An Empirically-Grounded Theory of Voting Based on Prosocial Motives

We propose a model of costly voting in a canonical spatial setting that reconciles key empirical features of electoral turnout. We integrate citizens with prosocial voting motives who realize higher payoffs from voting when (i) they expect more fellow supporters to vote, (ii) such votes collectively matter more for winning probabilities, and (iii) they care more about which candidate wins. We establish existence of Nash equilibria satisfying an additional condition that ensures equilibria correspond to those in a finite setting with many voters. We derive how electoral primitives—e.g., voter ideology, voting costs, elite polarization, expected closeness—affect outcomes such as turnout, turnout rates, and winning margins. We relate the predictions to empirical findings both from existing studies and our original data work. With all three prosocial motive components, our model reconciles the empirical regularities qualitatively and quantitatively; while existing models fail on multiple dimensions.