Cuando los procesos internos de una plataforma de comercio electrónico fallan, o cuando simplemente sus directivos o empleados toman decisiones equivocadas —digamos, enviando un pedido a una dirección incorrecta— un consumidor o un grupo de consumidores se ven perjudicados. Estos errores son fácilmente subsanables si la plataforma en cuestión tiene un buen proceso de atención al cliente.

Read the full piece here.

Executive Summary

A prominent theme in debates about US national privacy legislation is whether federal law should preempt state law. A federal statute could create one standard for markets that are obviously national in scope. Another approach is to allow states to be “laboratories of democracy” that adopt different laws so they can discover the best ones.

We propose a federal statute requiring states to recognize contractual choice-of-law provisions, so companies and consumers can choose what state privacy law to adopt. Privacy would continue to be regulated at the state level. However, the federal government would provide for jurisdictional competition among states, such that companies operating nationally could comply with the privacy laws of any one state.

Our proposed approach would foster a double competition aimed at discerning and delivering on consumers’ true privacy interests: market competition to deliver privacy policies that consumers prefer and competition among states to develop the best privacy laws.

Unlike a single federal privacy law, this approach would provide 50 competing privacy regimes for national firms. The choice-of-law approach can trigger competition and innovation in privacy practices while preserving a role for meaningful state privacy regulation.

Introduction

The question of preemption of state law by the federal government has bedeviled debates about privacy regulation in the United States. A prominent theme is to propose a national privacy policy that largely preempts state policies to create one standard for markets that are obviously national. Another approach is to allow states to be “laboratories of democracy” that adopt different laws, with the hope that they will adopt the best rules over time. Both approaches have substantial costs and weaknesses.

The alternative approach we propose would foster a double competition aimed at discerning and delivering on consumers’ true privacy interests: market competition to deliver privacy policies that consumers prefer and competition among states to develop the best privacy laws. Indeed, our proposal aims to obtain the best features—and avoid the worst features—of both a federal regime and a multistate privacy law regime by allowing firms and consumers to agree on compliance with the single regime of their choosing.

Thus, we propose a federal statute requiring states to recognize contractual choice-of-law provisions, so companies and consumers can choose what state privacy law to adopt. Privacy would continue to be regulated at the state level. However, the federal government would provide for jurisdictional competition among states, and companies operating nationally could comply with the privacy laws of any one state.

Unlike a single federal privacy law, this approach would provide 50 competing privacy regimes for national firms. Protecting choice of law can trigger competition and innovation in privacy practices while preserving a role for meaningful state privacy regulation.

The Emerging Patchwork of State Privacy Statutes Is a Problem for National Businesses

A strong impetus for federal privacy legislation is the opportunity national and multinational businesses see to alleviate the expense and liability of having a patchwork of privacy statutes with which they must comply in the United States. Absent preemptive legislation, they could conceivably operate under 50 different state regimes, which would increase costs and balkanize their services and policies without coordinate gains for consumers. Along with whether a federal statute should have a private cause of action, preempting state law is a top issue when policymakers roll up their sleeves and discuss federal privacy legislation.

But while the patchwork argument is real, it may be overstated. There are unlikely ever to be 50 distinct state regimes; rather, a small number of state legislation types is likely, as jurisdictions follow each other’s leads and group together, including by promulgating model state statutes.[1] States don’t follow the worst examples from their brethren, as the lack of biometric statutes modeled on Illinois’s legislation illustrates.[2]

Along with fewer “patches,” the patchwork’s costs will tend to diminish over time as states land on relatively stable policies, allowing compliance to be somewhat routinized.

Nonetheless, the patchwork is far from ideal. It is costly to firms doing business nationally. It costs small firms more per unit of revenue, raising the bar to new entry and competition. And it may confuse consumers about what their protections are (though consumers don’t generally assess privacy policies carefully anyway).

But a Federal Privacy Statute Is Far from Ideal as Well

Federal preemption has many weaknesses and costs as well. Foremost, it may not deliver meaningful privacy to consumers. This is partially because “privacy” is a congeries of interests and values that defy capture.[3] Different people prioritize different privacy issues differently. In particular, the elites driving and influencing legislation may prioritize certain privacy values differently from consumers, so legislation may not serve most consumers’ actual interests.[4]

Those in the privacy-regulation community sometimes assume that passing privacy legislation ipso facto protects privacy, but that is not a foregone conclusion. The privacy regulations issued under the Gramm-Leach-Bliley Act (concerning financial services)[5] and the Health Insurance Portability and Accountability Act (concerning health care)[6] did not usher in eras of consumer confidence about privacy in their respective fields.

The short-term benefits of preempting state law may come with greater long-term costs. One cost is the likely drop in competition among firms around privacy. Today, as some have noted, “Privacy is actually a commercial advantage. . . . It can be a competitive advantage for you and build trust for your users.”[7] But federal privacy regulation seems almost certain to induce firms to treat compliance as the full measure of privacy to offer consumers. Efforts to outperform or ace out one another will likely diminish.[8]

Another long-term cost of preempting state law is the drop in competition among states to provide well-tuned privacy and consumer-protection legislation. Our federal system’s practical genius, which Justice Louis Brandeis articulated 90 years ago in New State Ice v. Liebmann, is that state variation allows natural experiments in what best serves society—business and consumer interests alike.[9] Because variations are allowed, states can amend their laws individually, learn from one another, adapt, and converge on good policy.

The economic theory of federalism draws heavily from the Tiebout model.[10] Charles Tiebout argued that competing local governments could, under certain conditions, produce public goods more efficiently than the national government could. Local governments act as firms in a marketplace for taxes and public goods, and consumer-citizens match their preferences to the providers. Efficient allocation requires mobile people and resources, enough jurisdictions with the freedom to set their own laws, and limited spillovers among jurisdictions (effects of one jurisdiction’s policies on others).

A related body of literature on “market-preserving federalism” argues that strong and self-reinforcing limits on national and local power can preserve markets and incentivize economic growth and development.[11] The upshot of this literature is that when local jurisdictions can compete on law, not only do they better match citizens’ policy preferences, but the rules tend toward greater economic efficiency.

In contrast to the economic gains from decentralization, moving authority over privacy from states to the federal government may have large political costs. It may deepen Americans’ growing dissatisfaction with their democracy. Experience belies the ideal of responsive national government when consumers, acting as citizens, want to learn about or influence the legislation and regulation that governs more and more areas of their lives. The “rejectionist” strain in American politics that Donald Trump’s insurgency and presidency epitomized may illustrate deep dissatisfaction with American democracy that has been growing for decades. Managing a highly personal and cultural

issue like privacy through negotiation between large businesses and anonymous federal regulators would deepen trends that probably undermine the government’s legitimacy.

To put a constitutional point on it, preempting states on privacy contradicts the original design of our system, which assigned limited powers to the federal government.[12] The federal government’s enumerated powers generally consist of national public goods—particularly defense. The interstate commerce clause, inspired by state parochialism under the Articles of Confederation, exists to make commerce among states (and with tribes) regular; it is not rightly a font of power to regulate the terms and conditions of commerce generally.[13]

Preempting state law does not necessarily lead to regulatory certainty, as is often imagined. Section 230 of the Communications Decency Act may defeat once and for all the idea that federal legislation creates certainty.[14] More than a quarter century after its passage, it is hotly debated in Congress and threatened in the courts.[15]

The Fair Credit Reporting Act (FCRA) provides a similar example.[16] Passed in 1970, it comprehensively regulated credit reporting. Since then, Congress has amended it dozens of times, and regulators have made countless alterations through interpretation and enforcement.[17] The Consumer Financial Protection Bureau recently announced a new inquiry into data brokering under the FCRA.[18] That is fine, but it illustrates that the FCRA did not solve problems and stabilize the law. It just moved the jurisdiction to Washington, DC.

Meanwhile, as regulatory theory predicts, credit reporting has become a three-horse race.[19] A few slow-to-innovate firms have captured and maintained dominance thanks partially to the costs and barriers to entry that uniform regulation creates.

Legal certainty may be a chimera while business practices and social values are in flux. Certainty develops over time as industries settle into familiar behaviors and roles.

An Alternative to Preemption: Business and Consumer Choice

One way to deal with this highly complex issue is to promote competition for laws. The late, great Larry Ribstein, with several coauthors over the years, proposed one such legal mechanism: a law market empowered by choice-of-law statutes.[20] Drawing on the notion of market competition as a discovery process,[21] Ribstein and Henry Butler explained:

In order to solve the knowledge problem and to create efficient legal technologies, the legal system can use the same competitive process that encourages innovation in the private sector—that is, competition among suppliers of law. As we will see, this entails enforcing contracts among the parties regarding the applicable law. The greater the knowledge problem the more necessary it is to unleash markets for law to solve the problem.[22]

The proposal set forth below promotes just such competition and solves the privacy-law patchwork problem without the costs of federal preemption. It does this through a simple procedural regulation requiring states to enforce choice-of-law terms in privacy contracts, rather than through a heavy-handed, substantive federal law. Inspired by Butler and Ribstein’s proposal for pluralist insurance regulation,[23] the idea is to make the choice of legal regime a locus of privacy competition.

Modeled on the US system of state incorporation law, our proposed legislation would leave firms generally free to select the state privacy law under which they do business nationally. Firms would inform consumers, as they must to form a contract, that a given state’s laws govern their policies. Federal law would ensure that states respect those choice-of-law provisions, which would be enforced like any other contract term.

This would strengthen and deepen competition around privacy. If firms believed privacy was a consumer interest, they could select highly protective state laws and advertise that choice, currying consumer favor. If their competitors chose relatively lax state law, they could advertise to the public the privacy threats behind that choice. The process would help hunt out consumers’ true interests through an ongoing argument before consumers. Businesses’ and consumers’ ongoing choices— rather than a single choice by Congress followed by blunt, episodic amendments—would shape the privacy landscape.

The way consumers choose in the modern marketplace is a broad and important topic that deserves further study and elucidation. It nevertheless seems clear—and it is rather pat to observe—that consumers do not carefully read privacy policies and balance their implications. Rather, a hive mind of actors including competitors, advocates, journalists, regulators, and politicians pore over company policies and practices. Consumers take in branding and advertising, reputation, news, personal recommendations, rumors, and trends to decide on the services they use and how they use them.

That detail should not be overlooked: Consumers may use services differently based on the trust they place in them to protect privacy and related values. Using an information-intensive service is not a proposition to share everything or nothing. Consumers can and do shade their use and withhold information from platforms and services depending on their perceptions of whether the privacy protections offered meet their needs.

There is reason to be dissatisfied with the modern marketplace, in which terms of service and privacy policies are offered to the individual consumer on a “take it or leave it” basis. There is a different kind of negotiation, described above, between the hive mind and large businesses. But when the hive mind and business have settled on terms, individuals cannot negotiate bespoke policies reflecting their particular wants and needs. This collective decision-making may be why some advocates regard market processes as coercive. They do not offer custom choices to all but force individual consumers into channels cut by all.

The solution that orthodox privacy advocates offer does not respond well to this problem, because they would replace “take it or leave it” policies crafted in the crucible of the marketplace with “take it or leave it” policies crafted in a political and regulatory crucible. Their prescriptions are sometimes to require artificial notice and “choice,” such as whether to accept cookies when one visits websites. This, as experience shows, does not reach consumers when they are interested in choosing.

Choice of law in privacy competition is meant to preserve manifold choices when and where consumers make their choices, such as at the decision to transact, and then let consumers choose how they use the services they have decided to adopt. Let new entrants choose variegated privacy-law regimes, and consumers will choose among them. That does not fix the whole problem, but at least it doesn’t replace consumer choice with an “expert” one-size-fits-all choice.

In parallel to business competition around privacy choice of law, states would compete with one another to provide the most felicitous environment for consumers and businesses. Some states would choose more protection, seeking the rules businesses would choose to please privacy-conscious consumers. Others might choose less protection, betting that consumers prefer goods other than information control, such as free, convenient, highly interactive, and custom services.

Importantly, this mechanism would allow companies to opt in to various privacy regimes based on the type of service they offer, enabling a degree of fine-tuning appropriate for different industries and different activities that no alternative would likely offer. This would not only result in the experimentation and competition of federalism but also enable multiple overlapping privacy-regulation regimes, avoiding the “one-size-doesn’t-fit-all” problem.

While experimentation continued, state policies would probably rationalize and converge over time. There are institutions dedicated to this, such as the Uniform Law Commission, which is at its best when it harmonizes existing laws based on states’ experience.[24]

It is well within the federal commerce power to regulate state enforcement of choice-of-law provisions, because states may use them to limit interjurisdictional competition. Controlling that is precisely what the commerce power is for. Utah’s recent Social Media Regulation Act[25] barred enforcement of choice-of-law provisions, an effort to regulate nationally from a state capital. Federally backing contractual choice-of-law selections would curtail this growing problem.

At the same time, what our proposed protections for choice-of-law rules do is not much different from what contracts already routinely do and courts enforce in many industries. Contracting parties often specify the governing state’s law and negotiate for the law that best suits their collective needs.

Indeed, sophisticated business contracts increasingly include choice-of-law clauses that state the law that the parties wish to govern their relationship. In addition to settling uncertainty, these clauses might enable the contracting parties to circumvent those states’ laws they deem to be undesirable.[26]

This practice is not only business-to-business. Consumers regularly enter into contracts that include choice-of-law clauses—including regarding privacy law. Credit card agreements, stock and mutual fund investment terms, consumer-product warranties, and insurance contracts, among many other legal agreements, routinely specify the relevant state law that will govern.

In these situations, the insurance company, manufacturer, or mutual fund has effectively chosen the law. The consumer participates in this choice only to the same extent that she participates in any choices related to mass-produced products and services, that is, by deciding whether to buy the product or service.[27]

Allowing contracting parties to create their own legal certainty by contract would likely rankle states. Indeed, “we might expect governments to respond with hostility to the enforcement of choice-of-law clauses. In fact, however, the courts usually do enforce choice-of-law clauses.”[28] With some states trying to regulate nationally and some effectively doing so, the choice the states collectively face is having a role in privacy regulation or no role at all. Competition is better for them than exclusion from the field or minimization of their role through federal preemption of state privacy law. This proposal thus advocates simple federal legislation that preserves firms’ ability to make binding choice-of-law decisions and states’ ability to retain a say in the country’s privacy-governance regime.

Avoiding a Race to the Bottom

Some privacy advocates may object that state laws will not sufficiently protect consumers.[29] Indeed, there is literature arguing that federalism will produce a race to the bottom (i.e., competition leading every state to effectively adopt the weakest law possible), for example, when states offer incorporation laws that are the least burdensome to business interests in a way that arguably diverges from public or consumer interests.[30]

The race-to-the-bottom framing slants the issues and obscures ever-present trade-offs, however. Rules that give consumers high levels of privacy come at a cost in social interaction, price, and the quality of the goods they buy and services they receive. It is not inherently “down” or bad to prefer cheap or free goods and plentiful, social, commercial interaction. It is not inherently “up” or good to opt for greater privacy.

The question is what consumers want. The answers to that question—yes, plural—are the subject of constant research through market mechanisms when markets are free to experiment and are functioning well. Consumers’ demands can change over time through various mechanisms, including experience with new technologies and business models. We argue for privacy on the terms consumers want. The goal is maximizing consumer welfare, which sometimes means privacy and sometimes means sharing personal information in the interest of other goods. There is no race to the bottom in trading one good for another.

Yet the notion of a race to the bottom persists—although not without controversy. In the case of Delaware’s incorporation statutes, the issue is highly contested. Many scholars argue that the state’s rules are the most efficient—that “far from exploiting shareholders, . . . these rules actually benefit shareholders by increasing the wealth of corporations chartered in states with these rules.”[31]

As always, there are trade-offs, and the race-to-the-bottom hypothesis requires some unlikely assumptions. Principally, as Jonathan Macey and Geoffrey Miller discuss, the assumption that state legislators are beholden to the interests of corporations over other constituencies vying for influence. As Macey and Miller explain, the presence of a powerful lobby of specialized and well-positioned corporate lawyers (whose interests are not the same as those of corporate managers) transforms the analysis and explains the persistence and quality of Delaware corporate law.[32]

In much the same vein, there are several reasons to think competition for privacy rules would not succumb to a race to the bottom.

First, if privacy advocates are correct, consumers put substantial pressure on companies to adopt stricter privacy policies. Simply opting in to the weakest state regime would not, as with corporate law, be a matter of substantial indifference to consumers but would (according to advocates) run contrary to their interests. If advocates are correct, firms avoiding stronger privacy laws would pay substantial costs. As a result, the impetus for states to offer weaker laws would be diminished. And, consistent with Macey and Miller’s “interest-group theory” of corporate law,[33] advocates themselves would be important constituencies vying to influence state privacy laws. Satisfying these advocates may benefit state legislators more than satisfying corporate constituencies does.

Second, “weaker” and “stronger” would not be the only dimensions on which states would compete for firms to adopt their privacy regimes. Rather, as mentioned above, privacy law is not one-size-fits-all. Different industries and services entail different implications for consumer interests. States could compete to specialize in offering privacy regimes attractive to distinct industries based on interest groups with particular importance to their economies. Minnesota (home of the Mayo Clinic) and Ohio (home of the Cleveland Clinic), for example, may specialize in health care and medical privacy, while California specializes in social media privacy.

Third, insurance companies are unlikely to be indifferent to the law that the companies they cover choose. Indeed, to the extent that insurers require covered firms to adopt specific privacy practices to control risk, those insurers would likely relish the prospect of outsourcing the oversight of these activities to state law enforcers. States could thus compete to mimic large insurers’ privacy preferences—which would by no means map onto “weaker” policies—to induce insurers to require covered firms to adopt their laws.

If a race to the bottom is truly a concern, the federal government could offer a 51st privacy alternative (that is, an optional federal regime as an alternative to the states’ various privacy laws). Assuming federal privacy regulation would be stricter (an assumption inherent in the race-to-the-bottom objection to state competition), such an approach would ensure that at least one sufficiently strong opt-in privacy regime would always be available. Among other things, this would preclude firms from claiming that no option offers a privacy regime stronger than those of the states trapped in the (alleged) race to the bottom.

Choice of law exists to a degree in the European Union, a trading bloc commonly regarded as uniformly regulated (and commonly regarded as superior on privacy because of a bias toward privacy over other goods). The General Data Protection Regulation (GDPR) gives EU member states broad authority to derogate from its provisions and create state-level exemptions. Article 23 of the GDPR allows states to exempt themselves from EU-wide law to safeguard nine listed broad governmental and public interests.[34] And Articles 85 through 91 provide for derogations, exemptions, and powers to impose additional requirements relative to the GDPR for a number of “specific data processing situations.”[35]

Finally, Article 56 establishes a “lead supervisory authority” for each business.[36] In the political, negotiated processes under the GDPR, this effectively allows companies to shade their regulatory obligations and enforcement outlook through their choices of location. For the United States’ sharper rule-of-law environment, we argue that the choice of law should be articulate and clear.

Refining the Privacy Choice-of-Law Proposal

The precise contours of a federal statute protecting choice-of-law terms in contracts will determine whether it successfully promotes interfirm and interstate competition. Language will also determine its political salability.

Questions include: What kind of notice, if any, should be required to make consumers aware that they are dealing with a firm under a law regime not their own? Consumers are notoriously unwilling to investigate privacy terms—or any other contract terms—in advance, and when considering the choice of law, they would probably not articulate it to themselves. But the competitive dynamics described earlier would probably communicate relevant information to consumers even without any required notice. As always, competitors will have an incentive to ensure consumers are appropriately well-informed when they can diminish their rivals or elevate themselves in comparison by doing so.[37]

Would there be limits on which state’s laws a firm could choose? For example, could a company choose the law of a state where neither the company nor the consumer is domiciled? States would certainly argue that a company should not be able to opt out of the law of the state where it is domiciled. The federal legislation we propose would allow unlimited choice. Such a choice is important if the true benefits of jurisdictional competition are to be realized.

A federal statute requiring states to enforce choice-of-law terms should not override state law denying enforcement of choice-of-law terms that are oppressive, unfair, or improperly bargained for. In cases such as Carnival Cruise Lines v. Shute[38] and The Bremen v. Zapata Off-Shore Co.,[39] the Supreme Court has considered whether forum-selection clauses in contracts might be invalid. The Court has generally upheld such clauses, but they can be oppressive if they require plaintiffs in Maine to litigate in Hawaii, for example, without a substantial reason why Hawaii courts are the appropriate forum. Choice-of-law terms do not impose the cost of travel to remote locations, but they could be used not to establish the law governing the parties but rather to create a strategic advantage unrelated to the law in litigation. Deception built into a contract’s choice-of-law terms should remain grounds for invalidating the contract under state law, even if the state is precluded from barring choice-of-law terms by statute.

The race-to-the-bottom argument raises the question of whether impeding states from overriding contractual choice-of-law provisions would be harmful to state interests, especially since privacy law concerns consumer rights. However, there are reasons to believe race-to-the-bottom incentives would be tempered by greater legal specialization and certainty and by state courts’ ability to refuse to enforce choice-of-law clauses in certain limited circumstances. As Erin O’Hara and Ribstein put it:

Choice-of law clauses reduce uncertainty about the parties’ legal rights and obligations and enable firms to operate in many places without being subject to multiple states’ laws. These reduced costs may increase the number of profitable transactions and thereby increase social wealth. Also, the clauses may not change the results of many cases because courts in states that prohibit a contract term might apply the more lenient law of a state that has close connections with the parties even without a choice-of-law clause.[40]

Determining when, exactly, a state court can refuse to enforce a firm’s choice of privacy law because of excessive leniency is tricky, but the federal statute could set out a framework for when a court could apply its own state’s law. Much like the independent federal alternative discussed above, specific minimum requirements in the federal law could ensure that any race to the bottom that does occur can go only so far. Of course, it would be essential that any such substantive federal requirements be strictly limited, or else the benefits of jurisdictional competition would be lost.

The converse to the problem of a race to the bottom resulting from state competition is the “California effect”—the prospect of states adopting onerous laws from which no company (or consumer) can opt out. States can regulate nationally through one small tendril of authority: the power to prevent businesses and consumers from agreeing on the law that governs their relationships. If a state regulates in a way that it thinks will be disfavored, it will bar choice-of-law provisions in contracts so consumers and businesses cannot exercise their preference.

Utah’s Social Media Regulation Act, for example, includes mandatory age verification for all social media users,[41] because companies must collect proof that consumers are either of age or not in Utah. To prevent consumers and businesses from avoiding this onerous requirement, Utah bars waivers of the law’s requirements “notwithstanding any contract or choice-of-law provision in a contract.”[42] If parties could choose their law, that would render Utah’s law irrelevant, so Utah cuts off that avenue. This demonstrates the value of a proposal like the one contemplated here.

Proposed Legislation

Creating a federal policy to stop national regulation coming from state capitols, while still preserving competition among states and firms, is unique. Congress usually creates its own policy and preempts states in that area to varying degrees. There is a well-developed law around this type of preemption, which is sometimes implied and sometimes expressed in statute.[43] Our proposal does not operate that way. It merely withdraws state authority to prevent parties from freely contracting about the law that applies to them.

A second minor challenge exists regarding the subject matter about which states may not regulate choice of law. Barring states from regulating choice of law entirely is an option, but if the focus is on privacy only, the preemption must be couched to allow regulation of choice of law in other areas. Thus, the scope of “privacy” must be in the language.

Finally, the withdrawal of state authority should probably be limited to positive enactments, such as statutes and regulations, leaving intact common-law practice related to choice-of-law provisions.[44] “Statute,” “enactment,” and “provision” are preferable in preemptive language to “law,” which is ambiguous.

These challenges, and possibly more, are tentatively addressed in the following first crack at statutory language, inspired by several preemptive federal statutes, including the Employee Retirement Income Security Act of 1974,[45] the Airline Deregulation Act,[46] the Federal Aviation Administration Authorization Act of 1994,[47] and the Federal Railroad Safety Act.[48]

A state, political subdivision of a state, or political authority of at least two states may not enact or enforce any statute, regulation, or other provision barring the adoption or application of any contractual choice-of-law provision to the extent it affects contract terms governing commercial collection, processing, security, or use of personal information.

Conclusion

This report introduces a statutory privacy framework centered on individual states and consistent with the United States’ constitutional design. But it safeguards companies from the challenge created by the intersection of that design and the development of modern commerce and communication, which may require them to navigate the complexities and inefficiencies of serving multiple regulators. It fosters an environment conducive to jurisdictional competition and experimentation.

We believe giving states the chance to compete under this approach should be explored in lieu of consolidating privacy law in the hands of one central federal regulator. Competition among states to provide optimal legislation and among businesses to provide optimal privacy policies will help discover and deliver on consumers’ interests, including privacy, of course, but also interactivity, convenience, low costs, and more.

Consumers’ diverse interests are not known now, and they cannot be predicted reliably for the undoubtedly interesting technological future. Thus, it is important to have a system for discovering consumers’ interests in privacy and the regulatory environments that best help businesses serve consumers. It is unlikely that a federal regulatory regime can do these things. The federal government could offer a 51st option in such a system, of course, so advocates for federal involvement could see their approach tested alongside the states’ approaches.

[1] See Uniform Law Commission, “What Is a Model Act?,” https://www.uniformlaws.org/acts/overview/modelacts.

[2] 740 Ill. Comp. Stat. 14/15 (2008).

[3] See Jim Harper, Privacy and the Four Categories of Information Technology, American Enterprise Institute, May 26, 2020, https://www.aei.org/research-products/report/privacy-and-the-four-categories-of-information-technology.

[4] See Jim Harper, “What Do People Mean by ‘Privacy,’ and How Do They Prioritize Among Privacy Values? Preliminary Results,” American Enterprise Institute, March 18, 2022, https://www.aei.org/research-products/report/what-do-people-mean-by-privacy-and-how-do-they-prioritize-among-privacy-values-preliminary-results.

[5] Gramm-Leach-Bliley Act, 15 U.S.C. 6801, § 501 et seq.

[6] Health Insurance Portability and Accountability Act of 1996, Pub. L. No. 104-191, § 264.

[7] Estelle Masse, quoted in Ashleigh Hollowell, “Is Privacy Only for the Elite? Why Apple’s Approach Is a Marketing Advantage,” VentureBeat, October 18, 2022, https://venturebeat.com/security/is-privacy-only-for-the-elite-why-apples-approach-is-a-marketing-advantage.

[8] Competition among firms regarding privacy is common, particularly in digital markets. Notably, Apple has implemented stronger privacy protections than most of its competitors have, particularly with its App Tracking Transparency framework in 2021. See, for example, Brain X. Chen, “To Be Tracked or Not? Apple Is Now Giving Us the Choice,” New York Times, April 26, 2021, https://www.nytimes.com/2021/04/26/technology/personaltech/apple-app-tracking-transparency.html. For Apple, this approach is built into the design of its products and offers what it considers a competitive advantage: “Because Apple designs both the iPhone and processors that offer heavy-duty processing power at low energy usage, it’s best poised to offer an alternative vision to Android developer Google which has essentially built its business around internet services.” Kif Leswing, “Apple Is Turning Privacy into a Business Advantage, Not Just a Marketing Slogan,” CNBC, June 8, 2021, https://www.cnbc.com/2021/06/07/apple-is-turning-privacy-into-a-business-advantage.html. Apple has built a substantial marketing campaign around these privacy differentiators, including its ubiquitous “Privacy. That’s Apple.” slogan. See Apple, “Privacy,” https://www.apple.com/privacy. Similarly, “Some of the world’s biggest brands (including Unilever, AB InBev, Diageo, Ferrero, Ikea, L’Oréal, Mars, Mastercard, P&G, Shell, Unilever and Visa) are focusing on taking an ethical and privacy-centered approach to data, particularly in the digital marketing and advertising context.” Rachel Dulberg, “Why the World’s Biggest Brands Care About Privacy,” Medium, September 14, 2021, https://uxdesign.cc/who-cares-about-privacy-ed6d832156dd.

[9] New State Ice Co. v. Liebmann, 285 US 262, 311 (1932) (Brandeis, J., dissenting) (“To stay experimentation in things social and economic is a grave responsibility. Denial of the right to experiment may be fraught with serious consequences to the Nation. It is one of the happy incidents of the federal system that a single courageous State may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country.”).

[10] See Charles M. Tiebout, “A Pure Theory of Local Expenditures,” Journal of Political Economy 64, no. 5 (1956): 416–24, https://www.jstor.org/stable/1826343.

[11] See, for example, Barry R. Weingast, “The Economic Role of Political Institutions: Market-Preserving Federalism and Economic Development,” Journal of Law, Economics, & Organization 11, no. 1 (April 1995): 1 31, https://www.jstor.org/stable/765068; Yingyi Qian and Barry R. Weingast, “Federalism as a Commitment to Preserving Market Incentives,” Journal of Economic Perspectives 11, no. 4 (Fall 1997): 83–92, https://www.jstor.org/stable/2138464; and Rui J. P. de Figueiredo Jr. and Barry R. Weingast, “Self-Enforcing Federalism,” Journal of Law, Economics, & Organization 21, no. 1 (April 2005): 103–35, https://www.jstor.org/stable/3554986.

[12] See US Const. art. I, § 8 (enumerating the powers of the federal Congress).

[13] See generally Randy E. Barnett, Restoring the Lost Constitution: The Presumption of Liberty (Princeton, NJ: Princeton University Press, 2014), 274–318.

[14] Protection for Private Blocking and Screening of Offensive Material, 47 U.S.C. 230.

[15] See Geoffrey A. Manne, Ben Sperry, and Kristian Stout, “Who Moderates the Moderators? A Law & Economics Approach to Holding Online Platforms Accountable Without Destroying the Internet,” Rutgers Computer & Technology Law Journal 49, no. 1 (2022): 39–53, https://laweconcenter.org/wp-content/uploads/2021/11/Stout-Article-Final.pdf (detailing some of the history of how Section 230 immunity expanded and differs from First Amendment protections); Meghan Anand et al., “All the Ways Congress Wants to Change Section 230,” Slate, August 30, 2023, https://slate.com/technology/2021/03/section-230 reform-legislative-tracker.html (tracking every proposal to amend or repeal Section 230); and Technology & Marketing Law Blog, website, https://blog.ericgoldman.org (tracking all Section 230 cases with commentary).

[16] Fair Credit Reporting Act, 15 U.S.C. § 1681 et seq.

[17] See US Federal Trade Commission, Fair Credit Reporting Act: 15 U.S.C. § 1681, May 2023, https://www.ftc.gov/system/files/ftc_gov/pdf/fcra-may2023-508.pdf (detailing changes to the Fair Credit Reporting Act and its regulations over time).

[18] US Federal Reserve System, Consumer Financial Protection Bureau, “CFPB Launches Inquiry into the Business Practices of Data Brokers,” press release, May 15, 2023, https://www.consumerfinance.gov/about-us/newsroom/cfpb-launches-inquiry-into-the-business-practices-of-data-brokers.

[19] US Federal Reserve System, Consumer Financial Protection Bureau, List of Consumer Reporting Companies, 2021, 8, https://files.consumerfinance.gov/f/documents/cfpb_consumer-reporting-companies-list_03-2021.pdf (noting there are “three big nationwide providers of consumer reports”).

[20] See, for example, Erin A. O’Hara and Larry E. Ribstein, The Law Market (Oxford, UK: Oxford University Press, 2009); Erin A. O’Hara O’Connor and Larry E. Ribstein, “Conflict of Laws and Choice of Law,” in Procedural Law and Economics, ed. Chris William Sanchirico (Northampton, MA: Edward Elgar Publishing, 2012), in Encyclopedia of Law and Economics, 2nd ed., ed. Gerrit De Geest (Northampton, MA: Edward Elgar Publishing, 2009); and Bruce H. Kobayashi and Larry E. Ribstein, eds., Economics of Federalism (Northampton, MA: Edward Elgar Publishing, 2007).

[21] See F. A. Hayek, “The Use of Knowledge in Society,” American Economic Review 35, no. 4 (September 1945): 519–30, https://www.jstor.org/stable/1809376?seq=12.

[22] Henry N. Butler and Larry E. Ribstein, “Legal Process for Fostering Innovation” (working paper, George Mason University, Antonin Scalia Law School, Fairfax, VA), 2, https://masonlec.org/site/rte_uploads/files/Butler-Ribstein-Entrepreneurship-LER.pdf.

