Outcome-Independent Payoffs in Strategic Voting

We consider a game-theoretic model of voting in which players have an outcomeindependent component to their preferences. This outcome-independent component—even if arbitrarily small—can dramatically affect the set of Nash equilibria in voting games because it determines how voters behave when they are not pivotal. Given incomplete information, some weak restrictions on voter preferences, and a sufficiently large number of voters, there is a unique Bayesian Nash equilibrium in which every player votes according to the outcome-independent component of his preferences. Our model helps explain (1) why people vote when participation is optional and voting is costly, and (2) why public and secret voting may lead to different outcomes. ∗We thank Professor Stephen Salant for his devoted attention, careful guidance, and invaluable suggestions. We also thank Professor Lones Smith, Professor Scott Page, and Professor Tilman Börgers for their helpful comments. †University of Michigan, [email protected] ‡University of Michigan, [email protected]