U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000
FHWA > Policy > Transportation Investment: New Insights from Economic Analysis
The United States at the close of the 20th century is enjoying a resurgence of economic growth and vitality that scarcely could have been imagined a decade ago. In the past few years, growth in the nation's productivity (i.e., growth that is not accounted for by increases in the amount of labor or capital in use in the economy) has reached levels not seen in 20 years, levels that many economists feared would never be seen again. Alan Greenspan, the Chairman of the Federal Reserve System, is widely hailed as a national hero and is gracing the covers of our news magazines. Have the laws of economic gravity changed? Or must what has gone up come down again? As the American economy continues to expand, the limits of the nation's transportation infrastructure are being tested, calling into question the need for further investment. How much gravity will our transportation systems' limits have?
Transportation improvements foster improvements in the business environment. Cheaper transportation increases the size of markets. With larger markets, firms can realize greater economies of scale in production. Inputs to the production process may become cheaper because of lower transport costs. The size of the labor pool expands. Inventory management becomes easier. New land and new resources can be put to productive use, and greater specialization can occur.
With the building of the Interstate Highway System, the nation made a huge investment in transportation infrastructure during the 1950s and 1960s. As the Interstate Highway System neared completion in the early 1970s, transportation investment dropped, and it remained relatively smaller through the 1980s and 1990s. Since then, the nation's highway system has suffered increasingly from lower maintenance and increased traffic congestion, to which front-page stories in newspapers regularly attest.
The public clearly understands how much longer it takes to get to work today than it did 20 years ago. Is the public aware of the larger changes in economic activity spurred by the improvement of transportation systems? The slowdown in economic productivity that started in the 1970s (and that may or may not now be over) has been a puzzle for economists and the country for the past two decades. Once upon a time, economic historians considered the effect of canals and railroads on the economy. But it took until 1990 for research by David Aschauer, now of Bates College, and then work from Alicia Munnell of the Boston Federal Reserve Bank, to bring widespread attention to the potential for transportation systems, and public infrastructure more generally, to play a major role in promoting economic growth in the present.
Aschauer found huge benefits to the economy from public infrastructure investment and blamed the anemic performance of the economy in the 1970s and 1980s on the severe shortfall of such spending. His study and its clear call to action then came under intense scrutiny. Once put under a microscope, these claims lost a measure of their credibility; but the questions the study had provoked—why productivity growth had fallen and what effect public infrastructure investment, particularly transportation investment, had on the economy—did not go away.
Three years ago, the Eno Foundation held its first forum to address this second question. The Federal Highway Administration and the American Association of State Highway and Transportation Officials, through the National Cooperative Highway Research Program, asked the Eno Foundation to host a conference discussing the importance of transportation investment for the American economy. Transportation experts and policymakers from the private sector, public sector, and academia converged on Washington, DC, for a forum called Economic Returns from Transportation Investment.
The participants heard quantitative evidence of the importance of transportation investment from several parties. lshaq Nadiri presented the results of a nationwide study spanning four decades that showed solid rates of return for the Interstate Highway System. Randall Eberts of the Upjohn Institute provided corroboration with studies of the rates of return on highway investment for localities. Colin Gannon of the World Bank offered estimates that World Bank transportation investments also had high social value and generated substantial rates of return. Industry examples such as developments in logistics management at General Motors were cited to illustrate some of the changes brought about by infrastructure improvement. The group advocated bringing these kinds of findings to the attention of policymakers and the public.
Three years later, the completion of a new study by lshaq Nadiri presented an opportunity to conduct a second forum in which to revisit this important topic. The following report draws forth the most salient conclusions and intriguing speculations from that day.
In the first session of the forum, Nadiri presented the results of his most recent study of the contribution of highway capital to economic growth and productivity. In the second session, Arthur Jacoby and Charlie Han discussed their research exploring the connections between lshaq Nadiri's cost elasticity results and other transportation-related industry characteristics. Barbara Fraumeni described her recommendations for improving the quality of the highway capital stock data used by much of the empirical research. The third session featured analyses of state and regional highway investment programs including input and output methodologies and employment analysis. The final session of the forum tackled the nettlesome problem of communicating the results of research to the public and policymakers, particularly the issues of who to target and in what form the messages are best conveyed. The bulk of the report is divided into four parts, paralleling the events of the forum. After a brief conclusion, the appendices contain the text of selected papers discussed during the forum.
The growth of economic prosperity in America should not be taken for granted. For more than 100 years, from the end of the Civil War to the early 1970s, the American economy grew at a rate of nearly 3.5 % per year. In the 20 years that followed, through the end of the 1980s, the economy's rate of growth slowed by more than a percentage point to 2.4 % annually. Economic growth can be adjusted to account for changes in the use of labor in the economy to better measure the change in a nation's standard of living over a long period of time. This measure is called labor productivity, and it describes how much workers can produce in a given amount of time. When workers can produce more goods in the same amount of time, the savings in cost can be shared between the workers and the owners of capital. As a result, the standard of living rises. Unfortunately, while labor productivity grew at more than 2 % per year over that same 100-year period, labor productivity grew at only half that rate during the 1970s and 1980s.
The construction of the Interstate Highway System in the 1950s and 1960s coincided with a period of strong growth in the American economy. The nation's highway capital stock is a dollar value measure of the infrastructure in the national road and highway system. From 1952 to 1959, the nation's highway capital stock grew at an annual rate of 6.2 %. During the past 20 years of slower economic growth, the rate of addition to the nation's highway system has dropped markedly. Since 1982, the rate of growth has fallen to 1.2 % a year. Is this relationship between economic growth and the highway system merely coincidental? The increase in demand for transportation has continued unabated over the past 20 years. Congestion continues to increase, fewer large-scale transportation projects are started, and more funding for those projects that do start comes from state and local governments.
