Essays on the Econometrics of Games

This dissertation consists of three chapters related to the identification and estimation of games. The first chapter studies the identification of a class of semiparametric network games, which includes entry games, technology adoption with network effects, and social interactions as special cases. The main result is that, under weak assumptions that allow many issues of empirical relevance, the model is point identified. The second chapter studies the relation between the Bayesian and frequentist approaches to testing a one-sided hypothesis about a multivariate mean, which includes testing moment inequalities as a special case. Moment inequalities arise in many partially identified models, including games. The main result is that the Bayesian approach and the frequentist approach can give substantially different answers. The third chapter (with Elie Tamer) studies the nonparametric identification of best response functions in games. The main result is that these objects are generally only partially identified under standard assumptions from economic theory, but are point identified under sufficiently strong assumptions