Behaviorally Optimal Auction Design: Examples and Observations
用“层级k”思维模型替代均衡假设，研究两个对称竞拍者的最优密封拍卖设计，发现一级价格拍卖中层级k行为能改变最优保留价并超越迈尔森收入上限，一种利用非均衡信念的奇特拍卖可大幅超越该上限。
Abstract: This paper begins to explore behavioral mechanism design, replacing equilibrium by a model based on “level-k ” thinking, which has strong support in experiments. In representative examples, we consider optimal sealed-bid auctions with two symmetric bidders who have independent private values, assuming that the designer knows the distribution of level-k bidders. We show that in a first-price auction, level-k bidding changes the optimal reserve price and often yields expected revenue that exceeds Myerson’s (1981) bound; and that an exotic auction that exploits bidders ’ non-equilibrium beliefs can far exceed the revenue bound. We close with some general observations about level-k auction design.