This Policy Paper is part of the Digital Platforms & Democracy Project’s efforts to explain and disseminate ideas about regulation of major technology and digital platform companies. Click here to read more of their research and commentary.
The views expressed in Shorenstein Center Policy Papers are those of the author(s) and do not necessarily reflect those of Harvard Kennedy School or of Harvard University.
Policy Papers have not undergone formal review and approval. Such papers are included in this series to elicit feedback and to encourage debate on important issues and challenges in media, politics and public policy. Copyright belongs to the author(s). Papers may be downloaded and shared for personal use under the Shorenstein Center’s Open Access Policy. Please contact the Shorenstein Center with any republication requests.
It is now clear that antitrust enforcers are in full gear, from Europe to the United States and beyond, identifying actual or possible violations of law by companies like Google, Facebook, Amazon and Apple. As Europe leads efforts to implement remedies for violations, and numerous nations proceed with their investigations, we will increasingly have a better sense of how much antitrust will tame the growing dominance of a few enormous tech platforms over the global digital economy. Based on growing signs that these platforms are tipping toward monopoly in key market functions, it is very likely that antitrust is not enough of a solution without targeted regulation that opens markets to new competition.
Perhaps the most important change we need is competition-expanding regulations that address the problems antitrust cannot solve. A new expert regulator equipped by Congress with the tools to promote entry and expansion in these markets could actually expand competition to benefit consumers, entrepreneurship, and innovation. The regulator’s goal of actively promoting competition, not simply maintaining existing competition, is a significant distinction. Given the economic constraints described above, digital platforms require an additional jolt from a regulator to promote new competition. There is not enough competition here for us to merely “maintain” it. Congress must equip the regulator with pro-competition tools such as interoperability, non-discrimination, and merger review.
The regulator’s primary goal should be promoting competition. Competition and consumer choice will help to address many of the problems we face with digital platforms today. Though social welfare regulations are also required, in particular with regard to privacy, the power of these companies in the marketplace is the primary concern, and addressing it will have myriad benefits for consumers. The regulatory authority could be housed within an existing agency, such as the Federal Trade Commission (FTC) or Federal Communications Commission (FCC), or preferably should be a new expert body, focused on digital markets.
The explosive growth of Google in search and its suite of popular services, Facebook in social networking, and Amazon in online retailing illustrate how many digital markets have winner-take-all characteristics that tend to leave just one dominant player. Each of these companies have made extensive up-front investments to build platforms. Today, these platforms are characterized by network effects: consumers prefer to be where everyone else is congregating, and sellers prefer to sell where everyone else is congregating. They are characterized by strong economies of scale and economies of scope, which is due to their low marginal costs, and increasing profits based, inter alia, on control of data.
This combination means that these digital markets have large barriers to entry. The leaders in search, social networking, and other platforms have a large cost advantage from their scale of operations and a large leg-up from the scale of their data. Entrants cannot generally overcome these without either a similar installed base (network effects) or a similar scale (scale economies), both of which are difficult to obtain quickly and cost-effectively.
Additional barriers to entry are generated by consumers’ tendency to seek simple, one-stop-shopping on digital platforms. Consumers tend not to scroll down to see more search results, they agree to settings chosen by the service provider, they stick with one platform (“single-home”), and they generally take actions that favor the status quo and make it difficult for an entrant to attract consumers.
In addition, the role of data in the digital sector fuels the advantages of companies like Google, Facebook and Amazon. These companies’ ability to collect massive amounts of personal data of all types allows for targeted advertising to consumers. It appears that profits increase significantly as platforms gather more and more data about more and more people. This then creates even more advantages for these incumbent dominant service providers. These advantages are easily preserved by platforms that require consumers to agree to terms of service that are confusing and constantly changing, which prevents consumers from understanding how their personal data are being monetized.
In such environments, it is often more difficult to identify and quantify the types of harms that consumers and other market participants are likely to experience as a result of monopolies or near-monopolies. For example, digital platforms are characterized by offering “free” services, where the seller has chosen to set a monetary price of zero and may set other, non-monetary conditions or duties. It is likely that many digital markets have an equilibrium price that is negative — in other words, because of the value of the target advertising, the consumer’s data is so valuable that the platform would pay for it. However, this unusual negative price is very difficult to observe and value with precision, leaving consumers “overcharged.” This can make antitrust enforcement difficult. The non-monetary price of the service, be it for data use or advertising interruptions, may fluctuate as a result of anticompetitive conduct or an anticompetitive merger and be more complex for antitrust enforcers to measure than a monetary change in price would have been.
