Sequential Learning

Abstract

We develop a model in which two players sequentially and publicly examine a project. In our model the player who moves first can fabricate evidence to influence the second mover, which creates a moral hazard problem. We find that early strategic uncertainty can mitigate this problem. In particular, for intermediate prior beliefs about the project’s quality, the Pareto-dominant equilibrium is in mixed strategies and consists of an early stage in which evidence may be fabricated and a later stage in which evidence is always authentic. Our findings shed light on the dynamics of R&D, quality assurance, and drug approval. (JEL D82, D83, L15)