Why did Bergemann and Morris choose a different setting with respect to Kamenica and Gentzkow only to reach to the same results? However none of both approaches seem to present the problem in a perfect way. Is there any critique on the literature or any other paper that follows a third road to model Bayesian persuasion problem?
$\begingroup$
$\endgroup$
9
-
$\begingroup$ I'm not sure what you mean by "same results." Nor what "none of both approaches seem to present the problem in a perfect way" means. $\endgroup$– Michael GreineckerCommented Feb 18 at 18:36
-
$\begingroup$ Their conclusion is the same, no matter the approch. Namely, the celebrated outcome is the optimal information strucutre is the partial revelation of information (from my understanding) either with the one sender- receiver or the one sender - many receivers. They use different approachs, KG with the geometric approach of the concavyfication of the senders value function and BM with the linear programming and the analytical solution of the problem. The ``none of them presents the problem in a perfect way" has to do with the limitation in applications of both models $\endgroup$– Oliver QueenCommented Feb 19 at 10:32
-
$\begingroup$ I would say the central result of BG is to characterize the possible Bayes.Nash equilibria of a game in which the players might have information the analyst is unaware of. The central results of KG (already obtained by Aumann & Maschler) are the characterization of posterior distributions and concavification. These are very different results. If you want to know about tools that work better for some applications, you should mention these applications. I would not criticize hammers for being bad at screwing in screws. $\endgroup$– Michael GreineckerCommented Feb 19 at 11:13
-
$\begingroup$ To my opinion they both result to the same conclusion, that is that the optimal information structure is the partial revelation of information. I am in search of the tools I need, I do not know them. I am just coming to an conclusion, this does not mean that I do not understand the work of the authors, but yet it remains imperfect. A persuasion problem is more than just characterizing posterior beliefs or equilibrium concepts. Everything in economics and finance could be seen as a persuasion aspect, but still most of the literature turns around the same schematics. So this seems a bit odd. $\endgroup$– Oliver QueenCommented Feb 19 at 13:08
-
$\begingroup$ Your issue is that what these papers do does not correspond to the usual usage of "persuasion?" $\endgroup$– Michael GreineckerCommented Feb 19 at 13:16
|
Show 4 more comments