Dumb and Dumber State Construction Policies

By Stephen Herzenberg, Third and State

I’ve got an idea: let’s employ low-wage, low-skill, and sometime out-of-state workers on small and medium-sized state-funded construction projects, with no benefit to taxpayers and negative impacts on local economies.

Sound like a stupid idea? That’s because it is.

Here’s the backdrop: Pennsylvania’s prevailing wage law requires that workers on state-funded construction projects be paid a wage in line with what most other workers in their trade are paid within a certain geographical area.

Research in peer-refereed academic publications shows that the law could be called the quality construction law because it helps ensure the use of skilled workers on state projects. Where prevailing wage laws exist, training investment, worker experience, wages, benefits, and safety levels are all higher than where these laws do not exist.

Overall construction costs are the same with or without prevailing wage laws. The prevailing wage law, however, makes it impossible for contractors that employ low-wage, out-of-state workers to win bids on state projects: it ensures that jobs go to local workers, who spend their money at local businesses.

More middle-class jobs, stronger local economies, higher quality construction, no cost to taxpayers: what’s not to like?

Unfortunately, some members of the Pennsylvania Legislature seem unwilling to leave well enough alone. Through House Bill 1329, these lawmakers want to make the prevailing wage law to apply to less state-funded construction work. How so? By exempting projects of less than $185,000 from prevailing wage standards. Currently, the law applies to all state-funded projects of $25,000 or higher.

Why one wants a threshold at all is not clear; eight states don’t have one. They have, instead, a clear policy of supporting a high-wage, high-skill approach to all state-funded construction.

Of the 32 states that have a prevailing wage law, only three have a threshold as high as that proposed in House Bill 1329.

Raising the threshold may not be as stupid a policy as eliminating the prevailing wage law altogether, or carving big parts of public construction (e.g., school projects) out of the law. But it’s still dumb.

For a complete list of our research and commentary on Prevailing Wage see our Prevailing Wage Issue page.

Prevailing Wage Opponents Fail to Look at the Research

The Final Part of a Three-part Series on Prevailing Wage by Mark Price and originally published at Third and State. Read Part 1. Read Part 2.

In the first two posts of this series, I explained why the numbers being tossed around by advocates of repealing prevailing wage don’t add up. I explained that the claims of cost-savings are not based on any actual experience and that they represent the result of laughable hypothetical, or “what if,” calculations. 

This leads to the most important point that the Pennsylvania School Boards Association, the Pennsylvania State Association of Boroughs, the Harrisburg Patriot-News Editorial Board and others keep missing: we can do much better than a hypothetical when assessing the impact of prevailing wage laws.

There is a body of research that examines construction costs (and other construction outcomes, like safety, training investment, wages, benefits, etc.) in states with and without prevailing wage laws as well as in states that eliminated prevailing wage laws. We don’t have to conjecture what “might” happen: we can look at what did happen. The preponderance of the evidence shows that prevailing wage laws do not raise construction costs.

Back in the late 1990s, Pennsylvania actually ran this real-world experiment itself — we lowered our prevailing wage levels, particularly in rural areas. That means we can look at what happened to construction costs. What happened is the same thing that has happened in other places — lower prevailing wages did not translate into lower construction costs. 

Specifically, the Keystone Research Center’s 1999 study of this late 1990s Pennsylvania policy experiment examined changes in public school construction bids when Pennsylvania’s prevailing wages were lowered substantially in rural areas. Keystone found no association between the number of occupations in which the prevailing wage was lowered and the price per square foot of school construction bids. If anything, construction bids appeared to go up more in areas where prevailing wages were lowered more.

Advocates of repeal often point to sympathetic construction managers in the public sector who testify, based on their expertise, that prevailing wage laws raise costs. Not only did the Keystone study find no statistical evidence of a cost difference during the period wages were lowered, but the study highlighted two revealing instances of construction managers making wild predictions that just didn’t come true:


The recent experience of two Pennsylvania school districts show that even increases in legally mandated prevailing wage and benefits rates do not necessarily increase public construction costs. In March 1999, after two months of legal uncertainty about required prevailing wage levels, [the Pennsylvania Department of Labor and Industry] began issuing prevailing wage rates that were higher than the 1999 rates. The Blue Mountain School District, in Schuylkill County, was planning to renovate its high school.  In April 1999, the school district’s construction manager estimated that construction costs would increase by about $670,000 as a result of the higher prevailing wage and benefit rates. But when bids for the project were opened on May 6, the low bids, which were expected to be about $15.1 million, came in at only about $13.8 million, almost 9 percent below the anticipated level. And in April, bids for a middle school construction project in Tamaqua, which used the same prevailing wage and benefit rates as the Blue Mountain bids, also came in under budget estimates.


