All Roads Lead to Rome: Enforcing the Consumer Welfare Standard in Digital Media Markets
Thank you very much Judge Ginsburg, David, and Antonio, for the kind invitation to join you and this impressive array of speakers and panelists today. And thank you Chairman Pitruzzella for hosting us. It’s an honor to be here with you in Rome, the capitol of one of the greatest early empires in world history. It is also particularly fitting that the topic of this Jevons symposium is technology and media, two areas in which the early Roman Empire stood at the forefront of innovation. Of course, all of you are familiar with the stunning Coliseum nearby, home to many extraordinary spectacles. The early Romans had a knack for giving the crowds what they wanted to see. A more significant and lasting Roman innovation was its ability to manage an empire that expanded across much of the known world and remained largely stable for hundreds of years. The Romans accomplished that feat, in part, by creating one of the earliest networks in history: the complex array of roads across their territories. The Roman road system is one of the earliest reminders of the power of network effects.
By integrating its vast territory, Roman roads facilitated communications, trade, and—ultimately—a sense of Roman identity that persisted across the continent long after the fall of the Roman Empire. Network effects pushed the Roman Empire outward, but they also opened Rome up to greater risk from outside its borders. Rome eventually succumbed to the inevitable, which was invasion. Indeed, the Roman example is one of the earliest historical proofs that “monopoly”—even in network industries—can be inherently unstable over time.
The concept of network effects is not new, but the medium for creating them has evolved beyond the wildest dreams of the early Romans. Digital platform markets and the opportunities they provide for innovation have captured the imaginations of inventors, financiers, tech mavens, entertainers, and consumers across the globe. They have also drawn close attention, especially recently, from enforcers and regulators who see serious threats to established modes of business and competition.
As you are all aware, some critics assert that the antitrust consensus is not equipped to address competitive threats posed by new developments in technology—digital markets and platforms in particular. I don’t endorse that view. Indeed, last month at the University of Chicago Booth School of Business, I emphasized that the bipartisan antitrust consensus is flexible to challenges posed by digital platform markets because it can incorporate the latest economic wisdom in determining whether business practices or transactions are harmful to competition and consumers.
I focused my remarks in Chicago on one prong of the antitrust consensus—what I, and others, refer to as an “evidence-based” approach to antitrust law enforcement. That approach requires enforcement built on credible evidence that a practice harms competition and the American consumer, or in the case of merger enforcement, that it creates an unacceptable risk of doing so. Taking an evidence-based approach also means being open to persuasion. Where there is strong evidence of harm to competition, it is the duty of enforcers promptly and vigorously to investigate and enforce the antitrust laws.
Today, I would like to focus my comments on the other prong of the antitrust consensus—the “consumer welfare standard”—and how that standard is flexible to new business models generally and digital media markets, in particular.
The consumer welfare standard is premised on the idea that consumers benefit from free market competition because it increases economic efficiency, often in the form of lower prices or increased output. Outside the realm of naked horizontal agreements, courts and agencies in the United States traditionally analyze whether restraints or mergers may raise prices or reduce output.
With respect to price effects, antitrust lawyers and economists have developed certain rigorous tools for analyzing whether prices to consumers are supra-competitive. These tools have proven flexible to a variety of business contexts. Indeed, in the coming weeks, the Supreme Court in the American Express case may help clarify how to evaluate price effects in the context of two-sided platforms and whether the rule of reason and market definition principles ought to be re-tailored in order to do so.
Output effects, by comparison, have attracted less attention in the context of digital platform markets. That is likely because “output” is often a murky concept for a company that is a digital intermediary between market actors. For a social media network, should we measure output based on the number of users? How much time they spend using the network? The number of pictures they share or view? The number and length of political rants? The total number of “Likes”? It is not clear which of these metrics best captures “output” as a measure of how a digital platform offers a product in response to individual preferences.
I believe that three additional indicators of consumer welfare deserve greater attention in analyzing competitive effects in digital markets: innovation, choice, and quality. The Supreme Court and other courts, in describing antitrust as “a comprehensive charter of economic liberty,” often invoke these three concepts and note that innovation, consumer choice, and product quality constitute competitive effects that merit consideration in an antitrust analysis. Because these factors can be difficult to quantify, they often play a subsidiary role to price and output measures.
Given the challenges in defining and measuring output in platform markets, innovation, choice, and quality can serve as valuable metrics for competitive effects. They are all consistent with the Sherman Act’s overarching consumer welfare prescription.
First and foremost, innovation is central to consumer welfare. In a free market economy, new businesses emerge by offering consumers something new, rather than simply more of the same. For that reason, innovation is inherently disruptive, making it a target of entrenched business models that see existential threats from new entrants. Competition policy should encourage these threats to incumbents, not restrict them. The cycle of dynamic competition almost invariably accrues to the benefit of consumers.
Second, consumer choice can be an important metric for consumer welfare effects to the extent a practice or merger results in the elimination of a unique product offering. As enforcers, we must carefully analyze whether a company with market power uses that power in a manner that excludes an innovative product, service, or feature that customers desire. Likewise, concerted action—particularly in the context of standard setting organizations—can stymie consumer choice where competitors adopt self-serving standards that restrict consumers’ ability to seek out more attractive or less expensive alternatives.
