Cities worldwide have regulated peer-to-peer short-term rental platforms claiming that those platforms remove apartments from the long-term housing market, causing an in- crease in rents. Establishing and quantifying such a causal link is, however, challenging. We investigate two policy changes in Berlin to first assess how effective they were in regulating Airbnb, the largest online peer-to-peer short-term rental platform. We document that the policy changes reduced the number of entire homes listed on Airbnb substantially, by eight to ten listings per square kilometer. In particular the introduction of limitations on the misuse of regular rental apartments as short-term accommodations, also strongly reduced the average number of days per year that Airbnb listings are available for booking. In a second step, we then use this policy-induced change in Airbnb supply to assess the impact of Airbnb on rents in the city. Our results suggest that each nearby apartment on Airbnb increases average monthly rents by at least seven cents per square meter. This effect is larger for Airbnb listings that are available for rent for a larger part of the year. Further analyses suggest some effect heterogeneity across the city. In particular, areas with lower Airbnb density tend to be affected more by additional Airbnb listings.