In this paper incumbency effects in Federal Elections and Mayoral Elections in Germany are estimated using a quasi-experimental design which allows for causal inference under a set of rather mild assumptions. Relying on nonparametric and parametric estimation procedures and exploiting a recently...