You are on page 1of 20

1

An Analysis of House Bill 1073: Proposed Legislation to Alter Prevailing
Wages in the State of Washington

Submitted to
Billy R. Wallace, Jr.
Political & Legislative Director
Washington & Northern Idaho
District Council of Laborers

By
Kevin Duncan, Ph. D.
Professor of Economics
Colorado State University-Pueblo

February 16, 2015

2

About the Author
Kevin Duncan, Ph.D.
Kevin Duncan is Professor of Economics, Hasan School of Business, Colorado State University-Pueblo
where he specializes in labor, and regional economics. He received his
Ph.D. from the University of Utah in 1987 and his B.A. from the University
of California, Riverside in 1981. Duncan is the author of over 60 academic
papers and applied regional projects and is the winner of several honors and
awards including the Provost’s Award for Excellence in Teaching, the
Provost’s Award for Excellence in Scholarship, the Outstanding Faculty
Member Award for the Hasan School of Business, the Enterprise Rent-ACar Student Choice Award for Excellence in Teaching, the Dean’s Advisory
Council Award for Outstanding Faculty Member, as well as the Dean’s Award for Excellence in
Teaching. His research on prevailing wage laws has appeared in leading international and national peerreviewed journals such as Construction Management and Economics, Industrial and Labor Relations
Review, and Industrial Relations. He has provided expert testimony to the Colorado, Hawaii, and
Vermont state legislatures on policies related to construction labor markets. His research was referenced
by the California Senate President pro Tem, Darrel Steinberg, in support of SB7 (2013) that extends the
payment of prevailing wages on public works to charter cities. He has also provided data and analysis to
the Legislative Auditors Office during the review of Minnesota’s prevailing wage law. He has authored
numerous economic impact studies that examine the effect of California’s pharmaceutical industry,
Amtrak’s Southwest Chief, America’s Cup Races in San Diego, project labor and local hire agreements
(in California, Colorado, and Hawaii), state and municipal prevailing wage laws (in California and the
City of San Jose), the Colorado State Fair, CSU-Pueblo, the installation and operation of wind energy
towers, the nonprofit sector, as well as the impact of the proposed Colorado Amendment 61. He has
served on the Advisory Board for Economic Impact Analysis of the Colorado Nonprofit Association. He
teaches regional economics where his students learn economic impact analysis.

Table of Contents
Executive Summary……………………………………………………………3
Introduction: Current Prevailing Wage Policy and Proposed Changes.….4
Cost Impact of Prevailing Wages……………………………………………..9
Economic Impact of Prevailing Wages………………………………………17

3

Executive Summary
House Bill 1073 requires the use of a random stratified sampling methodology in the
determination of prevailing wage rates for state-funded construction and establishes fines for survey
recipients that do not submit requested wage data. Random stratified sampling is advantageous when
groups within the overall population differ significantly. This survey technique requires that each group
is sampled in its proportion to the overall population. The challenges in meeting this requirement and
using random stratified sampling in the determination of prevailing wages are illustrated with data and
other information on Washington’s construction industry. Surveying requires careful consideration and
resources. HB 1073 does not provide any guidance regarding the identification of subgroups or other
information about the construction industry that would be useful in random stratified sampling. If the
sampling is not conducted properly, prevailing wage rates will not be accurate and complaints will arise.
The current survey method used by the Department of Labor and Industries avoids the challenges,
expense, and problems associated with switching to a random stratified sampling approach. Concerns
about the current survey method can be addressed by increasing survey response rates. Rather than
introduce a new approach that fines recipients for not responding to survey requests, the legislature may
consider other methods of encouraging prevailing wage survey participation under the current method.
House Bill 1073 is motivated by the assumption that a decrease in prevailing wage rates is
associated with a decrease in construction costs. This assumption is not supported by research that is
based on statistical analyses of construction cost data. The preponderance of peer-reviewed research
indicates that prevailing wages do not cause higher construction costs. Other studies indicate that
construction costs do not decrease when prevailing wages decrease, or when the wage policy is repealed.
Research shows that construction wages and construction worker productivity are linked. When wages
decrease, unskilled construction workers replace skilled workers, less capital equipment is utilized, and
material and fuel costs increase in ways that offset the cost savings of lower wages. Information from the
U.S. Census Bureau, the most reliable and publicly available data on construction costs, indicates that
labor costs are approximately 18% and 25% of the overall costs of building schools and highways in
Washington. These are the types of projects most affected by the state’s prevailing wage law. The data
and research suggest that, since labor costs are a low percent of total construction costs, relatively small
changes in labor productivity and construction efficiency are needed to offset higher prevailing wages.
The purpose of Washington’s prevailing wage policy is to protect local wage standards from
being undercut when the infusion of government spending attracts low-wage contractors from other areas.
In this sense the prevailing wage law creates a level playing field by requiring that all contractors pay the
same wages. Economic impact studies indicate that prevailing wage laws are associated with increased
use of local contractors and construction workers. The spending of these individuals contributes to local
retail and service businesses, to additional area employment, and to local tax revenue. Prevailing wages
redirect tax dollars back into the local economy in a way that benefits businesses and employees that are
not directly related to the construction industry. In this way prevailing wages can be considered built-in
economic development policy. Research indicates that weakening the state’s prevailing wage law will
not reduce the cost of state-funded construction, but will reduce the economic benefits that are enjoyed by
the citizens of Washington.

