consumer price index(41,477),we multiply this lost income times the level of construction employment in Utah for each year.This lost income associated with a repeal,denominated in 1994 dollars, is shown in row 3. Row 4 shows the gained amount of employment associated with a fall in construction wages and earnings because of a repeal. Row 5 shows average construction worker income in each year(in 1994 dollars).Row 6 shows the gained income due to additional workers shown in row 4 multiplied by average construction worker income in row 5. Row 7 reports the difference between GROSS lost income due to lower earnings and gained income due to lower wages. This net lost income is the source of the lost income tax revenues reported in row 8. Utah's income tax rate is flat at 7.2 percent above modest exemptions and deductions. Utah's sales tax rate is 6.25 percent. For construction workers, it is conservative to assume an 80 percent marginal propensity to consume locally on items subject to sales tax. This means that as a construction worker's income rises by $1,000, that worker will spend $800 on local commodities subject to state sales taxes. This allows for 20 percent of additional income to go to savings or purchases not subject to sales taxes. (Food purchases are subject to sales taxes in Utah.) Row 9 reports lost sales tax revenues as a result of net lost income reported in row 7.Row 10 combines lost income and sales tax revenues. Rows 12 and 13 report in 1994 dollars the value of building and road construction in Utah not covered by the federal Davis-Bacon Act. Roughly 20 percent of road work in Utah is not covered by the federal prevailing wage law. Rows 16 through 21 calculate, again in 1994 dollars, hypothetical levels of construction cost savings associated with Utah's repeal of its prevailing wage law.These hypothetical savings range from 1 to 11 percent of total construction costs.Rows 23 to 28 subtract lost tax revenues from construction cost savings for the various hypothetical levels of cost savings. Rows 23 to 28 show that in Utah, at total construction cost savings of below 3 percent, the repeal of the state's prevailing wage law tended to increase state finance deficits. The loss in tax revenues associated with lost construction worker earnings exceeded likely gains in construction cost savings.At and above 5 percent in total construction cost savings,the repeal helped tip the balance of state finances into the surplus.Using the Congressional Budget Office's estimate of a 1.5 percent increase in construction cost savings plus 0.2 percent in paperwork,the state of Utah would have lost more in tax revenues than it gained in construction cost savings every year since it repealed its prevailing wage law in 1981. The likely effect of a Davis-Bacon repeal on federal budgets.For construction workers earning $20,000 to $40,000, federal marginal income tax rates range from 16 to 28 percent. There are no widely significant federal sales taxes. With these changes in mind, and using federal data for construction employment,we can use the above model to estimate the tax revenue effects of a repeal of Davis-Bacon(table 2.7). 28 Davis-Bacon Repeal Effects Table 2.6 The relation of hypothetical construction-cost savings to tax revenues I Year 1987 1988 1989 1990 loot 1992 1993 2 Employment 26676 24981 25868 27836 31528 34902 39715 3 Lost income ($39,397,044) ($36,893,746) ($38,203,731) ($41,110,216) ($46,562,828) $(51,545,795) ($58,653,981) 4 Gained 478 447 463 498 564 625 711 Employment 5 Income $26,206 $26,329 $25,940 $25,213 $25,166 $23,933 $23,041 6 Gained $12,513,453 $11,773,180 $12,011,379 $12,562,530 $14,202,408 $14,952,327 $16,379,981 Income 7 Net Lost ($26,883,591) ($25,120,566) ($26,192,352) ($28,547,686) ($32,360,421) ($36,593,468) ($42,274,000) Income 8 Lost Income ($1,881,851) ($1,758,440) ($1,833,465) ($1,998,338) ($2,265,229) ($2,561,543) ($2,959,180) Tax 9 Lost Sales ($1,344,180) ($1,256,028) ($1,309,618) ($1,427,384) ($1,618,021) ($1,829,673) ($2,113,700) Taxes 10 Total Lost ($3,226,031) ($3,014,468) ($3,143,082) ($3,425,722) ($3,883,250) ($4,391,216) ($5,072,880) Taxes I I Value of State-Financed Construction 12 Buildings $94,436,620 $78,089,603 $93,725,806 $78,661,056 $87,518,355 $108,325,018 $118,790,378 13 Roads $21,117,077 $9,824,176 $17,183,065 $11,970,161 $27,677,680 $14,337,135 $13,824,742 14 Total $115,553,697 $87,913,779 $110 908 871 $90,631,217 $115,196,035 $122,662 153 $132,615,120 15 Hypothetical Savings Constmction Costs 16 1% $1,155,537 $879,138 $1,109,089 $906,312 $1,151,960 $1,226,622 $1,326,151 17 3% $3,466,611 $2,637,413 $3,327,266 $2,718,936 $3,455,881 $3,679,865 $3,978,454 18 5% $5,777,685 S4,395,689 $5,545,444 $4,531,561 $5,759,802 $6,133,108 $6,630,756 19 7% $8,088,759 $6,153,964 $7,763,621 $6,344,185 $8,063,722 $8,586,351 $9,283,058 20 9% $10,399,833 $7,912,240 $9,981,798 $8,156,809 $10,367,643 $11,039,594 $11,935,361 21 11% $12,710,907 $9,670,516 $12,199,976 $9,969,434 $12,671,564 $13,492,837 $14,587,663 22 Net Gain or Loss in Tax Revenues 23 1% ($2,070,494) ($2,135,330) ($2,033,994) ($2,519,410) ($2,731,290) ($3,164,595) ($3,746,729) 24 3% $240,580 ($377,055) $184,184 ($706,786) ($427,369) ($711,352) ($1,094,426) 25 5% $2,551,654 $1,381,221 $2,402,361 $1,105,839 $1,876,551 $1,741,891 $1,557,876 26 7% $4,862,728 $3,139,497 $4,620,539 $2,918,463 $4,180,472 $4,195,134 $4,210,178 27 9% $7,173,802 $4,897,772 $6,838,716 $4,731,087 $6,484,393 $6,648,378 $6,862,481 28 11% $9,484,876 $6,656,048 $9,056,894 $6,543,712 $8,788,313 $9,101,621 $9,514,783 Davis-Bacon Repeal Effects 29 Table 2.7 Projected effect of a repeal of Davis-Bacon on the federal budget 1 Employment 6,000,000 2 Lost Income m to meot*$1,477) SS 2,000,000 3 Gained Employment 107,400 (Em to went*1.0179 4 Avg.Income in 1994 $27,373 5 Gained Income from New Employment $2 939,829,040 6 Net Lost Income $5,922,170,960 7 Lost Income Tax at Various Marginal Income Tax Rates 8 16%Marginal Rate $947,547,354 9 20%Marginal Rate $1,184,434,192 10 28%Marginal Rate $1 658 207,869 11 Value of Federal Construction $11,528,571,429 12 Hypothetical Savings in Construction 13 1% $115,285,714 14 3% $345,857,143 15 5% $576,428,571 16 7% $807,000,000 17 9% $1,037,571,429 18 11% $1,268142,857 19 Net Gain(Loss)in Budget 16%Marginal Rate 20%Marginal Rate 28%Marginal Rate 20 1% S832,261,639 $1,069,148 478 $1,542,922,154 21 3% ($601,690,211) ($838,577,049 51,312 350 726 22 5% $371,118,782) 5608,005,621 ($1,081,779,297) 23 7% ($140,547,354) ($377,434,192) ($851,207,869) 24 9% $90,024,075 S146,862,763) (S620,636,440 25 11% $320,595,504 $83,708,665 $390,065,012 With an employment level of 6 million construction workers and an average annual earning of$27,000,the lost income from lower wages exceeds the.gained income from increased employment This results in differing values of lost income tax revenues depending on the assumed marginal tax rate. With a value for.federal construction of$11.5 billion,the hypothetical savings on construction from axe-peal depends on-the assumed cost-sa Jones rate. At a marginal income tag rate of 16 percent,net budgetary savings.froma repeal occur oniy with cons#ructioncast savings rates above 5 percent. At a 20 percent marginal tax rate, net budgetary savings from a repeal occur only with construction cost savings rates above 9 percent. At a 28 percent marginal tax rate, net budgetary savings from a repeal never occur within the range of cost savings between 1 and 11;percent. In short,a repeal of the Davis-Bacon Act will hurt the federal budget deficit. 30 Davis-Bacon Repeal Effects There are approximately 6 million construction workers in the United States.61 Table 2.7,row 2 shows what would have been the loss in income that these construction workers would have experienced given the 1994 value(-$1,477)of our regression estimate of the effect of state repeals on construction income.Row 3 presents an estimate of increased national construction employment associated with lower wages. Row 4 presents average annual income for construction workers in 1994. Row 5 multiplies gained employment in row 3 times average income in row 4 to obtain the increase in total construction workers'income associated with a hypothetical repeal of the Davis- Bacon Act.Row 6 subtracts gained workers'income from new employment from lost income as a result of lower wages to yield net lost worker income resulting from a hypothetical repeal.Rows 8 through 10 present lost income tax revenues due to net lost income at three marginal tax rates of 16, 20 and 28 percent. In fiscal year 1990-91, the federal government spent $10.491 billion on construction.62 Row 11 presents this sum in 1994 dollars. Rows 13 through 18 present levels of hypothetical savings in construction costs associated with a repeal of Davis-Bacon. Recall that the Congressional Budget Office estimates total the savings to be 1.7 percent,but others have presented savings estimates between 0.5 percent and 11 percent.Rows 20 through 25 present the net effect on the federal budget of hypothetical construction cost savings at various projected rates minus tax revenue losses at various marginal tax rates. Rows 20 through 25 show that only at very low marginal tax rates and very high construction cost savings rates does the federal budget benefit from a repeal of Davis-Bacon. At a marginal tax rate of 20 percent and a construction cost savings rate of 3 percent,the federal budget loses$838 million annually in 1994 dollars based on the 1991 level of federal government expenditures on construction. Summary In Utah, the repeal of the state prevailing wage law led to an overheated bidding process which added uncertainty to the cost of state construction.In the decade before the repeal,cost overruns on state-financed road construction averaged 2 percent of accepted bids.In the decade after the repeal, average road construction cost overruns rose to 7 percent of the accepted bid. A closer inspection of the data showed that, after repeal,contractors tended to present bids at a lower percentage of the state engineer's estimate of project costs but that,after change orders,the projects ended up costing the state a higher percentage of the state engineer's project cost estimate than in the decade prior to repeal.After the Utah repeal,contractors shaved their bids to get state jobs and more than made up for low-ball bids with subsequent change orders. This caused the increased cost overruns. An econometric analysis controlling for variations in regional differences in construction earnings,variations in unemployment rates,and general trends in real earnings showed that the nine state repeals'effects on earnings was a loss of$1,477 in 1994 dollars. Econometric modeling also showed that construction employment rose in repeal states after repeal by about 1.7 percent. This employment increase appeared controlling for variations in unemployment and long-term trends in construction employment growth. Thus, in assessing the budget effect of repeals of prevailing wage.laws, we are able to do two things.First,balancing the overall loss of construction worker income resulting from lower average earnings against the overall gain in construction worker income resulting from higher construction employment, we are able to estimate the change in overall construction worker income and consequently the change in government tax revenues resulting from these repeals. Second, taking a very wide range of hypothetical construction cost savings,we are able to estimate the net gain or Davis-Bacon Repeal Effects 31 loss to government budgets associated with repeals. In Utah, given its structure of income and sales taxes,the state budget would benefit from its repeal of the prevailing wage law if construction cost savings were at or above 3 percent. At the Congressional Budget Office estimate of a 1.7 percent construction cost savings (including paperwork costs),the state of Utah's budget has annually lost money as a result of the repeal every year since the repeal. Whether the state budget has gained or lost from it repeal is an open question. It is certain that Utah construction workers have lost income,not only on public works employment but across the construction labor market. At the federal level, construction cost savings must be substantially higher to generate any budget benefit from a repeal of the Davis-Bacon Act because of the federal income tax structure.At the more conservative estimate of 3 percent construction cost savings with a 20 percent marginal tax rate and the 1991 level of federal construction spending(in 1994 dollars),the federal government would lose$838 million per year by repealing the Davis-Bacon Act. The justification often given for repealing the Davis-Bacon Act is that a repeal would help cut the federal deficit. That is incorrect. A repeal of Davis-Bacon would help raise the federal budget deficit.This is because the purpose and effect of a repeal is to lower the cost of wages on federally funded construction projects.But lower wages and earnings will not be isolated to federally financed public works. Earnings would decline across the entire construction labor market and the government would lose more in income tax revenues than it will gain in construction cost savings. 32 Davis-Bacon Repeal Effects III. The Effect of State Repeals of Prevailing Wage Laws on Training and Minority Participation in Training This chapter presents a case study of the effects of the repeal in 1981 of Utah's prevailing wage law The Utah repeal accelerated the decline in the unionization,construction earnings, and trainin p on umo , g g union share of the state's construction labor market,drove down average construction wages in the state,and decreased union apprenticeship training for construction.No public or private source has offset the decline in training. In response to the decline in union membership and training, contractors have reduced turnover in order to retain skilled workers and to minimize screening and training costs. In response not only to the decline in construction wages but also to the coincident decline in health and pension benefits,however,experienced construction workers are leaving their trades for careers in other industries. Thus, while construction firm turnover is on the decline, turnover in the industry is on the rise.63 This chapter examines also whether the Utah experience in training can be generalized to the eight other states that have repealed their prevailing wage laws in construction.The U.S.Department of Labor Bureau of Apprenticeship Training keeps state-by-state records on registered union and non-union apprenticeship programs in construction. These records suggest that what happened in Utah is typical of what has happened in other states after repeal of their prevailing wage laws. The ratio of apprentices to journeymen in construction is higher in states that retain their prevailing wage laws compared with states that never had such a law.The rate of apprenticeship training in states that repealed their prevailing wage laws was substantially higher before the repeal compared with after the repeal. This remains true even when one controls for regional differences in training rates,the effect of unemployment, and long-term trends in training. There are not many minority workers in Utah in construction, but nationally there are. ("Minority"here refers to nonwhites,male and female.)Some have argued that prevailing wage law repeals will open j ob opportunities for unskilled minority workers and lower the unemployment rate of minorities,relative to whites. However,there is no evidence to support this claim. Black-white unemployment ratios rose in repeal states after repeals. Black-white unemployment ratios tend to be slightly higher in states that have never had prevailing wage laws compared to states that have retained their laws. While repealing prevailing wage laws probably has not caused black-white unemployment ratios to go up,there is no evidence to suggest that a repeal of the Davis-Bacon Act would cause black-white unemployment ratios to decline. The repeal of prevailing wage laws has especially hurt the training of minorities. There are proportionately more minorities trained as construction apprentices in states that retain their prevailing wage laws compared with states that have never had such laws. In repeal states, the proportion of minorities trained in construction apprenticeship programs declines substantially after the repeals. This remains true after controlling for regional differences in relative training rates, unemployment,and long-term trends in minority training which are independent of state repeals of prevailing wage laws. The decline in minority participation in construction apprenticeships after repeal is tied to a decline in unionization.Union apprenticeship programs tend to be large.Apprenticeship coordinators move apprentices from contractor to contractor in order to broaden the experiences ofthe apprentice. Typically,because non-union apprenticeship programs tie the apprentice to one contractor,the non- union programs tend to be small, single-firm programs, as opposed to larger, 33 Davis-Bacon Repeal Effects joint programs.At the same time,affirmative action regulation of apprenticeship programs applies only to programs having five or more apprentices.With the repeal of prevailing wage laws,not only does formal apprenticeship training decline,but also remaining apprentices are found more often in smaller apprenticeship programs.Thus,one effect of state repeals of prevailing wage laws has been to move more apprenticeship training out from under the oversight of affirmative action regulation. The result has been a substantial decline in minority participation in the remaining apprenticeship training. The Effect of Repeal on Construction Unions and Wages When Utah repealed its prevailing wage law in construction,wages became a focus of competition between contractors bidding on state jobs. Many union contractors went non-union or double- breasted(with union and non-union subsidiaries) to match or beat the lower wages of non-union contractors, and other union contractors lost market share. Because construction employment was falling,many union members went non-union with their traditional employers to stay employed. The vice president of a large industrial and commercial general contracting firm in Utah noted that, after the repeal, There were a lot of union workers that carried their card in their shoe. They worked open shop until a union job came available.A lot of folks all of a sudden started to find homes over there[in the open shop]and never came back(personal interview,May 15, 1993). Consequently, in the short-run, at least, contractors that remained union did not have a significant labor productivity advantage over many of the newly non-union contractors. This effectively forced remaining union contractors out of much of the construction market. With the decline of union contractors,Utah construction union membership fell(fig.3.1).64 The decline in membership was accelerated by the 1982 recession. Union membership appeared to recover from the recession,but many dues-paying members were working open shop.With the onset of the next downturn in Utah construction in 1986,union membership began to fall steadily.These data are consistent with the story that union members working in the open shop eventually found a home there and quit paying their union dues. With the repeal of the prevailing wage law and the resulting decline in unionization in Utah, average wages in construction fell relative to the average Utah wage(fig.3.2).Construction wages, which had ranged from 120 to 125 percent of the average Utah wage before the construction boom of the 1970s, exceeded 130 percent during the boom. When construction employment growth stopped in the late 1970s,construction wages fell back toward the high end of their normal premium over average Utah wages. But with the repeal of the prevailing wage law, construction wages fell to a new lower range of 110 to 115 percent of the average wage in Utah.This is an across-the-board decline in construction wages and not isolated to union earnings nor the earnings of construction labor on public works. This relative decline in construction earnings in Utah is consistent with the overall decline in construction wages following repeal(chapter II). The data for Utah actually underestimate the effect of Utah's repeal on construction workers' earnings, in part because the data do not include the change in value of benefits. Davis-Bacon Repeal Effects 34 5500 Law Repealed in 1981 5000 ---- ------ -- --- --------- ------------------------------------------------------------------- m 4500 -- ------ -- ---------------------- ------------------------------------ c 0 4000 ------ ----------- - 0 1982 Recession 3500 ------------------------------------- ------------------------- - ------ ------------------------- c°� 3000 Union Members"Cary Their ------- ----------------- ------------------- ,� Cards in Their Shoes" 0 E2 2500 ------------- -- ------------------------- ---------- E 3 2000 Eventually, Members"Find a Home in the Open Shop" 1500 1977 1979 1981 1983 1985 1987 1989 Quarterly Membership Totals Figure 3.1 Union membership in construction in Utah, 1977-89 Source:Utah State Building and Construction Trades dues records. Union me tiers ip began,to dechrite with#lie pre �i g 1_a,, Nertbe oaset of tle=1982 recession. l�!