Ordinally Bayesian Incentive Compatible Stable Matchings

We study incentive issues related to two-sided one-to-one stable matching problem after weakening the notion of strategy-proofness to Ordinal Bayesian Incentive Compatibility (OBIC). Under OBIC, truthtelling is required to maximize the expected utility of every agent, expected utility being computed with respect to the agent’s prior beliefs and under the assumption that everybody else is also telling the truth. We show that when preferences are unrestricted there exists no matching procedure that is both stable and OBIC. Next preferences are restricted to the case where remaining single is the worst alternative for every agent. We show that in this case, if agents have uniform priors then the stable matchings generated by “deferred acceptance algorithms” are OBIC. However, for generic priors there are no matching procedures that are both stable and OBIC even with restricted preferences.