European Competition Policy in Digital: What’s Next?

August 2019

CPI Europe Column edited by Anna Tzanaki (Competition Policy International) & Juan Delgado (Global Economics Group) presents:

European Competition Policy in Digital: What’s Next? By Nicolas Petit (University of Liege & College of Europe)¹

This short note is about the future of European Union (“EU”) competition policy in digital industries. It is somewhat patent that in the global division of labor, the U.S. (and China) have specialized in creating large digital firms, while the EU’s comparative advantage has been in producing competition policies to discipline them. In past five years, the European Commission (“EC”) has issued three decisions against Google, and collected close to 10 billion USD from its shareholders, i.e. Google Search (Shopping), Google Androïd, and Google AdSense. It fined Facebook in 2017 for breach of merger processes. The EC is currently investigating whether Amazon is using sensitive data from retailers to promote its own product offerings. And it is also looking at Spotify’s complaint against Apple’s 30 percent fee, as well as alleged foreclosure from the Apple Watch.

A draft Stigler Center report issued in May called the European Commission the “global enforcer” against digital platforms exclusionary conduct. And in a recent paper economist Carl Shapiro invites us to “look to Brussels for much of the action.”²

So the question seems not so much whether European competition policy will be activist. It will be. The recent adoption of a Statement of Objections in proceedings against Amazon suggests that the EC is not yet ready to loosen the grip on digital firms.³ That said, one may expect a progressive recalibration of the EC’s competition policy mix, with more balance between antitrust enforcement in particular cases and rule-making/guidance initiatives. Moreover, we should not forget the EU courts who will ultimately be called to shed light on the legal tests applicable in the abovementioned cases, as defendants have sought annulment of the EC decisions on substantive, procedural, and remedial grounds.⁴ Readers will recall that  some EC decisions like Google Shopping have been criticized for their alleged failure to articulate a clear theory of liability in support of the case.

With this background, the purpose of the present piece is to address two speculative questions: (1) what will be the focus of EU competition policy in digital ?; and  (2) can we expect this policy to deliver consumer welfare?

1. What will be the Focus of EC Competition Policy in Digital Industries?

We can get a sense of what future European competition policy will look like from a report entitled “Competition Policy for the Digital Era” from May 2019. The report was commissioned to a group of three experts, independent from the EC. Over its 132 pages, there are many ideas. But a handful of them are aligned with the legal and policy preferences previously expressed by EC officials in published papers. In fact, close observers of the Brussels antitrust scene read the report as a JV between the three independent experts and the EC policy units.⁵

The first one concerns abuse of dominance cases. It looks technical. But it could have significant impact on antitrust enforcement and litigation. The idea is that in highly concentrated markets characterized by strong network effects, the burden of proof should shift on the incumbent, who should be asked to demonstrate the pro-competitiveness of its conduct. Otherwise, and absent a defendant showing of procompetitive purpose or effect, conduct ought to be deemed unlawful. This proposition marks a deviation from the rule of reason, evidenced-based approach increasingly taken by the Court of Justice of the EU in this area of the law, and it will thus necessitate a test case in Luxemburg. The practical significance of that procedural modification is the following: It is often complained that startups subject to platform anticompetitive conduct do not have the resources necessary to bring evidence-hungry complaints before courts and agencies, and go bust due to litigation fatigue. If the shift in the burden of proof happens, this could lead to increased enforcement activity levels before national agencies and antitrust courts across Europe.

The second is stricter merger policy against startup acquisitions by dominant platforms. The EC seems interested in controlling deals like Facebook/Instagram, where an incumbent with a dominant ecosystem buys a quickly growing startup that operates outside of its market.⁶ To date, these mergers are looked at as “conglomeral” transactions. In the absence of clear horizontal overlaps or risks of input foreclosure, they are presumptively innocuous. The proposed novel theory of harm is to think of the target firm as a horizontal competitor active within a same “technological space” or “users’ space.” Under the new test, the merger ought to be prohibited even if the platform has the ability and incentives to grow the startup, as long as it can be shown that “the target can grow as a self-standing competitive force if not acquired by the incumbent, or if other companies may be realistically interested.” The counterfactual scenario assumes that there is more social value to stand alone growth or M&A with another competitor than with a dominant firm. Little is said though of the empirical basis for the assumption that merging with a dominant platform is a second best option from a social welfare perspective.

