Website CV Paper
Fields
Theory, Behavioral & Experimental
Job market paper
Persuasion under Subjective Causal Models
Can a sender exploit the way a receiver interprets information? I study persuasion when the receiver interprets the sender's signal through a subjective causal model, which the sender knows but does not control. The answer depends on whether the receiver regards the signal as a consequence of the state or as one of its causes. In the first case, the sender never gains from a signal that the receiver misinterprets: the sender is weakly worse off than against a Bayesian receiver, and the receiver is never worse off than without information. In the second case, the receiver's posteriors across signal realizations might not average to the prior. The sender can then be strictly better off, and his gain is the receiver's loss: she can be harmed by information that she values.