On November 10, 2021, the European General Court (Court) issued its judgment in Case T-612/17 Google and Alphabet v Commission (Google Shopping).
The Court dismissed almost in its entirety the action brought by Google and Alphabet against the decision by the European Commission (Commission) of June 27, 2017, which found that Google had abused its dominant market position by favoring its own comparison shopping service (CSS) on its general results pages while demoting the results from competing CSSs. The Court also upheld the fine of €2.42 billion imposed on Google by the Commission. The judgment can be appealed to the Court of Justice of the European Union (CJEU).
Perhaps the most important aspect of this judgment is that the Court has now officially recognized that self-preferencing practices can constitute an abuse of dominance and has clarified that the legality of such practices does not have to be assessed in light of the traditional test relating to essential facilities or abusive refusal to provide access. Instead, self-preferencing can constitute an abuse on its own terms, arguably broadening the scope of Article 102 Treaty on the Functioning of the European Union (TFEU) and making it easier for the Commission to reach similar findings in the future.
Some of the key takeaways from this important and complex judgment are outlined below.
1. Self-preferencing practices recognized as an abuse of dominance
The Court determined that self-preferencing practices (described by the Court as “active favouring” and “active exclusionary practices”) could constitute an abuse of dominance, in practice confirming a new category of behavior that constitutes an abuse within the meaning of Article 102 TFEU.
The Court acknowledged some divergence in approach among member states in relation to self-preferencing practices, including that certain national courts have not recognized self-preferencing as an abuse of dominance. However, the Court made it clear that this does not preclude the Commission from reaching a different conclusion. It is for the national courts and authorities to follow the jurisprudence of the EU courts, not vice versa.
The Court determined that a finding of abusive self-preferencing requires a robust case-by-case effects analysis, taking into account “the particular circumstances of the practices in question.”
In other words, it is clear from the judgment that the context within which Google’s self-preferencing occurs was critical to the finding of abuse. It is in light of such specific circumstances that the Court referred to the principle of network neutrality (i.e., a general obligation of equal treatment on internet access providers), concluding that the principle “cannot be disregarded when analyzing the practices of an operator like Google,” which has the undisputed ultradominant position in a downstream market and resulting special responsibility not to allow its behavior to impair competition.
2. Self-preferencing practices do not have to be assessed in light of strict indispensability criteria
While the Court noted that Google’s general search service is akin to an essential facility, it concluded that Google’s conduct does not have to be assessed in light of traditional essential facilities and refusal to provide access case law (e.g., Bronner), which turns on whether access to a dominant provider’s products or services is “indispensable” to a competitor’s offering, such that a refusal to provide access will effectively eliminate competition. This is because in the Court’s view, Google’s conduct does not simply amount to a refusal to provide access. Instead, it involves “an unjustified difference in treatment” in the form of Google favoring its own CSS. Such self-preferencing conduct does not necessarily need to meet the conditions of indispensability and elimination of all competition to be found abusive. As such, the Court drew a distinction between two categories of cases:
a. “passive” (and explicit) refusal of access — for instance, where a dominant firm refuses access to an infrastructure that it deploys for its own needs, but it does not proactively engage in other strategies to disadvantage its rivals
b. “active” exclusionary practices — for instance, favoring of own products/services to the detriment of rivals (i.e., self-preferencing practices) where the abusive conduct does not “lie principally in the refusal as such” but instead involves active steps to discriminate against rivals
In relation to the first category of cases (i.e., passive refusals of access), the Court upheld strict Bronner indispensability and elimination of all competition criteria, recognizing that forcing a firm to conclude a contract with a third party, such as a competitor, can interfere with that firm’s freedom of contract and right to property and should therefore be confined only to exceptional circumstances.
In relation to the second category of cases (i.e., self-preferencing practices), the Court generally found that these cases could merit a more relaxed legal test that omits the indispensability criterion. While the Court did not set out a clear effects analysis in relation to such cases, it seems that such practices would be prohibited only when certain criteria are met. In particular:
- It is evident that only those with material market power should be subject to scrutiny in relation to potentially abusive self-preferencing practices. Indeed, the Court repeatedly refers to Google’s “superdominant” and “ultra-dominant” position on the market and “(almost worldwide) quasi-monopoly” which, in the Court’s view, has placed Google “under a stronger obligation not to allow its behaviour to impair genuine, undistorted competition.”
- Similarly, the Court repeatedly emphasized the importance of Google’s infrastructure (i.e., general search) for third parties to effectively compete in the market. Indeed, the judgment explicitly states that Google’s general search traffic is “akin to […] an essential facility”: It accounts for “a large proportion of the activity of [rival] comparison shopping services” and “no [other] viable alternative” exists.