4

Introduction: Current Prevailing Wage Policy and Proposed Changes.
Washington’s Prevailing Wage Policy is partly modeled after the federal Davis-Bacon
Act.1 The purpose of these acts is to create a wage floor to ensure that construction workers will
not see their wages and benefits undercut as a result of government spending practices. The
infusion of state or federal dollars into an area, along with a process that rewards low bids, may
depress wages by attracting contractors from other areas. These contractors may undercut local
wage standards by importing lower paid workers or by offering less pay to local workers. The
prevailing wage floor protects construction workers’ pay and benefits and establishes a level
playing field for contractors who are bidding on government projects. The State of Washington
currently uses a majority/average rule in determining prevailing wages. If the same wage rate is
paid for the majority of hours worked for a detailed job classification in the largest city in a
county, that rate is the prevailing wage. If there is no majority wage, then the average wage rate,
weighted for hours worked, is the prevailing wage.2 Under House Bill 1073, the current
majority/average method used in determining prevailing wage rates would be replaced by a
method where prevailing wages would be proportional to the percentages obtained from the
random stratified sample.3

1

See a description of the state law at: http://www.lni.wa.gov/IPUB/700-032-000.pdf. See The Davis-Bacon Act
Protecting Wage Equality Since 1931. Accessed at:
http://www.dol.gov/whd/programs/dbra/Survey/conformancefaq.htm.
2
See Washington State Legislature, WAC 296-127-019, Survey Methodology, accessed at:
http://apps.leg.wa.gov/WAC/default.aspx?cite=296-127-019.
3
For example, if 60% of the wages for the hours worked in a detailed job classification are the same (say, $20 per
hour) and the remaining 40% of wages paid for the same job classification are the same (say, $10 per hour), under
the current policy, the prevailing wage would be $20 per hour since this is the majority wage. Under HB 1073, the
prevailing wage would be proportional to the percentages of high and low wages, or it would be $16 = ($20 x 0.6) +
($10 x 0.4).

5

House Bill 1073, an act relating to improving the accuracy of the prevailing rate of wage,
would require the use of a random stratified sampling methodology in the determination of
prevailing wage rates on state-funded construction.4 Random stratified sampling is advantageous
when sub-populations, or groups within the overall population differ significantly. This
technique is widely used in political polling. For example, if a racial minority group tends to
vote differently than the majority, and if data from the U.S. Census Bureau indicates that the
minority represents 30% of the population, common sense and random stratified sampling tells
us that 30% of those surveyed should be from the minority group. A challenge in extending this
survey method to prevailing wages is the scarcity of accurate and current data needed to
determine the proportions of different wage earning groups in the construction industry. While it
is relatively easy to use demographic information from the U.S. Census to identify percentages
for minority groups, deriving the same information about the construction industry would require
the Washington Department of Labor and Industries to engage is substantially more surveying
and data collection.

Random stratified sampling is not a short-cut. For this sampling technique to be
effective, it is very important that the each group is sampled to its proportion to the overall
population. As is illustrated below, it is very challenging to obtain this information for the
Washington construction industry. 5 If the survey is not conducted properly, prevailing wage
rates will not be accurate and complaints will arise. Surveying requires careful consideration and
resources. House Bill 1073 does not provide any guidance regarding the identification of

4

See http://lawfilesext.leg.wa.gov/biennium/2015-16/Pdf/Bills/House%20Bills/1073.pdf.
Problems associated with the subjective decisions involved in identifying the subgroups for prevailing wages based
on random stratified sampling have been addressed by a previous Department of Labor and Industries statistician.
See the briefing paper by Miriam Moses accessed at:
http://rebound.org/pages/papers_detail.cfm?pageid=18&paperid=9.
5

6

subgroups or other information that can be used to identify wage differences in the construction
industry.

