Bayesian implementation with verifiable information

• 成果类型: Article
• 署名作者: Peralta, Esteban
• 署名单位: University of Michigan System; University of Michigan
• 刊物名称: GAMES AND ECONOMIC BEHAVIOR
• ISSN/ISSBN: 0899-8256
• DOI: 10.1016/j.geb.2019.04.003
• 发表日期: 2019
• 页码: 65-72
• 关键词: Mechanism design; Bayesian implementation; verifiable information

摘要

This note studies the class of allocations that are fully Bayesian implementable in the presence of verifiable information. I identify a condition, termed evidence distinguishability, that together with incentive compatibility is both necessary and sufficient for full Bayesian implementation in direct mechanisms that elicit verifiable information. I show, however, that every evidence distinguishable allocation that is incentive compatible when verifiable information is not elicited is Bayesian monotonic at every verifiable deception; namely, at every verifiable joint report of information at which the direct mechanism delivers an undesirable outcome. To the extent that requiring Bayesian monotonicity to hold at every verifiable deception is a demanding condition, only indirect mechanisms can exploit the presence of verifiable information to fully implement incentive compatible allocations. (C) 2019 Elsevier Inc. All rights reserved.