[23] See Henry N. Butler and Larry E. Ribstein, “The Single-License Solution,” Regulation 31, no. 4 (Winter 2008–09): 36–42, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1345900.

[24] See Uniform Law Commission, “Acts Overview,” https://www.uniformlaws.org/acts/overview.

[25] Utah Code Ann. § 13-63-101 et seq. (2023).

[26] O’Hara and Ribstein, The Law Market, 5.

[27] O’Hara and Ribstein, The Law Market, 5.

[28] O’Hara and Ribstein, The Law Market, 5.

[29] See Christiano Lima-Strong, “The U.S.’s Sixth State Privacy Law Is Too ‘Weak,’ Advocates Say,” Washington Post, March 30, 2023, https://www.washingtonpost.com/politics/2023/03/30/uss-sixth-state-privacy-law-is-too-weak-advocates-say.

[30] See, for example, William L. Cary, “Federalism and Corporate Law: Reflections upon Delaware,” Yale Law Journal 83, no. 4 (March 1974): 663–705, https://openyls.law.yale.edu/bitstream/handle/20.500.13051/15589/33_83YaleLJ663_1973_1974_.pdf (arguing Delaware could export the costs of inefficiently lax regulation through the dominance of its incorporation statute).

[31] Jonathan R. Macey and Geoffrey P. Miller, “Toward an Interest-Group Theory of Delaware Corporate Law,” Texas Law Review 65, no. 3 (February 1987): 470, https://openyls.law.yale.edu/bitstream/handle/20.500.13051/1029/Toward_An_Interest_Group_Theory_of_Delaware_Corporate_Law.pdf. See also Daniel R. Fischel, “The ‘Race to the Bottom’ Revisited: Reflections on Recent Developments in Delaware’s Corporation Law,” Northwestern University Law Review 76, no. 6 (1982): 913–45, https://chicagounbound.uchicago.edu/cgi/viewcontent.cgi?referer=&httpsredir=1&article=2409&context=journal_articles.

[32] Macey and Miller, “Toward an Interest-Group Theory of Delaware Corporate Law.”

[33] Macey and Miller, “Toward an Interest-Group Theory of Delaware Corporate Law.”

[34] Commission Regulation 2016/679, General Data Protection Regulation art. 23.

[35] Commission Regulation 2016/679, General Data Protection Regulation art. 85–91.

[36] Commission Regulation 2016/679, General Data Protection Regulation art. 56.

[37] See the discussion in endnote 8.

[38] Carnival Cruise Lines v. Shute, 499 US 585 (1991).

[39] The Bremen v. Zapata, 407 US 1 (1972).

[40] O’Hara and Ribstein, The Law Market, 8.

[41] See Jim Harper, “Perspective: Utah’s Social Media Legislation May Fail, but It’s Still Good for America,” Deseret News, April 6, 2023, https://www.aei.org/op-eds/utahs-social-media-legislation-may-fail-but-its-still-good-for-america.

[42] Utah Code Ann. § 13-63-401 (2023).

[43] See Bryan L. Adkins, Alexander H. Pepper, and Jay B. Sykes, Federal Preemption: A Legal Primer, Congressional Research Service, May 18, 2023, https://sgp.fas.org/crs/misc/R45825.pdf.

[44] Congress should not interfere with interpretation of choice-of-law provisions. These issues are discussed in Tanya J. Monestier, “The Scope of Generic Choice of Law Clauses,” UC Davis Law Review 56, no. 3 (February 2023): 959–1018, https://digitalcommons.law.buffalo.edu/cgi/viewcontent.cgi?article=2148&context=journal_articles.

[45] Employee Retirement Income Security Act of 1974, 29 U.S.C. § 1144(a).

[46] Airline Deregulation Act, 49 U.S.C. § 41713(b).

[47] Federal Aviation Administration Authorization Act of 1994, 49 U.S.C. § 14501.

[48] Federal Railroad Safety Act, 49 U.S.C. § 20106.

Executive Summary

We welcome the opportunity to comment on the European Commission’s call for contributions on competition in “Virtual Worlds”.[1] The International Center for Law & Economics (“ICLE”) is a nonprofit, nonpartisan global research and policy center founded with the goal of building the intellectual foundations for sensible, economically grounded policy. ICLE promotes the use of law & economics methodologies to inform public-policy debates and has longstanding expertise in the evaluation of competition law and policy. ICLE’s interest is to ensure that competition law remains grounded in clear rules, established precedent, a record of evidence, and sound economic analysis.

The metaverse is an exciting and rapidly evolving set of virtual worlds. As with any new technology, concerns about the potential risks and negative consequences that the metaverse may bring have moved policymakers to explore how best to regulate this new space.

From the outset, it is important to recognize that simply because the metaverse is new does not mean that competition in this space is unregulated or somehow ineffective. Existing regulations may not explicitly or exclusively target metaverse ecosystems, but a vast regulatory apparatus already covers most aspects of business in virtual worlds. This includes European competition law, the Digital Markets Act (“DMA”), the General Data Protection Act (“GDPR), the Digital Services Act (“DSA”), and many more. Before it intervenes in this space, the commission should carefully consider whether there are any metaverse-specific problems not already addressed by these legal provisions.

This sense that competition intervention would be premature is reinforced by three important factors.

The first is that competition appears particularly intense in this space (Section I). There are currently multiple firms vying to offer compelling virtual worlds. At the time of writing, however, none appears close to dominating the market. In turn, this intense competition will encourage platforms to design services that meet consumers’ demands, notably in terms of safety and privacy. Nor does the market appear likely to fall into the hands of one of the big tech firms that command a sizeable share of more traditional internet services. Meta notoriously has poured more than $3.99 billion into its metaverse offerings during the first quarter of 2023, in addition to $13.72 billion the previous calendar year.[2] Despite these vast investments and a strategic focus on metaverse services, the company has, thus far, struggled to achieve meaningful traction in the space.[3]

Second, the commission’s primary concern appears to be that metaverses will become insufficiently “open and interoperable”.[4] But to the extent that these ecosystems do, indeed, become closed and proprietary, there is no reason to believe this to be a problem. Closed and proprietary ecosystems have several features that may be attractive to consumers and developers (Section II). These include improved product safety, performance, and ease of development. This is certainly not to say that closed ecosystems are always better than more open ones, but rather that it would be wrong to assume that one model or the other is optimal. Instead, the proper balance depends on tradeoffs that markets are better placed to decide.

Finally, timing is of the essence (Section III). Intervening so early in a fledgling industry’s life cycle is like shooting a moving target from a mile away. New rules or competition interventions might end up being irrelevant. Worse, by signaling that metaverses will be subject to heightened regulatory scrutiny for the foreseeable future, the commission may chill investment from the very firms is purports to support. In short, the commission should resist the urge to intervene so long as the industry is not fully mature.

I. Competing for Consumer Trust

The Commission is right to assume, in its call for contributions, that the extent to which metaverse services compete with each other (and continue to do so in the future) will largely determine whether they fulfil consumers’ expectations and meet the safety and trustworthiness requirements to which the commission aspires. As even the left-leaning Lessig put it:

Markets regulate behavior in cyberspace too. Prices structures often constrain access, and if they do not, then busy signals do. (America Online (AOL) learned this lesson when it shifted from an hourly to a flat-rate pricing plan.) Some sites on the web charge for access, as on-line services like AOL have for some time. Advertisers reward popular sites; online services drop unpopular forums. These behaviors are all a function of market constraints and market opportunity, and they all reflect the regulatory role of the market.[5]

Indeed, in a previous call for contributions, the Commission implicitly recognized the important role that competition plays, although it frames the subject primarily in terms of the problems that would arise if competition ceased to operate:

There is a risk of having a small number of big players becoming future gatekeepers of virtual worlds, creating market entry barriers and shutting out EU start-ups and SMEs from this emerging market. Such a closed ecosystem with the prevalence of proprietary systems can negatively affect the protection of personal information and data, the cybersecurity and the freedom and openness of virtual worlds at the same time.[6]

It is thus necessary to ask whether there is robust competition in the market for metaverse services. The short answer is a resounding yes.

A. Competition Without Tipping

While there is no precise definition of what constitutes a metaverse—much less a precise definition of the relevant market—available data suggests the space is highly competitive. This is evident in the fact that even a major global firm like Meta—having invested billions of dollars in its metaverse branch (and having rebranded the company accordingly)—has struggled to gain traction.[7]

Other major players in the space include the likes of Roblox, Fortnite, and Minecraft, which all have somewhere between 70 and 200 million active users.[8] This likely explains why Meta’s much-anticipated virtual world struggled to gain meaningful traction with consumers, stalling at around 300,000 active users.[9] Alongside these traditional players, there are also several decentralized platforms that are underpinned by blockchain technology. While these platforms have attracted massive investments, they are largely peripheral in terms of active users, with numbers often only in the low thousands.[10]

There are several inferences that can be drawn from these limited datasets. For one, it is clear that the metaverse industry is not yet fully mature. There are still multiple paradigms competing for consumer attention: game-based platforms versus social-network platforms; traditional platforms versus blockchain platforms, etc. In the terminology developed by David Teece, the metaverse industry has not yet reached a “paradigmatic” stage. It is fair to assume there is still significant scope for the entry of differentiated firms.[11]

It is also worth noting that metaverse competition does not appear to exhibit the same sort of network effects and tipping that is sometimes associated with more traditional social networks.[12] Despite competing for nearly a decade, no single metaverse project appears to be running away with the market.[13] This lack of tipping might be because these projects are highly differentiated.[14] It may also be due to the ease of multi-homing among them.[15]

More broadly, it is far from clear that competition will lead to a single metaverse for all uses. Different types of metaverse services may benefit from different user interfaces, graphics, and physics engines. This cuts in favor of multiple metaverses coexisting, rather than all services coordinating within a single ecosystem. Competition therefore appears likely lead to the emergence of multiple differentiated metaverses, rather than a single winner.

Ultimately, competition in the metaverse industry is strong and there is little sense these markets are about to tip towards a single firm in the year future.

B. Competing for Consumer Trust

As alluded to in the previous subsection, the world’s largest and most successful metaverse entrants to date are traditional videogaming platforms that have various marketplaces and currencies attached.[16] In other words, decentralized virtual worlds built upon blockchain technology remain marginal.

This has important policy implications. The primary legal issues raised by metaverses are the same as those encountered on other digital marketplaces. This includes issues like minor fraud, scams, and children buying content without their parents’ authorization.[17] To the extent these harms are not adequately deterred by existing laws, metaverse platforms themselves have important incentives to police them. In turn, these incentives may be compounded by strong competition among platforms.

Metaverses are generally multi-sided platforms that bring together distinct groups of users, including consumers and content creators. In order to maximize the value of their ecosystems, platforms have an incentive to balance the interests of these distinct groups.[18] In practice, this will often mean offering consumers various forms of protection against fraud and scams and actively policing platforms’ marketplaces. As David Evans puts it:

But as with any community, there are numerous opportunities for people and businesses to create negative externalities, or engage in other bad behavior, that can reduce economic efficiency and, in the extreme, lead to the tragedy of the commons. Multi-sided platforms, acting selfishly to maximize their own profits, often develop governance mechanisms to reduce harmful behavior. They also develop rules to manage many of the same kinds of problems that beset communities subject to public laws and regulations. They enforce these rules through the exercise of property rights and, most importantly, through the “Bouncer’s Right” to exclude agents from some quantum of the platform, including prohibiting some agents from the platform entirely…[19]

While there is little economic research to suggest that competition directly increases hosts’ incentive to policy their platforms, it stands to reason that doing so effectively can help platforms to expand the appeal of their ecosystems. This is particularly important for metaverse services whose userbases remain just a fraction of the size they could ultimately reach. While 100 or 200 million users already comprises a vast ecosystem, it pales in comparison to the sometimes billions of users that “traditional” online platforms attract.

The bottom line is that the market for metaverses is growing. This likely compounds platforms’ incentives to weed out undesirable behavior, thereby complementing government efforts to achieve the same goal.

II. Opening Platforms or Opening Pandora’s Box?

In its call for contributions, the commission seems concerned that the metaverse competition may lead to closed ecosystems that may be less beneficial to consumers than more open ones. But if this is indeed the commission’s fear, it is largely unfounded.

There are many benefits to closed ecosystems. Choosing the optimal degree of openness entails tradeoffs. At the very least, this suggests that policymakers should be careful not to assume that opening platforms up will systematically provide net benefits to consumers.

A. Antitrust Enforcement and Regulatory Initiatives

To understand why open (and weakly propertized) platforms are not always better for consumers, it is worth looking at past competition enforcement in the online space. Recent interventions by competition authorities have generally attempted (or are attempting) to move platforms toward more openness and less propertization. For their part, these platforms are already tremendously open (as the “platform” terminology implies) and attempt to achieve a delicate balance between centralization and decentralization.

Figure I: Directional Movement of Antitrust Intervention

The Microsoft cases and the Apple investigation both sought or seek to bring more openness and less propertization to those respective platforms. Microsoft was made to share proprietary data with third parties (less propertization) and to open its platform to rival media players and web browsers (more openness).[20] The same applies to Apple. Plaintiffs in private antitrust litigation brought in the United States[21] and government enforcement actions in Europe[22] are seeking to limit the fees that Apple can extract from downstream rivals (less propertization), as well as to ensure that it cannot exclude rival mobile-payments solutions from its platform (more openness).

The various cases that were brought by EU and U.S. authorities against Qualcomm broadly sought to limit the extent to which it was monetizing its intellectual property.[23] The European Union’s Amazon investigation centers on the ways in which the company uses data from third-party sellers (and, ultimately, the distribution of revenue between those sellers and Amazon).[24] In both cases, authorities are ultimately trying to limit the extent to which firms can propertize their assets.

Finally, both of the EU’s Google cases sought to bring more openness to the company’s main platform. The Google Shopping decision sanctioned Google for purportedly placing its services more favorably than those of its rivals.[25] The separate Android decision sought to facilitate rival search engines’ and browsers’ access to the Android ecosystem. The same appears to be true of ongoing litigation brought by state attorneys general in the United States.[26]

Much of the same can be said of the numerous regulatory initiatives pertaining to digital markets. Indeed, draft regulations being contemplated around the globe mimic the features of the antitrust/competition interventions discussed above. For instance, it is widely accepted that Europe’s DMA effectively transposes and streamlines the enforcement of the theories harm described above.[27] Similarly, several scholars have argued that the proposed American Innovation and Choice Online Act (“AICOA”) in the United States largely mimics European competition policy.[28] The legislation would ultimately require firms to open up their platforms, most notably by forcing them to treat rival services as they would their own and to make their services more interoperable with those rivals.[29]

What is striking about these decisions and investigations is the extent to which authorities are pushing back against the very features that distinguish the platforms they are investigating. Closed (or relatively closed) platforms are forced to open up, and firms with highly propertized assets are made to share them (or, at the very least, monetize them less aggressively).

B. The Empty Quadrant

All of this would not be very interesting if it weren’t for a final piece of the puzzle: the model of open and shared platforms that authorities apparently favor has traditionally struggled to gain traction with consumers. Indeed, there seem to be vanishingly few successful consumer-oriented products and services in this space.

There have been numerous attempts to introduce truly open consumer-oriented operating systems in both the mobile and desktop segments. Most have ended in failure. Ubuntu and other flavors of the Linux operating system remain fringe products. There have been attempts to create open-source search engines, but they have not met with success.[30] The picture is similar in the online retail space. Amazon appears to have beaten eBay, despite the latter being more open and less propertized. Indeed, Amazon has historically charged higher fees than eBay and offers sellers much less freedom in the ways in which they may sell their goods.[31]

This theme is repeated in the standardization space. There have been innumerable attempts to impose open, royalty-free standards. At least in the mobile-internet industry, few (if any) of these have taken off. Instead, proprietary standards such as 5G and WiFi have been far more successful. That pattern is repeated in other highly standardized industries, like digital-video formats. Most recently, the proprietary Dolby Vision format seems to be winning the war against the open HDR10+ format.[32]

Figure II: Open and Shared Platforms

This is not to say that there haven’t been any successful examples of open, royalty-free standards. Internet protocols, blockchain, and Wikipedia all come to mind. Nor does it mean that we will not see more decentralized goods in the future. But by and large, firms and consumers have not yet taken to the idea of fully open and shared platforms. Or, at least, those platforms have not yet achieved widespread success in the marketplace (potentially due to supply-side considerations, such as the difficulty of managing open platforms or the potentially lower returns to innovation in weakly propertized ones).[33] And while some “open” projects have achieved tremendous scale, the consumer-facing side of these platforms is often dominated by intermediaries that opt for much more traditional business models (think of Coinbase in the blockchain space, or Android’s use of Linux).

C. Potential Explanations

The preceding section posited a recurring reality: the digital platforms that competition authorities wish to bring into existence are fundamentally different from those that emerge organically. But why have authorities’ ideal platforms, so far, failed to achieve truly meaningful success?

Three potential explanations come to mind. First, “closed” and “propertized” platforms might systematically—and perhaps anticompetitively—thwart their “open” and “shared” rivals. Second, shared platforms might fail to persist (or grow pervasive) because they are much harder to monetize, and there is thus less incentive to invest in them. This is essentially a supply-side explanation. Finally, consumers might opt for relatively closed systems precisely because they prefer these platforms to marginally more open ones—i.e., a demand-side explanation.

In evaluating the first conjecture, the key question is whether successful “closed” and “propertized” platforms overcame their rivals before or after they achieved some measure of market dominance. If success preceded dominance, then anticompetitive foreclosure alone cannot explain the proliferation of the “closed” and “propertized” model.[34]

Many of today’s dominant platforms, however, often overcame open/shared rivals, well before they achieved their current size. It is thus difficult to make the case that the early success of their business models was due to anticompetitive behavior. This is not to say these business models cannot raise antitrust issues, but rather that anticompetitive behavior is not a good explanation for their emergence.

Both the second and the third conjectures essentially ask whether “closed” and “propertized” might be better adapted to their environment than “open” and “shared” rivals.

In that respect, it is not unreasonable to surmise that highly propertized platforms would generally be easier to monetize than shared ones. For example, to monetize open-source platforms often requires relying on complementarities, which tend to be vulnerable to outside competition and free-riding.[35] There is thus a natural incentive for firms to invest and innovate in more propertized environments. In turn, competition enforcement that limits a platform’s ability to propertize their assets may harm innovation.

Similarly, authorities should reflect on whether consumers really want the more “competitive” ecosystems that they are trying to design. The European Commission, for example, has a long track record of seeking to open digital platforms, notably by requiring that platform owners do not preinstall their own web browsers (the Microsoft decisions are perhaps the most salient example). And yet, even after these interventions, new firms have kept using the very business model that the commission reprimanded, rather than the “pro-consumer” model it sought to impose on the industry. For example, Apple tied the Safari browser to its iPhones; Google went to some length to ensure that Chrome was preloaded on devices; and Samsung phones come with Samsung Internet as default.[36] Yet this has not ostensibly steered consumers away from those platforms.

Along similar lines, a sizable share of consumers opt for Apple’s iPhone, which is even more centrally curated than Microsoft Windows ever was (and the same is true of Apple’s MacOS). In other words, it is hard to claim that opening platforms is inherently good for consumers when those same consumers routinely opt for platforms with the very features that policymakers are trying to eliminate.

Finally, it is worth noting that the remedies imposed by competition authorities have been anything but successes. Windows XP N (the version of Windows that came without Windows Media Player) was an unmitigated flop, selling a paltry 1,787 copies.[37] Likewise, the internet-browser “ballot box” imposed by the commission was so irrelevant to consumers that it took months for authorities to notice that Microsoft had removed it, in violation of the commission’s decision.[38]

One potential inference is that consumers do not value competition interventions that make dominant ecosystems marginally more open and less propertized. There are also many reasons why consumers might prefer “closed” systems (at least, relative to the model favored by many policymakers), even when they must pay a premium for them.

Take the example of app stores. Maintaining some control over the apps that can access the store enables platforms to easily weed out bad actors. Similarly, controlling the hardware resources that each app can use may greatly improve device performance. Indeed, it may be that a measure of control facilitates the very innovations that consumers demand. Therefore, “authorities and courts should not underestimate the indispensable role control plays in achieving coordination and coherence in the context of systemic ef?ciencies. Without it, the attempted novelties and strategies might collapse under their own complexity.”[39]

Relatively centralized platforms can eliminate negative externalities that “bad” apps impose on rival apps and consumers.[40] This is especially true when consumers will tend to attribute dips in performance to the overall platform, rather than to a particular app.[41] At the same time, they can take advantage of positive externalities to improve the quality of the overall platform.

And it is surely the case that consumers prefer to make many of their decisions at the inter-platform level, rather than within each platform. In simple terms, users arguably make their most important decision when they choose between an Apple or Android smartphone (or a Mac and a PC, etc.). In doing so, they can select their preferred app suite with one simple decision. They might thus purchase an iPhone because they like the secure App Store, or an Android smartphone because they like the Chrome Browser and Google Search. Absent false information at the time of the initial platform decision, this decision will effectively incorporate expectations about subsequent constraints.[42]

Furthermore, forcing users to make too many “within-platform” choices may undermine a product’s attractiveness. Indeed, it is difficult to create a high-quality reputation if each user’s experience is fundamentally different.[43] In short, contrary to what antitrust authorities appear to believe, closed platforms might give most users exactly what they desire.

All of this suggests that consumers and firms often gravitate spontaneously toward both closed and highly propertized platforms, the opposite of what the commission and other competition authorities tend to favor. The reasons for this trend are still misunderstood, and mostly ignored. Too often it is simply assumed that consumers benefit from more openness, and that shared/open platforms are the natural order of things. Instead, what some regard as “market failures” may in fact be features that explain the rapid emergence of the digital economy.

When considering potential policy reforms targeting the metaverse, policymakers would be wrong to assume openness (notably, in the form of interoperability) and weak propertization are always objectively superior. Instead, these platform designs entail important tradeoffs. Closed metaverse ecosystems may lead to higher consumer safety and better performance, while interoperable systems may reduce the frictions consumers face when moving from one service to another. There is little reason to believe policymakers are in a better position to weigh these tradeoffs than consumers, who vote with their virtual feet.

III. Conclusion: Competition Intervention Would be Premature

A final important argument against intervening today is that the metaverse industry is nowhere near mature. Tomorrow’s competition-related challenges and market failures might not be the same as today’s. This makes it exceedingly difficult for policymakers to design appropriate remedies and increases the risk that intervention might harm innovation.

As of 2023, the entire metaverse industry (both hardware and software) is estimated to be worth somewhere in the vicinity of $80 billion, and projections suggest this could grow by a factor of 10 by 2030.[44] Growth projections of this sort are notoriously unreliable. But in this case, they do suggest there is some consensus that the industry is not fully fledged.

Along similar lines, it remains unclear what types of metaverse services will gain the most traction with consumers, what sorts of hardware consumers will use to access these services, and what technologies will underpin the most successful metaverse platforms. In fact, it is still an open question whether the metaverse industry will foster any services that achieve widespread consumer adoption in the foreseeable future.[45] In other words, it is not exactly clear what metaverse products and services the Commission should focus on in the first place.

Given these uncertainties, competition intervention in the metaverse appears premature. Intervening so early in the industry’s life cycle is like aiming at a moving target. Ensuing remedies might end up being irrelevant before they have any influence on the products that firms develop. More worryingly, acting now signals that the metaverse industry will be subject to heightened regulatory scrutiny for the foreseeable future. In turn, this may deter large platforms from investing in the European market. It also may funnel venture-capital investments away from the European continent.

Competition intervention in burgeoning industries is no free lunch. The best evidence concerning these potential costs comes from the GDPR. While privacy regulation is obviously not the same as competition law, the evidence concerning the GDPR suggests that heavy-handed intervention may, at least in some instances, slow down innovation and reduce competition.

The most-cited empirical evidence concerning the effects of the GDPR comes from a paper by Garrett Johnson and co-authors, who link the GDPR to widespread increases to market concentration, particularly in the short-term:

We show that websites’ vendor use falls after the European Union’s (EU’s) General Data Protection Regulation (GDPR), but that market concentration also increases among technology vendors that provide support services to websites…. The week after the GDPR’s enforcement, website use of web technology vendors falls by 15% for EU residents. Websites are relatively more likely to retain top vendors, which increases the concentration of the vendor market by 17%. Increased concentration predominantly arises among vendors that use personal data, such as cookies, and from the increased relative shares of Facebook and Google-owned vendors, but not from website consent requests. Although the aggregate changes in vendor use and vendor concentration dissipate by the end of 2018, we find that the GDPR impact persists in the advertising vendor category most scrutinized by regulators.[46]

Along similar lines, an NBER working paper by Jian Jia and co-authors finds that enactment of the GDPR markedly reduced venture-capital investments in Europe:

Our findings indicate a negative differential effect on EU ventures after the rollout of GDPR relative to their US counterparts. These negative effects manifest in the overall number of financing rounds, the overall dollar amount raised across rounds, and in the dollar amount raised per individual round. Specifically, our findings suggest a $3.38 million decrease in the aggregate dollars raised by EU ventures per state per crude industry category per week, a 17.6% reduction in the number of weekly venture deals, and a 39.6% decrease in the amount raised in an average deal following the rollout of GDPR.[47]

In another paper, Samuel Goldberg and co-authors find that the GDPR led to a roughly 12% reduction in website pageviews and e-commerce revenue in Europe.[48] Finally, Rebecca Janssen and her co-authors show that the GDPR decreased the number of apps offered on Google’s Play Store between 2016 and 2019:

Using data on 4.1 million apps at the Google Play Store from 2016 to 2019, we document that GDPR induced the exit of about a third of available apps; and in the quarters following implementation, entry of new apps fell by half.[49]

Of course, the body of evidence concerning the GDPR’s effects is not entirely unambiguous. For example, Rajkumar Vekatesean and co-authors find that the GDPR had mixed effects on the returns of different types of firms.[50] Other papers also show similarly mixed effects.[51]

Ultimately, the empirical literature concerning the effects of the GDPR shows that regulation—in this case, privacy protection—is no free lunch. Of course, this does not mean that competition intervention targeting the metaverse would necessarily have these same effects. But in the absence of a clear market failure to solve, it is unclear why policymakers should run such a risk in the first place.

In the end, competition intervention in the metaverse is unlikely to be costless. The metaverse is still in its infancy, regulation could deter essential innovation, and the commission has thus far failed to identify any serious market failures that warrant public intervention. The result is that the commission’s call for contributions appears premature or, in other words, that the commission is putting the meta-cart before the meta-horse.

 

[1] Competition in Virtual Worlds and Generative AI – Calls for contributions, European Commission (Jan. 9, 2024) https://competition-policy.ec.europa.eu/document/download/e727c66a-af77-4014-962a-7c9a36800e2f_en?filename=20240109_call-for-contributions_virtual-worlds_and_generative-AI.pdf (hereafter, “Call for Contributions”).

[2] Jonathan Vaian, Meta’s Reality Labs Records $3.99 Billion Quarterly Loss as Zuckerberg Pumps More Cash into Metaverse, CNBC (Apr. 26, 2023), https://www.cnbc.com/2023/04/26/metas-reality-labs-unit-records-3point99-billion-first-quarter-loss-.html.

[3] Alan Truly, Horizon Worlds Leak: Only 1 in 10 Users Return & Web Launch Is Coming, Mixed News (Mar. 3, 2023), https://mixed-news.com/en/horizon-worlds-leak-only-1-in-10-users-return-web-launch-coming; Kevin Hurler, Hey Fellow Kids: Meta Is Revamping Horizon Worlds to Attract More Teen Users, Gizmodo (Feb. 7, 2023), https://gizmodo.com/meta-metaverse-facebook-horizon-worlds-vr-1850082068; Emma Roth, Meta’s Horizon Worlds VR Platform Is Reportedly Struggling to Keep Users, The Verge (Oct. 15, 2022),
https://www.theverge.com/2022/10/15/23405811/meta-horizon-worlds-losing-users-report; Paul Tassi, Meta’s ‘Horizon Worlds’ Has Somehow Lost 100,000 Players in Eight Months, Forbes, (Oct. 17, 2022), https://www.forbes.com/sites/paultassi/2022/10/17/metas-horizon-worlds-has-somehow-lost-100000-players-in-eight-months/?sh=57242b862a1b.

[4] Call for Contributions, supra note 1. (“6) Do you expect the technology incorporated into Virtual World platforms, enabling technologies of Virtual Worlds and services based on Virtual Worlds to be based mostly on open standards and/or protocols agreed through standard-setting organisations, industry associations or groups of companies, or rather the use of proprietary technology?”).

[5] Less Lawrence Lessig, The Law of the Horse: What Cyberlaw Might Teach, 113 Harv. L. Rev. 508 (1999).

[6] Virtual Worlds (Metaverses) – A Vision for Openness, Safety and Respect, European Commission, https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/13757-Virtual-worlds-metaverses-a-vision-for-openness-safety-and-respect/feedback_en?p_id=31962299H.

[7] Catherine Thorbecke, What Metaverse? Meta Says Its Single Largest Investment Is Now in ‘Advancing AI’, CNN Business (Mar. 15, 2023), https://www.cnn.com/2023/03/15/tech/meta-ai-investment-priority/index.html; Ben Marlow, Mark Zuckerberg’s Metaverse Is Shattering into a Million Pieces, The Telegraph (Apr. 23, 2023), https://www.telegraph.co.uk/business/2023/04/21/mark-zuckerbergs-metaverse-shattering-million-pieces; Will Gendron, Meta Has Reportedly Stopped Pitching Advertisers on the Metaverse, BusinessInsider (Apr. 18, 2023), https://www.businessinsider.com/meta-zuckerberg-stopped-pitching-advertisers-metaverse-focus-reels-ai-report-2023-4.

[8] Mansoor Iqbal, Fortnite Usage and Revenue Statistics, Business of Apps (Jan. 9, 2023), https://www.businessofapps.com/data/fortnite-statistics; Matija Ferjan, 76 Little-Known Metaverse Statistics & Facts (2023 Data), Headphones Addict (Feb. 13, 2023), https://headphonesaddict.com/metaverse-statistics.

[9] James Batchelor, Meta’s Flagship Metaverse Horizon Worlds Struggling to Attract and Retain Users, Games Industry (Oct. 17, 2022), https://www.gamesindustry.biz/metas-flagship-metaverse-horizon-worlds-struggling-to-attract-and-retain-users; Ferjan, id.

[10] Richard Lawler, Decentraland’s Billion-Dollar ‘Metaverse’ Reportedly Had 38 Active Users in One Day, The Verge (Oct. 13, 2022), https://www.theverge.com/2022/10/13/23402418/decentraland-metaverse-empty-38-users-dappradar-wallet-data; The Sandbox, DappRadar, https://dappradar.com/multichain/games/the-sandbox (last visited May 3, 2023); Decentraland, DappRadar, https://dappradar.com/multichain/social/decentraland (last visited May 3, 2023).

[11] David J. Teece, Profiting from Technological Innovation: Implications for Integration, Collaboration, Licensing and Public Policy, 15 Research Policy 285-305 (1986), https://www.sciencedirect.com/science/article/abs/pii/0048733386900272.

[12] Geoffrey Manne & Dirk Auer, Antitrust Dystopia and Antitrust Nostalgia: Alarmist Theories of Harm in Digital Markets and Their Origins, 28 Geo. Mason L. Rev. 1279 (2021).

[13] Roblox, Wikipedia, https://en.wikipedia.org/wiki/Roblox (last visited May 3, 2023); Minecraft, Wikipedia, https://en.wikipedia.org/wiki/Minecraft (last visited May 3, 2023); Fortnite, Wikipedia, https://en.wikipedia.org/wiki/Fortnite (last visited May 3, 2023); see Fiza Chowdhury, Minecraft vs Roblox vs Fortnite: Which Is Better?, Metagreats (Feb. 20, 2023), https://www.metagreats.com/minecraft-vs-roblox-vs-fortnite.

[14]  Marc Rysman, The Economics of Two-Sided Markets, 13 J. Econ. Perspectives 134 (2009) (“First, if standards can differentiate from each other, they may be able to successfully coexist (Chou and Shy, 1990; Church and Gandal, 1992). Arguably, Apple and Microsoft operating systems have both survived by specializing in different markets: Microsoft in business and Apple in graphics and education. Magazines are an obvious example of platforms that differentiate in many dimensions and hence coexist.”).

