In this theoretical examination, Ryo Shirakawa revisits the standard assumption in market design that allocation mechanisms must be strategy-proof, thereby eliminating any incentives for participants to misstate their preferences. The author contends that this requirement may be excessively demanding, given that agents in real-world settings often lack the sophistication necessary to execute complex manipulations. Within the school choice framework, the study advances a weaker robustness criterion, arguing that mechanisms need only be resistant to a limited class of “simple” misreports. While no fully strategy-proof mechanism is shown to Pareto dominate the deferred acceptance (DA) mechanism, the analysis demonstrates that such an improvement is achievable under this relaxed condition through Kesten’s (2010) efficiency-adjusted deferred acceptance (EADA) mechanism. Extending the refinement to also exclude simple forms of coordinated misreporting yields characterizations of Kesten’s mechanism, offering practical contributions to the theory and application of market design. This is especially relevant in the context of school choice, where Abdulkadiroglu, Pathak, and Roth (2009) estimated, through simulations, that in New York City’s school choice system (which uses the DA) thousands of students could be made better off through a reallocation that would not make any student worse off.
- Journal Article: https://www.aeaweb.org/article...
- Working Paper: https://papers.ssrn.com/sol3/p...