Essays on information economics

My dissertation consists of three essays on information economics, each of which addresses a problem of incentive compatibility with the presence of asymmetric information. Chapter 1 proposes an incentive compatible self-fulfilling mechanism as a strategic foundation for rational expectations equilibrium in economies with asymmetric information. Chapter 2 studies a classical model of bank runs with asymmetric information and introduces a full-information mechanism that could uniquely implement the first-best allocation in an incentive compatible manner. Chapter 3 establishes positive results regarding the existence of ex ante incentive compatible core in exchange economies with asymmetric information and intransitive preferences. A more detailed summary of results is presented in order below. In Chapter 1, I extend the exact equivalence result between the allocations realized by self-fulfilling mechanisms and rational expectations equilibrium allocations in Forges and Minelli (1997) to a large finite-agent replica economy where different replicas of the same agent are allowed to receive different private information. To deal with the problem of incentive compatibility, i.e., an agent's incentive to use private information to strategic advantage, I apply the machinery of informational size developed in McLean and Postlewaite (2002) and establish two approximation results. The first result states that the allocation realized by any incentive compatible self-fulfilling mechanism is an approximate rational expectations equilibrium allocation. Conversely, the second result states that, given any rational expectations equilibrium satisfying a "uniform continuity" condition on the equilibrium price, one can construct an approximate incentive compatible self-fulfilling mechanism whose equilibrium allocation coincides with the rational expectations equilibrium allocation for all non-monetary commodities. Therefore, this chapter provides a strategic foundation for rational expectations equilibrium using an incentive compatible self-fulfilling mechanism. In Chapter 2, I study how banks and other financial intermediaries can use information disclosure to prevent self-fulfilling bank runs. I begin with a finite-agent version of Diamond and Dybvig (1983) with correlated liquidity shocks and sequential service. I propose a full-information mechanism that allows the intermediary to inform each investor about the withdrawal decisions of previous investors. Adding information disclosure to the efficient banking contract creates a withdrawal game with a feature of sequential signaling in the sense that each investor can signal his type through withdrawal decision. In particular, a patient depositor can signal his true type by choosing not to withdraw from the intermediary. This action of not withdrawal can "persuade" remaining investors, if patient, to do the same. Anticipating such responses from remaining investors, investors earlier in the sequence, if patient, also want to choose not to withdraw. As a consequence, I show that, under certain conditions on the underlying economy, all bank run equilibria fail to survive in a natural refinement called forward induction equilibrium (Cho, 1987) that generalizes the "intuitive" criterion in the classical signaling game. In other words, no bank run occurs as a reasonable equilibrium outcome in the full-information mechanism. Therefore, disclosing withdrawal information can promote financial stability. In Chapter 3, which is joint work with Richard P. McLean, we study two problems associated with exchange economies in which agents are asymmetrically informed about an underlying state and have intransitive preferences admitting skew-symmetric representations. We first address the existence of ex post incentive compatible approximately efficient allocations and then study the ex ante incentive compatible core for such exchange economies. We provide conditionsunder which the ex ante incentive compatible core and ex ante incentive compatible epsilon core are nonempty when agents are informationally small. Furthermore, we study replica economies and identify conditions under which the ex ante incentive compatible epsilon core is nonempty when the replica economy is large enough.Ph.D.Includes bibliographical reference