Office of Operations Freight Management and Operations

Highway Operations Spending as a Catalyst for Job Growth (Page 3 of 5)

General Approaches to Analyzing Employment Effects

Key Measures

This analysis seeks to determine the total effect on employment levels resulting from spending on highway operations. The key to this definition is the cause and effect relationship between highway operations spending and changes in the level of employment. Identifying and quantifying this relationship is the essence of any assessment effort.

The effect on total employment level is traditionally considered the combination of the direct, indirect, induced, and enabling effects of the spending. These classifications of effects are distinguished by the nature of their linkage to the original cause.

Take as an example the output effect of a highway infrastructure project. The project will require purchases of, say, concrete, steel, and asphalt, and will generate income for, say, construction workers. It can also be assumed that the project will lead to improvements in the highway network. Accordingly, the direct effect in this example is represented by the output required to satisfy the completion of the project (e.g., construction jobs, steel plant jobs); the indirect effect is represented by the output required to support the project construction (e.g., iron mining jobs); the induced effect is represented by the output required to satisfy demand spurred by empowered consumers (e.g., retail jobs); and the enabling effect is represented by the output produced in response to an improved highway network due to the highway project (e.g., trucking jobs).

In other words, the direct, indirect, and induced effects relate to the highway project funding as the original source of employment, while the enabling effects relate to the highway itself as the source of additional employment levels. The former (three) can be thought of as demand-side effects and the latter (one) a supply-side effect. Demand-side effects result directly from funds spent on a project, while supply-side effects arise from the improvements in productive processes realized upon project completion.

Highway operations projects differ from highway investment projects in many ways, including the initial spending or purchase patterns associated with each. For example, in a highway operations project, relatively more operational spending may go to labor while relatively more investment spending may go to materials as compared to a highway construction project. Nevertheless, highway operations projects affect labor and output in a manner similar to highway investment projects. Specifically, highway operations projects result in direct increases in employment. Furthermore, highway operations spending normally entails the direct purchase of materials such as computers, phones, and other office supplies. Labor is needed to produce the needed materials. These are all direct impacts on labor. Additional goods (and services) must be produced (and provided) to facilitate the production of materials directly supporting highway operations projects. The production of such additional goods and services requires more labor, which represents the indirect impact of a project on labor. Further indirect impacts may ripple through the economy. The increased employment supporting highway operations projects increases the wealth of affected workers and thus leads to more personal consumption spending on goods and services. The labor employed in producing the goods and services for which there is increased demand represents induced employment. Highway operations projects also facilitate more efficient personal transport and freight movement that enable more productive activities throughout the entire economy and result in changes in overall employment. Figure 1 provides an overview of the effect of highway operations spending on employment.

Figure 1. Employment Effect of Highway Operations Spending. Flow chart showing highway operations spending leads to purchase of goods of services and payment of wages and salaries, as well as improved travel and freight movement, which results in efficiency gain. This spending leads to increased industrial output, which results in additional employment.
Figure 1. Employment Effect of Highway Operations Spending

The direct effect works through two channels: direct purchase of goods and services and direct hire of labor (arrows labeled "1" on figure 1). Except for the direct hire, all employment impact works through industry output. Purchase of goods and services, efficiency gain, and consumption spending all lead to increased industrial output, which in turn results in additional employment of labor.

Alternative Methods

Historically, researchers have used three methodologies to estimate the effect of highway investments on the economy, including changes in employment levels. They are: 1) the Input-Output (IO) approach, 2) the structural model approach, and 3) the aggregate production/cost function approach. The different approaches require different input data and produce output in different forms. Each of the three approaches can be adapted for use in assessing the effect of highway operations spending.[5]

Input-Output Analysis

The standard IO approach lends itself most readily to the measurement of direct, indirect, and induced effects. In fact, these effects are often defined as inherent IO concepts because each corresponds exactly to one IO quantity. The direct effect is the vector of goods and services on which the initial spending is made; the indirect effect is the sum of total commodity multipliers for industries minus the direct effect; and the induced effect is the sum of total commodity multipliers for households. Commodity multipliers here refer to all multipliers except for income multipliers in the IO system with an endogenous household sector. A commodity multiplier is the additional output of the commodity that is produced in response, directly and indirectly, to one dollar's worth of initial spending. Since the initial spending triggers a chain reaction that affects all commodity production, all commodity multipliers must be summed to measure the total effect corresponding to each category of spending. Because a total multiplier includes both direct and indirect effect, the indirect effect must be isolated. With the exception of direct hire employment, total employment effect is determined through output effects, as shown in Figure 1. Output impact is thus used to illustrate these IO quantities. Only the commodity multipliers are included in output impact[6].