The Court also emphasized a “certain form of abnormality” that was “not necessarily rational” in Google favoring its own specialized results given “the rationale and value of a general search engine lie in its capacity to be open to results from external (third party) sources and to display these multiple and diverse sources on its general results pages.”
- The Court made it clear that self-preferencing practices can be prohibited only when they have, or are likely to have, anticompetitive effects. The Court further emphasized that “favouring and its effects [must be] properly established” for the self-preferencing conduct to be found abusive. Indeed, the Court refers to “material effects” and “material consequences” when describing Google’s practices, and the Commission did carry out an in-depth analysis of the actual anticompetitive effects of Google’s conduct, including in its decision a number of detailed charts illustrating a correlation between Google’s conduct and rapid drops in visibility and traffic received by its competitors. The Court’s comments clearly indicate that self-preferencing practices should be subject to a robust effects analysis. This would be in line with the comments made by judges at the hearing last year, who explicitly mentioned applying some kind of “relaxed” (but still sufficiently restrictive) Bronner-like test to these types of self-preferencing practices. It is significant that the Commission went beyond an examination of whether Google’s conduct was merely capable of having anticompetitive effects. Instead, it considered in detail the features of the relevant market and the actual impact of Google’s conduct therein.
We hope that if the case is appealed, the CJEU will provide more clarity as to which factors should be taken into account and how the analysis of effects should be conducted in practice.
3. High bar for dominant firms to demonstrate efficiencies
The Court reiterated that with respect to objective justifications for anticompetitive behavior, the burden of proof rests on the firm relying on such justifications to “convincingly” put them forward, particularly where the firm alone is aware of an objective justification or is naturally better placed than the Commission to disclose its existence and demonstrate its relevance. In this case, the Court found that Google failed to show that its conduct was objectively necessary or that it genuinely improved Google’s service to the benefit of users.
As regards efficiency gains, a firm should show that (a) any efficiency gains counteract any likely negative effects on competition and consumer welfare; (b) those gains have been, or are likely to be, brought about as a result of that conduct; (c) such conduct is necessary for the achievement of those gains in efficiency; and (d) such conduct does not eliminate effective competition by removing all or most existing sources of actual or potential competition. In consequence, a firm has to do more than put forward vague, general, and theoretical arguments or rely exclusively on its own commercial interests. In practice, such efficiencies may be difficult for firms to prove.
4. Market definition
The Court confirmed the Commission’s assessment that merchant platforms are not part of the market for CSSs and therefore endorsed the Commission’s view that there is little competitive pressure on Google from merchant platforms.
The Court reasoned that (i) CSSs do not sell products themselves but inform consumers about offers from online sellers, whereas consumers can purchase products directly from merchant platforms; and (ii) merchant platforms offer only a limited selection of goods, namely those of the platform itself and those of the merchants contracting with the platform. CSSs, on the other hand, offer products from the entire market. The Court recognized that sellers can and do offer their products on both CSS and merchant platforms. However, they do so on a complementary basis and not because they view each platform as interchangeable. The judgment illustrates that where platforms have similar uses, they can be argued to be complementary uses in order to separate services into different markets.
As regards the market for general search services, the Court considered that the Commission did not establish that Google’s conduct had had any (even potential) anticompetitive effects and therefore annulled the finding of an infringement in respect of that market alone.
The Court made its own assessment of the facts and concluded (a) first, that the annulment of the part of the Commission’s decision regarding the market for general search services has no effect on the amount of the fine, as the Commission did not take the value of sales on that market into consideration in order to determine the basic amount of the fine; and (b) second, that the infringement was of a particularly serious nature and was adopted intentionally, not negligently. As such, the Court concluded that the amount of the fine imposed on Google must remain the same.
A key point of contention was the alleged intentional nature of Google’s infringement. Google argued that given the novel nature of the Commission’s analysis (in particular, the Commission stated its decision was “a precedent which establish[ed] the framework for the assessment of the legality of this type of conduct”), Google could not be accused of either intentionally or negligently abusing its dominant position. Nonetheless, the Court found that in line with established case law, where an infringing firm could not have been unaware of the anticompetitive nature of its behavior, it acts intentionally. Google was aware of its dominant position in the market for general search services and must have known that its conduct undermined equality of opportunity among the various economic operators in the market, such that its conduct was capable of foreclosing its competitors or otherwise restricting competition.
This post is as of the posting date stated above. Sidley Austin LLP assumes no duty to update this post or post about any subsequent developments having a bearing on this post.