The Unintended Costs and Unfulfilled Promises of Concurrent Elections: A Natural Experiment on Turnout and Invalid Voting

Abstract

Holding two second-order elections simultaneously is expected to increase electoral participation. We exploit a natural experiment in which one group of Czech precincts was “as if” randomly assigned to holding subnational elections concurrently with senatorial ones. Using a unique data set containing variables on more than thirteen thousand precincts in five elections between 2000 and 2016, we detect a modest effect of concurrency only in the first election but no or inconsistent effect in the four subsequent contests. Furthermore, we report a strong effect of concurrency on invalid voting. We check for robustness using difference-in-differences design and matching techniques. Incongruent with existing theories, concurrency does not deliver on its promises and may come at a substantial cost to political representation. The surprising null effect on turnout is attributable to analyzing the effect of concurrency of the less salient on more salient elections.