This paper investigates the relation between changes in self-employment and changes in unemployment at the regional level in Spain in the period 1979-2001. We estimate a vec-tor autoregression model as proposed by Audretsch, Carree, van Stel and Thurik (2005) using a data base for Spanish regions. By estimating the model we are able to empirically distinguish between two directions of causality. On the one hand increases in self-employment may contribute to lower unemployment rates (the u91entrepreneurialu92 effect). On the other hand, higher unemployment rates may push individuals into self-employment, thereby contributing to higher self-employment rates (the u91refugeeu92 effect). In our analysis of these two effects we distinguish between higher and lower income regions within Spain. We find empirical support for the u91entrepreneurialu92 effect to exist, both in higher income and in lower income regions. As regards the u91refugeeu92 effect, the evidence is mixed. We find empirical support for this effect for higher income regions. Remarkably, we do not find evidence for a u91refugeeu92 effect in lower income regions of Spain, even though unemployment rates are on average higher in these regions. We argue that this may be partly related to a lack of incentives for unemployed individuals in these regions to find paid employment.
展开▼