Working Paper: CEPR ID: DP14075
Authors: Johannes Hrner; Nicolas Klein; Sven Rady
Abstract: This paper considers a class of experimentation games with Lévy bandits encompassing those of Bolton and Harris (1999) and Keller, Rady and Cripps (2005). Its main result is that efficient (perfect Bayesian) equilibria exist whenever players’ payoffs have a diffusion component. Hence, the trade-offs emphasized in the literature do not rely on the intrinsic nature of bandit models but on the commonly adopted solution concept (MPE). This is not an artifact of continuous time: we prove that such equilibria arise as limits of equilibria in the discrete-time game. Furthermore, it suffices to relax the solution concept to strongly symmetric equilibrium.
Keywords: Two-armed bandit; Bayesian learning; Strategic experimentation; Strongly symmetric equilibrium
JEL Codes: C73; D83
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| efficient perfect Bayesian equilibria exist in bandit games (C73) | improved experimentation outcomes (C90) |
| diffusion component (F29) | efficient perfect Bayesian equilibria exist in bandit games (C73) |
| strongly symmetric equilibria (SSE) (D50) | inefficiencies associated with MPE can be mitigated (D61) |
| diffusion component (F29) | efficiency of experimentation is contingent upon players' patience and nature of rewards (C90) |
| strongly symmetric equilibria (SSE) (D50) | equivalent total payoffs to those predicted by MPE (D79) |