Of course, anecdotes pro or con pale in comparison to careful statistical examination of large-scale data sets on actual construction costs. A study published in the Journal of Education Finance in spring 2002 explored the dependence of school construction costs across the United States from mid-1991 to mid-1999 on factors such as the state of the economy (measured by the level of unemployment), the size of the school, the season, and the existence of a prevailing wage law. The analysis found that public school construction costs:

  • rose 22% when the unemployment rate declined by half; 
  • fell 2.5% for bids accepted in the spring compared to bids accepted in the fall;
  • fell by 4.7% with a doubling of the school size, indicative of modest “economies of scale”; and 
  • did not go up or down a statistically significant amount based on the presence of a prevailing wage law. 

Another article from the Journal of Education Finance explored the impact of the establishment of prevailing wages in British Columbia at about 90% of the collectively bargained wage. This analysis, looking at a wide range of variables that potentially impact school construction costs, found that there was no statistically significant change in construction costs following establishment of a prevailing wage. 

In Michigan in the 1990s, school construction costs did not differ significantly during a period when the prevailing wage law was suspended temporarily compared to the period before and after. 

The reason researchers don’t observe differences in cost associated with prevailing wage laws is that higher wages in construction tend to reflect higher productivity. Family-sustaining wages, health coverage and good pensions attract and retain workers, leading to an accumulation of what economists call “human capital” — know-how that allows a skilled trades worker with years of experience to problem solve and do the job more quickly and right the first time. This know-how also translates into lower costs due to less need for supervisors and the higher retention of experienced workers which lowers recruitment and screening costs. Higher wages also promote the use of labor-saving technology and management practices that keep per-square-foot costs low.[1] 

While research finds that state prevailing wage laws do not significantly raise construction costs, these laws do lead to more investment in workforce training, lower injury rates, and higher wages and benefits (click here for a review). Thus, prevailing wage laws tend, over time, to lead to a more skilled and experienced workforce that is less likely to leave the industry, compensating for higher per-hour wage and benefit costs.

For a comprehensive review of the research literature on state prevailing wage laws, I highly recommend this work by Nooshin Mahalia. Unlike the Cato Institute journal, which the Patriot-News Editorial Board told us it relied on in supporting weakening the state prevailing wage law, Mahalia’s piece represents a full and careful review of all the literature. The Cato journal article, by contrast, ignores articles in peer-reviewed academic economic journals and relies on … wait for it … hypothetical calculations that support the ideological predisposition of the Cato Institute against regulation.

OK, we get it, that’s the Cato Institute’s excuse. What’s the excuse of the Pennsylvania School Boards Association or, for that matter, the Patriot-News Editorial Board, for ignoring the most credible research and evidence?


[1] Steven G. Allen, “Unionized Construction Workers are More Productive,” Quarterly Journal of Economics, 99(2) (May 1984):251-274; Kevin Duncan and Mark J. Prus, “Prevailing Wage Laws and Construction Costs: Evidence from British Columbia’s Skills Development and Fair Wage Policy, in Hamid Azari-Rad, Peter Philips, and Mark J. Prus (eds), The Economics of Prevailing Wage Laws (London: Ashgate Publishing Limited, 2005); Dale Belman and Paula B. Voos, “Prevailing Wage Laws in Construction: The Costs of Repeal to Wisconsin” (Milwaukee: The Institute for Wisconsin’s Future, 1995.)

Prevailing Wage Opponents Fail Labor Market Statistics 101

Part Two of a Three-part Series on Prevailing Wage by Mark Price, originally published at Third and State. Read Part 1.