�epreoere ` n as fast.as overall.gonstruction�-m tgext: Witli the 1985 downtarn: "Ok construction employment,union.inembership began a-steady-decrne Ae less than half its late-1970s peak. 35 Davis-Bacon Repeal Effects 135% — 40 Employment Stagnates � c 35 130/o --- --- -- - --- - - ------------ E ° --------------------------- t >, o CO i a 30 c 125% --------------- ---------------- ----------------------- -------- --- s 25 = c � 0 U 0 120% 1 i -- ------------ ---- > c I 20 0 Wages 0 115% ------ ------ ---- ---- ------- ---------- Fall 15 ' Q 110% 10 50 55 60 65 70 75 80 5 90 FLaw Repealed -�- Wages @%UT Median--I- Construction Employ Figure 3.2 Construction wages as a percentage of the Utah median wage and Utah construction employment Source: Utah Job Security,Division of Labor Market Information,Annual Report,table 5. Construction employment(measured in.theusal son;the rig-4t4hand Y axis)in Utah grew rapidly in the 1' 'T ,F1ugruw stpeda the:lhs"an+d eyelcul fluctuations became more p'rouomiced.Wages,(measu ed-as:a percentage of the median Utah wage)rabged l even 12f1 and 12 %of#WVOh median Wage prior to the construction boom of the 19'70s: These construction earnings rose above 130% of Utah's median wage income during the:boom. As the boom ended, construction wages moved down to.them normal range.With the repeal of Utah's prevailing wage laav in;190"wages plummeted. Davis-Bacon Repeal Effects 36 Typically,unionized construction workers receive better health and pension benefits than do non- unionized workers. Lower benefits, particularly health and pension benefits, contribute to the increase in overall labor turnover in and out of the construction industry in Utah. This increased occupational turnover,we will see,led to a younger,less trained,and less experienced labor force. The Relation between Repeals and Black Unemployment It has been argued that the Davis-Bacon Act was passed, in part, to restrict southern blacks from northern construction job opportunities. It is further claimed that the current high and rising ratio of black unemployment rates relative to white unemployment rates is partly due to restrictions that prevailing wage laws impose on the ability of unskilled black labor to compete with better skilled white labor. From these beliefs, it is argued that a repeal of the Davis-Bacon Act would lower black unemployment relative to white unemployment by opening up j obs for less-skilled black labor.65 These arguments are not directly supported by the available evidence. Black unemployment rates are separately collected for only five of the nine states that have repealed their state prevailing wage laws.Arizona,Idaho,New Hampshire,and Utah do not have large-enough black populations to generate meaningful unemployment statistics.However,Alabama,Colorado,Florida,Kansas,and Louisiana do have sufficient black populations to test the above argument. The ratio of black-to- white unemployment for five repeal states can be shown using state unemployment rates for white and blacks and white males and black males(fig. 3.3). In all cases, black unemployment rates are more than twice the rate of white unemployment. Before the repeal of state prevailing wage laws, however,the male black-to-white unemployment ratio and the overall black-to-white unemployment ratio were both less than their corresponding ratios after these states repealed their prevailing wage laws. This does not mean that the repeals caused the black-to-white unemployment ratios to rise. Black-to-white unemployment ratios were rising across the country in the 1980s in repeal states and elsewhere. The rise in the black-to-white unemployment ratios simply reflects this time trend.66 By comparing the states that retain their prevailing wage laws with those states that never had prevailing wage laws, we can eliminate the effect of time trends in black-to-white unemployment ratios.The black-to-white unemployment ratio and the male black-to-white unemployment ratio are both lower for states with prevailing wage laws compared to states without prevailing wage laws— averaging unemployment rates across states and years from 1974 to 1992 (fig.3.4).67 The male unemployment ratios in figure 3.4 are almost the same and statistically they are not different. In terms of employment,rather than unemployment,in 199014 percent of all persons employed in construction were minorities (here defined as nonwhites plus hispanics). In the 32 states which had prevailing wage laws, 14 percent of all construction workers were minority workers,and in the 9 states that had never had prevailing wage laws plus the 9 states which had repealed their laws, 14 percent of all construction workers were minority workers. In all states, minorities were under- represented in construction. The average minority population in states which had prevailing wage laws was 20 percent and the average minority population in states without prevailing wage laws in 1990 was 19 percent.68 Thus,minorities were under-represented in both state groupings. However, there is little here to suggest that repealing prevailing wage laws would ameliorate this under- representation. The construction employment prospects of minorities are quite similar in both states 37 Davis-Bacon Repeal Effects with and without prevailing wage laws regulating public construction. These data do not support the proposition that a repeal of the Davis-Bacon Act would ameliorate in any significant way the relative unemployment of blacks to whites. Davis-Bacon Repeal Effects 38 3 -------------- -------- o 2.5 2 a -------- --------- 0 Q m 1.5 C -------- -------- Z) m 1 ---- -------- m 0.5 After Repeals O-Z Before Repeals Males All Figure 3.3 The ratio of black-white unemployment in five repeal states before and after repeals Source:US DOL Geographical Profile of Employment and Unemployment 1974-92. Five repeal states-Alabama;Colorado,Florida,Kansas,and,Louisiana-have sufficient black populations to report a separate b aclz;unemplayment rate and a blaek,male unemploym e t TWO-1-'these five states, ft the decade prior to repeals, the-ratio of black to vczhite a cm loy nt rates was 2A& After repeals,the rant► rose to 2.61 which means black tinetnp eutswas oven 3-gtrer to relation to white unemployment. F—,tfia"les,t e-h ack=to� hi€te100001p? y dot ratio was 2.28 before repeals and 2.40-after'repeals. These ratios,are;based an unemployment rates for the entire state not simply construction. 'if repeals opened job opportunities for blacks,the effect is hidden. Black-white unemployment ratios rose throughout the 1980s and the rise is not due directly to the repeals. 39 Davis-Bacon Repeal Effects 3 ------------- --------- 0 2.5 -------- Cu E 2 -------- --------- 0 0- 1.5 m II -------- ------- Y U In 0.5 Never Had Law 0 Retained Law Males AID Figure 3.4 Black-white unemployment ratio for states that retained and never had state wage law Source:US DOL Geographical Profile of Employment and Unemployment 1974-92. Comparing the black-to-white unemployment ratio in states that retained`their state prevailing wage laws throughout the last 25 years with the ratio in those states that never had-state prevailing wage laws elinAnates the effect of a strong time trend-that shows.up in,before-and-after analysis. The male black to-white unemployment ratio is,sljghtiy higher in the states.that never had,prevailing wage laws compared with states that retained theirs.The" he difference'is-not statistically significant. The overate black-to-white unemployment ratio is significantly greater in the states that never hail a prevailing:Wage law,but this is because of: female unemployment differentials,which-are unlikely to be significantty affected by.construction eniployment patterns, Davis-Bacon Repeat Effects 40 A Decline in Training With the decline in union membership and in relative wages, training for construction in union apprenticeships and through vocational schools,declined in Utah.Union apprenticeships are tied to the availability of union jobs.For instance,unionized plumbers and pipe fitters in Utah,the United Association of Journeymen and Apprentices of the Plumbing and Pipe Fitting Industry of the United States and Canada,historically have attempted to maintain apprenticeship rates at 10 to 15 percent of the number of union journeymen plumbers in the state(fig. 3.5). As employment boomed in the 1970s, however, the union could not meet the demand for journeymen from union contractors. Consequently,the union increased apprenticeship rates to a peak of 25 percent in 1975.The boom persisted,but the backlog had been remedied. So the union lowered its apprenticeship rate back to normal ranges by 1978.Employment during the construction boom peaked in 1979 and membership in the plumbers and pipefitters' union peaked in 1981. With the repeal of the Utah prevailing wage law,the union dropped its apprenticeship rate to 10 percent,a historical low.Union membership fell slightly in 1982 and began a steeper decline in 1983.Faced with these sustained declines in membership,the union cut its apprenticeship rate even lower in 1986 and thereafter. Unions hit harder by declines in membership have scaled back their apprenticeship programs further. The carpenters'union, Utah locals 184 and 1498 of the United Brotherhood of Carpenters and Joiners of America, which graduated seventy in a class in 1977, graduated five in 1992. The Utah International Union of Bricklayers and Allied Craftsmen suspended its apprenticeship program altogether. The decline in union apprenticeship training in Utah has not been offset by a rise in other 0 sources of training. Because the repeal of Utah's prevailing wage law was motivated by a desire to limit state expenditures, state legislators were not eager to raise funding for state-sponsored vocational training. Although the number ofvocational graduates in construction grew in the 1970s,the construction labor force grew more rapidly.Thus,while the 1970s was the heyday of vocational training at Salt Lake Community College,vocational graduates as a percentage of the construction labor force had already begun to decline" The steady decline in state-supported vocational training as a percentage of the construction labor force through good times and bad supports the notion that the state has simply tried to get out of the business of vocational training in construction.The fall in union membership and wages has made construction a less attractive career.At the same time,unions are less able to train construction workers. As unions are weakened and community colleges drift toward academic offerings, the capacity to respond smoothly to an upsurge in construction jobs is undercut.And federally sponsored Job Corps vocational training is not in a position to fill in the gap. Federal revenues pay for Job Corps training in Utah at the Weber Basin and Clearfield centers. Federal funding in real terms for these centers has not expanded, but the Weber Basin Job Corps Center,which draws predominantly from the Utah population,has significantly cut its construction worker training throughout the 1980s. This center committed itself to changing from an all-male student population in 1980 to 50 percent female by 1990.To accommodate this switch, training for traditionally male occupations such as construction,have been scaled back 41 Davis-Bacon Repeal Effects 35% 1300 Membership Peaks in 1981 30% ----------------------------------------- ------ --------------------------------- 1200 a) E E c 2,9% Above Normal ----------------- -- ----------- -------- ---------------- 1100 a Training Rate E 0 6 20% ---------------------------------------- -------- --- --------------------------- 1000 ca o � 15% ------------------------------------- -------- --------- - ------------------------ 900 c 0 Cn a ------------ ----------------- m 10% ----- -------- 800 0 Q E o -------- ------------ ---- ----- 700 =1 Below Normal Training Rate 0 0 1961 1971 1981 1991 600 --�— Plumbers-number -- - Apprentices-percent Figure 3.5 Apprentice plumbers as a percentage of journeymen plumbers in Utah, 1961 to 1991 Source:Utah plumbers and pipe fitters locals' membership records. The plumbers' union in Utah has historically.attempted to train apprentices at a rate of 10 to 15 percent of their journeymen members. As employment boomed in the 19701s,the union could not meet journeyman:demand and consequently- expanded apprenticeship training rapidly. As the numbers of journeymen grew to meet demand, apprenticeship training was reduced to;normal rates. But:with the repeal of the state preailittg.wage iaw in 198i,upon mmnhersitip eclined and apprenticeship training rotesarere cut to all time lows. Davis-Bacon Repeal Effects 42 to accommodate new offerings in traditionally female occupations,such as office management and clerical work.Cement masonry and heavy-equipment training have been eliminated,and instruction in carpentry,painting, and brick laying has been cut in half The Clearfield Center has graduated approximately 100 construction trainees per year since the early 1970s.Fewer Clearfield graduates go into the Utah labor market,compared with Weber Basin graduates, because most of Clearfield's students are from out of state. Perhaps 10 percent of Clearfield'sgraduates go into the Utah labor market,but this percentage rises during periods of local labor shortage. It is estimated,however,that at most only 25 percent of Clearfield's graduates will stay in Utah. Even without union pressure, it is possible that a shortage of skilled construction workers in Utah will raise wages and induce a new generation of young people to enter construction vocational training for the industry.Nonetheless Utah is now in a building boom-when wages would normally rise- and annual earnings in construction relative to annual earnings for all Utahns continue to fall. In 1993,the most recent year for which data are available,the construction earnings premium fell to a historic new low of 103 percent of the average annual earnings for all non-agricultural workers in Utah.70 Utah is now in a building boom, one that has come quickly. High-quality training programs, which take time to create, are not in place to meet the demand. This adds an additional lag to the usual time it takes to train a skilled laborer. Utah's current boom has relied partly on using a less- skilled labor force(which partly accounts for the lower construction earnings premium)and partly on travelers from California,which is currently in a construction lull.Whether the Utah construction industry can rely,in the long run,on training systems for construction workers in California remains to be seen.A pick-up in California construction would quickly bleed away the skilled workers Utah is now attracting.This is one difference between state repeals of prevailing wage laws and a federal repeal of the Davis-Bacon Act. If construction cycles are not synchronized,it is at least possible,if a state is lucky, for one state to freely ride on the training systems of another state. A repeal of Davis-Bacon would create a nationwide decline in training. Under such a circumstance free riding on the training of another area would not be an option. Market Responses: Training, Turnover, and Careers The market in Utah has not successfully made up for the decline in union and state-sponsored training. At the national level, the non-union Associated Builders and Contractors (ABC) has attempted to replicate the union system of bargaining for hourly contributions to a training fund. It is difficult,however,to induce ABCs member contractors to include general training costs in their bids. Each contractor fears that his competitors will not include training costs. Thus, in an attempt to be the low-cost bidder, ABC contractors often refrain from including training costs despite the ABC initiative. Consequently, very little ABC training has occurred in Utah. In Utah, non-union apprenticeship programs operate, however, in the licensed trades of electricians and plumbers.In 1992,there were 846 non-union licensed apprentice electricians in Utah and 2,068 non-union journeymen.Thus,there are 4 apprentices for every 10 journeymen in the non- union sector.In contrast,there were 123 apprentices and 607 journeymen in the union sector in 1992, or 2 apprentices for every 10 journeymen. In the non-union sector,apprentices begin at around$6 per hour with no benefits. Over a four-year period,the state mandates that the apprentice wage rise 43 Davis-Bacon Repeal Effects to 80 percent of a journeyman's pay. In the union sector, apprentices begin at $7 per hour with an additional$3 in benefits. Their wages rise to$14 per hour plus$3 in benefits over five years.Non- union apprentices are sponsored by a particular contractor that oversees on-the-job training, and these apprentices take classwork at a participating community college.Union apprentices work under the direction of an apprenticeship coordinator,rotate among employers for on-the-job training,and take classes at community colleges and union apprenticeship centers. Roughly 90 to 95 percent of the union apprentices complete their programs and graduate to journeymen status,while only 15 to 20 percent of the non-union apprentices graduate. Given these rates, in four years,out of 846 non- union apprentices, we should expect 125 to 170 journeymen to be graduated. In five years in the union sector, out of 123 apprentices, 110 to 115 apprentices would graduate to journeymen electrician. Thus, while the non-union sector accounts for more than 85 percent of all electrician apprentices, it accounts for about 60 percent of journeymen graduates. Economic theory is consistent with this pattern wherein non-union apprentices are paid less and graduate at a lower rate than union apprentices. Economic theory posits that in the absence of marketwide institutions or government subsidies,individual workers will have to pay for their own on-the-job training when the skills learned are general to an industry and not specific and unique to the activities of a particular firm. The worker-learner pays for training by accepting a wage that is lower than the value to the firm of that worker's marginal product. By working for less than the worker's worth to the employer, the worker pays the employer for on-the-job training. That beginning non-union electrical apprentices earn$6 per hour while union apprentices earn $10 per hour(including benefits) is consistent with the theoretical proposition that non-union apprentices pay for their own training by taking a discounted wage below their marginal value to the contractor. Because the employer does not pay.much for non-union training,the theory suggests that the employer has no stake in the worker's training.If