Post by Erin Bibb (Glasgow alumna and Future Trainee Solicitor at Herbert Smith Freehills) and Magali Eben (Lecturer in Competition Law, University of Glasgow).
Google was met with disappointment in November 2021, after the EU’s General Court (‘GC’) delivered a judgment that upheld the European Commission’s 2017 decision imposing a €2.42bn fine on the company for ‘favouring its own comparison-shopping service over others’ in its search engine. This judgment was based on Article 102 TFEU, the abuse of dominance provision in EU competition law. Unsurprisingly, Google filed an appeal with the Court of Justice of the European Union (‘CJEU’) on 20 January 2022, with hopes that the CJEU will accord greater importance to the tech giant’s contribution to digital innovation. It will be a different story, however, as to whether the EU will rely on the decision when interpreting the highly anticipated Digital Markets Act (the Proposal for which was first published by the Commission in December 2020 and the adoption of which is imminent), and particularly the prohibition on ‘self-preferencing’ included in the proposal.
While we wait for the much-anticipated CJEU judgment as well as the final text of the Digital Markets Act (‘DMA’), it might be time to reflect on what guidance can be taken from the General Court’s judgment: for the abuse of self-preferencing by dominant undertakings under Article 102 TFEU, as well as for the self-preferencing conduct which will likely be included in the DMA.
Equal Treatment: Abuse and Legal Certainty
The GC gives some guidance as to how Google’s behaviour constituted a form of anticompetitive abuse under Article 102 TFEU. In the original decision, the Commission objected to ‘the more favourable positioning and display, in Google’s general results pages, of its own comparison shopping service compared to competing comparison shopping services’. This type of conduct became commonly known as ‘self-preferencing’. Although the GC confirmed the Commission’s view that ‘self-preferencing’ was an abusive practice of which Google was culpable, the GC’s reasoning actually differed from that in the Commission’s decision. The GC did not establish a per se ‘self-preferencing’ abuse or a standalone leveraging abuse ([162-163]), but rather focused on ‘equal treatment’ and how Google’s conduct was ‘abnormal’.
The judgment acknowledges a general obligation of ‘equal treatment’ ([155]) that must be honoured by ‘ultra-dominant’ undertakings ([180]) like Google. In order for a ‘self-preferencing’ practice to contravene such a principle, a dominant firm must be shown to have both promoted its own product and demoted those of its competitors ([261]), and in doing so have departed from ‘normal competition’ ([157-165]). The GC stressed that the effects of the two practices (the demotion and the promotion) must be analysed simultaneously, and not in isolation ([376]).
In theory, this judgment can be lauded as a prima facie clarification of the law for the identification of abusive self-preferencing under Article 102 TFEU, perhaps even as a step towards legal certainty for digital abuses of dominance under Article 102 TFEU. Moreover, some might argue that this judgment provides guidance for the identification of ‘self-preferencing’ practices prohibited by the DMA. Yet it can be questioned whether this is a realistic view.
Establishing a new form of abuse
Firstly, the Google v Commission appeal was very fact sensitive in refining the Commission’s new abuse under Article 102 TFEU. It is somewhat doubtful that this principle of ‘equal treatment’ is capable of being applied to a wide variety of digital products in an identical manner. Indeed, the GC reflected on the nature of Google Search’s business to make the claim that the ‘self-preferencing’ conduct was a departure from normal competition. It referred to the ‘universal vocation of Google’s general search engine’ and argued that the search engine’s ‘value and rationale … lie[s] in its capacity to be open to results from external (third-party) sources’ ([176-178]).
