X. SUMMARY, CONCLUSIONS AND DIRECTIONS FOR FUTURE RESEARCH

X. 1. Summary and Conclusions

The main goal of this report and the 1996 report is to analyze and measure the contribution of highway capital to private sector productivity growth. The approach developed here explicitly incorporates demand and supply forces, including the contribution of highway capital, which may affect productivity performance. We estimate the model using disaggregated data composed of 35-sectors in the U.S. economy for the period 1950-1991. The data include measures of gross output, material inputs (inclusive of energy), and private capital and labor. We also estimate demand and supply (cost) functions for each industry. We identify the determinants of productivity growth for each industry, including the contribution of highway capital, and we measure specifically the marginal benefit of highway capital to each industry.

To generate aggregate measures for the whole economy, we use a weighted sum of individual industry elasticities to obtain the aggregate elasticity measures for the whole economy. Using these estimates, we decompose total TFP growth into its various components and calculate the net social rate of return to highway capital. The rates of return on highway capital are compared to that of private sector capital.

In this report, we modify our 1996 analysis in several ways. First, we extend the sample period to cover the period 1950-91. The data set for the 35 sectors which comprises the entire US economy were revised significantly and in principle lead to a new set of estimates and results. Second, we explicitly introduce other infrastructure capital as another unpaid input in the production structure. We explicitly take into account the interaction of this type of capital with private sector inputs and highway capital at the industry and macroeconomy levels. Again, this innovation may lead to results different from those in our 1996 report.²³ Finally, we use a different functional form for the cost function, i.e. a translog instead of the Normalized Symmetric MacFadden functional form. We do this because it is now easier to empirically evaluate a more complicated and expanded model.

The specific quantitative results of this report are briefly summarized as follows:

• There is evidence of a mild degree of increasing returns to scale in most industries and at the national level. The marginal products of labor, capital and intermediate inputs vary across industries. The output elasticity of materials is in general the largest, followed by that of labor and capital inputs. More importantly, at both the industry level and the national level, the elasticity of private capital is larger than that of total highway capital by a factor of two times for the entire period and by a factor of about four times for the period 1981-91. This result is in sharp contrast to those reported in Aschauer (1990), Canning, Fay and Perotti (1993), and Fernald (1992). All of those studies imply that an additional dollar of public investment is substantially more productive than a corresponding dollar of private investment.

• Total highway capital contributes significantly to economic growth and productivity at the industry and national economy levels. This contribution varies across industries and over time. The magnitude of the elasticity of output with respect to total highway capital at the aggregate level is about 0.08, which is much smaller than comparable estimates reported in the literature and somewhat larger than our previous results.

• An increase in highway capital has an initial productivity effect: it reduces total cost for a given level of output for all industries and at the aggregate economy level. This productivity effect induces output expansion in all industries which in turn increases costs by requiring increases in input demands. When output level is allowed to vary, the productivity gains of highway capital offset the cost increases required by the output expansion.

• Total highway capital has a significant effect on employment, private capital formation and demand for materials inputs in all industries. The magnitude of these effects varies among the three inputs in a given industry and across industries. Given a level of output, an increase in highway capital leads to a reduction in demand for labor and materials and an increase in demand for private capital in all industries. This result differs from those stated in the 1996 report. There, we found that the effect on the demand for inputs differed for manufacturing and non-manufacturing industries. When the output level is allowed to increase in response to the costs saving effects of an increase in highway capital investment, the direction of the effects on the demand for labor, capital, and materials remain the same as noted earlier. However, the magnitude of negative elasticities of demand for labor and materials are substantially reduced in each industry. The elasticity of private capital to highway capital increases significantly due to the induced output expansion effect.

