The nation's highway system is in a constant state of change. Every year, new capacity is added to the network, and older infrastructure is replaced, reconstructed, or improved.

Thousands of construction workers owe their livelihoods to building and repairing our country's roads and bridges. Thousands of other workers in the economy are employed producing the equipment, construction materials, and other supplies required by highway projects.

The expansion, renewal, and improvement of our highway system require substantial public spending. The highway capital outlays by all levels of government in 1998 totaled more than $51.6 billion.¹ A large portion of spending originates from the federal government. Every year, Congress authorizes federal-aid highway funds for highway construction projects undertaken by the states. Expenditure of federal funds administered by the Federal Highway Administration (FHWA) during 1998 amounted to more than $20.3 billion for a variety of federal-aid projects.

The main reason for these expenditures is to maintain and improve the safe and efficient transportation system that is so crucial to America's prosperity. There is an additional economic benefit, however, in the form of the stimulus to state and regional economies arising directly from highway construction spending. These benefits are in the form of employment and income for the construction industry and for the industries supplying equipment and materials for highway projects.

How large is the economic stimulus provided by federal-aid projects? How many people are employed directly and indirectly? How much new income is generated? These are the questions asked by an ongoing study conducted collaboratively by FHWA and the Boston University Center for Transportation Studies.

The goal of the study is to quantitatively assess the direct, indirect, and induced economic effects of different categories of highway improvement projects. For example, consider the employment effects of a bridge construction project. All the people who work for the project or its subcontractors --construction workers, site engineers, equipment operators, etc. -- make up the direct employment effect.

But many others are employed by companies that provide materials, products, and services that are purchased to support the project. These include steelworkers, truck drivers, miners, and many more. Employment- generation effects travel backward along a chain of "production inputs." For example, if the bridge requires steel, not only are jobs created for steelworkers but also for ore and coal miners. Keeping track of all these indirect employment effects requires a detailed accounting of the "input-output" structure of all the industries in the economy.

Finally, workers for whom jobs are created have new income to spend on consumer goods and services. This, in turn, creates new jobs in industries such as retail and personal services, food processing, and the manufacturing of consumer goods. Calculation of this induced employment impact requires details on how workers distribute their earnings across different categories of goods and services.

Form 47: Mining a Data Source

The first step in measuring the economic benefits of construction spending is to produce a profile of expenditures for different types of highway improvement projects. This includes the number of employees and how much they earn; the tons of aggregate, cement, steel, paving mix, etc.; the linear feet of guard rail and noise barriers; and the expenditure on lighting and traffic signals.

Fortunately, there is a data resource that provides this sort of information. Form 47, Statement of Materials and Labor Used by Contractors on Highway Construction Involving Federal Funds, must be filed by the contractor upon completion of all federal-aid projects with construction costs exceeding $1 million. It was created for the purpose of tracking the economic impact of federal-aid highway expenditures and contains information on employment and quantities of various types of construction inputs.

To estimate the direct, indirect, and induced employment effects, a database containing Form 47 records for more than 10,000 federal-aid highway construction projects has been assembled. (An earlier study extracted data from the Form 47 database to create the HIGHWAY1 model that is currently used by FHWA to estimate direct employment generation from federal-aid highway expenditures.) The average usage factors in table 1, defined as the number of units purchased per $1 million of construction expenditure, were calculated from that database.

Table1 - Usage Factors for Highway Construction Materials and Labor

Weighted averages for federal-Aid Highway Construction Contracts over $1 Million reported as Completed During calendar years 1988-1996

The usage factors provide a "recipe" that makes it possible to estimate the total input requirements for a project with a given total construction cost. It is not reasonable, however, to expect the same recipe to apply to all projects. The typical input profile may vary over time and across parts of the country, and it will almost certainly vary across different types of highway improvement projects.

As an example of regional variation in input structure, labor hours per $1 million total construction costs are calculated separately for each of the nine former FHWA regions that were in effect prior to field restructuring in 1999. (The analysis is organized according to the former regional structure due to the largely historical nature of the Form 47 data.) Labor hours per $1 million total construction costs are shown in figure 1. We can see that the labor requirements for projects of equal value are much higher in the Southeast and South Central states (former regions 4 and 6) than in other parts of the country. This may reflect differences in construction methods, terrain, or the mix of improvement types (as shown in table 2) across regions. The significance of this variation for estimating economic impacts is that the direct employment-generation effect will be considerably higher for projects in the South and Southeast.

Input structure may vary over time as well. For example, figure 2, using 1996-constant dollars, shows the quantities of structural and reinforcing steel per million dollars of construction expenditure for projects with different completion dates. There has been a significant decline in the intensity of structural steel use from the 1980s to 1997. On the other hand, the usage rate of reinforcing steel remained rather steady until 1996 when it increased sharply. Again, there are a number of factors that may explain this, including changing construction techniques, the mix of steel and concrete bridges, and shifts in the mix of improvement types. These variations over time will have significant implications for the calculation of indirect employment effects and for the overall impact on industrial activity.

Finally, different improvement types are expected to have different input structures. For example, a resurfacing project might be more intensive in the use of paving materials than a bridge project, which might be more intensive in the use of steel. This means that the mix of improvement types in total expenditure on highway projects will affect the calculation of total economic impacts.