Recent research shows that investment in the national highway system has provided demonstrable benefits to the U.S. economy over the last several decades. However, public awareness of this benefit may be lacking. Part of the need for this forum is to communicate these promising research results in clear terms to the public. Ultimately, the country needs a vision of transportation investment for the postinterstate era. Two questions need to be successfully addressed in this forum:
Ishaq Nadiri, the Jay Gould Professor of Economics at New York University and a member of the National Bureau of Economic Research, explores the relationship between infrastructure investment and economic output. Nadiri presented the results of his most recent study of this complex relationship, Contributions of Highway Capital to Output and Productivity Growth in the U.S. Economy and Industries.
Nadiri explained that to understand research in this area, it is important to recognize the special characteristic of highway capital that sets it apart from private capital: Its influence is shared across firms in the economy. This special characteristic is called an externality. An externality is any kind of shared or additional effect of an activity on entities other than those who are direct parties to the activity. By capturing the externalities of highway capital—its effect on firms throughout the economy—studies measure the social rate of return from highway capital investment, that is, the value that the investment provides to society as a whole (rather than any private individual part of society). Calculating a social rate of return for such investments provides policymakers and the public with some ability to compare the relative value of government investment in different activities. Investments with higher social rates of return will provide more economic growth for a given amount of expenditure and benefit the nation most.
Ishaq Nadiri, New York University
Nadiri's study represents a significant advance in the field of applied econometric analysis. It is an important piece of evidence supporting the proposition that infrastructure investment matters economically. For years, economists have been struggling to achieve a consensus view about this relationship. Several wellknown studies in the late 1980s concluded that infrastructure investment was crucial to the rate of productivity growth. The influential works of David Aschauer and Alicia Munnell attributed most of the slowdown in economic productivity growth in the 1970s and 1980s to the lower rate of national infrastructure investment during this period. Before these studies appeared, few economists had empirically explored how public investment affected the nation's performance. However, when a flurry of new research followed the provocative Aschauer and Munnell findings in subsequent years, the new research showed that public infrastructure investment has an incredibly large effect on productive capacity. A number of methodological issues were raised about the modeling that underpinned the Aschauer and Munnell studies, and their strong conclusions were later cast into doubt by the academic community.
The Aschauer and Munnell approach to estimating the relationship between public investment and economic growth focused on using a production function as the basic model of behavior. In a production function, firms in the economy use labor and capital with a certain level of technological sophistication to produce specific amounts of output. Aschauer's insight was to recognize that firms use both private capital and public capital (highways, airports, power generation and transmission facilities, sewers and water treatment plants, and other such public infrastructure) in production. Looking at production aggregated across firms throughout the country in this way showed an extremely significant relationship between the amount of output the economy could produce and the level of public infrastructure available to the economy.
The main problem with the approach is that the results are hard to interpret. Does investment in public infrastructure lead to economic growth, or does growth give the public sector more ability to invest in infrastructure? The nature of the correlation is unclear. In growing economies, many of the underlying variables grow together at similar rates. The nature of the production-function approach and the aggregate nature of the data used in the Aschauer and Munnell studies did not address the issue of causation very well. Economic growth and infrastructure expansion may have occurred contemporaneously without any direct causal relationship or in response to some unexamined variables.
After taking into account the possible simultaneities and self-correlations in this model, several academic researchers concluded that infrastructure investment had a much smaller effect on the economy than Aschauer and Munnell concluded. Most of these subsequent studies focused on economic output in smaller geographic areas, usually states or regions but sometimes counties or metropolitan areas. More output in a small part of the country is less likely to lead to increased local infrastructure, because the Interstate Highway System program was a nationwide initiative. Indeed, studies conducted at smaller geographic levels found a much smaller economic influence for infrastructure investment. Unfortunately, studies performed on a smaller geographic scale leave open the possibility that the benefits of infrastructure in any given locality may spill over into neighboring regions. If such spillovers exist, then the benefits measured in these studies are too small. That is, the externality aspect of infrastructure investment is not fully captured.
Nadiri navigated through the problems associated with production functions by focusing on the cost of production rather than the level of productive output. Costs of production are clearer to interpret as being driven by changes in infrastructure rather than the other way around. Nadiri also focused on modeling specific industries rather than all economic activity in a specific geographic location to reinforce the interpretation of a causal relationship between public infrastructure and productivity. By encompassing industrial production throughout the nation, though, Nadiri made sure to capture all of the network spillover effects that might exist.
Nadiri's industry focus accounted for different industries using the highway capital they share very differently. Differences in highway capital use among industries have important implications for the way in which highway capital investment is financed. According to economic theory, those who receive most of the benefits of an investment should pay proportionally more of the costs. Nadiri's study explored these differences in the effect of highway capital on individual industries. Just as different industries use highway capital differently, their use of capital may have different implications for their patterns of employment, capital utilization, and materials use. These effects are important to the economy.
Nadiri's study started by separately estimating a cost function and a demand function for each industry. A cost function, like a production function, is a model of firm or industry behavior. A demand function is a model of consumer behavior with respect to different goods. In the cost functions used in this study, the cost for an industry depended on relative factor prices of labor, materials, and capital; quantity of output; disembodied technical change; and public capital services. Price increases for labor, materials, or capital increase industry costs, as do increases in industry output. Increases in public capital or technical change may lead to lower industry costs. In this study, Nadiri used a different functional form for the cost function than in his prior work, allowing for more complicated and expanded modeling. In the demand functions used for these industries, changes in demand varied with changes in the price of output relative to prices in the rest of the economy. Demand also varied with changes in aggregate income and population. As population and incomes grow, demand grows as well. When the price level for an industry increases, though, demand for the output of that industry falls.
