The Digital Markets Act: The Path to Overregulation

By Dr. Aurelien Portuese¹

“The best antidote to the disruptive power of innovation is overregulation,” once wrote Tim Wu, now the White House’s adviser on competition policy. Paradoxically, a long-time “anti-monopoly evangelist” Tim Wu has recently admitted that “Europe has been interested in promoting competition within its own economy. We are thinking about the same thing ourselves.” This European inspiration for U.S. lawmakers and the Biden administration has a name: the Digital Markets Act (“DMA”). But this European regulation is on a collision course with innovation, and looms in the hazards of overregulation.

I. Introduction: Overregulation By the DMA ­– An Inevitable Choice

On March 24, 2022, the European Union’s institutions have reached a political agreement on the DMA. Such agreement is both unsurprising and unfortunate. Unsurprising because all institutions of the European Union have agreed, from the first proposal of the DMA by the European Commission in December 2020, on the contours of the new rules about digital competition; institutional discussions have only increased, not decreased, the regulatory burden imposed on the targeted companies. Case in point: The European Parliament approved in December 2021 an earlier version of the DMA with 642 votes in favor, 8 against, and 46 abstentions–the overwhelming support for the DMA unsurprisingly will soon lead to its final adoption. The agreement is also unfortunate since the possible unintended consequences of the DMA have been widely documented over the last two years. These unintended consequences include fewer services from digital ecosystems, greater cybersecurity risks, poorer quality of data privacy because of data combination requirements, and also consumer harm generated out of prohibitions to offer services that may benefit consumers. And yet, European lawmakers have consistently ignored them as frivolous and unfounded critiques, despite the inevitability of the DMA’s unintended consequences.

As the European institutions are now heading toward the final vote of the DMA by the European Parliament and the Council, nothing can stop the DMA from being adopted. The European Commission’s Vice-President, Margrethe Vestager, expects the DMA to be fully applicable in October 2022. The European Commission is already encouraging the targeted companies to comply with this forthcoming regulation.

To paraphrase Tim Wu, the DMA will become the best antidote to the disruptive power of innovation: the DMA is undoubtedly an illustration of overregulation. Two provisions of the DMA, among many others, illustrate its over-regulatory reach, hence generating detrimental consequences on innovations essential for digital technologies that are critically lacking for European technology (“tech”) companies. The DMA lays down the grounds for breaking up and breaking into the platforms, thereby disregarding legitimate innovation arguments and property rights considerations.

II. Overregulation By the DMA: Breakups for Non-compliance

First, there is one provision of the future DMA that commentators overlook, and yet, it is paradoxically the most stringent provision of the regulation: Article 16 of the DMA. Contrary to the European lawmakers’ assertions, the DMA may lead to the breakup of tech companies. Indeed, the French Digital Minister Cedric O used former Federal Communications Commission Chair Tom Wheeler’s slogan when he publicly presented the DMA on March 24, 2022. In reference to regulating tech giants, referred to as “digital gatekeepers,” he said: “don’t break them up, break them open.” Despite this call against breaking up tech platforms, the DMA lays down the condition for such breakups–or “structural remedies” in antitrust parlance–to materialize.

Article 16 states that whenever the European Commission considers that a digital gatekeeper “has engaged in systematic non-compliance,” the Commission may decide to “impose on such gatekeeper any behavioural or structural remedies which are proportionate to the infringement committed and necessary to ensure compliance with this Regulation.” “Systematic non-compliance” in Article 16.2 means “at least three non-compliance decisions” issued “within a period of five years” by the Commission regarding obligations laid down in Articles 5 and 6 of the Regulation. Such “systematic non-compliance” may lead to structural remedies against the gatekeeper “where there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burdensome for the gatekeeper concerned than the structural remedy” (Article 16.2).

