Spatial convergence of regions revisited : a spatial maximum likelihood systems approach

This paper suggests that one should account for the endogeneity of important explanatory variables and the persistence of technology shocks when analyzing spatial convergence among regions.Specifically, it is argued that a systems approach is called for that includes the average growth rate and the initial income level as the endogenous variables. For 212 European regions the estimation results reveal a substantial correlation between the disturbances of the equation explaining initial income per capita and that of its subsequent average growth rate. Moreover, the estimated speed of convergence is found substantially higher in a systems framework. This holds true for both spatial conditional and unconditional convergence. -- spatial beta-convergence ; spatial solow model ; spatial systems maximum likelihood estimation ; European regions