To illustrate the challenges associated with the use of random stratified sampling in the
determination of prevailing wages in Washington, consider a simple example of stratification by
contractor size. This kind of grouping would be useful if wage rates vary substantially between
large, medium, and small contractors. The first challenge is to determine the percentage
breakdowns for large, medium, and small contractor establishments within a county.6 This
information can be obtained from the U.S. Census Bureau’s County Business Patterns.7 For
example, in Kittitas County, there are 181 construction contractors. There are 151
establishments with 1-4 employees, 23 contracting companies with 5-9 employees, and 6
companies with 10-19 employees. There are no contracting companies in this county with more
than 19 employees. The data for this county indicate that about 83% of contractors are very
small (those with 1-4 employees). However, in King County, only 69% of contractors have
between 1-4 employees and there are numerous contractors with more than 19 employees.
Under random stratified sampling, prevailing wages in Kittitas County would be influenced by
the large (83%) of very small contractors with 1-4 employees. Since the percentage of very
small contractors in King County is lower (69%), the prevailing wage rates in this county should
be based on this lower rate. This example illustrates how prevailing wages can be influenced by
the percentages of small contractors in a county. But, since the distribution of large, medium,
and small contracting companies varies between counties, the Department of Labor and

6

For example, if 20% of all firms are large, 30% are medium, and 50% are small, and if 100 firms are to be
surveyed in the county, then data should be collected from 20 large firms, 30 medium, and 50 small firms. The
firms to be surveyed in each group should be randomly selected.
7
Data from the County Business Patterns can be accessed at: http://www.census.gov/econ/cbp/index.html.

7

Industries would need to collect random stratified samples (based on varying percentages of
large, medium, and small establishments) from each county. The surveying needed to obtain this
information would demand considerable resources.

This simple example can be used to identify other problems associated with the use of
random stratified sampling. The data used in the example above are from 2012, the most recent
information available from County Business Patterns. Since that time the number of contractor
establishments has changed in Washington. A unique effect of the Great Recession was the
substantial decrease in the number of business establishments, particularly in the construction
industry. For example, before the last recession there were 201,340 contracting establishments
in Washington’s construction industry.8 By June of 2012 (about the time the data from the
County Business Patterns was collected), the number had fallen to 132,363. The most recent
data from June of 2014 indicates that the number of construction establishments has increased to
153,627. As the number of contracting firms has increased since the time of the collection of the
County Business Patterns data, it is very likely that the distribution of large, medium, and small
contractor establishments has also changed. Consequently, the percentages of large, medium,
and small contracting establishments used in determining prevailing wages under this method
will not be timely or accurate.

There are numerous other issues associated with the use of random stratified sampling.
The percentages of large, medium, and small contractor establishments available at the county
level may not reflect the distribution of large, medium, and small contractors that participate in
state-funded construction projects. For example, if prevailing wages in Kittitas County are

8

See the Quarterly Census of Employment and Wages (U.S. Department of Labor, Bureau of Labor Statistics)
accessed at: http://www.bls.gov/cew/.

8

influenced by 83% of contractors that are very small, and if very small contractors do not
participate in state-funded construction in this county, the prevailing wage rate will not be
accurate. The data described above are based on all construction within a county. Data
regarding the size of specialty trade contractors are not available. However, these data are
needed to determine prevailing wages for detailed job classifications. The example presented
above is based on one stratification (contractor size), but additional sub-groups may be needed.
Given the differences in wages earned by union and nonunion construction workers, additional
stratification may be desired that identifies contractors that are, and are not signatory to
collective bargaining agreements. While it is possible to collect data on contractor establishment
size, data on unionization rates are not available at the county level or at the level of detailed job
classifications.9 In this case, and in many others where data on group percentage breakdowns are
not available, the Department of Labor and Industries statistician would be ‘flying blind’ when
determining prevailing wages. Survey groups may also be identified based on the number of
contracts awarded, corporate income, and a variety of other factors. As more groups are
included the sample size in each group will become smaller since House Bill 1073 establishes a
7.5% net sample.10

9

The unionization data that is available at the state level for the construction industry is provided with a warning of
small sample sizes and the recommendation that the information be used cautiously. See unionstats.com accessed
at: http://unionstats.gsu.edu/. This issue illustrates a fundamental challenge associated with random stratified
sampling. If the percentages of union and nonunion construction workers for a detailed job classification in a county
are not known, then there is no guidance regarding the appropriate percentages used in determining prevailing
wages. Consequently, the sample is random as opposed to random and stratified.
10
HB 1073 calls for an initial random survey of 30% followed by a stratified survey of 25% of the initial
respondents, or 7.5% (25% x 30%). The issue of small sample sizes in also mentioned in the testimony by former
Department of Labor and Industries statistician Miriam Moses accessed at:
http://rebound.org/pages/papers_detail.cfm?pageid=18&paperid=9 .

