Essays in Industrial Organization
Teng, Xuan
2022
Abstract
One theme of industrial organization is to understand firm behavior and its implication on market outcomes. This dissertation studies two firm behaviors that are closely related to competition: self-preferencing in digital platforms and patent licensing relationships between competing manufacturers. Chapter 1 and Chapter 2 study self-preferencing on Apple App Store, a dominant dual-role platform for mobile applications. It is widely perceived that dual-role platforms, which sell both third-party and own products to consumers, may give preferential treatment to own products in search results to obtain competitive advantages over third-party suppliers. However, platforms deny it. Meanwhile, there is an intense antitrust debate over whether and how to regulate self-preferencing of dominant platforms. To shed light on the debate, this study addresses two research questions. Does there exist any self-preferencing on digital platforms? What are the consequences of self-preferencing on product quality and social welfare? I argue that self-preferencing leads to lower product quality and hurts both consumers and third-party suppliers. Meanwhile, the effect of self-preferencing is limited when there is confounded consumer preference. In Chapter 1, I study the demand for mobile applications and their search rankings on Apple App Store, in order to identify the existence of self-preferencing. To that end, I compile a new dataset of consumer search and purchase, app characteristics, and search rankings on Apple App Store from April 2018 to February 2020. To motivate the demand model, I examine the effect of an unanticipated search algorithm change that dropped some of Apple’s apps from top search results. Then, I develop a demand model explicitly accounting for consumer search; and a search ranking model that captures potential self-preferencing. Estimation results direct to self-preferencing: Apple’s apps are significantly more likely to show up in top search results compared to third-party applications, conditional on app quality, installation price, and title match with search terms. To examine the equilibrium welfare effect of self-preferencing, Chapter 2 develops and estimates an empirical model of update competition in the presence of potential self-preferencing. Then, based on the estimates in Chapter 1 and Chapter 2, I conduct counterfactual simulations and find that shutting down the identified self-preferencing increases average update frequency by 0.4 percent, consumer surplus by 0.2 percent, and third-party profits by 0.7 percent. One reason for the modest effects of self-preferencing is consumer preference: consumers are estimated to prefer Apple’s apps, which limits the extent of identified self-preferencing and thus its welfare effect. Chapter 3 studies how unobserved patent licensing between competing manufacturers leads to bias in predicted merger effects. Patent licensing between competing manufacturers is typically hard to observe by researchers and thus not considered in classical pricing competition models for empirical merger analysis. Nevertheless, patent licensing introduces alignment incentives in firms’ pricing decisions since licensors collect royalty revenues proportional to competing licensees’ sales. Omitting these incentives leads to over-estimated marginal costs and potentially introduces bias in the prediction of merger effects. How large are the estimation and prediction biases? We conduct numerical simulations using a Bertrand competition model that incorporates patent licensing between competitors to assess the biases for both licensor-licensee mergers and licensee-licensee mergers in various simulated markets. We find that omitting patent licensing leads to predicted merger effects that are i) sometimes opposite to the true merger effects and ii) typically over-predicted.Deep Blue DOI
Subjects
self-preferencing search algorithm endogenous product characteristics mobile applications patent licensing horizontal merger effects
Types
Thesis
Metadata
Show full item recordCollections
Remediation of Harmful Language
The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.
Accessibility
If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.