Ronny Razin (London School of Economics and Political Science)

“Persuasion with Correlation Neglect”

Abstract
We consider an information design problem in which a sender tries to persuade a receiver that has “correlation neglect”, i.e., fails to understand that signals might be correlated. We show that the sender can change the expected posterior of the receiver in any direction. When the number of signals the sender can send is large, she can approach her first best utility. We characterize for which environments full correlation is the optimal solution; in these cases we can use a modified problem and standard concavification techniques. We show that full correlation is optimal in the familiar case of binary utilities but also more generally when utilities are super-modular and when the number of signals is large. However, in some environments full correlation is not optimal and in those cases the optimal solution involves negative correlation.