9

The irony is that the current survey method used in the determination of prevailing wages
is capable of avoiding the challenges, expense, and problems associated with switching to a
random stratified sampling approach. With high survey response rates under the current method,
if there is no majority wage rate, the average prevailing wage will be proportional to the rates
paid by large, medium, and small contractors, or between union and nonunion contractors, etc.
Furthermore, this information can be collected without the subjective decisions required of the
Department of Labor and Industries statistician to identify sub-groups. Under the current method
all parties involved in state-funded construction have the right and responsibility to participate in
wage surveys. If a party feels that prevailing wages are not proportional to the true distribution
of wages in a county, this party can initiate improvements by participating in the survey and
encouraging others to do the same.
The issue of basing Washington’s prevailing wages on a random stratified sample has
surfaced previously.11 It is time to put the issue of random stratified sampling to rest unless the
legislature is prepared to fully fund a complete survey. Rather than introduce a new approach
that fines recipients for not responding to survey requests, the legislature may consider other
methods of encouraging prevailing wage survey participation under the current method.

Cost Impact of Prevailing Wages.
House Bill 1073 appears to be motivated by the assumption that lower prevailing wage
rates will be associated with lower construction costs. This assumption is not supported by peer-

11

See Senate Bill 5248 from the 2003 legislative session.

10

reviewed research that is based on examination of construction cost data and other publicly
available information on the construction industry.
First, labor costs are a low percent of total costs in the construction industry. The most
reliable data on construction labor costs can be obtained from the U.S. Census Bureau’s
Economic Census of Construction.12 These data are derived from a survey of construction
contractors in every state, every five years. For example, data from the most recent Economic
Census of Construction indicates that labor costs are approximately 18.2% of the net value of
construction for commercial and institutional building construction in Washington. 13 This is the
category that includes school construction. Also, labor costs are about 24.5% percent for
highway, street, and bridge construction in the state. These data are consistent with U.S. Census
Bureau information from other states. For example, Professor Peter Philips reports that labor

12

See the U.S. Census Bureau, Economic Census of Construction, Construction: Geographic Area Series: Detailed
Statistics for Establishments, accessed at:
http://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=ECN_2007_US_23A1&prodType
=table.
13
The Economic Census of Construction for 2007 does not report labor costs as a percent of total costs. This ratio
must be calculated based on other data. Here, labor cost as a percent of total construction cost is derived by dividing
total construction worker payroll, plus proportionally allocated total fringe benefits, by the net value of construction
work. The net value of construction is based on the value of work completed by a contractor, less the value of work
subcontracted to other contractors. The Economic Census of Construction defines construction worker payroll as
the gross earnings paid in the reporting year to all construction workers on the payroll of construction
establishments. It includes all forms of compensation such as salaries, wages, commissions, dismissal pay, bonuses,
and vacation and sick leave pay, prior to deductions such as employees' Social Security contributions, withholding
taxes, group insurance, union dues, and savings bonds. See Construction: Geographic Area Series: Detailed
Statistics for Establishments: 2007. Accessed at:
http://factfinder2.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=ECN_2007_US_23A1&prodTyp
e=table. The Economic Census of Construction defines the net value of construction as the receipts, billings, or
sales for construction work done by contractors, less the value of construction work subcontracted to others. The net
value of construction does not include contractor business receipts from retail and wholesale trade, rental of
equipment without operator, manufacturing, transportation, legal services, insurance, finance, rental of property and
other real estate operations, and other nonconstruction activities. Receipts for separately definable architectural and
engineering work for others are also excluded. Nonoperating income such as interest, dividends, the sale of fixed
assets, and receipts from other business operations in foreign countries are also excluded. See Construction:
Geographic Area Series: Detailed Statistics for Establishments: 2007. Accessed at:
http://factfinder2.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=ECN_2007_US_23A1&prodTyp
e=table.