• The marginal benefits of highway capital are positive in all but three fairly small industries. The magnitudes of these benefits, which can be interpreted as a measure of producers' "willingness to pay", varies considerably across industries and over time. We observe that the average sum of marginal benefits across all industries is about 0.294. This is somewhat higher than the figure reported in our 1996 report. However, the individual industry marginal benefits are much smaller than previously reported. In the current study, there are very few negative marginal benefits at the industry level, which accounts for the larger sum of marginal benefits for the entire economy than our previously reported results.

• We calculate the net social return to total highway capital by using the industry marginal benefit calculations and the cost of highway capital, while taking into account the distortionary effects of taxation to finance highway capital. The results indicate that the net social rate of return on total highway capital is high during the 1950s and the 1960s, but declines considerably in the 1980s. In the 1980s, the rates of return on total highway capital and private sector capital seem to have converged.

• The contribution of highway capital to TFP growth is positive in all industries. In our previous results, highway capital increased productivity growth mainly in manufacturing industries, but not in non-manufacturing industries. Our recent results show a more pervasive influence on TFP growth. However, the magnitudes of the contribution of highway capital to productivity growth varies across industries. At the aggregate level, highway capital's contribution to TFP growth is about 0.25. Compared with estimates previously reported in the literature, this figure is relatively small. However, this contribution is somewhat larger than the estimate provided in our 1996 report. The decomposition of TFP growth confirms previous results that the main contributor to productivity growth both at the industry and aggregate levels is exogenous demand (representing the effect of aggregate income and population growth). Relative prices and technical change also contributed to the growth of TFP as well, but their contributions are generally smaller and vary across industries.

X. 2. Directions for Future Research

There are a number of important issues that require further research:

1. In this study, we estimate the demand and cost function for each industry separately. A step forward would be to estimate the two functions jointly and explicitly account for cross equation restrictions.

2. Finding measures to account for quality changes in highway capital stock and intensity of use of the capital stock are the other considerations for future research. The quality adjustments require a careful analysis of the highway capital stock series taking into account the vintage and depreciation of different components of the highway capital. As far as the utilization of the highway capital is concerned, the challenge is to find ways to convert the stock of highway capital to service measures. This requires adjustments for utilization of the highway capital taking account of factors such as congestion, intensity of use of highway capital by industry, and level of business activity. These adjustments could have important effects on the results reported. Finding proper adjustments for the degree of utilization of the highway network is challenging work. There are conceptual and measurement issues that need to be addressed. However, it is likely that the averages and time profile of the industry marginal benefits as well as the rate of return on highway capital will be substantially affected by adjustments for the degree of utilization of the highway capital stock.

3. Another issue is to examine more closely the depreciation rate estimates that are used to generate the total highway capital. If the depreciation rates are not well measured then the results on marginal benefit, net social rate of return and productivity contribution of highway capital reported here will be affected. Analytical models are available to estimate the depreciation rate from available investment data.²⁴ In addition, availability of data on maintenance expenditures and other relevant data may allow estimating a more precise measure of the depreciation rate and thus better measures of highway capital stock.

4. A careful analysis is required to examine further the size and pattern of the implied taxes. What is important to note is that the benefits of highway capital vary across industries. The needs of different industries for highway services diverge over time and the degree of benefits of new highway capital expansion may differ considerably across industries. That is, public highway capital creates important distributional effects across industries. These effects need to be further examined to ascertain the sign and magnitudes of the industry marginal benefits. There are a number of possibilities to explore in the future. One possibility is to classify industries into more detailed categories to capture the inherent diversity within each sector. Another possibility is to refine highway capital data in such a way to incorporate the adjustments for quality and degree of congestion.

5. Finally, in this study we have concentrated on the benefits of highway capital to the production side of the economy. The welfare benefits of highway capital services to the consumers have not been addressed in this report. The magnitude of these benefits are likely to be significant. Efforts to account for the total effect of highway capital requires modeling the consumption sector of the economy and integrating it with the production sector in a general equilibrium model. Such an attempt, though extremely important, at present remain outside the scope of our research.

FHWA Home | HPTS Home | Feedback United States Department of Transportation - Federal Highway Administration