kshardpig.bib

@article{ks-jme2016,
title = "Solving dynamic public insurance games with endogenous agent distributions: Theory and computational approximation",
journal = "Journal of Mathematical Economics",
volume = "64",
pages = "77 - 98",
year = "2016",
issn = "0304-4068",
doi = "https://doi.org/10.1016/j.jmateco.2016.03.004",
url = "http://www.sciencedirect.com/science/article/pii/S0304406816300040",
author = "Timothy Kam and Ronald Stauber",
keywords = "Heterogeneity, Dynamic games, Bilinear programs, Approximation",
abstract = "We make two contributions in this paper. First, we extend the characterization of equilibrium payoff correspondences in history-dependent dynamic policy games to a class with endogenously heterogeneous private agents. In contrast to policy games involving representative agents, this extension has interesting consequences as it implies additional nonlinearity (i.e., bilinearity) between the game states (distributions) and continuation/promised values in the policymaker’s objective and incentive constraints. The second contribution of our paper is in addressing the computational challenges arising from this payoff-relevant nonlinearity. Exploiting the game’s structure, we propose implementable approximate bilinear programming formulations to construct estimates of the equilibrium value correspondence. Our approximation method respects the property of upper hemicontinuity in the target correspondence. We provide small-scale computational examples as proofs of concept."
}