To illustrate this, a sample of projects for which Form 47 data are available has been broken down by improvement types to show the variation in input structure. The results in table 2 show that there are significant variations in the intensity of labor and steel usage. For example, bridge replacement projects are almost 30 percent more labor-intensive than environment-related projects. The variations for steel usage are even greater because new bridge projects use more than seven times as much total steel per million-dollar expenditure as resurfacing projects. Since different input structures lead to different direct and indirect effects, the mix of improvement types is a key factor in determining economic impacts.

Capturing the Interconnections

Once a database of usage rates for individual projects broken down by year, region, and improvement type has been established, it will be possible to estimate the direct employment and purchases from materials-supplying industries associated with a particular type of project in a particular state. Estimating indirect impacts, however, introduces a new set of problems. If a project purchases lighting, for example, it will be necessary to know how much labor and other inputs are required to produce the lighting. How much labor, iron ore, and other inputs are needed to produce the steel used to manufacture the lighting? How much labor is needed to produce the iron ore? The series of connections goes back up the production chain.

It would seem almost impossible to capture all these interconnections. Fortunately, a set of input-output accounts that are available from the Department of Commerce have captured them already.^2,3

In these accounts, all production of goods and services is broken down into about 500 industries. For each industry, it is possible to see how much is purchased from every other industry for every dollar of production. The accounts include only a single industry to represent all highway construction activity. They will, therefore, be augmented using the Form 47 database to create a series of 14 separate highway construction industries for different types of improvements. The input-output relationships for these industries will also be adjusted for regional variations.

Figure 1 - Labor hours per million dollars (using 1996-constant dollars) of construction expenditure.

Applying the methodology of input-output analysis, direct and indirect impacts will be calculated for each improvement type.⁴ For example, it is possible to estimate the total labor hours and labor income generated by a proposed project by simply multiplying the total projected expenditure by the appropriate labor usage factors defined in hours and dollars, respectively. Using the full set of input-output accounts, it will be possible to estimate both direct and indirect employment effects.

In preliminary analyses, the combined value of direct and indirect employment-income generation was estimated to be 2.5 times as high as the direct effect for some improvement types. When induced effects are added, the multiplier effect will be even greater. Thus, an analysis of economic impacts that is limited to direct effects tells only a small part of the story.

Firgure 2 - Metric tons of steel per million dollars ( using 1996-constant dollars) of construction expenditure.

The Product

The final product of the study will be a software package to estimate the direct, indirect, and induced employment impacts of federal-

aid highway program expenditures and state matching funds. With this software, it will be possible to anticipate national and regional economic effects associated with changes in the distribution of funds across states and across types of highway improvement projects.

Furthermore, total employment impacts will be assessed in terms of the jobs created in specific industries. In this way, we can see which industries benefit -- both directly and indirectly -- from highway expenditures. Projections of employment and other inputsrequired on an industry-by-industry basis, perhaps as a result of changes in the program expenditure level or types of improvements undertaken, may also provide important insights into potential supply constraints. This is an issue of increasing policy importance in the current economic environment of high industrial capacity usage and concern over inflationary pressures.

Table 2 - Inputs per Million Dollars (using 1996-Constant Dollars) of construction Expenditure by Improvement Type*

*based on a sample of 1266 Form 47 records.

The Future

FHWA and Boston University researchers hope to extend the work in a number of fruitful directions. We hope to refine the analysis to produce estimates of employment generation at the state level. This will require extending the database to include road construction projects without federal aid. Sources of this information are being investigated.

Finally, examining the input-output structure of highway projects can tell us a lot about the rate of productivity improvement in the highway construction industry. The research team plans to explore a number of construction-related economic performance indicators based on Form 47 data.

References

1. Highway Statistics '98, (Table HF-10), Federal Highway Administration, Washington, D.C., 1999.
2. Anne M. Lawson. "Benchmark Input-Output Accounts for the U.S. Economy, 1992: Make, Use, and Supplementary Tables," Survey of Current Business, November 1997.
3. Anne M. Lawson. "Benchmark Input-Output Accounts for the U.S. Economy, 1992: Requirements Tables," Survey of Current Business, December 1997.
4. R.D. Miller and P.E. Blair. Input-Output Analysis: Foundations and Extensions, Prentice-Hall, 1985.

Dr. William P. Anderson is a professor of geography at Boston University and a member of the Boston University Center for Transportation Studies. He received his doctorate in geography from Boston University in 1984, and from 1983 to 1998, he was a faculty member in the Department of Geography at McMaster University in Hamilton, Ontario, Canada, where he was also director of the McMaster Institute for Energy Studies. His research interests are in transportation studies, economic geography, energy and environmental studies, and regional economic development. He has conducted studies for federal, state, and provincial agencies in the United States and Canada. He is a member of the editorial board of The Journal of Transportation and Statistics and Papers in Regional Science. He was formerly editor of Energy Studies Review and Canadian Journal of Regional Science.

Dr. Arthur C. Jacoby currently leads the research program for FHWA's Office of Transportation Policy Studies. His research interests are in the areas of infrastructure investment and pricing, statistical and econometric applications, and logistics systems analysis. He is an active member of the Transportation Research Board's Committee on Taxation and Finance and the Committee on Transportation Economics. He currently serves on the editorial board of Transportation Quarterly. Before joining FHWA in 1991, Jacoby was an associate professor of management at Embry-Riddle University, a research associate at The Pennsylvania Transportation Institute, and an instructor in the Department of Business Logistics at The Pennsylvania State University, from which he earned a doctorate in transportation economics and logistics.