Two sections of the DMA demonstrate that the threat of breakups is much more probable than what European lawmakers admit when hoping to reassure their American counterparts. First, violations of the DMA’s obligations can indeed occur unconsciously. Article 5 contains broad obligations hitting at the heart of the gatekeepers’ business models and intellectual property rights: The DMA obliges them to question their diversification strategies and give rivals access to their patented technologies. Article 6 obligations are purportedly “susceptible of being further specified”: In other words, Article 6 obligations are designed to be broad and indeterminate. Consequently, because of the indeterminacy of the DMA’s obligations, violations of the generalized obligations of Articles 5 and 6 may occur more often than gatekeepers can envisage.

Second, the three “non-compliance decisions,” or violations, do not have to occur successively, but they can be concomitant. This means that a gatekeeper can be subject to three simultaneous investigations for three different violations and receive in the course of a few months three decisions of non-compliance. Furthermore, these three violations can take place regarding the same product. For example, today, LinkedIn profiles and Outlook are bundled together so that users can click on email recipients to discover their social media profiles. By combining data from Outlook with data from LinkedIn, Microsoft may breach Article 5(a) of the DMA,² Article 6(a),³ and Article 6(d)⁴ of the DMA: the wrath of the DMA’s obligations and prohibitions may come all at once as the European Commission may issue three simultaneous non-compliance decisions, thereby putting Microsoft into a position of “systematic non-compliance.” Among hundreds of other similar examples, this example demonstrates that a possible breakup is only a few months away after the DMA is finally adopted.

Third, that Article 16.2 does not alone provide sufficient guardrails. Article 16.2 states that “the Commission may only impose structural remedies pursuant to paragraph 1 either where there is no equally effective behavioral remedy or where any equally effective behavioural remedy would be more burdensome for the gatekeeper concerned than the structural remedy.” Consequently, Article 16.2 provides for two situations when structural remedies can be imposed: when there is a lack of available behavioral remedy, or when behavioral remedies are excessively burdensome. Since there are always some sorts of behavioral remedies available, the first theoretical situation is purely rhetorical. The second situation, however, is more interesting as it reveals the lawmakers’ true perception of breakups. That is, behavioral remedies can sometimes be more cumbersome than structural remedies. In other words, breakups can be a more a proportionate solution than companies’ behavioral commitments.

This approach turns antitrust on its head, as structural remedies have historically been considered the solution of last resort, including by European regulators. Behavioral remedies are more proportionate than the highly controversial structural remedies: forcing firms to commit to certain behaviors remains less intrusive than breaking them up. But Article 16.2 reverses this commonsensical logic—breaking up companies becomes less disproportionate than regulating them. The European Commission can easily invoke the alleged proportionality now attached to breakups: Constant micro-managing companies can be said to be more cumbersome for regulators than a one-time company breakup.

For example, should a search engine rank in a “discriminatory” manner its search results, promoting its adjacent services rather than third-party services? Would it not be less “burdensome” to break up the company between its search engine function and its adjacent service, rather than for European Commission’s enforcers to check every update of the algorithm frequently? Indeed, would it not be easier to unbundle Google Shopping from Google search rather than requiring Google search to actively promote rivals of Google Shopping for “fairness” reasons? Breakups may appear more attractive to European enforcers regarding the administrability of the remedies. Moreover, it can be argued that it would be less “cumbersome” for the company to divest its adjacent services than to be subject to weekly or daily regulatory control of its algorithm’s update to ensure a lack of “discrimination.” Article 16.2’s guardrails are either purely rhetorical or wholly ineffective in preventing excessive use of structural remedies. The bottom line is: breakups are looming.

A legal analysis of Article 16 demonstrates that the likelihood of breakups is, contrary to European lawmakers’ vain reassurance, much higher than one can think. Paradoxically, breakups following the adoption of the DMA appear overlooked in the current discussion. This neglect underestimates the risks of overregulation that the DMA threatens concerning the preservation of innovation incentives and the preservation of property rights.

