Essays in Industrial Organization and Networks

Abstract

The three essays of this dissertation consider rich network-like environments. In these environments, properly modeling the intricate relationships between agents is the key to capturing the important economic mechanisms that take place. In the first essay I introduces a flexible Bayesian structural model suitable for the analysis of social learning, competition, and diffusion in environments where firms consider entry to a new market. The model is estimated exploiting a reform in Illinois that legalized slot machines, and empirically study how information and adoption diffuse through a network. I find that when small firms make adoption decisions they react to others’ outcomes, yet they do not use all the relevant information. Moreover, they learn from more firms than they compete with. Finally, increasing information availability or learning substantially increases both adoption and total profits in the market. The second essay is co-authored with Krishna Dasaratha and Ben Golub. The essay is a theoretical study of an environment in which agents that are embedded in a social network learn about a changing state of the world. We examine when agents can aggregate information well, responding quickly to recent changes. A key condition for good aggregation is that each individual's neighbors have sufficiently different types of private signal qualities. In contrast, when signals are homogeneous, aggregation is suboptimal on any network. We show that achieving good aggregation requires a sophisticated understanding of correlations in neighbors' actions. The model provides a Bayesian foundation for a tractable learning dynamic in networks, closely related to the DeGroot model, and offers new tools for counterfactual and welfare analyses. The third essay studies biases that experts could suffer from when they provide their opinions. This paper suggests a method on how to put values on many alternatives, in a case when there is a small number of bilateral comparisons. This method is then used to create a ranking of all Division-I college football teams that is based solely on wins and losses. Comparing this ranking to the official experts-based ranking, I show that the expert opinion ranking is biased towards teams with higher viewership and more star players. At the same time, another experts’ ranking that is not used by the NCAA is not biased towards teams with high TV viewership, suggesting that the bias of the experts may be due to some incentives they have, either explicit or implicit.