Some Strategic Aspects of Private Information: An Experimental Study

In this paper we experimentally analyze the use of private information in a class of zero-sum repeated games with incomplete information on one side and perfect monitoring. We analyze whether the information disclosure by the informed players, and its use by the uninformed players, matches the theoretical predictions. We consider two games that differ according to the amount of information that the informed player should optimally disclose: in the first game, the informed player should entirely conceal his information. In contrast, in the second game the informed player should fully disclose his information. We find that the flow of information is higher than predicted in the first game and lower than predicted in the second game. However, the use of information is strictly higher in the second game than in the first one. Uninformed subjects tend not to use the revealed information in the first game, and seem to misinterpret the revealed information in the second game.