In analyzing the disparities in regional employment growth in Germany, in the recent empirical literature the so called shift-share-regression models are frequently applied. However, these models usually neglect spatial interdependencies, even though such interdependencies are likely to occur on a regional level. Therefore, this paper focuses on the importance of spatial dependencies using spatial autocorrelation in order to analyze regional employment growth. Spatial dependency in the form of spatial lag, spatial error, and cross regressive models are compared. The results indicate that the exogenous variables’ spatial lag sufficiently explains the spatial autocorrelation of regional employment growth.
Zierahn, Ulrich (2012), The importance of spatial autocorrelation for regional employment growth in Germany, Jahrbuch für Regionalwissenschaft 32(1), 19-43. Download