the worker leaves,the employer does not lose any investment in the worker's human capital. So, the employer will tolerate high levels of turnover. Because the worker is receiving less than what the worker can earn in other jobs with no on-the-job training,the worker may be tempted to exit jobs with training when current personal budget needs become pressing.So,on both the employer side and the worker side,turnover is tolerated in the non- union sector. This view is consistent with the higher turnover rates among non-union apprentices, but other factors also contribute to the roughly 70 percentage point differential in non-union to union graduation rates. Because the non-union employer prices new hands at discounted wages that shield the employer from investing in the human capital of new workers, the employer does not screen new workers extensively to forestall subsequent turnover. The employer's failure to preselect new workers for aptitudes and attitudes consistent with a long-term attachment to construction work adds to the turnover among non-union construction apprentices. In contrast,the joint apprenticeship boards of unions and union contractors do considerable preselection for aptitude and attitude before letting a candidate into an apprenticeship program. This is because the union contractors and unions will invest in the union apprentices' training." In the non-union sector,workers may also leave apprenticeships if it becomes apparent that the employer offering training at a discounted wage is not delivering on that training promise to train. Because employers are able to discount wages of apprentices below their current worth to the employer, it is tempting to engage in bait-and-switch tactics whereby training is promised but not delivered. By saving on training costs,the employer can earn an additional profit from employing Davis-Bacon Repeal Effects 44 green hands at discounted wages. In the union sector, because employers and union journeymen invest in the training of the apprentices, bait-and-switch tactics are less attractive. Because the apprentices'wage is not discounted as much below what they could earn elsewhere,the apprentices are not as tempted to leave. Thus, the non-union sector must begin training five apprentices to graduate one journeyman,while the ratio in the union sector is close to one to one. While non-union contractors tolerate high levels of turnover among apprentices, with the decline in training and union membership, non-union Utah contractors have sought to reduce the turnover among trained journeymen. There has been a long-term decline in labor turnover in construction(fig.3.6).This long-term decline can be explained with a pooled,cross-sectional,time- series linear regression model, as can the differences in turnover rates in Utah by contractor type from 1956 to 1991 (table 3.1).Not surprisingly,this model shows that turnover was higher in years in which variations in monthly construction employment were great.It also shows that contractors with larger crews tolerated proportionately more turnover.Contractors employing more-expensive labor sought to reduce turnover.When union membership was a high percentage of the construction labor force,turnover was higher simply because contractors losing one good worker could turn to the hiringhall for a reasonable substitute at little additional cost. When vocational schools were graduating a large number of construction-trained students relative to the Utah construction labor market, contractors tolerated more turnover because the market had proportionately more trained substitutes.The numbers of union membership and vocational graduates have been on the decline, however.Thus,this regression model shows that,over time,contractors have responded by reducing the turnover among journeymen . Although turnover at the firm level has been on the decline, workers may be entering and leaving construction at higher rates than 20 years ago. In 1970, Utah construction workers, on 1 g g 72 construction boom had be un in Utah,the 4 ears old. B 199 0 before the recentg average,were 2 y y , age had fallen to 33 years.73 Much of this decline may be due to the construction expansion in the 1970s,which brought in a new generation of younger workers. But the decline in age may also be a result of both the decline in health and retirement benefits and the decline in relative wages associated with the decline in unions. Although non-union contractors increasingly are providing health and retirement benefits, especially to their key people, the health benefits tend to be more expensive for a given level of care and the retirement 401K plans lack the insurance component associated with union-defined benefit plans. National Trends in Registered Apprenticeship Training The U.S. Department of Labor, Bureau of Apprenticeship Training, monitors registered apprenticeship programs- union and non-union-in the construction industry. Data are available for 1975-78 and 1987-90.Not all states have reported to the Bureau of Apprenticeship Training for all years during these periods.Nonetheless,29 states did report registered construction apprentices for every one of those years. The states included 6 states that eventually repealed 45 Davis-Bacon Repeal Effects 24%- 22%- ------ -- --- -------------------------------------------------------------------------------------------------- 20%- ---- - --- -- ----------------- ---- ----------------------------------------------------------------------- 18%- -- ---- - - -------- - - ---- -- - ----------- ------------------------------------------------- - - ------ -- - - - --- --- --- - ------------------------------------------------ 16%- ------------- 14%- ------- - ------- ------------- -------------------------- -- -- ----------------- -------------- --- (D 12%- ---------------------------------------------------------------- ---- --- 10%- --------------------------------------------------------------------------------- - -- ------ -- - --------- 8%- ,-- --------------------------------------------------------------------------------------------- --------------- ---------_---------------- -- 6%- --------------------------- -----------------------------�;�401,r 50 55 60 65 70 75 80 85 90 YEARS All Employment —0-- Contract Const. Building Heavy&Hwy —X— Specialty Figure 3.6 Turnover in Utah's construction industry compared with all employment statewide Source:Utah Job Security,Division of Labor Market Information Annual Report,table 5. As the number of trained journeymen in Union hiring halls declines and the number of non-union journeymen declines,firms respond by reducing firm turnover.Later,t will he shown that while firm turnover in Utah construction is declining,career turnover is on the rise. Davis-Bacon Repeal Effects 46 Table 3.1.Linear regression model of turnover rate in construction in Utah. Source:Utah Job Security,Annual Report,Table 5. Dependent variable=firm turnover in constructions Actual Standardized Variable Coefficient Coefficient Union Members` 1.76 •24 New Vocational Graduates` 2.45 .20 Real Wage -.076 -.62 Seasonality 2.12 •15 Workers per Contractor .052 .40 (Constant) -1.88 --------------------- - ------- a The actual variable is ln(turnover/(1-turnover))to meet the technical requirement in linear regressions of having an unbounded dependent variable. b All independent variables are statistically significant at the 1%level. °As a percent of the construction labor force. Adjusted R Square= .24 Number of Cases=351 Time Period= 1956 to 1991 Contractor Type=4 digit SIC Contractors in Utah:tolerate higher:labor turnover when union membership is a high percentage of the labor force,and when new vocational school graduates are plentiful. Turnover is more common in years when monthly employment fluctuates a lot. Contractors are more willing to tolerate turnover among lower paid workers and contrptors with larger work crews,mustaccept'higher levels of turnover. Standardized-coe€fctents incieate thaat worker skill and:crew size have the largest impact on variations*jjmjIqer turnover rates while both the availability of union members-and new vocational graduates have1arger effects than seasonal fluctuations ht employment. 47 Davis-Bacon Repeal Effects their prevailing wage laws,4 states that never had prevailing wage laws,and 19 states that retained a state prevailing wage law throughout the period.These 29 states can be divided into the categories "repeal," "never-had," and"retained-law," for comparison(figs.3.7 and 3.8).No state had repealed its prevailing wage law by 1978. By the end of the first quarter of 1987, all nine repeal states had passed their repeals except Louisiana which repealed in 1988.The data for 1987 are for the summer of 1987, after Kansas had repealed in that year.74 In the"before"period,states that had prevailing wage laws- those that retained such a law and those that had not yet repealed theirs-typically trained a higher percentage of registered apprentices than the states that never had a prevailing wage law. For unknown reasons, the year 1976 is an exception to this pattern. During this pre-repeal period,the states that would eventually repeal their laws had as high or higher training rates compared with the states that kept their laws throughout the period. By 1987, training rates had fallen for all states,but they had fallen least in states that had retained their prevailing wage laws.By 1989,the states that had repealed their prevailing wage laws had training rates as low as the states that never had prevailing wage laws.This is clear evidence that repealing state prevailing wage laws lowers formal apprenticeship training. A simple analysis can help isolate the effect on training of repealing state prevailing wage laws from a general downward trend in construction apprenticeship training.Apprenticeship training rates for states that repeal their prevailing wage laws in the late 1970s and 1980s are presented as a percentage of the training rates of states that retained their prevailing wage laws(table 3.2,col. 2). Throughout the 1970s,before repeals,the repeal states had training rates that were at or above the average training rates for states that had and would keep their prevailing wage laws.After the repeals in the late 1980s,the repeal states had training rates that fell to as little as 63 percent of the training rates of states that kept their prevailing wage laws. By 1990,the repeal states had relative training rates that were as low as the states that never had prevailing wage laws. Thus, while training in construction has been falling for all states,the fall for repeal states has been the most precipitous and - setting time trends aside- the repeal states matched the training rates of the retaining states prior to repeal and fell to the rates of states never having had prevailing wage laws after the repeal.75 Unlike the simple analysis just presented,however, a multiple linear regression analysis can control for other factors, such as differences in state unemployment rates or regional differences in training(table 3.3).The dependent variable in the analysis is a transformation of the training rate for each state, where the training rate is calculated as registered apprentices as a percentage of all construction employees in a state and year. For technical reasons associated with the assumptions of linear regression analysis,the actual dependent variable is the natural log of the odds ratio of the training rate where the odds ratio is calculated as (the percent trained) divided by (one minus the percent trained).76 In the regression model,regional differences in training rates are controlled for with the regions corresponding to standard Bureau of Labor Statistics regional categorizations. Unemployment differences are controlled for by state and year.The data are for the years 1975-78 and 1987-90.The focus variable is REPEAL, a dummy variable equalling 1 once a state repeals its prevailing wage law.A second focus variable is NEVERHAD which equals zero for all states except for those nine states that never had a state prevailing wage law in construction. For those states, NEVERHAD equals 1. There are 297 observations in the data set. California, Davis-Bacon Repeal Effects 48 5.50%_ a) 5.000/(0-------- ----------------------------------------- -------------------------------------------------- _�5 0 4.50%------------ ------ ---------------- _------- ------- -------- --------------------------------- 4.00%- ---- ----------------- ----------------- .... ........... ------------------ 0 After Repeals - - - ------------------------- ------------------ 3.50%- ---------------------------------- ------ ID cL3.00%- ---------------------------------------------- -- - --------------------------- -------------- -0 2.50%- --------- ------------------ -- - ------------------------------------------ 2 Before Repeals M2.00%- ------------------------------------------------- ---- . ........ ----- ---------- ----- m 1.50% ------- --------- ------------------------------ 1.500Y(O ------------------ 1.00 1 1 75 76 77 78 87 88 89 60 Year [� Repeal w Never Had E3 Retained Figure 3.7 Apprenticeship training rates,by state groups,before and after repeals Source: U.S.Department of Labor,Bureaus of Labor Statistics and Apprenticeship Training. This figure shows ap-preatic-tsas,,apercentage of all construction--workers,in,29 sta$esgro,u, l)e.&by, state freatinefit off-prevadiingwage,laaw.In the four years before the" ' that repeal,of-stalto Vreval W i0tvagellaws,states Ia I wouldeventuaRy,repeal their laws�,h,'qd Ajoiappeeati I-e'hipstra I imugerates. I States thatwould-T-et0i,Otheir Except lin"ll-'M 40M that Itait-pre construction had rehitivelylow-trainIng rates. never or- v-1a g--p'faw Vt-fift co In all state-groupifts;I-ruibibg rates in the late 1980s,were-lower than training; rates in the late 1970s. However,after the several state-repeals,those states that retained their prevailing wage laws had relatively higher training rates. Those states that repealed their prevailing wage laws eventually had training rates that matched the states that had never had prevailing wage laws. 49 Davis-Bacon Repeal Effects 5.0% - Before Repeal -----------•-------------------------------------------------------------------- 4.5% 0 Retain Law 4.0% 4.3% n=2a 3.5% 3.a°io Cn n=221 --------Never Had Law (j 3.0% 0 ------ ------ ------------------------------ 2 5% ° After Repeal U n�6 2.0% Q -- ..... Q„ n=39 ------ ------ a) 1.0% ------ ------ ------ 0.5% Repeal States: States which: States which: Repeal States: Years=1975 78 and 1987-90 Figure 3.8 Apprenticeship training rates,by state Source: U.S.Department of Labor Bureaus of Labor Statistics and Apprenticeship Training. States are grouped here into four categories,repeal states before and after their repeals of prevailing wage laws,states that retained-their prevailing wage laws, and states that never had prevailing wage laws.This simple pattern shows that repealing or not having prevailing wage laws reduces formal training in construction.(Part of,this before-and-after picture is due to an overall.downward> trend in registered apprenticeship rates in construction overtime) Repeals hurt apprenticeship training because repeals hart unions. Non-union construction contractors do less training and less formal,high quality training. Davis-Bacon Repeal Effects 50 Delaware,the District of Columbia,Hawaii,and Rhode Island are omitted from the analysis because they did not report to the Bureau of Apprenticeship Training of the U.S.Department of Labor during the second period of our analysis. The model is a good fit of the data with an adjusted R2 of 45 percent,and all variables are statistically significant. The focus variable in the regression analysis REPEAL-a marker for states that repealed their prevailing wage laws-is negative.This means that-controlling for unemployment,time trends and regional differences in training-when states repeal their prevailing wage laws,the training rate goes down.At the mean training rate for the entire data set of 3.7 percent,this model indicates that repeals drove down training rates to around 2.1 percent.The NEVERHAD variable,marking states that have never had a prevailing wage law, is also negative and statistically significant but smaller than the REPEAL variable. This is because of a close correlation(about 40 percent)between never having had a prevailing wage law and being a southern state. This means the analysis could not fully distinguish between the hypothesis that training rates in the South were low because many of these states never had prevailing wage laws and the hypothesis that other reasons associated with being a southern state caused training rates to be low.The REPEAL effect was easier to pick up compared to the NEVERHAD effect, simply because the repeal states presented information about their training rates before and after each state repealed its prevailing wage law. Thus, looking at training rates from a variety of measures and methods of analysis, it is clear that state repeals of prevailing wage laws have significantly lowered formal,organized,and quality training of construction workers. The effect is to lower training rates by about 40 percent. When apprenticeship training falls as a result of repeals of state prevailing wage laws,minority participation in apprenticeship programs falls even farther(fig.3.9). Minorities comprise almost 20 percent of all construction apprentices in the repeal states in the years before repeal of state prevailing wage laws. In the same states, after repeal of their prevailing wage laws, minority participation in apprenticeship programs falls to just under 13 percent of all apprentices. While construction apprenticeship training is falling in these states by around 40 percent, the share of minorities in this downsized training also falls by about 36 percent. One reason for the decline in minority training is the decline in union training. In figure 3.9,the share of minorities in apprenticeship training appears the same for states that retain their prevailing wage laws and states that never had such laws,but this is an illusion. Many of the states that have never adopted prevailing wage laws are in the South where there is a high percentage of minorities in the overall state population(fig. 3.10). We account for that factor with the ratio of the minority percentage in construction apprenticeship programs,divided by the minority percentage in the state population.This ratio is 100 percent if the two percentages are equal. We call this the"minority reflection percentage"because it measures whether minorities in apprenticeships reflect minorities in the state population. In the repeal states before repeal, the minority reflection percentage was 107 percent, which means that the construction apprenticeship programs slightly over-represented minorities. After repeal,minority representation in apprenticeships fell to 85 percent of minority representation in the state population. In the states that retained their prevailing wage laws throughout the period under review,minority representation in apprenticeships just about mirrored minority representation in the state population(a ratio of 102 percent). But, in states that never 51 Davis-Bacon Repeal Effects Table 3.2 Training rates in repeal and never-had states as a percentage of training rates in states which retained their wage laws Repeal States States Never Hav ing Had Law 1 2 3 1975 106% 85% 1976 112% 109% 1977 100% 86% 1978 97% 81% 1987 87% 74% 1988 68% 12% 1989 70% 70% 1990 63% 60% e lames-have train rates p Exce t in 1976 the states that never had-Tr ng wig u$ which fall from 86 percent of the training rates-of'states which retain,their prevailing wage laws to 60 percent of the training rates of states which retain their law. Repeal states mirror the training rates of retaining states prior to their repeals. After the several repeals of state prevailing wage laws--from 1979 to 1988—the average training rate in repeal states falk10 63 percent of the training rates in states retaining their.