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 election. The electoral advantage is larger for fulltime mayors, increasing in municipality size, independent of the specific partisanship of the mayor and constant between 1945 and 2010. Moreover, it increases with local spending hikes and it is independent of municipal debt. I also illustrate the causal dynamic effects of the incumbent status on distant future elections and therefore evaluate the global properties of the LATE estimate. Finally, I show that the total effect is due to an effect on the probability that the party participates in the next election (about 40% of the total effect) and an effect on the vote share (about 60%).
mayor elections regression discontinuity design party incumbency advantage