Expansion of markets and the geographic distribution of economic activities: The trends in United States regional manufacturing structure, 1860-1987

The dissertation investigates the empirical dimension of externalities and increasing returns by examining the long-run experience of the American economy. If external economies are important, then there should be increasing regional concentration or specialization of industrial production over time as transportation costs fall. In order to test this hypothesis, indexes of regional specialization and of localization are used to establish the record of U.S. regional manufacturing structure from 1860 to 1987. As the U.S. regions integrated economically between 1860 and 1890, regional manufacturing structure became more specialized at both the 2 and 3 digit levels of industry aggregation. It continued to specialize through the turn of the century and plateaued around the early 1930s. Since then it has despecialized continuously and substantially through 1987 to the point where U.S. manufacturing is less regionally specialized today than it was in 1860. Moreover, if agricultural and service sectors are included, the convergence in employment structure is even more dramatic. While external economies or other sources of increasing returns might have been important in accounting for regional concentration during the late 19th and the early 20th centuries, their significance may have diminished considerably over time. The evidence suggests that there are limits to such economies and that basing long-run policies on their realization may prove inadvisable.