Voter turnout as a participation game: An experimental investigation

The paradox of voter turnout has intrigued economists for many years. In 1957, Downs formulated the problem by writing about the theoretical irrationality of an individual vote in the light of massive turnout in elections. There is a large literature of theoretical and empirical papers addressing the paradox. In the last decade game theoretica models have been developed which show that turning out to vote may be rational under in an instrumental sense under specific circumstances. In a game theoretic setting, the decision to vote or to abstain is seen as a case of strategic interaction between individuals. The general idea is that if everyone knows that it is rational for everyone to abstain, one might expect zero turnout, in which case it may be rational to vote. This paper presents an experimental analysis of participation games which simulate a general election. It concentrates on electoral systems with proportional representation. The results describe the game theoretic predictions with emphasis on the comparative statistics between such systems and ‘winner-takes-all’ elections. It finds that participation is larger in ‘winner-takes-all’ and that the size of the groups in the game does not appear to affect the participation rates.