Don’t Stop Believin’: Antitrust Enforcement in the Digital Era
Thank you for having me here today. I am pleased to join such an impressive group of legal academics, economists, and practitioners here in the Windy City.
I’m honored to provide the lunch keynote today, sandwiched between my friend and former Antitrust Modernization Committee colleague, Professor Dennis Carlton, whom you heard from this morning, and my friend, Professor Carl Shapiro, whom you will hear this afternoon. It is also a great honor to see my friend, former Prime Minister of Italy, and former EU Competition Commissioner Mario Monti, who was certainly a pioneer on issues of the economics of competition.
For me, it feels like I am standing on hallowed ground here at the University of Chicago, where some of the world’s leading antitrust scholars have played a crucial role in developing modern doctrine—Ed Levi, Aaron Director, Robert Bork (who was mentored here by Aaron Director), Frank Easterbrook, Richard Posner, DOJ Antitrust alumnus Diane Wood, Bill Landes, and Richard Epstein, just to name a few. At the Justice Department now, we have numerous graduates of the University of Chicago, including my Principal Deputy Andrew Finch, my Deputy for Litigation Don Kempf, and, of course, United States Solicitor General Noel Francisco. Our chief economist, Professor Luke Froeb, was a Kramer Foundation Fellow here, and one of our front-office counsel, Emma Burnham, is a University of Chicago Law graduate. This University’s contributions have helped set the groundwork for the broad antitrust consensus that still reigns today. I personally thank you and congratulate you for the contributions to our field of law and economics.
I am pleased to talk with you today about this antitrust consensus, and what it means for the digital world. The antitrust consensus, in my view, has two key components.
The first component is the consumer welfare standard, which condemns practices as unlawful where they harm competition in such a way that consumers suffer. The consumer welfare standard is the lodestar of antitrust enforcement, and a humble recognition that antitrust law, if misapplied, can have harmful consequences for those it intends to protect.
The second component—which is the focus of my remarks today—is what I and others call “evidence-based enforcement.” An evidence-based 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.
Outside the realm of naked horizontal restraints such as price fixing, bid rigging, and market allocation, antitrust demands evidence of harm or likely harm to competition, often weighed against efficiencies or procompetitive justifications. As enforcers, we exercise our prosecutorial role every day using this approach to review economic evidence to determine whether a restraint, on its face, violates the antitrust laws.
Taking an evidence-based approach to antitrust law should not be mistaken for an unwillingness to bring enforcement actions. In fact, where there is clear evidence of harm to competition, it is the duty of enforcers to promptly and vigorously prosecute the antitrust laws, and to refuse to settle for ineffective behavioral band-aids that fall short of curing the underlying threat to competition and consumers. Believe me, it is a lot easier to accept some pretty promise, issue a press release congratulating ourselves, and move on to the next matter. Convenience, however, is not in the oath we take to uphold the U.S. antitrust laws.
Evidence-based enforcement also requires a readiness to adapt our existing antitrust framework and tools to new or emerging threats to competition. It is an approach of openness to persuasion, and is adaptable to new business models and emerging technologies that create novel threats to competition and consumer welfare.
This brings me to the two topics of this conference: digital platforms and concentration.
I am aware that some in this audience believe that existing antitrust laws are ill-equipped to address competition issues that have arisen in the digital platform economy, and that as a result of the antitrust laws’ supposed lack of adaptability more generally, there have been harmful increases in industry concentration, along with a variety of other social ills.
Some lay the blame for this state of affairs at the feet of the consumer welfare standard itself. Others blame antitrust enforcers for falling short of taking action against companies or transactions that might harm competition. Enforcers do indeed deserve some blame, particularly for their willingness to settle for ineffective behavioral (or regulatory) fixes to mergers rather than challenging them when necessary. Yet, overall, I have reason for optimism.
Based on my collective experiences over the past 25 years, I believe the antitrust consensus approach is flexible to new business models in digital markets, and that there is no persuasive evidence that antitrust itself has “failed.” The antitrust consensus may fall short if enforcement agencies are unwilling to take action, based on credible evidence, or—even worse—if they impose ineffective fixes that transform law enforcement agencies into inefficient regulators, where free markets can do a more effective and more honest job.