The overwhelming weight of evidence based on the actual cost of public construction projects shows that prevailing wage laws do not raise costs. Therefore, advocates of repealing the law in Pennsylvania ignore this evidence. Instead of “evidence-based policy,” we have “lack-of-evidence-based policy.” Go figure.

Repeal advocates use a hypothetical calculation that makes assumptions about cost, rather than empirically examining the relationship between higher wages and total construction costs. (As discussed here, even these hypothetical cost estimates don’t make sense once you apply real world data to how much labor costs represent of total construction cost.)

Another key ingredient in the hypothetical calculations used by proponents of repeal is the claim made most recently by the Pennsylvania State Association of Boroughs (PSAB) that “the prevailing wage is 30 percent to 60 percent higher than the average wage for the same occupation.”

This claim is based on an update of a flawed calculation by the Commonwealth Foundation. It compares the prevailing wage levels by trade as set by the Pennsylvania Department of Labor & Industry with the average wages for construction occupations reported in Occupational Employment Statistics (OES). The prevailing wages are 30% to 60% higher than the OES averages. 

The problem is, the Commonwealth Foundation/PSAB calculation is the proverbial apples-to-oranges comparison: it measures different portions of the construction industry.

OES data include wages paid to workers employed in the residential construction sector – smaller, less-complex projects than prisons, bridges, schools and other state-financed construction. Residential construction relies on workers less skilled and experienced than those needed for larger state projects.

Indicative of this skill gap in Pennsylvania is the fact that construction workers employed in nonresidential construction – most of which is private sector, not public – earn 52% more than construction workers in the residential construction sector. In other words, the gap between the occupational prevailing wages set by the Pennsylvania Department of Labor & Industry and average construction wages reported by OES reflects the wage gap between residential and nonresidential construction.[1] (See Table 1 below.)

Using OES data to estimate the potential savings from repealing the prevailing wage law greatly overstates the potential cost savings. Actual wage levels on state construction projects without prevailing wage laws would not fall 30% to 60%. Assuming such savings and assuming that any drop in wages would have no impact on worker skill and productivity lead to false claims of large savings from repealing prevailing wage. 

Of course, why use hypothetical calculations at all when we have actual data on what construction costs with and without prevailing wage laws? In my next post, I will review the research literature that examines actual cost data over time and between states to estimate whether prevailing wage laws influence total project cost.

On Wednesday: Prevailing Wage Opponents Fail to Look at the Research

Table 1. Average Weekly Wages in Residential and Nonresidential Construction in Pennsylvania
Construction Specialty Average Weekly Wage: Residential Average Weekly Wage: Nonresidential % Difference in Residential & Nonresidential Wages
Construction† $771 $1,170 52%
Building construction $796 $1,223 54%
Heavy and civil engineering construction ND $1,262  
Poured concrete structure $688 $943 37%
Steel and precast concrete $1,100 $1,126 2%
Framing $644 $985 53%
Masonry $636 $957 50%
Glass and glazing $888 $1,078 21%
Roofing $660 $942 43%
Siding $730 $946 30%
Other building exterior $818 $822 0%
Electrical and wiring $887 $1,232 39%
Plumbing and HVAC contractors $816 $1,245 53%
Other building equipment $879 $1,233 40%
Drywall and insulation $751 $1,093 46%
Painting and wall covering $639 $985 54%
Flooring $625 $988 58%
Tile and Terrrazzo $686 $994 45%
Finish carpentry contractors $718 $935 30%
Other building finishing contractors $691 $870 26%
Site preparation contractors $756 $987 31%
All other specialty trade contractors $696 $977 40%
Note. ND=No data
† Average weekly wages for residential and nonresidential construction are an employment weighted average of the average weekly wages paid in each subsector.
Source. Keystone Research Center based on 2010 Quarterly Census of Employment and Wages data.


[1] Beyond the fact that they measure wages in a construction market that bears little resemblance to the public construction market, OES data have other limitations that make them unusable and inappropriate for purposes of comparison with prevailing wage rates. At the county level, OES data do not distinguish between construction occupations within the construction industry (residential plus nonresidential) and those in other industries. Construction occupations in all other industries – e.g., utilities and manufacturing, repair industries and building services – are all lumped together with the construction industry itself in the OES. As a result, at the local level, even more than the state level, OES data measure wages in a pool of jobs quite different from those on state construction work. The OES survey does not collect information on compensation paid in the form of health insurance and pension benefits. To overcome this limitation, PSAB uses a national data source to assume that fringe benefits represent 30.4% of reported wages in Pennsylvania.

