1. Eduardo Faingold and Yuliy Sannikov Equilibrium Degeneracy and Reputation Effects
This paper examines reputation in continuous time models where a noisy signal of the long-run player follows a diffusion process. Without "KWMR" types the equilibrium is completely degenerate and the long-run player is limited to the static Nash equilibrium payoff. With "KWMR" types the equilibrium is non-degenerate. The key idea is that the length of effective horizon for the "audience" of short-run player(s) is critical. Without "KWMR" types, in continuous time, reaction to the long-run player must occur continuously, and the diffusion is very noisy over short time intervals. With "KWMR" types longer term information matters - a reputation is not won in a day - and over a longer period of time the diffusion is much less noisy.
2. Qiang Fu and JingFeng Lu The Optimal Multi-Stage Contest
A Principal wants to maximize productive effort from a group of agents. The principal has a fixed budget to be allocated as prizes in some contest. This paper considers a general set of contests that potentially involve multiple knockout stages and analyzes the optimal multi-stage structure as well as the allocation of prizes over time. The effort-maximizing contest eliminates a single contestant in each period until two remain in the "finale" and reserves all prize money for the winner of the finale.
3. Jeffrey C. Ely Kludged
Probably we should try to avoid the editors reviewing each others papers - but this one I can't resist. Anyone who has ever written computer code realizes that patches accumulate, and as they accumulate it gets harder to write additional code. Eventually programmers tear the thing apart and start over again, sometimes with good results, sometimes (Microsoft Vista) with catastrophic ones. Evolution of biological organisms is limited to patches - evolutionary processes cannot start all over again at the bottom. This paper works out an explicit evolutionary model. Even with large mutations occuring infinitely often, behavior can be perpetually suboptimal. (My own thought: evolution produced computer programmers, who can start over again.)
4. Jean-Pierre Benoit and Juan Dubra Overconfidence
It has been firmly established in the experimental laboratory and in survey data that Garrison Keeler is right and everbody thinks that they are above average. It turns out that in a population of rational people with rational expectations and noisy data, this is exactly as theory predicts.
5. Emir Kamenica and Matthew Gentzkow Bayesian Persuasion
A DA who always wishes to convict structures a case to a judge who wishes to do the right thing. The DA can select the forensic tests to perform-- which must be truthfully reported-- such that the judge will rationally convict a larger fraction of those on trial than are actually guilty. This paper breathes new life into Aumann and Maschler's results for repeated games with incomplete information.