Highway infrastructure enters this system of equations by affecting each industry's cost of supply. An investment in highway infrastructure may have several different effects on the cost of supply. First, it may directly allow firms to realize lower costs of production—called the "productivity effect." Second, firms may change their use of inputs (labor, materials, and physical capital) in production if highway infrastructure is either a substitute for or complement to labor, intermediate inputs, or physical capital. This is called the "factor-demand effect." Third, cost reductions may lead to price reductions for consumers, and these price reductions may lead to an increase in the demand for output. In cost studies of the effect of infrastructure, levels of output are typically held constant. In this work, prices were allowed to fall when firms lowered their costs of production as a result of productivity improvements. This change shifted the cost curve downward, and industry output expanded to meet the increased demand at the lower price. This "output-expansion effect" is an additional benefit to the economy not captured in previous research. The net effect of highway investment will be the combination of the productivity, factor demand, and output expansion effects. To ensure that the results attributed to highway capital investment were accurate, other kinds of public infrastructure (e.g., ports, airports, water treatment plants, and sewers) were included separately in the model. In this way, significant findings for the effects of highway capital on costs were not influenced by the effects of other public capital expenditures on costs.
Ishaq Nadiri, New York University (left), and Curtis Wiley, Indiana Department of Transportation
Nadiri used data aggregated into 35 different industry groupings. Data were consolidated from data on more than 80 sectors of economic activity and spanned the late 1940s to 1991. Combining such an expansive data set with the cost function model provided a level of comprehensiveness and detail unmatched by previous studies.
Nadiri econometrically estimated all of the industry cost and demand functions through a series of stages to generate output and cost elasticities with respect to highway infrastructure. Elasticities are measures of responsiveness between two variables: changes of a measured variable associated with a single percentage change in a base variable. These output and cost elasticities measure how output and cost would change when the highway capital stock is increased by 1 %. With output levels held constant, almost all (32 of 35) of the industries saw positive benefits (cost reductions) from highway capital investment. In addition to this productivity improvement, industries also realized the output expansion effect of highway investment. When output was allowed to vary, the drop in costs allowed industries to lower the prices of their products. This price reduction prompted an expansion of output. The net effect of highway capital on total costs for the industries then became negligible because the increased production costs from the expansion in output offset the decrease in costs from the productivity improvement. Consumers (rather than firms) effectively captured the benefit of the highway capital effect as prices fell and industry output expanded.
In the course of his econometric estimation, Nadiri also examined the relationship between highway investment and the use of capital, labor, and materials by industries. Looking at the interaction of highway investment with these inputs to production, in all of the industries studied, highway capital investment was found to be a substitute for labor and materials and a complement for private capital. Increases in the highway capital stock reduced the need for labor and materials while apparently spurring the demand for capital to create and expand additional business. When output was allowed to expand to the decline in prices in the model, the substitution effect of highway capital investment diminished for labor and materials, and the complementary effect of capital increased. The complementary relationship between capital and highway investment is important to note-investment in one benefits from investment in the other. The slight substitutability with labor and materials indicates that highway expansion leads to greater efficiency in the use of these inputs. This relationship may offer clues to the underlying responses by firms to changes in transportation infrastructure.
| 1960–1969 | 1970–1979 | 1980–1991 | Average 1960–1991 |
|
|---|---|---|---|---|
| Highway Capital | 54% | 27% | 16% | 32% |
| Private Capital | 16% | 18% | 17% | 17% |
| Interest Rate | 5% | 8% | 10% | 8% |
When the results of the model were inspected by industry, some of the industries benefited from highway investment more than others. Service and transportation industries were the chief beneficiaries of more highway capital stock, while manufacturing industries also gained, but to a lesser degree. The result suggested that an expanding service sector in the economy might make future highway investment more important, although further work needs to be done before coming to firm conclusions in this regard. The distributional results can have important implications for the form of finance chosen for future highway capital investment. Because the benefits fall in certain patterns, it maybe advisable for payments for taxes or other methods of finance to be more closely aligned with the gain of benefits.
Nadiri next combined the estimated cost elasticities with calculations of the opportunity cost of providing highway capital to generate rates of return for each industry. These results were summed across industries to derive the social rate of return for the highway infrastructure investment. The effects of highway capital investment were found to be very substantial. A 1 % increase in highway capital stock reduced costs by 0.08 % (producing a cost elasticity of -0.08). The average social rate of return over the study period was a healthy 29 %. As table 1 shows, this rate of return is clearly larger than the private rate of return over the same period, providing strong evidence for the value of this investment.
However, when broken down into shorter periods of time, the rates of return varied. Benefits were highest during the earlier development of the Interstate Highway system, when the nation was underinvested in highway capital. After the 1960s, as the system matured and more investment was made, the rate of returned declined. The results for the most recent years, the 1980-1991 period, show that the rate of return was roughly comparable with that of the private sector.
Nadiri also analyzed the different contributions of changes in exogenous demand, prices, and highway capital to the nation's productivity growth. Total factor productivity measures how output in the economy changes when the amount of labor and capital used in production stays constant. Total factor productivity is typically affected by determinants such as exogenous demand for goods and services, which depends on the size of the population and the level of income (as opposed to endogenous demand, which depends on the price of the product). Other determinants include the relative prices for the different goods used as inputs into the production process. The remainder of the change not explained by these two factors is usually attributed to autonomous technological change. Autonomous technological change represents the myriad improvements in products and processes that increase production as they accumulate over time. In this model, the highway capital stock is added as an additional factor affecting productivity.
