Introduction
Big tech platforms possess characteristics that lead to entrenched structural power and a lack of competition. One notable feature is the existence of extreme economies of scale, arising from the minimal marginal costs associated with additional business users or end users. Furthermore, the services offered by these platforms exhibit strong network effects and an unprecedented capacity to connect business users with end users through their multi-sided nature. Other key characteristics include significant data advantages, lock-in effects, a limited ability for multi-homing, and vertical integration. These features lead to winner-takes-most (or all) outcomes and allow firms to leverage their market power into adjacent and new markets, further expanding the reach of their ecosystem. An additional issue is that the power of these companies stretches to encompass not just the market but also realms that are, usually, set apart from the competition law realm: one example is power in the political sphere, including by way of being able to target individual users, or by way of exerting power over the narratives at play. But even without this additional concern – and thus staying within classic competition law – there is reason to consider the efficacy of the current competition law tools.
Under competition law we have two main tools to address this power and tackle anti-competitive behaviour of big tech: abuse of dominance cases under Article 102 TFEU and merger review under Regulation 139/2004. More recently, the Digital Markets Act (DMA) was adopted to complement competition law and promote the contestability and fairness of digital markets. We contend, however, that these tools are ultimately not effective in countering the negative effects of the entrenched structural nature of big tech power, either due to their inherent limitations, or the way they are currently implemented. Thus, we believe that we need to rethink the tools available to the Commission and strengthen its ability to tackle structural power at its core. We argue that a greater focus on structural remedies, i.e. breaking up companies, might be the solution we are looking for.
In the next section we briefly outline the current tools at the Commission’s disposal and explain their limitations when it comes to addressing the entrenched and structural nature of the power of big tech companies. We then propose a number of ways to increase the Commission’s ability to address entrenched structures of power, both by improving (the use of) existing tools and by creating novel tools.
Limitations of our current tools
Article 102 TFEU
Article 102 TFEU prohibits abusive behaviour by firms holding a dominant position. While an abuse can be of an exploitative nature, most abuse of dominance cases concern practices having an exclusionary effect on actual or potential competitors. Upon finding an infringement, the Commission may, under Article 7 of Regulation 1/2003, impose on the undertaking “any behavioural or structural remedies” in order to bring the infringement to an end. The ability to impose remedies is subject to the principles of effectiveness and proportionality. Accordingly, when there is a choice, the least onerous remedy must be chosen (Case T‑111/08, Mastercard). Based on these principles, because of the intrusiveness of structural remedies, Article 7 establishes that they “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”. AG Kokott noted that behavioural remedies should be preferred, and structural remedies may only be imposed in exceptional cases (AG Opinion in Case C‑449/21, Towercast, para 63). The reliance on behavioural remedies, however, is often not particularly effective in reintroducing competition, as in the Microsoft I (Case COMP/C-3/37.792) and the Google Shopping (Case AT.39740).
The Commission could of course make use of the instrument of structural remedies, when finding an infringement under Article 102 TFEU. Indeed, if a structural remedy is the only way of bringing an infringement to an end, it will most likely be proportionate (Joined Cases C-241/91P and C-242/91P, Magill, para 91). In its press release concerning the Google AdTech case, the Commission stated that:
“in this particular case, a behavioural remedy is likely to be ineffective to prevent the risk that Google continues such self-preferencing conducts or engages in new ones. Google is active on both sides of the market with its publisher ad server and with its ad buying tools and holds a dominant position on both ends. Furthermore, it operates the largest ad exchange. This leads to a situation of inherent conflicts of interest for Google. The Commission’s preliminary view is therefore that only the mandatory divestment by Google of part of its services would address its competition concerns”.
However, using structural remedies also has inherent limitations. Importantly, following the case law, a structural remedy needs to fit the specific infringement and does not address the powerful position from which the anti-competitive behaviour stems. In fact, remedies must correspond with the theory of harm relied on in a case, in order to be proportionate (Case T-310/94, Gruber + Weber v Commission). This raises the question of whether the strict relationship between case-specific harm and structural remedies relating directly to that case-specific harm hinders their effectiveness. Below we come back to this point, when we focus on ways forward.
