The regional employment growth in Germany is characterized by huge disparities. Whereas institutional factors might explain the disparities of employment growth between nations, they can only account for a minor fraction of the regional employment growth. Instead the sectoral structure of employment is often seen as a major reason for regional disparities. An important attribute of the research conducted so far is that it concentrates on estimating shift-share-regression-models when controlling for the influence of the sectoral structure on employment growth. However, these models do not account for spatial interdependencies and treat regions as autarkies, though on a regional level such effects are very likely to occur. Against this background, the present article focuses on analyzing the role played by spatial interdependencies between regions in explaining their employment growth. By using spatial econometric methods this article emphasizes that regional employment growth is characterized by spatial autocorrelation, pointing to spatial interdependencies between the regions. This holds true also for mayor factors of regional employment such as wages and qualification. In this article three different models of spatial interdependencies are being compared: the spatial lag, spatial error and cross regressive model. While the spatial lag model controls for the influence of the value of the endogenous variable in neighboring regions (i.e. the spatial lag of the endogenous variable) on the endogenous variable in the observed region, the spatial error model estimates the influence of the spatial lag of the error term. Finally, the cross regressive model includes the spatial lag of the exogenous variables. These spatial interdependencies are integrated into the framework of the shift-share-regression-model to identify the relevant spatial interdependencies and to measure their influence on the regional employment growth in Germany. Preliminary results indicate that the spatial autocorrelation of the regional employment growth is to a large extend caused by the spatial lags of the exogenous variables. The spatial lag of the endogenous variable and the error term become insignificant, when the spatial lags of the exogenous variables are being accounted for. These effects can be interpreted as spatial spillover effects.
展开▼