In table 2, the main source of total factor productivity growth is change in exogenous demand. Over the sample period, exogenous demand accounts for roughly half of the total factor productivity change. Highway capital makes a sizable contribution to total factor productivity growth as well, roughly a quarter of the total. Relative price changes offer the least contribution, working slightly against growth in the study sample.
Nadiri's study presents a clear case for the historical value of infrastructure investment to the country's economic growth. The average social rate of return exceeds the average rate of return for private-sector investment. The study also finds a sufficiently large rate of return in the later years of highway investment to justify continued investment in the system, although the decline of the rate of return in the most recent period does raise some questions about the extent of the future benefits of highway capital.
One participant wondered if the smaller marginal benefits of more recent highway capital investment could be the byproduct of smaller proportional increases in investment spending in recent years. If so, larger levels of investment might generate higher returns. What remains unclear is whether appreciably larger investments could be made in the current U.S. highway system. Most believe that the system is mature enough that further large expansions would lead to smaller and smaller returns. The first highways in an area usually provide a much greaterbenefit than later ones, which tend to compete with the older highways.
Participants also speculated what the relationship should be between the rate of return on highway investment and the rate of return on private capital. If the rate of return on highway investment falls below the private rate of return, should that end most highway investment? Conversely, realizing an average rate of return lower than the private rate of return calls into question the advisability of investing tax dollars that could possibly be kept in private hands. One rejoinder is that the private rate of return is also an average and that private rates of return are also influenced by the risk inherent in the investment. At the same time, public capital and private capital are complements, not substitutes for each other. Highway investment maybe justified in its complementary role even at lower calculated rates of return.
| 1951–1960 | 1961–1970 | 1971–1980 | 1981–1991 | Average 1951–1991 |
|
|---|---|---|---|---|---|
| Total Factor Productivity | 0.64% | 0.88% | 0.10% | 0.30% | 0.48% |
| Exogenous Demand | 0.40% | 0.63% | 0.18% | -0.06% | 0.29% |
| Price Changes | -0.01% | 0.01% | -0.04% | 0.01% | -0.01% |
| Highway Capital | 0.35% | 0.30% | 0.07% | 0.04% | 0.19% |
One qualification to the results offered by Nadiri is that changes in consumer welfare are left unmeasured. Investing in public highway infrastructure may benefit consumers in a number of ways. Personal vehicle operating costs are likely to be lower as a result of more direct-trip routing and better-quality road surfaces. Similarly, drivers have faster commute times to work and better travel times to nonwork places. Highway users may enjoy an expansion in opportunities for recreational activities, a greater choice in the location of their residences and possibly even wider employment opportunity as improved highways offer greater spatial accessibility. Because these consumer benefits may well be large in and of themselves, the rate of return estimates in this study are likely to underestimate the true effect of these improvements on the entire economy. If this underestimate is small, then the average private rate of return and the social rate of return on highway investment are converging. This convergence is a warning that the need for further physical expansion of the highway network is waning. If the uncounted benefits to consumers are large, then the social rate of return from the commercial sector, even in the most recent period, is well above the private rate of return, making further highway investment highly desirable.
The group proposed several areas for improving research. One suggestion involved incorporating more geographic detail into the estimates. Since the Nadiri study identified the effect of infrastructure investment on individual industries, without regard to location, all of the spillover effects of location-specific construction were captured in the model. Individual states or regions also need tools to help them gauge how such consequences are related to transportation investment. Butbecause states and regions compete for economic development, the gain to one area may come partly at the expense of a loss in some other area. Economic growth, as measured by Nadiri, is the national effect of this process, and in effect it nets out the gains and losses of individual areas. But the redistribution of economic activity within the nation may be an important consequence of transportation investment as seen by a state or region, and better tools for this purpose are also needed.
The ability to estimate rates of return for highway investment in different areas or places within the country, or even for different industries in different places, is desirable. Such estimates would be valuable for two reasons. First, public officials have an interest in how they would be affected by transportation changes. Second, a comparison of local benefits estimates with similarly computed national ones would shed light on the size of any spillover effects. While the spillover effects of highway infrastructure are generally thought to be positive, negative spillovers also may exist. Additions to one part of the highway system could spur growth in an area at the expense of growth somewhere else. The sizes and kinds of spillovers influence the way that any investment should be financed. If spillovers are very small, localities can capture all of the benefits of their investment in their area. In this situation, coordinating finance and construction with other localities is not important. If spillovers are large, then financial support needs to come from a larger geographic entity. Local construction may need to be subsidized nationally if the net effects of spillovers are positive, or taxed if the net effect of spillovers is negative. Extending the idea of negative spillovers internationally, highway infrastructure investment may affect the ability of the United States to attract economic activity that may have otherwise taken place outside the country.
Investment in transportation may improve international competitiveness by making transportation costs, logistics costs, or the costs of other inputs of production less expensive relative to these costs in other counties. The theoretical underpinning for studying U.S competitiveness needs further elaboration. An empirical exploration might best be grounded in a particular industry or region.
The changing nature and intensity of transportation infrastructure use also pose conceptual challenges for measuring the infrastructure's benefit. Highways that initially carry little traffic often fill over time as they are used more. Highways that face too much demand may see their use fall because traffic congestion diminishes their service. Trucking deregulation clearly altered the transportation environment and the use of infrastructure. Technological innovations, such as smart highways, could change highway use even more in the future. Simply measuring an annual flow of dollar investment fails to capture the effect of changes in infrastructure use patterns on the economy. Researchers need to develop new measures to better address the flow of transportation services provided by the highway capital stock. (This topic is discussed more in the next section of the report.) A participant suggested studying the benefits of research and development spending for highways. This aspect of investment has generally been left untapped and could prove to be a good topic for future research.
