As previously reported in this Bulletin, on 6 February 2019, the German Bundeskartellamt (Federal Cartel Office – “FCO”) found that Facebook abused its dominance by improperly combining user data that it collected from various sources. As a remedy, the FCO imposed far-reaching restrictions on Facebook’s processing of user data, including a duty to obtain “voluntary consent” from consumers prior to any combination of data. Some celebrated the decision as an innovative approach to protect consumers by means of antitrust law beyond its classical focus on consumer welfare. This rendered the decision as a symbol for what has been dubbed “Hipster Antitrust”. The investigation even won the FCO the prestigious GCR award for global “Enforcement action of the year 2018”. Others, including the authors of this article, criticized that the decision might have over-stretched the limits of antitrust law.
With a preliminary ruling on the suspensive effect of Facebook’s appeal on the merits, on 26 August 2019 the Higher Regional Court of Dusseldorf (the “Court”) joined the latter group and squashed the FCO’s decision with harsh and unambiguous words.
This case reflects a current conflict faced by many antitrust authorities. They are subject to immense political pressure to be creative and innovative in using the general antitrust laws to do “something” against perceived shortcomings resulting from digital monopolies. Yet, at the same time they are also (and should be) bound by case-law on the limits of antitrust law established and upheld by courts that are not subject to the same degree of political pressure, and arguably less open to new theories.
However, this conflict and the recent Court ruling against the FCO likely do not mean the end of “Hipster Antirust”. Neither is Facebook “off the hook” for now. More recent investigations at the European level relating to Facebook’s data use as well as the current proposal of the German government for a reform of competition law rather suggest that the solution to the perceived challenges may be a stronger focus on using well-established theories of harm or in outright legislative moves (“Hipster Regulation”) to stretch the boundaries of antitrust law, at least in Europe.
The FCO’s Facebook Decision on Data Exploitation
Users who wish to join and use Facebook must accept its terms and conditions, including its Data Policy. These conditions allow Facebook to collect user data not only when users are signed in to Facebook (on its website or in an app), but also when they navigate through the Internet outside of this particular service. This includes when they use other Facebook-owned services such as WhatsApp and Instagram, and when they visit third-party websites that have embedded Facebook’s tools (for “like” or “share” buttons) or analytical services. It was a thorn in the FCO’s side that many users may not be aware that Facebook transmits and combines such user data from the different data pools. In its prohibition decision, the FCO concluded that Facebook’s collection and merging of user data constitutes an abuse of a dominant position under German antitrust law, and ordered some form of “internal unbundling” of different data pools.
Too Hipster for the Court – Ruling in Favor of Facebook
On 26 August 2019, the Court, composed of a panel of three judges, literally squashed the FCO’s decision. With strong words, the interim decision reveals that the Court has “serious doubts as to the legality of the appealed decision” (¶25). These “serious doubts” caused the Court to preliminarily order that Facebook’s appeal in the main proceedings would have a suspensive effect (Sec. 65 German Competition Act). Based on a summary assessment of the merits of the case, there is a strong likelihood that the Court will overturn the FCO’s decision in the main proceedings.
Leaving the issue of Facebook’s market dominance aside (¶27), the Court found it clear that the alleged conduct does not amount to an abuse of dominance. In particular, the Court rebutted the FCO’s findings that Facebook exploits consumers (¶¶26-86) and that the conduct may also be considered exclusionary conduct aimed at foreclosing actual or potential competitors (¶¶87-96).
No Exploitative Abuse
The Court assessed the potential exploitation of users based on the general clause contained in the prohibition of abuse of dominance (Sec. 19(1) German Competition Act). As a preliminary point, the Court clarified that, in general, “harm to consumer protection” may be a relevant object of Sec. 19 German Competition Act (¶28). Thus, the Court acknowledged that it is not ruled out that interfering with consumer protection standards may harm competition and thus may constitute an abuse, even if no competitors were foreclosed. This is not surprising given that European and German courts have consistently highlighted that the prohibition of abuse of dominance covers not only practices that harm consumers indirectly by foreclosing competition, but also conduct that directly harms consumers, for example excessive pricing.
Despite this case-law, the Court disagreed with the FCO that it had sufficiently demonstrated that Facebook’s conduct had any anti-competitive effects (¶30). The Court did not see any convincing theory of harm. Merely being covered by the overall goals of competition law is not enough. In summary, the Court determined as follows:
- The relevant consumer data is duplicable, and consumers are not economically weakened by any (permanent) abandonment of such data (¶31).
