Incumbents elected in proportional representation system struggle in majoritarian elections. Quasi-experimental evidence from Polish municipality elections

It is well-known that incumbents often enjoy an advantage in elections, even at the level of local legislatures (Dahlgaard 2016). A vast majority of the literature finds an incumbent advantage in plurality systems, while the evidence for the advantage in proportional representation systems is mixed (Jankowski and Müller 2021). In this paper, I ask if incumbents maintain their advantage when the electoral system changes. So far, this question remains unanswered due to the relative stability of electoral systems and limited opportunities for causal inference. Specifically, I test whether incumbents elected under proportional rule maintain their personal advantage if they need to compete for re-election in a majoritarian system (and vice versa). I exploit a series of electoral system changes in local elections in Poland as natural experiments. In five municipality elections in Poland in the years 2002-2018, the arbitrary threshold of 20,000 inhabitants divided municipalities into different electoral rules. This allows to apply the sharp Regression Discontinuity Design (RDD). I match incumbents in over 2,400 municipalities with candidates running in the next election and compare the electoral results of incumbents running in municipalities 'just below' and 'just above' the threshold. I find that incumbents elected via a proportional representation system struggle to be re-elected after the system is changed to single member districts. The system change does not substantially affect the share of incumbents who run in the next election, nor the share of votes received by incumbents. However, the incumbents do not sort to avoid competition with other incumbents. In effect, more than one incumbent runs in the same district twice as often where previously the proportional system had been applied than in municipalities where the plurality rule had been in place. By force, less of such incumbents can be re-elected when they compete against each other. I do not find similar effect in terms of turnover after the system is changed from majoritarian to proportional with gender quota. Also, I do not find any difference between plurality and proportional system, and between majoritarian and proportional system with gender quota. It is an important contribution, since settings allowing for a direct and causal comparison of electoral systems are scarce. Finally, contrary to the expectations based on the literature on gender and age composition of legislative bodies, the system changes do not affect the demographic composition of municipality councils. In sum, the findings of this paper point to the limited impact of electoral systems on legislative turnover and composition of local legislatures. In turn, they encourage further studies on the effects of electoral system changes. Dahlgaard, Jens Olav. 2016. ‘You Just Made It: Individual Incumbency Advantage under Proportional Representation’. Electoral Studies 44: 319–28. Jankowski, Michael, and Stefan Müller. 2021. ‘The Incumbency Advantage in Second-Order PR Elections: Evidence from the Irish Context, 1942–2019’. Electoral Studies 71: 102331.