Abstract: When policymakers implement mechanisms in real-world institutions, they often prefer strategy-proof mechanisms over manipulable ones. The Boston School Committee, for instance, replaced the Boston mechanism with the students-proposing deferred acceptance algorithm in July 2005 to eliminate students’ incentives to misrepresent their preferences over schools. However, strategy-proof mechanisms are not always immune to manipulations by potential coalitions, even if these coalitions are small and easy to coordinate. This danger is avoided under group strategy-proof mechanisms.

In this paper, we study group strategy-proofness of stable matching mechanisms in two-sided matching markets when both sides of the market have strategic agents. In the context of a one-to-one matching market, we show that incorporating strategy-proofness for any stable matching mechanism not only removes the incentive for individual agents to manipulate, but also eliminates the incentive for any group of agents to manipulate (the group may include agents from both sides of the market), therefore implying group strategy-proofness. We establish this result assuming that the domain is sufficiently varied. We also explore the possibility of extending our findings to many-to-one matching markets.