- The FCO did not establish that, as in cases of price abuses, excessive data had been given away. In fact, according to the Court, the FCO erred by not even determining any specific countervalue for such consumer data (¶32), revealing that the “value of data” is more difficult to substantiate than that of a monetary consideration.
- In any case, the Court determined that the FCO did not demonstrate a significant gap between what Facebook required from its users and what the social network would have been able to demand in a competitive environment (¶33).
Consequently, the only remaining question regarding any potential exploitative abuse is whether there is a breach of mandatory requirements imposed by data protection law (¶38). This is the central point of the judgment, where the fundamental disagreement between the competition enforcer and the Court becomes most obvious. The FCO relied on the concepts of “normative causality” and “causality of results” to establish that this breach of data protection laws entails some form of manifestation of market power. However, agreeing with the arguments previously raised in this Bulletin, the Court rejected the FCO’s approach, concluding that the FCO failed to demonstrate any causality between dominance and abuse.
- The FCO had contended that any breach of data protection laws would simultaneously constitute an abuse of dominance in as much as there is a sufficient link to the market dominance of the infringing undertaking. The Court did not accept that view. It did not even bother to examine the admissibility of the conduct under data protection law (even though the FCO went into great detail on this). Instead, based on a teleological interpretation of German antitrust law, the Court found that an exploitative abuse is ruled out in this case, because no strict causality and actual harm to competition can be proven (¶¶41 et seq.). The fact alone that a company with a dominant position breaches any ordinary law is insufficient to constitute an abuse of this dominance (¶42).
- The Court also rejected the FCO’s attempts to establish a “causality of results” or a “normative causality”, labelling them as solely goal-oriented and inconsistent with the spirit of the law (¶¶51-71). According to the Court, the FCO needs to adhere to the strict principle that to find an abuse, a company must have exploited its dominant position to impose specific conduct on other market participants. Otherwise, the prohibition of abuse would become vague, its application flawed, and governmental intervention unjustified (¶61).
No Exclusionary Abuse
As regards any exclusionary type of abuse vis-à-vis Facebook’s competitors in the market for social networks for private users, the Court ruled that the FCO did not remotely demonstrate that the competitive opportunities of rivals had been impaired (¶¶89 et seq.). The mere ability to distort competition does not suffice for an abuse claim (¶91). The FCO’s argument that Facebook’s conduct raises barriers for market entry was rejected (¶93). The Court rightly pointed at direct network effects being the most significant barriers to market entry (= the benefits of the Facebook network for its users increase with the total number of people connected to the network, as the increasing number of users increases the communication opportunities for each individual user). Conversely, the FCO did not establish that there is an additional need for processing and merging the data at issue (¶93). The Court equally rejected, for lack of reasoning, the FCO’s argument that there was a leveraging of market dominance from the market for social networks for private users to other markets such as for online advertising or consumer communication services (¶¶94 et seq.).
Instead, the FCO should have substantiated to which extent Facebook’s conduct allows the company to increase its advertising revenues and to defend its market position against new entries. In the authors’ view, these foreclosure effects should indeed have been the focus of the three-year investigation and 309 pages of the FCO decision. The investigation was not based on any formal complaint by an affected company. Such a complaint may have helped the FCO to frame a more convincing case on the basis of a solid theory of harm.
To be sure, the FCO may still win before the German Federal Court of Justice which now has to decide whether the Court war right in finding that the FCO’s case was meritless. Until then, Facebook does not have to remedy the alleged abuse as the suspensive effect of its appeal has been established.
The FCO’s Reaction to the Court Ruling
Almost immediately after the Court ruling, the FCO announced that it will exhaust all the legal procedures and appeals available to overturn the ruling, including proceedings before the Court of Justice of the European Union. According to Andreas Mundt, the FCO’s president, the “legal issues [of this case] are highly significant for the future state of competition in the digital economy,” and the FCO is “convinced that [it] can act in this area based on the existing antitrust law.” Mundt further justified using the EU General Data Protection Regulation as a benchmark by arguing that otherwise there would not have been a clear legal standard to establish the limits of allowed collection of data.
The Watchdog’s Tightrope Walk, and the Apparent Clash Between Politics, Enforcers, and the Courts
The FCO’s novel approach in the Facebook case had been welcomed by many (in particular data protection authorities), but also harshly criticized by others. The watchdog had answered the question as to whether its tool-box suffices for the digital age in the affirmative. And admittedly, by requiring an “internal unbundling” of Facebook’s data pools, the FCO broke new and innovative ground in antitrust enforcement.