[15] Id. at 134 (“Second, tipping is less likely if agents can easily use multiple standards. Corts and Lederman (forthcoming) show that the fixed cost of producing a video game for one more standard have reduced over time relative to the overall fixed costs of producing a game, which has led to increased distribution of games across multiple game systems (for example, PlayStation, Nintendo, and Xbox) and a less-concentrated game system market.”).

[16] What Are Fortnite, Roblox, Minecraft and Among Us? A Parent’s Guide to the Most Popular Online Games Kids Are Playing, FTC Business (Oct. 5, 2021), https://www.ftc.net/blog/what-are-fortnite-roblox-minecraft-and-among-us-a-parents-guide-to-the-most-popular-online-games-kids-are-playing; Jay Peters, Epic Is Merging Its Digital Asset Stores into One Huge Marketplace, The Verge (Mar. 22, 2023), https://www.theverge.com/2023/3/22/23645601/epic-games-fab-asset-marketplace-state-of-unreal-2023-gdc.

[17] Luke Winkie, Inside Roblox’s Criminal Underworld, Where Kids Are Scamming Kids, IGN (Jan. 2, 2023), https://www.ign.com/articles/inside-robloxs-criminal-underworld-where-kids-are-scamming-kids; Fake Minecraft Updates Pose Threat to Users, Tribune (Sept. 11, 2022), https://tribune.com.pk/story/2376087/fake-minecraft-updates-pose-threat-to-users; Ana Diaz, Roblox and the Wild West of Teenage Scammers, Polygon (Aug. 24, 2019) https://www.polygon.com/2019/8/24/20812218/roblox-teenage-developers-controversy-scammers-prison-roleplay; Rebecca Alter, Fortnite Tries Not to Scam Children and Face $520 Million in FTC Fines Challenge, Vulture (Dec. 19, 2022), https://www.vulture.com/2022/12/fortnite-epic-games-ftc-fines-privacy.html; Leonid Grustniy, Swindle Royale: Fortnite Scammers Get Busy, Kaspersky Daily (Dec. 3, 2020), https://www.kaspersky.com/blog/top-four-fortnite-scams/37896.

[18] See, generally, David Evans & Richard Schmalensee, Matchmakers: The New Economics of Multisided Platforms (Harvard Business Review Press, 2016).

[19] David S. Evans, Governing Bad Behaviour By Users of Multi-Sided Platforms, Berkley Technology Law Journal 27:2 (2012), 1201.

[20] See Case COMP/C-3/37.792, Microsoft, OJ L 32 (May 24, 2004). See also, Case COMP/39.530, Microsoft (Tying), OJ C 120 (Apr. 26, 2013).

[21] See Complaint, Epic Games, Inc. v. Apple Inc., 493 F. Supp. 3d 817 (N.D. Cal. 2020) (4:20-cv-05640-YGR).

[22] See European Commission Press Release IP/20/1073, Antitrust: Commission Opens Investigations into Apple’s App Store Rules (Jun. 16, 2020); European Commission Press Release IP/20/1075, Antitrust: Commission Opens Investigation into Apple Practices Regarding Apple Pay (Jun. 16, 2020).

[23] See European Commission Press Release IP/18/421, Antitrust: Commission Fines Qualcomm €997 Million for Abuse of Dominant Market Position (Jan. 24, 2018); Federal Trade Commission v. Qualcomm Inc., 969 F.3d 974 (9th Cir. 2020).

[24] See European Commission Press Release IP/19/4291, Antitrust: Commission Opens Investigation into Possible Anti-Competitive Conduct of Amazon (Jul. 17, 2019).

[25] See Case AT.39740, Google Search (Shopping), 2017 E.R.C. I-379. See also, Case AT.40099 (Google Android), 2018 E.R.C.

[26] See Complaint, United States v. Google, LLC, (2020), https://www.justice.gov/opa/pr/justice-department-sues-monopolist-google-violating-antitrust-laws; see also, Complaint, Colorado et al. v. Google, LLC, (2020), available at https://coag.gov/app/uploads/2020/12/Colorado-et-al.-v.-Google-PUBLIC-REDACTED-Complaint.pdf.

[27] See, e.g., Giorgio Monti, The Digital Markets Act: Institutional Design and Suggestions for Improvement, Tillburg L. & Econ. Ctr., Discussion Paper No. 2021-04 (2021), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3797730 (“In sum, the DMA is more than an enhanced and simplified application of Article 102 TFEU: while the obligations may be criticised as being based on existing competition concerns, they are forward-looking in trying to create a regulatory environment where gatekeeper power is contained and perhaps even reduced.”) (Emphasis added).

[28] See, e.g., Aurelien Portuese, “Please, Help Yourself”: Toward a Taxonomy of Self-Preferencing, Information Technology & Innovation Foundation (Oct. 25, 2021), available at https://itif.org/sites/default/files/2021-self-preferencing-taxonomy.pdf. (“The latest example of such weaponization of self-preferencing by antitrust populists is provided by Sens. Amy Klobuchar (D-MN) and Chuck Grassley (R-IA). They introduced legislation in October 2021 aimed at prohibiting the practice.2 However, the legislation would ban self-preferencing only for a handful of designated companies—the so-called “covered platforms,” not the thousands of brick-and-mortar sellers that daily self-preference for the benefit of consumers. Mimicking the European Commission’s Digital Markets Act prohibiting self-preferencing, Senate and the House bills would degrade consumers’ experience and undermine competition, since self-preferencing often benefits consumers and constitutes an integral part, rather than an abnormality, of the process of competition.”).

[29] Efforts to saddle platforms with “non-discrimination” constraints are tantamount to mandating openness. See Geoffrey A. Manne, Against the Vertical Discrimination Presumption, Foreword, Concurrences No. 2-2020 (2020) at 2 (“The notion that platforms should be forced to allow complementors to compete on their own terms, free of constraints or competition from platforms is a species of the idea that platforms are most socially valuable when they are most ‘open.’ But mandating openness is not without costs, most importantly in terms of the effective operation of the platform and its own incentives for innovation.”).

[30] See, e.g., Klint Finley, Your Own Private Google: The Quest for an Open Source Search Engine, Wired (Jul. 12, 2021), https://www.wired.com/2012/12/solar-elasticsearch-google.

[31] See Brian Connolly, Selling on Amazon vs. eBay in 2021: Which Is Better?, JungleScout (Jan. 12, 2021), https://www.junglescout.com/blog/amazon-vs-ebay; Crucial Differences Between Amazon and eBay, SaleHOO, https://www.salehoo.com/educate/selling-on-amazon/crucial-differences-between-amazon-and-ebay (last visited Feb. 8, 2021).

[32] See, e.g., Dolby Vision Is Winning the War Against HDR10 +, It Requires a Single Standard, Tech Smart, https://voonze.com/dolby-vision-is-winning-the-war-against-hdr10-it-requires-a-single-standard (last visited June 6, 2022).

[33] On the importance of managers, see, e.g., Nicolai J Foss & Peter G Klein, Why Managers Still Matter, 56 MIT Sloan Mgmt. Rev., 73 (2014) (“In today’s knowledge-based economy, managerial authority is supposedly in decline. But there is still a strong need for someone to define and implement the organizational rules of the game.”).

[34] It is generally agreed upon that anticompetitive foreclosure is possible only when a firm enjoys some degree of market power. Frank H. Easterbrook, Limits of Antitrust, 63 Tex. L. Rev. 1, 20 (1984) (“Firms that lack power cannot injure competition no matter how hard they try. They may injure a few consumers, or a few rivals, or themselves (see (2) below) by selecting ‘anticompetitive’ tactics. When the firms lack market power, though, they cannot persist in deleterious practices. Rival firms will offer the consumers better deals. Rivals’ better offers will stamp out bad practices faster than the judicial process can. For these and other reasons many lower courts have held that proof of market power is an indispensable first step in any case under the Rule of Reason. The Supreme Court has established a market power hurdle in tying cases, despite the nominally per se character of the tying offense, on the same ground offered here: if the defendant lacks market power, other firms can offer the customer a better deal, and there is no need for judicial intervention.”).

[35] See, e.g., Josh Lerner & Jean Tirole, Some Simple Economics of Open Source, 50 J. Indus. Econ. 197 (2002).

[36] See Matthew Miller, Thanks, Samsung: Android’s Best Mobile Browser Now Available to All, ZDNet (Aug. 11, 2017), https://www.zdnet.com/article/thanks-samsung-androids-best-mobile-browser-now-available-to-all.

[37] FACT SHEET: Windows XP N Sales, RegMedia (Jun. 12, 2009), available at https://regmedia.co.uk/2009/06/12/microsoft_windows_xp_n_fact_sheet.pdf.

[38] See Case COMP/39.530, Microsoft (Tying), OJ C 120 (Apr. 26, 2013).

[39] Konstantinos Stylianou, Systemic Efficiencies in Competition Law: Evidence from the ICT Industry, 12 J. Competition L. & Econ. 557 (2016).

[40] See, e.g., Steven Sinofsky, The App Store Debate: A Story of Ecosystems, Medium (Jun. 21, 2020), https://medium.learningbyshipping.com/the-app-store-debate-a-story-of-ecosystems-938424eeef74.

[41] Id.

[42] See, e.g., Benjamin Klein, Market Power in Aftermarkets, 17 Managerial & Decision Econ. 143 (1996).

[43] See, e.g., Simon Hill, What Is Android Fragmentation, and Can Google Ever Fix It?, DigitalTrends (Oct. 31, 2018), https://www.digitaltrends.com/mobile/what-is-android-fragmentation-and-can-google-ever-fix-it.

[44] Metaverse Market Revenue Worldwide from 2022 to 2030, Statista, https://www.statista.com/statistics/1295784/metaverse-market-size (last visited May 3, 2023); Metaverse Market by Component (Hardware, Software (Extended Reality Software, Gaming Engine, 3D Mapping, Modeling & Reconstruction, Metaverse Platform, Financial Platform), and Professional Services), Vertical and Region – Global Forecast to 2027, Markets and Markets (Apr. 27, 2023), https://www.marketsandmarkets.com/Market-Reports/metaverse-market-166893905.html; see also, Press Release, Metaverse Market Size Worth $ 824.53 Billion, Globally, by 2030 at 39.1% CAGR, Verified Market Research (Jul. 13, 2022), https://www.prnewswire.com/news-releases/metaverse-market-size-worth–824-53-billion-globally-by-2030-at-39-1-cagr-verified-market-research-301585725.html.

[45] See, e.g., Megan Farokhmanesh, Will the Metaverse Live Up to the Hype? Game Developers Aren’t Impressed, Wired (Jan. 19, 2023), https://www.wired.com/story/metaverse-video-games-fortnite-zuckerberg; see also Mitch Wagner, The Metaverse Hype Bubble Has Popped. What Now?, Fierce Electronics (Feb. 24, 2023), https://www.fierceelectronics.com/embedded/metaverse-hype-bubble-has-popped-what-now.

[46] Garret A. Johnson, et al., Privacy and Market Concentration: Intended and Unintended Consequences of the GDPR, Forthcoming Management Science 1 (2023).

[47] Jian Jia, et al., The Short-Run Effects of GDPR on Technology Venture Investment, NBER Working Paper 25248, 4 (2018), available at https://www.nber.org/system/files/working_papers/w25248/w25248.pdf.

[48] Samuel G. Goldberg, Garrett A. Johnson, & Scott K. Shriver, Regulating Privacy Online: An Economic Evaluation of GDPR (2021), available at https://www.ftc.gov/system/files/documents/public_events/1588356/johnsongoldbergshriver.pdf.

[49] Rebecca Janßen, Reinhold Kesler, Michael Kummer, & Joel Waldfogel, GDPR and the Lost Generation of Innovative Apps, Nber Working Paper 30028, 2 (2022), available at https://www.nber.org/system/files/working_papers/w30028/w30028.pdf.

[50] Rajkumar Venkatesan, S. Arunachalam & Kiran Pedada, Short Run Effects of Generalized Data Protection Act on Returns from AI Acquisitions, University of Virginia Working Paper 6 (2022), available at: https://conference.nber.org/conf_papers/f161612.pdf. (“On average, GDPR exposure reduces the ROA of firms. We also find that GDPR exposure increases the ROA of firms that make AI acquisitions for improving customer experience, and cybersecurity. Returns on AI investments in innovation and operational efficiencies are unaffected by GDPR.”)

[51] For a detailed discussion of the empirical literature concerning the GDPR, see Garrett Johnson, Economic Research on Privacy Regulation: Lessons From the GDPR And Beyond, NBER Working Paper 30705 (2022), available at https://www.nber.org/system/files/working_papers/w30705/w30705.pdf.

Executive Summary

We thank the European Commission for launching this consultation on competition in generative AI. The International Center for Law & Economics (“ICLE”) is a nonprofit, nonpartisan global research and policy center founded with the goal of building the intellectual foundations for sensible, economically grounded policy. ICLE promotes the use of law & economics methodologies to inform public-policy debates and has longstanding expertise in the evaluation of competition law and policy. ICLE’s interest is to ensure that competition law remains grounded in clear rules, established precedent, a record of evidence, and sound economic analysis.

In our comments, we express concern that policymakers may equate the rapid rise of generative AI services with a need to intervene in these markets when, in fact, the opposite is true. As we explain, the rapid growth of AI markets, as well as the fact that new market players are thriving, suggests competition is intense. If incumbent firms could easily leverage their dominance into burgeoning generative AI markets, we would not have seen the growth of generative AI unicorns such as OpenAI, Midjourney, and Anthropic, to name but a few.

Of course, this is not to say that generative AI markets are not important—quite the opposite. Generative AI is already changing the ways that many firms do business and improving employee productivity in many industries.[1] The technology is also increasingly useful in the field of scientific research, where it has enabled creation of complex models that expand scientists’ reach.[2] Against this backdrop, Commissioner Margrethe Vestager was right to point out that it “is fundamental that these new markets stay competitive, and that nothing stands in the way of businesses growing and providing the best and most innovative products to consumers.”[3]

But while sensible enforcement is of vital importance to maintain competition and consumer welfare, knee-jerk reactions may yield the opposite outcomes. As our comments explain, overenforcement in the field of generative AI could cause the very harms that policymakers seek to avert. For instance, preventing so-called “big tech” firms from competing in these markets (for example, by threatening competition intervention as soon as they embed generative AI services in their ecosystems or seek to build strategic relationships with AI startups) may thwart an important source of competition needed to keep today’s leading generative-AI firms in check. In short, competition in AI markets is important, but trying naïvely to hold incumbent tech firms back out of misguided fears they will come to dominate this space is likely to do more harm than good.

Our comment proceeds as follows. Section I summarizes recent calls for competition intervention in generative AI markets. Section II argues that many of these calls are underpinned by fears of data-related incumbency advantages (often referred to as “data-network effects”). Section III explains why these effects are unlikely to play a meaningful role in generative-AI markets. Section IV concludes by offering five key takeaways to help policymakers (including the Commission) better weigh the tradeoffs inherent to competition intervention in generative-AI markets.

I. Calls for Intervention in AI Markets

It was once (and frequently) said that Google’s “data monopoly” was unassailable: “If ‘big data’ is the oil of the information economy, Google has Standard Oil-like monopoly dominance—and uses that control to maintain its dominant position.”[4] Similar claims of data dominance have been attached to nearly all large online platforms, including Facebook (Meta), Amazon, and Uber.[5]

While some of these claims continue even today (for example, “big data” is a key component of the U.S. Justice Department’s (DOJ) Google Search and adtech antitrust suits),[6] a shiny new data target has emerged in the form of generative artificial intelligence (AI). The launch of ChatGPT in November 2022, as well as the advent of AI image-generation services like Midjourney and Dall-E, have dramatically expanded the public’s conception of what is—and what might be—possible to achieve with generative-AI technologies built on massive datasets.

While these services remain in the early stages of mainstream adoption and remain in the throes of rapid, unpredictable technological evolution, they nevertheless already appear to be on the radar of competition policymakers around the world. Several antitrust enforcers appear to believe that, by acting now, they can avoid the “mistakes” that were purportedly made during the formative years of Web 2.0.[7] These mistakes, critics assert, include failing to appreciate the centrality of data in online markets, as well as letting mergers go unchecked and allowing early movers to entrench their market positions.^[8] As Lina Khan, chair of the U.S. Federal Trade Commission (FTC), put it: “we are still reeling from the concentration that resulted from Web 2.0, and we don’t want to repeat the mis-steps of the past with AI”.^[9]

This response from the competition-policy world is deeply troubling. Rather than engage in critical self-assessment and adopt an appropriately restrained stance, the enforcement community appears to be champing at the bit. Rather than assessing their prior assumptions based on the current technological moment, enforcers’ top priority appears to be figuring out how to rapidly and almost reflexively deploy existing competition tools to address the presumed competitive failures presented by generative AI.^[10]

It is increasingly common for competition enforcers to argue that so-called “data-network effects” serve not only to entrench incumbents in those markets where the data is collected, but also confer similar, self-reinforcing benefits in adjacent markets. Several enforcers have, for example, prevented large online platforms from acquiring smaller firms in adjacent markets, citing the risk that they could use their vast access to data to extend their dominance into these new markets.^[11]

They have also launched consultations to ascertain the role that data plays in AI competition. For instance, in an ongoing consultation, the European Commission asks: “What is the role of data and what are its relevant characteristics for the provision of generative AI systems and/or components, including AI models?”^[12] Unsurprisingly, the FTC has likewise been bullish about the risks posed by incumbents’ access to data. In comments submitted to the U.S. Copyright Office, for example, the FTC argued that:

The rapid development and deployment of AI also poses potential risks to competition. The rising importance of AI to the economy may further lock in the market dominance of large incumbent technology firms. These powerful, vertically integrated incumbents control many of the inputs necessary for the effective development and deployment of AI tools, including cloud-based or local computing power and access to large stores of training data. These dominant technology companies may have the incentive to use their control over these inputs to unlawfully entrench their market positions in AI and related markets, including digital content markets.[13]

Certainly, it stands to reason that the largest online platforms—including Alphabet, Meta, Apple, and Amazon—should have a meaningful advantage in the burgeoning markets for generative-AI services. After all, it is widely recognized that data is an essential input for generative AI.^[14] This competitive advantage should be all the more significant, given that these firms have been at the forefront of AI technology for more than a decade. Over this period, Google’s DeepMind and AlphaGo and Meta’s have routinely made headlines.[15] Apple and Amazon also have vast experience with AI assistants, and all of these firms use AI technology throughout their platforms.[16]

Contrary to what one might expect, however, the tech giants have, to date, been largely unable to leverage their vast data troves to outcompete startups like OpenAI and Midjourney. At the time of writing, OpenAI’s ChatGPT appears to be, by far, the most successful chatbot,[17] despite the large tech platforms’ apparent access to far more (and more up-to-date) data.

In these comments, we suggest that there are important lessons to glean from these developments, if only enforcers would stop to reflect. The meteoric rise of consumer-facing AI services should offer competition enforcers and policymakers an opportunity for introspection. As we explain, the rapid emergence of generative-AI technology may undercut many core assumptions of today’s competition-policy debates, which have largely focused on the rueful after-effects of the purported failure of 20^th-century antitrust to address the allegedly manifest harms of 21^st-century technology. These include the notions that data advantages constitute barriers to entry and can be leveraged to project dominance into adjacent markets; that scale itself is a market failure to be addressed by enforcers; and that the use of consumer data is inherently harmful to those consumers.

II. Data-Network Effects Theory and Enforcement

Proponents of tougher interventions by competition enforcers into digital markets often cite data-network effects as a source of competitive advantage and barrier to entry (though terms like “economies of scale and scope” may offer more precision).^[18] The crux of the argument is that “the collection and use of data creates a feedback loop of more data, which ultimately insulates incumbent platforms from entrants who, but for their data disadvantage, might offer a better product.”^[19] This self-reinforcing cycle purportedly leads to market domination by a single firm. Thus, it is argued, for example, that Google’s “ever-expanding control of user personal data, and that data’s critical value to online advertisers, creates an insurmountable barrier to entry for new competition.”^[20]

Right off the bat, it is important to note the conceptual problem these claims face. Because data can be used to improve the quality of products and/or to subsidize their use, the idea of data as an entry barrier suggests that any product improvement or price reduction made by an incumbent could be a problematic entry barrier to any new entrant. This is tantamount to an argument that competition itself is a cognizable barrier to entry. Of course, it would be a curious approach to antitrust if competition were treated as a problem, as it would imply that firms should under-compete—i.e., should forego consumer-welfare enhancements—in order to inculcate a greater number of firms in a given market simply for its own sake.[21]

Meanwhile, actual economic studies of data-network effects have been few and far between, with scant empirical evidence to support the theory.^[22] Andrei Hagiu and Julian Wright’s theoretical paper offers perhaps the most comprehensive treatment of the topic to date.[23] The authors ultimately conclude that data-network effects can be of different magnitudes and have varying effects on firms’ incumbency advantage.^[24] They cite Grammarly (an AI writing-assistance tool) as a potential example: “As users make corrections to the suggestions offered by Grammarly, its language experts and artificial intelligence can use this feedback to continue to improve its future recommendations for all users.”^[25]

This is echoed by other economists who contend that “[t]he algorithmic analysis of user data and information might increase incumbency advantages, creating lock-in effects among users and making them more reluctant to join an entrant platform.”^[26] Crucially, some scholars take this logic a step further, arguing that platforms may use data from their “origin markets” in order to enter and dominate adjacent ones:

First, as we already mentioned, data collected in the origin market can be used, once the enveloper has entered the target market, to provide products more efficiently in the target market. Second, data collected in the origin market can be used to reduce the asymmetric information to which an entrant is typically subject when deciding to invest (for example, in R&D) to enter a new market. For instance, a search engine could be able to predict new trends from consumer searches and therefore face less uncertainty in product design.^[27]

This possibility is also implicit in Hagiu and Wright’s paper.[28] Indeed, the authors’ theoretical model rests on an important distinction between within-user data advantages (that is, having access to more data about a given user) and across-user data advantages (information gleaned from having access to a wider user base). In both cases, there is an implicit assumption that platforms may use data from one service to gain an advantage in another market (because what matters is information about aggregate or individual user preferences, regardless of its origin).

Our review of the economic evidence suggests that several scholars have, with varying degrees of certainty, raised the possibility that incumbents may leverage data advantages to stifle competitors in their primary market or in adjacent ones (be it via merger or organic growth). As we explain below, however, there is ultimately little evidence to support such claims. Policymakers have, however, been keenly receptive to these limited theoretical findings, basing multiple decisions on these theories, often with little consideration given to the caveats that accompany them.^[29]

Indeed, it is remarkable that, in its section on “[t]he data advantage for incumbents,” the “Furman Report” created for the UK government cited only two empirical economic studies, and they offer directly contradictory conclusions with respect to the question of the strength of data advantages.[30] Nevertheless, the Furman Report concludes that data “may confer a form of unmatchable advantage on the incumbent business, making successful rivalry less likely,”[31] and adopts without reservation “convincing” evidence from non-economists that have no apparent empirical basis.[32]

In the Google/Fitbit merger proceedings, the European Commission found that the combination of data from Google services with that of Fitbit devices would reduce competition in advertising markets:

Giving [sic] the large amount of data already used for advertising purposes that Google holds, the increase in Google’s data collection capabilities, which goes beyond the mere number of active users for which Fitbit has been collecting data so far, the Transaction is likely to have a negative impact on the development of an unfettered competition in the markets for online advertising.^[33]

As a result, the Commission cleared the merger on the condition that Google refrain from using data from Fitbit devices for its advertising platform.^[34] The Commission will likely focus on similar issues during its ongoing investigation of Microsoft’s investment into OpenAI.[35]

Along similar lines, the FTC’s complaint to enjoin Meta’s purchase of a virtual-reality (VR) fitness app called “Within” relied, among other things, on the fact that Meta could leverage its data about VR-user behavior to inform its decisions and potentially outcompete rival VR-fitness apps: “Meta’s control over the Quest platform also gives it unique access to VR user data, which it uses to inform strategic decisions.”^[36]

The DOJ’s twin cases against Google also implicate data leveraging and data barriers to entry. The agency’s adtech complaint charges that “Google intentionally exploited its massive trove of user data to further entrench its monopoly across the digital advertising industry.”[37] Similarly, in its search complaint, the agency argues that:

Google’s anticompetitive practices are especially pernicious because they deny rivals scale to compete effectively. General search services, search advertising, and general search text advertising require complex algorithms that are constantly learning which organic results and ads best respond to user queries; the volume, variety, and velocity of data accelerates the automated learning of search and search advertising algorithms.[38]

Finally, updated merger guidelines published in recent years by several competition enforcers cite the acquisition of data as a potential source of competition concerns. For instance, the FTC and DOJ’s newly published guidelines state that “acquiring data that helps facilitate matching, sorting, or prediction services may enable the platform to weaken rival platforms by denying them that data.”^[39] Likewise, the UK Competition and Markets Authority (CMA) warns against incumbents acquiring firms in order to obtain their data and foreclose other rivals:

Incentive to foreclose rivals…

7.19(e) Particularly in complex and dynamic markets, firms may not focus on short term margins but may pursue other objectives to maximise their long-run profitability, which the CMA may consider. This may include… obtaining access to customer data….^[40]

In short, competition authorities around the globe have been taking an increasingly aggressive stance on data-network effects. Among the ways this has manifested is in basing enforcement decisions on fears that data collected by one platform might confer a decisive competitive advantage in adjacent markets. Unfortunately, these concerns rest on little to no empirical evidence, either in the economic literature or the underlying case records.

III. Data-Incumbency Advantages in Generative-AI Markets

Given the assertions canvassed in the previous section, it would be reasonable to assume that firms such as Google, Meta, and Amazon should be in pole position to dominate the burgeoning market for generative AI. After all, these firms have not only been at the forefront of the field for the better part of a decade, but they also have access to vast troves of data, the likes of which their rivals could only dream when they launched their own services. Thus, the authors of the Furman Report caution that “to the degree that the next technological revolution centres around artificial intelligence and machine learning, then the companies most able to take advantage of it may well be the existing large companies because of the importance of data for the successful use of these tools.”[41]

To date, however, this is not how things have unfolded—although it bears noting these markets remain in flux and the competitive landscape is susceptible to change. The first significantly successful generative-AI service was arguably not from either Meta—which had been working on chatbots for years and had access to, arguably, the world’s largest database of actual chats—or Google. Instead, the breakthrough came from a previously unknown firm called OpenAI.

OpenAI’s ChatGPT service currently holds an estimated 60% of the market (though reliable numbers are somewhat elusive).[42] It broke the record for the fastest online service to reach 100 million users (in only a couple of months), more than four times faster than the previous record holder, TikTok.^[43] Based on Google Trends data, ChatGPT is nine times more popular worldwide than Google’s own Bard service, and 14 times more popular in the United States.^[44] In April 2023, ChatGPT reportedly registered 206.7 million unique visitors, compared to 19.5 million for Google’s Bard.^[45] In short, at the time we are writing, ChatGPT appears to be the most popular chatbot. The entry of large players such as Google Bard or Meta AI appear to have had little effect thus far on its market position.^[46]

The picture is similar in the field of AI-image generation. As of August 2023, Midjourney, Dall-E, and Stable Diffusion appear to be the three market leaders in terms of user visits.^[47] This is despite competition from the likes of Google and Meta, who arguably have access to unparalleled image and video databases by virtue of their primary platform activities.^[48]

This raises several crucial questions: how have these AI upstarts managed to be so successful, and is their success just a flash in the pan before Web 2.0 giants catch up and overthrow them? While we cannot answer either of these questions dispositively, we offer what we believe to be some relevant observations concerning the role and value of data in digital markets.

A first important observation is that empirical studies suggest that data exhibits diminishing marginal returns. In other words, past a certain point, acquiring more data does not confer a meaningful edge to the acquiring firm. As Catherine Tucker put it following a review of the literature: “Empirically there is little evidence of economies of scale and scope in digital data in the instances where one would expect to find them.”^[49]

Likewise, following a survey of the empirical literature on this topic, Geoffrey Manne and Dirk Auer conclude that:

Available evidence suggests that claims of “extreme” returns to scale in the tech sector are greatly overblown. Not only are the largest expenditures of digital platforms unlikely to become proportionally less important as output increases, but empirical research strongly suggests that even data does not give rise to increasing returns to scale, despite routinely being cited as the source of this effect.^[50]

In other words, being the firm with the most data appears to be far less important than having enough data. This lower bar may be accessible to far more firms than one might initially think possible. And obtaining enough data could become even easier—that is, the volume of required data could become even smaller—with technological progress. For instance, synthetic data may provide an adequate substitute to real-world data,[51] or may even outperform real-world data.[52] As Thibault Schrepel and Alex Pentland surmise:

[A]dvances in computer science and analytics are making the amount of data less relevant every day. In recent months, important technological advances have allowed companies with small data sets to compete with larger ones.^[53]

Indeed, past a certain threshold, acquiring more data might not meaningfully improve a service, where other improvements (such as better training methods or data curation) could have a large impact. In fact, there is some evidence that excessive data impedes a service’s ability to generate results appropriate for a given query: “[S]uperior model performance can often be achieved with smaller, high-quality datasets than massive, uncurated ones. Data curation ensures that training datasets are devoid of noise, irrelevant instances, and duplications, thus maximizing the efficiency of every training iteration.”[54]

Consider, for instance, a user who wants to generate an image of a basketball. Using a model trained on an indiscriminate range and number of public photos in which a basketball appears surrounded by copious other image data, the user may end up with an inordinately noisy result. By contrast, a model trained with a better method on fewer, more carefully selected images, could readily yield far superior results.[55] In one important example:

[t]he model’s performance is particularly remarkable, given its small size. “This is not a large language model trained on the whole Internet; this is a relatively small transformer trained for these tasks,” says Armando Solar-Lezama, a computer scientist at the Massachusetts Institute of Technology, who was not involved in the new study…. The finding implies that instead of just shoving ever more training data into machine-learning models, a complementary strategy might be to offer AI algorithms the equivalent of a focused linguistics or algebra class.[56]

Platforms’ current efforts are thus focused on improving the mathematical and logical reasoning of large language models (LLMs), rather than maximizing training datasets.^[57] Two points stand out. The first is that firms like OpenAI rely largely on publicly available datasets—such as GSM8K—to train their LLMs.^[58] Second, the real challenge to create cutting-edge AI is not so much in collecting data, but rather in creating innovative AI-training processes and architectures:

[B]uilding a truly general reasoning engine will require a more fundamental architectural innovation. What’s needed is a way for language models to learn new abstractions that go beyond their training data and have these evolving abstractions influence the model’s choices as it explores the space of possible solutions.

We know this is possible because the human brain does it. But it might be a while before OpenAI, DeepMind, or anyone else figures out how to do it in silicon.^[59]

Furthermore, it is worth noting that the data most relevant to startups in a given market may not be those data held by large incumbent platforms in other markets, but rather data specific to the market in which the startup is active or, even better, to the given problem it is attempting to solve:

As Andres Lerner has argued, if you wanted to start a travel business, the data from Kayak or Priceline would be far more relevant. Or if you wanted to start a ride-sharing business, data from cab companies would be more useful than the broad, market-cross-cutting profiles Google and Facebook have. Consider companies like Uber, Lyft and Sidecar that had no customer data when they began to challenge established cab companies that did possess such data. If data were really so significant, they could never have competed successfully. But Uber, Lyft and Sidecar have been able to effectively compete because they built products that users wanted to use—they came up with an idea for a better mousetrap. The data they have accrued came after they innovated, entered the market and mounted their successful challenges—not before.[60]

The bottom line is that data is not the be-all and end-all that many in competition circles make it out to be. While data may often confer marginal benefits, there is little sense these are ultimately decisive.[61] As a result, incumbent platforms’ access to vast numbers of users and data in their primary markets might only marginally affect their AI competitiveness.

A related observation is that firms’ capabilities and other features of their products arguably play a more important role than the data they own.[62] Examples of this abound in digital markets. Google overthrew Yahoo, despite initially having access to far fewer users and far less data; Google and Apple overcame Microsoft in the smartphone OS market despite having comparatively tiny ecosystems (at the time) to leverage; and TikTok rose to prominence despite intense competition from incumbents like Instagram, which had much larger user bases. In each of these cases, important product-design decisions (such as the PageRank algorithm, recognizing the specific needs of mobile users,^[63] and TikTok’s clever algorithm) appear to have played a far more significant role than initial user and data endowments (or lack thereof).