Over the past several decades, antitrust law has responded to new and innovative products and markets to protect against novel threats to the competitive process. Enforcement agencies have developed a strong expertise as new types of assets emerge and consumer preferences shift, and have brought successful antitrust challenges to dismantle barriers to competition.
The landmark United States v. Microsoft case is illustrative. The Antitrust Division’s theory of harm in that case may have been cutting edge, yet the framework the en banc D.C. Circuit applied did not mark a seismic shift in how we think about antitrust law. The en banc court of appeals noted that “there is no consensus among commentators on the question of whether, and to what extent, current monopolization doctrine should be amended to account for competition in technologically dynamic markets characterized by network effects.” That was in 2001, and it rings true today. Rather than “amending” antitrust doctrine, the court applied established principles of market definition and monopolization under the Sherman Act to a product for which there was little agency or judicial experience.
The Microsoft case came at a pivotal moment in the expansion of technology and digital networks. We were on the brink of the Internet revolution with the launch of an efficient web browser developed by Marc Andreessen and his startup, Netscape. The Microsoft case proved that an evidence-based antitrust enforcement approach can be flexible in its application to new types of assets and markets—in that case, the computer code and software markets.
As enforcers, each day we confront similar challenges in adapting our antitrust tools to new business models. With each new challenging market or transaction, novel but tested economic modeling is necessary. We rely heavily on the work of economists deploying the most sophisticated economic techniques.
Yet, some ask, how can we stay at the forefront of technological change so that we can protect competition before it is too late?
A basic concept in economics that is helpful in the new area of high-tech platforms is the theory of revealed preference. It is a seemingly straightforward concept: individuals’ choices, including what they decide to purchase or forego, indicate their preferences. We can infer the relative values consumers place on alternatives by how they behave—that is, based on what they purchase, what they sell, and how they make trade-offs.
The late Nobel Laureate economist Paul Samuelson, who was principally responsible for revealed preference theory, described it as follows: “By comparing the costs of different combinations of goods at different relative price situations, we can infer whether a given batch of goods is preferred to another batch; the individual guinea-pig, by his market behavior, reveals his preference pattern—if there is such a pattern.”
This concept is particularly helpful in understanding dynamic or emerging markets, including digital markets. That is, economic value is demonstrated by consumers’ willingness to pay for goods and services.
With recent congressional attention and rising consumer awareness, we may be witnessing a revealed preference for personal data privacy in the digital economy. Ten years ago, few of us may have cared whether companies had access to our data. We would give up our private data freely in exchange for access to a “free” platform—to connect to friends, or to strangers. But the public now recognizes that, for many such platforms, advertisers on the other side of the platform pay handsomely for that access.
Indeed, Franklin Foer characterizes data as the “new oil.” He writes: “It’s the record of our actions: what we read, what we watch, where we travel over the course of a day, what we purchase, our correspondence, our search inquiries, the thoughts we begin to type and then delete.” Eric Schmidt, in 2010 when he was CEO of Google, put a finer point on it, saying “We know who you are. We know where you’ve been. We can more or less know what you’re thinking about.” Of course, a year earlier, Schmidt offered that “if you have something that you don’t want anyone to know, maybe you shouldn’t be doing it in the first place.”
With a greater understanding of the nature and scope of data collection and sharing, today, the mentality regarding free platforms may well be changing. I know that for me, it has changed.
How many of you would pay $1 a year so that your social media data would never be shared with a third party? $12 a year? $100 a year?
Your answer today is likely different than what it was in 2008, and what it may be in 2028. And not everyone’s answer is the same: like any good or service, there is a trade-off between what you give up in exchange for what you obtain. That trade-off may be different across the population. Some may be willing to give up access to their data or privacy for free, whereas others may demand some valuable service—like access to a technological platform—before they do so. Still others may view the benefits of access to a digital platform as inadequate compensation for the loss of control over their private information. From an economic standpoint, the key takeaway is that revealed preference for data privacy could signal the emergence of a new market asset.