11

costs range between 17% and 20% for selected building types in Kentucky.14 I have reported
elsewhere that labor costs are approximately 22% of the net value of construction for highway,
street, and bridge construction in Colorado.15 Consequently, changes in prevailing wage rates
affect a very small percent of total project cost.
Second, the preponderance of peer-reviewed research finds that prevailing wages are
unrelated to total construction costs. This report will focus on those studies that have examined
the effect of prevailing wages on building schools and highways indicates that prevailing wages
are unrelated to construction costs. For example, in examinations of thousands of public schools
built in states with and without prevailing wage standards, professors Azari-Rad, Philips, and
Prus find that building schools with prevailing wages are no more costly than building
comparable schools without the wage requirement.16 In the early 1990s the Province of British
Columbia introduced a prevailing wage standard that allows for a unique “natural experiment”
regarding the introduction of the wage standard within a jurisdiction. This policy was
comparable in strength to the Washington state prevailing wage policy.17 The effect of this
policy has been extensively examined. For example, professors Bilginsoy and Philips compare
the cost of building public schools before and after the introduction of the British Columbian

See Peter Philips, “Kentucky’s Prevailing Wage Law: An Economic Impact Analysis,” January 2014.
See Kevin Duncan, “The Effect of Federal Davis-Bacon and Disadvantaged Business Enterprise Regulations on
Highway Maintenance Costs,” Industrial and Labor Relations Review, January, 2015, Vol. 68, No. 1, pp. 212-237.
Accessed at: http://ilr.sagepub.com/content/68/1.toc.
16
See Hamid Azari-Rad, Peter Philips and Mark Prus, “State Prevailing Wage Laws and School Construction
Costs.” Industrial Relations, 2003, Vol. 43, pp. 445-457 and Hamid Azari-Rad, Peter Philips and Mark Prus,
”Making Hay When It Rains: The Effect Prevailing Wage Regulations, Scale Economies, Seasonal, Cyclical and
Local Business Patterns Have On School Construction Costs.” Journal of Education Finance, 2002, Vol.27, pp.
997-1012.
17
See Kevin Duncan, Peter Philips, and Mark Prus, “Using Stochastic Frontier Regression to Estimate the
Construction Cost Inefficiency of Prevailing Wage Laws,” Engineering, Construction and Architecture
Management, 2012, Vol. 19, No. 3, pp. 320-334.
14
15

12

wage policy and report that schools built under the wage regulations were no more expensive
than schools that were not covered by the policy.18
Along with professors Philips and Prus, I have examined the effect of the British
Columbian policy on the cost and productivity of building schools. For example, we compare
the cost of building public schools covered by the wage policy to the cost of building private
schools that were not covered by the policy. Public schools were approximately 40% more
expensive to build than comparable private schools before and after the wage policy.19 One
explanation of stable construction costs with the introduction of prevailing wages is that the
productivity or efficiency of construction increases along with wage rates. We find evidence of
this trend. For example, average efficiency for all public school construction in British
Columbia was 95% during the early and mid 1990s. Construction efficiency on public schools
covered by the first stage of the SDFW was 87%. Technical efficiency on projects covered by
the expansion of the British Columbian wage policy, 17 months later, was 99.8%.20 These
results indicate that the introduction of this prevailing wage law was associated with an
interruption in the efficiency of construction. But, contractors restored overall efficiency in a
relatively short period of time.
In an examination of Pennsylvania’s prevailing wage requirement, Mr. Keller and
Professor Hartman find that the wage policy adds, on average, 2.25% to the cost of building

See Cihan Bilginsoy and Peter Philips, “Prevailing Wage Regulations and School Construction Costs: Evidence
from British Columbia.” Journal of Education Finance, 2000, 24, 415- 432.
19
See Kevin Duncan, Peter Philips, and Mark Prus, “Prevailing Wage Regulations and School Construction Costs:
Cumulative Evidence from British Columbia” Industrial Relations, 2014, Vol. 53, No. 4, pp.593-616.
24
See Kevin Duncan, Peter Philips, and Mark Prus, “The Effects of Prevailing Wage Regulations on Construction
Efficiency in British Columbia,” International Journal of Construction Education and Research, 2009, Vol. 5,
No.1, pp. 63-78.
18

