Google’s behaviour in the advertising technology sector (Ad Tech) is under attack or, at the very least, it is being intensely scrutinised on both sides of the Atlantic. Google is already subject to several antitrust suits in the U.S. lodged against it by the Department of Justice (DoJ) and individual States. In the European Union, life is not much easier. On 14 June, the European Commission by sending Google a Statement of Objections (SO), gave notice that it also takes the preliminary view that Google’s behaviour in the Ad Tech sector has breached (EU) antitrust rules.
What is particularly remarkable is the remedy that is being considered by the U.S. as well as now the EU antitrust authorities. Both favour (or seem to favour) a structural remedy, sometimes referred to as a ‘break up remedy’. The U.S. DoJ is requesting a divestiture of, at minimum, the Google Ad Manager suite along with any additional structural relief as needed to cure any anticompetitive harm. It seems, as I have anticipated earlier, that this has emboldened the Commission to do the same. The Commission states in its press release that its preliminary view is that “only the mandatory divestment by Google of part of its services would address its competition concerns”. In this regard, Executive Vice-President Vestager noted in her remarks on the SO that the Commission has been in close contact with the DoJ as regards the case, and that the Commission and the DoJ share the same view as to what is good for competition and, ultimately, also how to best remedy the issues.
This contribution will shortly address the alleged infringement. It will, however, primarily focus on the remedy that is considered by the Commission. It will discuss the Commission’s remedial power and highlight two aspects of the remedy that seem particularly interesting: its aim to prevent future infringements and the fact that it must address conflicts of interest. An abridged description of the Ad Tech sector, the Ad Tech Stack and Google’s position can be found in my previous post on the same topic, see here.
Google’s alleged infringement(s): the EU perspective
According to its press release, the Commission has preliminarily found that Google, contrary to Article 102 TFEU, has abused its dominant position on the markets for i) publisher ad servers and ii) for programmatic ad buying tools for the open web, by favouring its own ad exchange service (AdX). Although it seems that the Commission considers ‘favouring’ as including a multitude of (complex) actions by Google, it has grouped it under two headings (Executive Vice-President Vestager spoke of two infringements).
Firstly, Google allegedly favoured AdX in the ad selection auction run using its publisher ad server (DFP) by, inter alia, informing AdX in advance of the value of the best bid from competitors which it had to beat to win the action. Secondly, Google also allegedly favoured AdX by using its ad-buying tools (Google Ads and DV360) by, inter alia, ensuring that Google Ad was avoiding competing ad exchanges and was mainly placing bids on AdX, making AdX the most attractive ad exchange.
The Commission’s preliminary findings, as far as can be distilled from its press release, show similarity with the conclusions of the DoJ. Like the Commission, the latter found that Google was dominant on the publisher ad servers and the advertiser ad networks stages of the Ad Tech Stack. In addition, the DoJ also considers that Google’s behaviour involved illegal preferential treatment of its own Ad Tech products.
There are also differences between the findings of the DoJ and the Commission’s preliminary conclusions (which may potentially be explained by a different market definition and a different antitrust regime). Contrary to the DoJ, the Commission does not (preliminary) consider that Google is dominant at the ad exchange stage of the Ad Tech Stack. Moreover, whilst the Commission seems to target solely Google’s alleged favouring practices, the DoJ takes issue with a broader array of practices including certain Google acquisitions, tying practices, manipulative practices, and the drawing up of restrictive platform rules.
The Commission’s power to impose antitrust remedies
The Commission’s preliminary view is that solely a divestiture will remedy Google’s alleged breach of EU antitrust rules. Only the divestment by Google of part of its services would address its competition concerns.
Article 7 of Regulation 1/2003 empowers the Commission to impose upon an infringing undertaking “any behavioural or structural remedies” in order to bring the infringement, properly established by the Commission, to an end. Regulation 1/2003 hence uses the dichotomy between behavioural and structural remedies. These concepts, however, have not been (clearly) defined in EU legislation nor in EU case law. Divestiture remedies are nonetheless generally considered a prime example of structural remedies.
