| This dissertation consists of a collection of three essays on the application and estimation of behavioral models in microeconomics, particularly in the field of industrial organizations. The focus of the first paper is on developing a methodology for dealing with behavioral model uncertainty in structural oligopoly models. Within the structural framework, the purpose of a behavioral model is to establish a relationship between observations of equilibrium outcomes and the structural parameters of primitive functions. It is well recognized that being an essential part of the identification strategy, the particular choice of a behavioral model embodies a highly influential, yet largely arbitrary, set of assumptions. The methods developed here are founded in Bayesian model averaging techniques and provide a practically and conceptually desirable way of accommodating behavioral model uncertainty in structural estimation. Moreover, a substantial feature of this approach is that it yields straightforward model comparison through the model posterior distribution. These methods are applied to estimate the parameters of the industry demand curve and firms' cost functions in oligopoly markets (e.g. marginal costs, markups, etc.). Five models of oligopoly behavior are considered: non-cooperative, two variations of cooperative with unobserved demand shocks, and two variations of cooperative with observed demand shocks. The specific industry analyzed is the 1800s railroad cartel, commonly known as the Joint Executive Committee, which is widely familiar to industrial organizations economists. The results indicate that the algorithm methods preforms quite well in correctly identifying cooperative behavior, as well as offer a clear view of the way in which model averaging resolves conflicting in inference arising from competing behavioral models.; The second paper develops a simple model to analyze the behavior of duopolist firms learning about their own demand by observing the pricing decisions of better-informed firms. This learning is accomplished through Bayesian updating. Consequently, a simple model is constructed in which firms choose prices with mixed strategies in equilibrium. This result allows for non-trivial learning on the part of the less-informed firms and possible deception by the better-informed firms. The final paper describes an approach used in the nonparametric estimation functions subject to regularity conditions. Such conditions are often necessary to guarantee that the estimated functions, such as supply and demand curves for example, reflect appropriate behavior of economic agents (e.g. behavior that is consistent with economic theory). The proposed approach is shown to be strongly founded in the familiar Bayesian nonparametric estimation. Besides being practically appealing, it also exhibits attractive theoretical properties, relative to currently available methods. Consequently, a Gibbs sampling algorithm is derived to nonparametrically estimate unknown functions subject to slope and convexity restrictions. The application and effectiveness of this algorithm is analyzed with a Monte Carlo study. |
