Biased Experts, Costly Lies, and Binary Decisions

Abstract

Decision makers lacking crucial specialist know-how often consult with better informed but biased experts. In our model the decision maker’s choice problem is binary and her preferred option depends on the state of the world unknown to her. The expert observes the state and sends a report to the decision maker. His bias is such that he prefers the same decision for all states. Lying about the state leads to a cost that increases in the size of the lie. As a function of the size of the expert’s bias and the decision maker’s prior about the underlying state, three kinds of equilibrium behavior occur. In each case equilibrium consists of separating and pooling segments, and the decision maker takes the expert’s preferred decision for some states for which she would not take this decision had she observed the state herself. The model has a variety of applications and extends to situations in which the decision maker may be naive and take the report by its face value, and to situations with multiple experts and uncertainty about the size of the expert’s bias.