The participants made several other points during the discussion. Some noted that the benefits generated by highway capital investment have not come without additional costs to the environment. The costs of additional pollution, urban sprawl, loss of natural habitat and environment for plants and animals, and other environmental damage should be incorporated into future work. Other participants pointed out that policymakers and the public often desire more up-to-date results than the ones presented. In response, some of the data experts in the forum reminded the group that data of the sort used in Nadiri's study take at least three years to collect, process, and disseminate before the analytical work even has a chance to start, so change in this regard is unlikely without changes in the data collection process.
Nadiri's research has verified the value of the expansion of the Interstate Highway System over the past four decades. These refinements will help future work define the best kinds of investment for the future.
Charlie Han, Bureau of Transportation Statistics (left), and Arthur Jacoby, Federal Highway Administration
The forum highlighted several other questions to explore and new refinements for future research projects. Using Nadiri's cost elasticities for individual industries, Arthur Jacoby of the Federal Highway Administration and Charlie Han of the Bureau of Transportation Statistics analyzed the correlation of these cost elasticities with statistical measures to explore possible explanations for the difference in impact of infrastructure on industries. Barbara Fraumeni of the Department of Commerce, Bureau of Economic Analysis, described some of the findings of her project to improve the quality of infrastructure data by clarifying and refining the many assumptions that go into the construction of such data. In the ensuing discussion, participants identified several other potential improvements to the tools of economic analysis.
Ishaq Nadiri found that lower costs of production for industries correlated with increases in the highway capital stock. Searching for more information about the mechanisms through which highway capital investment reduces industry costs, Jacoby and Han tried to link the industry cost elasticities estimated in the Nadiri study with other industry characteristics. Starting with the premise that improvements in highway systems reduce the cost of transporting goods and the time it takes to deliver them, Jacoby and Han concluded that the responsiveness of industry costs to highway capital should be related to industry expenditures for transport.
Charlie Han, Bureau of Transportation Statistics, Arthur Jacoby, Federal Highway Administration, Dennis Lebo, Pennsylvania Department of Transportation, and James March, Federal Highway Administration (left to right)
Jacoby and Han first tested the hypothesis that industries that were extensive users of transportation spent large sums and were most likely to benefit from improvements to highway systems. However, using 1992 data on industry transportation use from the new Transportation Satellite Accounts developed jointly by the Bureau of Transportation Statistics and the Bureau of Economic Analysis, as well as Nadiri's 1981-1991 cost elasticity estimates, they found no significant relationship between the two variables. In fact, the correlation showed a small positive relationship between the two variables. Industries in which total production costs responded more to changes in the highway capital stock tended to have a lower cost of transportation relative to their total costs.
This result led them to refine their study. Highway capital investment can lower not only the cost of transport, but also the time of transport. Therefore, benefits might have accrued to industries not only according to their transportation use, but also according to the importance of the whole logistics process for the industry. The logistics function manages the movement of products over space and time, procuring the raw materials of production, keeping any necessary inventories, and delivering the final product to the customer. With this insight, their hypothesis evolved: The more important that logistics are to an industry, the more sizable the cost elasticity with respect to highway infrastructure would be.
A good piece of evidence supporting the importance of logistics costs to firms is the rise of just-in-time inventory practices. Just-in-time inventory practices demand speed and reliability in freight transportation. Using a suitably constructed proxy for industry logistics cost intensity, Jacoby and Han indeed found a significant negative relationship between their proxy and Nadiri's cost elasticity estimates. Sizable production cost reductions stemming from highway investment identified by Nadiri were associated with industries that had a greater reliance on the logistics function.
This work by Jacoby and Han points out that the evolution of the logistics function within businesses may offer important insights into the benefits gained from highway capital investment. It also serves as a guidepost for future research into the mechanisms by which transportation system improvements are transformed into productivity benefits. Insights into the behavior of firms may provide us with a better ability to forecast the kinds of transportation investment that will be most beneficial in the future. The relative importance of changes in transportation cost, speed, reliability, or access may make some investments more desirable than others.
As mentioned earlier in the discussion of Nadiri's research, one of the difficulties facing studies of highway investment effects is the question of how to capture the level of service provided by the stock of highway capital. Investment dollars must be transformed into transportation infrastructure. These physical networks need to be translated into a flow of transportation services.
Barbara Fraumeni has been leading an effort to produce a quality-adjusted public highway capital stock for the Federal Highway Administration. Her work has examined the different components that make up the transportation capital stock and how they fit together as a system to provide transportation services. In her presentation, Fraumeni recommended three major ways to improve highway capital stock measurements:
The difference between a productive capital stock and a wealth capital stock is perhaps best illuminated by the example of a light bulb. A light bulb gives off a certain amount of light while it works. That amount of light is constant as long as it works. However, the light bulb has a limited working lifetime and failure is all at once. Say that the light bulb is expected to work for 12 months. The productive stock value of the capital good light bulb is constant at any given point during the year. Until it stops working, it produces the same illumination. However, if a rational buyer were to offer to purchase the light bulb, the amount that person would be willing to pay for the light bulb would depend on how much expected working life the light bulb had left. After six months of use, this wealth stock measure of the light bulb's value would be about half of what it cost originally.