Merger review
The EU has a system of ex ante merger review, in which those mergers creating a dominant position leading to a significant impediment of effective competition may be prohibited or be the subject of a remedy which redresses the anti-competitive problem. In this setting, structural remedies may also be imposed. However, the option of imposing structural remedies in a merger control situation falls short when it comes to addressing the power of big tech corporations. Firstly, and importantly for our argument, intervening in new mergers does not solve the problem of structural entrenched power. Secondly, after the judgment of Illumina v Commission (Joined Cases C‑611/22 P and C‑625/22 P), we still do not have a clear way to review acquisitions of start-ups when these fall under the notification thresholds, which would, at minimum, not strengthen already entrenched positions. Thirdly, the merger control regime more generally has difficulties including a proper assessment of (the impact of) innovation and how to incorporate the trajectories of innovation in the ex-ante nature of merger control. Below we also propose a way forward in how to better deal with innovation.
Proposed solutions
We propose several ways forward. The first three ways focus on improving (the use of) existing tools. The other two focus on creating novel tools. We provide them as a set of distinct solutions, but of course, a combined use would be possible and – if entrenched structures of power are to be addressed as such – preferred.
Article 102 TFEU
Firstly, as to Article 102 TFEU, where the option of structural remedies is legally possible but narrow in scope, it would be possible to consider either an interpretation of the ‘proportionality’ requirement where it is possible to take into account a series of infringement-decisions against the same company, or for the same behaviour addressed at another company.
The notion of proportionality is, in essence, that the punishment fits the crime: in such cases, akin to the option provided for by Article 18 DMA (see below), the Commission would consider the effectiveness of past remedies as element in the proportionality of imposing structural remedies. This loosens the tie between specific infringement and remedy. This might also lead to considering a ‘harsher’ remedy – for example one in which not merely one branch, or one business line is divested – than follows from the latest infringement-decision. The question is, of course, whether the legal basis for such a widening of the scope can be found in the notion of proportionality, which has been – and logically so – been interpreted restrictively by the Courts.
Conversely, it might also be possible to think of construing the same issue in a novel, but coherent theory of harm. Here, one would build upon the acceptance that protecting the ‘structure of the market’ is an accepted goal for the use of competition law – such as envisioned in Continental Can and also mentioned as such a goal in T-Mobile. As a thought experiment, the theory of harm could be ‘a series of continuing infringements stemming from structures of entrenched market power’. This would tie to the possibility of imposing a structural remedy that would be closer to ‘breaking up’ than a structural remedy that is limited to ‘divestiture of a specific part’ and thus, also, be closer to limiting a powerful market position directly.
Merger review
Secondly, as to merger control, several tweaks to the system are possible. A first tweak is to introduce a (segment of the) merger control regime that does stretch to mergers that now fall below the thresholds, if not on the EU-level then at least on the national levels. The Commission is looking into this, as Teresa Ribera has been tasked with addressing the risks of killer acquisitions. A second tweak is to consider theories of harms that go beyond the current theories of harm, and focus on – as Lisanne Hummel argues in her forthcoming dissertation – a ‘diversity of the innovation ecosystem’ theory of harm. This would increase scrutiny of innovation-focused acquisitions by making use of economic theories such as complexity theory, or evolutionary economics. A third tweak would entail to also strengthen the use of structural remedies in merger control, akin to the notion introduced when discussing Article 102 TFEU: a series of small acquisitions might need to be redressed by a wider imposition of structural divestiture before a merger could be given the green light.
The Digital Markets Act
Thirdly, the DMA was adopted to close gaps in competition law in relation to digital markets, by creating ex-ante obligations for gatekeepers. The notion of gatekeeper has elements of the issue at stake in this blog-post: the entrenched positions of power of big technology companies. The DMA can prevent gatekeepers from exploiting their position of power, by imposing unfair terms and by further decreasing market contestability. However, akin to Article 102 TFEU, the DMA does not address powerful positions directly. Greater fairness and contestability in digital markets are pursued by controlling gatekeepers’ behaviour.
The Commission has the power to impose structural remedies under Article 18 DMA; it may do so following “a market investigation for the purpose of examining whether a gatekeeper has engaged in systematic non-compliance”. More precisely, the Commission can impose “any behavioural or structural remedies which are proportionate and necessary to ensure effective compliance with this Regulation”, if a market investigation shows that a gatekeeper has systematically infringed its obligations under the DMA and that it has “maintained, strengthened, or extended its gatekeeper position”. Systematic non-compliance entails that there have been at least three non-compliance decisions against a gatekeeper within a period of eight years.