Secondly, the GC judgment may have evoked the idea of ‘neutrality’ in search, but within a context of competitive harm. The GC judgment includes a noteworthy reference to Regulations 2015/2120 and 531/2012, which cover net neutrality of open internet access services and public communication networks ([180]). Although the GC states that these references are there only ‘for the sake of completeness’, it considers that the choice made by the legislature to impose general obligations of equal treatment ‘cannot be disregarded when analysing the practices of an operator like Google on the downstream market’, regardless of ‘whether or not legislation calls … for such non-discriminatory access to online search results’. This is because Google has an ‘undisputed ultra-dominant position’ and therefore a ‘special responsibility not to allow its behaviour to impair genuine undistorted competition’ ([180]). This notion of special responsibility, linked to the existence of a dominant position, is one well known to competition scholars and an essential element of abuses of dominance in EU competition law (though its scope remains open to debate). It can be asked whether the GC may be widening the scope of special responsibility to include an element of social responsibility: in essence, it expects that a company like Google would adhere to neutrality principles in the absence of explicit regulation to that effect, because according to the GC the nature of its business model is – and should remain – open. Arguably, the DMA’s focus on ‘gatekeepers’ and their obligations evokes a similar notion of heightened responsibilities for platform providers in a key position. What is important to point out, however, is that it is not sufficient under Article 102 TFEU to evidence discrimination of some kind: it is necessary to show that this conduct is capable of anti-competitive effects. It is perhaps therefore only possible to draw minimal guidance from the judgment, given that prohibited acts of ‘self-preferencing’ in the DMA do not require harmful effects to competition to be shown: it is a broader prohibition.
Thirdly, the GC judgment steers clear of using the term ‘self-preferencing’ when establishing how, in this particular situation, the conduct at hand could harm competition. The careful wording and its case-specific nature prompts a pressing question: where does the alleged abuse of ‘self-preferencing’ fit in in the wider competition and regulatory framework? The Commission included it as a form of harmful conduct in its Proposal for a Digital Markets Act, whose Article 6(1)(d) imposes an obligation on ‘gatekeepers’ to ‘refrain from treating more favourably in ranking services and products offered by the gatekeeper itself … and apply fair and non-discriminatory conditions to such ranking’. The Commission’s drafting is asynchronous with the finding of the GC, which focused on promotion of own services and demotion of competing results with anti-competitive effects.
A disjuncture between the proposed legislation and case law presents a question. In theory, the DMA’s ‘self-preferencing’ might be interpreted as internal discrimination or an ‘unjustified difference in treatment’, and attempt to align with the GC’s take on ‘equal treatment’. However, the DMA and Article 102 TFEU have different aims, with the DMA promoting ‘fairness and contestability’, whose relationship has not yet been clarified. The DMA’s text specifically states that the measures taken to adhere to it have to be evaluated in light of its objectives. Though the DMA acts ex ante and competition law ex post, it is not unimaginable that practices established to conform with the DMA may conflict with those required under Article 102 TFEU, and vice versa. Enforcement agencies, both at Union and domestic level, will be faced with determining which authority takes precedence in case of such conflict.
In an unpredictable online environment, it is vital that any chance to promote legal certainty is grasped with both hands. Given that the Commission’s decision raised questions regarding legal certainty, many speculated, and perhaps hoped, that the GC’s appeal judgment and the DMA would offer solutions. Arguably, cementing legal certainty by way of the DMA is even more important, given that it has been formulated as an ex-ante instrument, the purpose of which was to construct a mechanism that prohibits harmful behaviour before such abuse is allowed to occur. Yet, as the DMA’s formulation of self-preferencing does not match that which the GC set out in its judgment as an obligation of equal treatment, it would seem that definitive legal certainty has not yet been reached. Perhaps this is something that the CJEU will aim to address in Google’s upcoming appeal.
As such, if the EU Legislature itself claims that it seeks to increase and safeguard the fundamental principle of legal certainty through the creation of the DMA, it would be prudent to either interpret the wording of Article 6(1)(d) to reflect the GC’s judgment or clarify the relationship between the two. Without it, the European Commission, Parliament and Council will have created a conscious dichotomy between an abuse of ‘self-preferencing’ contained in the DMA and ‘equal treatment’ in case law. Instead, they could provide uniformity to the area, bringing a general obligation of ‘equal treatment’ to the forefront of enforcement.
If the EU were to adopt the GC’s two-pronged approach in its interpretation of the DMA, it is necessary to consider objective justifications to this behaviour. This might lead one to ask whether a gatekeeper would ever be able to justify the promotion of its own products, without demoting those of its competitors, if it were to create specific benefits? As the ban on self-preferencing is in Article 6 of the DMA (and thus subject to being ‘further specified’), there is apt opportunity for the Commission to re-evaluate its position on self-preferencing and create an opportunity for isolated product promotion to stand as an objective justification. Yet this seems highly unlikely, both in light of the limited justifications foreseen in the text of the DMA and in light of the EU’s continued promotion of a wide range of services offered by a variety of competitors.