The third one is antitrust spirited data regulation. It is not a secret that the EC is not convinced by the economic idea whereby firms with market power can collect excessive data from users. It is well known that the EC was not particularly enthusiastic about the German 2019 case against Facebook, plausibly on the ground that firms without market power equally exploit users’ behavioral biases to coerce them into excessive terms of use (note that this economic perspective on data irks European privacy officials charged with GDPR enforcement, who believe that competition officials should take a more moral perspective on data minimization).⁷ By contrast, the EC seems to accept the notion that personal and non-personal data possession raise entry barriers for both competitors and complementors, and in turn necessitate data access, sharing, and interoperability remedies. Practicalities like monitoring of continuous data access or pricing disputes however plead against anything other than exceptional antitrust enforcement. The EC report logically proposes data regulation or soft law guidance through business review letters, though it does not demarcate clearly how to think of the division of labor between regulation and antitrust enforcement.

These are, in a nutshell, the priority items on the wish list of EC antitrust officials. Note that ideas often aired in the U.S. policy discussion have zero chance of success in Europe. Chris Hughes’ and others’ dream of a GAFA breakup will not come from Brussels. There are demanding legal tests to meet to dismember firms in European competition law.⁸ Nor do we see much demand for institutional reform or procedural innovation, beyond faster investigations.

2. Will Activist EC Competition Policy in Digital Deliver Consumer Welfare?

Can we assess the economic impact of the EC activist competition policy in digital industries? Short of clear remedial impact in the various closed cases to date, the EC’s motivation and justification have often boiled down to moral high ground antitrust rhetoric, like “fairness” or “trust.” In reality, the answer to whether activist competition policy delivers consumer welfare necessitates to discuss several hard questions.

The first is the following: Can we be confident that activist antitrust policy against platforms is socially beneficial, absent a clear understanding of the relationship between market structure, business models, and consumer surplus in digital markets? On this, a recent paper by Brynjolfsson, Collis & Eggers cuts new ground.⁹ The authors use choice experiments to estimate the consumer surplus created by digital goods with zero price. Concretely, their study asked users how much compensation they would require to forego the digital good. Though the study does not compare monopolized digital goods to competitive ones, it estimates the consumer surplus generated by platforms like Facebook or digital services like search engines. The study finds that the monthly median consumer surplus for a Facebook user is 48 USD. Emerging experimental work relativizes policy arguments that characterize users as unpaid data subjects. Moreover, it suggests that platforms that profit on other sides of the market earn Ricardian rents, not monopoly ones (otherwise, they would charge for the service).

Second hard question: Are Google, Facebook, and other digital platforms active in online advertising markets actually engines of intense non-price competition in the economy (and thus enablers of competitive markets)? In traditional economics, advertisement is modeled as a form of non-price competition. So, if demand (and supply) of online advertisement grows (as it has historically done), should we not consider (i) that this is a sign that there is increased non-price competition between advertisers; and (ii) that by providing the technological infrastructure that makes non-price competition possible, tech giants actually enable competitive markets? This, in turn, calls into turns on its head several sacred cows of European competition policy. For indeed, the argument could imply that tech giants’ rivalry is an instantiation of the “wasteful competition” argument, where excessive competition leads to oversupply of socially harmful activities, like excessive advertisement or personal data extraction.

Third hard question: What can we learn from past antitrust cases? Say we accept the commonly heard claim that big antitrust cases like U.S. v. Microsoft¹⁰ helped create firms like Google and other tech giants. We still fail to understand the causal, let alone, functional relationship between the antitrust case and subsequent social benefits like new firm entry, technological innovation, and dynamic efficiency. Did these firms enter on the expectation that Microsoft would be deterred (by antitrust litigation) to squash them? Or because of a corporate culture change at Microsoft? Or because Microsoft was distracted by protracted antitrust processes (in the U.S., but also most likely by subsequent proceedings elsewhere in the world)? On this, retrospective studies of past antitrust cases would certainly help improve the stock of expert knowledge, and assist the design of antitrust remedies, though we have to be aware that they will not yield unambiguous results.