Similarly, dominant platforms tend to harm quality of service in other ways that are equally difficult to discern and yet involve use of detailed personal data to manipulate users’ preferences. For example, digital businesses that use high-dimensional, large datasets to explore every nook and cranny of consumers many behavioral traits and biases in real time are able to frame, nudge and provide default settings that direct a consumer to the choice that is most profitable for the platform. Such platforms can analyse a user’s data in realtime to determine when the user is in an emotional “hot state” and then offer targeted sales. By exploiting the user, these tactics reduce the quality of the zero-price content the user experiences on the platform in very subtle ways, making it difficult to quantify a precise level of harm.
A common way for companies that obtain a dominant share of service on a platform to protect their dominance and/or to increase profits is to make all necessary complements to platform services themselves or position themselves as necessary bottlenecks between partners and customers. By attempting to maintain complete control over the user relationship, dominant platforms can limit the possibility for independent complementary services to gain meaningful traction and challenge the platform’s power. Similarly, dominant platforms often use exclusive contracts, bundling or technical incompatibilities to restrict entry of competitors. When such practices succeed, investors become wary of putting money behind an independent start-up that would directly or indirectly seek to challenge a dominant platform. VC’s will tend to put money behind companies that seek to be acquired by a dominant platform at an early stage, reducing opportunities for disruptive investment and innovation.
The costs of this situation for innovation, though hard to quantify exactly, are clearly significant. It is time for our government to assess and implement tools that could promote competition in our digital marketplace. Much has been written, rightly, about the importance of increasing enforcement by our antitrust agencies, the Federal Trade Commission (FTC) and the Department of Justice (DOJ). These agencies should not be afraid to take on difficult cases. In a recent report I co-authored, the Stigler Report, we argue for increased enforcement, while also demonstrating why and how to improve our antitrust laws, the Sherman Act and the Clayton Act, to address the narrowing they have suffered from decades of judicial pushback and inconsistent enforcement.11Final Report of the Stigler Committee on Digital Platforms, 2019, available at https://research.chicagobooth.edu/-/media/research/stigler/pdfs/market-structure-report.pdf. Let’s examine the three key tools with which Congress must equip a regulator in order to promote competition in digital platforms: interoperability, non-discrimination, and merger review.
First, the agency should be authorized to require dominant platforms to be interoperable with other services, so competitors can offer their customers access to the dominant network. For example, if Facebook, with its dominant position in social networking and ownership of Instagram and WhatsApp, were required to allow Snapchat users and similar alternative platforms to communicate with their Facebook friends easily using these other services, Facebook’s network effect advantages would be reduced and competition could more easily expand.
In many of the services that we use the most, we’ve come to take interoperability for granted, and often don’t even realize how essential it is. For telephone networks, the 1996 Telecom Act included interoperability requirements, which it referred to as “interconnection,” between competing carriers.22Fundamental errors in the FCC’s implementation of the Act, together with unchecked anticompetitive misbehavior by incumbent firms, failed to protect and promote competition, instead allowing many small upstarts to go out of business. The vigorous competition Congress had envisioned during passage of the 1996 Telecom Act largely failed to materialize. We must learn from these mistakes and set up clear rules with a quick dispute resolution process. The Act built upon Federal Communications Commission (FCC) regulation, setting forth a regulatory regime of duties to connect, of parity in quality between connections offered to the incumbent’s own affiliates and competitors. This duty to deal created the possibility of more competition, and allows for there to be just one national (indeed, global) telephone network, made up of thousands of independent carriers, including multiple competitors in many geographic markets. Allowing interconnection to the dominant network was also a crucial component of the breakup of AT&T. Many early internet services, as well, such as email and the world wide web, were designed in such a way as to permit many different entities to interoperate. Today, online platforms that benefit from network effects and control an important market bottleneck are also appropriate targets for an interoperability rule.
Of course, a rule requiring interoperability depends on strong privacy protections either as part of the rule or guaranteed by another statute such as comprehensive privacy legislation. However, it’s also important that privacy improvement efforts don’t inadvertently make interoperability harder or impossible, for example by banning any transfer of data from one company to another. The data protection and data empowerment tools that must be joined with interoperability should be the responsibility of the same regulator or carefully coordinated across two agencies.
Creating open interoperability regimes for the digital economy is a complex task that should be undertaken by an expert regulator, not generalist law enforcers. A regulator is especially useful for a tool like this because it will require technical detail, frequent updates, and speedy dispute resolution to make sure the interoperability requirement actually promotes competition effectively. Antitrust enforcers, focused on competition, are not well positioned to effectuate user intent and protect users’ personal data.
