Information and Persuasion in Competitive Markets and Politics

Abstract

This dissertation studies the complex interplay between information and competition, using a variety of formal models from general equilibrium and game theory. I first study two models that extend a baseline, general equilibrium model. In the chapter "Investment without Coordination Failures", I prove that, even with incomplete markets, the prospect of competitive markets in the future provides sufficient information for decision-makers today to choose efficient investment levels. In "On the Informational Efficiency of Decentralized Price Formation", written jointly with Rafael R. Guthmann, we prove that decentralized search economies which approximate a competitive economy in terms of allocation do not approximate the competitive economy in terms of the required information. We prove that a model of decentralized market-makers can approximate the informational efficiency of a competitive economy. I then move on to study two different models, using tools from the recent field of information design. Again, I focus on the role of competition. In "Price Competition and the Use of Consumer Data", I study how changes in the information available to competing firms affects the distribution of consumer and producer surplus. I prove that, unlike under a monopoly, complete information--which generates first-degree price discrimination--is optimal for consumers. I conclude by studying a model where people compete through information. I study a voting model where two political parties compete in an election by persuading voters of their candidate's quality. In the unique equilibrium, both parties design campaigns that generate uniform distributions of beliefs. The equilibrium means that voters are equally likely to have a range of beliefs about the candidate's quality after the campaign: some voters believe the candidate is perfect, some believe the candidate is terrible, and some believe everything in between. The level of disagreement is driven by competition; each party does not want to design a campaign that is easy to beat.