This paper provides empirical evidence on the party incumbency advantage in mayoral elections in Germany. Using a regression discontinuity design on a data set of about 25,000 elections, I estimate a causal incumbency effect of 38-40 percentage points in the probability of winning the next mayor...