Mechanism Design and Communication Costs

Abstract

This paper focuses on implementation issues in environments where it may be costly for the players to send certain messages. We develop an approach allowing to characterize the set of implementable outcomes in such environments, and then apply it to derive optimal mechanisms. The key elements of our approach are the absence of any restrictions on the communication structure in a mechanism and the ability of the principal to screen the agents not only on the basis of their preferences over the outcomes, but also on the basis of their communication abilities. A number of interesting implications for the monopoly regulation, signaling and screening is derived. In particular, we show that a monopoly may not want to exclude low-valuation consumers if some consumers in the population are not able to misrepresent their valuations, and why the employers may prefer to screen applicants via multiple rounds of interviews rather than via menus of contracts. Our findings also provide a justification for privacy laws.