Delayed perfect monitoring in an infinitely repeated discounted game is modelled by allocating the players to a connected and undirected network. Players observe their immediate neighbors’ behavior only, but communicate over time the repeated game’s history truthfully throughout the network....

We consider multi-stage games, where at each stage, players receive private signals about past and current states, past actions and past signals, and choose actions. We fully characterize the distributions over actions, states, and signals that obtain in any (sequential) communication...

Games with imperfect information often feature multiple equilibria, which depend on beliefs off the equilibrium path. Standard selection criteria such as passive beliefs, symmetric beliefs or wary beliefs rest on ad hoc restrictions on beliefs. We propose a new selection criterion that imposes...

The paper shows that Perfect Bayesian equilibria need not be unique in the strategic communication game of Crawford and Sobel (1982). First, different equilibrium partitions of the state space can have equal cardinality, despite fixed prior beliefs. Hence, there can be different equilibrium...