°law. This is a simple wad.of viewing a raughly 44 a° renticeslii trutni . caused,b state percent drop in rgg2stered construction _pp p, g y repeals of their prevailing wage laws. Davis-Bacon Repeal Effects 52 Table 3.3 Training rates in repeal and never-had states as a percentage of training rates in states that retained their wage laws,1975-78 and 1987-90 Source:US DOL,BLS and BAT. Dependent Variable=Log of the Odds Ratio of the Percent Apprentices Effect on Independent Variables Percent Trained Region 1 -1.11 Region 2 -0.99 Region 3 -0.77 Region 4 -0.81 Region 5 -1.18 Region 6 -1.10 Region 7 -0.53 Region 8 -0.55 Time trend -0.02 State unemployment rate 0.04 Marker for states never having had law (NEVERHAD) -0.13 Marker for states once they repealed their law(REPEAL) -0.44 Constant -0.78 -------------------------------------------------------------------Adjusted R square =.45 Number of Cases =297 Years =1975-78 and 1987-90 All variables are statistically significant at the 1% level except the marker for states never having had a prevailing wage law. That variable is signifcant at the 10% level. Region 1: CT MA NH RI VT ME Region 2: NY NJ DC PA DE MD Region 3: WI IL IN OH MI Region 4: ND SD MO MN KS IA NE Region 5: WV VA NC SC GA FL Region 6: TX OK NM AZ AS MS LA AR TN KY 53 Davis-Bacon Repeal Effects 20.0%-/ Before Repeal ------------------------------------------------------------------------------ 18.0%-/ E2 ' 16.0%- ------ ----- ----------------------------- 14.0%-/ 15.1% .j n=221 ----- ---- 15.4% After Repeal 0 -- (n ------ 12.0%- ----- ----- ----- 2.5% ----- 15 10.0%- 8.0%- CI_ ----- CL < 6.0%- C: ----- ----- a) 4.0% 2.0%-/ 0.0%_/ I Repeal States: States which: States which: Repeal States: I Years=1975-78 and�790 Figure 3.9 Minorities as a percentage of all construction apprentices by state groups Source:U.S.Department of Labor Bureaus of Labor Statistics and Apprenticeship Training In repeal states,before repeat of their prevailing wage laws in construction, minority participation in registered apprenticeship programs averaged 19.4 percent of A apprentices.After the repeals,minority participation fell to 12.5 percent of all apprentices.The n=28 and n=46're-fer to the number of state-year observationsStates vaft in,eachgroup..:S that prev that never had prevailing wage Laws had,roixgkly the same rate Of-aflitatity participation throughout 19754#,and 1-198740. On average;,;ho wever,,pqp:uIaU'xms of the states that never had prevailing,wage,laws had higher proportions,of minorities. Davis-Bacon Repeal Effects 54 120% Before Repeal Retain Law ------------------------------------------------------ 100% no28° 102% ever Had Law Pater Repeal n=221 ----- cc ----- C U 80% 83% 85% Q) n�6 n=39 0- ----- ..... c U 60% m ------ ----- ' 40% L O 1✓ C 20°fo 0 0/oRepeal States: States which: States which: Repeal States: 100%=MinoritY App. Reflect State Pop. Years=1975-78 and 1987-90 Figure 3.10 Ratio of the percentage of minorities in construction divided by the percentage of minorities in the state population,by state groups Source: U.S.Department of Labor,BAT and BLS. This figure presents the ratio of the percentage of minorities in construction apprenticeships divided by the percentage of minorities in the state population. This ratio allows us to measure whether minorles are under=represented u construction apprenticeship programs.A,ratio of l00%wanld show that the proportion of minority apprenticeships in each groop of states elaetty reflects the minority as part of the state,population:ltilinorityparticipaton Ih construction apprenticeships:mirrored the state population in repeal states prior to repeal. Ian fact,minorities were slightly over-represented at 107,percent.In states that retained their prevailing wage laws throughout the period(1975-78 and W7-90), minorities again were very slightly over-represented at 102 percent.In repeal states,after the repeals, in contrast, minority participation in apprenticeships fell to levels that seriously under-represented minorities (85 percent) and resembled the under-representation characteristic throughout the preiod of states that never had prevailing wage laws (83 percent). Non-union apprenticeship programs tend' to be small and do not fall within the oversight of affirmative action guidelines- which may be why the repeals have led to an mider-representation of minorities in apprenticeships. 55 Davis-Bacon Repeal Effects had prevailing wage laws,minority representation rates averaged 83 percent throughout the period. Thus,both repealing states prior to repeal and"retaining"states throughout the period had minority participation in construction apprenticeships that mirrored the state population. In contrast, both repealing states after the repeal and states which never had prevailing wage laws had substantially under-represented minority participation in construction apprenticeships. Summary Employment in construction is inherently unstable,because the industry fluctuates cyclically and seasonally-and firms expand and contract their employment as they win and lose job bids.Unions have acted like a flywheel in the industry,creating career workers when there were only casual jobs. Unions did this by facilitating the movement of journeymen from employer to employer and minimizing the employers'transaction and screening costs for the training. Unions also lowered training turnover by providing a mechanism whereby employers and journeymen could rationally invest in the human capital of apprentices.This raised the wages of apprentices so they would stay with training and induced the union and employers to promote the passage of apprentices to journeymen in order to preserve their investment.Unions also encouraged the career attachment of trained journeymen by providing relatively high wages and additional wages in the form of health and retirement insurance,which are increasingly attractive to workers as they age.By creating career jobs in a casual labor market, unions created the institutions needed to make human capital investment a rational market activity. With the decline of unions in Utah,the formation and preservation of human capital skills have become less-rational. Self-investment by apprentices becomes more precarious as the differential between the apprentices'wage and alternative wages in other industries widens.It simply becomes more reasonable for apprentices to leave construction if unforeseen personal budget problems emerge.The high turnover among non-union apprentices represents in the aggregate a considerable loss of human capital to the construction industry,even though it is not a loss the employer or the state pays for directly. With the lowering of construction wages, it becomes reasonable for young construction workers to limit the amount of human capital they invest in themselves. With the worker'slower stake in construction skills and with the disappearance of wages in the form of health and old-age insurance,it becomes more reasonable for journeymen construction workers to abandon the construction field when they start families.This represents an additional loss of built-up human capital. Contractors in Utah have attempted to minimize the effect of this increased skill volatility within the industry by encouraging firm attachment.Still,despite initiatives,such as profit-sharing, 401 K plans, and health insurance,designed to attach key workers to a firm, construction turnover remains well above the average for the Utah labor market. In short, union decline has meant the decline of the career worker within Utah construction, a diminution in incentives to invest in construction skills, and an increased loss of accumulated human capital as apprentices and journeymen leave the trades.Although the loss of human capital and career jobs in this industry does not appear as a private cost on the ledgers of any contractor,the industry and society at large pay a price for the loss of financially secure occupations in construction.Not only is quality in the industry put at risk when human capital stocks are allowed to dwindle, but the quality of social life is imperiled when we dismantle the institutions that generate stable jobs out of unstable working conditions. Davis-Bacon Repeal Effects 56 This instability is mirrored in the continuing decline of construction wages in Utah.Despite a return to boom times in Utah construction,construction worker earnings continue to fall relative to average annual earnings in the state.Utah's construction boom has had to piggy-back on the training of California construction workers.Whether Utah can continue this free ride is uncertain. What is certain is that there is no free ride from the effects of a federal repeal of Davis-Bacon. Experience from state repeals indicates that formal apprenticeship training in construction will fall by about 40 percent if Davis-Bacon is repealed.If state experiences are predictive,this will hurt minority workers most. In states that repealed their prevailing wage laws, minority participation in apprenticeship programs fell from reflecting each state's minority population to significantly under-representing minorities.This pattern is consistent with states that have never had prevailing wage laws.Although states that retain their prevailing wage laws have minority participation in apprenticeships that reflects their state populations, states that have never had prevailing wage laws have minority participation rates that are only about 80 percent of the rates in which minorities are present in the state population. From chapter II we have seen that a repeal of the Davis-Bacon Act will lower construction wages and earnings.That finding is consistent with the case study of Utah presented in this chapter. We have also seen that a repeal will significantly reduce training in construction.It may well be that as the stock of human capital falls in construction and as the jobs market becomes casual and turbulent,more minority workers will obtain jobs.But they will not obtain training as they do now n in the states that retain their prevailing wage laws and they will not be entering into occupations that offer a middle-class income with benefits. 57 Davis-Bacon Repeal Effects Davis-Bacon Repeat eat Effects 58 IV. Construction Safety Put at Risk Construction is dangerous work.In fact,it is the nation's most dangerous industry.According to the U.S. Bureau of Labor Statistics: • More than 900 construction workers are killed each year—3 to 5 per workday. • 510,500 work-related injuries and illnesses occur annually—almost 2,000 cases per day." • 204,800 cases involve lost work days,for a total of 4.6 million days lost from work per year. A recitation of the hazards associated with construction work, however, cannot ignore the substantial variability of accidents and their consequences across job sites and institutional environments. Accidents and injuries are the product of a complex interaction between worker and environment,and injuries will be either fostered or limited,depending on how well this interaction promotes safety. This chapter focuses on the effect of the repeal of state prevailing wage laws on injuries in construction. The focus on safety rather than overall health, at this juncture,is strictly a concession to the paucity of reliable data on illnesses related to construction. Why might the repeal of a state prevailing wage law affect the safety record in construction? How does the presence or absence of such a law alter the important interaction of worker and environment?Certain parameters are key to the incidence of injury.For instance,construction work is more dangerous when workers are untrained and inexperienced. Stresses associated with a lack of job security, the pace of work, and the possible avenues for grievance all feed into the critical interaction of work and environment on any job site. In Utah, following the 1981 repeal of the state's prevailing wage law, trainingdeclined as the ti labor market was going into recession see chapter III . The lack of training and construction ( p ) g g widespread use of inexperienced workers began to surface as the construction economy rebounded. One experienced pipe fitter recalls of that era: Contractors were using inexperienced people with no training.They had no training program to begin with,they were hiring people off the street with no experience in the trade.What they would do is everyone that got hired on one project that did not have a history or work experience on a construction job,they had to wear a red sticker on their hard hat. They had to wear that for 30 days.Well everywhere you would looked there were red stickers everywhere.I estimate that about 40 or 50%of the people had one on their hat.They called them"hamburger kids." —Pipe fitter, Salt Lake City, 1994 Lack of training and inexperience are not the only sources of work injuries. In Utah there was a greater sense of job insecurity after the repeal of the state's prevailing wage law and the related decline in union work.Without union security,ex-union workers with training and experience found themselves taking chances they would not have taken prior to the repeal.One union worker who was forced to take work in the open shop recalls: I got hurt in 1986. There was a great deal of pushing to get the job done.I was working with an older man that came out of retirement.He was about 70 years old.We were waiting for a cherry- picker to move some pipe. We were waiting for a couple of hours, because they laid off some operators. After two hours of waiting, two hours of superintendents eyeballing us, I went and 59 Davis-Bacon Repeal Effects walked under the piece of pipe, which weighed 253 lbs. I carried it over to the structure,but I didn't see because the snow was covering a hole in the ground. I stepped in it, it was about 14 inches deep and 2 feet across.I pulled muscles in my back,pulled some discs in my back.What I was thinking of at the time was, I can't afford to lose this job. All these guys walking by me looking at me,I thought we better get this pipe in there some way. I was nervous,I should not have done it but I did. —Union pipe fitter,Salt Lake City, 1994 Why Prevailing Wage Repeals Lead to Increased Injury Rates We can postulate,based on studies of safety and health in the construction industry,why repeal of the state prevailing wage laws is associated with increases in injury rates.Take as the first premise these telling facts: • The rate of injuries "decreases substantially as length of service increases."" • Large,experienced employers in construction have injury rates that are 80%below small-to- medium-size contractors. Repeals of state prevailing wage laws have altered construction labor markets in those states in several ways that affect job site safety: 1.The bidding process has become cutthroat. 2.Workers are less likely to make a career of construction work. 3.Even as experienced workers are leaving the industry in increasing numbers,apprenticeship training has declined. Cutthroat competitiveness in contracting.In Utah,the repeal ofthe state's prevailing wage law led to a burgeoning of start-up contractors with limited track records(chapter II).These new entrants joined existing contractors in a heated bidding process for state contracts that resulted in lower bids, but ultimately higher costs,as a percentage of the state engineer's estimate of the job cost.Cutthroat competitiveness,in other words,resulted in increased cost overruns.Inexperience at the firm level, small size,and cost pressures all contribute to compromised safety on the job. Because of their relative inexperience, new firms tend to face greater on-site coordination problems than firms with longer track records. Such problems can add to costs, but also directly endanger safety.Problems in coordination,perhaps related to delivery of materials and equipment, or in scheduling work with subcontractors, lead to greater uncertainty with respect to the construction schedule.Uncertainty is a breeder of safety risk,as workers can less easily anticipate and plan for the daily contingencies of work. New entrants in the industry also are generally smaller in size than established firms. Smaller firms have worse safety records than larger firms, in part because of greater laxity of enforcement of safety rules and the relative absence of formal safety programs. Of greatest importance,however,is the firm's reaction to increased pressure to cut costs in the face of intensified competition and cost overruns.There is a tendency to speed up work and cut back on safeguards in the face of such pressures. Workforce turnover.When state prevailing wage laws were repealed,worker turnover increased significantly,as the industry found it harder to retain workers for long-term careers(see chapter III). Repeals resulted in a decline in the union share of the construction labor market, driving down average construction wages in the state and decreasing union apprenticeship training for Davis-Bacon Repeal Effects 60 construction.In response to the decline in union membership and training,contractors attempted to reduce turnover—to retain skilled workers and to minimize screening and training costs. Still,the decline in wages and in health and pension benefits drove experienced construction workers from their trades for careers in other industries.Thus,while construction firm turnover is on the decline, turnover in the whole industry is on the rise. Those who now work on federally funded Davis-Bacon projects are more likely to be union trained because of the demanding nature of these large, civil engineering jobs. They are likely to know more about new processes and changes in technology, and they are more likely to have graduated from certified apprenticeship programs. In states that retain their prevailing wage law—compared with those that never had such a law or repealed such a law—the proportion of construction workers receiving training is higher and injury rates are lower.A decline in wages and benefits leads to a flood of inexperienced workers into the industry as well as a decline in skilled,experienced workers needed to supervise the recruits and to assure that they work safely. Decline in the skill base of the construction labor market Experience is a major determinant of safe work performance—and productivity.Training of skilled construction workers is normally conducted through apprenticeship training programs, most of which are operated by unions and employers through joint trust funds. An integral part of this training is learning on the job while properly supervised. In that way, workers learn from experience while on a variety of projects. Among other things,apprentices are trained to identify and correct ergonomic problems,to detect physical hazards, and to detect the presence or release of hazardous chemicals. Knowledge about safety and health hazards,appropriate protective measures,and hazard communication methods are all important elements that apprenticeship programs provide. When little Davis-Bacon acts are repealed, training and apprenticeship programs decline and the skill base of workers erodes (chapter 11I). Without employer incentives to continue apprenticeship programs, knowledge of proper safety and health procedures declines as well. Summary. The combination of these factors—cutthroat competition,a decline in training,and an erosion of career attachments to the industry—affects the safety-related skill and experience base of the construction labor force. Workers become more injury-prone and know less about the kinds of risks they are taking. Furthermore, as the workforce becomes less skilled and its wages in construction decline, workers are forced to take more safety risks to simply make a living. Furthermore,contractors caught in the competitive speed-up often press their workers to speed up and take more chances. Workers are put at increased risk in an already hazardous industry. 