中文摘要
预留系统通过创建带有优先受益人排序的类别, 将疫苗、学校学位或政府职位等稀缺资源配置给特定重要群体。以往的方法假定存在基线优先次序, 且要么采用硬性预留, 要么采用软性预留。我们提出的阈值预留模型支持相互独立的优先排序、硬性与软性预留的混合以及类别间的重叠, 从而既能容纳受益人指定又能容纳资格约束, 为政策制定者在平权行动体系中提供了更大的灵活性。迭代最大-最大分配机制 (Iterative Max-in-Max Assignment Mechanism, IMMAM) 在一个宽泛的适用域内实现了所有合意的性质: 它尊重类别内的优先次序, 最大化被利用的资源数量, 然后按字典序最大化受益人分配。IMMAM 具有路径独立性, 这保证了在多机构同时配置资源的情形下的稳定性。我们利用路径独立性证明了合意的比较静态结果, 并提高了机制的计算效率。我们概述了本框架在新冠 (COVID-19) 疫苗分配中的应用。
Abstract
Reserve systems allocate scarce resources –such as vaccines, school seats, or government positions– to essential groups by creating categories with prioritized beneficiaries. Previous approaches assumed baseline priorities and featured either hard or soft reserves. The threshold reserve model we propose supports independent priority orderings, a mixture of hard and soft reserves, and overlapping categories, thereby accommodating both beneficiary designations and eligibility constraints, and providing policymakers with greater flexibility in affirmative action systems. The Iterative Max-in-Max Assignment Mechanism (IMMAM) achieves all desirable properties in a broad domain: it respects priorities within categories, maximizes the number of resources utilized, and then lexicographically maximizes beneficiary assignments. IMMAM is path independent, ensuring stability in settings with multiple institutions allocating resources. We leverage path independence to prove desirable comparative statics and to improve the computational efficiency of our mechanism. We outline applications of our framework in the context of COVID-19 vaccine allocation.