查看文章

This paper introduces the concept of``self-justified equilibria" as a tractable alternative to rational expectations equilibria in stochastic general equilibrium models with heterogeneous agents. A self-justified equilibrium is a temporary equilibrium where, in each period, agents trade in assets and commodities to maximize the sum of their current utility and expected future utilities that are forecasted based on current endogenous variables and the current exogenous shock. The agents' forecasting functions are assumed to lie in a compact, finite-dimensional set of functions, and the forecasts constitute the best L^ 2 approximation to a selection of the temporary equilibrium correspondence. We provide sufficient conditions for the existence of self-justified equilibria, and we develop a simulation-based computational method to approximate them numerically, even if the underlying state space is very high-dimensional and the problem suffers from the curse of dimensionality. The concept of a self-justified equilibrium allows us to systematically weigh off the complexity of forecasting functions with the accuracy of forecasts. To illustrate the advantages of our concept, we focus on a convenient, theoretically sound, and numerically tractable case where we couple polynomials of a fixed degree to active subspaces to model the agents' forecasts. We solve annually-calibrated overlapping generations models with aggregate shocks and show that a one-dimensional active subspace often exists. Consequently, even in nonlinear, stochastic 60-dimensional models, approximating one-dimensional forecasting functions in the appropriate subspace can lead to highly …