The most crucial step in the IO method is the estimation of the direct output effect, which is often presented as a vector of goods and services directly purchased with the total spending. Multiplying a standard input structure or spending structure by the total spending on goods and services provides the vector of direct output effect. The spending structure differs between different types of operations. Using the direct output effect vector, the standard Leontief equation[7] can be used to estimate the indirect effect. The induced effect is calculated using the same procedure, with an adjustment to the commodity and industry-based IO tables to include a household sector and labor commodity in the intermediate section. The direct effect vector must also include an element for household income, which is equal to the direct spending on labor for highway operations.

The relationship between upstream and downstream industries may be quantified using a supply-side IO approach. While increased demand by a downstream industry certainly will stimulate more production by an upstream industry, increased production by an upstream industry may also stimulate production in the downstream industry. In the case of highway operations, an improved highway system may serve to reduce the level of friction between upstream and downstream industries and lead to reductions in the cost of finished goods, and ultimately serve to stimulate economic growth. This is particularly true in a supply-constrained economy. This supply-side effect may be called the enabling effect. As the following discussion will show, this supply-side measure does not reveal the full magnitude of the enabling effect.

Structural Model

The IO approach focuses on the linkages among industries in the form of input and output. Industry increases production in response to an increase in demand. Likewise, industry increases input requirements in preparation for or in response to increased demand for production. Although this interaction between and among industries is important, there are other influences on industry supply and demand. Individual companies may respond in different ways to an event such as a technological breakthrough, or the opening of a new market. This type of event can be thought of as an enabling effect, and is particularly important in assessing the effect of highway operations spending. Improvements to the highway system tend to result in market expansion. The IO approach does not capture this type of enabling effects.

A structural model is a more appropriate approach than the IO model. A structural model accounts for the interactivity between economic variables. In other words, a structural model provides a framework for definition of underlying economic relationships and makes possible the simultaneous determination of a wide variety of related variables.

For example, a more efficient highway system will likely result in reduced transportation costs. Industries will respond to cheaper transportation by reorganizing their logistics chains, buying more transportation services and reducing inventories and storage facilities. These changes will result in changes in employment levels across a number of industries. A structural model can be designed to capture this effect by including equations relating highway operational characteristics with variables for transportation cost and inventory and storage, for example. Specifying and estimating valid structural equations between and among these variables are challenging tasks, but a structural modeling approach recognizes the relationships.

In theory, a structural model can be designed for the whole economy or for any subset: a Computable General Equilibrium (CGE) model seeks to explain the former, and a partial equilibrium model for the transportation sector is an example of the latter. A CGE model is needed to capture the full magnitude of an impact felt throughout the entire economy, and may include a full set of IO relations as components. The specification of other components of a CGE model depends on specific issues at hand, data availability, and modeling strategies.

Production/Cost Function

A production function relates total output to total inputs. In the case of a transportation project, inputs include the elements discussed earlier in this report, including labor costs, private capital investment, and public transportation infrastructure and operations investment, among others, while outputs may be measured by the number of new employees, for example. For each unit of input, the production function determines the level of output to be expected. A cost function, on the other hand, relates total cost to total output (as measured by new jobs, for example.) For each unit of output gained, the cost function determines the level of cost per unit of output produced. Both production and cost functions are essentially single equation approaches. Each relates a variable affected by changes to another variable in one equation. Two or more production and/or cost functions may be combined (aggregated) to explain chain-reaction effects. Aggregate production functions are often used to examine private output growth and productivity changes. In contrast, cost functions are often used to estimate cost reduction and productivity effects and to derive factor demand.[8]

As noted above, both production and cost functions can be derived (estimated) for the economy as a whole, for an individual industry, or for a group of related industries. The key to the accurate assessment of the effect of highway operations spending on highway employment is the treatment of highway operations spending as an independent variable in the primary production function.

To review, in addition to the types of obstacles that must be overcome with any methodology, such as model specification and data development, the use of production and cost functions as an approach for employment change assessment poses further challenges. For example, the relationship between changes in inputs and/or cost (e.g., operations spending) and changes in output (e.g., ton-miles) must be properly determined before the ultimate output metric (e.g., new jobs) and the relationship between that metric and output level (e.g., ton-miles) can be defined.

