Economists' interest in the question of public infrastructure productivity has grown steadily since the 1980s. This paper reviews the literature on this topic with a particular focus on transportation's economic impact. Cumulative evidence reveals that, first, estimates of the elasticity of output with respect to public capital have declined over time and are currently indistinguishable from zero. Second, highways have local negative spillover effects that arise from economic activities being drawn to infrastructurerich locations at the expense of adjacent areas. Third, transportation infrastructure is subject to congestion, which reduces the productivity of such infrastructure even when stocks remain constant. Finally, highways consistently enhance the productivity of manufacturing firms even when they do not do so for firms in other sectors. Is transportation economically productive? The answer to this question is no small matter. Over the past half century, the United States has built the world's most extensive transportation system, the bulk of which involved laying in over 160,000 miles of Interstate and other National Highway System roads (1). Justifications for creating this expanse of public infrastructure have evolved over the years. A common thread, however, is the notion that transportation infrastructure makes the economy more productive. Over the past 20 years, a growing number of researchers have studied the validity of this justification through a much broader question: is total government capital, usually defined as transportation infrastructure, water facilities, and sewer systems, economically productive? Although this broader question may cloud the waters somewhat, lessons learned from this body of research have highly focused implications for the transportation professions for a number of reasons. First, transportation comprises the single largest component of most public capital stock estimates used in studies. The Bureau of Economic Analysis (2) estimated that in 2003, transportation infrastructure accounted for 35% of the estimated $5.5 trillion in total state and federal (nonmilitary) fixed assets. Other large categories of public capital included education structures (19%); water, power, and sewer facilities (15%); and commercial, office, and residential structures (12%). Fraumeni (3) reviews these and other estimates, revealing the many ways that public capital can be measured and categorized. Second, there is a strong theoretical and intuitive connection between transportation and productivity. An enormous literature on

Economists’ interest in the question of public infrastructure productivity has grown steadily since the 1980s. This paper reviews the literature on this topic with a particular focus on transportation's economic impact. Cumulative evidence reveals that, first, estimates of the elasticity of output with respect to public capital have declined over time and are currently in-distinguishable from zero. Second, highways have local negative spillover effects that arise from economic activities being drawn to infrastructure-rich locations at the expense of adjacent areas. Third, transportation infrastructure is subject to congestion, which reduces the productivity of such infrastructure even when stocks remain constant. Finally, highways consistently enhance the productivity of manufacturing firms even when they do not do so for firms in other sectors.

Decentralization, the intrametropolitan movement of people and activity away from dense centers toward less developed suburbs, occurs when dispersive forces are greater than agglomerative forces for any metropolitan area. Two mathematical tools, the regional dispersion index and the Hoover index, are used to measure the movement of population, employment, and retail activity in the Hartford, Connecticut, metropolitan region from 1900 to 2000. The approach used differs from other research in that it does not assume a parametric density function, and it allows decentralization for a multicentered region to be measured. Between 1900 and 1930, population concentrated toward Hartford and other regional centers. The trend reversed around 1940, and for the remainder of the century population decentralized away from both Hartford and regional centers. Economic activity bore the same overall pattern with a lag. The results presented here are consistent with existing theories of decentralization and previous empirical work.