Parties, Majorities, Incumbencies : Four essays in political economics

Sammanfattning: This thesis consists of four self-contained essays on political economics. The first paper studies the impact of having one party in government versus a coalition of parties, where the exogenous variation in the type of government stems from close election outcomes. It uses a new algorithm to detect these close elections in multi-party systems to answer this question. Based on data from more than 2,000 municipalities in the German state of Bavaria, it finds that single-party governments spend more, not less as is often concluded in the theoretical and empirical literature. The second paper uses the same method of detecting close elections to extract exogenous variation, but looks at the political power of parties and its effect on tax policies. It finds significant effects of party power that are mostly in line with expectations. The third paper looks at the transition of voters between parties in three consecutive elections for the state parliament in Bavaria, and infers parties’ ideological positions from these transition flows. After estimating the transition matrices with a method based on maximum entropy, it uses these matrices to compute a distance matrix and uses multi-dimensional scaling to place parties in a policy space. The resulting positions of parties are plausible, consistent across both transition periods, and comparable to those estimated with other methods. The final paper studies the heterogeneity in the advantage of incumbent district candidates in German federal and state parliament elections. In particular, it looks at the party in government, and how that affects the incumbency advantage of district candidates. It finds that an incumbency effect only exists (for both major parties) if the center-left SPD is in government, a heterogeneity that is robust across different specifications and jurisdictions.