Model Used in Analyzing the Effect of Highway Operations Spending on Employment

An IO model was used to estimate the effect highway operations spending has on employment. There are two reasons for choosing an IO model for this analysis. First, the model facilitates the estimation of direct, indirect, and induced effects on employment of highway operations expenditures. Second, the IO model produces consistent estimates within an integrated system, it provides industry-level details, and it serves as an information base as well as an analytical tool.

The IO model is one of the most frequently used methods in economic analysis and has been widely used in analyzing highway-related infrastructure and maintenance investments. Therefore, the use of an IO model in this analysis has the added benefit of providing estimates of employment generation that are conceptually comparable with those for other highway investment projects.

Model Structure

Using the Transportation Satellite Account[9] as its basis, the IO model is comprised of five interrelated components (Figure 2).

  • An IO "make" table provides information on goods and services produced by each industry in the U.S. economy.
  • An IO "use" table presents information on goods and services used by each industry in the U.S. economy. Thus, the "make" and "use" tables together provide a complete characterization of inter-industry relationships and serve as the basis for all IO analyses including assessment of employment effects.
  • The spending structure vector contains information on purchase or spending patterns for highway operations. This vector consists of the same number of commodities as the IO "use" table. Each value represents the share of spending on the corresponding commodity for highway operation purposes.
  • The employment-output ratios vector provides information on employment impact per unit of output for each industry. The vector contains a separate element for each industry in the IO "use" table (industrial multipliers) with each element's value equal to the unit employment level for the corresponding industry.
  • Average employee compensation value provides single or multiple parameters for converting payments to employees, including wages or salaries and benefits, into jobs directly created through highway operations projects.

Figure 2. Model Structure. Drawing of the IO model, composed of five interrelated components: the IO make table, IO use table, spending structure vector, employment-output ratios vector, and average employee compensation.
Figure 2. Model Structure

The left-most box and arrow in the graphic above are displayed with dotted lines to signify that basic input data from the underlying IO tables will not be transparent to the end users of the model. The reason for this is as follows: First, basic IO "make" and "use" tables do not have direct use to the end users except for providing input data to derive multipliers. Second, deriving multipliers from basic IO tables involves complicated matrix manipulations. Third, multipliers do not fluctuate very much over time provided that the analytical period is not too far removed from the IO base period. Nevertheless, the IO tables should be provided along with other components in Figure 2 to enforce data consistency and facilitate data updating.

Research Scope and Data Requirements

The above model includes all aspects of highway operations employment and provides general guidance. Because of limited resources, the induced and enabling effects were not considered in this analysis, as has been the case with most previous studies of similar kind. The employment effects estimated in this analysis include direct hires and indirect hires realized through the purchase of goods and services explicitly supporting highway operations. The analysis is further limited to highway operations expenditures on state-administered highways.

Based on the basic model and the research scope, the following are data required to estimate the employment effects of spending on highway operations.

  • Industrial multipliers
  • Spending structure of highway operations activities
  • Employment-output ratios by industry
  • Average employee compensation in each highway operations activity
  • Expenditures on highway operations

Except for industrial multipliers, which are available from the Transportation Satellite Accounts, there are no readily available data for other required elements. Therefore, the spending structures of highway operation activities, employment-output ratios by industry, and average employee compensation must be developed from a variety of data sources.

The development of the spending structures of highway operations activities is a major challenge, requiring input as to what goods and services are used in those activities. There are potentially two ways to overcome this difficulty. One way is to conduct a survey of spending structures underlying highway operations activities across states. This approach is time consuming and in fact impractical given resource constraints. The other way is to borrow the spending structures of similar industries from the Bureau of Economic Analysis (BEA) input-output analysis. The latter approach has been adopted in this analysis.

Another challenge is the development of employment-output ratios by industry. Although the effect on output comes directly from the analysis within an IO model, additional data are needed to link the effect to employment. Since IO models produce the effects at the industry level, it is highly preferable to build the output-employment links at the industry level as well. Fortunately, data on industrial output and employment are available for the development of employment-output ratios.

Finally, average employee compensation is required for model implementation. This element is independent of the underlying IO model. It is necessary for the estimation of the number of jobs directly generated by the payment of employee compensation.