13

public schools in the Keystone State.21 In an examination of 3,000 schools built nationwide,
professors Vincent and Monkkonen report that construction costs are from 8% to 13% higher in
states with prevailing wage policies.22 However, both of these studies have errors that result in
prevailing wage cost estimates that are too high. To estimate the effect of prevailing wages on
total construction costs, Keller and Hartman compare labor costs under open-shop conditions to
labor costs with prevailing wage rates, assuming that the number of labor hours and labor
productivity do not change with wage rates. This assumption is contrary to what research tells
us. For example, professors Blankenau and Cassou find that the use of skilled and unskilled
construction labor is very sensitive to wage rates.23 When construction wage rates increase,
more skilled and productive construction workers are used instead of less skilled workers.
Professors Balistreri, McDaniel, and Wong also find that when wages increase and more skilled
construction workers are employed, more capital equipment and machinery is used in
construction.24 Consequently, when construction wages increase, for whatever reason, more
productive workers are used along with more equipment. These changes also alter hours
worked. Therefore, the labor cost and prevailing wage cost effect reported by Keller and
Hartman is too high. On the other hand, professors Vincent and Monkkonen do not take into
account changes in construction costs over the business cycle in their measurement of the
prevailing wage cost effect. Professors Azari-Rad, Philips and Prus find that when the

See Edward Keller and William T. Hartman, “Prevailing Wage Rates: the Effects on School
Construction Costs, Levels of Taxation, and State Reimbursements.” Journal of Education Finance, 2001, Vol.27,
pp. 713-728.
22
See Jeffrey Vincent and Paavo Monkkonen, “The Impact of State Regulations on the Cost of
Public School Construction,” Journal of Education Finance, 2010, Vol. 35, No. 4, spring, pp. 313-330.
23
See William Blankenau and Steven Cassou, “Industry Differences in the Elasticity of
Substitution and Rate of Biased Technological Change Between Skilled and Unskilled Labor.” Applied Economics,
2011, Vol. 43, pp. 3129-3142.
24
See Edward Balistreri, Christine McDaniel and Eina Vivian Wong, “An Estimation of U.S. IndustryLevel Capital-Labor Substitution Elasticities: Support for Cobb-Douglas.” The North American Journal of
Economics and Finance, 2003, Vol. 14, No. 3, 343-356.
21

14

unemployment rate in a state increases (doubles) during a recession, construction costs decrease
by 21%. 25 Consequently, if states with prevailing wage requirements also have lower
unemployment rates, the prevailing wage cost estimate reported by Vincent and Monkkonen is
too high.
Less research has been conducted on the effect of prevailing wage rates on highway
construction costs. However, the results of this research are consistent with the preponderance of
research on school construction costs. I have compared the cost of federal and state-funded
highway resurfacing projects in Colorado.26 While state and federal projects are built to the
same quality and safety standards, highway projects funded by the federal government are also
covered by Davis-Bacon prevailing wage requirements and by Disadvantaged Business
Enterprise requirements.27 While federal highway resurfacing projects are more expensive, these
projects are also larger and more complex. When complexity and size differences are taken into
account, there is no statistically significant difference in the costs of state or federally funded
projects. Results of this study also indicate that the additional regulations on federal projects
have no effect on the level of bid competition, and important determinant of construction costs.
My other research, in progress, indicates that contractors do not lower their bids when they
switch from more regulated federally funded highway resurfacing projects, to less regulated

See Hamid Azari-Rad, Peter Philips and Mark Prus, ”Making Hay When It Rains: The Effect Prevailing Wage
Regulations, Scale Economies, Seasonal, Cyclical and Local Business Patterns Have On School Construction
Costs.” Journal of Education Finance, 2002, Vol.27, pp. 997-1012.
26
See Kevin Duncan, “The Effect of Federal Davis-Bacon and Disadvantaged Business Enterprise Regulations on
Highway Maintenance Costs,” Industrial and Labor Relations Review, January, 2015, Vol. 68, No. 1, pp. 212-237.
Accessed at: http://ilr.sagepub.com/content/68/1.toc.
27
This policy establishes targets for the participation of subcontracting firms owned by minority and socially
disadvantage groups in federally funded construction. See U.S. Department of Transportation DBE Program
accessed at: http://www.dot.gov/osdbu/disadvantaged-business-enterprise.
25