61 Davis-Bacon Repeal Effects A Comparison of Injury Rates The U.S. Bureau of Labor Statistics' annual Occupational Injuries and Illness Survey reports accidents by state and year. Construction injuries vary by the type of work being done. We will analyze these BLS data for plumbers and pipe fitters employed by specialty contractors in the Standard Industrial Classification(SIC) 171. This specialty trade has injury rates in the mid-range of rates for construction and this trade is often employed on public works. For pipe fitters in 1978-91, states that had state prevailing wage laws averaged 13.83 injuries for every 100 workers employed(fig. 4.1). In addition,in the states that repealed prevailing wage laws, injury rates for plumbers and pipe fitters before repeal was slightly less (13.54 per 100 workers)than the injury rates in other states with state prevailing wage laws.By contrast,states that never had state prevailing wage laws had higher injury rates(14.74 per 100 workers)and the repeal states, after they repealed the prevailing wage laws had the highest injury rates of 15.41 per 100 workers. These increases in injuries resulted in a similar increase in workdays lost per worker.79 It is possible that injury rates might differ between states for reasons other than changes in legal status. The union pipe fitter who got hurt in Utah in 1986 slipped partly because of snowy conditions.Perhaps factors associated with safety unrelated to repeal coincidentally worsened after repeal.We controlled for factors such as regional differences in weather,time trends in injury rates, and the effects of unemployment in a multiple regression analysis of construction injuries among plumbers. This approach permitted us to isolate the effect on safety of changes strictly associated with the repeal of state prevailing wage laws. We modeled injury cases per worker as a function of geographic regions,the unemployment rate,a time trend,and the legal status of state prevailing wage laws(table 4.1).Three measures of injury rates are reported: injury cases per worker(col. 2); serious injury cases per worker, defined as injury cases that required time off from work (col. 3); and the number of lost work days per worker(col.4). In all three models,our focus variable,the act of repealing a state prevailing wage, has a positive coefficient.This means that as the states repealed their prevailing wage laws,injury rates went up according to all three measures. In our model, the dependent variables are logged. This allows for a straightforward interpretation of the repeal variable as a percent increase in injury rates. So,as these states repealed their laws, the injury case rate went up by 14 percent, the serious injury case rate went up by 15 percent and the work days lost per worker per year went up by 12 percent.All of these findings are. statistically significant. All other things being equal, states that have never had prevailing wage laws also have higher injury rates for plumbers and pipe fitters in the construction industry.In terms of injuries per worker and serious injuries per worker, our results indicate that states that never had prevailing wage laws affecting construction had a statistically significant 5 to 9 percent higher rate compared with states that have prevailing wage laws.80 The Cost of Injuries The costs of injuries in the construction industry are staggering.Of the nation's$62 billion spent on workers' compensation, approximately 30% goes for construction-related injuries and illnesses, Davis-Bacon Repeal Effects 62 16.00% ---------------------------------------------- ----- ----- 15.41% n=38 14.00% 1n4.G7°4% 12.00% n-30 Y ----- ----- .... 0 10.00% ------ ------ ----- ----- 0 8.00% U 6.00% c ----- ------ ----- ---- 4.00% ------ ------ ----- ...... 2.00% 0.00°l0 Before Repeal Have Law Never Had Law After Repeal Figure 4.1 Injury rates in construction by status of prevailing wage law Source:U.S.Department of Labor,Bureau of Labor Statistics. uty,rates><n constructivo were relatively low in the nine repeal states prior to D repeal(1,3.54 percent). After,the various repeals,.iupry rates,on average,rose-;to 15.41 percent. In flW32 states "t have retained-prevailing wage laws,i0ory rates,ha-ve been and,remain relatively low. In nine sta eS 1h haNN=never had state prevaingge laws,iry:rats ere an€l renttn tve� Haigh,`I:he notation "'W' refers. wthe p Hers of state-ye al(Net tihttsAU leach group.por instance,:there were 2,0 state year combinations for states that had prevailing wage laws a 1978-91. 63 Davis-Bacon Repeal Effects Table 4.1 Regression model of the effect of state repeals on injury rates for plumbers and pipefitters Source: US DOL,BLS. Dependent variable:log of injury rate for plumbers and pipe fitters (by year and state) Serious Days Cases Cases Lost Per Worker Per Worker Per Worker (1) (2) (3) (4) (Constant) -1.21 -2.16 -6.85 Region 1 -0.39 -0.41 -0.10» Region 2 -0.27 -0.29 0.14' Region 3 -0.46 -0.70 -0.35 Region 4 -0.40 -0.65 -0•44 Region 5 -0.34 -0.49 -0.29 Region 6 -0.33 -0.40 -0.13* Region 7 -0.32 -0.64 -0.43 Region 8 -0.18 -0.26 -0.25 Time Trend -0.02 0.00' 0.01 Unemployment -0.18 -0.19 -0.04* Never Had Law 0.09 0.07 0.05' Repealed Law 0.14 0.15 0.12 Adjusted RZ 35% 49% 16% Observations 350 313 350 Not statistically significant. (Regions are standards BLS categories). In columns{2);(3) and(4),we report three models;of injury rates,the first for injury cases per worker(2),the second for serious in ies per worker(3)and the last for days lost per worker<(3). Controlling,for real derences in injury rates,=general.trndsninlureso+r tie--anda �n. tte:unemployment rags,all.threepesiinluies:or htgern sti(teaftal~r+gealed heir prevailing wage lawsA states,':that.never had such lbws:Inyepeal states,injury rates climb from 1.2 to,15 percent competed to the rat, ',prior to repeals. Davis-Bacon Repeal Effects 64 or roughly$20 billion.This,for a construction labor force which represents but 5 to 6 percent of the whole U.S.labor force.In addition to the direct costs of workers'compensation,there are numerous industry-related indirect costs connected to work-related injuries or deaths. These include job shutdowns and retraining of workers. According to the Construction Industry Institute,"even when the estimates of claims are deleted from cost data, indirect costs still exceed the direct costs."" Based on the our regression model of the experience of the nine states that repealed their prevailing wage laws, we project that national injury rates82 will increase by around 15% if the Davis-Bacon Act is repealed. What this will mean in terms of safety is: • There will be 30,000 new cases of lost-time injuries each year,accounting for 675,000 days lost from work. • Workers' compensation costs will increase by about$3 billion per year. • Because Davis-Bacon construction accounts for approximately 10 percent of all construction, it is estimated that repeal of the Davis-Bacon Act would increase federal construction costs by $300 million per year in direct,workers'-compensation-related costs alone,and indirect costs would double this figure. The numbers might prove larger, because a Davis-Bacon repeal in the wake of state repeals may have a larger impact on the construction industry. Summary The institutional context of work is critical to worker health and safety.State prevailing wage laws, on the surface,have little to do with worker health and safety. But such repeal has fundamentally altered an institutional context that was more conducive to workplace safety. Repeals of state prevailing wage laws,therefore,have had hidden effects.Because the bidding process becomes overheated;because contractors,as a group,take less responsibility for training and safety;because workers feel less secure on the job;and because the workforce becomes less attached to and experienced with construction work; construction becomes more dangerous. Safety in an already relatively dangerous industry is put at risk by the repeal of prevailing wage laws. 65 Davis-Bacon Repeal Effects Davis-Bacon Repeal Effects 66 V. Conclusion The Effects of the Repeal of Prevailing Wage Laws The federal system of government in the United States is sometimes called"democracy's workshop." The diverse experiences of the 50 states afford a valuable window for assessing the successes and failures of public policies. Between 1979 and 1988,nine states repealed their prevailing wage laws regulating the construction of public works. These legislative changes enable us to examine the before-and-after pictures of the effects of such repeals.Nine other states never had prevailing wage laws governing public construction,while the remaining thirty-two states retained prevailing wage laws. These "never-had" states and "retaining" states give us additional perspectives on what it means to keep or repeal prevailing wage laws. Legislators are often forced to act on theory; this is one instance where they can act on facts and experience.The experience of the last 20 years in the application and removal of state prevailing wage laws on public construction offers insight into the prospective effects of further state repeals or the proposed repeal of the federal Davis-Bacon Act. The Goals of State Prevailing Wage Laws Prevailing wage laws were first enacted at the state level. Kansas passed the first prevailing wage law on public works in 1891 as part of legislation mandating the eight-hour work day. Prevailing wage laws were central to a larger effort to improve working conditions for American citizens.The notion was that child labor laws should enable children to be in school and the eight-hour work day should help allow workers time to spend with their families. The proponents of prevailing wage legislation wanted to prevent the government from using its purchasing power to undermine the wages of its citizens. It was believed that the government should set an example, by paying the wages prevailing in a locality for each occupation hired by government contractors to build public projects. Before the Great Depression,Arizona,Idaho, Kansas,Massachusetts,Nebraska,New Jersey, New York, and Oklahoma passed prevailing wage laws regulating state building and road construction. In 1931, Congress passed the Davis-Bacon Act. Soon thereafter, 18 additional states adopted prevailing wage laws.After World War II and until 1982, 15 more states passed prevailing wage laws. All of these laws raised the question: what was meant by a prevailing wage? The Definition of a Prevailing Wage Wages in local labor markets often have a peculiar distribution.Particularly where there are unions, but also in other circumstances,the highest wage in a local labor market is often the most commonly found wage rate.Even when the highest wage occurs most often,however,it will not be the average wage simply because the lower wages—however few or many—for that occupation will bring the average wage down. Prevailing wage laws are intended to get the government out of the business of pulling down wages. The dilemma is that if the state pays the average wage, it will automatically undercut the most commonly found wage. Alternatively, if government pays the highest wage found, it will always be pulling the average wage up.When is the highest wage sufficiently common that it should be called the prevailing wage rate,even though it will never be the average wage? Davis-Bacon Repeal Effects 67 In the federal law, this dilemma was resolved by a threshold rule. This rule stated that if the most commonly found wage rate,to the penny,accounted for more than 30 percent of all wages for an occupation in a local labor market, that was the prevailing wage even though it was not the average wage.On the other hand,if the most commonly found wage rate accounted for less than 30 percent of all wages for an occupation in a local area,the average wage rate prevailed. In 1985, the Reagan administration revised the rule and raised the threshold to 50 percent. Today,Davis-Bacon wage rates are the average rate for an occupation in a local labor market except, in roughly one-third of the cases,where 50 percent of the wages in that area are precisely the same. If more than half of all workers in an occupation in an area make the same wage,that wage rate— even if it is above the average — is said to prevail. But two-thirds of the time the average wage prevails. Modern opposition to prevailing wage laws is usually founded on one of two objections.Some people oppose the idea of the government agreeing in advance to pay the average wage rate for workers in specific occupations in a local area.This criticism is completely at odds with the original purpose of prevailing wage legislation, which was to prevent the government from hiring labor at below-standard rates.Other critics object to paying a prevailing wage that is greater than the average wage in the locality.The premise of this second objection has lost a great deal of its force in recent decades. As a result of the adoption of the 50 percent threshold, and the additional fact that unionization in the construction labor market has fallen from 70 percent to about 25 percent in the last three decades, there are far fewer cases in which the wages rates determined as prevailing are greater than the average rate. The Financial Costs of State Repeals Lower wages for all construction workers. Supporters of Utah's 1981 repeal of its prevailing wage law recognized that repeal would lower construction wages.They maintained,however,that the money saved on public works construction justified the government's indirectly lowering the wages and earnings of some of its citizens. And, indeed, construction earnings did fall. In Utah, construction workers, who through the 1950s, 1960s, and 1970s earned 120 to 130 percent of the average non-agricultural wage in the state,saw their wages fall steadily after repeal.By 1993,Utah construction workers were earning only 103 percent of the average annual earnings in Utah, even though Utah was then experiencing a massive construction boom, in which construction wages normally go up. This earnings decline affected all Utah construction workers—whether union or non-union,whether employed on publicly or privately financed projects. Taking the nine repeal states as a whole,the average annual earnings of construction workers in these states fell from$24,317 (in 1991 dollars) per year before the repeals to $22,148 after the repeals. This is simple but compelling evidence that repeals of state prevailing wage laws have lowered construction wages. A more complex analysis confirms this general observation. Using multiple linear regression analysis,we isolated the earnings effects of the state repeals while controlling for the business cycle, regional differences in wages and unemployment,and long-term trends in earnings and employment that are not associated with repeals of prevailing wage laws. We found that the nine repeals cost construction workers in those states$1,477(in 1994 dollars)per worker every year since state repeal. This was about a 5 percent drop in construction earnings attributable to each state's repeal of its prevailing wage law on public works. A slight increase in construction employment. Proponents of state repeals maintained that the lowering of wages would be offset by an increase in construction employment. While high-paid, 68 Davis-Bacon Repeal Effects high-skilled workers would be hurt by a repeal,it was believed,low-paid,low-skilled workers would have more job opportunities in construction. Repeal proponents were right that cheaper construction labor would lead to an increase in construction employment. Again using regression analysis, we found that the repeal states experienced a 1.7 percent increase in construction employment that would not have occurred without these repeals. This was an unfavorable trade-off from the standpoint of workers,however,as their wages fell by 5 percent overall while their employment rose by less than 2 percent. It turned out to be a tough trade-off for government budget-watchers as well. Lost tax revenues. As a group,construction workers lost income,because their wages dropped by 5 percent and their total employment rose by less than 2 percent.This caused the government to lose substantial tax revenues. In recent years, the state of Utah has lost $3 million to $5 million annually in sales tax and income tax revenues because it repealed its prevailing wage law in construction. Increased construction cost overruns. Cost overruns are a hidden cost of repealing prevailing wage laws.In Utah,cost overruns resulted from an over-heated bidding process in which contractors, shaved their bids in an urgent effort to obtain government contracts. After the repeal,winning bids on state jobs came in lower than ever before, but the final job costs were a higher percentage of original estimates than ever before (chapter 2, fig. 2.3). Having underbid jobs, contractors and subcontractors would arrange change orders to get the jobs done or simply walk away from badly underbid jobs and leave the state to pick up the pieces. In Utah, cost overruns on the construction of state roads tripled in the 10 years after repeal,compared with the 10 years before 83 The bottom linefor Utah's budget.The Congressional Budget Office estimates that,should the federal Davis-Bacon Act be repealed, the federal government might save a total of 1.7 percent on its construction costs.This savings might even be less.84 Using an even more conservative figure of 3 percent to estimate what Utah saved in construction costs by repealing its prevailing wage law,we calculate that the Utah state budget almost—but not quite—broke even.Balancing construction cost savings against lost tax revenues,in two of the years since 1987 the Utah budget saved more money in construction costs than it lost in tax revenues. In five of the years since 1987,the state lost more in tax revenues than it saved in construction costs(fig. 5.1).In either case,the net savings or losses were small compared with the lost earnings of Utah's citizens (table 2.6, row 3). But construction workers—and the industry—were to lose more than money when these repeals were enacted. Davis-Bacon Repeal Effects 69 $4 ------------------------------------- ------------------------------------------ $3 --------------------------------------- ------------------------------------------- $2 --------------------------------------- -------- --------------------------------- C c o $Q --------- ------------------ ----------------------------------------------------- M {$2) ---------- --------------------------------------- ---------------------_-------- ---------- ------------------------------------------------------------------------ ($3) � ) Lost Taxes Saved Costs Budget Loss Figure 5.1 Average annual income-tax revenue loss and construction cost savings and net effect of repeal for Utah, 1987 to 1993,in 1994 dollars Source:Table 2.6. On average.,the repeal of Utah's prevailing wage law has cast the state budget $400,000 per year froth 1987 to 1993. This figure has been rising,and reached$1 million in 1993. Should the;federal prevailing wage law be repealed,the gap between lost federal tax dollars and construction cost'savings will be greater. This is partly because>a Davis-Bacon repeal would affect more construction and more workers,but also because1he federal government income tax rate is higher than Utah's. Obviously,the higher the income tax rate,the greater the tag 70 Davis-Bacon Repeal Effects Other Costs of State Repeals A less-skilled labor force Unions and union contractors do the lion's share of worker training in the construction industry. Some very large non-union contractors do their own training,but most non-union contractors hire out-of-work union-trained construction workers and workers who have learned a trade on a catch-as-catch-can basis. Most non-union contractors are not big enough to afford to train and retain their own labor force. Contractors,understandably, are afraid that in the first slack period,the workers they trained will leave them and work for their competitors. Unions historically have compensated for this market failure by inducing union contractors to share the burden of training