Data Sources Used in the Analysis

The types and sources of data used in this analysis are listed below:

  • Expenditures on highway operations by activity: U.S. Federal Highway Administration (2001) Highway Statistics.
  • Industrial multipliers (2-digit IO industry level): U.S. Bureau of Transportation Statistics and Bureau of Economic Analysis (2000) Transportation Satellite Accounts for 1996.
  • Spending structure in highway operations: U.S. Bureau of Economic Analysis (1992b) Input-Output Tables; and U.S. Federal Highway Administration (1998) Test and Evaluation Project No. 28: Anti-icing Technology, Field Evaluation Report (Publication No.: FHWA-RD-97-132).
  • National employment by industry: U.S. Bureau of Labor Statistics: National Employment, Hours, and Earnings (Employees on Nonfarm Payroll by Industry (2-digit SIC) in 1996, seasonally adjusted).
  • Industrial output (2-digit IO industry level): U.S. Bureau of Economic Analysis (2000) Transportation Satellite Accounts for 1996, The Make of Commodities by Industry table.
  • Average annual wage by occupation: U.S. Bureau of Labor Statistics, Occupational Employment Statistics: 2000 National Occupational Employment and Wage Estimates.
  • Average annual wage by industry: U.S. Bureau of Labor Statistics, Occupational Employment Statistics: 2000 National Industry-Specific Occupational Employment and Wage Estimates (2000 National 3-digit SIC Estimates for SICs 071 to 903).
  • Benefit-compensation ratios: U.S. Bureau of Labor Statistics, National Compensation Survey: Employer Cost for Employee Compensation.
  • State and local government expenditure: U.S. Bureau of the Census, State and Local Government Finance Data.
  • State and local government employment: U.S. Bureau of the Census (2000b) State and Local Government Public Employment and Payroll Data.
  • Full-time sworn state police officers: U.S. Bureau of Justice Statistics (2000) Law Enforcement Management and Administrative Statistics, 1999: Data for Individual, State and Local Agencies with 100 or More Officers.
  • Average employee compensation in toll collection: Illinois State Toll Highway Authority (2000) Alternatives for Restructuring the Tollway System: A Report to the Governor.

Keys Steps in Data Processing

This analysis includes the direct effect of operations spending on employment and the indirect effect on employment through supporting industries. The number of jobs derived in this analysis is the number of full-time job equivalents rather than full-time job positions. The temporal scope of data used in this analysis is approximately 10 years. It is assumed that no changes in technology affecting the highway operations have occurred.

The major procedures undertaken in the analysis are as follows:

  • Converted SIC-based industrial employment into two-digit IO level industrial employment.
  • Developed the employment-output ratios by industry.
  • Adjusted the industrial multipliers from TSA for 1996 in order to match the developed employment-output ratios by industry.
  • Developed the spending structure, i.e., the input structure of each activity including traffic control operations, snow and ice removal, toll collection, other services, traffic supervision, highway safety and driver education, vehicle inspection and vehicle size and weight enforcement, general administration, and research and planning.
  • Disaggregated the total U.S. highway operations spending in each activity into corresponding categories of inputs.
  • Determined the average annual employee compensation for each activity.
  • Calculated direct hiring based on the derived labor cost and average annual employee compensation for each activity.
  • Calculated the indirect impact of each activity: First, calculated the impact of input purchases on industrial output. Second, multiplied the industrial output change with employment-output ratios to obtain the number of jobs created through the purchases of commodities and services.
  • Obtained the number of jobs by summing the number of direct hires and the number of jobs created through the purchases of goods and services.
  1. A general discussion of these approaches is provided below for conceptualization purposes. This report only provides estimates of traditional impact measures based on IO approach.
  2. Going from output impact to employment impact and handling direct employment impact through direct hire requires technical steps outside the IO framework. Discussion follows below.
  3. In the Leontief model, there are a finite number of industries producing a finite number of different products, such that the input equals the output, or consumption equals production.
  4. One disadvantage of both production and cost function approaches is that neither models the underlying mechanism that relates highway investments to cost reductions or productivity increases. Factors such as mobility, accessibility, and business inventory management are intuitive elements within the mechanisms. Importantly, however, solid conceptual and empirical understanding is currently lacking in that area.
  5. The Transportation Satellite Account (TSA) is an extension of the Input-Output tables of the United States. Developed by BTS and BEA, it identifies and measures transportation activities by establishments engaged in transportation as a secondary activity (in-house transportation) and those engaged in transportation as a primary activity (for-hire). TSA identifies detailed expenditure flows with their commensurate material and labor flows. Expenditures on labor, material, and other components for operations can be extracted from the TSA by industry and translated into number of jobs.

previous | next
Office of Operations