Elias Tsakas

Department of Economics
Maastricht University

 Home   Curriculum Vitae   Research   Teaching   Pictures   Research Links   Other Links 
Noisy persuasion
(with Nikolas Tsakas)

Abstract.
We study Bayesian persuasion when communication between the sender and the receiver is noisy. Our first result proves that an optimal signal for the sender always exists. Our proof does not rely on the standard concavification technique, which as we show cannot be used in the presence of noise. Thus, the standard condition for persuasion (Kamenica and Gentzkow, 2011) is only necessary but not sufficient when communication is noisy. Our second result shows that noise always makes the sender worse off. However, it is not necessarily the case that the sender’s expected utility is decreasing in the communication channel’s informativeness (a la Blackwell). In order for the latter to be true additional assumptions on the channels or on the preferences need to be imposed. Finally, we present various applications of our theory, with a special focus on cases where  the sender benefits from strategically increasing the amount of noise via delegation and/or intermediation.