Contracting Experts with Unknown Cost Structures

Abstract

We investigate the problem of a principal contracting an expert to provide a probability forecast for a binary event. Experts can research this event at a cost unknown to the principal. We present a truthful and efficient mechanism for the principal's problem of contracting an expert. This results in the principal contracting the best expert to do work equivalent to having the second best expert in-house. We discuss several extensions to this mechanism. The contracts in [9] are used to generalize our mechanism to non-binary events. We consider how the mechanism is affected when the principal and experts have a maximum acceptable risk and cannot afford to exceed a certain budget. Finally, we discuss the result of the experts changing their belief before the mechanism - either due to a signal they received or due to costly research that they carry out.