Dresden Discussion Paper Series in Economics No. 08/01
We analyze the spatial interaction among regions in North America and in Western Europe. We use a gravity model extended by a spatial correlation structures where data allows to evaluate the level of impact and the length of the spatial tail. This allows us to address to effects external to the gravity model: level of impact of neighboring regions on the region and size of the cluster of regions. We find that the methodology employed improves the statistical quality of results and their economic interpretation. We conclude that national borders matter and that North America is more polarized in the sense of connected clusters whereas regions Europe externalities are more evenly distributed. We argue that this relates to different types of institutional arrangements with effects on the spatial division of labor.