Wealth stocks, such as those compiled by the Bureau of Economic Analysis, take into account the depreciation of the wealth stock value. Therefore, they can be misleading when trying to capture the productive effects of highway capital stocks. Fraumeni calculates that her productive stock estimates for highways can be as much as 40 % higher than the Bureau of Economic Analysis wealth stock estimate.
Fraumeni's second recommendation for improvement addresses the difficulties of assuming the degree to which highway capital stock investment takes the form of new construction or rehabilitation. These different kinds of investment activities have differently weighted component expenditures for grading, pavement, rightof-way, and structures. Typically, expenditures for grading and acquiring right-of-way will constitute a larger proportion of new construction costs. These construction items will have much longer productive lifetimes than, for instance, the pavement surface. Historically, capital expenditures have been for new construction. In the last two decades, investment has taken the form of more reconstruction and rehabilitation, with a greater proportion of expenditures for pavement and lower share of expenditures on grading and right-of-way. Differentiating between new construction spending and capital maintenance spending is important for translating highway investment flows into accurate capital stock estimates.
Barbara Fraumeni, Bureau of Economic Analysis
Fraumeni's final recommendation for improving the quality of highway capital stock data is to develop a benchmark for capital stock estimates. Because some components of capital stock have very long lifetimes, using a perpetual inventory calculation without a benchmark (or very long time horizon) can lead to underestimates. The problem is most likely to affect state and local estimates, for which the time span covered by the data usually tends to be short.
All of these refinements lead to the development of a better measure of the potential productive services that the highway capital stock could provide.
Ideally, this improved capital stock measure would be converted into an actual flow of productive transportation services. Congestion and road usage influence the amount of productive services produced by the capital stock, as does the location of the facility. The value of the physical inventory of highways needs to take into account which parts are heavily used portions of the network and which are not, and this varies sharply from one part of the network to another. For example, some heavily used routes may actually be devalued because their level of service has deteriorated because of congestion. Other underused routes may have a low value because they are not generating a high flow of productive transportation services based on low utilization. Incorporating the degree of highway use into capital stock measures would clarify the value of the transportation services that our infrastructure system provides. That is, measuring the value of the light instead of the light bulb would help to dispel the shadows surrounding the future of highway investment.
Clyde Pyers of the Maryland Department of Transportation and Gregory Bischak from the Appalachian Regional Commission offered state and regional analyses of highway investment programs. Both studies found reasonable rates of return on recent investments, pointing to both short-term job creation and construction spending benefits and longer-term industry cost reductions and net travel savings. The discussion afterward reemphasized points made in the discussion of Ishaq Nadiri's study and offered a few additional insights.
Gregory Bishak, Appalachian Regional Council
The Maryland State Highway Administration conducted a study to quantify the contribution of highways to the Maryland economy. The analysis measured both the short-term effects of highways on employment and output and the long-term productivity gains that accrue to individual industries throughout the state.
The effects of Maryland highways on employment and output in the state were calculated using an input-output model of the regional economy over the 1991 to 1996 period. An estimate of the state's economic output without the benefit of the highway investment was compared with the level of output with the investment. Spending was broken down by industry to allow calculation of the resulting changes in employment. These results were linked to a model of the Maryland economy, prepared by RESI, a research institute of Towson University, with equations for employment and output at the industry level and linkages to components of population and income.
The long-term productivity effects were calculated using cost functions for nine industry groups from 1982 to 1996. Similar to the Nadiri study, annual industry costs were taken to be a function of wages, the price of capital goods, the stock of highway capital, and a proxy for technological progress (the functional form used was different). Savings in industry costs associated with increases in the highway capital stock were estimated econometrically.
The findings of the Maryland study were reasonably large and of a magnitude generally consistent with Nadiri's nation-wide estimates. In the short term, the study found that the roughly $1 billion annual highway system investment from 1991 to 1996 yielded 23,000 additional jobs and roughly a $2.7 billion increase in the output of Maryland goods and services. Most of the jobs created were estimated to be in the construction sector. From this greater business activity, the increase in state tax revenues alone recovered about one-fifth of the initial outlay. Over the long haul, the study estimated an annual rate of return on highway investment of 17 %. The rate of return was higher in the early 1980s, with a slight decline in the general trend moving toward 1996, for which the return was calculated as 17.2%. In contrast to the Nadiri study, the bulk of the cost savings accrued to manufacturing firms rather than services or retail trade (manufacturing production costs declined on average by more than 120 of the total 170 per highway dollar invested). In relation to economic growth, highway investment was found to be responsible for 10 % of productivity growth and 45 % of total growth in the state.
The Appalachian Regional Commission also examined the economic development effect of its highway system additions, known as the Appalachian Development Highway System (ADHS). The ADHS is a planned, 3,000-plus-mile network of highways extending through 13 states from Mississippi to New York. To date, nearly 80 % of the system has been completed. Wilbur Smith Associates performed the analysis for the commission, identifying 165 counties with 1,400 miles of highway as likely to be most affected by the system.
The Wilbur Smith study estimated changes in population, employment, wages, and value-added in the region as a result of the ADHS, using a regional economic model developed by Regional Economic Models, Inc. The study also examined the travel efficiency benefits of the highway system, such as reduced travel times, lower vehicle operating costs, and fewer accidents, as well as the broader economic development effects from increased roadside business, tourism, and an overall improvement in competitiveness in the region. Results were generated by comparing data from the historical record with an estimate of traffic volumes without the ADHS.