While structural remedies could potentially be far-reaching, in the form of forced divestiture, comparably to Article 102, they need to “proportionate and necessary in order to maintain or restore fairness and contestability as affected by the systematic non-compliance”. Structural remedies seem to be particularly suitable when it comes to obligations that regulate conflict of interest and access to infrastructure in case of vertically integrated gatekeepers. In these cases, imposing structural remedies would clearly remove gatekeepers’ incentives to advantage their own services.
We call for a broad interpretation of Article 18, to ensure the effectiveness of the DMA’s enforcement mechanism across all DMA obligations. In particular, we argue that structural remedies should not necessarily have to be closely tied to obligations previously infringed by the gatekeeper to meet the criteria of proportionality and necessity. A broad interpretation is justified by the significant market power of gatekeepers and the guarantees for proportionality and necessity already embedded in Article 18, i.e. the requirement for systematic non-compliance and that the gatekeeper still holds or has further consolidated its gatekeeping position, despite the enforcement actions taken by the Commission (DMA, recital 75). Thus, it should be sufficient to show that a gatekeeper has frustrated more broadly the objectives of the DMA, i.e. contestability and fairness, to impose structural remedies to stop further infringements.
The New Competition Tool (NCT)
Fourthly, we strongly support the idea that undergirds the NCT. The NCT is a market investigation tool that was proposed by the European Commission some years ago, morphed and reformed into different ideas, and has recently been picked up again in the Draghi report. Its purpose is to enhance the ability to address structural competition issues that cannot be adequately handled under existing EU competition law. The NCT offers several key benefits that complement and extend current competition law. One advantage of the NCT is that it does not concentrate on wrongdoings of individual undertakings, but instead targets structural issues that hinder competition. By focusing on the broader market features, such as high barriers to entry and concentration of power, the NCT seeks to improve the overall competitive environment, rather than addressing isolated anti-competitive behaviour. By addressing the underlying causes of competition problems, this approach aims to deliver longer-term solutions for market failures. Furthermore, the NCT promises more timely interventions, helping to prevent prolonged market distortions. While current competition investigations can be lengthy and reactive, the NCT aims to allow quicker action to rectify competition problems and prevent harm to markets and consumer welfare.
The NCT does represent a clear solution to the problem presented here, as it is designed to address deeper structural issues that create persistent challenges to competition, such as entrenched market power or systemic imbalances. It will allow the Commission to impose structural remedies, if these are needed to remedy the problems identified on specific markets. When it comes to big tech companies, it might very well be that structural remedies will be necessary, as experience has shown that behavioural remedies are often ineffective in restoring competition.
Interestingly, Draghi suggests that one specific situation which calls for the use of the NCT is when past enforcement has been ineffective, both in terms of infringement decisions and merger review. He suggests that parties to competition decisions can be required to report metrics used for evaluating ex post enforcement. This would facilitate future assessments of competition and intervention with the NCT, if needed. In this way, the NCT could be a good complement to traditional competition law enforcement, representing a safety net in case enforcement under Article 102 has been unsuccessful.
The notion of power
Finally, and fifthly, we propose to reconsider the notion of power more fundamentally. Here, we draw attention to the notion that the power of some very large (technology) companies stretches beyond the mere market (as argued elsewhere). This power also covers ((infra-) structures of) political power and discursive power, which impacts plurality of media-voices, the quality and accessibility of the digital public sphere, and democratic structures. The roles of platform users as consumer (or content-creator) and as citizen are, here, not neatly separable. Taking both these elements into account in merger control and the NCT would mean that also structural remedies that are necessary to protect plurality of media-voices, the digital public sphere, and democratic structures come more firmly into view.
Conclusion
The structural power wielded by big tech companies poses threats to our markets and society. Their characteristics – economies of scale, network effects, data advantages, and multi-sided platform dynamics – enable these firms to entrench their dominance, expand into adjacent markets, and influence realms beyond traditional market competition. While the tools available to the Commission, including Article 102 TFEU, merger reviews, and the Digital Markets Act, provide a foundation for addressing anti-competitive behaviour, they fall short in curbing the structural nature of this power. This underscores the urgent need to rethink and enhance the regulatory framework. We have argued that a greater emphasis on structural remedies, with the possibility of breaking up dominant firms, may offer a more effective means to restore competition and limit the far-reaching power of big tech. Our proposals are not merely along the lines of a ‘better use’ of structural remedies, but also include novel theories of harm, and reconsider the notion of power more fundamentally to protect the EU’s constitutional values.
________________________
To make sure you do not miss out on regular updates from the Kluwer Competition Law Blog, please subscribe here.
