Market potential and spatial autocorrelation in the european regions

Resumen

The concept of market potential developed by Harris (1954) has several shortcomings but it has proved its practical utility even when compared with more sophisticated versions derived from the wage equation of the New Economic Geography. We stress the ability of this variable to summarize regional markets accessibility and their spatial distribution. We specially highlight a feature of the market potential/market access concept in its foreign or external form (without computing own region�s market potential). Namely, External Market Potential is a non-standardized inverse distance spatial lag of the regional internal markets. Considered together, internal and external market potential, a variable of Harris�s Market Potential is able to capture the core-periphery pattern of economic activity and per capita Gross Value Added in the European regions. This long-distance spatial dependence is different from the short-distance spatial dependence emphasized by Spatial Econometrics, with its standardized spatial weights for a reduced number of neighbours. But capturing these spatial patterns in a regression with a variable of Market Potential comes at a cost of spatial autocorrelation. We show how the Spatial Econometrics procedures allow a correction of the spatial autocorrelation keeping the interpretation of Market Potential in terms of the accessibility to the markets and as an indicator of peripherality. So, both types of spatial dependences should be considered together.