Indeed, in recent weeks, we have witnessed the attention paid to the role of consumer data aggregated by digital platforms, and how this data might be used or misused by the companies collecting it. Of course, for sophisticated followers of these business models, none of this is surprising. Going back at least ten years, political campaigns have relied on social media data to hone their messaging and voter outreach efforts. Lest we forget, it was not long ago that we were heralding the role of social media, particularly Twitter, as a potent force of political messaging and voter outreach to challenge authoritarian rule throughout much of the Middle East.
As public attention has been drawn to the practice of collecting data, there is a heightened concern about the value of privacy and the value of consumer data. This concern is no longer limited to privacy advocates and policymakers who have sounded the alarm for years, only for their concerns to fall on deaf ears.
We may now be witnessing a fundamental change in the marketplace. As preferences shift—or as users of digital platforms become better informed about the consequences of sharing their private data—we could be entering a world in which a substantial number of consumers view their data as an asset they won’t part with for free. At a basic level, the bargain between platform and user appears to be shifting. At the margins (and with a generational shift), some may opt out of certain networks because they simply don’t see the bargain as working in their favor. Time will tell.
As a result, the overwhelming incentives to design technologies to maximize the collection and use of personal information may be shifting, and with that shift companies are designing technologies that respond to our revealed preferences for privacy. Some call it “privacy by design,” but in essence it is the simple idea that companies can respond to changing demands by designing and marketing technologies with privacy as the default setting.
There is some burgeoning evidence to support the idea of an emerging market for data privacy. Users of online platforms now are exercising greater control of their data through modifications to their security and privacy settings. News reports also suggest that many users are tuning out social media networks that they believe do not provide adequate protection of their personal data. And new services are emerging that respond to these changing preferences.
Indeed, Facebook COO Sheryl Sandberg has floated the idea that customers would need to pay a premium price for total privacy. And the popularity of Snapchat, which posed some competitive threat to Facebook, was largely based on its so-called “vanishing” chats. In other markets, we already have examples of consumers making choices that favor privacy. Music streaming services like Pandora and Spotify, for example, have basic subscriptions with ads and premium subscriptions without ads. Obviously, the incentive to gather data of users on ad-free premium subscriptions is not the same as it is to gather data of users who receive the service free with ads.
What role does antitrust law play in all of this? In the current environment, there is an arms race among some enforcers to use the strongest remedies available to combat the perceived problem, reminding me of the H.L. Mencken quote that “there is always a well-known solution to every human problem” that is “neat, plausible, and wrong.”
For example, some argue that U.S. or international agencies should simply declare that data is the new digital currency, that online platforms have been exploiting data without consent, that loss of informational control is anti-competitive, and then impose eye-popping penalties by multiplying some measure of data value by the size of the customer base. I wouldn’t call that an antitrust remedy stemming from an evidence-based analysis. Such an approach ignores the economic nuance of revealed preference—that is, not every customer values their data, or their privacy, the same way. Money has face value, but privacy cannot yet be measured in nominal terms, and varies according to the type and utility of the data used.
A more effective approach grounded in evidence would be to ensure that there are sufficient incentives for new competitive entry in markets for platforms with network effects, while ensuring that customers who prefer established business models still have access to the product or service they want.
Many experts have testified that imposing rigid regulations or a new (and likely irreversible) statutory scheme may not be a necessary or viable solution for fostering competition. Like other government entities, competition authorities and consumer protection agencies are at risk of regulatory capture. There is a long history of incumbent companies manipulating regulatory schemes to create entry barriers. A monopolist may be willing to incur a cost for maintaining its monopoly, if the same cost is applied to new competitors. In that event, consumers don’t win.
That is not to say that antitrust law and economics should play no role. Quite the opposite. When any new asset class emerges, companies will try to gain monopoly power over that asset. This can cut two different ways.
On one hand, in certain platform markets involving network effects, there may be barriers to entry or a tendency toward a single firm emerging as the sole winner. Antitrust enforcers may need to take a close look to see whether competition is suffering and consumers are losing out on new innovations as a result of misdeeds by a monopoly incumbent.
On the other hand, of course, antitrust economists understand that monopolies often can be fragile, and they give rise to new entrants. Indeed, it is the prospect of gaining lawful monopoly power that provides incentives to innovate and create products consumers want. We seek to encourage competition for the market—that is, dynamic competition to innovate and provide new products or services that customers demand. In reaction to controversies over how companies use or misuse personal data, we may see the emergence of new competitors in the digital space that attract customers based on promises of user privacy. Neither antitrust agencies nor incumbent monopolists should be allowed to kill off this type of competition. Antitrust enforcement should ensure that the markets allow for new, more efficient, more innovative competitors to enter.