The power of the Commission to impose (behavioural and/or structural) remedies is constrained by the principles of effectiveness and proportionality. First, pursuant to Article 7, the Commission may impose remedies “to bring the infringement effectively to an end”. If the Commission decides to impose remedies, it must ensure that the remedies are effective. Second, the remedies imposed by the Commission must be proportionate. This requirement also follows directly from Article 7 which states that the Commission may impose remedies “which are proportionate to the infringement committed and necessary to bring the infringement effectively to an end”.
According to the case law, the requirement of proportionality demands that the remedies adopted by the Commission must not exceed the limits of what is appropriate and necessary in order to attain the objectives legitimately pursued by it. When there is a choice between several appropriate remedies, recourse must be had to the least onerous means, and the disadvantages caused must not be disproportionate to the aims pursued (Mastercard, para 323). If a remedy is the only way of bringing an infringement to an end, it will most likely be proportionate (Magill, para 91).
Article 7 dedicates a segment specifically to structural remedies. It adds a rather confusing and unfortunate passage which notes that “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”. This passage has led some to believe that there is some sort of presumption in favour of imposing behavioural remedies. Advocate General Kokott referred in this regard to the “primacy” of behavioural remedies and noted that the Commission may impose structural remedies only in exceptional cases (Opinion AG Kokott in Towercast, para 63).
The idea of preference for behavioural remedies may, however, be questioned. Albeit the Commission stays away from imposing structural remedies – indeed, the Commission has never unilaterally imposed such remedies – Article 7 does not seem to create a (legal) preference for behavioural remedies. In fact, the quoted passage from Article 7 seems to merely articulate the consequences of the application of the principles of effectiveness and proportionality which the Commission must take into account when considering any remedy.
Firstly, the fact that structural remedies can only be imposed where there is no equally effective behavioural remedy is a logical consequence of the principle of effectiveness, which forces the Commission to pick the remedy that is most effective. Second, the passage of Article 7 that notes that structural remedies can only be imposed where any equally effective behavioural remedy would be more burdensome for the undertaking concerned than the structural remedy follows logically from the principle of proportionality, which dictates that the Commission must impose the least burdensome remedy (the ‘least restrictive measure’). The quoted passage, therefore, does not seem to add a whole lot.
The Commission, notwithstanding the lack of a preference for behavioural remedies, has until today only once imposed a divestiture remedy in an infringement decision. In ARA Foreclosure the undertaking concerned agreed to transfer part of its household collection infrastructure which it owned to one or several buyers independent of it. In this case, the divestiture remedy was, however, the result of a cooperative infringement procedure as the undertaking and the Commission agreed to the divestment in return for a lowering of the fine in the final infringement decision. In the context of this cooperative procedure, the undertaking not only agreed to divestiture but also declared that the remedy was proportionate. This is important because this way it becomes very challenging for the undertaking, during a potential action for annulment, to successfully claim that the remedy is disproportionate.
Preventing future infringements
What ultimately matters when assessing a possible remedy is its effectiveness and proportionality in a specific case, taking into account the circumstances and particularities of the case. In order to assess a remedy’s effectiveness and proportionality – particularly when assessing whether there are less restrictive measures that may attain the same goal equally well – it is necessary to ascertain the remedy’s objective(s). It is therefore interesting that, as seems to follow from the Commission’s press release, the main or at least an important objective of the divestiture remedy is “to prevent the risk that Google continues its self-preferencing conducts or engages in new ones”. The Commission’s remedial goal thus seems to be to prevent future (similar) infringements. Accordingly, the Commission seeks a prophylactic remedy.
Preventing future anticompetitive conduct is, in principle, a legitimate enforcement objective. Fines, for example, also aim to prevent future breaches of the antitrust rules since they aim to deter undertakings (specifically the undertaking that has been fined as well as other undertakings more generally) from engaging in anticompetitive behaviour. Similarly, remedies may also prevent a repetition of the conduct complained of (Commercial Solvents, para 46).