Prevailing Wage Opponents Fail the Laugh Test

Part One of a Three-part Series on Prevailing Wage by Mark Price and originally published at Third and State.

Prevailing wage laws have long operated nationally and in states as a check against the tendency of the construction industry to degenerate into destructive wage and price competition. Such competition can drive skilled and experienced workers from the industry, reduce productivity and quality, and lead to poverty-level jobs, all without saving construction customers any money.

In an exhaustive review of the research on the impact of prevailing wages on contracting costs, Nooshin Mahalia concluded:

At this point in the evolution of the literature on the effect of prevailing wage regulations on government contract costs, the weight of the evidence is strongly on the side that there is no adverse impact. Almost all of the studies that have found otherwise use hypothetical models that fail to empirically address the question at hand. Moreover, the studies that have incorporated the full benefits of higher wages in public construction suggest that there are, in fact, substantial, calculable, positive benefits of prevailing wage laws.

Although the weight of evidence suggests prevailing wage laws do not raise costs, advocates for repealing the law in Pennsylvania continue to repeat some version of the following:

Prevailing Wage law also harms taxpayers, as it forces them to pay higher labor costs on public construction projects. Construction companies forced to pay union-inflated wages and benefits will pay upward of 30 percent more in labor costs for identical work on private sector projects. This adds a little more than 20 percent to the cost of every taxpayer-funded construction project – resulting in an estimated $1 billion cost for state and local taxpayers each year.

– Matthew J. Brouillette
President & CEO of the Commonwealth Foundation
March 22, 2011 

What is the source of this 20% saving claim? One source is Nathan Benefield, the research director of the Commonwealth Foundation, in this 2009 blog post.  

Benefield compared wages as measured in Occupational Employment Statistics (OES) to the Pennsylvania Department of Labor and Industry’s prevailing wage and concluded that prevailing wage rates are on average 37% higher (in a future post, I will address this issue directly). To estimate how this difference will affect total cost, Benefield assumed that labor represents 45% of total cost.

There are several problems with this analysis. 

Today I will address the first: it fails the laugh test. According to the 2007 Economic Census of Construction, labor costs represent no more than 24% of total construction costs in Pennsylvania. In road and bridge construction, much of which is funded by the state and thus impacted by prevailing wage regulations, labor costs represent no more than 21% of total cost.

If labor were 45% of total costs, and cutting wages and benefits were to have no impact on worker productivity, it would be possible to achieve a 20% reduction in total costs with a 37% reduction in labor costs. 

But only in Commonwealth Foundation World do labor costs account for 45% of construction costs. Tables 1A and 1B below show that if you use actual data on labor’s share of total cost, the wage declines necessary to achieve a 20% reduction in total construction cost are impossibly – laughably – large.  

In all construction, for example, labor costs are 24% of total costs. With labor costs at 24% of project costs, wages must fall by 70% to wring 20% out of total production cost, from $36.30 to $10.89 per hour for a carpenter in Philadelphia – that’s below even what Benefield claims Philadelphia carpenters make when not on a prevailing wage projects (see Table 1A).

With labor costs equal to 21% of project costs, as on road and bridge construction, wages would have to fall 80% to lower total costs by 20% (see Table 1B). Cement masons in Dauphin County employed on a road project would see their wages fall from $34.76 per hour to $6.95 per hour (the minimum hourly wage is currently $7.25).