Whether antitrust enforcement makes sense in this area is therefore an evidence-based question: are consumers who believe that their data is digital currency facing a monopoly seller or monopsony buyer with structural barriers to competition and entry?
That is the question enforcers should answer. Admittedly, it is a complex question and we need studies and new and innovative thinking—exactly the kind of thinking that occurs and is fostered at institutions such as the University of Chicago.
The antitrust consensus approach is flexible enough to accommodate new and evolving economic wisdom, so long as it is grounded in empirical evidence. If there is an evidentiary basis for an enforcement action against any platform, and if that enforcement action is built on a sound economic foundation, antitrust agencies should fulfill their duties to the American consumer.
Even though economic analysis can adjust to evolving business models, there are still criticisms that existing antitrust tools or doctrines are ill-equipped to handle more nuanced threats to competition that may exist in digital platform markets. Some scholars have taken aim at the predatory pricing doctrine in particular, which in their view condones below-cost pricing models that sweep away competitors and raise the prospect of future price increases. A number of commentators have endorsed this critique.
It is worth recounting exactly why predatory pricing claims, in particular, face such a high bar. The Supreme Court in Brooke Group Ltd. v. Brown & Williamson Tobacco Corporation recognized the serious danger of condemning business practices that are pro-consumer. Indeed, a predatory pricing claim is the most dangerous tool in an enforcer’s or litigant’s toolbox because it seemingly would condemn the type of action that is at the heart of competition: charging a lower price than one’s competitor. As the Court explained, “it would be ironic indeed if the standards for predatory pricing liability were so low that antitrust suits themselves became a tool for keeping prices high.” The Court therefore required a showing of a “dangerous probability of recoupment,” as well as pricing “below an appropriate measure of its rival’s costs.”
Some view this test as exceedingly difficult to meet, but that is precisely the point of the doctrine. The fact that predatory pricing claims frequently fall short does not mean the doctrine is inadequate. It means that the doctrine is doing its work.
At the same time, I believe that, as enforcers, we should be open and receptive to empirical evidence that companies in digital markets may be engaging in predatory pricing or other exclusionary conduct to drive out competition and cause long-run harm to consumers. I am encouraged by the fact that theoretical work has recently emerged on this front. For example, as a Yale Law student, Lina Khan wrote an interesting student note in The Yale Law Journal that has drawn significant attention to how predatory pricing or other exclusionary conduct claims might apply to digital platforms. Her note surely has been subject to some criticism, but we should encourage fresh thinking on how our legal tools apply to new digital platforms. We need more thinking—diverse thinking—about these questions. And, we need a civil discourse on this topic.
Contrary to what some critics suggest, however, I believe that the antitrust consensus need not be discarded in order for a predatory pricing claim to be successful. What we need instead is a sound body of evidence that dynamic competition will be harmed, such that consumers will suffer in the long-run by the practices we are concerned about.
In the context of digital platforms, an evidence-based approach is critical to protecting innovation. Disruptive competition and new business models are healthy, and should be encouraged. It is rational, pro-competitive, and pro-consumer for new entrants to charge low prices. Predation that harms dynamic competition, however, should not be condoned.
Another approach to exclusionary conduct in digital markets that might be worthy of consideration is an “economic sense” test within the Brooke Group framework. That is, there may be an antitrust violation where a business practice makes no economic sense other than to harm competition and subsequently recoup short-term losses. To channel Sherlock Holmes, if you rule out all pro-competitive justifications through a careful empirical analysis, all that’s left may be anticompetitive action.