There’s also the question of how far interoperability should extend. In the telephone networks example, the essential nature of the network and the Telecom Act’s mandate to promote affordable access to all Americans drove the need for network equipment and infrastructure to work seamlessly. In order to simply promote competition and consumer choice in the diverse digital platform market, that exacting level of interoperability may not be necessary. For example, the old “Find My Friends” feature that Facebook used to offer when a Facebook user joined another social network provides a certain level of interoperability, making it easier for users to link up with friends who are already using the new network. But the users of the new network would still be isolated from friends who did not switch. A further level of interoperability would allow communication from Facebook onto the new network. However, sometimes competing social networks will offer vastly different features, and this diversity is great for competition. A key component of Twitter, for example is short messages. It may dampen innovation if every new feature must be perfectly mirrored across networks. Policymakers should take on this issue, and decide, with the help of experts and relevant stakeholders, what level of interoperability should be required. This would include which types of data should be part of an interoperability mandate, and under which conditions.
Competing against an incumbent digital platform is hard. But it can happen in two ways: head-to-head platform entry and/or expansion from one vertical to many. Recently we have experienced very little market entry into areas where one platform has gained enormous market share, like Google in search and Facebook in social networking. Therefore, it is important to assess other ways in which competition may grow.
Online platforms know that companies that use their platform can “disintermediate” them by connecting directly with the consumer, effectively cutting out the platform middleman. It was Microsoft’s fear of disintermediation that led them to compete unfairly against Netscape. They knew their operating system would become less important if people did most of their computing inside one of their apps, in this case the browser. Also, online platforms know that a company that competes with them in one vertical can expand to compete in other verticals, becoming stronger as they take advantage of synergies from the multiple verticals. This means that for platforms, the companies that use the platform are also potential competitors. As a result of this competitive dynamic, some platforms have the incentive and ability to discriminate in ways that may harm competition. The platform has a variety of mechanisms it can use to disadvantage companies that pose a competitive threat, including its access to transaction data, its prioritization of search results, its allocation of space on the page, etc. In the most extreme versions of this behavior, antitrust can prevent abuse, but it is less useful to prevent many subtle discriminatory practices.
Congress should authorize the new regulator to monitor and ban discrimination by digital platforms with bottleneck power in favor of their own services and against their competitors who rely on their platform to reach customers. This is another tool that particularly requires speedy adjudication and an expert regulator . There is a difficult line-drawing problem in identifying which aspects of business are features of a platform, and which are products competing on the platform. For example, an app store may be an essential part of a smartphone operating system, so preferencing the operating system’s own app store by having it pre-loaded on the phone may not be appropriately understood as “discrimination.” On the other hand, a grocery store is probably not an inseparable part of an e-commerce platform, so preferencing Whole Foods over a competing grocery retailer on the Amazon Marketplace might be a good example of discrimination that would be subject to the rule. The slow pace and complexity of antitrust litigation does not lend itself to fast-paced digital markets where discrimination can quickly make or break a competitive outcome.
Similarly, the agency should be authorized to ban certain “take it or leave it” contract terms. For example, a dominant digital platform may require that in order to be on the platform, companies doing business with them must turn over customer data for the dominant platform to use however it pleases. This effectively bundles the service the companies need with data sharing that could undermine their competitive market position. By prohibiting these practices, we can give potential competitors a fighting chance.
Another major concern with digital platforms is their acquisition of potential competitors. Acquisitions of potential or nascent competitors are often small, even falling below the value threshold for pre-merger notification of the competition authorities under the Hart Scott Rodino (HSR) Act. It is very difficult to effectively assess how likely such companies in adjacent markets are to truly be potential competitors to the acquiring digital platform. The small size or lack of pre-existing direct competition of these types of mergers can make it much harder for antitrust enforcement agencies to block them, even if there are indications the merger may be anticompetitive. Markets move quickly and a competitor’s window of opportunity to gain traction against the incumbent is short, making mergers an even more effective tactic at preventing competition, and making effective merger enforcement even more important.
Thus, the regulator should also have the power to review and block mergers, concurrently with the existing antitrust agencies. For particularly important industries, like communications, energy, and national security, we have an additional merger review structure in addition to antitrust. The FCC, FERC, and CFIUS processes run concurrently with the DOJ or FTC merger review process. Similarly, digital platforms that have become essential in our economy and society and face inadequate competition require merger review under a new and different standard, in addition to traditional antitrust review.
The new regulator would have a different standard than the antitrust agencies. This different standard should place a higher burden on dominant platforms to demonstrate overall benefits to society that antitrust enforcers do not have the tools to thoroughly measure. It should assess mergers involving platforms with bottleneck power, and it should only allow those mergers that actually expand competition. Also, there should be no size limit for mergers to warrant pre-merger review by the agency. Any acquisition by a platform with bottleneck power should be reviewed for its competitive impact. This would prevent increased concentration of power when the company being purchased is too small or the competitive consequences are too uncertain. Mergers that provide no clear competitive benefit would be blocked. The standard also must take account of the particular ways that competition happens in digital platforms. For example, non-horizontal mergers may be particularly harmful here due to the economies of scope in data-driven platforms, as well as the importance of interoperability between complementary products.