The remedy imposed by the Commission in the potential infringement decision should thus be effective and proportionate in relation to preventing the infringement, duly established by the Commission, from being repeated by Google. It may, however, be queried how far the power to prevent antitrust infringements stretches. Put differently, what is the Commission allowed to prevent? What behaviour may it prevent? It may be hard to argue that where the Commission has found that a dominant undertaking engaged in predatory pricing the Commission may (also) impose remedies that prevent the undertaking from tying two of its products, notwithstanding that this may potentially be anticompetitive too.
In this regard, the way the Commission frames the infringement may be particularly significant. It is clear from the Commission’s press release as well as from the remarks by Executive Vice-President Vestager on the SO that, in order to describe the (possible) antitrust infringements, the Commission uses the rather general terms favouring and self-preferencing instead of (or alongside) the specific forms of behaviour that Google has engaged in (like the Commission seems to have done in Google Shopping, which arguably also involved favouring). By arguing that the infringement committed by Google involved favouring or self-preferencing it may be easier to legitimise that a broad array of specific forms of behaviour that can amount to favouring or self-preferencing must be prevented in order to ward off similar infringements by Google in the future. Since favouring or self-preferencing can involve many specific forms of behaviour, the Commission may be able to argue that a behavioural remedy (which is often more tailored to the specific forms of behaviour that infringed the antitrust rules) could find it difficult to prevent favouring or self-preferencing and is, therefore, likely to be ineffective. One form of specific favouring behaviour may be too easily replaced by a different, slightly distinct form. Taking such a view, the Commission will be able to argue that divestiture is required as other (behavioural) remedies are very likely to be ineffective. It may be distilled from Vice-President Vestager’s remarks on the SO that such an argument is indeed being made.
Conflicts of interest
Apart from the fact that favouring or self-preferencing may encompass many different specific and complex forms of behaviour – all of which may be difficult to address and hence to prevent by a behavioural remedy – a more fundamental factor that, according to the Commission, seems to warrant a divestiture remedy is the presence of “pervasive” conflicts of interests currently faced by Google in the Ad Tech sector. It seems that the Commission, similar to the U.S. DoJ, considers this to be the root of the problem.
Google is dominant in respect of the technology used by publishers to offer advertising space for sale as well as the leading tools used by advertisers to buy advertising space. Additionally, its ad exchange, which matches publishers with advertisers each time an ad space is sold, has a very strong position. Google’s structure, thus, spans across the whole Ad Tech Stack. It has ownership of the entire value chain. The Commission is concerned that Google’s structure therefore inherently incentivises it to abuse its dominant position(s). This is even more worrisome given the characteristics of the Ad Tech sector. In particular, ad exchanges are two-sided markets and Google is dominant on both sides (sell-side as well as buy-side) of this market. It is for this reason that Vice-President Vestager noted that this case is not ‘just another leveraging case’.
In this regard, as already alluded to in my previous contribution, there is an argument to be made that Regulation 1/2003 envisioned structural remedies, especially for cases such as the one at hand. Recital 12 states that “changes to the structure of an undertaking as it existed before the infringement was committed would only be proportionate where there is a substantial risk of a lasting or repeated infringement that derives from the very structure of the undertaking” (underlining added).
One can, nonetheless, question whether a situation such as that referred to by Recital 12 presents itself here. In this respect, one could point out that Google’s structure does not force it to abuse its dominant position(s) and that, consequently, (potential) abuses do not follow directly from its structure. Put simply, although Google may have an incentive to engage in anticompetitive behaviour, it does have a choice whether to do so or not. One could, moreover, argue that even though Google faces conflicts of interest, such a situation has, in and of itself, not been found to breach the antitrust rules (Opinion AG Rantos in Super League, paras 48, 134).
To impose or not to impose
It should, however, be stressed that the Commission’s preliminary views are… well… preliminary. The fact that the Commission is currently considering a divestiture remedy does not mean that in a potential infringement decision, it will indeed impose such a remedy. The Commission’s views in this respect might be influenced by the proceedings and discussions on remedies in its ‘twin’ U.S. case. Its views on the possible remedy may also be affected by Google’s reply to the SO or the subsequent discussions with that company or by discussions it may have with other market participants. A lot may still happen on both sides of the Atlantic.