Enforcers should be receptive to fresh thinking in this area, and I encourage experts in academia, private practice, and the public policy sphere to rise to the occasion and enlighten us. But rather than suggesting new laws or resurrecting tools we have discarded after decades of experience, it would be particularly helpful for the bar and the academy to ask how the current analytical tools can address the emerging challenges of digital platforms based on existing jurisprudence. Just as the Department of Justice, the EU, and the D.C. Circuit addressed the then-emerging issue in the Microsoft case, the antitrust laws are flexible enough to address competition issues from emerging platforms. As enforcers who bear the burden of proof, we must demand hard evidence of anticompetitive harm before rushing to condemn as unlawful business practices that fuel innovation and thus dynamic competition. An agency or court considering a Section 2 claim under the Sherman Act must be careful not to kill the golden goose of innovation.
An evidence-based approach to antitrust law also should be self-reflective, in order to assess the real-world results of the antitrust consensus. In this regard, critics of the consumer welfare standard have their hearts in the right place. Yet critics of the current regime may be forgetting to subject their preferred approach to a hard, evidence-based evaluation of its own. I refer specifically to claims that industry concentration has dramatically increased, with the antitrust consensus as the primary culprit.
Many of these critiques date back to a 2016 “issue brief” issued by the Council of Economic Advisors asserting evidence of “a decline of competition” and “increased industry concentration.” Academics, commentators, and journalists made similar claims in subsequent months. As a side note, my favorite has to be the John Oliver segment on concentration that aired on the HBO network this past fall. Most of the concentration claims rely on data from the U.S. Census Bureau, yet those data are overly broad aggregations of economic activity. The concentration of an industry is not the same thing as the concentration of an antitrust market. No available data demonstrate increasing concentration of markets, as economists and antitrust enforcers define them.
My point is consistent with a similar point made by Professor Carl Shapiro, former Deputy Assistant Attorney General of the Antitrust Division and a member of the CEA in the Obama Administration a few years before the issue brief was released. In a forthcoming article, he explains why the evidence on concentration cited by the CEA is “not informative regarding the state of competition.” A further explanation of why that is so has just been posted on the Social Science Research Network by Antitrust Division economist Greg Werden and Deputy Assistant Attorney General Luke Froeb.
The shortcomings of the Census data as a measure of concentration have been well established for decades. Nearly thirty years ago, Antitrust Division economists compared the breadth of the Census “industries” to the relevant markets defined in the Division’s merger complaints, finding that the Census industries were much broader—often more than a hundred times broader. Today, the Census data differs slightly from the earlier dataset, but it still suffers the same shortcomings as any meaningful metric of “concentration.”
If one compares the commerce of relevant markets to the commerce in corresponding industries, one can readily recognize the discrepancy. For 32 of the 44 relevant markets examined for mergers filed between 2013 and 2015, the least aggregated Census industry measure was more than a hundred times broader than the relevant antitrust market. The excessive aggregation of Census data makes them uninformative on trends in market concentration. Among other things, they treat all mergers within a broad industry the same for purposes of “concentration,” even if a merger does not increase share in any relevant antitrust market.
For example, in a recent enforcement action, the United States brought suit to enjoin a merger of two movie theater chains operating in the antitrust markets defined as “the exhibition of first-run, commercial movies” in Voorhees and Somerdale, New Jersey; East Louisville, Kentucky; Western Forth Worth, Texas; and Greater Denton, Texas. The effects of the proposed merger, and hence the four alleged relevant markets, were local. The United States entered into a structural remedy requiring divestures that prevented all but a tiny increase in concentration for one of the four markets. Obviously, the market for movie theaters in Western Fort Worth, Texas is not remotely the same as the national movie theater industry. And yet the narrowest category in the Census data is “motion picture theaters” in the United States. The merger did increase concentration within the relevant Census industry, but for antitrust purposes the level of national concentration is meaningless in movie exhibition, just as it is in much of the economy.
To be sure, any casual observer or consumer will notice that market concentration is increasing in certain sectors, and the CEA report has a level of basic intuitive appeal. Yet it takes a tremendous leap of logic to conclude that antitrust law enforcement is the cause of these particular trends. Competition enforcement can do little to alter the natural evolution of industry structure in which some firms thrive and grow while others languish or fail. Indeed, market concentration can increase when antitrust enforcement is working as it should. Rather than a failure of antitrust, concentration may be the byproduct of healthy competition, as the most innovative and efficient firms grow and attract customers.
