The role of information in repeated games

We show that in repeated interactions the avenues for effective provision of incentives depend crucially on the type of information players observe. We establish this conclusion for general repeated games in which players learn information via a continuous-time stationary process that has a continuous multi-dimensional Brownian component and a Poisson component and in which the players act frequently. The Poisson jumps can be used to effectively provide incentives both with transfers and value burning while continuous Brownian information can be used to provide incentives only with transfers. † We thank Dilip Abreu, Eduardo Faingold, Drew Fudenberg, Ben Golub, Patrick Hummel, David Levine, Yuval Salant and Larry Samuelson and three anonymous referees for helpful comments . We also thank seminar participants at Harvard, University of Rochester, Cowles 75 Anniversary Conference, SED 2006, WEM 2006. * Sannikov: Department of Economics, University of California at Berkeley, [email protected]. Skrzypacz: Stanford Graduate School of Business, [email protected].