III. Overregulation By the DMA: “Break-ins” for Proprietary Assets

Furthermore, not only does the DMA seek to break up tech platforms, but European lawmakers also want to break into these tech platforms by way of the DMA. Indeed, Articles 5 and 6 of the DMA lay down a wide range of obligations and prohibitions applicable to the targeted digital gatekeepers. These rules intend to create “fairness” in the way gatekeepers interact with trading partners and “contestability” in the markets within which these gatekeepers operate. These obligations allow for break-ins by business users and competitors: They purportedly enable businesses to enter the operating systems of gatekeepers. As economists note about the DMA’s obligations, it is unclear whether the benefits of the ex-ante obligations will outweigh their costs. The break-ins provided in Articles 5 and 6 are considerable regarding the implications for innovation and regulatory overreach as they allow rivals to unfairly extract rents through the DMA’s obligations.

For example, nondiscriminatory rules suggest algorithmic accountability, which means the enforcers would constantly micromanage the gatekeepers’ algorithmic formulas to ascertain the absence of “unfair” or “discriminatory” treatments of business users. Algorithm-driven companies may “self-prefer” their products to offer consumers and business users synergies across various services. For instance, Google may place Google Meet in Gmail to provide consumers with an alternative, easily accessible video-conferencing program. Also, Amazon may promote products using its delivery services as these services will enable the company to optimize delivery times for consumers. Additionally, “self-preferencing” can reveal consumer preferences, as illustrated by the fact that consumers generally prefer private label products (i.e., generic products) because of brand loyalty and lower prices. For these reasons, the European Commission’s own experts concluded that self-preferencing practices are “not abusive per se, but [should be] subject to an effects test.” And yet, against its own experts’ advice but also against general economic knowledge recognizing the desirability of self-preferencing, the DMA prohibits “self-preferencing” per se.

Article 6(d) of the DMA bans self-preferencing by gatekeepers with no justification available. The version of the DMA approved by the European Parliament on December 15, 2021 indeed states that gatekeepers have to “not treat more favourably in ranking or other settings, services and products offered by the gatekeeper itself or by any third party belonging to the same undertaking compared to similar services or products of third-party and apply transparent, fair and non-discriminatory conditions to such third-party services or products.” Such an aggressive stance on the platforms’ business models has one objective: to break into the platform and unbundle complementary products and services that the platform has introduced as a result of innovation.

The inevitable consequence would be for the targeted platforms to refrain from introducing new products and services because of the risks of engaging in “self-preferencing,” irrespective of likely lost or foregone consumer benefits. Targeted platforms may no longer engage in fierce competition through innovation since the European Commission enforcers may eventually assess their ability to innovate and challenge other industries’ incumbents as anticompetitive. One example would be for digital gatekeepers to compete less with financial institutions since their ability to innovate in terms of financial payments (e.g., Apple Pay, Google Pay, Amazon Pay, Facebook Pay, etc.) may be perceived as “self-preferencing” despite lower fees and the greater choice these innovative services offer to consumers and business users. In other words, the DMA breaks into tech platforms with the likely consequence of ossifying markets with entrenched positions–i.e., the opposite result of the DMA’s stated objectives.

Another illustration of the DMA’s intent to break into the platforms and dictate alternative business models is provided with Article 11.1b, voted on by the European Parliament on December 15, 2021. This article states that “[t]he gatekeeper shall not engage in any behaviour discouraging interoperability by using technical protection measures, discriminatory terms of service, subjecting application programming interfaces to copyright or providing misleading information.” In conjunction with article 11.1a, this article means that gatekeepers must not discourage maximal interoperability even in services which are not the “core platform services” identified by the DMA. In other words, gatekeepers, at all times and for all services, have to ensure interoperability of their services with any business users’ services and devices irrespective of considerations related to intellectual property rights (i.e., copyright) or “technical protection measures” (i.e., cybersecurity). To illustrate, this “anti-circumvention” requirement may lead Amazon, for example, to interoperate its cloud services with any domestic or foreign cloud provider, and Amazon cloud services could not impose technical protection measures designed to prevent data leakages, nor could it impose confidential requirements related to the intellectual property rights ascribed to its cloud technology. The unintended consequences for end-users appear blatant—cybersecurity threats and lack of data protection. Additionally, the unintended consequences for the platform’s innovation will eventually arise at the expense of high-quality business users and all end-users (i.e., investment deterrence on the technology because of constant neglect of intellectual property rights may undermine consumer and business users’ experience).

