Advocacy and Dynamic Delegation

An advocate for a special interest provides information to an uninformed planner for her to consider in making a sequence of important decisions. Although the advocate may have valuable information for the planner, it is is also known that the advocate is biased and will distort his advice if necessary to influence the planner's decision. Each time she repeats the problem, however, the planner learns about the accuracy of the advocate's recommendation, mitigating some of the advocate's incentive to act in a self-serving manner. We propose a theory of dynamic delegation to explain why planners do sometimes rely on information provided by advocates in making decisions. The interaction takes place in two phases, a communication phase, followed by a sequence of decisions and learning by the planner. We ∑first establish that the capability to delegate dynamically is a necessary condition for influential communication in this setting, and characterize the optimal dynamic delegation policy. Next, we show that a planner may prefer to consult an an advocate rather than a neutral adviser. Finally, we demonstrate how an advocate gains influence with a decision maker by making his preferences for actions unpredictable. Our results have implications for a variety of real world interactions including regulation, organization, and whistleblowing.