Yet, aside from the abysses of German law on abuse of dominance, one rather general point of critique remains: the enforcer being focused only on results and instrumentalized by politics. Instead of sailing in the safe waters of established case-law (and focusing on cases of genuine concern to competitors), the German antitrust watchdog acted overly courageous in the international “enforcement race” between the different authorities (so the accusation). By encroaching into matters of data protection, it missed the opportunity to render a stronger case based on genuine antitrust law.
To be sure, both the FCO decision and Facebook’s win before the Court somehow mirror the general risks and uncertainties of applying antitrust law in the digital era. The setback should not have a chilling effect on the FCO’s activities for fear of false-positives. On the contrary, an authority that never tries to test the boundaries of antitrust law will never reach its potential.
However, it is obvious that the abuse of dominance prohibition should, first, not be instrumentalized as a gap filler to sort out existing deficiencies in other fields of law, and second, not become an overly vague concept. Arguably, an extensive acceptance of the concept of “normative causality” is flawed and unadvisable for reasons of legal certainty. The convincing and well-functioning concept of abuse of dominance must not be degenerated into a universal legality audit vis-à-vis dominant market players. The focus needs to remain on keeping markets open for competition.
Outlook: Pending Facebook Investigations and Legislative Proposals in Europe
Even if the German Federal Court sides with the Dusseldorf court, this will not have been the last shot by a competition enforcer against digital giants, and Facebook in particular, on the basis of novel theories of harm. A sudden end to “Hipster Antitrust” appears unlikely. A statement of a senior judge at the Court of Justice of the European Union further backed up calls for a broader understanding of antitrust law. In March 2019, the German judge at this court, Thomas von Danwitz contended that “systematic infringements” of data protection or privacy rules could be sanctioned as an “abuse” under Article 102 TFEU. In his view, a “too-narrow conception of competition law that blanks out privacy concerns” should be avoided because of the “similar and complementary objectives” pursued by both fields of law.
EU Facebook Marketplace Investigation
The European Commission is currently scrutinizing Facebook’s rollout of the Facebook Marketplace and its use of data to strengthen its online classified ads business. Based on a formal antitrust complaint for an alleged abuse of dominance submitted in 2018, the EU’s competition watchdog already has issued two rounds of formal requests for information to market participants. The first questionnaire of June 2019 inquired about the economic particularities of providing online classifieds ads and Facebook’s activities on this market. The second questionnaire of September 2019, in turn, went into greater details of Facebook’s alleged tie of social media advertising customers to its classified ads platform. The recent move of Facebook to open Marketplace Search Results to ads may further escalate the competition concerns.
Amongst others, the Commission investigates how Facebook uses and shares data and its access to data through interfaces such as login tools and social plug-ins. Unlike the German probe, however, this investigation (rightly) much more focuses on the competitive advantages that data can confer on Facebook vis-à-vis its online advertising rivals. Similar to the Commission’s investigations Google Search (Shopping), Google for Jobs, Amazon Marketplace, and with regard to Apple’s App Store the case ultimately deals with the incentives and capability of a dominant platform to foreclose competition on adjacent but connected platforms markets.
French Probe into Facebook’s Advertising Practices
Triggered by a formal complaint of AdTech firm Criteo in September 2019, the French Autorité de la concurrence is currently investigating whether Facebook undermines competition in online advertising by “favoring its own services”. According to a company statement, Criteo “believes that the gradual exclusion of companies from the Facebook platform has adversely impacted the diversity of the online advertising industry”.
German Legislative Proposal Removing any Requirement of Strict Causality Between Dominance and Exploitation
Last but not least, the Facebook case discussed in this article also made it into the legislative proposal for the 10th reform of the German Competition Act, which was published in October 2019. Referring to the Dusseldorf Court decision discussed above, the draft bill contains a provision to the effect that there should no longer be any need for a strict causality between abuse and dominance in any type of exploitative conducts. According to the reasoning of the draft bill (and backing up the FCO), an abusive “exploitation […] may, therefore, not only occur where this results in a monetary damage, but also, for example, in the unjustified disclosure of personal data which is subject to special legal protection”.
For the reasons set out above, such an amendment may cause more fundamental problems than it solves. However, the proposal illustrates that ultimately the tensions between policy driven enforcers and case-law-bound courts may (only) be solved by the legislator. Where there is a political will, there is also a way. Thus, “Hipster Antitrust” may soon develop into “Hipster Regulation.”