This anti-circumvention provision is further detailed in Recital 32 of the latest version of the DMA adopted on December 15, 2021. This recital states that the DMA’s obligations

“should apply to any behaviour by a gatekeeper, irrespective of its form and irrespective of whether it is of a contractual, commercial, technical or any other nature, insofar as it could, in practice, have an equivalent object or effect to the practices that are prohibited under this Regulation. Such behaviour includes the design used by the gatekeeper, the presentation of end-user choices in a non-neutral manner, or using the structure, function, or manner of operation of a user interface or a part thereof to subvert or impair user autonomy, decision-making, or choice.”

This “neutrality” would lead gatekeepers to offer products and services in a way where they not only disfavor their own products, but also where they actively promote their rivals’ products and services at the expense of the gatekeepers’ incentives to innovate and compete.

The mere likelihood that a practice of the gatekeeper could lead consumers to choose the gatekeeper’s services will be subject to prohibition under this anti-circumvention clause. For instance, any sign-in suggestions (e.g., Apple, Google, or Facebook sign-ins suggestions) may be prohibited as they are “technical” practices that may “impair” consumers’ “choice” in a “non-neutral manner”; even suggestions among a few competing sign-in alternatives may still appear to be non-neutral since other email providers are discriminated against and excluded from the sign-in suggestions. In other words, if Google suggests that consumers sign in with Google accounts, Apple accounts, or Facebook accounts, companies such as Twitter, Alibaba, or many other rivals may complain that these choices are non-neutral since their names do not appear in the suggestions. Given the impossibility of defining “neutrality,” Google may become incentivized to no longer propose its sign-in solutions at the expense of product innovation and possibly consumer preferences.

The DMA intends to break into the way gatekeepers interact with business users and end users, irrespective of intellectual property rights considerations and irrespective of the ways consumers benefit from digital ecosystems so that gatekeepers are subject to “non-discriminatory” requirements that enforcers define, determine, and implement without predictability for these platforms. Such uncertainty as to whether new products and services can be considered as violating the numerous prohibitions and obligations of the DMA will irremediably lead gatekeepers to lower the pace of introducing new products and services in their ecosystems at the expense of the disruptive innovation which has characterized the digital economy. Ideally, European lawmakers want, through the DMA, to treat gatekeepers as public utilities, subjecting them to neutrality requirements— much like the General Court of the European Union has recently considered that Google should be treated as an “essential facility.”

To be sure, European lawmakers have added to possible breakups of digital gatekeepers a range of prohibitions and obligations which will effectively lead enforcers to break into the platform’s ecosystem, business model and, ultimately, property rights. Treating innovative companies as public utilities is not without consequences for either innovation incentives or consumer benefits.

IV. Conclusion: Overregulation By the DMA – Does it Whither Innovation?

Breakups and break-ins are both highly controversial, and yet they are the legal force at the heart of the DMA. The populist “big-is-bad” has its comeback, as breakups aim at shrinking large-scale enterprises. In contrast, break-ins aim at sharing critical assets legitimately created to redistribute power from large companies to (not so) small ones. Both objectives undermine the fundamental protection of intellectual property rights. Thus, they annihilate the incentives for innovation at the potential expense of consumers and only in the name of following a populist stance against large platforms.

The regulatory pattern of overregulation is undeniable: the DMA adds one layer of stringent regulation atop national regulations without removing regulatory barriers to innovation. Indeed, the premise for the DMA was initially that national regulatory barriers burden European tech companies to compete and thrive: The European institutions had to complete the Digital Single Market. In the traditional narrative of the European Commission regarding the completion of the internal market, such completion would only take place by 1) removing national regulatory barriers and 2) adopting a European-wide regulation harmonizing national rules with the benefit of scale. The DMA was supposed to solve the so-called “fragmentation” of the Digital Single Market due to numerous national regulatory obstacles.

