In our digital economy, privacy has taken center stage. Given that spotlight, we have already seen regulatory intervention into markets with the EU’s GDPR and DMA. (More generally, GDPR and DMA are part of a larger body of regulation that the EU has passed or is contemplating passing to address large platforms. (See Márton Sulyok’s “How to Tackle IT?” published on this blog.) While the verdict is still out, the early empirical evidence strongly suggests that whatever its privacy benefits, the GDPR has had negative economic consequences.
Because the large tech platforms that tend to be in the bullseye of regulators on both sides of the Atlantic give their products away and live off consumer information, a conventional wisdom that has arisen is that market power becomes manifest through degraded privacy protections. In other words, the assertion is that, when platforms have more market power, they lower their privacy quality. Yet, in a recent article, Antitrust & Privacy: It’s Complicated, our empirical results challenge this conventional wisdom. In this blog post, we contribute to the debate surrounding personal data protection that has already been started by Bianka Maksó and Lilla Kiss also on this blog.
Privacy and antitrust have been on a collision course for some time now. For instance, in the U.S., an executive order from the president denounced dominant online platforms for using their market power “to gather intimate personal information that they can exploit for their own advantage.” The chair of the U.S. Federal Trade Commission (FTC) has expressed concern that “[m]onopoly power […] can enable firms to degrade privacy without ramifications.” In the EU, the story is the same. For example, the German Bundeskartellamt brought a case against Facebook based on the theory that violating consumers’ privacy right under the GDPR gave Facebook a data advantage the helped cement its dominant position.
On a superficial level, the negative relationship between privacy quality and market power sounds “right”—after all, we often hear that “if the product is free, then the product is me.” This leads to the following testable hypothesis: if data is the price that we pay for using these free platforms, market power will become manifest through lower levels of privacy.
In our paper, we address this hypothesis both theoretically and empirically. On a theoretical level, equating privacy and price is problematic for several reasons. First, while privacy is a “normal good,” in that, all else equal, consumers prefer more privacy to less, how consumers value privacy relative to other products is uncertain. Specifically, the “Privacy Paradox” suggests that, although consumers profess to care deeply about their privacy in surveys, their revealed behavior suggests otherwise. The root cause of this paradox is the subject of considerable debate. But whether rational choice, asymmetric information, or cognitive biases are to blame is beside the point—if privacy does not drive consumers’ marketplace choices, then privacy is not a relevant dimension of competition.
Second, unlike price, user data is an input into a larger production process to produce some type of output. That is, unlike a monopolist who enjoys increased profits immediately when they exercise market power by reducing the quality of their product (and, hence, the monopolist’s costs), a firm can profit from increased levels of data collection only by taking an action to monetize them. And this monetization process provides benefits, typically through customization of advertisements or services (e.g., recommendation engines in streaming services or bespoke workouts in fitness apps). Thus, the relationship between the collection of user data and consumer welfare is not necessarily negative—again, unlike in the case for price. Finally, contrary to popular opinion, there is no general economic result that establishes a relationship between greater competition and product quality. To the extent that we view privacy as a dimension of quality, the result carries through—there is no a priori reason to assume that competition is more likely to result in better privacy protection than monopoly.
The relationship between competition intensity and privacy quality is further complicated for multisided platforms, which cater to both users and advertisers/sellers. Put simply, while users may value more privacy, advertisers/sellers value less user privacy. A platform balances the competing incentives of these two groups. Moreover, if we consider that users themselves may benefit from more personalized content generated from user data, then the story is even more complicated.
Theory can only take you so far, however. What is happening in the real world? Namely, what is the empirical relationship between market power and privacy quality? Surprisingly, little work has been done to answer this question. We attempt to fill that void with our study. We examined the relationship between various measures of market concentration—an imperfect proxy for market power, but one used by competition authorities throughout the world—and privacy levels for mobile apps on the Google Android platform and popular websites.
For mobile apps, we measure privacy quality using PrivacyGrade.org, which is a third-party assessment of app quality from a group of researchers at Carnegie Mellon University. Our results suggest no relationship exists between privacy grades and our proxies for market power, that is, market shares based on Google Play Store categories, and market concentration, i.e., the Herfindahl-Hirschman Index (HHI). We also find a robust, negative relationship between privacy and app quality ratings, consistent with a tradeoff between privacy and other dimensions of product quality that consumers value.
For websites, we measure privacy quality using DuckDuckGo’s privacy ratings for websites in thirty-seven website categories (e.g., Search, Health, News). These website categories are from SimilarWeb. While these categories do not necessarily correspond with “relevant product markets” used in competition law, they represent independently created grouping of sites that are based on content tags and website self-identification. Again, the results suggest no relationship between privacy ratings and market concentration measures.
Combined, our empirical results cast serious doubt on the validity of the conventional wisdom that firms exercise market power by reducing privacy, and also suggest that app developers use consumer data to enhance the quality of their products. What this means for competition policy is that antitrust law appears to be a poor vehicle to address perceived privacy problems. To the extent that the marketplace is failing to produce optimal levels of privacy, we suggest that consumer protection aimed at increasing consumer access to information and the firms’ ability to credibly commit to higher privacy quality promises is likely to be the better policy tool.
The presumption that privacy and market power are linked is neither supported by theory nor empirics, which suggests that bringing high-profile antitrust cases against large platforms is unlikely to result in higher levels of privacy protection. The relationship between privacy and market power is complicated, and as such, the debate surrounding competition law and privacy could benefit from an injection of both nuanced theoretical considerations and more empirical evidence.
James C. Cooper is Professor of Law and Director, Program on Economics & Privacy, Antonin Scalia Law School, George Mason University; previously served as Deputy Director of Economic Analysis in the Bureau of Consumer Protection, U.S. Federal Trade Commission.
John M. Yun is Associate Professor of Law and Deputy Executive Director, Global Antitrust Institute, Antonin Scalia Law School, George Mason University; previously served as an Acting Deputy Assistant Director in the Bureau of Economics, Antitrust Division, U.S. Federal Trade Commission.