Projection Methods for Economies with Heterogeneous Agents

In this paper we develop a general methodology for solving models with heterogeneous agents by projection methods. Our approach is solely based on the functional forms of agents' optimal policy rules and on a functional condition on the endogenous stationary distribution. Solving simultaneously the optimal policy rules and the distribution, this paper provides a new methodology for computing equilibria in which the distribution of wealth and income is a part of a social planner's optimization problem. We do not impose any additional restrictions or assumptions on the equilibrium allocations. Compared to other computational methods, it does not suffer from the curse of dimensionality and provides an efficient tool for computing models of economies with a continuum of heterogeneous agents with several endogenous and exogenous state variables. We illustrate the algorithm on a standard model with uninsurable idiosyncratic risk from labor income. The approximate solution is highly accurate, especially for the distribution function. This method can be used to compute equilibria in economies with heterogeneous agents in which the distribution of wealth and income is a part of a government's optimization problem