Table 1. Changes in total cost as a function of the share of labor cost
A: Benefield’s calculation assuming labor represents 30% of total construction cost in PA
  With P.W. Without P.W. % Change
Labor Cost $240,000 $72,000 -70.00%
Non-Labor Cost $760,000 $760,000  
Total Cost $1,000,000 $832,000 -20.19%
B: Benefield’s calculation assuming labor represents 21% of total cost in road & bridge construction
  With P.W. Without P.W. % Change
Labor Cost $210,000 $42,000 -80.00%
Non-Labor Cost $790,000 $790,000  
Total Cost $1,000,000 $832,000 -20.19%
Note. The calculation of percent change is sensitive to the point of reference. The percent change in labor cost is calculated relative to the labor cost with a prevailing wage. The percent change in total cost is calculated in reference to total cost without a prevailing wage. To switch the point of reference is unconventional but necessary to remain consistent with the manner in which Benefield made his calculations. Calculating the percent change in total cost in the cases above where total cost with a prevailing wage is the point of reference results in a 17% decline in total costs.

Recall that Benefield’s laughable savings projections are the basis for the claim that the state can save $1 billion per year. Don’t count on it. 

On Tuesday: Prevailing Wage Oponents Fail Labor Market Stats 101

Must Reads: State of The Union, Stimulus and Austerity Economics PA Style

A blog post by Mark Price, originally published at Third and State.

Tonight President Obama will deliver his State of the Union Address to Congress. We are expecting the President to recommend an extension through the end of 2012 of extended unemployment insurance benefits and the payroll tax credit. It looks as though a major theme in the address – besides the catch phrase “built to last” – will be conventional policies aimed at reducing inequality, such as increased spending/tax credits for education and training.

Education and training are important and fruitful means of reducing inequality, but they fall well short of what’s needed to reduce the degree of inequality we now face.  A more forceful step in the direction of reducing inequality would include raising the minimum wage and making it easier for workers to form and join unions. We don’t expect to hear the President call for either of those changes.

The President will propose paying for his new initiatives with higher taxes on wealthy households. As with education and training, restoring some sense of fairness to the tax code is a laudable goal but longer-lasting reductions in inequality will only come from policies that allow the pre-tax wages of more Americans to rise as the size and wealth of our economy grows.

Manufacturing, energy, job training and middle-class growth will be the cornerstones of President Barack Obama’s speech tonight as he takes to the nation’s grandest political stage for the annual address on the state of the union, according to senior advisers.

We are slowly getting details of a settlement of allegations of fraud by banks during the housing bubble. Dean Baker notes this morning that the deal is said to include immunity from prosecution for banking executives in exchange for mortgage relief paid for by investors (not the banks). It’s good to be a banker.

The Philadelphia Inquirer reports this morning that the association that represents construction contractors who mainly compete for work in the non-residential construction sector is expecting essentially no change in the number of workers they will employ in 2012. Non-residential construction makes up roughly two-thirds of all construction employment in Pennsylvania. Also of note in the article: 62% of Pennsylvania contractors surveyed reported relying on some stimulus-related work. Remember that factoid next time you hear someone claim stimulus spending had no effect on the economy.

Construction employment will go up – very slightly – in 2012, contractors predicted in a survey released Monday by the Associated General Contractors of America…

The survey notes that many contractors relied on stimulus-funding projects over the past years, but few expect to perform much stimulus-funded work in 2012.

In Pennsylvania, for example, 62 percent of those surveyed had stimulus work, with most of them assigning the majority of their workers to those projects. But in 2012, only one in five expects stimulus work.

More news of property tax hikes, teacher layoffs and larger class sizes – this time out of Dauphin County.

The Central Dauphin School Board Monday night approved a $155.4 million preliminary budget for 2012-13 that could mean higher taxes, larger class sizes or furloughs of as many as 50 district employees.

The Patriot-News Editorial Board notes that the asset tests for food stamps proposed by the Corbett administration are unwise and likely to punish many rural families.

Creating an asset test for food stamps in Pennsylvania is the wrong approach…

Given the economic woes many families are facing with at least one parent – sometimes both – out of a job, the car rule hardly makes sense. This is especially true in rural parts of the state. Reliable transportation is critical to achieving financial independence, and in many families that means parents having two decent cars to drive.

The other issue is the $2,000 limit in savings. Families struggling to get out of poverty are likely to be trying to save money, build up funds to help them pay off bills, make a security deposit on an apartment or catch up on mortgage payments. It makes no sense to compel people to potentially liquidate funds to be able to put food on the table.

Hunger is a problem in our state, and many people rely on food stamps to solve it.