The Value of Preference Information in Agency Relationships.

Standard models of agency theory assume that the principal has considerable information about the preferences of the agent. Once this assumption is relaxed, the question arises whether the principal should try to obtain additional information about the agent's preferences. In this paper we introduce the concept of a Value of Preference Information (VPI), which describes the benefits to the principal from obtaining additional information about the agent's preferences. We show analytically that the VPI is non-negative and that the VPI will not decrease when the principal's information structure is refined. Computational methods are used to study factors influencing the VPI. The results of these experiments show a strong relationship between the entropy of the principal's information structure and the VPI. It is thus possible to evaluate an information structure independently of the decision problem delegated to the agent.