All of this suggests that the early success of OpenAI likely has more to do with its engineering decisions than what data it did (or did not) own. Going forward, OpenAI and its rivals’ ability to offer and monetize compelling stores offering custom versions of their generative-AI technology will arguably play a much larger role than (and contribute to) their ownership of data.[64] In other words, the ultimate challenge is arguably to create a valuable platform, of which data ownership is a consequence, but not a cause.

It is also important to note that, in those instances where it is valuable, data does not just fall from the sky. Instead, it is through smart business and engineering decisions that firms can generate valuable information (which does not necessarily correlate with owning more data).

For instance, OpenAI’s success with ChatGPT is often attributed to its more efficient algorithms and training models, which arguably have enabled the service to improve more rapidly than its rivals.^[65] Likewise, the ability of firms like Meta and Google to generate valuable data for advertising arguably depends more on design decisions that elicit the right data from users, rather than the raw number of users in their networks.

Put differently, setting up a business so as to extract and organize the right information is more important than simply owning vast troves of data.[66] Even in those instances where high-quality data is an essential parameter of competition, it does not follow that having vaster databases or more users on a platform necessarily leads to better information for the platform.

Indeed, if data ownership consistently conferred a significant competitive advantage, these new firms would not be where they are today. This does not mean that data is worthless, of course. Rather, it means that competition authorities should not assume that the mere possession of data is a dispositive competitive advantage, absent compelling empirical evidence to support such a finding. In this light, the current wave of decisions and competition-policy pronouncements that rely on data-related theories of harm are premature.

IV. Five Key Takeaways: Reconceptualizing the Role of Data in Generative-AI Competition

As we explain above, data (network effects) are not the source of barriers to entry that they are sometimes made out to be. The picture is far more nuanced. Indeed, as economist Andres Lerner demonstrated almost a decade ago (and the assessment is only truer today):

Although the collection of user data is generally valuable for online providers, the conclusion that such benefits of user data lead to significant returns to scale and to the entrenchment of dominant online platforms is based on unsupported assumptions. Although, in theory, control of an “essential” input can lead to the exclusion of rivals, a careful analysis of real-world evidence indicates that such concerns are unwarranted for many online businesses that have been the focus of the “big data” debate.[67]

While data can be an important part of the competitive landscape, incumbents’ data advantages are far less pronounced than today’s policymakers commonly assume. In that respect, five main lessons emerge:

1. Data can be (very) valuable, but beyond a certain threshold, those benefits tend to diminish. In other words, having the most data is less important than having enough;
2. The ability to generate valuable information does not depend on the number of users or the amount of data a platform has previously acquired;
3. The most important datasets are not always proprietary;
4. Technological advances and platforms’ engineering decisions affect their ability to generate valuable information, and this effect swamps effects stemming from the amount of data they own; and
5. How platforms use data is arguably more important than what data or how much data they own.

These lessons have important ramifications for competition-policy debates over the competitive implications of data in technologically evolving areas.

First, it is not surprising that startups, rather than incumbents, have taken an early lead in generative AI (and in Web 2.0 before it). After all, if data-incumbency advantages are small or even nonexistent, then smaller and more nimble players may have an edge over established tech platforms. This is all the more likely given that, despite significant efforts, the biggest tech platforms were unable to offer compelling generative-AI chatbots and image-generation services before the emergence of ChatGPT, Dall-E, Midjourney, etc.

This failure suggests that, in a process akin to Clayton Christensen’s “innovator’s dilemma,”[68] something about the incumbent platforms’ existing services and capabilities was holding them back in those markets. Of course, this does not necessarily mean that those same services or capabilities could not become an advantage when the generative-AI market starts addressing issues of monetization and scale.[69] But it does mean that assumptions about a firm’s market power based on its possession of data are off the mark.

Another important implication is that, paradoxically, policymakers’ efforts to prevent Web 2.0 platforms from competing freely in generative AI markets may ultimately backfire and lead to less, not more, competition. Indeed, OpenAI is currently acquiring a sizeable lead in generative AI. While competition authorities might like to think that other startups will emerge and thrive in this space, it is important not to confuse desires with reality. While there currently exists a vibrant AI-startup ecosystem, there is at least a case to be made that the most significant competition for today’s AI leaders will come from incumbent Web 2.0 platforms—although nothing is certain at this stage. Policymakers should beware not to stifle that competition on the misguided assumption that competitive pressure from large incumbents is somehow less valuable to consumers than that which originates from smaller firms.

Finally, even if there were a competition-related market failure to be addressed in the field of generative AI (which is anything but clear), it is unclear that the remedies being contemplated would do more good than harm. Some of the solutions that have been put forward have highly ambiguous effects on consumer welfare. Scholars have shown that, e.g., mandated data sharing—a solution championed by EU policymakers, among others—may sometimes dampen competition in generative-AI markets.[70] This is also true of legislation like the General Data Protection Regulation (GDPR), which makes it harder for firms to acquire more data about consumers—assuming such data is, indeed, useful to generative-AI services.[71]

In sum, it is a flawed understanding of the economics and practical consequences of large agglomerations of data that lead competition authorities to believe that data-incumbency advantages are likely to harm competition in generative AI markets—or even in the data-intensive Web 2.0 markets that preceded them. Indeed, competition or regulatory intervention to “correct” data barriers and data network and scale effects is liable to do more harm than good.

 

[1] See, e.g., Michael Chui, et al., The Economic Potential of Generative AI: The Next Productivity Frontier, McKinsey (Jun. 14, 2023), https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-AI-the-next-productivity-frontier.

[2] See, e. g., Zhuoran Qiao, Weili Nie, Arash Vahdat, Thomas F. Miller III, & Animashree Anandkumar, State-Specific Protein–Ligand Complex Structure Prediction with a Multiscale Deep Generative Model, 6 Nature Machine Intelligence, 195-208 (2024); see also, Jaemin Seo, Sang Kyeun Kim, Azarakhsh Jalalvand, Rory Conlin, Andrew Rothstein, Joseph Abbate, Keith Erickson, Josiah Wai, Ricardo Shousha, & Egemen Kolemen, Avoiding Fusion Plasma Tearing Instability with Deep Reinforcement Learning, 626 Nature, 746-751 (2024).

[3] See, e.g., Press Release, Commission Launches Calls for Contributions on Competition in Virtual Worlds and Generative AI, European Commission (Jan. 9, 2024), https://ec.europa.eu/commission/presscorner/detail/en/IP_24_85.

[4] Nathan Newman, Taking on Google’s Monopoly Means Regulating Its Control of User Data, Huffington Post (Sep. 24, 2013), http://www.huffingtonpost.com/nathan-newman/taking-on-googlesmonopol_b_3980799.html.

[5] See, e.g., Lina Khan & K. Sabeel Rahman, Restoring Competition in the U.S. Economy, in Untamed: How to Check Corporate, Financial, and Monopoly Power (Nell Abernathy, Mike Konczal, & Kathryn Milani, eds., 2016), at 23 (“From Amazon to Google to Uber, there is a new form of economic power on display, distinct from conventional monopolies and oligopolies…, leverag[ing] data, algorithms, and internet-based technologies… in ways that could operate invisibly and anticompetitively.”); Mark Weinstein, I Changed My Mind—Facebook Is a Monopoly, Wall St. J. (Oct. 1, 2021), https://www.wsj.com/articles/facebook-is-monopoly-metaverse-users-advertising-platforms-competition-mewe-big-tech-11633104247 (“[T]he glue that holds it all together is Facebook’s monopoly over data…. Facebook’s data troves give it unrivaled knowledge about people, governments—and its competitors.”).

[6] See, generally, Abigail Slater, Why “Big Data” Is a Big Deal, The Reg. Rev. (Nov. 6, 2023), https://www.theregreview.org/2023/11/06/slater-why-big-data-is-a-big-deal; Amended Complaint at ¶36, United States v. Google, 1:20-cv-03010- (D.D.C. 2020); Complaint at ¶37, United States V. Google, 1:23-cv-00108 (E.D. Va. 2023), https://www.justice.gov/opa/pr/justice-department-sues-google-monopolizing-digital-advertising-technologies (“Google intentionally exploited its massive trove of user data to further entrench its monopoly across the digital advertising industry.”).

[7] See, e.g., Press Release, European Commission, supra note 3; Krysten Crawford, FTC’s Lina Khan Warns Big Tech over AI, SIEPR (Nov. 3, 2020), https://siepr.stanford.edu/news/ftcs-lina-khan-warns-big-tech-over-ai (“Federal Trade Commission Chair Lina Khan delivered a sharp warning to the technology industry in a speech at Stanford on Thursday: Antitrust enforcers are watching what you do in the race to profit from artificial intelligence.”) (emphasis added).

[8] See, e.g., John M. Newman, Antitrust in Digital Markets, 72 Vand. L. Rev. 1497, 1501 (2019) (“[T]he status quo has frequently failed in this vital area, and it continues to do so with alarming regularity. The laissez-faire approach advocated for by scholars and adopted by courts and enforcers has allowed potentially massive harms to go unchecked.”);
Bertin Martins, Are New EU Data Market Regulations Coherent and Efficient?, Bruegel Working Paper 21/23 (2023), https://www.bruegel.org/working-paper/are-new-eu-data-market-regulations-coherent-and-efficient (“Technical restrictions on access to and re-use of data may result in failures in data markets and data-driven services markets.”); Valéria Faure-Muntian, Competitive Dysfunction: Why Competition Law Is Failing in a Digital World, The Forum Network (Feb. 24, 2021), https://www.oecd-forum.org/posts/competitive-dysfunction-why-competition-law-is-failing-in-a-digital-world.

[9] See Rana Foroohar, The Great US-Europe Antitrust Divide, FT (Feb. 5, 2024), https://www.ft.com/content/065a2f93-dc1e-410c-ba9d-73c930cedc14.

[10] See, e.g., Press Release, European Commission, supra note 3.

[11] See infra, Section II. Commentators have also made similar claims; see, e.g., Ganesh Sitaram & Tejas N. Narechania, It’s Time for the Government to Regulate AI. Here’s How, Politico (Jan. 15, 2024) (“All that cloud computing power is used to train foundation models by having them “learn” from incomprehensibly huge quantities of data. Unsurprisingly, the entities that own these massive computing resources are also the companies that dominate model development. Google has Bard, Meta has LLaMa. Amazon recently invested $4 billion into one of OpenAI’s leading competitors, Anthropic. And Microsoft has a 49 percent ownership stake in OpenAI — giving it extraordinary influence, as the recent board struggles over Sam Altman’s role as CEO showed.”).

[12] Press Release, European Commission, supra note 3.

[13] Comment of U.S. Federal Trade Commission to the U.S. Copyright Office, Artificial Intelligence and Copyright, Docket No. 2023-6 (Oct. 30, 2023), at 4, https://www.ftc.gov/legal-library/browse/advocacy-filings/comment-federal-trade-commission-artificial-intelligence-copyright (emphasis added).

[14] See, e.g. Joe Caserta, Holger Harreis, Kayvaun Rowshankish, Nikhil Srinidhi, & Asin Tavakoli, The Data Dividend: Fueling Generative AI, McKinsey Digital (Sep. 15, 2023), https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-data-dividend-fueling-generative-ai (“Your data and its underlying foundations are the determining factors to what’s possible with generative AI.”).

[15] See, e.g., Tim Keary, Google DeepMind’s Achievements and Breakthroughs in AI Research, Techopedia (Aug. 11, 2023), https://www.techopedia.com/google-deepminds-achievements-and-breakthroughs-in-ai-research; See, e.g., Will Douglas Heaven, Google DeepMind Used a Large Language Model to Solve an Unsolved Math Problem, MIT Technology Review (Dec. 14, 2023), https://www.technologyreview.com/2023/12/14/1085318/google-deepmind-large-language-model-solve-unsolvable-math-problem-cap-set; see also, A Decade of Advancing the State-of-the-Art in AI Through Open Research, Meta (Nov. 30, 2023), https://about.fb.com/news/2023/11/decade-of-advancing-ai-through-open-research; see also, 200 Languages Within a Single AI Model: A Breakthrough in High-Quality Machine Translation, Meta, https://ai.meta.com/blog/nllb-200-high-quality-machine-translation (last visited Jan. 18, 2023).

[16] See, e.g., Jennifer Allen, 10 Years of Siri: The History of Apple’s Voice Assistant, Tech Radar (Oct. 4, 2021), https://www.techradar.com/news/siri-10-year-anniversary; see also Evan Selleck, How Apple Is Already Using Machine Learning and AI in iOS, Apple Insider (Nov. 20, 2023), https://appleinsider.com/articles/23/09/02/how-apple-is-already-using-machine-learning-and-ai-in-ios; see also, Kathleen Walch, The Twenty Year History Of AI At Amazon, Forbes (July 19, 2019), https://www.forbes.com/sites/cognitiveworld/2019/07/19/the-twenty-year-history-of-ai-at-amazon.

[17] See infra Section III.

[18] See, e.g., Cédric Argenton & Jens Prüfer, Search Engine Competition with Network Externalities, 8 J. Comp. L. & Econ. 73, 74 (2012).

[19] John M. Yun, The Role of Big Data in Antitrust, in The Global Antitrust Institute Report on the Digital Economy (Joshua D. Wright & Douglas H. Ginsburg, eds., Nov. 11, 2020) at 233, https://gaidigitalreport.com/2020/08/25/big-data-and-barriers-to-entry/#_ftnref50; see also, e.g., Robert Wayne Gregory, Ola Henfridsson, Evgeny Kaganer, & Harris Kyriakou, The Role of Artificial Intelligence and Data Network Effects for Creating User Value, 46 Acad. of Mgmt. Rev. 534 (2020), final pre-print version at 4, http://wrap.warwick.ac.uk/134220) (“A platform exhibits data network effects if, the more that the platform learns from the data it collects on users, the more valuable the platform becomes to each user.”); see also, Karl Schmedders, José Parra-Moyano, & Michael Wade, Why Data Aggregation Laws Could be the Answer to Big Tech Dominance, Silicon Republic (Feb. 6, 2024), https://www.siliconrepublic.com/enterprise/data-ai-aggregation-laws-regulation-big-tech-dominance-competition-antitrust-imd.

[20] Nathan Newman, Search, Antitrust, and the Economics of the Control of User Data, 31 Yale J. Reg. 401, 409 (2014) (emphasis added); see also id. at 420 & 423 (“While there are a number of network effects that come into play with Google, [“its intimate knowledge of its users contained in its vast databases of user personal data”] is likely the most important one in terms of entrenching the company’s monopoly in search advertising…. Google’s overwhelming control of user data… might make its dominance nearly unchallengeable.”).

[21] See also Yun, supra note 19 at 229 (“[I]nvestments in big data can create competitive distance between a firm and its rivals, including potential entrants, but this distance is the result of a competitive desire to improve one’s product.”).

^[22] For a review of the literature on increasing returns to scale in data (this topic is broader than data-network effects) see Geoffrey Manne & Dirk Auer, Antitrust Dystopia and Antitrust Nostalgia: Alarmist Theories of Harm in Digital Markets and Their Origins, 28 Geo Mason L. Rev. 1281, 1344 (2021).

[23] Andrei Hagiu & Julian Wright, Data-Enabled Learning, Network Effects, and Competitive Advantage, 54 RAND J. Econ. 638 (2023).

^[24] Id. at 639. The authors conclude that “Data-enabled learning would seem to give incumbent firms a competitive advantage. But how strong is this advantage and how does it differ from that obtained from more traditional mechanisms…”.

^[25] Id.

^[26] Bruno Jullien & Wilfried Sand-Zantman, The Economics of Platforms: A Theory Guide for Competition Policy, 54 Info. Econ. & Pol’y 10080, 101031 (2021).

^[27] Daniele Condorelli & Jorge Padilla, Harnessing Platform Envelopment in the Digital World, 16 J. Comp. L. & Pol’y 143, 167 (2020).

[28] See Hagiu & Wright, supra note 23.

^[29] For a summary of these limitations, see generally Catherine Tucker, Network Effects and Market Power: What Have We Learned in the Last Decade?, Antitrust (2018) at 72, available at https://sites.bu.edu/tpri/files/2018/07/tucker-network-effects-antitrust2018.pdf; see also Manne & Auer, supra note 22, at 1330.

[30] See Jason Furman, Diane Coyle, Amelia Fletcher, Derek McAuley, & Philip Marsden (Dig. Competition Expert Panel), Unlocking Digital Competition (2019) at 32-35 (“Furman Report”), available at https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/785547/unlocking_digital_competition_furman_review_web.pdf.

[31] Id. at 34.

[32] Id. at 35. To its credit, it should be noted, the Furman Report does counsel caution before mandating access to data as a remedy to promote competition. See id. at 75. That said, the Furman Report does maintain that such a remedy should certainly be on the table because “the evidence suggests that large data holdings are at the heart of the potential for some platform markets to be dominated by single players and for that dominance to be entrenched in a way that lessens the potential for competition for the market.” Id. In fact, the evidence does not show this.

^[33] Case COMP/M.9660 — Google/Fitbit, Commission Decision (Dec. 17, 2020) (Summary at O.J. (C 194) 7), available at https://ec.europa.eu/competition/mergers/cases1/202120/m9660_3314_3.pdf at 455.

^[34] Id. at 896.

[35] See Natasha Lomas, EU Checking if Microsoft’s OpenAI Investment Falls Under Merger Rules, TechCrunch (Jan. 9, 2024), https://techcrunch.com/2024/01/09/openai-microsoft-eu-merger-rules.

^[36] Amended Complaint at 11, Meta/Zuckerberg/Within, Fed. Trade Comm’n. (2022) (No. 605837), available at https://www.ftc.gov/system/files/ftc_gov/pdf/D09411%20-%20AMENDED%20COMPLAINT%20FILED%20BY%20COUNSEL%20SUPPORTING%20THE%20COMPLAINT%20-%20PUBLIC%20%281%29_0.pdf.

[37] Amended Complaint (D.D.C), supra note 6 at ¶37.

[38] Amended Complaint (E.D. Va), supra note 6 at ¶8.

^[39] Merger Guidelines, US Dep’t of Justice & Fed. Trade Comm’n (2023) at 25, available at https://www.ftc.gov/system/files/ftc_gov/pdf/2023_merger_guidelines_final_12.18.2023.pdf.

[40] Merger Assessment Guidelines, Competition and Mkts. Auth (2021) at  ¶7.19(e), available at https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1051823/MAGs_for_publication_2021_–_.pdf.

[41] Furman Report, supra note 30, at ¶4.

[42] See, e.g., Chris Westfall, New Research Shows ChatGPT Reigns Supreme in AI Tool Sector, Forbes (Nov. 16, 2023), https://www.forbes.com/sites/chriswestfall/2023/11/16/new-research-shows-chatgpt-reigns-supreme-in-ai-tool-sector/?sh=7de5de250e9c.

^[43] See Krystal Hu, ChatGPT Sets Record for Fastest-Growing User Base, Reuters (Feb. 2, 2023), https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01; Google: The AI Race Is On, App Economy Insights (Feb. 7, 2023), https://www.appeconomyinsights.com/p/google-the-ai-race-is-on.

^[44] See Google Trends, https://trends.google.com/trends/explore?date=today%205-y&q=%2Fg%2F11khcfz0y2,%2Fg%2F11ts49p01g&hl=en (last visited, Jan. 12, 2024) and https://trends.google.com/trends/explore?date=today%205-y&geo=US&q=%2Fg%2F11khcfz0y2,%2Fg%2F11ts49p01g&hl=en (last visited Jan. 12, 2024).

^[45] See David F. Carr, As ChatGPT Growth Flattened in May, Google Bard Rose 187%, Similarweb Blog (Jun. 5, 2023), https://www.similarweb.com/blog/insights/ai-news/chatgpt-bard.

^[46] See Press Release, Introducing New AI Experiences Across Our Family of Apps and Devices, Meta (Sep. 27, 2023), https://about.fb.com/news/2023/09/introducing-ai-powered-assistants-characters-and-creative-tools; Sundar Pichai, An Important Next Step on Our AI Journey, Google Keyword Blog (Feb. 6, 2023), https://blog.google/technology/ai/bard-google-ai-search-updates.

^[47] See Ion Prodan, 14 Million Users: Midjourney’s Statistical Success, Yon (Aug. 19, 2023), https://yon.fun/midjourney-statistics; see also Andrew Wilson, Midjourney Statistics: Users, Polls, & Growth [Oct 2023], ApproachableAI (Oct. 13, 2023), https://approachableai.com/midjourney-statistics.

^[48] See Hema Budaraju, New Ways to Get Inspired with Generative AI in Search, Google Keyword Blog (Oct. 12, 2023), https://blog.google/products/search/google-search-generative-ai-october-update; Imagine with Meta AI, Meta (last visited Jan. 12, 2024), https://imagine.meta.com.

^[49] Catherine Tucker, Digital Data, Platforms and the Usual [Antitrust] Suspects: Network Effects, Switching Costs, Essential Facility, 54 Rev. Indus. Org. 683, 686 (2019).

^[50] Manne & Auer, supra note 22, at 1345.

[51] See, e.g., Stefanie Koperniak, Artificial Data Give the Same Results as Real Data—Without Compromising Privacy, MIT News (Mar. 3, 2017), https://news.mit.edu/2017/artificial-data-give-same-results-as-real-data-0303 (“[Authors] describe a machine learning system that automatically creates synthetic data—with the goal of enabling data science efforts that, due to a lack of access to real data, may have otherwise not left the ground. While the use of authentic data can cause significant privacy concerns, this synthetic data is completely different from that produced by real users—but can still be used to develop and test data science algorithms and models.”).

[52] See, e.g., Rachel Gordon, Synthetic Imagery Sets New Bar in AI Training Efficiency, MIT News (Nov. 20, 2023), https://news.mit.edu/2023/synthetic-imagery-sets-new-bar-ai-training-efficiency-1120 (“By using synthetic images to train machine learning models, a team of scientists recently surpassed results obtained from traditional ‘real-image’ training methods.).

[53] Thibault Schrepel & Alex ‘Sandy’ Pentland, Competition Between AI Foundation Models: Dynamics and Policy Recommendations, MIT Connection Science Working Paper (Jun. 2023), at 8.

[54] Igor Susmelj, Optimizing Generative AI: The Role of Data Curation, Lightly (last visited Jan. 15, 2024), https://www.lightly.ai/post/optimizing-generative-ai-the-role-of-data-curation.

[55] See, e.g., Xiaoliang Dai, et al., Emu: Enhancing Image Generation Models Using Photogenic Needles in a Haystack, ArXiv (Sep. 27, 2023) at 1, https://ar5iv.labs.arxiv.org/html/2309.15807 (“[S]upervised fine-tuning with a set of surprisingly small but extremely visually appealing images can significantly improve the generation quality.”); see also, Hu Xu, et al., Demystifying CLIP Data, ArXiv (Sep. 28, 2023), https://arxiv.org/abs/2309.16671.

[56] Lauren Leffer, New Training Method Helps AI Generalize like People Do, Sci. Am. (Oct. 26, 2023), https://www.scientificamerican.com/article/new-training-method-helps-ai-generalize-like-people-do (discussing Brendan M. Lake & Marco Baroni, Human-Like Systematic Generalization Through a Meta-Learning Neural Network, 623 Nature 115 (2023)).

^[57] Timothy B. Lee, The Real Research Behind the Wild Rumors about OpenAI’s Q* Project, Ars Technica (Dec. 8, 2023), https://arstechnica.com/ai/2023/12/the-real-research-behind-the-wild-rumors-about-openais-q-project.

^[58] Id.; see also GSM8K, Papers with Code (last visited Jan. 18, 2023), available at https://paperswithcode.com/dataset/gsm8k; MATH Dataset, GitHub (last visited Jan. 18, 2024), available at https://github.com/hendrycks/math.

^[59] Lee, supra note 57.

[60] Geoffrey Manne & Ben Sperry, Debunking the Myth of a Data Barrier to Entry for Online Services, Truth on the Market (Mar. 26, 2015), https://truthonthemarket.com/2015/03/26/debunking-the-myth-of-a-data-barrier-to-entry-for-online-services (citing Andres V. Lerner, The Role of ‘Big Data’ in Online Platform Competition (Aug. 26, 2014), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2482780.).

[61] See Catherine Tucker, Digital Data as an Essential Facility: Control, CPI Antitrust Chron. (Feb. 2020), at 11 (“[U]ltimately the value of data is not the raw manifestation of the data itself, but the ability of a firm to use this data as an input to insight.”).

[62] Or, as John Yun puts it, data is only a small component of digital firms’ production function. See Yun, supra note 19, at 235 (“Second, while no one would seriously dispute that having more data is better than having less, the idea of a data-driven network effect is focused too narrowly on a single factor improving quality. As mentioned in supra Section I.A, there are a variety of factors that enter a firm’s production function to improve quality.”).

^[63] Luxia Le, The Real Reason Windows Phone Failed Spectacularly, History–Computer (Aug. 8, 2023), https://history-computer.com/the-real-reason-windows-phone-failed-spectacularly.

[64] Introducing the GPT Store, Open AI (Jan. 10, 2024), https://openai.com/blog/introducing-the-gpt-store.

^[65] See Michael Schade, How ChatGPT and Our Language Models are Developed, OpenAI, https://help.openai.com/en/articles/7842364-how-chatgpt-and-our-language-models-are-developed; Sreejani Bhattacharyya, Interesting Innovations from OpenAI in 2021, AIM (Jan. 1, 2022), https://analyticsindiamag.com/interesting-innovations-from-openai-in-2021; Danny Hernadez & Tom B. Brown, Measuring the Algorithmic Efficiency of Neural Networks, ArXiv (May 8, 2020), https://arxiv.org/abs/2005.04305.

[66] See Yun, supra note 19 at 235 (“Even if data is primarily responsible for a platform’s quality improvements, these improvements do not simply materialize with the presence of more data—which differentiates the idea of data-driven network effects from direct network effects. A firm needs to intentionally transform raw, collected data into something that provides analytical insights. This transformation involves costs including those associated with data storage, organization, and analytics, which moves the idea of collecting more data away from a strict network effect to more of a ‘data opportunity.’”).

[67] Lerner, supra note 60, at 4-5 (emphasis added).

[68] See Clayton M. Christensen, The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (2013).

[69] See David J. Teece, Dynamic Capabilities and Strategic Management: Organizing for Innovation and Growth (2009).

[70] See Hagiu & Wright, supra note 23, at 23 (“We use our dynamic framework to explore how data sharing works: we find that it in-creases consumer surplus when one firm is sufficiently far ahead of the other by making the laggard more competitive, but it decreases consumer surplus when the firms are sufficiently evenly matched by making firms compete less aggressively, which in our model means subsidizing consumers less.”); see also Lerner, supra note 60.

[71] See, e.g., Hagiu & Wright, id. (“We also use our model to highlight an unintended consequence of privacy policies. If such policies reduce the rate at which firms can extract useful data from consumers, they will tend to increase the incumbent’s competitive advantage, reflecting that the entrant has more scope for new learning and so is affected more by such a policy.”); Jian Jia, Ginger Zhe Jin, & Liad Wagman, The Short-Run Effects of the General Data Protection Regulation on Technology Venture Investment, 40 Marketing Sci. 593 (2021) (finding GDPR reduced investment in new and emerging technology firms, particularly in data-related ventures); James Campbell, Avi Goldfarb, & Catherine Tucker, Privacy Regulation and Market Structure, 24 J. Econ. & Mgmt. Strat. 47 (2015) (“Consequently, rather than increasing competition, the nature of transaction costs implied by privacy regulation suggests that privacy regulation may be anti-competitive.”).

Dear Justice Guerrero and Associate Justices,

In accordance with California Rule of Court 8.500(g), we are writing to urge the Court to grant the Petition for Review filed by Petitioner Gilead Sciences, Inc. (“Petitioner” or “Gilead”) on February 21, 2024, in the above-captioned matter.

We agree with Petitioner that the Court of Appeal’s finding of a duty of reasonable care in this case “is such a seismic change in the law and so fundamentally wrong, with such grave consequences, that this Court’s review is imperative.” (Pet. 6.) The unprecedented duty of care put forward by the Court of Appeal—requiring prescription drug manufacturers to exercise reasonable care toward users of a current drug when deciding when to bring a new drug to market (Op. 11)—would have far-reaching, harmful implications for innovation that the Court of Appeal failed properly to weigh.

If upheld, this new duty of care would significantly disincentivize pharmaceutical innovation by allowing juries to second-guess complex scientific and business decisions about which potential drugs to prioritize and when to bring them to market. The threat of massive liability simply for not developing a drug sooner would make companies reluctant to invest the immense resources needed to bring new treatments to patients. Perversely, this would deprive the public of lifesaving and less costly new medicines. And the prospective harm from the Court of Appeal’s decision is not limited only to the pharmaceutical industry.

We urge the Court to grant the Petition for Review and to hold that innovative firms do not owe the users of current products a “duty to innovate” or a “duty to market”—that is, that firms cannot be held liable to users of a current product for development or commercialization decisions on the basis that those decisions could have facilitated the introduction of a less harmful, alternative product.

Interest of Amicus Curiae

The International Center for Law & Economics (“ICLE”) is a nonprofit, non-partisan global research and policy center aimed at building the intellectual foundations for sensible, economically grounded policy. ICLE promotes the use of law and economics methodologies and economic learning to inform policy debates. It also has longstanding expertise in evaluating law and policy relating to innovation and the legal environment facing commercial activity. In this letter, we wish to briefly highlight some of the crucial considerations concerning the effect on innovation incentives that we believe would arise from the Court of Appeal’s ruling in this case.[1]

The Court of Appeal’s Duty of Care Standard Would Impose Liability Without Requiring Actual “Harm”

The Court of Appeal’s ruling marks an unwarranted departure from decades of products-liability law requiring plaintiffs to prove that the product that injured them was defective. Expanding liability to products never even sold is an unprecedented, unprincipled, and dangerous approach to product liability. Plaintiffs’ lawyers may seek to apply this new theory to many other beneficial products, arguing manufacturers should have sold a superior alternative sooner. This would wreak havoc on innovation across industries.

California Civil Code § 1714 does not impose liability for “fail[ing] to take positive steps to benefit others,” (Brown v. USA Taekwondo (2021) 11 Cal.5th 204, 215), and Plaintiffs did not press a theory that the medicine they received was defective. Moreover, the product included all the warnings required by federal and state law. Thus, Plaintiffs’ case—as accepted by the Court of Appeal—is that they consumed a product authorized by the FDA, that they were fully aware of its potential side effects, but maybe they would have had fewer side effects had Gilead made the decision to accelerate (against some indefinite baseline) the development of an alternative medicine. To call this a speculative harm is an understatement, and to dismiss Gilead’s conduct as unreasonable because motivated by a crass profit motive, (Op. at 32), elides many complicated facts that belie such a facile assertion.

A focus on the narrow question of profits for a particular drug misunderstands the inordinate complexity of pharmaceutical development and risks seriously impeding the rate of drug development overall. Doing so

[over-emphasizes] the recapture of “excess” profits on the relatively few highly profitable products without taking into account failures or limping successes experienced on the much larger number of other entries. If profits were held to “reasonable” levels on blockbuster drugs, aggregate profits would almost surely be insufficient to sustain a high rate of technological progress. . . . If in addition developing a blockbuster is riskier than augmenting the assortment of already known molecules, the rate at which important new drugs appear could be retarded significantly. Assuming that important new drugs yield substantial consumers’ surplus untapped by their developers, consumers would lose along with the drug companies. Should a tradeoff be required between modestly excessive prices and profits versus retarded technical progress, it would be better to err on the side of excessive profits. (F. M. Scherer, Pricing, Profits, and Technological Progress in the Pharmaceutical Industry, 7 J. Econ. Persp. 97, 113 (1993)).

Indeed, Plaintiffs’ claim on this ground is essentially self-refuting. If the “superior” product they claim was withheld for “profit” reasons was indeed superior, then Plaintiffs could have expected to make a superior return on that product. Thus, Plaintiffs claim they were allegedly “harmed” by not having access to a product that Petitioners were not yet ready to market, even though Petitioners had every incentive to release a potentially successful alternative as soon as possible, subject to a complex host of scientific and business considerations affecting the timing of that decision.

