Political Accountability: A Stochastic Control Approach

We study a political agency model with career concerns in a two-period continuous time stochastic framework where politicians differ on their competence in managing public sector. Competence is unobserved and learnt over time in a Bayesian fashion through the observation of the economy wealth. Citizens are expected utility maximizers and vote at the end of the first period for the incumbent or a challenger randomly chosen among the population depending on the amount of the public wealth delivered by the former. The incumbent politician chooses the extent of public intervention in the economy and a rent-seeking behaviour which reduces the economy wealth and is not observed by voters (moral hazard). According to traditional literature, we find support for the idea that elections lead politicians to be opportunistically more aligned with voters' preferences.