Using a natural experiment from Germany, we show that temporary place-based subsidies generate persistent effects on economic density. We identify employment and capital formation as main channels for higher income per square kilometer. As the spatial regression discontinuity design allows us to control for all spatially-continuous determinants of agglomeration (e.g. homemarket effects, knowledge spillovers), we attribute an important role to capital formation in explaining persistent spatial patterns of economic activity. However, estimates of externalities at the treatment border point to small net effects of the policy. We find strong evidence that pre-treatment land owners have benefitted substantially from the program and that transfers have shown larger effects in high-density places. Finally, accounting for regional subsidies raises the causal effect of market access for economic development as identified in Redding and Sturm (2008) by about 45 percent.
place-based policy regional policy economic geography persistence regression discontinuity locational advantage land value capitalization