Essays in Industrial Organization and Econometrics

Abstract

This thesis includes two essays in Industrial Organization and one in Econometrics. The first two essays study the returns to the scale of data available to firms and its impact on market structure, using actual firm data from an online retailer that optimizes its product display on the basis of revealed preference. While the first chapter presents an approach that is directly empirical looking at the quality of out-of-sample predictions, the second, which is the main chapter of my dissertation, is model-based, allowing for a competitive market analysis and accounting for reoptimization of learning based on firm size. Both papers find moderate to significant returns to data depending on a firm's position on the learning curve. The third essay shows that in settings with interval-censored data, which is common for instance with survey data, information on the marginal distribution of said regressors can substantially tighten identification regions when combined with restrictions on the regression function such as monotonicity and shape.