Related, the Court of Appeal’s decision rests on the unfounded assumption that Petitioner “knew” TAF was safer than TDF after completing Phase I trials. This ignores the realities of the drug development process and the inherent uncertainty of obtaining FDA approval, even after promising early results. Passing Phase I trials, which typically involve a small number of healthy volunteers, is a far cry from having a marketable drug. According to the Biotechnology Innovation Organization, only 7.9% of drugs that enter Phase I trials ultimately obtain FDA approval.[2] (Biotechnology Innovation Organization, Clinical Development Success Rates and Contributing Factors 2011-2020, Fig. 8b (2021), available at https://perma.cc/D7EY-P22Q.) Even after Phase II trials, which assess efficacy and side effects in a larger patient population, the success rate is only about 15.1%. (Id.) Thus, at the time Gilead decided to pause TAF development, it faced significant uncertainty about whether TAF would ever reach the market, let alone ultimately prove safer than TDF.

Moreover, the clock on Petitioner’s patent exclusivity for TAF was ticking throughout the development process. Had Petitioner “known” that TAF was a safer and more effective drug, it would have had every incentive to bring it to market as soon as possible to maximize the period of patent protection and the potential to recoup its investment. The fact that Petitioner instead chose to focus on TDF strongly suggests that it did not have the level of certainty the Court of Appeal attributed to it.

Although conventional wisdom has often held otherwise, economists generally dispute the notion that companies have an incentive to unilaterally suppress innovation for economic gain.

While rumors long have circulated about the suppression of a new technology capable of enabling automobiles to average 100 miles per gallon or some new device capable of generating electric power at a fraction of its current cost, it is rare to uncover cases where a worthwhile technology has been suppressed altogether. (John J. Flynn, Antitrust Policy, Innovation Efficiencies, and the Suppression of Technology, 66 Antitrust L.J. 487, 490 (1998)).

Calling such claims “folklore,” the economists Armen Alchian and William Allen note that, “if such a [technology] did exist, it could be made and sold at a price reflecting the value of [the new technology], a net profit to the owner.” (Armen A. Alchian & William R. Allen, Exchange & Production: Competition, Coordination, & Control (1983), at 292). Indeed, “even a monopolist typically will have an incentive to adopt an unambiguously superior technology.” (Joel M. Cohen and Arthur J. Burke, An Overview of the Antitrust Analysis of Suppression of Technology, 66 Antitrust L.J. 421, 429 n. 28 (1998)). While nominal suppression of technology can occur for a multitude of commercial and technological reasons, there is scant evidence that doing so coincides with harm to consumers, except where doing so affirmatively interferes with market competition under the antitrust laws—a claim not advanced here.

One reason the tort system is inapt for second-guessing commercial development and marketing decisions is that those decisions may be made for myriad reasons that do not map onto the specific safety concern of a products-liability action. For example, in the 1930s, AT&T abandoned the commercial development of magnetic recording “for ideological reasons. . . . Management feared that availability of recording devices would make customers less willing to use the telephone system and so undermine the concept of universal service.” (Mark Clark, Suppressing Innovation: Bell Laboratories and Magnetic Recording, 34 Tech. & Culture 516, 520-24 (1993)). One could easily imagine arguments that coupling telephones and recording devices would promote safety. But the determination of whether safety or universal service (and the avoidance of privacy invasion) was a “better” basis for deciding whether to pursue the innovation is not within the ambit of tort law (nor the capability of a products-liability jury). And yet, it would necessarily become so if the Court of Appeal’s decision were to stand.

A Proper Assessment of Public Policy Would Cut Strongly Against Adoption of the Court of Appeal’s Holding

The Court of Appeal notes that “a duty that placed manufacturers ‘under an endless obligation to pursue ever-better new products or improvements to existing products’ would be unworkable and unwarranted,” (Op. 10), yet avers that “plaintiffs are not asking us to recognize such a duty” because “their negligence claim is premised on Gilead’s possession of such an alternative in TAF; they complain of Gilead’s knowing and intentionally withholding such a treatment….” (Id).

From an economic standpoint, this is a distinction without a difference.

Both a “duty to invent” and a “duty to market” what is already invented would increase the cost of bringing any innovative product to market by saddling the developer with an expected additional (and unavoidable) obligation as a function of introducing the initial product, differing only perhaps by degree. Indeed, a “duty to invent” could conceivably be more socially desirable because in that case a firm could at least avoid liability by undertaking the process of discovering new products (a socially beneficial activity), whereas the “duty to market” espoused by the Court of Appeal would create only the opposite incentive—the incentive never to gain knowledge of a superior product on the basis of which liability might attach.[3]

And public policy is relevant. This Court in Brown v. Superior Court, (44 Cal. 3d 1049 (1988)), worried explicitly about the “[p]ublic policy” implications of excessive liability rules for the provision of lifesaving drugs. (Id. at 1063-65). As the Court in Brown explained, drug manufacturers “might be reluctant to undertake research programs to develop some pharmaceuticals that would prove beneficial or to distribute others that are available to be marketed, because of the fear of large adverse monetary judgments.” (Id. at 1063). The Court of Appeal agreed, noting that “the court’s decision [in Brown] was grounded in public policy concerns. Subjecting prescription drug manufacturers to strict liability for design defects, the court worried, might discourage drug development or inflate the cost of otherwise affordable drugs.” (Op. 29).

In rejecting the relevance of the argument here, however, the Court of Appeal (very briefly) argued a) that Brown espoused only a policy against burdening pharmaceutical companies with a duty stemming from unforeseeable harms, (Op. 49-50), and b) that the relevant cost here might be “some failed or wasted efforts,” but not a reduction in safety. (Op. 51).[4] Both of these claims are erroneous.

On the first, the legalistic distinction between foreseeable and unforeseeable harm was not, in fact, the determinative distinction in Brown. Rather, that distinction was relevant only because it maps onto the issue of incentives. In the face of unforeseeable, and thus unavoidable, harm, pharmaceutical companies would have severely diminished incentives to innovate. While foreseeable harms might also deter innovation by imposing some additional cost, these costs would be smaller, and avoidable or insurable, so that innovation could continue. To be sure, the Court wanted to ensure that the beneficial, risk-reduction effects of the tort system were not entirely removed from pharmaceutical companies. But that meant a policy decision that necessarily reduced the extent of tort-based risk optimization in favor of the manifest, countervailing benefit of relatively higher innovation incentives. That same calculus applies here, and it is this consideration, not the superficial question of foreseeability, that animated this Court in Brown.

On the second, the Court of Appeal inexplicably fails to acknowledge that the true cost of the imposition of excessive liability risk from a “duty to market” (or “duty to innovate”) is not limited to the expenditure of wasted resources, but the non-expenditure of any resources. The court’s contention appears to contemplate that such a duty would not remove a firm’s incentive to innovate entirely, although it might deter it slightly by increasing its expected cost. But economic incentives operate at the margin. Even if there remains some profit incentive to continue to innovate, the imposition of liability risk simply for the act of doing so would necessarily reduce the amount of innovation (in some cases, and especially for some smaller companies less able to bear the additional cost, to the point of deterring innovation entirely). But even this reduction in incentive is a harm. The fact that some innovation may still occur despite the imposition of considerable liability risk is not a defense of the imposition of that risk; rather, it is a reason to question its desirability, exactly as this Court did in Brown.

The Court of Appeal’s Decision Would Undermine Development of Lifesaving and Safer New Medicines

Innovation is a long-term, iterative process fraught with uncertainty. At the outset of research and development, it is impossible to know whether a potential new drug will ultimately prove superior to existing drugs. Most attempts at innovation fail to yield a marketable product, let alone one that is significantly safer or more effective than its predecessors. Deciding whether to pursue a particular line of research depends on weighing myriad factors, including the anticipated benefits of the new drug, the time and expense required to develop it, and its financial viability relative to existing products. Sometimes, potentially promising drug candidates are not pursued fully, even if theoretically “better” than existing drugs to some degree, because the expected benefits are not sufficient to justify the substantial costs and risks of development and commercialization.

If left to stand, the Court of Appeal’s decision would mean that whenever this stage of development is reached for a drug that may offer any safety improvement, the manufacturer will face potential liability for failing to bring that drug to market, regardless of the costs and risks involved in its development or the extent of the potential benefit. Such a rule would have severe unintended consequences that would stifle innovation.

First, by exposing manufacturers to liability on the basis of early-stage research that has not yet established a drug candidate’s safety and efficacy, the Court of Appeal’s rule would deter manufacturers from pursuing innovations in the first place. Drug research involves constant iteration, with most efforts failing and the potential benefits of success highly uncertain until late in the process. If any improvement, no matter how small or tentative, could trigger liability for failing to develop the new drug, manufacturers will be deterred from trying to innovate at all.

Second, such a rule would force manufacturers to direct scarce resources to developing and commercializing drugs that offer only small or incremental benefits because failing to do so would invite litigation. This would necessarily divert funds away from research into other potential drugs that could yield greater advancements. Further, as each small improvement is made, it reduces the relative potential benefit from, and therefore the incentive to undertake, further improvements. Rather than promoting innovation, the Court of Appeal’s decision would create incentives that favor small, incremental changes over larger, riskier leaps with the greatest potential to significantly advance patient welfare.

Third, and conversely, the Court of Appeal’s decision would set an unrealistic and dangerous standard of perfection for drug development. Pharmaceutical companies should not be expected to bring only the “safest” version of a drug to market, as this would drastically increase the time and cost of drug development and deprive patients of access to beneficial treatments in the meantime.

Fourth, the threat of liability would lead to inefficient and costly distortions in how businesses organize their research and development efforts. To minimize the risk of liability, manufacturers may avoid integrating ongoing research into existing product lines, instead keeping the processes separate unless and until a potential new technology is developed that offers benefits so substantial as to clearly warrant the costs and liability exposure of its development in the context of an existing drug line. Such an incentive would prevent potentially beneficial innovations from being pursued and would increase the costs of drug development.

Finally, the ruling would create perverse incentives that could actually discourage drug companies from developing and introducing safer alternative drugs. If bringing a safer drug to market later could be used as evidence that the first-generation drug was not safe enough, companies may choose not to invest in developing improved versions at all in order to avoid exposing themselves to liability. This would, of course, directly undermine the goal of increasing drug safety overall.

The Court of Appeal gave insufficient consideration to these severe policy consequences of the duty it recognized. A manufacturer’s decision when to bring a potentially safer drug to market involves complex trade-offs that courts are ill-equipped to second-guess—particularly in the limited context of a products-liability determination.

Conclusion

The Court of Appeal’s novel “duty to market” any known, less-harmful alternative to an existing product would deter innovation to the detriment of consumers. The Court of Appeal failed to consider how its decision would distort incentives in a way that harms the very patients the tort system is meant to protect. This Court should grant review to address these important legal and policy issues and to prevent this unprecedented expansion of tort liability from distorting manufacturers’ incentives to develop new and better products.

[1] No party or counsel for a party authored or paid for this amicus letter in whole or in part.

[2] It is important to note that this number varies with the kind of medicine involved, but across all categories of medicines there is a high likelihood of failure subsequent to Phase I trials.

[3] To the extent the concern is with disclosure of information regarding a potentially better product, that is properly a function of the patent system, which requires public disclosure of new ideas in exchange for the receipt of a patent. (See Brenner v. Manson, 383 U.S. 519, 533 (1966) (“one of the purposes of the patent system is to encourage dissemination of information concerning discoveries and inventions.”)). Of course, the patent system preserves innovation incentives despite the mandatory disclosure of information by conferring an exclusive right to the inventor to use the new knowledge. By contrast, using the tort system as an information-forcing device in this context would impose risks and costs on innovation without commensurate benefit, ensuring less, rather than more, innovation.

[4] The Court of Appeal makes a related argument when it claims that “the duty does not require manufacturers to perfect their drugs, but simply to act with reasonable care for the users of the existing drug when the manufacturer has developed an alternative that it knows is safer and at least equally efficacious. Manufacturers already engage in this type of innovation in the ordinary course of their business, and most plaintiffs would likely face a difficult road in establishing a breach of the duty of reasonable care.” (Op. at 52-3).

Dear Chair Khan, Commissioners Slaughter and Bedoya, and General Counsel Dasgupta,

The International Center for Law & Economics (ICLE) respectfully submits this letter in response to your June 29, 2023, NPRM regarding amendments to the premerger notification rules that implement the Hart-Scott-Rodino Antitrust Improvements Act (HSR Act) and to the Premerger Notification and Report Form and Instructions.

ICLE is a nonprofit, nonpartisan research center working to promote the use of law & economics methodologies to inform public policy debate. We have a long history of participation in regulatory proceedings relating to competition and antitrust law, including recent revisions to the merger guidelines[1] and the proposed revisions to the HSR premerger notification process.[2] We are consistently grateful for the opportunity to participate in proceedings such as these.

We write to express our concern about an important omission in the FTC’s proposed changes to the premerger notification form: its failure to address the requirements of the Regulatory Flexibility Act (RFA).[3] We appreciate your interest in this matter and the opportunity to share our concern with your offices.

This concern involves two legislative frameworks: the HSR premerger notification process and the requirements of the RFA. Under the HSR Act’s amendments to the Clayton Act, firms engaging in mergers above a statutorily defined minimum value[4]—including many that would involve smaller businesses to which the RFA applies—are required to provide information about a proposed merger to the FTC and the Department of Justice (DOJ) before the transaction can close. To bolster information gathering in merger enforcement, the FTC (with concurrence from the DOJ) proposed an extensive set of amendments to the filing process outlined by the HSR Act.[5] The proposed changes to the HSR process would dramatically expand the disclosure obligations for merging companies that meet the minimum valuation threshold.[6]

Under the RFA, federal agencies “shall prepare and make available for public comment an initial regulatory flexibility analysis. . . describ[ing] the impact of the proposed rule on small entities,”[7] except where “the head of the agency certifies that the rule will not, if promulgated, have a significant economic impact on a substantial number of small entities.”[8] If the agency believes the amendment will not have a substantial impact on small businesses, then it must provide “the factual basis” for its conclusion.[9]

The statement certifying that the proposed HSR changes in the NPRM won’t affect small businesses reads, in full:

Because of the size of the transactions necessary to invoke an HSR Filing, the premerger notification rules rarely, if ever, affect small entities. The 2000 amendments to the Act exempted all transactions valued at $50 million or less, with subsequent automatic adjustments to take account of changes in Gross National Product resulting in a current threshold of $111 million. Further, none of the proposed amendments expands the coverage of the premerger notification rules in a way that would affect small entities. Accordingly, the Commission certifies that these proposed amendments will not have a significant economic impact on a substantial number of small entities.[10]

Unfortunately, this is insufficient to satisfy the requirements of the RFA. Although the FTC stresses the $111 million HSR threshold to assert that small entities will not be affected, the Small Business Administration (SBA) “generally defines a small business as an independent business having fewer than 500 employees.”[11] The SBA also offers more detailed, industry-specific identification of small businesses.[12] Indeed, the NPRM cites to the SBA’s own standards, but these, too, do not align with the FTC’s “factual statement,” and it is not evident that the Commission sufficiently delved into those standards to understand their relevance for the size thresholds under the HSR Act.

Even a quick review of the SBA’s “Small Business Size Standards by NAICS Industry” table reveals that the SBA classifies the size of a firm based on either annual receipts or number of employees,[13] depending upon the characteristics of their industry.[14] Neither of these, it should go without saying, is the same as a “size-of-transaction” threshold under the HSR Act. Nor does the NPRM’s “factual basis” statement contain information sufficient to determine that there is any correlation between the SBA’s size thresholds and the size of a transaction (which typically represents something between the discounted present value of a firm’s expected returns under new ownership and current ownership over an indefinite time period).

Despite the FTC’s claim that the $111 million deal threshold will ensure that small businesses are not substantially affected, the agency’s own data from 2022 shows that nearly a quarter of all HSR filings covered transactions involving firms with sales of $50 million or less.[15] The same data shows that, out of the 3029 reported transactions in 2022, 513 involved firms with between $50 and $100 million in sales and 305 with between $100 and $150 million in sales. Here, again, the SBA’s metrics for identifying small businesses bear emphasis: where the SBA relies on dollar values instead of employee headcounts to define small businesses at all, it does so based on annual average receipts, not on the overall value of the firm.

This distinction underscores a point made in a letter filed by the App Association, a trade group representing small technology firms, that it is important not to conflate valuation with size.[16] A company, such as an innovative tech startup, can have a small number of employees but a high value based on projected sales, intellectual property, and forthcoming products. Indeed, the App Association notes that a number of its members are already subject to HSR disclosures and that that number can only increase under the proposed amendments.[17]

To provide further context regarding whether many of these deals involve small businesses, a 2013 CrunchBase dataset showed that the average successful American startup sold for $242.9 million.[18] Furthermore, the FTC’s 2022 HSR report highlights at least one challenged transaction involving a small business: Meta/Within.[19] Within, a virtual-reality startup with 58 employees, was acquired by Meta for $400 million.[20] Not only was there an HSR filing, but the FTC attempted to challenge the transaction—and lost in district court.[21]

Small businesses are clearly burdened by the HSR premerger notification requirements—and this burden would only increase under the proposed changes. By the FTC’s own estimate the new requirements would quadruple the hours required to prepare an HSR filing and raise costs by $350 million. By other, more realistic estimates, that increase in work hours would entail a cost of more than $1.6 billion[22]—or, indeed, considerably more.[23] There is no question that drastically increasing the cost of merger filings will make it much harder for small businesses to merge or be acquired, which is a primary form of success for small businesses.

Indeed, the NPRM’s proposed changes are, in part, specifically designed to affect small businesses. “Acquisitions of small companies can cause harm, including in sectors where competition occurs on a local level. . . . Thus, the Commission proposes several changes to expand the requirements for information related to prior acquisitions beyond what is currently required by Item 8.”[24] Furthermore, “given the difficulties in determining the value of small or nascent companies, the Commission believes it would be less burdensome for filers to report all acquisitions. . . .”[25]  Indeed, the FTC is aware of the potential burden on small businesses that such an approach would entail, but nevertheless aims to ensure that its proposed rules “still captur[e] acquisitions of entities worth less than $10 million.”[26]

And there is yet a further problem: These concerns take into account only the direct costs the NPRM would impose on small businesses. But, as the National Federation of Independent Business highlighted in 2023, several agencies have arguably failed to comply with the RFA by failing to consider indirect effects on small businesses.[27] Obviously, there is no such analysis provided here—and, indeed, as noted above, a clear intent of the NPRM is to affect the likelihood of small-business acquisition by reducing the incentive for firms to serially acquire small businesses. Doing so, of course, reduces funders’ incentives to invest in startups and small businesses and raises these companies’ cost of capital. Arguably that increase is itself a direct cost, but certainly its indirect effect is incredibly significant to the health of small businesses in the U.S.

The dramatic changes to the HSR premerger notification requirements proposed by the FTC have already created substantial uncertainty within the antitrust bar. Procedural defects such as failing to comply with the requirements of the RFA increase the likelihood that any rules adopted by the FTC will be challenged in court. This would increase the uncertainty (and thus the cost) surrounding the HSR process. This would be an unfortunate outcome. Fortunately, it is one that can be avoided if the FTC addresses these issues prior to finalizing its proposed rules.

[1] Geoffrey A. Manne, Dirk Auer, Brian Albrecht, Eric Fruits, Daniel J. Gilman, & Lazar Radic, Comments of the International Center for Law and Economics on the FTC & DOJ Draft Merger Guidelines, International Center for Law and Economics (Sept 18, 2023), https://laweconcenter.org/resources/comments-of-the-international-center-for-law-and-economics-on-the-ftc-doj-draft-merger-guidelines/.

[2] Brian Albrecht, Dirk Auer, Daniel J. Gilman, Gus Hurwitz, & Geoffrey A. Manne, Comments of the International Center for Law & Economics on Proposed Changes to the Premerger Notification Rules, International Center for Law and Economics (Sept 27,2023), https://laweconcenter.org/resources/comments-of-the-international-center-for-law-economics-on-proposed-changes-to-the-premerger-notification-rules/.

[3] 5 U.S.C. §§ 601-612 (2018).

[4] 15 U.S.C. § 18a(a)(2) (2018).

[5] NPRM, 88 FR 42178 (Jun. 29, 2023).

[6] See id. at 42208 (estimating the hours and expenses required to comply with the new rules). According to antitrust practitioners, however, the NPRM’s estimate likely substantially underestimates the true burden and cost of the proposed rules. See, e.g., Sean Heather, Antitrust Experts Reject FTC/DOJ Changes to Merger Process, Chamber of Commerce (Sept 19, 2023), https://www.uschamber.com/finance/antitrust/antitrust-experts-reject-ftc-doj-changes-to-merger-process.

[7] 5 U.S.C. § 603(a) (2018).

[8] 5 U.S.C. § 605(b) (2018).

[9] Id.

[10] NPRM, 88 FR 42178, 42208 (Jun. 29, 2023).

[11] Frequently Asked Questions, U.S. Small Bus. Admin. Off. of Advoc. (2023), https://advocacy.sba.gov/wp-content/uploads/2023/03/Frequently-Asked-Questions-About-Small-Business-March-2023-508c.pdf.

[12] See 13 CFR § 121.101, et seq. (1996)

[13] 13 CFR § 121.201 (2024).

[14] Indeed, the SBA’s standards entail a review of a wide range of such characteristics. See 13 CFR § 121.102 (1996) (“SBA considers economic characteristics comprising the structure of an industry, including degree of competition, average firm size, start-up costs and entry barriers, and distribution of firms by size. It also considers technological changes, competition from other industries, growth trends, historical activity within an industry, unique factors occurring in the industry which may distinguish small firms from other firms, and the objectives of its programs and the impact on those programs of different size standard levels.”).

[15] See Fed. Trade Comm’n and Dept of Just., Hart-Scott-Rodino Annual Report (2022), at Table IX, available at https://www.ftc.gov/system/files/ftc_gov/pdf/FY2022HSRReport.pdf.

[16] See Letter from Morgan Reed, President of App Association, to Lina Khan, Chair of Fed. Trade. Comm’n and Members of Congress (Feb 1, 2024), available at https://actonline.org/wp-content/uploads/App-Association-HSR-RFA-Ltr-1-Feb-2024-1.pdf.

[17] See id.

[18] See Mark Lennon, CrunchBase Reveals: The Average Successful Startup Raises $41M, Exits at $242.9M, TechCrunch (Dec 14, 2013), https://techcrunch.com/2013/12/14/crunchbase-reveals-the-average-successful-startup-raises-41m-exits-at-242-9m.

[19] See Fed. Trade Comm’n and Dept of Just., Hart-Scott-Rodino Annual Report (2022), available at https://www.ftc.gov/system/files/ftc_gov/pdf/FY2022HSRReport.pdf.

[20] See, e.g., Within (Virtual Reality) Overview, Pitchbook (last visited Feb. 29, 2024), https://pitchbook.com/profiles/company/117068-59#overview.

[21] In the Matter of Meta/Zuckerberg/Within, Fed. Trade Comm’n Docket No. 9411 (Aug. 11, 2022), https://www.ftc.gov/legal-library/browse/cases-proceedings/221-0040-metazuckerbergwithin-matter.

[22] See Albrecht, et al., supra note 2, at 7 (“The U.S. Chamber of Commerce conducted ‘a survey of 70 antitrust practitioners asking them questions about the proposed revisions to the HSR merger form and the new draft merger guides.’ Based on average answers from the survey respondents, the new rules would increase compliance costs by $1.66 billion, almost five times the FTC’s $350 million estimate.”).

[23] See id. (“For the current rules, the average survey response puts the cost of compliance at $79,569. Assuming there are 7,096 filings (as the FTC projects for FY 23), the total cost under the current rules would be $565 million. Under the new rules, the average survey response estimates the expected cost of compliance to be $313,828 per transaction, for a total cost of $2.23 billion.”) (emphasis added).

[24] NPRM, 88 FR 42178, 42203 (Jun. 29, 2023).

[25] Id. at 42204 (emphasis added).

[26] Id. (emphasis added).

[27] See Rob Smith, The Regulatory Flexibility Act: Turning a Paper Tiger Into a Legitimate Constraint on One-Size-Fits-All Agency Rulemaking, NFIB Small Business Legal Center (May 2, 2023), https://strgnfibcom.blob.core.windows.net/nfibcom/NFIB-RFA-White-paper.pdf (collecting examples).

I. Introduction

It was once (and frequently) said that Google’s “data monopoly” was unassailable: “If ‘big data’ is the oil of the information economy, Google has Standard Oil-like monopoly dominance — and uses that control to maintain its dominant position.”^[1] Similar epithets have been hurled at virtually all large online platforms, including Facebook (Meta), Amazon, and Uber.^[2]

While some of these claims continue even today (for example, “big data” is a key component of the U.S. Justice Department’s (“DOJ”) Google Search and AdTech antitrust suits),^[3] a shiny new data target has emerged in the form of generative artificial intelligence. The launch of ChatGPT in November 2022, as well as the advent of AI image-generation services like Midjourney and Dall-E, have dramatically expanded people’s conception of what is, and what might be, possible to achieve with generative AI technologies built on massive data sets.

While these services remain in the early stages of mainstream adoption and are in the throes of rapid, unpredictable technological evolution, they nevertheless already appear on the radar of competition policymakers around the world. Several antitrust enforcers appear to believe that, by acting now, they can avoid the “mistakes” that were purportedly made during the formative years of Web 2.0.^[4] These mistakes, critics assert, include failing to appreciate the centrality of data in online markets, as well as letting mergers go unchecked and allowing early movers to entrench their market positions.^[5] As Lina Khan, Chair of the FTC, put it: “we are still reeling from the concentration that resulted from Web 2.0, and we don’t want to repeat the mis-steps of the past with AI”.^[6]

In that sense, the response from the competition-policy world is deeply troubling. Instead of engaging in critical self-assessment and adopting an appropriately restrained stance, the enforcement community appears to be chomping at the bit. Rather than assessing their prior assumptions based on the current technological moment, enforcers’ top priority appears to be figuring out how to deploy existing competition tools rapidly and almost reflexively to address the presumed competitive failures presented by generative AI.^[7]

It is increasingly common for competition enforcers to argue that so-called “data network effects” serve not only to entrench incumbents in the markets where that data is collected, but also confer similar, self-reinforcing benefits in adjacent markets. Several enforcers have, for example, prevented large online platforms from acquiring smaller firms in adjacent markets, citing the risk that they could use their vast access to data to extend their dominance into these new markets.^[8] They have also launched consultations to ascertain the role that data plays in AI competition. For instance, in an ongoing consultation, the European Commission asks: “What is the role of data and what are its relevant characteristics for the provision of generative AI systems and/or components, including AI models?”^[9] Unsurprisingly, the U.S. Federal Trade Commission (“FTC”) has been bullish about the risks posed by incumbents’ access to data. In comments submitted to the U.S. Copyright Office, for example, the FTC argued that:

The rapid development and deployment of AI also poses potential risks to competition. The rising importance of AI to the economy may further lock in the market dominance of large incumbent technology firms. These powerful, vertically integrated incumbents control many of the inputs necessary for the effective development and deployment of AI tools, including cloud-based or local computing power and access to large stores of training data. These dominant technology companies may have the incentive to use their control over these inputs to unlawfully entrench their market positions in AI and related markets, including digital content markets.^[10]

Against this backdrop, it stands to reason that the largest online platforms—including Alphabet, Meta, Apple, and Amazon — should have a meaningful advantage in the burgeoning markets for generative AI services. After all, it is widely recognized that data is an essential input for generative AI.^[11] This competitive advantage should be all the more significant given that these firms have been at the forefront of AI technology for more than a decade. Over this period, Google’s DeepMind and AlphaGo and Meta’s have routinely made headlines.^[12] Apple and Amazon also have vast experience with AI assistants, and all of these firms use AI technology throughout their platforms.^[13]

Contrary to what one might expect, however, the tech giants have, to date, been unable to leverage their vast data troves to outcompete startups like OpenAI and Midjourney. At the time of writing, OpenAI’s ChatGPT appears to be, by far, the most successful chatbot^[14], despite the fact that large tech platforms arguably have access to far more (and more up-to-date) data.

This article suggests there are important lessons to be learned from the current technological moment, if only enforcers would stop to reflect. The meteoric rise of consumer-facing AI services should offer competition enforcers and policymakers an opportunity for introspection. As we explain, the rapid emergence of generative AI technology may undercut many core assumptions of today’s competition-policy debates — the rueful after-effects of the purported failure of 20^th-century antitrust to address the allegedly manifest harms of 21^st-century technology. These include the notions that data advantages constitute barriers to entry and can be leveraged to project dominance into adjacent markets; that scale itself is a market failure to be addressed by enforcers; and that the use of consumer data is inherently harmful to those consumers.

II. Data Network Effects Theory and Enforcement

Proponents of tougher interventions by competition enforcers into digital markets often cite data network effects as a source of competitive advantage and barrier to entry (though terms like “economies of scale and scope” may offer more precision).^[15] The crux of the argument is that “the collection and use of data creates a feedback loop of more data, which ultimately insulates incumbent platforms from entrants who, but for their data disadvantage, might offer a better product.”^[16] This self-reinforcing cycle purportedly leads to market domination by a single firm. Thus, for Google, for example, it is argued that its “ever-expanding control of user personal data, and that data’s critical value to online advertisers, creates an insurmountable barrier to entry for new competition.”^[17]

Right off the bat, it is important to note the conceptual problem of these claims. Because data is used to improve the quality of products and/or to subsidize their use, the idea of data as an entry barrier suggests that any product improvement or price reduction made by an incumbent could be a problematic entry barrier to any new entrant. This is tantamount to an argument that competition itself is a cognizable barrier to entry. Of course, it would be a curious approach to antitrust if this were treated as a problem, as it would imply that firms should under-compete — should forego consumer-welfare enhancements—in order to bring about a greater number of firms in a given market simply for its own sake.^[18]

Meanwhile, actual economic studies of data network effects are few and far between, with scant empirical evidence to support the theory.^[19] Andrei Hagiu and Julian Wright’s theoretical paper offers perhaps the most comprehensive treatment of the topic.^[20] The authors ultimately conclude that data network effects can be of different magnitudes and have varying effects on firms’ incumbency advantage.^[21] They cite Grammarly (an AI writing-assistance tool) as a potential example: “As users make corrections to the suggestions offered by Grammarly, its language experts and artificial intelligence can use this feedback to continue to improve its future recommendations for all users.”^[22]

This is echoed by other economists who contend that “[t]he algorithmic analysis of user data and information might increase incumbency advantages, creating lock-in effects among users and making them more reluctant to join an entrant platform.”^[23]

Crucially, some scholars take this logic a step further, arguing that platforms may use data from their “origin markets” in order to enter and dominate adjacent ones:

First, as we already mentioned, data collected in the origin market can be used, once the enveloper has entered the target market, to provide products more efficiently in the target market. Second, data collected in the origin market can be used to reduce the asymmetric information to which an entrant is typically subject when deciding to invest (for example, in R&D) to enter a new market. For instance, a search engine could be able to predict new trends from consumer searches and therefore face less uncertainty in product design.^[24]

This possibility is also implicit in the paper by Hagiu and Wright.^[25] Indeed, the authors’ theoretical model rests on an important distinction between within-user data advantages (that is, having access to more data about a given user) and across-user data advantages (information gleaned from having access to a wider user base). In both cases, there is an implicit assumption that platforms may use data from one service to gain an advantage in another market (because what matters is information about aggregate or individual user preferences, regardless of its origin).

Our review of the economic evidence suggests that several scholars have, with varying degrees of certainty, raised the possibility that incumbents may leverage data advantages to stifle competitors in their primary market or adjacent ones (be it via merger or organic growth). As we explain below, however, there is ultimately little evidence to support such claims.

