Political Accountability: A Stochastic Control Approach

Abstract

In a democracy elections are the primary mechanism to discipline politicians. Indeed, policy-makers care for being in office and this affects their policy choices; for instance, they refrain from rent-extraction in order to be re-elected and benefit from future rents. Hence, elections provide implicit incentives that allow voters to align politicians’ preferences with their own ones. This role is crucial because constitutions do not offer explicit incentive schemes (cf. Persson et al., 1997), that is, forms of compensation based on some performance measure as may happen in a relationship between employers and employees.

Early political agency models, such as Barro (1973) and Ferejohn (1986), describe the disciplining effect of elections assuming that voters are backward-looking (i.e., re-election is a reward for incumbent’s past performance) and that the incumbent and the challenger are identical. This implies that even a small change in voters’ preferences makes them leave their announced voting rule.