Antitrust enforcement requires the antitrust consensus to be subject to constant re-evaluation and scrutiny to maintain its foothold. But a “guilt by correlation” approach implying a causal connection between enforcement standards and specious concentration measures does not advance the conversation. Enforcers need to be more vigilant, but they also need to be humble.
More importantly, critics of purportedly increasing concentration have not made an affirmative, evidence-based case that revising existing antitrust standards would be an effective remedy. Nor have they demonstrated that a concentration-focused overhaul of the antitrust laws would benefit competition and consumers. Tempting as it is to hope for a lower standard of proof for proving antitrust violations, existing evidentiary standards serve a purpose—they ensure that impartial courts serve as a check on government power.
A related criticism some assert against the antitrust consensus is that that it has neglected, or even given rise to, rising economic inequality, stagnant wages, unemployment, and a concentration in political power in the hands of private market actors. Some of these critics go on to argue that antitrust should be recalibrated specifically to address these issues. As a matter of causation, I am skeptical that competition policy can be isolated as a contributing cause to any of these wider trends.
It has long been observed that the statutory framework that the Antitrust Division and FTC enforce already provides enormous latitude. The Sherman Act deems illegal “[e]very contract, combination . . . or conspiracy, in restraint of trade or commerce,” and the Clayton Act condemns transactions whose “effect . . . may be substantially to lessen competition.” The judiciary disciplines enforcement agencies—and antitrust law more generally—by demanding proof to establish a violation. This evidence-based standard ensures that free market competition and economic liberty will not fall victim to the whims of Washington bureaucrats and changing populist rhetoric.
I view the narrow focus on competition and consumers as a feature, not a bug. The burden of curing our economic woes rests on the shoulders of policymakers and agencies with the institutional capacity and statutory mandate to tackle other complex issues. Shifting this responsibility to antitrust enforcers would require us to make trade-offs between competition and non-competition goals on a case-by-case basis. I view this as dangerous. It would threaten to disrupt the bipartisan economic consensus that has emerged by making antitrust a political tool that changes significantly depending on the party in power.
Of course, enforcers carrying out the antitrust consensus approach are not free from blame. In fact, looking back over the past hundred years, American antitrust agencies likely have made more enforcement mistakes than any of their foreign counterparts. Yet we still carry the mantle of leadership as the global economy encounters more challenges from digitization. That is because our evidence-based approach to antitrust means learning from failure and continuing to take action where harm to competition arises.
Despite our efforts to learn from the past, we have recently failed on repeated occasions to address the harm imposed by some anticompetitive mergers. I refer to the imposition of so-called behavioral remedies on horizontal or vertical transactions that, if cleared outright, would substantially lessen competition. At best, the enforcement intuition behind imposing behavioral consent decrees is to strike a compromise that might preserve the efficiencies from a merger while hoping that the merged company will not act on its natural economic incentives to harm competition. At worst, enforcers may be attracted to behavioral remedies as a convenient way of saving face when clearing politically unpopular mergers. Or, worse yet, as an opportunity to advance goals with only an attenuated connection to competition.
Both of these instincts are misguided, and they fly in the face of an evidence-based approach to enforcement. Behavioral remedies for anticompetitive mergers often are prone to failure because they seek to cure on paper the enormous financial incentives that motivate profit-maximizing businesses. The empirical support on behavioral remedies demonstrates that they are dubious because, too often, they fail to constrain the ability of merged companies to act on those incentives, and they overestimate the willingness of aggrieved parties to come forward and risk the consequences of an enforcement action. Do harmed companies want to challenge their necessary future business partners, whose ongoing business relationships they need, in an adversarial proceeding, such as arbitration or a decree enforcement action?
One institutional change we are implementing to try to address enforcement of DOJ Antitrust decrees, both past and future, is the creation of the Office of Decree Enforcement, with the sole goal to ensure compliance with, and enforcement of, these decrees. We will provide more details in coming weeks.
American consumers deserve better than unenforceable fixes. The stakes of digital markets are enormous. That is why we are all here today. That is why Congress held hearings earlier this month in Washington, D.C. And why the Senate held them almost exactly 20 years ago in the Microsoft hearings. If there is clear evidence of harm to competition in digital platforms, enforcers must take vigorous action and seek remedies that protect American consumers, so that free markets or consumers don’t instead bear the risk of failure.