Policymakers, however, have largely been receptive to these limited theoretical findings, basing multiple decisions on these theories, often with little consideration of the caveats that accompany them.^[26] Indeed, it is remarkable that, in the Furman Report’s section on “[t]he data advantage for incumbents,” only two empirical economic studies are cited, and they offer directly contradictory conclusions with respect to the question of the strength of data advantages.^[27] Nevertheless, the Furman Report concludes that data “may confer a form of unmatchable advantage on the incumbent business, making successful rivalry less likely,”^[28] and adopts without reservation “convincing” evidence from non-economists with apparently no empirical basis.^[29]

In the Google/Fitbit merger proceedings, the European Commission found that the combination of data from Google services with that of Fitbit devices would reduce competition in advertising markets:

Giving [sic] the large amount of data already used for advertising purposes that Google holds, the increase in Google’s data collection capabilities, which goes beyond the mere number of active users for which Fitbit has been collecting data so far, the Transaction is likely to have a negative impact on the development of an unfettered competition in the markets for online advertising.^[30]

As a result, the Commission cleared the merger on the condition that Google refrain from using data from Fitbit devices for its advertising platform.^[31] The Commission will likely focus on similar issues during its ongoing investigation into Microsoft’s investment into OpenAI.^[32]

Along similar lines, the FTC’s complaint to enjoin Meta’s purchase of a virtual-reality (VR) fitness app called “Within” relied, among other things, on the fact that Meta could leverage its data about VR-user behavior to inform its decisions and potentially outcompete rival VR-fitness apps: “Meta’s control over the Quest platform also gives it unique access to VR user data, which it uses to inform strategic decisions.”^[33]

The U.S. Department of Justice’s twin cases against Google also raise data leveraging and data barriers to entry. The agency’s AdTech complaint that “Google intentionally exploited its massive trove of user data to further entrench its monopoly across the digital advertising industry.”^[34] Similarly, in its Search complaint, the agency argues that:

Google’s anticompetitive practices are especially pernicious because they deny rivals scale to compete effectively. General search services, search advertising, and general search text advertising require complex algorithms that are constantly learning which organic results and ads best respond to user queries; the volume, variety, and velocity of data accelerates the automated learning of search and search advertising algorithms.^[35]

Finally, the merger guidelines published by several competition enforcers cite the acquisition of data as a potential source of competitive concerns. For instance, the FTC and DOJ’s newly published guidelines state that “acquiring data that helps facilitate matching, sorting, or prediction services may enable the platform to weaken rival platforms by denying them that data.”^[36] Likewise, the UK Competition and Markets Authority (“CMA”) warns against incumbents acquiring firms in order to obtain their data and foreclose other rivals:

Incentive to foreclose rivals…

7.19(e) Particularly in complex and dynamic markets, firms may not focus on short term margins but may pursue other objectives to maximise their long-run profitability, which the CMA may consider. This may include… obtaining access to customer data….^[37]

In short, competition authorities around the globe are taking an aggressive stance on data network effects. Among the ways this has manifested is in basing enforcement decisions on fears that data collected by one platform might confer a decisive competitive advantage in adjacent markets. Unfortunately, these concerns rest on little to no empirical evidence, either in the economic literature or the underlying case records.

III. Data Incumbency Advantages in Generative AI Markets

Given the assertions canvassed in the previous section, it seems reasonable to assume that firms such as Google, Meta, and Amazon would be in pole position to dominate the burgeoning market for generative AI. After all, these firms have not only been at the forefront of the field for the better part of a decade, but they also have access to vast troves of data, the likes of which their rivals could only dream when they launched their own services. Thus the authors of the Furman Report caution that “to the degree that the next technological revolution centres around artificial intelligence and machine learning, then the companies most able to take advantage of it may well be the existing large companies because of the importance of data for the successful use of these tools.^[38]

At the time of writing, however, this is not how things have unfolded — although it bears noting these markets remain in flux and the competitive landscape is susceptible to change. The first significantly successful generative AI service was arguably not from either Meta—which had been working on chatbots for years and had access to, arguably, the world’s largest database of actual chats—or Google. Instead, the breakthrough came from a previously unknown firm called OpenAI.

OpenAI’s ChatGPT service currently holds an estimated 60% of the market (though reliable numbers are somewhat elusive).^[39] It broke the record for the fastest online service to reach 100 million users (in only a couple of months), more than four times faster than the previous record holder, TikTok.^[40] Based on Google Trends data, ChatGPT is nine times more popular than Google’s own Bard service worldwide, and 14 times more popular in the U.S.^[41] In April 2023, ChatGPT reportedly registered 206.7 million unique visitors, compared to 19.5 million for Google’s Bard.^[42] In short, at the time of writing, ChatGPT appears to be the most popular chatbot. And, so far, the entry of large players such as Google Bard or Meta AI appear to have had little effect on its market position.^[43]

The picture is similar in the field of AI image generation. As of August 2023, Midjourney, Dall-E, and Stable Diffusion appear to be the three market leaders in terms of user visits.^[44] This is despite competition from the likes of Google and Meta, who arguably have access to unparalleled image and video databases by virtue of their primary platform activities.^[45]

This raises several crucial questions: how have these AI upstarts managed to be so successful, and is their success just a flash in the pan before Web 2.0 giants catch up and overthrow them? While we cannot answer either of these questions dispositively, some observations concerning the role and value of data in digital markets would appear to be relevant.

A first important observation is that empirical studies suggest data exhibits diminishing marginal returns. In other words, past a certain point, acquiring more data does not confer a meaningful edge to the acquiring firm. As Catherine Tucker puts it, following a review of the literature: “Empirically there is little evidence of economies of scale and scope in digital data in the instances where one would expect to find them.”^[46]

Likewise, following a survey of the empirical literature on this topic, Geoffrey Manne & Dirk Auer conclude that:

Available evidence suggests that claims of “extreme” returns to scale in the tech sector are greatly overblown. Not only are the largest expenditures of digital platforms unlikely to become proportionally less important as output increases, but empirical research strongly suggests that even data does not give rise to increasing returns to scale, despite routinely being cited as the source of this effect.^[47]

In other words, being the firm with the most data appears to be far less important than having enough data, and this lower bar may be accessible to far more firms than one might initially think possible.

And obtaining enough data could become even easier — that is, the volume of required data could become even smaller — with technological progress. For instance, synthetic data may provide an adequate substitute to real-world data^[48] — or may even outperform real-world data.^[49] As Thibault Schrepel and Alex Pentland point out, “advances in computer science and analytics are making the amount of data less relevant every day. In recent months, important technological advances have allowed companies with small data sets to compete with larger ones.”^[50]

Indeed, past a certain threshold, acquiring more data might not meaningfully improve a service, where other improvements (such as better training methods or data curation) could have a large effect. In fact, there is some evidence that excessive data impedes a service’s ability to generate results appropriate for a given query: “[S]uperior model performance can often be achieved with smaller, high-quality datasets than massive, uncurated ones. Data curation ensures that training datasets are devoid of noise, irrelevant instances, and duplications, thus maximizing the efficiency of every training iteration.”^[51]

Consider, for instance, a user who wants to generate an image of a basketball. Using a model trained on an indiscriminate range and number of public photos in which a basketball appears, but is surrounded by copious other image data, the user may end up with an inordinately noisy result. By contrast, a model trained with a better method on fewer, more-carefully selected images, could readily yield far superior results.^[52] In one important example,

[t]he model’s performance is particularly remarkable, given its small size. “This is not a large language model trained on the whole Internet; this is a relatively small transformer trained for these tasks,” says Armando Solar-Lezama, a computer scientist at the Massachusetts Institute of Technology, who was not involved in the new study…. The finding implies that instead of just shoving ever more training data into machine-learning models, a complementary strategy might be to offer AI algorithms the equivalent of a focused linguistics or algebra class.^[53]

Current efforts are thus focused on improving the mathematical and logical reasoning of large language models (“LLMs”), rather than maximizing training datasets.^[54] Two points stand out. The first is that firms like OpenAI rely largely on publicly available datasets — such as GSM8K — to train their LLMs.^[55] Second, the real challenge to create cutting-edge AI is not so much in collecting data, but rather in creating innovative AI training processes and architectures:

[B]uilding a truly general reasoning engine will require a more fundamental architectural innovation. What’s needed is a way for language models to learn new abstractions that go beyond their training data and have these evolving abstractions influence the model’s choices as it explores the space of possible solutions.

We know this is possible because the human brain does it. But it might be a while before OpenAI, DeepMind, or anyone else figures out how to do it in silicon.^[56]

Furthermore, it is worth noting that the data most relevant to startups operating in a given market may not be those data held by large incumbent platforms in other markets, but rather data specific to the market in which the startup is active or, even better, to the given problem it is attempting to solve:

As Andres Lerner has argued, if you wanted to start a travel business, the data from Kayak or Priceline would be far more relevant. Or if you wanted to start a ride-sharing business, data from cab companies would be more useful than the broad, market-cross-cutting profiles Google and Facebook have. Consider companies like Uber, Lyft and Sidecar that had no customer data when they began to challenge established cab companies that did possess such data. If data were really so significant, they could never have competed successfully. But Uber, Lyft and Sidecar have been able to effectively compete because they built products that users wanted to use — they came up with an idea for a better mousetrap. The data they have accrued came after they innovated, entered the market and mounted their successful challenges — not before.^[57]

The bottom line is that data is not the be-all and end-all that many in competition circles rather casually make it out to be.[58] While data may often confer marginal benefits, there is little sense these are ultimately decisive.^[59] As a result, incumbent platforms’ access to vast numbers of users and data in their primary markets might only marginally affect their AI competitiveness.

A related observation is that firms’ capabilities and other features of their products arguably play a more important role than the data they own.^[60] Examples of this abound in digital markets. Google overthrew Yahoo, despite initially having access to far fewer users and far less data; Google and Apple overcame Microsoft in the smartphone OS market despite having comparatively tiny ecosystems (at the time) to leverage; and TikTok rose to prominence despite intense competition from incumbents like Instagram, which had much larger user bases. In each of these cases, important product-design decisions (such as the PageRank algorithm, recognizing the specific needs of mobile users,^[61] and TikTok’s clever algorithm) appear to have played a far greater role than initial user and data endowments (or lack thereof).

All of this suggests that the early success of OpenAI likely has more to do with its engineering decisions than the data it did (or did not) own. And going forward, OpenAI and its rivals’ ability to offer and monetize compelling stores offering custom versions of their generative AI technology will arguably play a much larger role than (and contribute to) their ownership of data.^[62] In other words, the ultimate challenge is arguably to create a valuable platform, of which data ownership is a consequence, but not a cause.

It is also important to note that, in those instances where it is valuable, data does not just fall from the sky. Instead, it is through smart business and engineering decisions that firms can generate valuable information (which does not necessarily correlate with owing more data).

For instance, OpenAI’s success with ChatGPT is often attributed to its more efficient algorithms and training models, which arguably have enabled the service to improve more rapidly than its rivals.^[63] Likewise, the ability of firms like Meta and Google to generate valuable data for advertising arguably depends more on design decisions that elicit the right data from users, rather than the raw number of users in their networks.

Put differently, setting up a business so as to generate the right information is more important than simply owning vast troves of data.^[64] Even in those instances where high-quality data is an essential parameter of competition, it does not follow that having vaster databases or more users on a platform necessarily leads to better information for the platform.

Given what precedes, it seems clear that OpenAI and other generative AI startups’ early success, as well as their chances of prevailing in the future, hinge on a far broader range of factors than the mere ownership of data. Indeed, if data ownership consistently conferred a significant competitive advantage, these new firms would not be where they are today. This does not mean that data is worthless, of course. Rather, it means that competition authorities should not assume that merely possessing data is a dispositive competitive advantage, absent compelling empirical evidence to support such a finding. In this light, the current wave of decisions and competition-policy pronouncements that rely on data-related theories of harm are premature.

IV. Five Key Takeaways: Reconceptualizing the Role of Data in Generative AI Competition

As we explain above, data (network effects) are not the source of barriers to entry that they are sometimes made out to be; rather, the picture is far more nuanced. Indeed, as economist Andres Lerner demonstrated almost a decade ago (and the assessment is only truer today):

Although the collection of user data is generally valuable for online providers, the conclusion that such benefits of user data lead to significant returns to scale and to the entrenchment of dominant online platforms is based on unsupported assumptions. Although, in theory, control of an “essential” input can lead to the exclusion of rivals, a careful analysis of real-world evidence indicates that such concerns are unwarranted for many online businesses that have been the focus of the “big data” debate.^[65]

While data can be an important part of the competitive landscape, incumbent data advantages are far less pronounced than today’s policymakers commonly assume. In that respect, five main lessons emerge:

1. Data can be (very) valuable, but past a certain threshold, the benefits tend to diminish. In other words, having the most data is less important than having enough;
2. The ability to generate valuable information does not depend on the number of users or the amount of data a platform has previously acquired;
3. The most important datasets are not always proprietary;
4. Technological advances and platforms’ engineering decisions affect their ability to generate valuable information, and this effect swamps the effect of the amount of data they own; and
5. How platforms use data is arguably more important than what data or how much data they own.

These lessons have important ramifications for competition-policy debates over the competitive implications of data in technologically evolving areas.

First, it is not surprising that startups, rather than incumbents, have taken an early lead in generative AI (and in Web 2.0 before it). After all, if data-incumbency advantages are small or even nonexistent, then smaller and more nimble players may have an edge over established tech platforms. This is all the more likely given that, despite significant efforts, the biggest tech platforms were unable to offer compelling generative AI chatbots and image-generation services before the emergence of ChatGPT, Dall-E, Midjourney, etc. This failure suggests that, in a process akin to Christensen’s Innovator’s Dilemma,^[66] something about their existing services and capabilities was holding them back in those markets. Of course, this does not necessarily mean that those same services/capabilities could not become an advantage when the generative AI market starts addressing issues of monetization and scale.^[67] But it does mean that assumptions of a firm’s market power based on its possession of data are off the mark.

Another important implication is that, paradoxically, policymakers’ efforts to prevent Web 2.0 platforms from competing freely in generative AI markets may ultimately backfire and lead to less, not more, competition. Indeed, OpenAI is currently acquiring a sizeable lead in generative AI. While competition authorities might like to think that other startups will emerge and thrive in this space, it is important not to confuse desires with reality. For, while there is a vibrant AI-startup ecosystem, there is at least a case to be made that the most significant competition for today’s AI leaders will come from incumbent Web 2.0 platforms — although nothing is certain at this stage. Policymakers should beware not to stifle that competition on the misguided assumption that competitive pressure from large incumbents is somehow less valuable to consumers than that which originates from smaller firms.

Finally, even if there were a competition-related market failure to be addressed (which is anything but clear) in the field of generative AI, it is unclear that contemplated remedies would do more good than harm. Some of the solutions that have been put forward have highly ambiguous effects on consumer welfare. Scholars have shown that mandated data sharing — a solution championed by EU policymakers, among others — may sometimes dampen competition in generative AI markets.^[68] This is also true of legislation like the GDPR that make it harder for firms to acquire more data about consumers — assuming such data is, indeed, useful to generative AI services.^[69]

In sum, it is a flawed understanding of the economics and practical consequences of large agglomerations of data that lead competition authorities to believe that data-incumbency advantages are likely to harm competition in generative AI markets — or even in the data-intensive Web 2.0 markets that preceded them. Indeed, competition or regulatory intervention to “correct” data barriers and data network and scale effects is liable to do more harm than good.

[1] Nathan Newman, Taking on Google’s Monopoly Means Regulating Its Control of User Data, Huffington Post (Sep. 24, 2013), http://www.huffingtonpost.com/nathan-newman/taking-on-googlesmonopol_b_3980799.html.

[2] See e.g. Lina Khan & K. Sabeel Rahman, Restoring Competition in the U.S. Economy, in Untamed: How to Check Corporate, Financial, and Monopoly Power (Nell Abernathy, Mike Konczal, & Kathryn Milani, eds., 2016), at 23 (“From Amazon to Google to Uber, there is a new form of economic power on display, distinct from conventional monopolies and oligopolies…, leverag[ing] data, algorithms, and internet-based technologies… in ways that could operate invisibly and anticompetitively.”); Mark Weinstein, I Changed My Mind — Facebook Is a Monopoly, Wall St. J. (Oct. 1, 2021), https://www.wsj.com/articles/facebook-is-monopoly-metaverse-users-advertising-platforms-competition-mewe-big-tech-11633104247 (“[T]he glue that holds it all together is Facebook’s monopoly over data…. Facebook’s data troves give it unrivaled knowledge about people, governments — and its competitors.”).

[3] See generally Abigail Slater, Why “Big Data” Is a Big Deal, The Reg. Rev. (Nov. 6, 2023), https://www.theregreview.org/2023/11/06/slater-why-big-data-is-a-big-deal/; Amended Complaint at ¶36, United States v. Google, 1:20-cv-03010- (D.D.C. 2020); Complaint at ¶37, United States v. Google, 1:23-cv-00108 (E.D. Va. 2023), https://www.justice.gov/opa/pr/justice-department-sues-google-monopolizing-digital-advertising-technologies (“Google intentionally exploited its massive trove of user data to further entrench its monopoly across the digital advertising industry.”).

[4] See e.g. Press Release, European Commission, Commission Launches Calls for Contributions on Competition in Virtual Worlds and Generative AI (Jan. 9, 2024), https://ec.europa.eu/commission/presscorner/detail/en/IP_24_85; Krysten Crawford, FTC’s Lina Khan warns Big Tech over AI, SIEPR (Nov. 3, 2020), https://siepr.stanford.edu/news/ftcs-lina-khan-warns-big-tech-over-ai (“Federal Trade Commission Chair Lina Khan delivered a sharp warning to the technology industry in a speech at Stanford on Thursday: Antitrust enforcers are watching what you do in the race to profit from artificial intelligence.”) (emphasis added).

[5] See e.g. John M. Newman, Antitrust in Digital Markets, 72 Vand. L. Rev. 1497, 1501 (2019) (“[T]he status quo has frequently failed in this vital area, and it continues to do so with alarming regularity. The laissez-faire approach advocated for by scholars and adopted by courts and enforcers has allowed potentially massive harms to go unchecked.”);
Bertin Martins, Are New EU Data Market Regulations Coherent and Efficient?, Bruegel Working Paper 21/23 (2023), available at https://www.bruegel.org/working-paper/are-new-eu-data-market-regulations-coherent-and-efficient (“Technical restrictions on access to and re-use of data may result in failures in data markets and data-driven services markets.”); Valéria Faure-Muntian, Competitive Dysfunction: Why Competition Law Is Failing in a Digital World, The Forum Network (Feb. 24, 2021), https://www.oecd-forum.org/posts/competitive-dysfunction-why-competition-law-is-failing-in-a-digital-world.

[6] Rana Foroohar, The Great US-Europe Antitrust Divide, FT (Feb. 5, 2024), https://www.ft.com/content/065a2f93-dc1e-410c-ba9d-73c930cedc14.

[7] See e.g. Press Release, European Commission, supra note 5.

[8] See infra, Section II. Commentators have also made similar claims. See, e.g., Ganesh Sitaram & Tejas N. Narechania, It’s Time for the Government to Regulate AI. Here’s How, Politico (Jan. 15, 2024) (“All that cloud computing power is used to train foundation models by having them “learn” from incomprehensibly huge quantities of data. Unsurprisingly, the entities that own these massive computing resources are also the companies that dominate model development. Google has Bard, Meta has LLaMa. Amazon recently invested $4 billion into one of OpenAI’s leading competitors, Anthropic. And Microsoft has a 49 percent ownership stake in OpenAI — giving it extraordinary influence, as the recent board struggles over Sam Altman’s role as CEO showed.”).

[9] Press Release, European Commission, supra note 5.

[10] Comment of U.S. Federal Trade Commission to the U.S. Copyright Office, Artificial Intelligence and Copyright, Docket No. 2023-6 (Oct. 30, 2023) at 4, available at https://www.ftc.gov/legal-library/browse/advocacy-filings/comment-federal-trade-commission-artificial-intelligence-copyright (emphasis added).

[11] See, e.g. Joe Caserta, Holger Harreis, Kayvaun Rowshankish, Nikhil Srinidhi, and Asin Tavakoli, The data dividend: Fueling generative AI, McKinsey Digital (Sept. 15, 2023), https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-data-dividend-fueling-generative-ai (“Your data and its underlying foundations are the determining factors to what’s possible with generative AI.”).

[12] See e.g. Tim Keary, Google DeepMind’s Achievements and Breakthroughs in AI Research, Techopedia (Aug. 11, 2023), https://www.techopedia.com/google-deepminds-achievements-and-breakthroughs-in-ai-research; See e.g. Will Douglas Heaven, Google DeepMind used a large language model to solve an unsolved math problem, MIT Technology Review (Dec. 14, 2023), https://www.technologyreview.com/2023/12/14/1085318/google-deepmind-large-language-model-solve-unsolvable-math-problem-cap-set/; See also A Decade of Advancing the State-of-the-Art in AI Through Open Research, Meta (Nov. 30, 2023), https://about.fb.com/news/2023/11/decade-of-advancing-ai-through-open-research/; See also 200 languages within a single AI model: A breakthrough in high-quality machine translation, Meta, https://ai.meta.com/blog/nllb-200-high-quality-machine-translation/ (last visited Jan. 18, 2023).

[13] See e.g. Jennifer Allen, 10 years of Siri: the history of Apple’s voice assistant, Tech Radar (Oct. 4, 2021), https://www.techradar.com/news/siri-10-year-anniversary; see also Evan Selleck, How Apple is already using machine learning and AI in iOS, Apple Insider (Nov. 20, 2023), https://appleinsider.com/articles/23/09/02/how-apple-is-already-using-machine-learning-and-ai-in-ios; see also Kathleen Walch, The Twenty Year History Of AI At Amazon, Forbes (July 19, 2019), https://www.forbes.com/sites/cognitiveworld/2019/07/19/the-twenty-year-history-of-ai-at-amazon/?sh=1734bcb268d0.

[14] See infra Section III.

[15] See e.g. Cédric Argenton & Jens Prüfer, Search Engine Competition with Network Externalities, 8 J. Comp. L. & Econ. 73, 74 (2012); Mark A. Lemley & Matthew Wansley, Coopting Disruption (February 1, 2024), https://ssrn.com/abstract=4713845.

[16] John M. Yun, The Role of Big Data in Antitrust, in The Global Antitrust Institute Report on the Digital Economy (Joshua D. Wright & Douglas H. Ginsburg, eds., Nov. 11, 2020) at 233, available at https://gaidigitalreport.com/2020/08/25/big-data-and-barriers-to-entry/#_ftnref50. See also e.g. Robert Wayne Gregory, Ola Henfridsson, Evgeny Kaganer, & Harris Kyriakou, The Role of Artificial Intelligence and Data Network Effects for Creating User Value, 46 Acad. of Mgmt. Rev. 534 (2020), final pre-print version at 4, available at http://wrap.warwick.ac.uk/134220) (“A platform exhibits data network effects if, the more that the platform learns from the data it collects on users, the more valuable the platform becomes to each user.”). See also Karl Schmedders, José Parra-Moyano & Michael Wade, Why Data Aggregation Laws Could be the Answer to Big Tech Dominance, Silicon Republic (Feb. 6, 2024), https://www.siliconrepublic.com/enterprise/data-ai-aggregation-laws-regulation-big-tech-dominance-competition-antitrust-imd.

[17] Nathan Newman, Search, Antitrust, and the Economics of the Control of User Data, 31 Yale J. Reg. 401, 409 (2014) (emphasis added). See also id. at 420 & 423 (“While there are a number of network effects that come into play with Google, [“its intimate knowledge of its users contained in its vast databases of user personal data”] is likely the most important one in terms of entrenching the company’s monopoly in search advertising…. Google’s overwhelming control of user data… might make its dominance nearly unchallengeable.”).

[18] See also Yun, supra note 17 at 229 (“[I]nvestments in big data can create competitive distance between a firm and its rivals, including potential entrants, but this distance is the result of a competitive desire to improve one’s product.”).

^[19] For a review of the literature on increasing returns to scale in data (this topic is broader than data network effects) see Geoffrey Manne & Dirk Auer, Antitrust Dystopia and Antitrust Nostalgia: Alarmist Theories of Harm in Digital Markets and Their Origins, 28 Geo Mason L. Rev. 1281, 1344 (2021).

[20] Andrei Hagiu & Julian Wright, Data-Enabled Learning, Network Effects, and Competitive Advantage, 54 RAND J. Econ. 638 (2023) (final preprint available at https://andreihagiu.com/wp-content/uploads/2022/08/Data-enabled-learning-Final-RAND-Article.pdf).

^[21] Id. at 2. The authors conclude that “Data-enabled learning would seem to give incumbent firms a competitive advantage. But how strong is this advantage and how does it differ from that obtained from more traditional mechanisms….”

^[22] Id.

^[23] Bruno Jullien & Wilfried Sand-Zantman, The Economics of Platforms: A Theory Guide for Competition Policy, 54 Info. Econ. & Pol’y 10080, 101031 (2021).

^[24] Daniele Condorelli & Jorge Padilla, Harnessing Platform Envelopment in the Digital World, 16 J. Comp. L. & Pol’y 143, 167 (2020).

[25] See Hagiu & Wright, supra note 21.

^[26] For a summary of these limitations, see generally Catherine Tucker, Network Effects and Market Power: What Have We Learned in the Last Decade?, Antitrust (Spring 2018) at 72, available at https://sites.bu.edu/tpri/files/2018/07/tucker-network-effects-antitrust2018.pdf. See also Manne & Auer, supra note 20, at 1330.

[27] See Jason Furman, Diane Coyle, Amelia Fletcher, Derek McAuley & Philip Marsden (Dig. Competition Expert Panel), Unlocking Digital Competition (2019) at 32-35 (“Furman Report”), available at https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/785547/unlocking_digital_competition_furman_review_web.pdf.

[28] Id. at 34.

[29] Id. at 35. To its credit, it should be noted, the Furman Report does counsel caution before mandating access to data as a remedy to promote competition. See id. at 75. That said, the Furman Report does maintain that such a remedy should certainly be on the table because “the evidence suggests that large data holdings are at the heart of the potential for some platform markets to be dominated by single players and for that dominance to be entrenched in a way that lessens the potential for competition for the market.” Id. In fact, the evidence does not show this.

^[30] Case COMP/M.9660 — Google/Fitbit, Commission Decision (Dec. 17, 2020) (Summary at O.J. (C 194) 7), available at https://ec.europa.eu/competition/mergers/cases1/202120/m9660_3314_3.pdf at 455.

^[31] Id. at 896.

[32] See Natasha Lomas, EU Checking if Microsoft’s OpenAI Investment Falls Under Merger Rules, TechCrunch (Jan. 9, 2024), https://techcrunch.com/2024/01/09/openai-microsoft-eu-merger-rules/.

^[33] Amended Complaint at 11, Meta/Zuckerberg/Within, Fed. Trade Comm’n. (2022) (No. 605837), available at https://www.ftc.gov/system/files/ftc_gov/pdf/D09411%20-%20AMENDED%20COMPLAINT%20FILED%20BY%20COUNSEL%20SUPPORTING%20THE%20COMPLAINT%20-%20PUBLIC%20%281%29_0.pdf.

[34] Amended Complaint (D.D.C), supra note 4, at ¶37.

[35] Amended Complaint (E.D. Va), supra note 4, at ¶8.

^[36] US Dep’t of Justice & Fed. Trade Comm’n, Merger Guidelines (2023) at 25, https://www.ftc.gov/system/files/ftc_gov/pdf/2023_merger_guidelines_final_12.18.2023.pdf.

[37] Competition and Mkts. Auth., Merger Assessment Guidelines (2021) at  ¶7.19(e), https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1051823/MAGs_for_publication_2021_–_.pdf.

[38] Furman Report, supra note 28, at ¶4.

[39] See e.g. Chris Westfall, New Research Shows ChatGPT Reigns Supreme in AI Tool Sector, Forbes (Nov. 16, 2023), https://www.forbes.com/sites/chriswestfall/2023/11/16/new-research-shows-chatgpt-reigns-supreme-in-ai-tool-sector/?sh=7de5de250e9c.

^[40] See Krystal Hu, ChatGPT Sets Record for Fastest-Growing User Base, Reuters (Feb. 2, 2023), https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/; Google: The AI Race Is On, App Economy Insights (Feb. 7, 2023), https://www.appeconomyinsights.com/p/google-the-ai-race-is-on.

^[41] See Google Trends, https://trends.google.com/trends/explore?date=today%205-y&q=%2Fg%2F11khcfz0y2,%2Fg%2F11ts49p01g&hl=en (last visited, Jan. 12, 2024) and https://trends.google.com/trends/explore?date=today%205-y&geo=US&q=%2Fg%2F11khcfz0y2,%2Fg%2F11ts49p01g&hl=en (last visited Jan. 12, 2024).

^[42] See David F. Carr, As ChatGPT Growth Flattened in May, Google Bard Rose 187%, Similarweb Blog (June 5, 2023), https://www.similarweb.com/blog/insights/ai-news/chatgpt-bard/.

^[43] See Press Release, Meta, Introducing New AI Experiences Across Our Family of Apps and Devices (Sept. 27, 2023), https://about.fb.com/news/2023/09/introducing-ai-powered-assistants-characters-and-creative-tools/; Sundar Pichai, An Important Next Step on Our AI Journey, Google Keyword Blog (Feb. 6, 2023), https://blog.google/technology/ai/bard-google-ai-search-updates/.

^[44] See Ion Prodan, 14 Million Users: Midjourney’s Statistical Success, Yon (Aug. 19, 2023), https://yon.fun/midjourney-statistics/. See also Andrew Wilson, Midjourney Statistics: Users, Polls, & Growth [Oct 2023], ApproachableAI (Oct. 13, 2023), https://approachableai.com/midjourney-statistics/.

^[45] See Hema Budaraju, New Ways to Get Inspired with Generative AI in Search, Google Keyword Blog (Oct. 12, 2023), https://blog.google/products/search/google-search-generative-ai-october-update/; Imagine with Meta AI, Meta (last visited Jan. 12, 2024), https://imagine.meta.com/.

^[46] Catherine Tucker, Digital Data, Platforms and the Usual [Antitrust] Suspects: Network Effects, Switching Costs, Essential Facility, 54 Rev. Indus. Org. 683, 686 (2019).

^[47] Manne & Auer, supra note 20, at 1345.

[48] See e.g. Stefanie Koperniak, Artificial Data Give the Same Results as Real Data—Without Compromising Privacy, MIT News (Mar. 3, 2017), https://news.mit.edu/2017/artificial-data-give-same-results-as-real-data-0303 (“[Authors] describe a machine learning system that automatically creates synthetic data—with the goal of enabling data science efforts that, due to a lack of access to real data, may have otherwise not left the ground. While the use of authentic data can cause significant privacy concerns, this synthetic data is completely different from that produced by real users—but can still be used to develop and test data science algorithms and models.”).

[49] See e.g. Rachel Gordon, Synthetic Imagery Sets New Bar in AI Training Efficiency, MIT News (Nov. 20, 2023), https://news.mit.edu/2023/synthetic-imagery-sets-new-bar-ai-training-efficiency-1120 (“By using synthetic images to train machine learning models, a team of scientists recently surpassed results obtained from traditional ‘real-image’ training methods.).

[50] Thibault Schrepel & Alex ‘Sandy’ Pentland, Competition Between AI Foundation Models: Dynamics and Policy Recommendations, MIT Connection Science Working Paper (Jun. 2023), at 8.

[51] Igor Susmelj, Optimizing Generative AI: The Role of Data Curation, Lightly (last visited Jan 15, 2024), https://www.lightly.ai/post/optimizing-generative-ai-the-role-of-data-curation.

[52] See e.g. Xiaoliang Dai, et al., Emu: Enhancing Image Generation Models Using Photogenic Needles in a Haystack , ArXiv (Sep. 27, 2023) at 1, https://ar5iv.labs.arxiv.org/html/2309.15807 (“[S]upervised fine-tuning with a set of surprisingly small but extremely visually appealing images can significantly improve the generation quality.”). See also Hu Xu, et al., Demystifying CLIP Data, ArXiv (Sep. 28, 2023), https://arxiv.org/abs/2309.16671.

[53] Lauren Leffer, New Training Method Helps AI Generalize like People Do, Sci. Am. (Oct. 26, 2023), https://www.scientificamerican.com/article/new-training-method-helps-ai-generalize-like-people-do/ (discussing Brendan M. Lake & Marco Baroni, Human-Like Systematic Generalization Through a Meta-Learning Neural Network, 623 Nature 115 (2023)).

^[54] Timothy B. Lee, The Real Research Behind the Wild Rumors about OpenAI’s Q* Project, Ars Technica (Dec. 8, 2023), https://arstechnica.com/ai/2023/12/the-real-research-behind-the-wild-rumors-about-openais-q-project/.