and to share each other's apprentices. In Utah,the repeal of the prevailing wage law led to a dramatic decline in union apprenticeship programs because the repeal led to a dramatic decline in local construction unions.Having repealed the prevailing wage law,the state was not inclined to pour money into local community colleges and vocational training centers to make up the difference. At first after the repeal in 1981, the Utah construction economy limped along in the trough of a business cycle so the absence of quality training systems was not strongly missed.Non-union contractors hired out-of-work union members and the older generation of construction workers provided a relatively skilled labor force in the open shop of non-union construction. In the last three years, however, Utah has experienced a massive construction boom. Few training systems were in place to meet this boom. Utah has filled the gap by relying on traveling construction workers from California, which is in a construction slump. Utah has also relied on a less-skilled labor force.Whether Utah will be able to continue to rely on California workers remains to be seen; if California's economy picks up, many of the skilled California travelers will likely return home to the increased wages there. Utah's experience with declining availability of construction training was not unique. Comparing the decade before repeals to the decade after repeals,union and non-union apprenticeship rates in construction fell by more than half in the nine states that repealed their prevailing wage laws. States that retained their prevailing wage laws did not lose ground in apprenticeship training and states that never had prevailing wage laws had relatively low training rates in construction throughout the period. The repeal of prevailing wage laws thus had the indirect effect of reducing training and hindering the formation of a skilled labor force. When unions declined in the wake of repeal, only state government could have picked up the pieces. The cost of expanded state-financed vocational training is a hidden cost of repealing prevailing wage laws. So far,it is a hidden cost that few repeal states have been willing to pay. Slowed economic gains by minority workers. A faltering stock of human skills in construction is not the only non-monetary cost that resulted from state repeals of prevailing wage laws. Construction used to be one of the few blue-collar occupations left where a worker lacking a college education could earn a middle-class income.Nationwide,the average construction income in 1994 was $27,500. Becoming a skilled construction worker was a road out of poverty for minority workers.Before the nine state repeals,participation by minority group members—male and female nonwhites—in construction apprenticeships mirrored the minority populations in each state. In the repeal states before the repeal of their prevailing wage laws, minorities accounted for almost 20 percent of all construction apprentices. After repeal, minority participation fell to 12.5 percent of all construction apprentices. Thus, after these repeals, minorities became significantly under-represented in construction apprenticeships. One reason for this decline is that union apprenticeship programs usually enrolled dozens of Davis-Bacon Repeal Effects 71 tied to single employers apprentices.Non-union apprenticeship programs tended to be smaller,often g involving no more than one,two,or three apprentices.Affirmative action regulations do not cover apprenticeship programs of fewer than five apprentices. So the union programs had to fill out affirmative action plans and follow affirmative action guidelines,while the smaller programs did not. When the repeals drove the union programs into decline,minority workers lost the most. For instance, the percentage of minority apprentices in construction, which reflected the minority proportion in each state's population before repeal,declined in the repeal states(fig. 5.2).Minority construction workers may still enter the industry but they are less likely to receive full formal training in the absence of prevailing wage legislation.Although it has been suggested that repeal of Davis-Bacon would lower black unemployment relative to white unemployment by opening up jobs for less-skilled black labor,85 the data do not support such a claim(see chapter 3,figs. 3.3 and 3.4). Nor is there evidence suggesting that a repeal would increase the proportion of minorities in the construction labor force. In 1990, the percentage of minority construction workers among all construction workers was virtually the same in the 32 states with prevailing wage laws compared to the 18 states without prevailing wage laws. Thus, repeal means that minority workers will begin construction work in unskilled jobs and get their training, if at all, on a catch-as-catch-can basis. Furthermore, minorities will enter an industry that is less able to provide a secure blue-collar,middle-class income.Repealing prevailing wage laws has therefore cut off an important road for minorities into the middle class.Without skills training,workers are less productive;without safety training,they are at greater risk of injury in an already dangerous profession. Increased work-related injury rates. All construction workers in the nine repeal states have been put at increased physical risk by the repeal of the several state prevailing wage laws. Injury rates in construction in the nine repeal states have risen by 15 percent after repeal,even controlling for other factors such as unemployment,trends in construction safety,and differences in work safety experiences by region. The decline in apprenticeship training and the rise in construction career turnover are probable causes of this increased injury rate.Other research has found this to be so.The Department of Labor found that the rate of injuries "decreases substantially as length of service increases."86 Construction firms in Utah (and perhaps elsewhere) have sought to stem the tide of increased injuries by reducing firm turnover at least for key workers. This may have dampened the deleterious effects of less formal training and increased career turnover,but on net,injuries are still rising. If the experience in these states can be extended to the nation, a repeal of Davis-Bacon would result in 76,000 additional construction workplace injuries annually.About 30,000 of these injuries would be serious,requiring time off to recover.More than 675,000 work days would be lost.These new injuries would occur because workers would be less well-trained and because they would have fewer on-the-job protections against contractors who are in a hurry. Workers, of course, suffer directly from these occupational injuries—in their physical well- being and in their wallets. Increased injury rates also lead to increased costs for contractors, who must pay higher worker's compensation premiums. And, as consumers of construction services, local,state,and federal governments pay a share of those higher worker's compensation premiums. 72 Davis-Bacon Repeal Effects 120% -------------------------------------------------------------- 100% ---- ---------------- ----------- 80% c a� a� 60% ------ ----------- a) m 40% ---------------- ---------- 20% 0% Before Repeals After Repeals Under 100% = Under-representation Figure 5.2 The percentage of minority construction apprentices,divided by the percentage of minority state population state--the minority reflection percentage--for nine repeal states In the nice repeal states where separate data were available on Kminority populations,in the decatde lieforeTepeai,minority apprenti-ces were°slightly over- represented relatiive ti#their p�rtion.atthe state-papptatirtn. The mtnQrity- reflection percentage was I07 percent. In the+decade=aftOr t w repeals,the minority reflection percentagefellto$5 percent,indicating Agnificantly under- represented minorities.5outce: Figure 3.10. Davis-Bacon Repeal Effects 73 Billions of Dollars $0.5 $0.0 w 0 ---------------------------------- ------------ --------- ($0.5) E i= --------- ----------------------------------------------------------------- ($1.0) ($1.5) Lost Taxes Saved Costs Net Loss Figure 5.3 Estimated effect of a repeal of the Davis-Bacon Act on income-tax revenues, construction costs and total budget Source:Table 2.7. WIMM The Congressional Budget Office estimates that the federal government would save a total of 1.7 percent in>construction costs from a repeal of Davis-Bacon.This chart uses the more conservative cost savingss estimate of 3 percent.At a 3 percent construction cost savings,with a marginal income tag:rate 020 percent and federal construction expenditures at their 1991 bevel(gin 1094 dollars,a,repeal of Davis-Bacon would cost the federal govern ore nt-$i.2 hilli on Ain incase tax revenues.The federal government would save$W million lin cunstraction costs and the federal budget would lose,on net,MIS million. 74 Davis-Bacon Repeal Effects Estimated Effect of A Davis-Bacon Repeal Democracy'sworkshop has given us an opportunity.The nine states that repealed their"little"Davis- Bacon Acts offer a chance to estimate the likely consequences of the repeal of the federal Davis- Bacon Act. Based on this study,we project the following. First, construction earnings would drop if the federal law was repealed. Collectively, for all construction workers,this would mean a loss of almost$5 billion per year in real terms every year. As a result of lower wages in construction,federal income tax collections would fall by roughly$1 billion per year.Projected cost savings on federally purchased construction almost certainly would be less. (fig. 5.3). Second,we estimate that formal training in construction could fall by 40 percent.The industry would move from one of skilled blue-collar workers earning a middle-class income to a much-less- skilled labor force earning substantially lower wages.Minority access to good training likely would fall even farther than overall training rates. Contractors would be using more construction workers and paying less for them,but the less-skilled workers would be building less and adding less value to building projects. Purchasers of construction services would not necessarily profit from lower- wage labor if that labor is also less skilled. This is a potential lose-lose situation. Utah was able to patch together a large-enough construction labor force after its repeal of prevailing wage law. Contractors in Utah rode freely on the training systems in place in California. But the country as a whole cannot go on a similar free ride. If Davis-Bacon is repealed and construction training nationally declines sharply,the United States will not be a small state like Utah turning to California for its rescue. Nationally, there will be no place to turn. Is the federal government prepared to spend the money to establish its own apprenticeship programs in construction?Alternately,will the government induce or require contractors to join into cooperative training programs? If prevailing wage legislation is repealed, it is likely that some additional measures will be needed to ensure occupational training for the construction industry. Last,but not least,we estimate that the construction job site would produce 30,000 additional serious injuries yearly.These injuries would add a large but still-undetermined financial cost to the ultimate price of repeal. It goes without saying that the public benefits from a bidding process that lowers construction costs. But the bidding process must be kept within certain bounds, to prevent consequences that could lead to increased—rather than decreased—public and societal costs. Competitive pressures tempt contractors to cut corners on quality. States and communities employ building inspectors to assure that quality is maintained.Historically,unions have assumed the role of"building inspector" for safety and training in the construction industry. The role of unions. Employment in construction is inherently unstable because the industry fluctuates cyclically and seasonally. Firms expand and contract employment as they win and lose job bids. A worker rarely has a long-term attachment to one employer in the industry, and the construction union may be the only stable,work-related institution the worker knows.Construction unions act like a flywheel in the industry,creating career opportunities out of a casual labor market. Unions do this by facilitating the movement of journeymen from employer to employer and minimizing the employers' transaction and screening costs. Unions lower training turnover by providing a way for employers and journeymen to rationally invest in the human capital of apprentices.Collectively bargained agreements create wage incentives for apprentices to stay with training programs, and also cause their employers to promote the workers'passage to journeyman status. Unions also encourage the career attachment of trained Davis-Bacon Repeal Effects 75 journeymen by providing relatively high wages and health and retirement insurance, which is increasingly attractive to workers as they age. By creating career jobs in a casual labor market, unions create the institutions needed to make human capital investment a rational market activity. With the lowering of construction wages,young construction workers will limit the amount of human capital they invest in themselves. With a lower stake in construction skills and the disappearance of wages in the form of health and old-age insurance,it becomes more reasonable for many journeyman construction workers to abandon construction work entirely when they start families. This is an additional loss of built-up human capital. The loss of a career. Contractors have attempted to minimize the effect of this increased skill volatility in the industry by encouraging attachment of workers to their firms.Still,despite initiatives such as profit-sharing, 401K plans, and health insurance to bind key workers to the firm, construction firm turnover remains high. It appears that the decline of unions has been associated with the decline of the career worker in construction, a diminution in incentives to invest in construction skills, and an increasing loss of accumulated human capital as apprentices and journeymen leave the trades. The loss of human capital and career jobs in this industry does not appear as a private cost on the ledgers of any single contractor.Nonetheless,the industry and society at large pay a price for the loss of middle-class occupations in construction. Not only is quality in the industry at risk when human capital stocks are allowed to dwindle, but the quality of our society is imperiled when we dismantle the institutions that generate stable employment out of unstable working conditions. The construction industry is turbulent.Caught in a perennial boom-bust cycle,characterized by fleeting relationships between small contractors and subcontractors, and driven by short-term strategies of free-riding on the training of others,the construction industry is a market failure waiting to happen. The turmoil in the construction labor market has traditionally been tempered by prevailing wage legislation and labor unions. Absent these institutions, it is unclear how — or whether—the market will regularly and carefully train workers,or assure safety and health on the job site,or provide training opportunities for minority workers,or offer the incomes needed to make construction an attractive career. Government purchases account for 20 percent of all construction in the United States.For the last six decades and more,the government has contributed to the stability in construction labor markets by requiring contractors to pay the wage rates that already prevail in a local areas. Today, voices are urging the government to use its purchasing powers to reduce construction costs at the expense of worker incomes. Such a strategy has a very real cost for workers,the industry,and the government.When nine states chose this path,the results were significantly lower construction wages,slightly higher construction employment,a tripling of cost overruns on public works, an across-the-board 15-percent increase in construction injuries, a 40 percent decrease in apprenticeship training,and an even further decline in minority apprenticeship training.All this was sacrificed to save an estimated 1.7 percent in state construction costs.Even that savings was squandered by the loss in state tax revenues from an impoverished construction labor force—a poor bargain indeed. 76 Davis-Bacon Repeal Effects I References Chapter III Azair-Rad, Hamid, Peter Philips, and Anne Yeagle, "The Effects of the Repeal of Utah's Prevailing Wage Law on the Labor Market in Construction to Sheldon Friedman, et al., eds.,Restoring the Promise ofAmerican Labor Law, Cornell University, ILR Press, Ithaca,New York, 1994, 207-22. Jones, S. Garth. 1979. Letter to the Editor.Deseret News, February 23, 1979. Thieblot,Armand J,Jr. 1986.Prevailing Wage Legislation. U.S. Bureau of Apprenticeship Training, State and National Apprenticeship Program Survey (SNAPS) 1975 to 1978, and Apprenticeship Registration Actions (AMS Report 18) 1987 to 1990. U.S.,Bureau of the Census. 1970 Census of Population. U.S., Bureau of the Census. 1987 Census of Construction. U.S., Bureau of the Census, 1989-1991 Current Population Survey. U.S. Department of Labor, Geographical Profile of Employment and Unemployment, 1974-92. U.S., House of Representatives, 69th Congress, 2nd Session, H.R. 17069,"Hours of Labor and Wages on Public Works'; Hearings Before the Committee on Labor, Feb 18, 1927. Utah, House of Representatives,House Audio Tapes, Feb 1979. I Davis-Bacon Repeal Effects 83 End Notes 1. Topeka State Journal,February 24, 1891,co1.4,p.4. 2.Topeka Daily Capital, February 25, 1891,p.1. 3. Topeka State Journal,February 25, 1891,col. 3-4,p.1. 4.Sixth Annual Report of the Bureau of Labor and Industrial Statistics,126. 5.Sixth Annual, 215. 6.Sixth Annual, 124. 7.L. 1891 Ch. 114 p.192-193. 8.In reply to the question of needed legislation,most workers polled by the Kansas Bureau of Labor and Industry, cited a lack of enforcement of the eight-hour law or complained of long hours.(Twelfth Annual Report of the Kansas Bureau of Labor and Industrial Statistics, 1896,88-89.)Similarly,the Kansas Bureau of Labor and Industrial Statistics reported complaints of noncompliance with the eight-hour law.P.E.Cook,of the A.R.U.Local No.57,stated that the"[eight hour]law is being openly violated by the corporate and private parties on public work...."(Fourteenth Annual Report of the Kansas Bureau of Labor and Industrial Statistics, 1898,204.) 9.Sixteenth Annual Report,272-78.From July 1900 to June 1901,33 cases were won upholding the law. 10. Thirteenth Annual Report of the Bureau of Labor Statistics of the State of New York, 1895,515-37;Fourteenth Annual Report of the Bureau of Labor Statistics of the State of New York 1896,802. 11.U.S.Congress, 1927.Although it has recently been asserted that the Alabama workers were black,there is no direct evidence supporting this claim.Evidence to support this claim comes from third-hand sources,several years after the fact and not referring directly to this incident at all. (See George F.Will,"It's time to repeal the Davis- Bacon Act,"Deseret News,Feb.5, 1995.)In Alabama,in 1930,only 34 percent of all brick and tile layers, carpenters,electricians,painters,plasterers,cement masons,plumbers and construction laborers were black.Thus, absent direct evidence to the contrary,the odds are that these Alabama workers were primarily white. U.S.,Bureau of the Census,Fifteenth Census of the United States,Population Volume IV"Occupations by States",G.P.O., 1933, Table 12,p. 121. 12.Hearings Before the Committee on Labor,House of Representatives-Seventy-First Congress.January 31, 1931. Bacons proposal was re-introduced in 1930 as H.R.9232 by Congressman Elliot W.Sproul from Illinois,while Bacon proposed a complementary bill. 13.Armand J.Thieblot Jr.,Prevailing Wage Legislation: The Davis-Bacon Act, State "Little Davis-Bacon Acts," The Walsh-Healey Act, and The Service Contract Act.Philadelphia:The Wharton School,1986,p.8. 