CANCELLED: Encouraging a Go-Getter

About this event

This paper considers a decision maker who receives a reward if he takes a costly action above an unknown threshold. A designer controls the information available to the decision maker about the threshold. The designer’s objective is increasing in the decision maker’s action. Our objective is to characterise the information she would optimally disclose. We show that two properties of optimal disclosure policies are robust to the specification of the designer’s preferences. First, the decision maker achieves the reward exactly when he would have optimally chosen to do so if he had complete information about the threshold. Second, the decision maker remains indifferent between receiving the designer’s message or acting on the basis of his prior. An important determinant of the structure of optimal policies is the designer’s attitude towards risk. We present explicit solutions in the two polar cases in which the designer is either risk-loving or risk-averse in the decision maker’s action. A contribution of the paper is the characterisation of optimal persuasion schemes in an environment in which actions and states live in a continuum. This seminar is not hybrid.