Robust Market Design

点击翻译按钮获取中文摘要

The mechanism design literature of the last thirty years has been very successful. A wide range of institutional design questions, ranging from simple trading rules to political constitutions, have been analyzed and elegant optimal solutions have been obtained. Market design has become more important in many economic arenas, both because of new insights from theory and developments in information technology. Yet there has been an unfortunate disconnect between the general theory and the applications/empirical work: mechanisms that work in theory or are optimal in some class of mechanisms often turn out to be too complicated to be used in practise. Practitioners have then often been led to argue in favor ad hoc restriction which lead to simpler but apparently sub-optimal mechanisms. The PIs will conduct research on mechanism design to answer this question. They model robustness considerations by relaxing common knowledge assumptions in such a way that stronger solution concepts and simpler mechanism are endogenously created as robust optimal solution. In the first part of the project they consider design problems in which the objective of the designer is to maximize expected revenues and encompasses optimal monopoly pricing and optimal auctions. They introduce robustness consideration by allowing the principal to be uncertain about agents' beliefs and higher order beliefs about other agents' types. Subsequently, they will pursue a local version of the robustness analysis by requiring that the true distribution over valuations is within a range of a model distribution. In the second part they consider full rather than partial implementation. As private information may enable the agents to coordinate behavior in many different ways, the designer is concerned that there exists equilibrium behavior by the agents which does not realize his objective. The robust design then seeks to prevent equilibria, in the form of collusive behavior among the agents, which fail to realize the objective of the designer. This research conceptualizes the notion of robustness for a wide class of mechanism design problems. It provides analytical instruments to evaluate how sensitive design rules are to strategic uncertainty. If successful, the techniques will provide solutions to specific design questions such as auctions and other allocation rules. The theory of mechanism design is used with increasing frequency to solve real-life trading problems. It is then of foremost interest to understand how robust the theoretical solutions are to small variations in the environment. The goal of this project is to deepen our understanding of the performance of mechanisms such as bilateral trading and auctions rules. If successful, the techniques and solutions suggested will lend themselves to improve the performance of mechanisms and decrease their sensitivity to the specific features of the environment.