By Norbert Schanne, RĂ¼diger Wapler and Antje Weyh, 2008


We forecast unemployment for the 176 German labour-market districts on a monthly basis. Because of their small size, strong spatial interdependencies exist between these regional units. To account for these as well as for the heterogeneity in the regional development over time, we apply different versions of an univariate spatial GVAR model. When comparing the forecast precision with univariate time-series methods, we find that the spatial model does indeed perform better or at least as well. Hence, the GVAR model provides an alternative or complementary approach to commonly used methods in regional forecasting which do not consider regional interdependencies.