The new regulator should also be responsible for consumer protection regulations relating to digital platforms, such as privacy protections for users. These rules may also have pro-competitive benefits. For example, if the incredibly detailed data dossiers that the large platforms collect on their users are significantly curtailed by data protection legislation that limits collection and use of personal data, it may be easier for smaller or new companies which don’t have access to data to compete. But these rules are also crucially important to protect users’ rights and people’s freedom from the type of control that detailed data collection can provide to companies.
To which types of companies should these regulations apply? Some of the regulations, such as limits on data collection and use, are not related to levels of competition and therefore must apply broadly to be effective. Some others, like the requirement of non-discrimination, need only apply to especially powerful digital platforms that have the incentive and opportunity to disadvantage competitors. Identifying which platforms are powerful enough to be subject to those rules will require some additional work by the agency. Using the definitions of market power from the jurisprudence of antitrust is likely not sufficient. Instead, the regulator would need to make a determination of which companies hold important “bottlenecks” in the marketplace. This might be because they hold the buying power of many, many customers, so that anyone who wants to sell must be on their platform to reach those customers. Or it might be because they have a monopoly on a key product that’s complementary to many others, creating lock-in for a suite of related products as well. The Stigler Report describes “Bottleneck power” as a situation where buyers or sellers primarily single-home (rely upon a single service provider), which makes obtaining access to those buyers or sellers prohibitively costly without the bottleneck. It might be more efficient for the regulator to assess companies for their bottleneck status at regular intervals. Or the regulator may prefer to assess them at the point of investigation, but have that assessment remain “sticky” for a period of time. Either way, it will be important for companies to know in advance if they are deemed to be bottlenecks, so that they will know which rules apply to them. This will improve predictability for the companies, and more importantly it will improve compliance with the law, since companies will know they risk an enforcement action by engaging in certain proscribed behavior.
In addition, it is important for this new regulatory authority not to duplicate or interfere with any other agency Congress has already authorized to review mergers or otherwise establish policy goals governing companies that become digital platforms. The purpose of a new agency is simply to fill in gaps and supplement antitrust enforcement. 32 It is important not to create a no-man’s land where neither regulation nor antitrust are applicable. After the Verizon v. Trinko, 540 U.S. 398 (2004), case found the antitrust savings clause in the 1996 Telecom Act insufficient to protect antitrust enforcement in that area, we must be especially careful on this point. Legislation to regulate dominant digital platforms should clearly state that it is not intended to supersede antitrust. Antitrust enforcers and other agencies can share dual authority with different review standards and goals, accounting for each other’s work, minimizing inconsistencies, and not necessarily taking priority over one other. Antitrust should remain in full force, with new digital bottleneck legislation filling in the gaps.
In the Stigler Report we point out that it will also be important to improve our general purpose antitrust laws to make sure that powerful companies are following the basic rules of fair competition. For many years, antitrust enforcement has been too limited. This may be one of the reasons that digital platforms have been allowed to become so powerful. Antitrust law must recalibrate the balance it strikes between the risks of over-enforcement and under-enforcement. Certain types of business conduct that were previously thought to be benign are now understood to be anticompetitive. Knowing this, Congress should revise aspects of antitrust doctrine that were adopted explicitly in order to minimize the risk of over-enforcement, and change presumptions to offer less demanding proof requirements on antitrust enforcers. Congress should make it easier for antitrust enforcement to take place against unilateral refusals to deal, predatory pricing, and anticompetitive product design, as well as cases where anticompetitive conduct targets one side of a two-sided market. Antitrust law should relax the proof requirements imposed upon plaintiffs in some circumstances, and in some circumstances it should reverse burdens of proof. Lastly, antitrust needs to work together with regulation. The presence of regulation should not be used to create a penumbra where neither regulation nor antitrust are applied to harmful behavior.
To be clear, even with these changes, antitrust will not be able to resolve the problem of lack of competition in digital platforms. Antitrust law targets specific violations of the law, not broadly applicable market conditions like the ones described here. Antitrust law exists to stop anticompetitive behavior and anticompetitive mergers, but it does not have a mandate to actually increase competition or challenge monopolies that arise through hard-nose competition. Stopping antitrust violations will not change the underlying tendency for digital markets to tip toward few or one dominant player. To jumpstart more competition and give it a fighting chance of surviving, we need a regulator working closely with antitrust enforcers and applying tools that go beyond antitrust remedies.