Fourth hard question: What kind of policy can assist the creation of a European GAFA? Many in Brussels who support active antitrust enforcement in digital markets in their day job dream at night of a European Google or Facebook. Not long ago, a German minister talked of a European AI “Airbus.” Last time I checked, we had not found the recipe. But aren’t we faced with a puzzling natural experiment? For the past 30 years, the U.S. has given birth to GAFA under a weak antitrust enforcement paradigm; and in the past 20 years, China has given birth to BAT under a policy paradigm where centralized cooperation prevails over decentralized competition. Of course, it would be hasty to infer that deviations from mainstream antitrust policies help create superstar firms in tech.¹¹ There are too many confounding factors: labor and capital market institutions, cultural and social norms, demographic and linguistic homogeneity, size of population sets, etc. At the same time, must domestic consumer welfare driven antitrust policy be sensitive or insensitive to the growth of large home tech firms when foreign incumbents enjoy first mover advantages?¹² This issue could lead to original developments in Europe like M&A restrictions or forced data sharing against U.S. and China tech, as we hear increased calls for strong industrial policy and digital sovereignty.

Conclusion

One should expect sustained European antitrust policy in digital markets in upcoming years. And certainly more, beyond antitrust. Many of the previous European Parliament Members (“MEPs”) who worked on tech have just been reelected, and especially the German conservative group who were vocal supporters of the content industry and the press against Google. There is an expectation that they will push the envelope on a number of regulatory initiatives. Some of these reforms may have a much more adverse impact on tech platforms than the one-year-old GDPR. One that should already make the digital industry shiver is a possible attempt to discard the limited liability regime applicable to passive online intermediaries for illegal users’ activities under e-commerce Directive 2000/31/EC.¹³ In her recently announced political platform EC president elect Ursula Von der Leyen announced an upcoming Digital Services Act that will upgrade “liability and safety rules for digital platforms, services and products.” As Carl Shapiro says, “look to Brussels.”

Click here for a PDF version of the article

1 Professor of Law, University of Liege and College of Europe; Visiting Fellow, Stanford University Hoover Institution. This paper complies with ASCOLA Transparency and Disclosure Declaration. I have nothing to disclose.

2 Protecting Competition in the American Economy: Merger Control, Tech Titans, Labor Markets.

3 In fact, the Commission seems to be methodically targeting every letter of the GAFAM acronym.

4 As far as we know, Google has appealed the three EC decisions before the General Court.

5 On the conference circuit, people refer to it as the “Vestager” report.

6 We are neither talking of essentially horizontal acquisitions a la Google/DoubleClick or Microsoft/LinkedIn where the target was already grownup.

7 The EU Commissioner declared to Bloomberg “I don’t think that it can serve as a template” for EU action since the case sits “in the zone between competition law and privacy” and was partly based on German law.”

8 Pursuant to Article 7(1) of Regulation 1/2003, “Structural remedies can only be imposed either where there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burdensome for the undertaking concerned than the structural remedy.”

9 Erik Brynjolfsson, Avinash Collis & Felix Eggers, “Using Massive Online Choice Experiments to Measure Changes in Well-being,” [2019] Proceedings of the National Academy of Sciences, 116 (15) 7250-7255.

10 United States v. Microsoft Corporation, 253 F.3d 34 (D.C. Cir. 2001).

11 See, however, Levitt, T. (1952), The dilemma of antitrust aims: comment. The American Economic Review, 42(5), 893-895 (asking whether we should replace the Sherman Act “by a system of laws that recognize the technological inevitability of monopoly and oligopoly in modern industry,” and perhaps change the liability focus to other issues?).

12 For a version of the argument in the U.S., see J. New, The Hipster Antitrust Movement Could Undermine American AI, https://www.datainnovation.org/2019/05/the-hipster-antitrust-movement-could-undermine-american-ai/.

13 See, in particular, Article 12 to 15, Directive 2000/31/EC of the European Parliament and of the Council of 8 June 2000 on certain legal aspects of information society services, in particular electronic commerce, in the Internal Market (“Directive on electronic commerce”) (July 17, 2000) OJEU L 178 Vol. 43.