Identification and estimation of incomplete information games with multiple equilibria

In games, the multiplicity of equilibria poses a challenge for identification and estimation. The existing literature typically abstracts from this multiplicity by assuming that the data are generated from a single equilibrium. Instead of imposing such restrictions, this paper provides sufficient conditions to non-parametrically identify payoff primitives in finite action games with incomplete information, while allowing for multiple equilibria. I then propose a two-step estimator and illustrate its finite-sample performances via Monte Carlo simulations. Furthermore, I study the strategic interaction among radio stations when choosing different time slots to air commercials. I indeed find evidence to support the existence of multiple equilibria.