Abstract
We study a strategic experimentation game with exponential bandits, in which experiment outcomes are private. The equilibrium amount of experi-mentation is always weakly higher than with public outcomes. Moreover, in pure equilibria, experimentation is never significantly below the socially optimal level and may be strictly higher. We provide a tight bound on the extent of over-experimentation. The analysis rests on a novel encouragement effect.
Keywords
Strategic experimentation; exponential bandit; private payoffs;
Reference
Jérôme Renault, Eilon Solan, and Nicolas Vieille, “Strategic experimentation with privately observed payoffs”, Theoretical Economics, 2026, forthcoming.
Published in
Theoretical Economics, 2026, forthcoming
