A principal-agent model of altruistic redistribution, with some implications for fiscal federalism

Abstract

Economists may have taken an unduly limited view of the role that subnational governments (states, counties, cities) ought to play in the redistribution of income. When information is incomplete, subnational redistribution may reduce agency costs. A principal-agent model is proposed, in which taxpayers want to redistribute only to the deserving poor, where deservingness is defined in terms of an unobservable, effort, and an imperfectly observed stochastic shock. In the resulting equilibria, the taxpayer’s optimal benefit strategies are shown to be inconsistent with uniformity of benefit in a large country.