^[55] Id. See also GSM8K, Papers with Code (last visited Jan. 18, 2023), available at https://paperswithcode.com/dataset/gsm8k; MATH Dataset, GitHub (last visited Jan. 18, 2024), available at https://github.com/hendrycks/math.

^[56] Lee, supra note 55.

[57] Geoffrey Manne & Ben Sperry, Debunking the Myth of a Data Barrier to Entry for Online Services, Truth on the Market (Mar. 26, 2015), https://truthonthemarket.com/2015/03/26/debunking-the-myth-of-a-data-barrier-to-entry-for-online-services/ (citing Andres V. Lerner, The Role of ‘Big Data’ in Online Platform Competition (Aug. 26, 2014), available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2482780.).

[58] See e.g., Lemley & Wansley, supra note 18, at 22 (“Incumbents have all that information. It would be difficult for a new entrant to acquire similar datasets independently….”).

[59] See Catherine Tucker, Digital Data as an Essential Facility: Control, CPI Antitrust Chron. (Feb. 2020) at 11 (“[U]ltimately the value of data is not the raw manifestation of the data itself, but the ability of a firm to use this data as an input to insight.”).

[60] Or, as John Yun puts it, data is only a small component of digital firms’ production function. See Yun, supra note 17, at 235 (“Second, while no one would seriously dispute that having more data is better than having less, the idea of a data-driven network effect is focused too narrowly on a single factor improving quality. As mentioned in supra Section I.A, there are a variety of factors that enter a firm’s production function to improve quality.”).

^[61] Luxia Le, The Real Reason Windows Phone Failed Spectacularly, History–Computer (Aug. 8, 2023), https://history-computer.com/the-real-reason-windows-phone-failed-spectacularly/.

[62] Introducing the GPT Store, Open AI (Jan. 10, 2024), https://openai.com/blog/introducing-the-gpt-store.

^[63] See Michael Schade, How ChatGPT and Our Language Models are Developed, OpenAI, https://help.openai.com/en/articles/7842364-how-chatgpt-and-our-language-models-are-developed; Sreejani Bhattacharyya, Interesting innovations from OpenAI in 2021, AIM (Jan. 1, 2022), https://analyticsindiamag.com/interesting-innovations-from-openai-in-2021/; Danny Hernadez & Tom B. Brown, Measuring the Algorithmic Efficiency of Neural Networks, ArXiv (May 8, 2020), available at https://arxiv.org/abs/2005.04305.

[64] See Yun, supra note 17 at 235 (“Even if data is primarily responsible for a platform’s quality improvements, these improvements do not simply materialize with the presence of more data—which differentiates the idea of data-driven network effects from direct network effects. A firm needs to intentionally transform raw, collected data into something that provides analytical insights. This transformation involves costs including those associated with data storage, organization, and analytics, which moves the idea of collecting more data away from a strict network effect to more of a ‘data opportunity.’”).

[65] Lerner, supra note 58, at 4-5 (emphasis added).

[66] See Clayton M. Christensen, The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail (2013).

[67] See David J. Teece, Dynamic Capabilities and Strategic Management: Organizing for Innovation and Growth (2009).

[68] See Hagiu and Wright, supra note 21, at 4 (“We use our dynamic framework to explore how data sharing works: we find that it in-creases consumer surplus when one firm is sufficiently far ahead of the other by making the laggard more competitive, but it decreases consumer surplus when the firms are sufficiently evenly matched by making firms compete less aggressively, which in our model means subsidizing consumers less.”). See also Lerner, supra note 58.

[69] See e.g. Hagiu & Wright, id. (“We also use our model to highlight an unintended consequence of privacy policies. If such policies reduce the rate at which firms can extract useful data from consumers, they will tend to increase the incumbent’s competitive advantage, reflecting that the entrant has more scope for new learning and so is affected more by such a policy.”); Jian Jia, Ginger Zhe Jin & Liad Wagman, The Short-Run Effects of the General Data Protection Regulation on Technology Venture Investment, 40 Marketing Sci. 593 (2021) (finding GDPR reduced investment in new and emerging technology firms, particularly in data-related ventures); James Campbell, Avi Goldfarb, & Catherine Tucker, Privacy Regulation and Market Structure, 24 J. Econ. & Mgmt. Strat. 47 (2015) (“Consequently, rather than increasing competition, the nature of transaction costs implied by privacy regulation suggests that privacy regulation may be anti-competitive.”).

As Greg Werden has noted, the process of defining the relevant market in an antitrust case doesn’t just finger which part of the economy is allegedly affected by the challenged conduct, but it also “identifies the competitive process alleged to be harmed.” Unsurprisingly, plaintiffs in such proceedings (most commonly, antitrust enforcers) often seek to set exceedingly narrow parameters for relevant markets in order to bolster their case. In the extreme, these artificially constrained definitions sketch what can only be called “gerrymandered” markets—obscuring rather than illuminating the competitive processes at issue.

This unfortunate tendency is exemplified in the Federal Trade Commission’s (FTC) recent complaint against Amazon, which describes two relevant markets in which anticompetitive harm has allegedly occurred: (1) the “online superstore market” and (2) the “online marketplace services market.” Because both markets are exceedingly narrow, they grossly inflate Amazon’s apparent market share and minimize the true extent of competition. Moreover, by lumping together wildly different products and wildly different sellers into single “cluster markets,” the FTC misapprehends the nature of competition relating to the challenged conduct.

Read the full piece here.

Introduction

Market definition is a critical component of any antitrust case. Not only does it narrow consideration to a limited range of relevant products or services but, perhaps more importantly, it specifies a domain of competition at issue in an antitrust case—that is, the nature of the competition between certain firms that might (or might not) be harmed by the conduct of the defendant. As Greg Werden has characterized it:

Alleging the relevant market in an antitrust case does not merely identify the portion of the economy most directly affected by the challenged conduct; it identifies the competitive process alleged to be harmed.[1]

Unsurprisingly, plaintiffs—not least, antitrust agencies—are often tempted to define artificially narrow markets in order to reinforce their cases (sometimes, downright ridiculously so[2]). The consequence is not merely to artificially inflate the market significance of the firm under scrutiny, although it does do that; it is also to misapprehend and misdescribe the true nature of competition relevant to the challenged conduct.

This unfortunate trend—allegations of harm to artificially constrained and gerrymandered markets—is exemplified in the Federal Trade Commission’s (FTC) recent proceedings against Amazon.

The FTC’s complaint against Amazon describes two relevant markets in which anticompetitive harm has allegedly occurred: (1) the “online superstore market” and (2) the “online marketplace services market.”[3]

Unfortunately, both markets are excessively narrow, thereby grossly inflating Amazon’s apparent market share and minimizing the true extent of competition. Moreover, the FTC’s approach to market definition here—lumping together wildly different products and wildly different sellers into single “cluster markets”—grossly misapprehends the nature of competition relating to the challenged conduct.

First, the FTC’s complaint limits the online-superstore market to online stores only, and further limits it to stores that have an “extensive breadth and depth”[4] of products. The latter means online stores that carry virtually all categories of products (“such as sporting goods, kitchen goods, apparel, and consumer electronics”[5]) and that also have an extensive variety of brands within each category (such as Nike, Under Armor, Adidas, etc.).[6] In practice, this definition excludes leading brands’ private channels (such as Nike’s online store),[7] as well as online stores that focus on a particular category of goods (such as Wayfair’s focus on furniture).[8] It also excludes the brick-and-mortar stores that still account for the vast majority of retail transactions.[9] Firms with significant online and brick-and-mortar sales might count, but only their online sales would be considered part of the market.

Second, the online-marketplace-services market is limited to online platforms that provide access to a “significant base of shoppers”;[10] a search function to identify products; a means for the seller to set prices and present product information; and a method to display customer reviews. This implies that current Amazon sellers can’t reach consumers through mechanisms that don’t incorporate all these specific functions, even though consumers regularly use multiple services and third-party sites that accomplish the same thing (e.g., Google Shopping, Shopify, Instagram, etc.)[11] Moreover, it implies that these myriad alternative channels do not constrain Amazon’s pricing of its services.

Documents identified in the complaint do appear to demonstrate that Amazon pays substantial attention to competition from online superstores and online marketplaces. But cherry-picked business documents do not define economically relevant markets.[12] At trial, Amazon will doubtless produce a host of ordinary-course documents that show significant competition from a wide array of competitors on both sides of its retail platform. The scope of competition that the FTC sketches—based on a few documents from among tens of thousands—is a public-relations and litigation tactic, but not remotely the full story.

Third, the FTC’s casual use of “cluster markets,” which lump together distinct types of products and different types of sellers into single markets, may severely undermine the commission’s case. It’s one thing to group, say, all recorded music into a single market (despite the lack of substitutability between, say, death metal and choral Christmas music), but it’s another thing entirely to group batteries and bedroom furniture into a single “market,” just because Amazon happens to facilitate sales of both.

Fourth and finally, it is notable that the relevant markets alleged in the FTC’s complaint draw a distinct line between the seller and buyer sides of Amazon’s platform. Implicit in this characterization is the rejection of cross-market effects as a justification for Amazon’s business conduct. Some of the FTC’s specific concerns—e.g., the alleged obligation imposed on sellers to use Amazon’s fulfillment services to market their products under Amazon’s Prime label—have virtually opposite implications for the seller and buyer sides of the market. Arbitrarily cordoning off such conduct to one market or the other based on where it purportedly causes harm (and thus ignoring where it creates benefit) mangles the two-sided, platform nature of Amazon’s business and would almost certainly lead to its erroneous over-condemnation.[13]

Ultimately, what will determine the scope of the relevant markets will be economic analysis based on empirical data. But based on the FTC’s complaint, public data, and common sense (the best we have to go on, for now), it seems implausible that the FTC’s conception of distinct, and distinctly narrow, relevant markets will comport with reality.

An artificially narrow and gerrymandered market definition is a double-edged sword. If the court accepts it, it’s much easier to show market power. But the odder the construction, the more likely it is to strain the court’s credulity. The FTC has the burden of proving its market definition, as well as competitive harm. By defining these markets so narrowly, the FTC has ensured it will face an uphill battle before the courts.

I.           The Alleged ‘Online Superstore’ Market

A first weakness of the FTC’s suit pertains to the alleged “online superstore market.” This market definition excludes the following: (1) brick-and-mortar retailers, (2) brick-and-mortar sales by firms that do considerable business online and in-person, and (3) online retailers that don’t meet the definition of a “superstore.”[14] The FTC’s market definition also excludes sales of perishable grocery items.[15] The agency argues that consumers don’t consider these other types of retailers to be substitutes for online superstores.[16] This seems dubious, and the FTC’s complaint does little to dispel the doubt.

To see how the market definition tilts the balance, consider the FTC’s allegation that Amazon dominates the online-superstore market with approximately 82% market share.[17] That is, Amazon is reported to have approximately 82% market share (in gross merchandise value, or “GMV”), provided we exclude perishables, and consider the market to comprise solely U.S. online sales by Amazon, Walmart, Target, and eBay, but no other vendors. Note, for example, that Walmart, Target, and Costco all have both online and in-person sales at brick-and-mortar stores, but Costco’s online sales are excluded from the online-superstore category, presumably due to their relatively limited scale and scope. But counting both online and in-person sales, it turns out that twelve-month trailing revenue at Costco is reported to be more than double that of Target, which is included in the FTC’s online-superstore category.[18] Amazon’s share of overall online retail is substantial, but it’s much smaller (37.6%) than its share of a purported market that comprises Amazon, Walmart online, Target online, eBay, and nobody else.[19] Indeed, if one includes total retail sales, then Walmart leads Amazon, not vice versa.[20] And while e-commerce may be substantial and growing, it still represents only about 15% of U.S. retail.[21]

There are countless examples where consumers cross-shop online and offline—televisions and other electronics, clothing, and sporting goods (among many others) spring to mind. Indeed, most consumers would surely be hard-pressed to identify any product they’ve purchased from Amazon that they have not, at some point, also purchased from an offline or non-superstore retailer.

Defining a market with reference to a single retailer’s particular product offering—that is, by a single channel of distribution—is unlikely to “identif[y] the competitive process alleged to be harmed.”[22] In fact, for consumers, it doesn’t identify a product at all, and ends up excluding a host of competing sellers that offer economic substitutes for the products consumers actually buy.[23] By failing to do so, the FTC’s purported market definition is woefully deficient in describing the scope of competition: “Including economic substitutes ensures that the relevant product market encompasses ‘the group or groups of sellers or producers who have actual or potential ability to deprive each other of significant levels of business.’”[24]

A.        Brick-and-Mortar Competes with Amazon Because Shopping Is Not the Same Thing as Consuming

While it may be that some consumers do not consider offline vendors or non-superstores to be substitutes, it does not follow that such rivals don’t impose competitive constraints on online superstores.

If a hypothetical monopolist raises prices, some consumers—perhaps many, perhaps even most—may switch to a brick-and-mortar retailer. That may be enough to constrain the monopolist’s pricing. How many might switch, and the extent to which that constrains pricing, are empirical questions, but there is no question that some consumers might switch: retail multi-homing is common.

And the constraints on switching are far weaker than the FTC claims. The complaint observes that 1) brick-and-mortar retailers are less convenient because it takes time to go to a physical store, 2) stores are not open for shopping at all hours, and 3) consumers may have to visit multiple stores to buy the necessary items.[25]

Online shopping is almost certainly quicker than offline—at least, once one is sitting in front of a computer with Internet access. But the complaint seems to conflate shopping with consuming.

Even with Amazon’s impressive fulfillment and delivery network, if a consumer needs a product that very moment or even that day, a brick-and-mortar retailer may be preferable. The same may be true in circumstances in which a consumer wants to see a product in person, try on clothing, consult an experienced salesperson, etc. And while some consumers may enjoy shopping, they may or may not prefer the experience of online shopping.

More generally and more to the point, consumers purchase goods to use and consume them. Online stores may be “always open,” but shipping and delivery are not instantaneous. That one can shop online at all hours may be convenient, but it may do nothing to hasten the ability to consume the items purchased.

Meanwhile, brick-and-mortar retailers typically have websites that show their inventory and pricing online. Consumers can, accordingly, comparison shop across e-commerce and brick-and-mortar vendors, even when the brick-and-mortar retailers have closed for the evening.

B.         ‘Depth and Breadth’ Isn’t Solely Available from Superstores, and Consumers Buy Products, Not Store Types

Consumers within the “online superstore market” may be able to prevent a hypothetical monopolist from raising prices by switching to other online channels that don’t qualify as a “superstore,” as defined by the FTC.

For example, if a consumer is looking for sporting goods, she can shop at an online superstore, or she can shop at Dick’s online, REI online, or Bass Pro online, all of which have an exceptional “depth and breadth” of items.[26] Alternatively, if the consumer is shopping for a Columbia Sportswear jacket, in addition to the sporting-goods retailers listed, she can also shop on Columbia’s website[27] or at any other online-clothing retailer that carries Columbia jackets (e.g., Macy’s or Nordstrom[28]).

The complaint anticipates and responds to this concern by saying that non-superstore online retailers (as well as brick-and-mortar retailers) lack the depth and breadth of products sold by superstores.[29] But so what? For many consumers, Amazon purchases are made one (or a few) item(s) at a time. When consumers need a bolt cutter, they log in and order it, and when they need a pair of sneakers the next day, they log in and order that. They don’t wait to buy the bolt cutter until they are ready to buy sneakers (i.e., people don’t typically log in to Amazon with a shopping list and purchase multiple items at the same time, except perhaps for perishable groceries, which are excluded from the proposed market). Whether the consumer is buying one item or three or five, a purchase that bundles products across the broad scope of the online-superstore market is not at all the norm.

Indeed, part of the purported advantage of online shopping—when it’s an advantage—is that consumers don’t have to bundle purchases together to minimize the transaction costs of physically visiting a brick-and-mortar retailer. Meanwhile, another part of the advantage of online shopping is the ease of comparison shopping: consumers don’t even have to close an Amazon window on their computers to check alternatives, prices, and availability elsewhere. All of this undermines the claim that one-stop shopping is a defining characteristic of the alleged market.

Data are hard to come by (and the data will ultimately demonstrate whether and to what extent the complaint portrays reality), but public sources indicate that the average number of units per transaction is less than three (admittedly, this is worldwide, and for all online e-commerce, not just Amazon).[30] This does not suggest that shoppers demand extensive “depth and breadth” each time they shop online.

Meanwhile, important lacunae in Amazon’s offerings belie the notion that it offers a true “depth and breadth” that transcends competitive constraints from other retailers. The fact that Nike, on the seller side, doesn’t view Amazon as an essential marketplace[31]—in other words, it believes it has plenty of alternative, competing channels of distribution—has important consequences for the FTC’s market definition on the consumer side. It’s difficult to conceive of a retailer offering anything approaching a comprehensive “depth and breadth” of footwear without offering any Nike shoes. For consumers who buy shoes, Amazon is hardly a unique outlet, and finding even a minimally suitable range of options requires shopping elsewhere, either in combination with Amazon or in its stead.

But the implications are even greater. Because the FTC has grouped sales of all products together—not just footwear or even apparel—and defined the relevant market around that broad clustering of disparate products, can it really be said that Amazon is a “one-stop-shop” at all if it doesn’t offer Nike shoes?

The example may seem trivial, but it aptly illustrates the inherent error in defining the product market essentially by the offerings of a single entity. Necessarily, those offerings will be unique and affected by a host of seller/buyer interactions specific to that company. And in many cases, those specific inclusions and exclusions may be significantly more important than the simple number of SKUs on offer (which is essentially the basis for including Walmart and Target online, but excluding, say, Costco online from the FTC’s “superstores” market).

Further, despite its repeated reliance on “depth and breadth,” the complaint ignores e-commerce aggregators, which allow consumers to search products and pricing across an incredible variety of retailers. Google Shopping is, of course, the most notable example—and, for such a prominent example, curiously absent from the complaint. Through Google Shopping—among other sites—consumers can see extensive results in one place for almost any product, including across all categories and across many brands (the breadth-and-depth factors relied upon by the complaint). Indeed, while many product searches today begin at Amazon, a huge amount of online shopping takes place via Google.[32]

Moreover, online shoppers regularly use third-party sites to research (shop) for products, and these, too, aggregate information from across a huge range of sources. As Search Engine Land reports:

Reviews and ratings can make or break a sale more than any other factor, including product price, free shipping, free returns and exchanges, and more.

Overall, 77% of respondents said they specifically seek out websites with reviews—and this number was even higher for Gen Z (87%) and millennials (81%).[33]

While Amazon is where consumers most often read reviews (94%), other retail websites (91%), search engines (70%), brand websites (68%), and independent review sites (40%) are all significant.[34] And yet, despite their manifest importance in the competitive process of online retail, the FTC’s complaint entirely dismisses the significance of shopping aggregators and non-Amazon, product-review sources.

II.         The Alleged ‘Online Marketplace Services’ Market

The complaint is similarly flawed when it assesses the scope of competition from the point of view of sellers.

The complaint endeavors to distinguish and exclude from the market for online marketplace services all other methods by which a seller can market and sell its products to end consumers. For instance, the complaint distinguishes online marketplaces from online retailers where the seller functions as a vendor (i.e., it transfers title to the retailer) and those where sellers provide their own storefronts or sell directly through social media and other aggregators using “software-as-a-service” (“SaaS”) to market products (e.g., Shopify and BigCommerce).[35]

The complaint alleges that neither operating as a vendor nor utilizing SaaS is “reasonably interchangeable”[36] with online marketplace services—the key language from the Brown Shoe case.[37] But merely saying so does not make it true. Service markets can display differentiated competition, just as product markets do. Superficial—and even significant—differences among services do not, in themselves, establish that they are not competitors.

First, where sellers operate as vendors by transferring title to another party to sell the product (either online or at a brick-and-mortar retailer), they could very well constrain the costs that a hypothetical monopolist imposes on sellers. For example, if a hypothetical monopolist increased prices or decreased quality for selling a product, why would Nike not transfer its products away from the monopolist and toward Foot Locker, Macy’s, or any other number of retailers where Nike operates as a vendor? Or why not rely on Nike’s own website, selling directly to the consumer? In fact, Nike has already done this. In 2019, Nike stopped selling products to Amazon because it was dissatisfied with Amazon’s efforts to limit counterfeit products.[38] Instead, Nike opted to sell directly to its consumers or through its other retailers (both online and offline, of course).

The same can be said for sellers without well-known brands or those who opt to use SaaS to sell their products. Certainly, there are differences between SaaS and online-marketplace services, but that doesn’t mean that a seller can’t or won’t use SaaS in the face of increased prices or decreased quality from an online marketplace. Notably, Shopify claims to be the third-largest online retailer in the United States, with 820,000 merchants selling through the platform.[39] It’s remarkable that it is completely absent from the FTC’s market definition.

Also remarkable is that he FTC’s complaint alleges that SaaS providers are not in the relevant market because:

SaaS providers, unlike online marketplace service providers, do not provide access to an established U.S. customer base. Rather, merchants that use SaaS providers to establish direct-to-consumer online stores must invest in marketing and promotion to attract U.S. shoppers to their online stores.[40]

This is remarkable because a significant claim in the FTC’s complaint is that Amazon has “degraded” its service by introducing sponsored search results, “litter[ing] its storefront with pay-to-play advertisements,” and allegedly requiring (some would say enabling…) sellers to pay for marketing and promotion.[41] It’s unclear why the need to invest in marketing and promotion to attract shoppers to one’s online storefront is qualitatively different than the need to invest in marketing and promotion to attract shoppers to one’s products on Amazon’s platform.

Indeed, the notion that large platforms like Amazon simply “provide access” to consumers glosses over the immense work that such access entails. Amazon and similar platforms (including, of course, SaaS providers) make significant investment in designing and operating user interfaces, matching algorithms, marketing channels, and innumerable other functionalities to convert undifferentiated masses of consumers and sellers into a functional retail experience. Amazon’s value for sellers in providing access to customers must be balanced by the reality that, in doing so, large “superstores” like Amazon also necessarily put a large quantity of disparate sellers in the same unified space.

For obvious reasons, sellers don’t necessarily value selling their products in the same location as other sellers. They do, of course, want access to consumers, but the “marketplace” or “superstore” aspects of Amazon simultaneously impedes that access by congesting it with other sellers and products (and consumers seeking other products). A specialized outlet may, in fact, offer the optimal sales environment: all consumers seeking the seller’s category of goods (but somewhat fewer consumers), and fewer sellers impeding discovery and access (though more selling the same category of goods). A furniture seller may have dozens of online outlets (and, of course, many offline outlets, catalog sales, decorator sales, etc.), and there is little or no reason to think that, by virtue of also offering batteries, clothes, and bolt cutters, Amazon offers anything truly unique to a furniture seller that it can’t get by selling through another distribution channel with a different business model.

The complaint relies heavily on this notion that online-marketplace services deliver a large customer base that cannot be matched by selling as a vendor or using SaaS. (It is entirely unclear if the FTC considers single-category online marketplaces like Wayfair to be in the “online marketplace services” market, a topic to which I return below in the “cluster markets” discussion; it is clear the FTC doesn’t consider Wayfair part of the “online superstores market.”).[42] Again, in this context, the complaint ignores e-commerce aggregators and how they affect sellers’ ability to access customers. Through Google Shopping, consumers can see extensive results for almost any product, including across all categories and across many brands. And Google aggregates product listings without charging the seller.[43] Thus, through Google Shopping, a seller can access a large consumer base that may constrain a hypothetical monopolist in the online-marketplace-services market.

And Google Shopping is not alone. Selling through social media has boomed. According to one source, Instagram is an online-shopping juggernaut.[44] Among other things:

• 130 million people engage with shoppable Instagram posts monthly;
• 72% of users say they made a purchase based on something they saw on Instagram;
• 70% of Instagram users open the app in order to shop; and
• 81% of Instagram users research new products and services on the platform.[45]

Sellers on Instagram can use Meta’s “Checkout on Instagram”[46] service to process orders directly on Instagram, as well as logistics services like Shopify or ShipBob to manage their supply chains and fulfill sales,[47] replicating the core functionality of a vertically integrated storefront like Amazon.

The bottom line is that Amazon is not remotely the only (or, in many cases, even the best) place for sellers to find, market, and sell to consumers. Its superficial differences from other distribution channels are just that: superficial.

III.       Cluster Markets

One of the most important problems with the FTC’s alleged relevant markets is that they treat all products and all sellers the same. They effectively assume that consumers shop for bolt cutters the same way they shop for furniture, and that Adidas sells shoes the same way that drop-shippers sell toilet paper.

Courts have recognized that such an approach—using “cluster markets” to assess a group of disparate products or services in a single market—can be appropriate for the sake of “administrative[ ]convenience.” As the 6th U.S. Circuit Court of Appeals noted in Promedica Health v. FTC, “[t]his theory holds, in essence, that there is no need to perform separate antitrust analyses for separate product markets when competitive conditions are similar for each.”[48]

A second basis for clustering is the “transactional-complements” theory, relabeled by the 6^th Circuit as the “‘package-deal’ theory.”[49] This approach clusters products together for relevant market analysis when “‘most customers would be willing to pay monopoly prices for the convenience’ of receiving certain products as a package.”[50]

For example, it may be appropriate to refer to a “market for recorded music” even though consumers of music by Taylor Swift probably exert little or no competitive pressure on the price or demand for recordings of, say, Cannibal Corpse. Thus, in the EU’s 2012 clearance (with conditions) of the Universal Music Group/EMI Music merger, the Commission determined that, although classical music may present somewhat different competitive dynamics, there was no basis for defining separate markets by artist or even by genre.[51]

Hospital mergers provide another classic example.[52] Labor and delivery services are not a substitute for open-heart surgery, but the FTC nonetheless frequently defines a market as “inpatient general acute care services” or something similar because of the similar relationship of each to a hospital’s organization and administration, as well as the fact that payers typical demand such services (and hospitals typically provide such services) in combination (even though patients, of course, do not consume them together).

The Supreme Court put its imprimatur on the notion of a cluster market in Philadelphia National Bank, accepting the lower court’s determination that “commercial banking” constituted a relevant market because of the distinctiveness, cost advantages, or consumer preferences of the constituent products.[53]

A.        Assessing Cluster Markets

Widespread use (and the occasional fairly serious analysis) of cluster markets notwithstanding, it is worth noting that the economic logic of such markets is, at best, poorly established.

In the UMG/EMI case, for example, the Commission rested on the following factors in concluding that markets should not be separated out by genre (let alone by artist):

The market investigation showed that, by and large, a segmentation of the recorded music market based on genre is not appropriate. First, the borders between genres are often blurred and artists and songs can fit within several genres at the same time. Second, several customers also underline that placing of a song or an album into a specific genre is entirely subjective. Third, a vast majority of customers indicated that they purchase and sell all genres of music.[54]

These facts may all be true, but they do little to permit the inference drawn. Indeed, the first two factors arguably refer only to administrability, not economic reality, and the third is woefully incomplete (e.g., it says little about a potential monopolist’s ability to raise prices if price increases can be passed on to end-consumers in some genres but not others). While the frailties of the market determination may not ultimately have mattered in that case (after all, the parties got their merger, and the Commission presumably brought the strongest case it could), such casual conclusions may well prove problematic elsewhere and do little to advance the logic of the cluster-markets concept.

Similar defects plague the Supreme Court’s endorsement of the theory in PNB. The Court suggests some reasons why, even in its own telling, “some commercial banking products or services”[55] may be insulated from competition, but that still leaves open the possibility that others aren’t, and that the relevant insulating characteristics could be eroded by simple product repositioning, different pricing strategies, or changes in reputation and brand allegiance.

In fact, the defendants in PNB argued before the district court that:

commercial banking in its entirety is not a product line. Rather, they submit it is a business which has two major subdivisions—the acceptance of deposits in which the bank is the debtor, and the making of loans in which the bank is the creditor. Both of these major divisions are further divided by distinct types of deposits and loans. As to many of these functions, there are different types of customers, different market areas, and, most importantly, different types of competitors and competition. With the possible exception of demand deposits, there is an identical or effective substitute for each one of the services which a commercial bank offers.[56]

The court, however, rejected these arguments with little more than a wave of the hand (a conclusion that was then simply accepted by the Supreme Court):

It seems quite apparent that both plaintiff’s and defendants’ positions have some merit. However, it is not the intention of this Court to subdivide a commercial bank into certain selected services and functions. An approach such as this, carried to the logical extreme, would result in many additional so-called lines of commerce. It is the conglomeration of all the various services and functions that sets the commercial bank off from other financial institutions. Each item is an integral part of the whole, almost every one of which is dependent upon and would not exist but for the other. The Court can perceive no useful purpose here in going any further than designating commercial banking a separate and distinct line of commerce within the meaning of the statute. It is undoubtedly true that some services of a commercial bank overlap, to some degree, with those of certain other institutions. Nevertheless, the Court feels quite confident in holding that commercial banking, viewed collectively, has sufficient peculiar characteristics which negate reasonable interchangeability.[57]

None of this response goes to the question of how users of commercial-banking services consume them. Instead, it essentially takes the superficial marketing distinction as economically dispositive, despite the acknowledgment that economic substitutes for the constituent products exist. It is, of course, possible that, in PNB, the error was not outcome determinative; perhaps none of the overlap between commercial banks and other providers of commercial lending is significant enough to change the analysis. But this is not a rigorous defense of the notion.

In a few cases, a more rigorous econometric analysis has been used to establish the viability of cluster markets. Consider, for example, the FTC’s successful challenge of the proposed Penn State Hershey Medical Center/Pinnacle Health System merger.[58] At issue there were the likely effects of a merger for certain services provided by general acute care (GAC) hospitals—that is, a range or “cluster” of services sold to commercial health plans in a defined geographic area covering roughly four counties in central Pennsylvania. Two small community hospitals offered some of the same acute care services, and various clinics and group practices provided some of the primary and secondary care services in the cluster.

At the same time, there was evidence that commercial health plans needed to negotiate for coverage over a range of GAC services that other providers could not offer, and that the merging parties competed on price in such negotiations with commercial health plans. Copious econometric evidence—analysis of price data and patient-draw data—substantiated the FTC’s market definition, bolstered by an amicus brief filed by more than three dozen experts in antitrust, competition, and health-care economics.[59]

All of this supported the FTC’s argument that the provision of GAC services constituted a single “cluster market”—and the 3rd U.S. Circuit Court of Appeals agreed, overturning a flawed geographic-market definition initially adopted by the district court.[60] That is, the agency didn’t merely waive its hands at an impression of ways that certain hospital services were similar to each other; rather, it provided detailed economic analysis of the price competition at issue for a specific range of GAC hospital services.

Notably, in that case, there were specific, identifiable consumers—commercial health plans—that were negotiating prices for a diverse “cluster” of GAC services. An individual patient will not, we hope, need to shop for oncology, cardio-thoracic surgery, a hip replacement, and ob-gyn services at the same time. But a health plan typically considers all of those and more. The same dynamic is not, of course, applicable in the Amazon case.

Perhaps the best example of the rigorous defense of cluster markets came in the first Staples/Office Depot merger matter, where ordinary-course documents played a role in the FTC’s review, but were by no means core to the staff’s analysis.[61] The FTC Bureau of Economics applied considerable econometric analysis of price data to establish that office superstore chains constrained each other’s pricing in a way that other vendors of office supplies did not.[62] That analysis of price effects (as evidence of likely merger effects and as evidence on behalf of the FTC’s market definition) is not apparent in the district court’s opinion enjoining the transaction.[63] But it figured heavily in the FTC’s presentation of the case and, presumably, in the commission’s internal decision to bring the case.

Two things are particularly notable about the cluster markets employed in Staples/Office Depot. First is that the exercise was undertaken at all. That is, it was assumed to be a crucial question whether other types of retailers (those with fewer products or catalog-only sales) constrained the pricing power of office-supply “superstores.” Second, the groupings of products analyzed were based on detailed analyses of pricing and price sensitivity over identified products, not superficial, subjective impressions of the market. The same was likewise the case in the Penn State Hershey hospital case mentioned above, and in other hospital-merger cases.

These types of evidence and analyses are simply not in evidence in the FTC’s case against Amazon—certainly not as they’ve presented it thus far.

B.         The Problem of Cluster Markets in the FTC’s Amazon Complaint

The FTC’s approach to market definition in Amazon appears in sharp contrast with prior cases involving what were, arguably, valid cluster markets and somewhat narrow market definitions.

