Digital Platforms and Antitrust Law

• Publication year: 2021

98 Nebraska L. Rev. 272. Digital Platforms and Antitrust Law

Digital Platforms and Antitrust Law

Keith N. Hylton(fn*)

ABSTRACT

This Article is about "big data" and antitrust law. Big data, for my purposes, refers to digital platforms that enable the discovery and sharing of information by consumers, and the harvesting and analysis of consumer data by the platform. The obvious example of such a platform is Google. The big platforms owe their market dominance not to anticompetitive conduct, but to economies of scale. This Article discusses three types of anticompetitive conduct associated with digital platforms: kill zone expropriation, acquisition of nascent rivals, and denial of access to data. There is nothing so unusual about digital platforms that would require a reform of the antitrust laws. Some are described as two-sided markets, but this designation, even after Ohio v. American Express Co., should not present an obstacle to the application of antitrust law.

Keywords: digital platform competition, platform acquisitions, big data and antitrust, kill zone competition, data and competition, Amex, two sided markets, data monopolies, network effects and competition, data as property

TABLE OF CONTENTS

I. Introduction .......................................... 273

II. Platforms ............................................. 274

III. Competition Issues .................................... 279
A. Kill Zone Expropriation ........................... 279
B. Acquisition of Nascent Rivals ...................... 284
C. Denial of Access to Data ........................... 286

IV. Antitrust Law ........................................ 287

V. Conclusion ............................................ 296

1

I. INTRODUCTION

This Article is about big data and antitrust law. The first question I should address is precisely what I mean by the term "big data." A search of the term on the internet reveals amorphous definitions, leaving one with the impression that big data is a conceptual category that includes several different things.

Big data, for my purposes, refers to digital platforms that enable the discovery and sharing of information by consumers, and the harvesting and analysis of consumer data by the platform. Because of the enormous quantities of data, the platforms must use algorithms and often machine learning to process the data. Thus, big data includes, in this definition, a reference to the sophisticated processes by which data are analyzed.

The obvious example of such a digital platform is Google. Consumers search for information on Google, and Google uses the information from consumer searches to improve its own search algorithms and to provide information to advertisers on consumer preferences. Google tracks the activities of consumers, as far as possible, to build out its database.

The focus here is on competition issues associated with digital platforms.(fn1) However, competition is a broad concept itself. There are many noncompetition issues associated with digital platforms, such as privacy, which, on further inspection, are capable of being described as competition issues after all.(fn2) This Article takes a broad brush to the topic and sweeps in noncompetition issues where relevant.

The largest platforms owe their market dominance not to anticompetitive conduct, but to economies of scale. Still, three types of anticompetitive conduct are associated with digital platforms: kill zone expropriation, acquisition of nascent rivals, and denial of access to data.(fn3) There is nothing so unusual about the platforms that would require a reform of the antitrust laws. Some have been described as

2

two-sided markets, but this designation, even after Amex,(fn4) should not present an obstacle to the application of antitrust law.

Part II below discusses the economics of digital platforms and the processes that have generated economies of scale. Scale benefits platform consumers, but it also precipitates injuries to consumers in the form of privacy invasions, online harassment, and intellectual property infringement. Part III examines anticompetitive theories likely to be asserted against digital platforms. Part IV discusses the antitrust law applicable to these theories.

II. PLATFORMS

The digital platforms with which most of us are familiar are Google and Facebook. Google has a dominant market share in online search in every country except Russia and China.(fn5) Facebook has roughly two billion users-almost half of the entire population of internet users.(fn6) These large customer bases enable Google and Facebook to amass large quantities of data on billions of consumers.(fn7) The data are used to improve services offered on the digital platforms. Google, for example, uses its data on consumers to enable advertisers to target consumers who are most likely to purchase the advertised products. Data also

3

enable Google to continuously improve the quality of its search algorithms.(fn8)

For digital platforms such as Google and Facebook, size results mostly from economies of scale. As the number of consumers increases, Google can amass more data and use the data to improve its search experience and the accuracy of advertising matches. In other words, as the platform expands and obtains more data, the cost of providing a given unit of service (e.g., the answer to a search query) falls. Sophisticated data processing methods, such as machine learning and artificial intelligence, further enhance the scale economies effect.

There are two processes giving rise to economies of scale in the digital platform markets. One is fixed costs. Much of the costs of providing Google's services reside in the employment of engineers, and Google has consistently bid for the highest quality of this form of labor. Increasingly, digital platforms are also hiring content moderators to make sure that the material presented to consumers does not unduly tarnish their brands. The costs of engineers are fixed, in the sense that the volume of the search service Google provides is not highly sensitive to the number of engineers the firm hires, after it reaches a sufficient scale to generate the service. In other words, once a given number of engineers produce the search machine, consumers can use the machine anywhere from zero to an uncountable number of times per day, and the cost to Google is largely the same. The cost of content moderators is probably more dependent on the number of users of the service, but even here the elasticity of moderator cost to intensity of consumer use is well below one.

The other source of economies of scale is data. Data enable engineers to improve the quality of the search process at Google. Thus, each consumer, by providing data, lowers the unit cost of servicing the next consumer.(fn9) The intensity of this effect is magnified by the use of sophisticated data processing methods. Google, therefore, has an interest in investing resources into the enhancement of data processing methods, such as machine learning and artificial intelligence. The demand for such processes from the digital platforms, in turn, bids up the wages of researchers who specialize in data processing methods, bidding some of them out of their academic positions and into industry.

4

Google also has an interest in investing resources into the creation and harvesting of more data on its consumers. It is interesting to note that for many years after Google entered the market, observers wondered how the service would become a source of revenue-that is, how search could be monetized.(fn10) That question has been answered: Google's annual revenue in 2018 was $137 billion.(fn11) Data have proven to be monetizable. Because data are monetizable, Google gathers as much data as it can on users through their own search inquiries and through monitoring their activities online. Using the acquired data, Google generates a detailed map of the preferences, personal characteristics, and location of every internet search consumer.

In addition to the scale economies effect, the law provides a subsidy that has aided the expansion of digital platforms. Under Section 230 of the Communications Decency Act,(fn12) digital platforms such as Google are treated as information intermediaries rather than publishers.(fn13) This relieves platforms of tort liability for harms caused by their activities. Another safe harbor from liability is provided by Section 512 of the Digital Millennium Copyright Act, which protects platforms from copyright liability if they take down an infringing item after notification of the infringement.(fn14) In the absence of these broad immunity shields, platforms such as Google would have to consider their potential liability based on defamation and related tort theories, or copyright law. These costs would be directly related to scale and would compel platforms to monitor their information services in order to minimize liability. The cost of monitoring would constrain the platforms' incentive to expand. The platforms would have to be much more protective than they are now against harms created by platform speech. Since privacy is also protected by tort law, platforms would also have to be more vigilant toward the privacy of consumer data transferred to other entities.

5

Although this Article is about competition issues, the foregoing features of digital platforms have generated important problems currently...