1. Per Krusell and Anthony A. Smith, Jr. Consumption Savings Decisions with Quasi-Geometric Discounting
A consumption-savings problem of an infinitely lived consumer with beta-delta preferences is shown to have many Markov perfect equilibria. In particular, the consumer's capital holdings may converge to any point in a wide interval.
2. George J. Mailath and Stephen Morris Repeated Games with Almost-Public Monitoring
This paper provides positive and interpretable results for repeated games where monitoring is private (i.e. each player privately observes a noisy signal of opponent actions), but close to perfect. The main result shows that if in a (strict) perfect public equilibrium of a game with public monitoring, players condition their behavior only on a finite history, then such an equilibrium also exists in close-by games with private monitoring. The paper gives several instructive examples, illustrating how the restriction on memory implies that players have approximate common knowledge of the state of the game.
3. Christopher Harris and David Laibson Instantaneous Gratification
Worried that hyperbolic discounting is unusable? That there are a continuum of equilibria? This paper gives a clean usable formulation of hyperbolic discounting in continuous time with an illustrative application to a savings problem.
4. Brian Skyrms Signals, Evolution and the Explanatory Power of Transient Information
Talk about consumers' surplus; Skyrms shows that in evolutionary dynamic models of games, talk can be extremely valuable, even though it is cheap. In games like the stag hunt or Nash bargaining game with multiple Nash equilibria, a cheap talk phase can create new polymorphic equilibria, and change the stability and the basins of attraction of equilibria in the base game.
5. Sandeep Baliga and Tomas Sjostrom Arms Races and Negotiations
A "solution" to Schelling's burglar's dilemma. Players are randomly matched to play a two-person game where a player's type is private information. For some low types, defect is dominant strategy, while for higher types "cooperate" is a best response if and only if the probability that one's match cooperates is high enough. In equilibrium without talk, mutual distrust feeds on itself and all defect; but with pregame cheap talk, there is a remarkable partially separating equilibrium that maintains a good deal of cooperation.