November 7, 2013

November 7, 2013 | Issue Brief on Transportation

Additional Infrastructure Spending Would Employ Few New Workers

Prominent Members of Congress, such as Senate Budget Committee chair Patty Murray (D–WA), have called for substantially increasing infrastructure spending to create jobs. They claim that spending tens of billions of dollars repairing bridges and roads would significantly reduce unemployment and stimulate the economy.[1]

These calls misunderstand the nature of infrastructure construction work.[2] Infrastructure projects are capital intensive, not labor intensive. Road and bridge construction requires a relatively small number of highly skilled workers using advanced equipment and machinery. Across the U.S., just over 300,000 Americans work in highway, street, or bridge construction—less than the population of Wichita, Kansas.[3]

In total, such construction employs just 0.2 percent of all workers.[4] This includes less skilled employees on the worksites. Even doubling the number of jobs in the sector would have only small effects on overall employment.[5]

Highly Skilled Occupations

Infrastructure construction requires significant human capital as well as physical capital. Many workers on these jobs need advanced skills to effectively and safely use construction equipment. An unemployed residential drywall installer cannot simply start building bridges or highways.


Table 1 shows typical apprenticeship requirements for various jobs in infrastructure construction. It takes several years to fully train new employees. A grade and paving equipment operator, for example, needs three years and between 4,000 and 6,000 hours of on-the-job-training. A structural ironworker requires four years and 6,400 hours of on-the-job training.

Most Unemployed Workers Lack These Skills

The collapse of the housing bubble caused residential construction employment to drop sharply since 2007. The recession had a much smaller effect on infrastructure employment. Of the net 2.1 million job drop in construction employment between 2007 and 2012, only 55,000 occurred in highway, street, and bridge construction.

Unsurprisingly, then, very few unemployed workers have the skills necessary to work on infrastructure projects. Just 8.5 percent of the unemployed previously worked in an occupation heavily utilized in highway, street, or bridge construction.[6] Even fewer worked in the most highly skilled construction occupations. One-fifth of workers on highway, street, and bridge projects operate heavy construction equipment. Less than one in 200 unemployed workers previously worked in such jobs.

Similarly, Bureau of Labor Statistics data shows just 13,500 unemployed cement masons, concrete finishers, and terrazzo workers in the entire country.[7] The vast majority of the unemployed would require retraining before working on infrastructure projects.

Reallocating Jobs, Not Creating Them

Additional infrastructure spending would consequently employ relatively few unemployed workers. Instead, federal construction contractors would hire the skilled workforce they need away from private construction projects. New jobs created would come primarily at the expense of other jobs in the private sector.

Exactly this happened with the stimulus. Stimulus funds did not go primarily to unemployed workers. Instead, the government hired workers and firms with the necessary skills for their construction projects. Surveys found that more workers on stimulus projects were hired away from other companies than were previously unemployed.[8] As researchers at George Mason University found:

Six of the organizations we interviewed, primarily engineering firms, said that there was little or no change in their work level due to the stimulus. They were niche firms with services in high demand. When they took [stimulus] work, they were turning down other work. These six were an extreme version of what many firms told our teams: The lunch wasn’t nearly as free as advertised. Tradeoffs mattered, and skilled firms and workers were scarce even in a world of 10 percent unemployment.[9]

Additional infrastructure spending would do more to shuffle jobs around than reduce unemployment.

Infrastructure in Historically Good Shape

New federal infrastructure spending would provide questionable value for taxpayers. As a share of gross domestic product, infrastructure spending has remained fairly stable for the past two decades.[10] Further, the states best know their highway and bridge maintenance and rehabilitation priorities. Funneling resources from the states to Washington and back to the states promotes a Washington-centric approach to infrastructure needs. Spending decisions should be made by states based on the condition of existing infrastructure and future capacity needs—not by Washington on the basis of economic conditions.

The interstate highway system is wearing out and will require reconstruction, and states will face repairs or reconstruction of their bridges and roads. Yet the country’s infrastructure is not crumbling before the traveling public’s very eyes. These future infrastructure needs do not constitute a national crisis in need of massive federal cash infusions. In fact, America’s infrastructure quality has improved markedly over the past two decades.

The Federal Reserve Bank of Chicago analyzed Federal Highway Administration data and reports:

Since the mid-1990s, our nation’s interstate highways have become indisputably smoother and less deteriorated.… [They] were smoother in 2006 than in any other year since 1980.… In view of this finding, accelerated expenditures on improving road surfaces are unlikely to yield significant direct benefits unless they are carefully targeted to specific interstate segments that are in need of improvement.[11]

Similarly, the number of structurally deficient bridges has fallen steadily since 1992.[12] While 21.7 percent of bridges had structural deficiencies in 1992, by 2012 that number fell in half to 11.0 percent.

