A TEST OF ALTERNATIVE MODELS OF INDUSTRIAL DECENTRALIZATION IN THE UNITED STATES (INDUSTRIAL DEVELOPMENT, CAPITAL DYNAMICS, REGIONAL ECONOMICS, PRODUCT CYCLE, INCOME INEQUALITIES).

摘要

There is much speculation and debate concerning the manner in which regional industrial development inequalities are eliminated in an industrialized economy. Two classes of paradigms have been proposed to account for this process--models of hinterland-spread and models of hierarchical diffusion. The objective of this research is to empirically determine which paradigm performed best in modeling the lessening of regional industrial development inequalities in the post-World War II United States.;Results of choroplethic maps and multiple regression analyses of employment density distributions of individual years (1947, 1958, 1967, 1977) and decadal changes support a hinterland-spread of industrial development. Moreover, the regression residuals were examined for positive spatial autocorrelation which, in general, was found to be mitigated in the regression models. Simple and partial correlation analyses were performed to identify salient determinants of industrial development and which reaffirmed the suggestion that the dominant process of industrial change was hinterland-spread.;The theoretical underpinnings of both paradigms are discussed and the weaknesses and strengths of either are critically evaluated. Moreover, an alternative theoretical basis for either paradigm is proposed in which the dynamics of capital investment in the manufacturing sector is argued to result in the spatial division of production and nonproduction facilities. Production activities are shown to decentralize from the industrial heartland and into the hinterland more extensively and earlier than nonproduction activities. In this research, hinterland-spread models and hierarchical diffusion models are examined in light of changing spatial distributions of production and nonproduction manufacturing employment per 100 square miles by state. Independent variables proposed in this research to account for these changing spatial distributions include adjacent state employment, log normal distance from New York, an urban population concentration index, state rural population percentages, and indices of regional industrial diversity and labor force productivity.