A Retrospective Voting Model With Heterogeneous Politicians

This paper extends the performance‐based voting models of Barro (1973), Ferejohn (1986), and Austen‐Smith and Banks (1990) by including heterogeneous politicians. The introduction of heterogeneous politicians has three major implications. First, the act of voting must now simultaneously address both “moral hazard” and “adverse selection” problems. Accordingly, this paper demonstrates that voters will usually have to trade off the twin goals of (i) securing the optimal performance from a given type of politician, and (ii) insuring that only the best types of politicians are selected to office. Second, both equilibrium reelection rates and the tenure‐performance relationship are shown to depend on the value of holding office. For example, when the value of holding political office is large relative to other employment opportunities. incumbents will experience very high reelection rates: and longer‐serving politicians will perform poorer, on average, than shorter‐serving politicians. Finally, we demonstrate the surprising result that increased heterogeneity of politicians can make voters worse off.