Models Of Competition In The U. S. Motion Picture Industry

Abstract

My dissertation builds and estimates models of competition for the U.S. motion picture industry. The motion picture industry provides a rich setting for exploring the effect of competition. First, there is a rapid refresh cycle with new titles being the mainstay of industry activity. Modeling new releases adds to estimation complexity due to more stringent data demands. Second, the industry sees seasonal changes in demand and competition. The frequent entry of new products leads to an emphasis on release timing and pricing. Third, the presence of multiple generations of technological formats leads to time varying cannibalization and substitution across channels. While I study the motion pictures industry, insights generated and methods developed are generalizable to other industries with multiple-channel distribution, critical entry and pricing decisions, and technology platform transitions. Specifically, I model three elements of competition in this industry. In essay 1, I model how movies compete with other movies in the same channel and other channels (e.g. theaters, rentals, purchases, etc.). I allow for movies within any channel to be substitutes or complements for movies in the same and other channels. In essay 2, I model how release date and price matter for movie competition; the application is to DVD sales channel. In essay 3, I model how studios compete in format wars. The application is to studios planning whether to release movies in DVD and/or VHS and at what price, in the days when the DVD format was growing, and studios had the choice of waiting for the rivals to do the work of subsidizing the new format. My dissertation delivers insights on managing marketing mix variables across formats, accounting for seasonal demand, competition and cost differences. The models both allow inference of payoffs, and provide counterfactuals for future actions of managers, accounting for competition. The model frame works described provide methodological advances for model specification and estimation in other industries allowing generalizability. In particular, I contribute to the literature in dynamic games by introducing partial and complete information estimators for games with multiple potential equilibriums played in the data.