Theoretical economic geography has received significant new impetus from so-called 'new economic geography' theory, and this provides the basis of the spatial econometric model developed in this paper. The model, which is seen as part of an attempt to bridge the gap that presently exists between this new body of theory and reality, is fitted using 3SLS to data on manufacturing productivity growth data for 178 regions of the EU covering the periods 1975-81, 1981-89, 1989-95. The resulting estimates point to the continued existence over time of increasing returns, and thus provides support for one of the main tenets of new economic geography theory. The results also highlight urbanization, peripherality, the initial level of technology and across-region spillovers as determinants of regional productivity growth variations, operating via the rate of technical progress and labour efficiency variations. There is no evidence that these variables are diminishing in significance systematically over time, leading to the conclusion that spatial polarization in productivity will persist. The paper points to the need for further testing of fundamental assumptions of the underlying theory, and for developments in the theory to accommodate interdependence between model parameters and the socio-economic context within which the process is operating.
- economic geography
- spatial econometrics
- EU manufacturing productivity