Rational Voter Theory: An Agent-Based Model

Contemporary democratic theory typically assumes electors use some sort of limited rationality to determine their voting behavior. However, political scientists debate whether electors use a rational calculus to determine whether or not to vote. In particular, some argue that rational electors would choose to abstain under most conditions, and thus that rational decision theory cannot account for the observed high turnout levels in large elections. One early game-theoretic model of voter turnout yielded both high- and low-turnout equilibria, but only the low-turnout equilibria was robust to conditions of imperfect information. Explaining the observed high turnouts in large electorates remains a significant challenge for rational-decision theoretic approaches to voter behavior. However, the rational-instrumental account nicely predicts a suite of empirically-observed marginal effects. We used agent-based modeling to investigate whether the rational-instrumental account is truly inconsistent with high turnout values. We found that the rational-instrumental account does not predict widespread abstention. While there are no high-turnout equilibria, the model exhibits a cyclic dynamic which can sustain significant levels of turnout for sustained periods. The model also predicts many of the empirically-observed marginal effects associated with the rational-instrumental voter hypothesis.