15

projects funded by the State of Colorado.28 This finding also provides evidence that
construction costs are not affected by federal Davis-Bacon and Disadvantaged Business
Enterprise policies.
Of particular relevance to House Bill 1073, is my current research on highway
resurfacing projects that examines the cost effect of a change in prevailing wages from union to
average wage and benefit rates.29 For example, from at least the mid 1990s to April of 2002,
prevailing wage and benefit rates for the detailed job classifications involved in highway
resurfacing projects in Colorado were based on union rates. From April 2002 until the next
prevailing wage survey in the fall of 2011, average wage and benefit rates prevailed. This
change applied to 11 of the 13 detailed job classifications involved in highway resurfacing and
represented an average 18% decrease in total hourly compensation for these categories. Despite
this substantial decrease in the overwhelming majority of the wages paid for highway
resurfacing, there was no corresponding decrease in the cost of federally funded resurfacing
work relative to comparable state-funded projects.
Other researchers have also found that construction costs do not decrease when prevailing
wage rates decrease, or when state-level prevailing wage laws are appealed. For example,
Professor Wial examined the effect of a change in Pennsylvania’s prevailing wage survey and
wage determination.30 Before the survey change in the mid 1990s, union wage and benefit rates
usually prevailed in most counties. After the change, union rates continued to prevail in some

See Kevin Duncan, “Do Federal Davis-Bacon and Disadvantaged Business Enterprise Regulations Affect
Aggressive Bidding? Evidence from Highway Procurement Auctions.” Currently under publication consideration
at the Journal of Public Procurement.
29
See Kevin Duncan, “Do Construction Costs Decrease When Davis-Bacon Prevailing Wages Change from Union
to Average Rates?” Working Paper, Colorado State University-Pueblo.
30
See Howard Wial, “Do Lower Prevailing Wages Reduce Public Construction Costs,” Keystone Research Center,
July, 1999.
28

16

counties, but switched to lower rates in other counties. Wials’ examination of these changes on
school construction costs indicates that, while lower wage and benefit rates were intended to
save taxpayers money, there was no measureable relative cost impact.
In an examination of construction costs in Kentucky, Michigan, and Ohio during periods
in the 1990s when prevailing wage policies for school projects changed within these states,
Professors Philips finds that there was no statistically significant difference in school
construction costs associated with a change in prevailing wage policies. 31 Professor Philips also
reports that the value added per construction worker, a measure of labor productivity, is 14%
higher in states with prevailing wage laws, construction job-related disabilities are 12% higher in
states without prevailing wages, and repeal of prevailing wages is associated with a substantial
decrease in the kind of apprenticeships that are associated with future productivity growth.32
Taken together, the studies examining the effect of decreases in or the elimination of
prevailing wages reveal that these changes are not associated with reduced construction costs.
Why would this occur? As described above, the research by professors Blankenau, Cassou,
Balistreri, McDaniel, and Wong indicate that as construction wages decrease, so does the use of
skilled construction workers as well as the use of equipment. Both of these changes tend to
decrease construction worker productivity. While wage rates decrease on state-funded projects,
when prevailing wages are decreased or eliminated, construction worker labor productivity
decreases in a way that increases construction costs.

All of these findings are reported in Peter Philips, “Kentucky’s Prevailing Wage Law,” January 2014.
When comparing construction industry outcomes in states with and without prevailing wages, it is important to
recognize that the differences cannot be entirely attributed to the wage policy. Rather, prevailing wage standards are
part of a set of integrated and complementary institutions that contribute to a construction workforce that is trained,
productive, stable, and where the construction industry finances more of pension and health benefits instead of
shifting these costs to the rest of society.
31
32

17

Another way to illustrate the effect of changes in prevailing wage rates on construction
costs is to examine component costs between states with and without meaningful state-level
wage policies. Along with Mr. Alex Lantsberg, I have used data from the Economic Census of
Construction to compare construction cost components between states with differing wage
policies. We find that in states with weak or no prevailing wage requirements, construction
worker labor costs and fringe benefits are each lower by approximately two percentage points
compared to states with average or strong prevailing wage policies.33 Value added per
construction worker is about 10% lower in these states with weak or no prevailing wages. The
combined costs of materials, fuels, and equipment rentals are approximately 3 percentage points
higher in states without meaningful prevailing wage standards. These data suggest that higher
material and fuel expenses may be a consequence of the increased use of less productive labor in
those states with less than average prevailing wage laws. Regardless, the data from the
Economic Census of Construction suggests that states without effective prevailing wage laws
have lower labor costs, but also have lower labor productivity and other construction cost
components that are higher.

