Empirics on the causal effects of rent control in Germany

This paper analyzes empirically the effects of a second generation rent control. We investigate the consequences of an uncommon policy intervention in the German housing market in 2015. We rely on a difference-and-differences setup, augmented with elements of a discontinuity-in-time design, to identify the causal impact of rent control on controlled and uncontrolled prices, land values, and the short-run supply of (rental) housing. We exploit an exception for newly built units and compare these units to young regulated units in order to measure the relative effect of the regulation on these groups. We are able to decompose this total effect into a negative effect on regulated and a positive effect on unregulated units, by exploiting spatio-temporal variation in treatment and the _ne-grained temporal variation of the data. We document positive effects on rents and prices of unregulated units and negative effects on regulated units. Intra-market variation is not available for identifying effects on land values and supply. We thus use propensity score weighting and trimming for selecting comparable treatment and control units. We find a robust positive impact of the regulation on land values that is qualitatively and quantitatively in line with the results for rents. We then document that the rent control regime led to more demolitions of small residential units (single- and two-family homes demolished with the purpose of making room for a new residential building), but we do not find an effect for larger buildings. We take this as a first sign of positive long-run supply effects. We also provide evidence that landlords more often chose to sell rather than let vacant rental units and that maintenance effort decreased under rent control. Overall, these results _t the predictions of a standard comparative-static representation of a second-generation rent control, which sheds a more favorable light on rent control.