14.Laws of Florida 1979 ch79-14 HB730. 15.Thieblot, 163. 16.Laws of Alabama 1979 Act no.79-123 H362 Letson. 17.Thieblot, 151. Davis-Bacon Repeal Effects 77 18.Mark Erlich,Labor at the Ballot Box: The Massachusetts Prevailing Wage Campaign of 1988(Philadelphia: Temple University Press, 1989),33. 19.Nelson,Richard R."State labor legislation enacted in 1981"U.S.Department of Labor Employment Standards Administration Division of State Employment Standards,p. 1 (office memo). 20.Salt Lake City Tribune,Jan.23, 1981,B2. 21.Salt Lake City Tribune January 23, 1981 B2. 22.Salt Lake City Tribune Jan. 16, 1981,A6,col 1. 23.Salt Lake City Tribune,January 16, 1981 (A6,col 1). 24.Arizona laws 1984 S.C.R.No.1001. 25. The Phoenix Gazette, "Little Davis-Bacon Act is Ruled Unconstitutional," Sept. 12, 1979,p.A-1. 26.Thieblot, 12. 27.Idaho Session Laws Ch 3 HB7. 28.Thieblot, 165. 29.Colorado Revised Statutes 1985 Article 16,8-16-101. 30.Thieblot, 159-60. 31.New Hampshire Ch 117(SB71). 2.Thieblot, , 184. 33.Kansas Ch 186 S.B.112. 34.Louisiana Act No.18 H.B.2. 35.Erlich,4 and 6. 36.Erlich,7. 37.Regional Information Group of Data Resources,"Executive Summary of the Study of the Economic Impact of Repeal of the Massachusetts Prevailing Wage Law"Lexington,MA,August 1988. 38.The Massachusetts Foundation for Economic Research,The Peculiar Prevailing Wage Law(I I I Cabot Street Needham,MA 02194),March 1988, 39. Massachusetts Foundation for Economic Research, 10. 40.Data Resources, 1. 41. Data Resources,4. 78 Davis-Bacon Repeal Effects 42.Erlich, 101, 111,and 112.. 43.Senator Harrison Williams,United State Senate Hearing Before the Subcommittee on Housing and Urban Affairs,First Session on Oversight to Examine the Administration of the Davis-Bacon Act,Washington:GPO,1979, 1-2. 44.With the increasing prevalence of market-recovery agreements between unions and contractors,more often there are multiple union wage rates for a single occupation in a local labor market. This means that,even when uni onization rates are above 50 percent,there is not always a single union wage rate that counts for 50 percent of workers in the market. Thus it is even less likely that the union rate will be the prevailing rate. 45.Steven Allen,"Declining Unionization in Construction:The Facts and the Reasons,"Industrial and Labor Relations Review,41,No.3(April 1988),343-59. 46.Further research in this area is required. We will not know for certain that the Utah repeal of its prevailing wage law was the cause of the subsequent cost overruns until road construction costs in other repeal states are studied. Then the regression modeling that is used elsewhere in this study can be applied to the issue of cost overruns. 47. Current Population Survey. this issue an exam 48.Let us first investigate sue usingle. Assume that union wages are$16.30 per hour and non- union union wages are$10 per hour. When average wages decline by 7.5 percent,and if non-union wages remain the same,what is the percentage decline in union wages? pr:$10-00 r " Wag 7.5% Union Non-Union " W. r overall Wage Wage decline in Percent Percent W Decline Decline 0.13 $10.00 70.87 .87 $10.82 $10.00 0.13 $10.00 $10.00 38.70%, 0% Here W.is the union wage,r.is the percentage of the construction workforce that is unionized,W.is the non-union wage,r"is the percentage of the construction workforce that is not unionized,and Wa,g is the average wage in construction.This table shows that the percentage decline in union wages must be almost 40 percent. 79 Davis-Bacon Repeal Effects 49.Assume that non-union wages are$10 per hour and there is initially a wage differential of 20 percent between the union and non-union workers. This implies that the union wages are$12 per hour.If unions represent 13 percent of the construction labor force,average wages decline by 7.5 percent,and the wage differential is eradicated,what is the percentage decline in non-union wages? 80 Davis-Bacon Repeal Effects I Wu ru W. ra Wang 7.5%decline Percent Percent in Wavg Decline Decline in Union in Non- Wage Union W. $12.00 0.13 $10.00 0.87 $10.26 $9.49 $10.00 0.13 $10.00 0.87 $10.00 $9.49 0.13 $9.49 0.87 $9.49 21% 5.1% The percentage decline in union wages must be 5 percent(from$10 per hour to$9.49 per hour). 50.Assume there is an initial wage differential of 20 percent between the union and non-union sectors.After the repeal of a prevailing wage law,assume the union-non-union wage differential decreases to 10 percent.Now let us investigate the effect of a 7.5 percent overall fall in wages.Assume that non-union wages are$10 per hour and the wage differential between the union and non-union workers falls to 10 percent.This means that the union wages are about$11 per hour.If average wages decline by 7.5 percent and the wage differential remains unchanged,what is the percentage decline in union and non-union wages? Wu ru W" r" Wavg 7.5% Percent Percent decline in Decline Decline Wavg in Union in Non- Wa e Union W $12.00 0.13 $10.00 0.87 $10.26 $9.49 $11.00 0.13 $10.00 0.87 $10.13 $10.31 0.13 $9.37 0.87 $9.49 14.1% 6.3% The percentage decline in union wages is 14 percent(from$12 per hour to$10.31 per hour),and the percentage decline in non-union wages is 6 percent(from$10 per hour to$9.37 per hour). 51.Unless described differently,figures are given in 1991 dollar amounts. 52.The data are provided in four-digit detail of the Standard Industrial Classification(SIC)code.Data are from U.S.Department of Labor,Bureau of Labor Statistics,Office of Earnings and Employment Statistics,Data Analysis Section,Special Tape XC4057,provided by Darrell E.Carr. 53."Secular"trends refers to trends in earnings that are not due to fluctuations in the business cycle nor due to the state repeals of prevailing wage laws. 54.Controlling for contractor type is a conservative procedure.Overall construction earnings may decline as a result of a shift in the mix of construction worker type.We are focusing on the decline of earnings within trades instead of any decline resulting from a shift in the mix of trades.Additional earnings losses may be calculated associated with a shift to a mix of less skilled workers. This is one reason why the regression estimate of earnings decline is lower than the simple estimate which does include the effect of changing crew mixes within the states construction industries. Unemployment rates are for each state for each year. Davis-Bacon Repeal Effects 81 55.This is an annual earnings average by SIC group. When earnings are weighted by the number of workers in each group,earnings fall to slightly below$25,000. 56. Technical details:This regression model was tested on 27,778 observations.Control(dummy)variables for 26 detailed 4-digit standard industry code(SIC)classifications were included in the regression model but not reported in the table.Each coefficient reported in column(2)of the table is statistically significant except the control for the mountain states region.(This means that the estimated regional effect on annual construction earnings for the mountain region of-$79 is small and probably not different from zero.)The unemployment rates for 1991 for the example states shown in table 2.3 were 4.9 for Utah,5.0 for Georgia,and 5.9 for Maryland.The model is a generalized least-squares weighted regression with the weight being the square root of the annual average employment in each SIC industry for that year. The W is 0.73,which means the model is a good fit of the data. 57. The model also estimates a negative effect on annual earnings of$1173 associated with raising the threshold for construction contracts covered by a prevailing wage law to$500,000 or more. This suggests that at some point raising the threshold has a similar effect to repealing the law altogether. However,this result is based on experience from only two states,Maryland with a$500,000 threshold and Oklahoma with a$600,000 threshold. We could not find negative effects on earnings from lower thresholds in the$100,000 to$400,000 range which leads us to be cautious about this result. A conservative interpretation of this result may be that thresholds below$500,000 have a minimum impact on construction earnings while thresholds above$500,000 have progressively more negative effects on earnings. 58.The Standard Industrial Classification(SIC)code for construction consists of detailed categories of general contractors such as commercial and residential general contractors,detailed categories of heavy and highway contractors,and detailed categories of specialty subcontractors such as masonry and carpentry. 59.Robert D.Reischauer,Congressional Budget Office Testimony,before the Subcommittee on Labor Standards, Labor,U.S.House of Representatives,May 4, 1993, Occupational Health and Safety,Committee on Education and Lab P p.4. 60. Robert D.Reischauer,Congressional Budget Office Testimony,before the Subcommittee on Labor Standards, Occupational Health and Safety,Committee on Education and Labor,U.S.House of Representatives,May 4, 1993, p.4-5. 61.U.S.Department of Labor,Bureau of Statistics,Employment and Earnings,December, 1994,Table B-3 for November 1994,p.55. 62.U.S.Department of Commerce,Economics and Statistics Administration,Bureau of the Census,Government Finances, 1990-91, Series GF/91-5,Table 8,column 6,p.9. 63.Some of the material in this chapter concerning the Utah case originally appeared in Hamid Azari-Rad,Peter Philips and Anne Yeagle,"The Effects of the Repeal of Utah's Prevailing Wage Law on the Labor Market in Construction,"in Sheldon Friedman,et al.,eds.,Restoring the Promise of American Labor Law,Cornell University, ILR Press,Ithaca,New York, 1994,207-22. 64.These data are based on quarterly per capita dues contributions to the Utah AFL-CIO Building and Construction Trades Council.These payments underestimate union membership because of under-reporting of membership from participating locals as well as other exemptions and withdrawals of locals. 65.George F.Will,"It's time to repeal the Davis-Bacon Act",Deseret News,February 5, 1995. 82 Davis-Bacon Repeal Effects 66.Indeed,an unreported multiple linear regr ession model tested whether changes in the male black-white unemployment ratio could be associated with state repeals of prevailing wage laws or the fact that a state never had such a law. This model controlled for time trends in the male black-to-white unemployment ratio,regional differences in unemployment ratios,and changes in the level of unemployment. While the model found a strong time trend in the black-to-white unemployment ratio and significant regional differences in the ratio,there was no statistically significant relationship between either the repeal of prevailing wage laws or the complete absence of prevailing wage laws and the black-to-white unemployment ratio.In short,there is no statistical connection one way or the other between the status of prevailing wage laws and the relative unemployment of blacks and whites. 67.These state demographic unemployment rates are from U.S.Department of Labor,Bureau of Labor Statistics, Geographic Profile of Employment and Unemployment, 1974 to 1992. 68.U.S.Department of Commerce,Bureau of the Census, 1990 Census of Population,Detailed Population Characteristtcs and General Population Characteristics,GPO, 1992. 69.Data available on request. 70.Utah Department of Employment Security,Labor Market Information and Research,Annual Report of Labor Market Information, 1993,table 5,Salt Lake City, 1994. 71.Hamid Azari-Rad,Peter Philips and Anne Yeagle,"The Effects of the Repeal of Utah's Prevailing Wage Law on the Labor Market in Construction,"in Sheldon Friedman et al.,eds.,Restoring the Promise of American Labor Law.Cornell University,ILR Press, 1994,207-22. 72.U.S.Bureau of the Census, 1970 Census of Population. 73.U.S.Bureau of the Census, 1970 Census of Population. 74.Figures 3.7 and 3.8 include all states for which any data are available,except California,Delaware,the District of Columbia,Hawaii,and Rhode Island- for which there are no Bureau of Apprenticeship Training data for the second period. We exclude these states and the District of Columbia(for the same reason). 12.We do not know what accounts for the unusually high training rate for"never-had"states in 1976.This anomaly disappears when average training rates by decades are compared. 76.This transformation into the log of an odds ratio meets the normality assumptions of linear regression analysis. The technique used is generalized least-squares regression,with the regression weighted by the square root of (percent trained)times(one minus percent trained)times(state employment). 77.Latent illnesses resulting from exposure to toxic materials are responsible for an uncounted and thus undetermined additional number of injuries and illnesses—the costs of which are home as reduced productivity, ruined lives for workers and their families,and burdens on workers'compensation and other social security systems.For a mix of reasons,there are no reliable estimates on the number of such illnesses. 78.C.Culver,M.Marshall,and C.Connolly,Construction Accidents: The Workers'Compensation Data Base, 1985-1988, Washington,DC,OSHA Office of Construction Engineering, 1992. 79.In figure 4.1,n refers to the number of observations in each state-law category. For instance,there were 230 state-year combinations for states that had prevailing wage laws throughout the period. 83 Davis-Bacon Repeal Effects 80.In the case of lost workdays per injury,the reported result is of the expected sign,but not statistically significant. 81.Jimmie Hinze,Indirect Costs of Construction Accidents,Seattle:The University of Washington,1992, 14. 82.Because of small numbers,there are no reliable estimates on how repeal would affect death rates. Thus,we cannot calculate the projected increase in fatalities due to repeal. If,however,they were to be affected at the same magnitude as are injuries,we would expect an increase of 130 to 150 fatalities per year. 83.Utah,Department of Transportation,"Final Estimates Processed for Payments, 1970-74 data published in 1985 and 1994 reports. 84.The savings are so small because labor costs on public works are only roughly 25 percent of total costs.If you cut those labor costs by 10 percent,you have cut total costs by only 2.5 percent. 85.George F.Will,"It's time to repeal the Davis-Bacon Act",Deseret News,February 5, 1995. 86 Charles Culver,Michael Marshall,and Constance Connolly,Construction Accidents: The Workers' Compensation Data Base, 1985-1988. Office of Construction and Engineering,OSHA, 1992. 84 Davis-Bacon Repeal Effects PrevaningWage Laws and School Construction Costs Study on Davis Bacon and Virginia As presented at the Michigan Prevailing Wage Symposium March 1999 By Peter Philips, Ph.D. Summary: A study of school construction costs from 1992-1998 for 104 schools found that with the payment of prevailing wages average costs were $99 per square foot. when prevailing wages were not paid, the average cost was $104. DAVIS BACON AND VIRGINIA Page 1 of 15 THE DAVIS-BACON ACT: Are prevailing wages too high for Virginia's construction workers? The Davis-Bacon Act was passed in 1931 to maintain community wage standards,support local economic stability,and protect taxpayers from sub-standard work quality on federal projects.The Act requires that contractors on federally funded construction projects pay the locally"prevailing wage. Prevailing wages are determined by wage surveys conducted by the Department of Labor.If over 50 percent of workers in a trade in a local area earn the same wage,this is the prevailing wage.Usually large numbers of workers earn the same wage only when the wage is collectively bargained,meaning the workers are represented by a labor union.If no single wage is paid to a majority of workers,the prevailing wage is defined as the area average. Davis-Bacon wage decisions are issued in four categories:residential,building,heavy,and highway construction.The Building and Construction Trades Department studied Davis-Bacon wage decisions in the building category,which includes schools,for ten trades in seven of Virginia's largest cities and counties to determine whether paying Davis-Bacon wages for school construction would result in inflated construction pay in that state. City/County Area covered by Davis-Bacon wage decision Charlottesville Charlottesville City,Albemarle, Fluvannna,and Greene Counties Danville Danville City,Pittsylvania County Roanoke Roanoke and Salem Cities,Roanoke and Botetourt Counties Richmond Richmond City and Henrico County Prince William County Manassas and Manassas Park Cities, Prince William County Fairfax County Falls Church City, Fairfax County Norfolk-Virginia Beach Norfolk,Virginia Beach,Portsmouth,and Chesapeake Cities The following table lists Davis-Bacon wage decisions for ten trades in the seven metropolitan areas,as well as annual incomes derived from the hourly rates.The number of hours logged by a construction worker annually has never been accurately documented,either by government statistical agencies or independent researchers.The conventional wisdom in the construction industry puts yearly hours worked at 1500-1600.Because of weather and work availability,construction workers almost never work the typical 50 weeks per year.The upper limit of 1600 hours,or 40 weeks at 40 hours weekly,was used to calculate an annual income based on Davis-Bacon wages. Davis-Bacon Wage Decisions,April 1999 Hourly Wages and(Annual Incomes) Norfolk-Virginia Prince William Fairfax _ Beach Charlottesville Roanoke Richmond County County Danville Laborer 7.38 9.60 5.59 8.09 9.17 8.35 6.67 (11,808) (15,360) (8,944) (12,944) (14,672) (13,360) (10,672) Painter 10.33 10.51 11.23 12.50 13.10 8.00 (16,528) (16,816) (17,968) (20,000) (20,960) (12,800) Carpenter 11.24 11.98 8.62 13.29 14.93 15.37 9.97 (17,984) (19,168) (13,792) (21,264) (23,888) (24,592) (15,952) Cement Mason 10.09 11.87 -- 12.02 14.75 13.97 8.50 (16,144) (18,992) (19,232) (23,600) (22,352) (13,600) Ironworker 10.21 17.20 17.50 14.08 17.75 20.53 17.75 (16,336) (27,520) (28,000) (22,528) (28,400) (32,848) (28,400) Backhoe Operator 10.72 11.93 - 10.69 13.14 13.88 10.07 (17,152) (19,088) (17,104) (21,024) (22,208) (16,112) http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 2 of 15 Bricklayer 15.24 14.63 12.55 17.15 16.55 15.00 16.00 (24,384) (23,408) (20,080) (27,440) (26,480) (24,000) (25,600) Electrician 12.00 18.00 8.40 19.55 25.65 25.65 9.85 (19,200) (28,800) (13,440) (31,280) (41,040) (41,040) (15,760) Plumber 12.00 18.15 10.02 13.56 23.93 16.07 9.57 (19,200) (29,040) (16,032) (21,696) (38,288) (25,712) (15,312) Sheet Metal 12.07 12.09 7.56 10.54 23.68 23.68 9.83 (19,312) (19,344) (12,096) (16,864) (37,888) (37,888) (15,728) Collectively bargained rates are bolded. ARE DAVIS-BACON RATES TOO HIGH?Davis-Bacon Wages vs. Community Standards One what basis of comparison can Davis-Bacon wages be deemed unreasonable?Two community income standards were chosen against which to compare Davis-Bacon wages:average income in the community and the poverty line.How do incomes based on Davis-Bacon wages compare to incomes of the average worker in the community?Can a worker keep a family of four out of poverty on an income based on Davis-Bacon wages? When comparing incomes from Davis-Bacon wages to average incomes in the community,construction workers fare poorly.Davis-Bacon annualized income was compared to the average income of all workers in the appropriate Metropolitan Statistical Area(MSA)(i).In all seven local areas,construction workers earning Davis-Bacon wages had lower incomes than the average worker in over 80 percent of the 10 trades compared.In 87 percent(58 of 67)of individual wage decisions studied,workers earning Davis-Bacon rates earn less than the average local worker.(2) http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 3 of 15 Number of Occupations (of 10) where Davis Bacon Inca Than Average Income in the Community 10/10 10 �r 9f 1 9110 9 8f 1 8/10 8 7 6/? 