The study estimated that, from its beginning in 1965, the ADHS generated 16,000 additional jobs in the region, with a net increase of 42,000 jobs forecast by 2015. By sector, employment increased most for the construction industry in the first 10 years. However, by 1995, the employment gains were felt most heavily in retail trade and services, a pattern that was forecast to continue through 2015. In general, all sectors of employment showed some level of employment benefit. On a return basis, a dollar's investment in the ADHS generated $1.18 in efficiency benefits. For the more comprehensive economic development benefit measure, which includes tourism and competitiveness effects, the return climbs to $1.32 of benefits for the average dollar invested. Looking at individual corridors within the system, efficiency returns ranged from 5 to 10 %. The study also provided an estimate of $2.7 billion as the benefit from increased competitiveness for the entire 1965–2015 period.
The magnitude of the results from both studies is consistent with the results of Nadiri's national study, demonstrating the value of highway infrastructure investment at the state and regional level as well. The main concern about the interpretation of these regional study results was the extent of any negative spillovers not captured in the model. Did regional highway investments create benefits for one area at the expense of other areas? Several participants generally held that these improvements did not constitute a zero-sum game; the gains in productivity provide large enough increases to justify any moves by existing business and increase the competitiveness of the country as a whole. Work performed by Marlon Boarnet of the University of California at Irvine shows a degree of tradeoffs from highway investment between counties within California. However, more work needs to be done to better identify these interrelationships between localities and the extent to which local highway projects represent economic gains or merely result in shifts in the location of economic activity—transfers from one area to another.
A related issue for area or regional studies is the difficulty in assessing the influence of changes in the area not under study. While the transportation system within the area can be controlled for, changes outside the area may also affect the region's infrastructure. Transportation investments may complement or substitute for those in different areas. A more complete understanding of exactly how individual firms are affected by changes in highway infrastructure may help solve this outstanding issue.
As with the Nadiri study, these two studies most likely underestimate the actual rates of return for their investments. Including consumer benefits from cheaper, faster, and safer highway travel would certainly drive the estimates higher. Ideally, studies also would capture other economic effects such as benefits from improved accessibility to an area, the substitution of products produced locally for goods that would otherwise be imported, and increases in the economies of scale of production. While the Appalachian study did attempt to account for tourism benefits, its authors believe they may have undermeasured this effect. The degree to which all of these effects would alter the calculated benefits for transportation investment is currently unknown.
Several participants raised questions about the implications of the rates of return found in area studies. If such a study finds a rate of return a few percentage points below the typical private-sector rate of return, does it mean that the highway investment should not have been made? Ishaq Nadiri was quick to point out that highway investments are complementary to private investment.
Others questioned the kind of relationship that should be expected between estimates from the national study and regional studies. Should the regional studies be lower because they are not able to capture benefits from the infrastructure investment that accrue outside their area? Different areas are home to different kinds of industries. Gaps of a few percentage points between regions or between a regional and a national estimate are to be expected, since different industries benefit differently.
Trying to compare the two regional studies highlights the need for consistent data. If data could be collected in a consistent manner from state to state, states or regions could use a common framework for their analyses. The necessary research methodology could be developed and then shared, drastically reducing the time and cost of producing this sort of study for lcalities. It would also provide a better basis for comparing the benefits of investing in different areas.
After all the attention given to pouring pavement in these research studies, participants in the discussion also wanted to discuss ways in which highway investments other than construction can lead to greater transportation system capacity and efficiency. For example, the great changes in freight transportation that followed the deregulation of trucking demonstrate the importance of highway system management.
The use of technology may also be necessary to increase the capacity and safety of highways. Developing and implementing intelligent vehicle highway systems (smart highways) is a promising possibility for the future. Even investing more in the thickness of the pavement itself may generate additional benefits that outweigh the costs.
Nor should attention be limited solely to highways. Rail, sea, and air transportation continue to play vital roles in the nation's transportation system. These modes and their interconnections could become even more important in the future. Ultimately, more research and creative thinking will be needed to uncover new ways to make the best use of our existing transportation systems and highway capital stock.
After much of the forum focused on recent studies of the relationship between highway investment and economic growth and how to improve future work, a spirited session addressed the importance of communicating these results to policymakers and the public. Miles Friedman of the National Association of State Development Agencies and Terrence Mulcahy from the Wisconsin Department of Transportation drew the group into a wide-ranging discussion of the role that economic analyses play in the decision-making process. Comments highlighted (1) the difference between developing new programs and implementing additional local projects, (2) specific groups to target for communication and outreach, and (3) the roles of state and local transportation agencies.
Friedman discussed the evolution taking place in state economic development agencies and the effect that change has on the role of transportation investment in state development. State development agencies are reaching out to encompass wider geographical areas and are branching out to take on a wider range of development activities (to which the rise in regional development groups, both multistate and multicounty, attests). Development agencies no longer target just heavy industry. They also spur technology development, foster small business growth, promote tourism, and tackle international competitiveness issues. With this expansion in the scope of their responsibility, state development agencies increasingly view themselves as long-term strategic planners, coordinating and facilitating a number of different activities, including investment in the transportation system. Accordingly, transportation decisions should be folded into the planning mix. In this light, transportation investment is best cast as part of other longterm development efforts rather than as the enhancement of a set of facilities.
To illustrate his point, Friedman told the story of a town in Kansas that decided to revive its sagging downtown economy by advertising and promoting a historic fort located nearby. After hiring a professional marketing company to publicize the fort, the number of tourists visiting the fort soared. The downtown, however, remained empty. Upon further reflection, the town leaders realized that the transportation system needed to be part of their solution. They widened the 10-mile highway between downtown and the fort, offered a shuttle service between the two areas, and promoted special events downtown. Only after they considered all of the different facets of their problem, including the transportation environment, did they finally achieve the success they desired in revitalizing an important part of their community.
