Party Incumbency in the German Bundestag: Evidence from Regression Discontinuity Design

Abstract

This study investigates the party incumbency advantage for district candidates in German federal elections, through a Regression Discontinuity Design (RDD). For this purpose, it replicates and applies a theoretical framework to analyse election results for district representatives in federal elections in Germany from 1976 to 2021, where closely contested elections induce local randomization. This allows drawing inference on the causal effect of incumbency. This study finds that the centre-right Union has an incumbency advantage of roughly 2 percentage points in vote share and roughly 2 percentage points in probability to win over her largest rival, the centre-left SPD in the elections of district representatives. The effects for both major parties are robust and induce an overall incumbency advantage for district candidates in the magnitude of roughly 15 percentage points in probability to win. The analysis of individual candidate data shows that randomization is given and that no perfect sorting for full-time politicians of either major party is feasible.