【主题】Strategy-Proof Allocations with Punishment
【时间】2021年10月26日 星期二 14:30-16:00
【摘要】A planner chooses an allocation of a divisible resource and charges to agents based on their reported type. We discover and describe the set of dominant strategy incentive compatible (strategy-proof) mechanisms when the planner has the ability to observe the true type of agents ex-post and punish those agents who misreported their type. This class of mechanisms depends on the punishment function available for the planner to use and expands previous characterizations of incentive compatible mechanisms when punishment was no available. For any punishment function, an optimal mechanism for the planner is characterized as the convex combination of two mechanisms resembling the first-price and second-price mechanisms. When the planner has ability to select the punishment function, the minimal punishment necessary to achieve incentive compatibility and first-best efficiency is provided.