Economic Impact of Prevailing Wages.
As explained above, the purpose of Washington’s prevailing wage policy is to protect
wage standards from being undercut by low-wage contractors from other areas. In this sense the
prevailing wage law creates a level playing field by requiring that all contractors pay the same
wage rates. Economic impact studies indicate that more local contractors and construction
workers are employed when prevailing wages apply. This creates a benefit to local economies as

See Kevin Duncan and Alex Lantsberg, “Building the Golden State: The Economic Impacts of California’s
Prevailing Wage Policy.” To be released by SmartCitiesPrevail.org, February 2, 2015.
33

18

local tax dollars, used to finance capital construction projects, are also used to employ more
contractors and construction from the area. The spending of these individuals contributes to
local retail and service businesses, to additional area employment, and to local tax revenue.
My examination of library construction in Santa Clara County, California indicates that,
when prevailing wages applied, 39% of contractors employed on these projects were located in
the county.34 When prevailing wages did not apply, only 23% of contractors had county
business addresses. Since local contractors are more likely to hire county-resident construction
workers, prevailing wages redirect local tax dollars into the economy. The examination of 16
library projects financed by the City of San Jose (located in Santa Clara County), indicates that,
if these projects had not been built under the city’s prevailing wage standard, economic activity
in Santa Clara County would decrease by approximately $11 million. Employment of countyresident construction workers would decrease by about 80 jobs. With reduced spending in the
county, employment in local retail and service industries would decrease by approximately 25
jobs. With a reduction in economic activity, county sales and property tax revenue would
decrease by about $128,000. This economic impact analysis is based on the IMPLAN inputoutput software. This is the leading economic impact software and is based on observed
spending patterns.35
As described above, Mr. Lantsberg and I have examined differences in construction
spending between states with average or strong prevailing wage standards to those states with
weak or without construction wage policies. In addition to the differences in wage and material
costs described above, states with average and strong prevailing wage laws have more

See Kevin Duncan, “An illustration of the Impact on the Santa Clara County Economy of Repealing the
Prevailing Wage Policy of the City of San Jose.” Submitted to Working Partnerships USA, February 11, 2011.
35
See http://implan.com/.
34

19

subcontracting conducted by in-state establishments. For example, 96.5% of subcontracting is
completed by in-state businesses in Western states with meaningful prevailing wage laws. In
those western states with weak or no prevailing wage requirements, in-state subcontracting is
92.4%. Higher levels of in-state subcontracting are associated with the retention of more state
tax dollars and increased economic activity in states with average and strong prevailing wage
laws.
Most states employ contract award processes based on the low bid. In the absence of
prevailing wage requirements, this practice is associated with market failure that drives
construction worker wages and benefits down. As explained above, states with weak or no
prevailing wage laws have labor costs and fringe benefit rates that are each lower by two
percentage points. On the other hand, the profits of contracting companies in these states are
higher by 0.5%. These data suggest that prevailing wage laws correct the market failure
associated with awarding contracts based on low bids by altering the distribution of wage and
profit income. The IMPLAN software can be used to measure the impact of this redistribution.
Because individuals with lower incomes spend more in a region, and more of the spending of
higher income earners leaks out of the area economy, higher wages and lower profits are
associated with increased economic activity. When construction workers earn higher pay,
spending increases in local retail and service industries which typically include a large number of
small businesses.
To illustrate the economic impact of the redistribution of wage and profit income, the
greater use of in-state subcontractors, and other spending changes associated with prevailing
wage laws, we consider California’s construction industry if this state were to switch from the
characteristics of the average state with at least average prevailing wage requirements to the

20

average state without meaningful prevailing wage laws. If this switch were to take place,
economic activity in the Golden State would decrease by $1.4 billion and state-wide employment
would decrease by about 17,500 jobs.36
Prevailing wages redirect tax dollars back into the local economy in ways that benefit
businesses and employees that are not directly related to the construction industry. In this way
prevailing wages can be considered built-in economic development policy. Legislators may
chose to weaken the state’s prevailing wage law by altering the survey method. If so, research
tells us that the cost of state-funded construction will not decrease, but productivity will, and jobrelated injuries in the industry will increase. The earnings of a greater number of construction
workers in the state will decrease as will economic activity. Workers and small businesses that
are not directly related to the construction industry will experience increased unemployment and
reduced sales. The earnings of a few construction contractors will increase. If the legislature is
interested in creating a benefit for a few at the expense of many, this can be achieved by
weakening Washington’s prevailing wage policy.

36

Others have also examined the economic impact of prevailing wage laws. See Alison Quesada, Frank Manzo,
Dale Belman and Robert Bruno, “A Weakened State: The Economic and Social Impacts of Repeal of the Prevailing
Wage Law in Illinois.” University of Illinois AT Urbana-Champaign, School of Labor and Employment Relations,
Labor Education Program.