Working Paper: NBER ID: w30111
Authors: Leonid Kogan; Indrajit Mitra
Abstract: We propose a general simulation-based procedure for estimating quality of approximate policies in heterogeneous-agent equilibrium models, which allows to verify that such approximate solutions describe a near-rational equilibrium. Our procedure endows agents with superior knowledge of the future path of the economy, while imposing a suitable penalty for such foresight. The relaxed problem is more tractable than the original, and results in an upper bound on agents’ welfare. Our method applies to various solution algorithms. We illustrate our approach in two applications: the incomplete-markets model of Krusell and Smith (1998) and the heterogeneous firm model of Khan and Thomas (2008).
Keywords: Heterogeneous-Agent Models; Equilibrium Models; Welfare Loss; Approximate Policies
JEL Codes: C02; C18; C63; C68; E00; E37; G1
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| modifying the original problem to include perfect foresight (D84) | derive a more tractable problem (C60) |
| derive a more tractable problem (C60) | higher value function in the modified scenario (D46) |
| higher value function in the modified scenario (D46) | establish an upper bound on welfare (D69) |
| the gap between the upper bound and the welfare under the approximate solution (D69) | indicates the degree of suboptimality (C61) |
| a smaller gap (Y60) | signals that the approximate equilibrium is near-rational (C62) |
| their method reveals potentially large welfare losses (D69) | approximation quality can significantly impact welfare outcomes (C60) |