Addressing the large inefficiencies generated by the Deferred Acceptance (DA) mechanism requires priority violations, but which ones are justifiable? The leading approach is to ask individuals if they consent to waive their priority ex-ante. We develop an alternative question-free solution, in which a priority violation is justifiable whenever the affected student either (i) directly benefits from the improvement, or (ii) is unimprovable under any assignment that Pareto-dominates DA. This endogenous justifiability criterion permits improvements unattainable by the leading consent-based mechanism under any consent structure. We provide a ``just below cutoffs'' mechanism that always finds a strongly justifiable matching whenever DA's outcome is inefficient, and build on it to construct a polynomial-time algorithm that expands justifiable improvements iteratively, converging to a DA improvement that cannot be Pareto-improved by any justifiable matching without strictly expanding the beneficiary set. Finally, we prove theoretically that both the ex-ante consent and the endogenous justifiability frameworks have important limitations in reaching Pareto-efficient outcomes, and use simulations to quantify how binding these constraints are in practice.

翻译：推导因延迟接受（DA）机制产生的巨大低效问题需要优先权违规，但哪些违规是可辩护的？主流方法是询问个体是否同意在事前放弃其优先权。我们提出了一种替代性的无问题解决方案，其中当受影响的学生要么（i）直接从改进中受益，要么（ii）在任何帕累托优于DA的分配下均无法改进时，该优先权违规便是可辩护的。这种内生可辩护性标准允许在任意同意结构下，实现由主流基于同意的机制无法达成的改进。我们提出了一种“略低于阈值”机制，该机制总是在DA结果低效时找到强可辩护的匹配，并在此基础上构建了一个多项式时间算法，迭代扩展可辩护的改进，最终收敛到一个DA改进方案，该方案无法通过任何严格扩大受益者集合的可辩护匹配实现帕累托改进。最后，我们从理论上证明了事前同意框架与内生可辩护性框架在达到帕累托有效结果方面均存在重要局限性，并通过模拟量化了这些约束在实际中的限制程度。