Has such fragmentation of the Digital Single Market been solved with the DMA proposal? Not at all. Contrary to the initial and laudable objective of the DMA, which was the completion of the Digital Single Market through the non-fragmentation of digital rules across Member States, the DMA does not even try to address such fragmentation. This insincere objective is highly problematic for two reasons.

First, from a legal perspective, the DMA lacks the legal basis necessary for its adoption. In other words, it does not achieve the purpose which justifies its proposal and adoption. The DMA has Article 114 of the Treaty on the Functioning of the European Union as a legal basis, which enables the EU to adopt “measures for the approximation of the provisions laid down by law, regulation or administrative action in Member States which have as their object the establishment and functioning of the internal market.” That is, the EU can legislate under Article 114 TFEU only for harmonization purposes. The proposed rule must reduce the regulatory fragmentation within the EU.

But the DMA does nothing to mitigate such regulatory fragmentation. Adding an EU-wide rule and letting Member States adopt their own rules paradoxically increases regulatory fragmentation. Such legal inconsistency–i.e., the legal basis’s objective of reduced fragmentation with the realized objective of increased fragmentation–has led prominent legal scholars to conclude that the DMA may be illegal under European treaties.

Second, from an economic perspective, it is well-known that digital companies can thrive and compete only if they reap off the scale economies inherent to the network effects at play in the digital economy: absent scale, tech platforms are bound to fail, struggle, and ultimately be outcompeted by larger-scale enterprises better equipped for the innovation-driven economy we live in. Consequently, European tech companies will move to the United States to scale-up given the size of the American market and the size of the market for venture capital. There is a consensus that only scale can save the European digital economy. Thus, national regulatory barriers need to be removed.

And yet, the DMA does nothing to help European companies compete in the digital economy, contrary to what European lawmakers might wishfully think that the DMA pursues. In other words, the DMA may very well overregulate and harm a few American tech platforms, altogether harming these platforms’ business partners and millions of consumers. But ironically, the DMA does not help European companies to thrive, since the main obstacle for these companies remains unchanged and could worsen with the DMA–namely, the regulatory fragmentation of the Digital Single Market.

These unfortunate outcomes arising from the DMA proposal and its soon adoption lead to one conclusion: overregulation harms innovation. As a case of overregulation, the DMA will irremediably reduce innovation. The disruptive power of innovation— a force of change and progress—is subdued when a regulation like the DMA adopts a precautionary, risk-averse approach to innovative companies and unleashes radical solutions such as breakups and “break-ins.” In a time of greatly needed transatlantic cooperation, the adoption of the DMA will have lasting consequences on the regulatory divergence of tech platforms between the European Union and the United States.

Click here for a PDF version of the article

1 Director, The Schumpeter Project on Competition Policy, Information Technology and Innovation Foundation; Adjunct Professor, Global Antitrust Institute, George Mason University.

2 Article 5(a) states that digital gatekeepers must “refrain from combining and cross-using personal data sourced from these core platform services with personal data from any other services offered by the gatekeeper or with personal data from third-party services, and from signing in end users to other services of the gatekeeper in order to combine personal data, unless the end-user has been presented with the specific choice in an explicit and clear manner, and has provided consent in the sense of Regulation (EU) 2016/679…”

3 Article 6(a) states that digital gatekeepers must “for its own commercial purposes, and the placement of third-party advertising in its own services, refrain from combining personal data for the purpose of delivering targeted or microtargeted advertising, except if a clear, explicit, renewed, informed consent has been given to the gatekeeper in line with the procedure laid down in the Regulation (EU) 2016/679 by an end-user that is not a minor.”

4 Article 6(d) states that digital gatekeepers must “not treat more favourably in ranking or other settings, services and products offered by the gatekeeper itself or by any third party belonging to the same undertaking compared to similar services or products of third party and apply transparent, fair and non-discriminatory conditions to such third party services or products…”