Nice Matchings: Equalizing Externalities as a Solution Concept for Two-Sided Matching

Abstract

Our main goal is to develop a solution concept for two-sided matching settings without money called nice matchings. Nice matchings equalize externalities, with respect to some scaling of agents’ preferences. We make a case for nice matchings over the GS algorithm, both with respect to maximizing welfare, and the fairness of the resulting allocation. We also apply the APEX algorithm to two-sided matchings. Our primary purpose is to use the APEX algorithm as an algorithm for finding nice matchings. It may also be a subsequent analytical tool for proving the existence of nice matchings, and for analyzing strategic incentives of agents when a mechanism outputs nice matchings.