The Microsoft case is illustrative. By “integrating” Internet Explorer and Windows, Microsoft both “prevented OEMs from pre-installing other browsers and deterred consumers from using them.” In particular, Microsoft excluded Internet Explorer from the utility permitting users to add or remove programs, and commingled browser and operating system code. Taken together, these actions bound “Internet Explorer to Windows with . . . technological shackles,” which “discourag[ed] OEMs from distributing rival products” that consumers might choose over Microsoft’s default option. Microsoft offered no valid justification—technological or otherwise—for these actions, leading to the conclusion that its actions “protected its operating system monopoly from a middleware threat” that would increase consumer choice. Importantly, the competition from middleware would diminish the monopoly Microsoft enjoyed in the operating system market as a result of its hold on application programming interface protocols and the network effects that went with it. The en banc D.C. Circuit, accordingly, unanimously upheld the district court’s finding that depriving consumers of competitive alternatives to its monopoly operating system constituted anticompetitive effects in violation of Section 2 of the Sherman Act.
Of course, we must be careful not to condemn the elimination of choice as inherently suspect—particularly in the context of merger review. After all, the goal of the antitrust laws is to protect competition, not competitors. A substantial number of mergers result in fewer choices in the marketplace, yet they nevertheless pose no serious concerns to competition, or indeed may be clearly procompetitive. Where a merger would give the combined entity the incentive and ability to undermine innovative competitors offering new product choices, however, there may be a strong ground for an enforcement action. This is particularly true if the result is to limit innovative choices that threaten a less efficient incumbent business.
Third, product quality can be a useful barometer for whether a merger or business practice harms competition to the detriment of consumers. Courts recognize that a reduction in competition can result in less innovation, and less of a need to provide a high quality product. The notion of “quality” is not limited—it incorporates more than simply the shininess of the new product or the box it arrives in. Particularly for media and technology companies, quality is best captured as the entire customer experience.
Although “quality” may be hard to measure with precision, there are a number of metrics for analyzing the quality of customer experience. In 2003, an article in Harvard Business Review popularized the concept of the “net promoter score,” or “NPS.” NPS measures whether customers or users, on average, are likely to recommend a product to a friend or a stranger (a positive score) or to badmouth the product (a negative score). The idea quickly took hold, particularly as online social media emerged as a dominant method of communication and significant source of network effects—every Facebook or Twitter user could now be a powerful promoter or detractor of a product. According to one company that specializes in NPS measures, streaming media companies (like Pandora or Spotify) rank among the highest performers, and—not surprisingly—cable TV and internet service providers rank among the lowest.
Academics and enforcers should consider whether tools such as NPS and similar benchmarks are useful for measuring quality as a byproduct of competition. For example, in markets where NPS scores are available, we may ask as part of our comprehensive analysis: has a merger lowered or increased the merged company’s scores; or does a proposed merger raise the risk of reducing quality? It could be interesting to observe whether NPS scores over time may be correlated with increases or declines in competition.
Just as measures of consumer welfare warrant flexibility in the context of digital markets, as enforcers we should also be mindful in understanding how barriers to entry enable or facilitate exclusionary conduct that harms consumers. One traditional method of considering entry barriers focuses on fixed costs that a new entrant must incur—that is, it asks whether the risk of upfront investment is too high and the likelihood of success too low, thus deterring entry. In digital markets, entry barriers often appear to be quite low. A new software product can be revolutionary and gain popularity with relatively minimal investment in physical infrastructure, and certainly no regulatory barriers such as the need for spectrum licenses. At the same time, powerful network effects can create their own implicit barriers to entry in markets where a dominant player has clearly emerged, making traction for a new entrant difficult to achieve regardless of its efficiency or quality.
These insights are relevant to claims of exclusionary conduct, including predatory pricing. Predatory pricing claims face a high standard, as the U.S. Supreme Court requires a showing of a “dangerous probability of recoupment,” as well as pricing “below an appropriate measure of [a] rival’s costs.” There has nevertheless been a growth in academic work attempting to re-examine the foundations for predatory pricing claims in digital markets.
As I explained in Chicago, a careful, evidence-based approach is necessary to evaluating these claims, in order to minimize the risk of false positives. Fresh thinking, including theoretical and empirical work in this area, is warranted. A key question that enforcers should answer is whether the presence of network effects can dilute the ability of new competitors to emerge or whether they pose a significant competitive challenge to entrenched incumbents. Network effects may work in different ways in different markets, but in certain markets, they may operate such that there is a higher risk—or an increased probability—of recoupment following a period of below-cost pricing.
To conclude, I encourage further research and civil debate on these important issues, as we are having here. As enforcers, we must be very careful in our enforcement actions to ensure that we don’t punish the very competitors who have won the race we have encouraged them to compete in. At the same time, we must be vigilant in utilizing the tools provided to us within a sound antitrust and economic framework. In particular, we should not hesitate to bring an action where there is evidence of harm to competition manifested through higher prices, lower output, reduced innovation, undue restrictions in consumer choice, or a serious deterioration in quality. Failure to enforce the antitrust laws in a timely manner may result in heavy-handed government regulation later. Rarely a preferred result to free market competition.
We have and should maintain a flexible, dynamic consumer welfare standard that is well-equipped to face threats to competition in media markets in the digital age.