Repeal Davis–Bacon

Instead of spending more, Congress should clear red tape, prioritize current appropriations more effectively, and empower states to spend their federal gas tax dollars on priorities of their own choosing, not those of federal bureaucrats.

The Davis–Bacon Act, for example, requires contractors to pay union wage rates and use union work rules on federally funded construction projects. This inflates the cost of federal construction projects by almost 10 percent.[13]

Repealing the Davis–Bacon Act would allow for the repair of approximately one-tenth more bridges and roads at no additional cost to taxpayers. Eliminating the need to apply union work rules—such as maximum numbers of apprentices at a worksite—would also enable employers to train new workers more quickly.

Little Effect on Unemployment

Supporters argue increased infrastructure spending would create jobs and boost the economy. These arguments have little empirical justification. Infrastructure projects require more physical and human capital than brute labor.

Consequently, most workers hired on new federal construction projects would come from existing projects—not unemployment lines. Additional infrastructure spending would do little to reduce unemployment.

—James Sherk is Senior Policy Analyst in Labor Economics in the Center for Data Analysis at The Heritage Foundation. The author thanks Heritage research associate Emily Goff and intern Nick Dau for their assistance with the research for this report.

Technical Appendix

The Heritage Foundation used data from the 2012 Current Population Survey (CPS) and May 2012 Occupational Employment Statistics (OES) to estimate the proportion of unemployed workers who previously worked in occupations heavily involved in infrastructure.

OES data was used to identify the occupations that account for 80 percent of highway, street, and bridge construction employment, including every occupation that accounted for at least 2.5 percent of total highway, street, and bridge employment.

The CPS asks unemployed workers what occupation they previously worked in. However, the CPS uses a different occupational coding system than the OES, which employs the Standard Occupational Classification (SOC).

The Heritage Foundation used a Bureau of Labor Statistics crosswalk to match the occupations identified by the SOC to the corresponding CPS occupation code (see table below). Analysis of CPS data revealed that 8.5 percent of unemployed workers in 2012 previously worked in these occupations.

About the Author

James Sherk Research Fellow, Labor Economics
Center for Data Analysis

Show references in this report

[1]Jonathan Weisman, “Senate Passes $3.7 Trillion Budget, Setting Up Contentious Negotiations,” The New York Times, March 23, 2013, (accessed October 31, 2013).

[2]This paper abstracts from the debate over whether government spending has a multiplier effect in the economy. Whether it does or not, spending on infrastructure creates fewer direct jobs than other projects would.

[3]U.S. Department of Labor, Bureau of Labor Statistics, “Occupational Employment Statistics,” May 2012, (accessed November 5, 2013).

[4]Heritage Foundation calculations using data from the Bureau of Labor Statistics, “Occupational Employment Statistics” and “Current Employment Statistics”/Haver Analytics. Note that the payroll survey reported 133.7 million Americans with jobs in 2012.

[5]While additional workers would obtain jobs supplying materials for the projects, this would also apply if Congress spent the money elsewhere (such as defense procurement) or left the resources in the private sector to invest.

[6]Heritage Foundation calculations using data from the 2012 Current Population Survey. See appendix for details.


[8]Garrett Jones and Daniel Rothschild, “Did Stimulus Dollars Hire the Unemployed? Answers to Questions about the American Recovery and Reinvestment Act,” Mercatus Center, September 2011, (accessed November 5, 2013).

[9]Garrett Jones and Daniel Rothschild, “No Such Thing As Shovel Ready: The Supply Side of the Recovery Act,” Mercatus Center, September 2011, p. 12, (accessed November 5, 2013).

[10]Federal Reserve Bank of St. Louis, “Graph: Total Public Construction Spending (TLPBLCONS)/(10*Gross Domestic Product (GDP)),” (accessed November 5, 2013).

[11]Jeffrey Campbell and Thomas Hubbard, “The State of Our Interstates,” Chicago Fed Letter, No. 264 (July 2009), (accessed October 31, 2013).

[12]Chris Edwards, “Infrastructure Investment,” Cato Institute, August 2013, (accessed November 5, 2013).

[13]Sarah Glassman et al., “The Federal Davis–Bacon Act: The Prevailing Mismeasure of Wages,” Beacon Hill Institute, February 2008, (accessed November 5, 2013).