Sequential Decision Problems and Asymmetric Information

Abstract

The study of sequential decision processes has been important for theories of statistical decisionmaking, dynamic programming, optimal stopping, search, and investment decisions. These literatures have emphasized the benefits of incrementalism and the resulting value of information. However, the costs arising from asymmetries in access to newly-arriving information have been uniformly ignored. This paper presents a preliminary inquiry into the costs of information in sequential economic decisionmaking problems. An investment project with several stages is modeled, in which the project manager and the investors have differential access to new information about project value. An optimal contract is derived specifiying financing terms and project management rules. Two general conclusions are developed: first, asymmetric information in sequential decision problems can be quite costly; second, modeling investment decisions as sequential problems with asymmetric information can lead to optimal contracts that are simple, robust and realistic. Applications to a broad range of economic questions are suggested.