Adaptively Rational Retrospective Voting

Abstract

Since the seminal work of Key (1966), Kramer (1971), and Nordhaus (1975), retrospective voting has been a major component of voting theory. However, although these views are alive empirically (Lewis-Beck and Stegmaier, 2000; Franzese, 2002; Hibbs, 2006), most theorizing assumes rational citizens. We suspect that Key had a less heroic view of voter cognition, and we formalize his verbal theory accordingly. Our model is based on two axioms: if an incumbent performed well (above voter A’s aspiration) then A becomes more likely to vote for the incumbent; A is less likely to do so if the incumbent performed poorly (below A’s aspiration). We then prove that such citizens, though lacking ideologies, endogenously develop partisan voting tendencies. This result is robust against perceptual errors (citizens evaluating an incumbent’s performance incorrectly). We also show that the best-informed voters, who perceive performance most accurately, are the most partisan.