Types of Incumbency and Their Electoral Effects in Germany’s Mixed-member System

• In Event: Electoral Institutions

Abstract

The literature on incumbency effects finds vote premiums for incumbent candidates in various institutional contexts. It has been shown that incumbency effects vary over electoral systems, with larger effects in single member districts than in closed list systems. In mixed-member electoral systems, however, incumbency can mean different things: having won the local district race in the first-past-the-post (FPTP) tier; or having entered parliament via the party list in the PR tier. Notably, in compensatory systems such as Germany, those two categories are not disjunct.
Should we expect large incumbency effects in the plurality/FPTP tier, and smaller effects in the proportional (PR) tier within the electoral system, analogous to the dynamics in the pure systems? Or does incumbency produce distinct dynamics in a mixed-system? This question has not been explicitly addressed so far. Previous research in this area has solely focused on the effect of district incumbents, but disregarded the existence of a second type of incumbents, who won their seat via the PR tier. The questions I address in this paper are not only for the electoral benefit of each type of incumbency. I’m also investigating whether the effects are only relevant for the respective tier, or whether they spill over onto the other tier.
Furthermore, the pattern of competition in local district races involving different types of incumbents and non-incumbents offers fruitful grounds to analyze distinct aspects of incumbency effects. For example, the peculiarities of the German system offer the opportunity to separate the effect of office holder benefits from political experience and success on the local level by analyzing PR list incumbents who have not competed in the past district race, but possess benefits from holding office. Seizing the opportunity to study those aspects separately also yield insights into the more general mechanisms that underlie incumbency advantages in other systems.
This research is based on a large panel data set covering all district candidates at German Bundestag elections. I apply panel fixed-effects regression models to analyze the effects of different types and aspects of incumbency. Inference is thus based on variation in vote shares within the units of analysis, that is candidates, rather than having to rely on variation between candidates that might lead to biased estimates due to unobserved heterogeneity and self-selection.

Author

• Anna Sophie Kurella, Universität Mannheim