Impacts on Innovation
Innovation and harm in the Judgment
The GC noted that Google enjoys an ‘open’ infrastructure as a general search engine, owing to its ability to accept results from third party sources and its strategic exploitation of considerable network effects and economies of scale as a result. From the GC’s commentary from paragraph 177 of the judgment onwards, it seems that digital entities are arguably subject to a higher standard of scrutiny, given their intrinsic ability to act as a bridge between multiple third parties.
The GC has set a significant precedent for reasons other than those discussed above. By imposing an obligation of equal treatment on Google, the GC greatly expands the long-recognised special responsibility imposed on dominant undertakings not to abuse their position in their relevant market. With Google being widely regarded as the dominant general search engine in its market, it is unsurprising that it has devised strategies to maintain its position, including by remaining attractive to consumers. Google argued that the introduction of its comparison-shopping units provided consumers with an enhanced experience, providing readily available visual prompts and reviews in search results. As the GC noted, however, conduct which entails a product or service improvement may still be abusive if it is capable of having anticompetitive effects ([188]). Since the improvement can only be considered at the objective justification stage [(188]) rather than when assessing the abuse itself, the opportunity to consider the innovation efforts by the dominant platform are very narrow.
Innovation as an objective justification
Even when considering objective justifications for the conduct, the Judgment made rather short shrift with innovation arguments put forward by Google. The GC seemed to argue that that Google had not demonstrated to the Commission that internet users want or expect the service to be offered in this way (as search results from a single source) nor that the product design was, as a whole, necessary to achieve the alleged improvements ([560-562]).
Moreover, Google had an obligation to show that the pro-competitive effects it put forward outweighed the negative effects on competition, which the GC considered it had failed to do ([560]). Although the GC recognised that Google’s activity may have represented ‘pro-competitive service improvements’ from Google’s Shopping Units ([560], [587]), the GC confirmed that the Commission was not obliged to examine the actual effects of Google’s conduct on competition, even with regards to innovation for the benefit of the consumer ([443]). Even if the GC or Commission had taken these factors into account in their analyses, the GC made clear that this did not ‘outweigh’ the negative effects on competition created by Google’s apparent disregard for the principle of equal treatment ([588]), consisting of a decrease in traffic to competing comparison shopping products which had the potential to foreclose them from the market.
This is somewhat surprising, given that the GC admits that innovation is particularly ‘key’ to commercial success in the digital sector ([171]). If innovation is the cornerstone of the digital industry, should this welfare-enhancing factor not have been at the forefront of the GC’s analysis, or perhaps even given the benefit of forming an objective justification? Aside from learning where digital undertakings stand in relation to equal treatment, perhaps the GC’s judgment cements Europe’s stance as a jurisdiction that values the protection of competition through maintaining or increasing the number of competitors, rather than concentrated with the few large-scale innovators. Although admirable, it is important to ask how that objective fits with the current and future development of innovation in digital markets, and the benefits to consumers that it delivers.
Google itself and other dominant undertakings in the digital sphere will no doubt feel increased pressure not to breach this additional ‘arm’ of their special responsibility following the GC’s judgment. As a result, there is a risk that these companies that inherently revolutionise their markets will intentionally slow their individual innovation efforts out of fear of being faced with expensive repercussions for implementing novel strategies. It has been argued that this approach to innovation incentivises innovation by smaller undertakings, and indeed it seems that this is the argument which the Commission finds most convincing in its decisional practice. Yet perhaps there is also scope for the CJEU to consider the incentives to innovate of dominant undertakings and the quality and form of innovation they are able to provide.
It would therefore be a welcome development to see the CJEU strike a justiciable balance between meaningful consumer welfare-enhancing innovations and truly damaging anticompetitive conduct when hearing Google’s appeal, not just when assessing objective justifications, but particularly when considering the harmful effects of the conduct. As it is a recurring central focus of many online businesses, it will be interesting to note whether the CJEU grasps this pivotal opportunity to definitively demonstrate the value derived from innovation in the European digital economy.
Yet it is not only national and European Courts that will benefit from the clarity provided by the CJEU. In fact, those enforcing the DMA will undoubtedly seek guidance from the upcoming judgment in terms of the extent to which innovation could be prioritised going forward. Given that one underlying objective of the new Proposal was to achieve more innovation in a competitive online environment, it will be particularly interesting to see how much direction the Commission will take from the CJEU when enforcing the DMA.