Although the Amazon case is only at the complaint stage, of course, no factors or analysis similar to those adduced in the hospital and office-superstore cases discussed above are present in the FTC’s complaint against Amazon. Indeed, the complaint offers no evidence that the FTC considered the possibility that different products and different sellers would need to be considered separately (the FTC certainly saw no need to preemptively defend its clustering in the complaint). Instead—and consistent with the apparent assumption that Amazon and its particular characteristics are virtually unique—the complaint appears to assume that if Amazon offers a grouping of products, or if Amazon offers services to different types of sellers, this constitutes an economically rigorous “relevant market.” (Spoiler alert: It does not.)

Such an assumption would seem to need some defense. Certainly, a customer buying a bolt cutter will not consider buying a sneaker to be a reasonable alternative; it is clearly not on the basis of demand substitution that the FTC lumps these products together.[64] Instead, similar competitive conditions across products are implicit in the FTC’s alleged markets. But are competitive conditions sufficiently similar across products sold on Amazon to justify clustering them?

1.           Buyer-side clustering

Conditions vary considerably across the broad swath of products sold on Amazon. For some products sold at online superstores, brick-and-mortar retailers are a much closer substitute. Conceivably, consumers may prefer buying shoes at a brick-and-mortar retailer so that they can try them on, making physical retail a closer substitute for sneakers than for, say, a toilet brush, where very few consumers will demand to try the brush for balance before buying it. And surely consumers may be more willing to buy well-established brands (Nike, Gucci, etc.) directly from the brand’s website than a lesser-known brand sold at an online superstore.

Furniture, for example, is bought and sold in vastly different ways than, say, batteries (by consumers with different preferences for service and timing, by retailers with different relationships with manufacturers, through different channels of distribution, etc.). Whatever the merits to consumers of bundling purchases together from an “online superstore,” it is likely the case that they far less often bundle furniture purchases with other purchases than they do batteries. And surely consumers far more often seek to buy furniture offline or after testing it out in person than they do batteries. Vertically integrated furniture stores like IKEA have certainly done much to “commoditize” the production and sale of furniture in recent decades, but the market remains populated mostly by independent furniture showrooms, traditional manufacturers, and catalog and decorator sales. The same cannot be said for batteries, of course.

It also seems unlikely that consumers purchase Amazon’s proffered products in bundles meaningfully distinct from those they purchase elsewhere. People shopping for kitchen pantry items may well bundle their purchases of these items together. But in the vast majority of cases, they can get that same bundle from a grocery store, even though the grocery store carries many fewer SKUs overall. There is no analog to commercial health plans negotiating prices for a particular “cluster” of hospital services in Amazon’s case—and even if there were, it is certain that any number of other stores can match the actual clusters in which people regularly buy products from Amazon.

2.           Seller-side clustering

The problem of false clustering is even more acute on the seller side in the alleged “online marketplace services” market. Sellers on Amazon comprise at least two distinct types. On the one hand are brands and manufacturers that have a limited range of their own products to offer. These sellers are not resellers of others’ goods, but product creators or brands that use Amazon to sell “direct to consumer” the same sort of products they might otherwise have to sell through a retail intermediary. Within this group there is a further distinction between large, known brands and entrepreneurs selling a unique product (or maybe a few unique products) of their own creation out of their proverbial garage.[65]

On the other hand are retailers—resellers—that offer a wide range of products, none of which they manufacture themselves, but which they may purchase in bulk from manufacturers or offer through drop-shipping. The seller is an intermediary between the actual maker or seller of the product and the customer (in this case, marketing and reaching customers through another intermediary: Amazon). Here, again, there is a further distinction between intermediaries that are virtually invisible or interchangeable pass-throughs of others’ goods and those that attempt to add some value by establishing their own private-label brands or by acting as a trusted intermediary that offers a curated set of products.

Each of these types of sellers has a different demand for the various services bundled by Amazon, and a different set of available alternatives to Amazon. They often compete in different markets, have different relationships with manufacturers, and have differing sets of internal capacities necessitating the purchase of different services (or the purchase of different services in different relative quantities), and entailing a different ability to evaluate their need for different services and differing degrees of reliance on Amazon to complement their capacities. Moreover, the competitive ramifications of constraining each’s ability to sell on Amazon (or increasing the price to do so) is considerably different.

This last point is most obvious when considering the effect on drop-shippers of a possible increase in price on Amazon. What would be the competitive effects if a particular drop-shipper of, say, toilet paper were somehow precluded from Amazon, or harmed by using it? In that case, the seller is largely irrelevant (or worse—simply an additional source of markup). The relevant question is not whether a particular seller can profitably sell the product: “The antitrust laws… were enacted for ‘the protection of competition not competitors.’”[66] Rather, the relevant question is whether the manufacturer of the product can access consumers, and whether consumers can access competing sellers. In the case of toilet paper (or virtually anything else drop-shipped), the answer is manifestly yes. Drop shippers of Charmin could probably disappear completely from Amazon, and consumers would still be able to buy it at competitive prices from Amazon, among a host of competing options, and Proctor & Gamble would have no trouble reaching consumers.

3.           Implications

The implication of all this is that it seems highly dubious that furniture and batteries (to take just one example) face similar enough competitive conditions across online superstores for them to be grouped together in a single “cluster market.” While there may be superficial similarities in the website or technology connecting buyers and sellers, the underlying economics of production, distribution, and consumption seem to vary enormously.

The complaint offers no evidence to support the assertion of similar competitive conditions; no analysis of cross-elasticities of demand or supply across product categories; and no empirical evidence that a price increase for, say, furniture, could be offset by increased sales of batteries. Nor does the complaint consider more granular markets—like furniture, or sporting goods, or books—that would better capture these critical differences.

Indeed, it’s quite possible that narrower markets would demonstrate that Amazon faces real competition in some areas but not others. Grouping disparate products together risks obscuring situations where market power—and thus potentially anticompetitive effects from Amazon’s conduct—might exist in some product spaces but not others. The failure to properly define the relevant market for antitrust analysis doesn’t inherently imply a particular outcome; it just means no outcome can properly be determined.

The FTC offers no defense for clustering beyond the mere fact that Amazon offers these varied products on its platform. Yet selling through a common intermediary hardly establishes that the underlying competition is sufficiently similar to warrant single-market treatment, let alone that common conduct toward sellers affects all products and sellers equally. If the FTC cannot empirically defend treating distinct products as competitively interchangeable, as transactional complements, or as having the same competitive conditions, its case may collapse under the weight of its own market gerrymandering.

IV.      Out-of-Market Effects

This leaves a final question about the two markets defined in the complaint: can and should they really be considered separately, when conduct in each market has significant effects in the other? My colleagues and I intend to address this question more broadly and in more detail in the future (and, indeed, have already begun to do so[67]). For now, I will share a few tantalizing thoughts about this issue.

If Amazon’s practices vis-à-vis sellers cause the sellers to lower their prices, improve the quality of the products available through the marketplace, or otherwise lower costs and whittle down the seller’s profits, then consumers would benefit. Similarly, if Amazon’s practices with sellers improve the quality of consumers’ experience on its marketplace, then consumers would also benefit. The question is whether gain on one side should offset any harms on the other.

The FTC contends that the markets should be considered separately, despite acknowledging (and even trying to bolster its case with) the reality that the two sides of Amazon’s platform have important effects on each other:

Feedback loops between the two relevant markets further demonstrate the critical importance of scale and network effects in these markets. While the markets for online superstores and online marketplace services are distinct, an online superstore may operate an online marketplace and offer associated online marketplace services to sellers. As a result, the relationship and feedback loops between the two relevant markets can create powerful barriers to entry in both markets.[68]

Despite this, the FTC will likely contend that out-of-market efficiencies are not cognizable. That is, benefits to consumers in the online-superstore market that flow from harm in the online-marketplace-services market do not apply (i.e., harm is harm, and it doesn’t matter if it benefits someone else). This approach, however, presents some obvious problems.

If platforms undertake conduct to maximize the overall value of the platform (and not merely the benefits accruing to any one side in particular), it is inevitable that some decisions will impose constraints on some users in order to maximize the value for everyone. Indeed, the FTC attempts to disparage “Amazon’s flywheel” as a mechanism for exploiting its dominance.[69] For Amazon, meanwhile, that “flywheel” encompasses the importance of ensuring value on one side of the platform in order to increase its value to the other side:

A critical mass of customers is key to powering what Amazon calls its “flywheel.” By providing sellers access to significant shopper traffic, Amazon is able to attract more sellers onto its platform. Those sellers’ selection and variety of products, in turn, attract additional shoppers.[70]

But at times, maximizing the value of the platform may entail imposing constraints on sellers or buyers. Unfortunately, some of these practices are the precise ones the FTC complains of here. Limiting access to the “Buy Box” by sellers of products that are available for less elsewhere, for example, ensures that consumers pay less and builds Amazon’s reputation for reliability;[71] bundling Prime services may mean some consumers pay for services they don’t use in order to get fast shipping, but it also attracts more Prime customers, enabling Amazon to raise revenue sufficient to guarantee same-, one-, or two-day shipping and providing a larger customer base for the benefit of its sellers.[72]

The bifurcated market approach also conflicts with the Supreme Court’s holding in Ohio v. American Express.[73] In Amex, the Court held that there must be net harm to both sides of a two-sided market (like Amazon) before a violation of the Sherman Act may be found. And even the decision’s critics recognize the need to look at effects on both sides of the market (whether they are treated as a single market, as in Amex, or not).[74]

The complaint itself seems to provide enough fodder to suggest that Amazon’s marketplace should be treated as a two-sided market, which the Supreme Court defined as a “platform [that] offers different products or services to two different groups who both depend on the platform to intermediate them.”[75] The complaint is replete with allegations of a “feedback loop” between the two markets, and it does appear that the consumers depend on the sellers and vice versa.

The economic literature shows that two-sided markets exhibit interconnectedness between their sides. It would thus be improper to consider effects on only one side in isolation. Yet that is what artificially narrow market definitions facilitate—letting plaintiffs make out a prima facie case of harm in one discrete area. This selective focus then gets upended once defendants demonstrate countervailing efficiencies outside that narrow market.

But why define markets so narrowly if weighing interrelated effects is ultimately essential? Doing so seems certain to heighten false-positive risks. Moreover, cabining market definitions and then trying to “take account” of interdependencies is analytically incoherent. It makes little sense to start with an approach prone to missing the forest for the trees, only to try correcting the distorted lens part way into the analysis. If interconnectedness means single-market treatment is appropriate, the market definition should match from the outset.

But I think the FTC is aiming not for the most accurate approach, but for the one that (it believes) simply permits it to ignore procompetitive effects in other markets, despite its repeated acknowledgment of the “feedback loops” between them.[76] Certainly, FTC Chair Lina Khan is well aware of the possible role that Amex could play, and has even stated previously that she believes Amex does apply to Amazon.[77] Instead, the agency is hoping (incorrectly, I believe) that the Court’s decision in Amex won’t apply, and that its decisions in PNB and Topco will ensure that each market be considered separately and without allowance for “out-of-market” effects occurring between them.[78] Such an approach would make it much easier for the FTC to win its case, but would do nothing to ensure an accurate result.

The district court in Amex, in fact, took a similar approach (finding in favor of the plaintiffs), holding that the case involved “two separate yet complementary product markets.”[79] Citing Topco and PNB, the district court asserted that, “[a]s a general matter . . ., a restraint that causes anticompetitive harm in one market may not be justified by greater competition in a different market.”[80] Similarly, Justice Stephen Breyer, also citing Topco, concluded in his Amex dissent that a burden-shifting analysis wouldn’t incorporate consideration of both sides of the market: “A Sherman Act §1 defendant can rarely, if ever, show that a procompetitive benefit in the market for one product offsets an anticompetitive harm in the market for another.”[81]

Some scholars assert that PNB and Topco apply to preclude offsetting, “out-of-market” efficiencies in monopolization cases, but it is by no means clear that the PNB limitation applies in Sherman Act cases. As a matter of precedent, PNB applies only to mergers evaluated under the Clayton Act. And the claim that the Court in Topco has extended the holding in PNB to the Sherman Act rests (at best) on dicta.[82]

It is true that the Court limited Amex to what it called “transaction” markets.[83] But courts are almost certainly going to have to deal with interrelated effects that occur in less-simultaneous markets, and they will almost certainly have to do so either by extending Amex’s single-market approach, or by accepting out-of-market efficiencies in one market as relevant to the antitrust analysis of an ostensibly distinct market on the other side of the platform. The FTC’s Amazon complaint presents precisely this dynamic.

Legal doctrine aside, ignoring benefits in one interconnected market while focusing on harms in another will lead to costly overdeterrence of procompetitive conduct.

Indeed, the FTC’s complaint identifies not just ambiguous conduct (conduct that may constrain one side but benefit the other side and the platform overall), but it points to the very act of providing benefits to consumers as a means of harming competition.[84]

What if Amazon makes it harder for new entrants on the “marketplace” side to enter profitably, because it offers benefits on the consumer side that most competitors can’t match? The FTC would have you believe that is a harm, full stop, because of the seller-side effect. But that would also effectively mean that simply increasing efficiency and lowering prices would amount to harm, because it would also make it harder for new entrants to match Amazon. How can conduct that provides a clear benefit to consumers constitute an antitrust harm?[85]

In essence, the FTC maintains this illogical position by cordoning off the two sides of Amazon’s platforms into separate markets and then asserting that benefits in one cannot justify “harms” in the other, despite recognizing the close interrelatedness between the two markets:

Sellers who buy marketplace services from Amazon provide much of the product selection that helps Amazon attract and keep its shoppers. As more shoppers turn to Amazon for its product selection, more sellers use its platform to gain access to its ever-expanding consumer base, which attracts more shoppers, and so on. . . . The interplay between Amazon’s shoppers and sellers increases barriers to new entry and expansion in both relevant markets and limits existing rivals’ ability to compete. In this way, scale builds on itself, and is cumulative and self-reinforcing.[86]

This is artificial and nonsensical. What Amazon does is maximize the value of the platform to the benefit of all users, on net. That some of those benefits accrue at certain times to only one set of users cannot be taken to undermine the value of Amazon’s overall, long-term platform-improving conduct.

Finally, it is worth noting that, even where nominal market distinctions across platform users have been argued by plaintiffs and upheld by courts, analysis of anticompetitive effects has generally turned to out-of-market effects.

Consider the famous case of Aspen Skiing Co. v. Aspen Highlands Skiing Corp. In that case, analyzing the competitive effect of the defendant’s conduct regarding access by a competitor to an “all Aspen” ski pass required looking at effects in the output market for downhill skiing, as well as the input market for mountain access needed to provide those tickets.[87] Indeed, as the Court noted, “[t]he question whether Ski Co.’s conduct may properly be characterized as exclusionary cannot be answered by simply considering its effect on Highlands. In addition, it is relevant to consider its impact on consumers and whether it has impaired competition in an unnecessarily restrictive way.”[88] If Aspen Skiing were evaluated as the FTC seeks in this case, there would be two distinct markets at issue, and harm could be proven by assessing the effect on the input market alone, regardless of the effect on consumers.

Indeed, especially where vertically related markets are involved (which is, of course, how the two sides of Amazon’s platform are related), courts have recognized that weighing effects on competition requires a cross-market perspective across both upstream and downstream segments.

Conclusion

The FTC’s proposed market definitions in its case against Amazon exhibit several critical flaws that undermine the complaint. The alleged “online superstore” and “online marketplace services” markets are excessively narrow, excluding manifest competitors and alternatives. The FTC improperly groups together distinctly different products and sellers into questionable “cluster markets” without empirical evidence to support treating them as economically integrated. And the complaint arbitrarily cordons the two markets off from each other, despite acknowledging their interconnectedness, likely in a deliberate effort to avoid weighing out-of-market efficiencies and procompetitive effects flowing between them.

Ultimately, the burden lies with the FTC to defend these narrow market definitions as economically sound. But based on the limited information available thus far, the proposed markets appear to be gerrymandered to suit the FTC’s case, rather than reflective of actual competitive realities.

Whether deliberately tactical or not, the problems with the FTC’s market definition invite skepticism regarding the overall merits of the agency’s case. If the relevant markets prove indefensible upon fuller examination of the facts, the theory of harm in the case may well collapse. At a minimum, the FTC faces an uphill battle if its case indeed rests more on artful pleading than rigorous economics.

 

[1] Gregory J. Werden, Why (Ever) Define Markets? An Answer to Professor Kaplow, 78 Antitrust L.J. 729, 741 (2013) (emphasis added).

[2] See, e.g., Josh Sisco, The FTC Puts Your Lunch on Its Plate, Politico (Nov. 21, 2023), https://www.politico.com/news/2023/11/21/feds-probe-10b-deal-for-subway-sandwich-chain-00128268.

[3] Complaint, F.T.C., et al. v. Amazon.com, Inc., Case No. 2:23-cv-01495-JHC (W.D. Wa., Nov. 2, 2023) at ¶¶ 119-208, available at https://www.ftc.gov/legal-library/browse/cases-proceedings/1910129-1910130-amazoncom-inc-amazon-ecommerce (“Amazon Complaint”).

[4] Id. at ¶ 124.

[5] Id.

[6] Id.

[7] Nike Store (last visited Dec. 6, 2023), https://www.nike.com.

[8] Wayfair (last visited Dec. 6, 2023), https://www.wayfair.com.

[9] E-Commerce Retail Sales as a Percent of Total Sales (ECOMPCTSA), FRED Economic Data (last updated Nov. 17, 2023), https://fred.stlouisfed.org/series/ECOMPCTSA.

[10] Amazon Complaint, supra note 3, at ¶ 185.

[11] See, e.g., How Google Shopping Works, Google (last visited Dec. 6, 2023), https://support.google.com/faqs/answer/2987537; Shopify Official Website, Shopify (last visited Dec. 6, 2023), https://www.shopify.com/; Instagram Shopping, Instagram (last visited Dec. 6, 2023), https://business.instagram.com/shopping.

[12] See Geoffrey A. Manne & E. Marcellus Williamson, Hot Docs vs. Cold Economics: The Use and Misuse of Business Documents in Antitrust Enforcement and Adjudication, 47 Ariz. L. Rev. 609 (2005).

[13] For a discussion of this problem in the context of mergers (but with relevance to market definition in Section 2 cases), see Daniel J. Gilman, Brian Albrecht and Geoffrey A. Manne, The Conundrum of Out-of-Market Effects in Merger Enforcement, Truth on the Market (Jan. 16, 2024), https://truthonthemarket.com/2024/01/16/the-conundrum-of-out-of-market-effects-in-merger-enforcement.

[14] See Amazon Complaint, supra note 3, at ¶ 117.

[15] See id. at ¶ 163.

[16] See id. at ¶ 123 (“Online superstores offer shoppers a unique set of features”).

[17] See id. at ¶ 171. (“Other commercially available data, including recently reported statistics from eMarketer Insider Intelligence, a widely cited industry market research firm, confirms Amazon’s sustained dominance across this same set of companies, with an estimated market share of more than 82% of GMV in 2022.”).

[18] See Matthew Johnston, 10 Biggest Retail Companies, Investopedia (last updated May 8, 2023), https://www.investopedia.com/articles/markets/122415/worlds-top-10-retailers-wmt-cost.asp.

[19] Stephanie Chevalier, Market Share of Leading Retail E-Commerce Companies in the United States in 2023, Statista (Nov. 6, 2023), https://www.statista.com/statistics/274255/market-share-of-the-leading-retailers-in-us-e-commerce.

[20] See Matthew Johnston, supra note 18.

[21] See E-Commerce Retail Sales as a Percent of Total Sales, supra note 9.

[22] Werden, supra note 1, at 741.

[23] See Geoffrey A. Manne, Premium Natural and Organic Bulls**t, Truth on the Market (Jun. 6, 2007), https://truthonthemarket.com/2007/06/06/premium-natural-and-organic-bullst (“[E]conomically relevant market definition turns on demand elasticity among consumers who are often free to purchase products from multiple distribution channels, [and] a myopic focus on a single channel of distribution to the exclusion of others is dangerous.”).

[24] Hicks v. PGA Tour, Inc., 897 F.3d 1109, 1120-21 (9th Cir. 2018) (citing Newcal Indus., Inc. v. Ikon Office Sol., 513 F.3d 1038, 1045 (9th Cir. 2008)).

[25] See Amazon Complaint, supra note 3, at ¶¶ 128-33.

[26] Dick’s Sporting Goods (last visited Dec. 6, 2023), https://www.dickssportinggoods.com; REI Co-op Shop (last visited Dec. 6, 2023), https://www.rei.com; Bass Pro Shops (last visited Dec. 6, 2023), https://www.basspro.com/shop.

[27] Jackets, Columbia (last visited Dec. 10, 2023), https://www.columbia.com/c/outdoor-jackets-coats.

[28] See Columbia Coats & Jackets, Macy’s (last visited Dec. 10, 2023), https://www.macys.com/shop/womens-clothing/womens-coats/Brand/Columbia?id=269; Women’s Columbia Coats, Nordstrom (last visited Dec. 10, 2023), https://www.nordstrom.com/browse/women/clothing/coats-jackets?filterByBrand=columbia.

[29] See Amazon Complaint, supra note 3, at ¶¶ 148-59.

[30] See Daniela Coppola, Average Number of Products Bought Per Order Worldwide from January 2022 to December 2022, Statista (Feb. 1, 2023), https://www.statista.com/statistics/1363180/monthly-average-units-per-e-commerce-transaction.

[31] See Khadeeja Safdar, supra note 38.

[32] Google Product Discovery Statistics, Think with Google (last visited Dec. 6, 2023), https://www.thinkwithgoogle.com/marketing-strategies/search/google-product-discovery-statistics (“49% of shoppers surveyed say they use Google to discover or find a new item or product”). Also notable, “51% of shoppers surveyed say they use Google to research a purchase they plan to make online.” Product Research Statistics, Think with Google (last visited Dec. 6, 2023), https://www.thinkwithgoogle.com/marketing-strategies/search/product-research-search-statistics.

[33] See Danny Goodwin, 50% Of Product Searches Start on Amazon, Search Engine Land (May 16, 2023), https://searchengineland.com/50-of-product-searches-start-on-amazon-424451.

[34] Id.

[35] See Shopify (last visited Dec. 6, 2023), https://www.shopify.com; BigCommerce (last visited Dec. 6, 2023), https://www.bigcommerce.com.

[36] Amazon Complaint, supra note 3, at ¶ 198 (“SaaS providers’ services are not reasonably interchangeable with online marketplace services.”).

[37] See Brown Shoe Co., Inc. v. United States, 370 U.S. 294, 325 (1962) (“The outer boundaries of a product market are determined by the reasonable interchangeability of use or the cross-elasticity of demand between the product itself and substitutes for it.”).

[38] See, e.g., Khadeeja Safdar, Nike to Stop Selling Directly to Amazon, Wall Street J. (Nov. 13, 2019), https://www.wsj.com/articles/nike-to-stop-selling-directly-to-amazon-11573615633.

[39] See Tomas Kacevicius (@intred), Twitter (Jun. 19, 2019, 7:05 PM), https://x.com/intred/status/1141527349193842688?s=20 (“[M]ore than 820K merchants are currently using #Shopify, making it the 3rd largest online retailer in the US.”).

[40] Amazon Complaint, supra note 3, at ¶ 199 (emphasis added).

[41] Id. at ¶ 5.

[42] See infra Section III.

[43] Juozas Kaziukenas, Google Shopping Is Again an E-Commerce Aggregator, Marketplace Pulse (Apr. 28, 2020), https://www.marketplacepulse.com/articles/google-shopping-is-again-an-e-commerce-aggregator.

[44] See Mohammad. Y, Instagram Commerce Statistics and Shopping Trends in 2023, OnlineDasher (last updated Sep. 19, 2023), https://www.onlinedasher.com/instagram-shopping-statistics.

[45] Id.

[46] Checkout on Instagram, Instagram for Business (last visited Dec. 7, 2023), https://business.instagram.com/shopping/checkout.

[47] See Shopify Fulfillment Network, Shopify (last visited Dec. 6, 2023), https://www.shopify.com/fulfillment; Outsourced Fulfillment, ShipBob (last visited Dec. 7, 2023), https://www.shipbob.com/product/outsourced-fulfillment.

[48] Promedica Health Sys., Inc. v. Fed. Trade Comm’n, 749 F.3d 559, 565 (6th Cir. 2014).

[49] Id. at 567.

[50] Id. (quoting 2B Areeda, Antitrust Law, ¶ 565c at 408).

[51] See EU Commission, Universal Music Group / EMI Music, Case No. COMP/M.6458, Decision, 21 September 2012, ¶¶ 141-58.

[52] See, e.g., In the Matter of HCA Healthcare/Steward Health Care System, FTC Docket No. 9410 (Jun. 2, 2022), available at https://www.ftc.gov/legal-library/browse/cases-proceedings/2210003-hca-healthcaresteward-health-care-system-matter.

[53] U.S. v. Philadelphia Nat. Bank, 374 U.S. 321, 356 (1963) (“PNB”) (“We agree with the District Court that the cluster of products (various kinds of credit) and services (such as checking accounts and trust administration) denoted by the term ‘commercial banking,’ composes a distinct line of commerce.”).

[54] Universal Music Group / EMI Music, supra note 51, at ¶ 141.

[55] PNB, 374 U.S. at 356 (emphasis added).

[56] United States v. Philadelphia National Bank, 201 F. Supp. 348, 361 (E.D. Pa. 1962).

[57] Id. at 363.

[58] In the Matter of Penn State Hershey Medical Center and Pinnacle Health System, FTC Docket No. 9368 (Dec. 7, 2015), available at https://www.ftc.gov/system/files/documents/cases/151214hersheypinnaclecmpt.pdf.

[59] Consent Brief of Amici Curiae Economics Professors in Support of Plaintiffs/Appellants Urging Reversal, FTC v. Penn State Hershey Medical Center, et al., Case No. 16-2365 (3^rd Cir., Jun. 8, 2016), available at https://www.hbs.edu/ris/Profile%20Files/Amicus%20Brief%20in%20re%20Hershey-Pinnacle%20Proposed%20Merger%206.2016_e38a4380-c58b-4bb4-aecd-26fc7431ecba.

[60] Fed. Trade Comm’n v. Penn State Hershey Med. Ctr., 838 F.3d 327 (3d Cir. 2016).

[61] Complaint, FTC v. Staples Inc. and Office Depot, Inc., Case No. 1:97CV00701 (D.D.C., Apr. 10, 1997), available at https://www.ftc.gov/legal-library/browse/cases-proceedings/9710008-staples-inc-office-depot-inc.

[62] See Orley Ashenfelter, David Ashmore, Jonathan B. Baker, Suzanne Gleason, & Daniel S. Hosken, Empirical Methods in Merger Analysis: Econometric Analysis of Pricing in FTC v. Staples, 13 Int’l J. Econ. of Bus. 265 (2006).

[63] F.T.C. v. Staples, Inc., 970 F. Supp. 1066 (D.D.C. 1997).

[64] And, for at least one court, this is the only basis on which a cluster market is appropriate. See Green Country Food v. Bottling Group, 371 F.3d 1275, 1284 (10th Cir. 2004) (“A cluster market exists only when the ‘cluster’ is itself an object of consumer demand.”) (citing Westman Comm’n Co. v. Hobart Int’l, Inc., 796 F.2d 1216, 1221 (10th Cir. 1986) (rejecting cluster market approach where cluster was not itself the object of consumer demand)).

[65] For example, successful Chinese food product startup Fly By Jing was started by one woman in 2018. She sells only her own products and does so not only on Amazon, but also on her own website and, among countless other places, Costco. See Fly By Jing Amazon Storefront, Amazon.com (last visited Dec. 8, 2023), https://www.amazon.com/stores/page/F2C02352-02C6-4804-81C4-DEA595C644DE; Fly By Jing (last visited Dec. 8, 2023), https://flybyjing.com/shop; Fly By Jing (@flybyjing), Instagram (Feb. 22, 2022), https://www.instagram.com/reel/CaSnvVzlkUW/ (“Sichuan Chili Crisp Now in Costco”).

[66] Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 488 (1977) (quoting Brown Shoe, 370 U.S. at 320).

[67] See Gilman, Albrecht & Manne, supra note 13.

[68] Amazon Complaint, supra note 3, at ¶ 119.

[69] Id. at ¶ 9.

[70] Id. at ¶ 215.

[71] Id. at ¶ 269.

[72] Id. at ¶ 218.

[73] 138 S. Ct. 2274 (2018) (“Amex”).

[74] See, e.g., Michael Katz and Jonathan Sallet, Multisided Platforms and Antitrust Enforcement,127 Yale L.J. 2142 (2018). Katz and Sallet criticize the concept of treating both sides of a two-sided market in one relevant market: “Because users on different sides of a platform have different economic interests, it is inappropriate to view platform competition as being for a single product offered at a single (i.e., net, two-sided) price.” Id. at 2170. But they also contend that effects on both sides must be considered: “[In order] to reach sound conclusions about market power, competition, and consumer welfare, any significant linkages and feedback mechanisms among the different sides must be taken into account.” Id.

[75] Amex, 138 S. Ct. at 2280.

[76] See Amazon Complaint, supra note 3, at ¶¶ 119, 176, 179, 209, 215, & 217.

[77] Lina Khan, The Supreme Court Just Quietly Gutted Antitrust Law, Vox (Jul. 3, 2018), https://www.vox.com/the-big-idea/2018/7/3/17530320/antitrust-american-express-amazon-uber-tech-monopoly-monopsony (“On the surface, the Court’s language [in Amex] suggests that the special rule would apply to Amazon’s marketplace for third-party merchants.”).

[78] PNB, 374 U.S. 321; United States v. Topco Associates, Inc., 405 U.S. 596 (1972) (“Topco”).

[79] United States, et al. v. Am. Express Co., et al., 88 F. Supp. 3d 153, 171 (E.D.N.Y. 2015).

[80] Id., 88 F. Supp. 3d at 247 (citing Topco, 405 U.S. at 610; PNB, 374 U.S. at 370).

[81] Amex, 138 S. Ct. at 2303 (quoting Topco, 405 U.S. at 611).

[82] See Geoffrey A. Manne, In Defence of the Supreme Court’s ‘Single Market’ Definition in Ohio v American Express, 7 J. Antitrust Enf. 104, 115-17 (2019) (“The Court in Topco cited PNB in dictum, not for a doctrinal proposition relating to the operation of the rule of reason, but for a general, conceptual point about the asserted difficulty of courts adjudicating between conflicting economic rights. . . . Nowhere does the Court in Topco suggest that it is inappropriate within a rule-of-reason analysis to weigh out-of-market efficiencies against in-market effects.”).

[83] Ohio v. Am. Express Co., 138 S. Ct. at 2280 (“Thus, credit-card networks are a special type of two-sided platform known as a ‘transaction’ platform. The key feature of transaction platforms is that they cannot make a sale to one side of the platform without simultaneously making a sale to the other.”) (citations omitted).

[84] See, e.g., Amazon Complaint, supra note 3, at ¶ 222 (“Amazon’s restrictive all-or-nothing Prime strategy artificially heightens entry barriers because rivals and potential rivals cannot compete for shoppers . . . solely on the merits of their online superstores or marketplace services. Instead, they must enter multiple unrelated industries to attract Prime subscribers away from Amazon or incur substantially increased costs to convince Prime subscribers to sign up for a second shipping subscription or otherwise pay for shipping a second time. This substantial expense significantly constrains the number of firms who have any meaningful chance to compete against Amazon and raises the costs of any that even try. . . . Amazon’s restrictive strategy artificially heightens barriers to entry, such that an equally or even a more efficient or innovative rival would be unable to fully compete by offering a better online superstore or better online marketplace services.”).

[85] See Brian Albrecht, Is Amazon’s Scale a Harm?, Truth on the Market (Oct. 13, 2023), https://truthonthemarket.com/2023/10/13/is-amazons-scale-a-harm/.

[86] Amazon Complaint, supra note 3, at ¶¶ 214 & 216.

[87] In Aspen Skiing, the “jury found that the relevant product market was ‘[d]ownhill skiing at destination ski resorts,’” Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585, 596 n.20 (1985). The conduct at issue, however, occurred on the input side of the market.

[88] Id. at 605 (emphasis added).