Discussion Paper No. 942THE IMPACT OF MONITORINGIN INFINITELY REPEATED GAMES:PERFECT, PUBLIC, AND PRIVATEMasaki AoyagiV. BhaskarGuillaume R. FrechetteJuly 2015The Institute of Social and Economic ResearchOsaka University6-1 Mihogaoka, Ibaraki, Osaka 567-0047, Japan
The Impact of Monitoring in Infinitely RepeatedGames: Perfect, Public, and Private*Masaki AoyagiOsaka UniversityV. BhaskarUniversity of Texas at AustinGuillaume R. Fr´echetteNew York UniversityJuly 27, 2015AbstractThis paper uses a laboratory experiment to study the effect of a mon-itoring structure on the play of the infinitely repeated prisoner’s dilemma.Keeping the stage game fixed, we examine the behavior of subjects when in-formation about past actions is perfect (perfectmonitoring), noisy but public(publicmonitoring), and noisy and private (privatemonitoring). We find thatthe subjects sustain cooperation in every treatment, but that their strategiesdiffer substantially in the three treatments. Specifically, we observe that thestrategies are more complex under public and private monitoring than underperfect monitoring. We also find that the strategies under private monitoringare more lenient than under perfect monitoring, and less forgiving than underpublic monitoring.JEL classification: C72, C73, C92Key words: Infinitely repeated games, monitoring, perfect, public, pri-vate, experiments*We wish to thank Matthew Embrey and Sevgi Yuksel as well as seminar participants at IASPrinceton, Southern Methodist University, WZB, Stanford, Toulouse, Science Po, Harvard Univer-sity, the University of Montr´eal, University of Heidelberg, the University of Munich, the Max-PlankInstitute, the University of Indiana, and conference participants at the North American Meeting of theESA, the Social Dilemmas conference, and the SITE conference for helpful comments and sugges-tions that improved the paper. We would further like to thank the National Science Foundation (grantSES-SES-1225779), the CV Starr Center and the Center for Experimental Social Science (CESS) forresearch support. We are responsible for all errors.1
1IntroductionMany economic situations involve repeated interactions among players who do notknow for sure what actions other players have chosen, or what information theyhave observed about those actions. In a pioneering work, Stigler (1964) studies amodel of secret price cutting and points out the difficulty faced by the firms whoattempt to collude under (imperfect)private monitoring,where information aboutthe past actions (price choices) of other players is noisy and privately observed.As Kandori (1992) notes, and by now is well recognized, private monitoring im-plies the absence of a coordination device for the players, and hence the lack ofa recursive structure in the repeated game. In this sense, there is a fundamentaldifference between private monitoring and the more traditional environments ofperfect monitoring,where information about past actions is perfect (and public),and (imperfect)public monitoring,where the information is noisy but public. The-ory shows that any equilibrium that sustains a positive degree of cooperation underprivate monitoring must entail the play of an intricate mixed strategy.
Upload your study docs or become a
Course Hero member to access this document
Upload your study docs or become a
Course Hero member to access this document
End of preview. Want to read all 39 pages?
Upload your study docs or become a
Course Hero member to access this document