Miles Friedman, National Association of State Development Agencies
The larger political context in which transportation projects are presented becomes much more important as strategic planning becomes broader and more interconnected. New challenges will arise as transportation investments face the reality of competing with social programs. Environmental concerns and perceptions of favoritism to business can also be stumbling blocks for new projects. At the same time, the bigger picture surrounding transportation investment can provide new opportunities. Good transportation systems can connect people with jobs. New projects can encourage livable communities and spur business development.
Terrence Mulcahy of the Wisconsin Department of Transportation reinforced the need to view transportation in context. The chief concerns of communities in Wisconsin are the effects of transportation on local businesses, economic development, and the quality of life in the state. Roads, railroads, and airports are built to achieve important economic and social objectives, and transportation agencies are responsible for ensuring that these facilities serve the needs of citizens and businesses. In Mulcahy's view, one of the state highway department's primary roles is to provide a forum in which different communities can communicate their experiences to each other. Supporting Friedman's remarks, Mulcahy also urged transportation agencies to embrace the sometimes-disparate elements of commerce, tourism, and the environment.
How do research studies influence the public and policymakers? The reasonably good recent rates of return estimated in Nadiri's study provide some justification for implementing new transportation investment policies. Studies that show evidence of huge benefits from particular kinds of new programs, such as the Aschauer and Munnell results, may be readily apparent to the public or key beneficiaries. At the same time, when the benefits are not dramatic, information about returns alone may not be sufficiently compelling to the general public and legislators. With many competing interests in the public arena, gaining attention for transportation investment may be difficult.
Terrence Mulcahy, Wisconsin Department of Transportation (left), and Madeleine Bloom, Federal Highway Administration
The importance of success stories—anecdotes that vividly and succinctly convey the effect of a specific transportation project in a particular place—cannot be overstated. Linda Thelke of the Wisconsin Department of Transportation provided the group with a powerful example. The Harley-Davidson Company is an important employer in her home state of Wisconsin, providing high-quality, good paying jobs to local workers. In meetings with government officials, the president of Harley-Davidson has emphasized the importance to the company of good access to the transportation system. Good transportation allows the company to run with 35-minute delivery windows for moving products between their plants. With this kind of transportation system performance, the company can keep their costs lower and stay competitive without having to move their plants and jobs overseas to lower-wage countries. As Curt Wiley, the chair of the forum and the Commisioner of the Indiana Department of Transportation, put it, "[0]ne Harley-Davidson in a governor's ear probably takes the place of whatever study there was."
Case studies and individual stories often prove to be effective calls to action. Stories of successful transportation projects can provide the edge needed to generate wider appeal for larger programs. However, the group did not fully resolve whether studies and stories have different influences on national programs, state programs, and local projects. One tendency is to argue for programs using studies and to offer anecdotes to facilitate local projects, partially as a result of better data availability for larger-scale programs and funding constraints for smaller-scale projects.
Either studies or stories can provide a starting point for discussions of the pros and cons of projects or programs. In the current transportation environment of moderate benefits from investment, the question of how to effectively communicate results really becomes a question of how best to tie together these local anecdotes and program-wide estimates.
One solution is to use stories and studies in tandem. One participant likened the relationship between the use of national studies and anecdotes to using logic and emotion—a successful combination of both makes for the most compelling argument. National and state studies may provide a context for local anecdotes and at the same time reinforce them.
Another solution is to make the results more closely address the questions that policymakers and the public want answered. People want to know about transportation's relationship not only with growth, but also with quality of life, community development, and local commerce. Because of its role as a facilitator of activity, transportation analysis should be incorporated into the strategic planning of a state or region and tied into its economic development. Different regional development priorities will lead to different research needs—the effects on urban versus rural areas, manufacturing versus service-sector businesses, retail versus wholesale trade, and tourism versus job access. A closer match between the community's concerns and the scope of research may make studies more effective.
Not only the style and content, but also the vehicle of a transportation agency's message, must be considered. Groups such as the press and the business community are worth particular attention because of their influence. Miles Friedman found that thorough briefings with the press facilitated better coverage and a better ability to carry the messages to the public. The business community is also viewed as a vital group with which to communicate. The business community has a general level of influence with the public and policymakers. Its opinion tends to carry even more weight because it is significantly affected by highway investment. The forum participants were more divided on the usefulness of having a very persuasive and visible individual as a proponent. Better outreach to the press and the business community will help the effectiveness of whatever kinds of studies and anecdotes transportation agencies use.
The next wave of transportation research must move outward from the center created by Ishaq Nadiri's work. In one direction, new research needs to delve further into the effects of transportation investment on businesses, examining firms' responses to changes in transportation systems. Research also must better understand the geographic variation in the return to transportation investment, grappling to measure the spillover effects and interrelationships between investment in different locations. Such detailed information will help determine the kinds of infrastructure investments that hold the most promise for the future.
In a second direction, though, research needs to move beyond the realm of infrastructure investment to consider the full range of transportation policies that could be used to improve the nation's productivity. Policies that can bring new technologies to bear on the problem of congested highways or other new approaches to the management and operation of our infrastructure may ultimately prove more attractive than infrastructure expansion.
Transportation researchers should not forget to reach out to the public and seek to embrace a whole range of relatively new concerns: competitiveness, job access, tourism, and the environment. Policymakers and the public are the ultimate consumers of transportation research, and the most effective research efforts will be those that discuss the kinds of issues in which these consumers are most interested. Effectively communicating new transportation research means being sophisticated in using the right combination of study results and compelling stories to reach policymakers and the public.