6 5 4 3 ri - 1, a � T 1 m 1 0 Charlottesville Danville Roanoke Richmond Prince William Fairfax Comparing Davis-Bacon wages converted into annual income against the poverty line reveals even more surprising results.(3)While critics of Davis-Bacon claim that prevailing es are inflated almost a third of the wage rates studied would leave a family of four below the P g wages g poverty line.In every city,Davis-Bacon wages leave at least one trade below the poverty line. And wages for every trade except the bricklayers fell under the poverty line in at least one local area. (1) 1996 data,the most recent available from BLS,was adjusted by the CPI to 1999 dollars. (2)Ten trades in 7 cities/counties were analyzed yielding 70 wage decisions.However,Roanoke was missing wage decisions for 3 trades,leaving 67. (3)The state of Virginia uses the federal poverty guideline for determining financial eligibility for federal and state entitlement programs.In 1999,this poverty line was set at$16,700 annually for a family of four. httT)://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 4 of 15 Number of Occupations (of 10) Living in Poverty Under Davis Bacon Income 10 n 9 8!t 8 7 6 517 5 !10 4 1167 31% 3 2 1/1 1!1 1l10 1!1 1 0 Charlottesville Danville Roanoke Richmond Pnnce VWlliam Fairfax Norfolk-Virginia Average Beach In Charlottesville,none of the 10 trades studied had collectively bargained prevailing wages,and Davis-Bacon wages for every trade result in less annual income than the average worker.In addition,4 of 10 craft wages fall below the poverty guideline at 1600 annual hours. Annual Incomes in Charlottesville: Davis Bacon vs.Poverty Line and Average Community Income commun ity y measures $30,000 ®non-union wage 26.956 24,384 $25.000 19,200 192200 19,312 $20'000 17,984 17,152 16,144 16,336 16.528 16,700 $15,000 11,808 $10,000 $5 000 $0 Laborer Cement Ironworker Painter poverty Backhoe Carpenter Electrician Plumber Sheet Bricklayer average Mason line Operator Metal income Poverty line:1999 Federal poverty guideline,also used by Virginia State Dept.of Social Services Average community income:Average worker income for Charlottesville MSA.1996 BLS data adjusted to 1999- - Beach area is slightly better.Only laborer rates fall below the poverty line,but 8 out of 10 trades earn The situation 1n the Norfolk-Virginiag y Y less than the average worker in the area. httn://www-buildinsztrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 5 of 15 Annual Incomes in Norfolk-Virginia Beach: Davis Bacon vs.Poverty Line and Average Community Income ■community measures $35,� m union wage ®non-union wage 28,5M 28,800 29,040 $30,000 27,520 $25,000 23,408 18,992 19,098 19,168 19,344 $20,000 16,700 16,816 15,360 $13,000 $10,000 $S,000 ' $0 Laborer poverty&ne Painter Cement Backhoe Carpenter Sheet Metal Bricklayer Ironworker average Electrician Plumbers Mason Operator income Poverty line:1999 Federal poverty guideline,also used by Virginia State Dept of Social Services Average community income:Average worker income for Norfolk-Virginia Beach MSA 1996 BLS data adjusted to 1999. In Roanoke,Davis-Bacon wage decisions were made for only 7 of the 10 trades studied.Five of these 7 earn below the poverty line and 6 earn less than the average worker. Annual Incomes in Roanoke: Davis Bacon vs.Poverty Line and Average Community Income $30,000 28,000 m community measures a union wage ®non-union wage 25,786 $25,000 20,080 $20,000 1 16,032 6,700 13,792 $IS,13�1 13,440 12 096 $10,000 9,944 $5,000 $0 Laborer Sheet Metal Electrician Carpenter Plumbers poverty has Bricklayer average income Ironworker Poverty line:1999 Federal poverty guideline,also used by Virginia State Dept.of Social Services Average community income:Average worker income for Roanoke MSA.1996 BLS data adjusted to 1999. In Richmond,the laborer once again earns below the poverty line.Nine trades earn less than the average worker;only the electrician's union rate is higher. httn•//www_huildingtrades-ora/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 6 of 15 Annual Incomes in Richmond: Davis Bacon vs.Poverty Line and Average Community Income ■community measures $35,000 ■union wage ®non-union wage 30.112 31,280 $30,000 27,440 $25,000 22,528 21,264 21,696 19,232 $20,000 16,700 16,864 17,104 17y68 $15, 12,944 $10,000 $5,000 $0 Laborer poverty line Sheet Metal Backhoe Painter Cement Carpenter Plumbers ironworker Bricklayer average Electrician Operator Mason income Poverty line:1999 Federal poverty guideline,also used by Virginia State Dept of Social Services Average community income:Average worker income for Richmond-Petersburg MSA.1996 BLS data adjusted 10 1999. In both Prince William and Fairfax counties,the laborer is left below the poverty line.Unionized trades fare better,but 8 trades in Prince William and 9 trades in Fairfax County earn less under Davis-Bacon than the average worker in the Washington,D.C.MSA. e Prince William County: Annual Incomes�n P Davis Bacon vs.Poverty Line and Average Community Income $45.000 ®community measures 41,040 ®union wage 37,888 38.155 38.288 $40 000 m non-union wage $35,000 28,400 $30.000 26,480 23,6� 23,388 $25,000 20,000 21,024 $20,000 16,700 14,672 $15 000 $10,000 $5,000 $0 Laborer poverty Painter Backhoe Cement Carpenter Bricklayer Ironworker Sheet average Plumbers Electrician line Operator Mason Metal income Poverty line:1999 Federal poverty guideline,also used by Virginia State Dept.of Social Services Average community income:Average worker income for Washington,D.C.MSA.1996 8LS data adjusted to 1999. http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 7 of 15 Annual Incomes in Fairfax County: Davis Bacon vs.Poverty Line and Average Community Income $45,000 a community measures 41,040 ®union wag a 37.8W 38,155 $40,000 a non-union wage $35,000 32,848 $30,000 $25,000 24 000 24,592 2r',712 20,960 9� 22,21� 22,352 $20,000 16,700 $15,000 13.360 $10,000 $5,000 $0 Laborer poverty Painter Backhoe Cement Bricklayer Carpenter Plumbers Ironworker Sheet average Electrician line Operator Mason Metal income Poverty line:1999 Federal poverty guideline,also used by Virginia State Dept.of Social Services Average community income:Average worker income for Washington,D.C.MSA 1996 8LS data adjusted to 1999. Davis-Bacon wages for 8 trades in Danville leave workers below the poverty line.Only 2 trades,the bricklayers and the ironworkers,earn more than the poverty line and the average worker. Annual Incomes in Danville: Davis Bacon vs.Poverty Line and Average Community Income ®community measures $30,000 ®union wage 28 400 ®non-union wage 25,600 $25,000 23,249 $20,000 15,312 15,728 15,760 15,952 16,112 16,700 $13,000 13,600 12,800 10,672 �P $10,000 $5,000 $0 Laborer Painter Cement Plumbers Sheet Metal Electrician Carpenter Backhoe poverty line average Bricklayer Ironworker Mason Operator income Poverty line:1999 Federal poverty guideline,also used by Virginia State Dept.of Social Services Average community income:Average worker income for Danville MSA 1996 BLS data adjusted to 1999. http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 8 of 15 UNION RATES KEEP DAVIS-BACON EARNINGS ABOVE THE POVERTY LINE How do collectively bargained rates compare with average-based rates in keeping workers out of poverty?While 40 percent of the Davis- Bacon wages based on average wages keep workers'earnings under the poverty line,7 percent of collectively bargained Davis-Bacon rates do so.When Davis-Bacon wages are based on union contracts,communities benefit because workers are able to provide the financial "Davis-Bacon rates can fail to provide a living wage;without union protection and collective .Even"inflated g support a family needs en P PP Y bargaining,tax dollars can contribute to substandard compensation. Union vs. Non-Union Davis-Bacon Rakes: How do they compare to the poverty line? p t 00% an5 90% ®Below poverty line •Above overt line 80% 70% g 2152 d 60% o m 50% 0 m 6 40% c m a 30% 20% n 10% 0% Union rates Non-union rates HOURLY WAGES UNDER DAVIS-BACON ARE LOWER THAN PRODUCTION WAGES In comparing Davis-Bacon incomes to community standards,hourly Davis-Bacon wages were annualized because the poverty line and average worker income data are reported as annual figures.Because of overtime and uncertainty in work patterns,it is also desirable to compare Davis-Bacon wages to other hourly rates for a comparison that does not require an assumption of hours worked in a year. One of the few industry sectors for which hourly wage data is available locally is manufacturing.Comparing average hourly wages for production workers in the MSAs of the selected Virginia areas shows that at current hourly rates,production workers generally earn more than construction workers under Davis-Bacon.In 60 percent of wages studied,workers paid at Davis-Bacon wages earn less than production workers in their area.This is especially surprising given that construction workers typically cannot work for several weeks, sometimes several months,of the year. http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 9 of 15 Number of occupations (of 10)where Davis Bacon Hourly Wage Rate is Lower Than Average Production Wage Rate 10 9 1 8l10 8 7l10 7 Ol67 60% 6 5 4 10 3 2 1 - 0 Charlottesville Danville Roanoke Richmond Prince William Fairfax Norfolk-Virginia Average Beach Average production wage by MSA.1998 data from BLS. In Charlottesville,production wages are comparable to Davis-Bacon wages,falling just below the median of the 10 construction trades studied. Davis-Bacon Hourly Wages vs.Production Wages in Charlottesville $18 a production 15.24 $16 in non-union wage $14 12.00 1200. 12.07 $12 11.24 10.09 10.21 10.33 10.57 10.72 $10 7.38 $6 $4 $2 $0 Laborer Cement Ironworker Painter Production Backhoe Carpenter Electrician Plumber Sheet Metal Bricklayer Mason Worker Operator Average hourly production waye for 1998 for Charlottesville N1SA Data from BLS. Non-union Davis-Bacon rates in the Norfolk-Virginia Beach area are lower than production wages,while unionized wages are usually were higher. http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 10 of 15 Davis-Bacon Hourly Wages vs.Production Wages in Norfolk-Virginia Beach $20 1720 12.00 18.15 $18 ■production ®union wage $16 ®non-uruonwage 14.63 1495 $14 11.87 1193 1198 12.09 $12 1051 9.60 $10 $4 $2 Laborer Painter Cement Backhoe Carpenter Sheet Metal Bricklayer Production Ironworker Electrician Plumber Mason Operator Worker Average hourly production wage for 1998 for Norfolk Virginai Beach MSA Data from BLS. In Roanoke,construction workers fare quite poorly compared to those in the production industry.Only the ironworker,under a collectively bargained wage rate,earns more than the average production employee. Davis-Bacon Hourly Wages vs.Production Wages in Roanoke $20 17 50 $IS ■production E union wage $16 ®non-union wage 1489 $14 1235 $12 10.02 $10 8.40 8A2 7.56 $8 5.59 $6 $4 $2 $0 Laborer Sheet Metal Electrician Carpenter Plumber Bricklayer Production Ironworker Worker Average hourly production wage for 1998 for Roanoke MSA.Data from BLS. production workers.Workers in the other 8 Marl in Richmond only the bricklayer and electrician have higher hourly wages thanp Similarly y y g Y construction trades earn less. http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 11 of 15 Davis-Bacon Hourly Wages vs.Production Wages in Richmond $20 $18 ®Production 17,15 a union wage 16.16 El non-union wage $16 13.29 1336 14.08 $14 1123 12.02 $12 1054 10,69 $10 8.09 $4 $2 Laborer Sheet Metal Backhoe Painter Cement Carpenter Plumber Ironworker Production Bricklayer Electrician Operator Mason Worker Average hourly production wage for 1998 for Richmond-Petersburg MSA.Data from BLS. In Prince William County,wages for only 3 trades are less than average production wages. Davis-Bacon Hourly Wages vs.Production Wages in Prince William County $26 23.68 23.93 $24 ■production ®union wage $22 ®non-union wage $20 17.73 $18 16.55 $16 14.47 14.75 14.93 $14 13.14 1230 $12 $10 9.17 $8 $6 -, $4 $2 $0 Laborer Painter Backhoe Production Cement Carpenter Bricklayer Ironworker Sheet Metal Plumber Electrician Operator Worker Mason Average hourly production wage for 1998 for Washington,DC MSA.Data from BLS. In Fairfax County,production workers earn just below the median of the 10 construction trades.Fairfax and Prince William counties have more collectively bargained Davis-Bacon rates than most of the other local regions studied,possibly explaining why construction workers in these counties do relatively well compared to production workers. http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 12 of 15 Davis-Bacon Hourly Wages vs.Production Wages in Fairfax County $26 23.68 $24 ■production a union wage in non-union wage 2033 E20 $18 15.00 15.37 16.07 $16 13.10 13.88 1397 14.`C) $14 $12 $10 835 $6 $4 $2 so Laborer Painter Backhoe Cement Production Bricklayer Carpenter Plumber Ironworker Sheet Metal Electrician Operator Mason Worker Average hourly production wage for 1998 for Washington,DC MSA.Data from BLS. In Danville,8 of 10 trades earn less under Davis-Bacon wages than production workers in the same area.Only bricklayers and ironworkers have higher hourly wages. Davis-Bacon Hourly Wages vs.Production Wages in Danville $20 Eta e production 17.75 ■union wage E16 ®non-union wage 16.00 $14 12.85 $12 9.57 9.83 9.85 9.97 10.07 $10 8.50 8.00 6.67 $6 E4 $2 Laborer Painter Cement Plumber Sheet Metal Electrician Carpenter Backhoe Production Bricklayer Ironworker Mason Operator Worker Average hourly production wage for 1998 for Danville MSA Data from BLS. DAVIS-BACON WAGES NA TIONALL Y AND IN VIRGINIA ARE MOSTL Y NON-UNION RATES The most common misconception of the Davis-Bacon Act is that prevailing wages are synonymous with union rates.In fact,the majority http://www.butldmgtrades.org/davls.htm 7/8/fl2 DAVIS BACON AND VIRGINIA Page 13 of 15 of Davis-Bacon wage determinations are based on local averages,not collectively bargained,or union,rates.In 1994,29 percent of all US counties used prevailing wage rates taken only from union contracts,while 48 percent of counties used rates exclusively based on averages, and 23 percent of counties had rates based on both. Percentage of US Counties where Prevailing Wages are Union, Non-Union,and Mixed 60% ai 50% 0 40% c� 0 30% a� 20% ku 10% a 0% exclusively union exclusively non-union mixed In Virginia's largest cities and counties,Davis-Bacon building wages are largely non-union.In Charlottesville,Danville,and Roanoke,well under 10 percent of all wages decisions for building construction(not just the ten trades discussed earlier)are based on collective bargaining,and no areas have a majority of collectively bargained rates. Percentage of Davis-Bacon Building Wage Decisions that are Union Rates, by City/County �E so o% 70.0% 60.0% � e e li 40.0% 30.0% — 20 0% 10.0% Charlottesville Danville Roanoke Richmond Prince William Fairfax Norfolk-Virginia Beach Wage Decisions as of 4F99. Percentage of Davis-Bacon Building Rates that are Collectively Bargained Collectively Total wage Percentage union bargained wage decisions decisions City/County Charlottesville 1 24 4.2% Danville 1 18 5.6% Roanoke 1 15 6.7% http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 14 of 15 Richmond 4 29 13.8% Prince William 9 28 32.1% Fairfax 13 35 37.1% Norfolk-Virginia Beach 12 28 42.9% Because of the misconception that prevailing wages are union wage rates,the Davis-Bacon Act is often criticized for inflating the cost of P projects rojects by overpaying construction workers.Both assumptions,however,are wrong.Most Davis-Bacon wage decisions are not collectively bargained rates,and these wages are not excessive or unreasonable.In the Virginia wage rates studied by the Department, many of these wages result in incomes lower than the community average.Almost one-third of wage decisions leave workers unable to support a family of four at the typical 1600 annual hours worked in construction.Comparing hourly wages to production workers,Davis- Bacon rates are similar or slightly lower. RECENT FINDINGS ON THE EFFECTS OF PREVAILING WAGE LAWS Several recent academic studies have shown that prevailing wage laws do not have the negative financial impact presumed by critics of Davis-Bacon.Three papers by Peter Philips and Mark Prus conclude that prevailing wages do not increase construction costs.All three studies specifically studied school construction costs.Additional studies by Dale Belman and Paula Voos and Garth Magnum et al find negative impacts on government budgets when prevailing wage laws are not applied,due to lower quality workmanship and lost tax revenues. Presentation on Prevailing Wage Laws Peter Philips,Ph.D.,University of Utah n Prevailing Wage Symposium,March 1999 Presented at Michigan Preva , g g g YmP Studied school construction costs in Michigan with and without prevailing wage legislation(state prevailing wage law was suspended for two years). Found no statistically significant difference in school construction costs when prevailing wage was suspended. Prevailing Wage Laws and School Construction Costs:An Analysis of Public School Construction in Maryland and the Mid Atlantic States Mark Prus,Ph.D.,SUNY Cortland Prince George's County City Council,January 1999 Compared school construction costs in three states with prevailing wage laws that apply to school construction(Delaware,Pennsylvania, and West Virginia)with costs in two states without laws(North Carolina and Virginia). no statistically significant di fference in costs associated with controlling for regional differences in cost of living,found t y gn After c g g S, prevailing wage regulations. Square Foot Construction Costs for Newly Constructed State and Local Schools, Offices,and Warehouses in Nine Southwestern and Intermountain States Peter Philips http://www.buildingtrades.org/davis.htm 7/8/02 DAVIS BACON AND VIRGINIA Page 15 of 15 Legislative Education Study Committee of the New Mexico State Legislature,September 1996 Compared square foot construction costs in five states with prevailing wage legislation(New Mexico,Texas,Oklahoma,Wyoming,and Nevada)with four states without(Arizona,Utah,Idaho,and Colorado). • Found that elimination of prevailing wage laws does not save on public construction costs. • Prevailing wage laws appear to promote collective bargaining and apprenticeship training and may consequently lower public construction costs. Prevailing Wage Laws in Construction: The Costs of Repeal to Wisconsin Dale Belman and Paula Voos The Institute for Wisconsin's Future,January 1996 Examined fiscal impact of repealing both the Davis-Bacon Act and Wisconsin's state prevailing wage law. • Tax revenue losses would exceed the small wage savings resulting from prevailing wage law repeal,and Wisconsin would experience a net deficit.Repealing the state law would cost the state$8.4 million.Davis-Bacon repeal would cost the state$2.8 million.If both were repealed,the state would suffer an annual shortfall of$11.2 million. • Higher injury rates,declining apprenticeship training,and reduced minority access to training would also occur with repeal. Losing Ground:Lessons from the Repeal of Nine"Little Davis-Bacon"Acts Garth Magnum,Peter Philips,Norm Waitzman,and Anne Yeagle University of Utah,February 1995 Studied nine states(Alabama,Arizona,Colorado,Florida,Idaho,Kansas,Louisiana,New Hampshire,and Utah)that repealed prevailing wage laws. • Found in Utah(the state studied most intensively)that repeal had a negative effect on the state budget.The loss of construction earnings led to reduced income and sales tax revenues. • Cost overruns on road construction in Utah tripled after repeal. • In the nine states,repeal led to reduced training by 40 percent,reduced minority representation in training programs,and increased injuries by 15 percent. The authors conclude that,if the federal Davis-Bacon Act were repealed: • Federal income tax revenues would drop by at least$1 billion per year in real terms. • There would be 76,000 additional workplace injuries in construction annually,with more than 675,000 work days lost each year. http://www.buildingtrades.org/davis.htm 7/8/02