Public Interest and EU Competition Law

Abstract

While European Union (EU) competition law has long been understood as a variety of public interest law, the extent to which the rules can be applied directly to advance noneconomic public interest-oriented goals is more contentious. This contribution considers whether and how such concerns can be accommodated within the framework of Articles 101 and 102 of the Treaty on the Functioning of the European Union (TFEU). It considers both the conventional approach to addressing public interest concerns within the analytical structure of the antitrust rules and also how broader public interest objectives have shaped recent EU-level enforcement efforts in three key sectors: the liberalizing public utilities markets, the pharmaceutical sector, and the digital economy.

The function of [the EU competition] rules is precisely to prevent competition from being distorted to the detriment of the public interest[.] ¹

I. Introduction

Competition law has long been understood as a variety of public interest law, broadly construed. ² Whether stemming from its capacity to increase the size of the pie available to all, or specifically to increase access to necessities for the most vulnerable, the task of taming private market power almost inevitably relies upon justifications related to the broader public interest. ³ Yet in a contemporary landscape where even mainstream scholars are calling for an end to capitalism or at least its progressive rethinking, ⁴ the simple pursuit of an economically oriented understanding of consumer welfare has faced criticism as being both ill-conceived and inadequate. ⁵ This contribution explores the capacity of EU competition law, in particular, to pursue more specific public interest objectives over and above its established task of “making markets work better.” ⁶

The term “public interest” is inevitably somewhat nebulous and imprecise. Within the EU, the newly revived debate has been prompted by the prohibition of the proposed Alstom/Siemens merger, ⁷ a decision heavily criticized by the certain Member States on the basis that it fails to take due account of wider industrial policy considerations. ⁸ Environmental protection and the desirability of ensuring sustainability in modes of production is another area of focus, ⁹ chiming with policy debates regarding the necessity to tackle the incipient climate crisis. Similarly, concerns about societal inequality, ¹⁰ unfair labor practices, ¹¹ and a host of problems associated with the increasingly pervasive digital economy ¹² —from violations of data privacy to the spread of fake news—have made their mark in discussions regarding the optional scope and goals of contemporary competition law. This article adopts an equally expansive definition of public interest, to encompass any application of the competition rules which embraces values that extend beyond the conventional (if disputed) wisdom that “given its economic character, competition law aims, in the final analysis, to enhance efficiency.” ¹³

This article focuses on the EU antitrust rules, namely Articles 101 and 102 of the Treaty on the Functioning of the European Union (TFEU). Public interest concerns may feed into the enforcement of these rules in multiple ways. Rival economic actors may enter into outwardly restrictive agreements in furtherance of noneconomic public policy objectives, such as the promotion of animal welfare ¹⁴ or access to justice. ¹⁵ To what extent can these latter goals be taken into account when considering the compatibility of the underlying coordination with Article 101? A dominant undertaking may impose a pricing policy which has the effect of both foreclosing rivals but also of enhancing social inclusion, ¹⁶ or conversely one which reflects an economically defensible monopoly profit for successful innovation but nonetheless prevents access for the most socially precarious. ¹⁷ The question, again, is whether these wider public interest considerations can feed into the otherwise narrowly tailored assessment of whether the relevant conduct should be classified as abusive under Article 102. Moreover, where other fields of law exist precisely to protect the claimed public interest concerns, such as data protection or consumer protection rules, when might (non)compliance with parallel regulatory requirements be a pertinent antitrust consideration?

The conventional wisdom is that Articles 101 and 102 focus on the protection of “competition as such.” This comparatively expansive concept links to the role played by the competition rules in developing and reinforcing the single market structure but, conversely, does not extend to noneconomic public interest justifications. Yet EU competition law cannot avoid the inconvenient reality that antitrust, if interpreted broadly and without reference to the underlying context, might serve to inhibit or even prohibit large swathes of activity that the average European would consider to be very worthwhile indeed. This article explores this tension, considering both the orthodoxy, but also going beyond it to suggest a variety of ways in which broader public interest concerns may feed in, more obliquely and typically without explicit acknowledgment, to the Commission’s enforcement practice. The resulting legal position is quite distinct from the assumption that public interest concerns are accommodated within the competition rules primarily through an approach of derogating from the application of the latter. Yet, the examples considered are both context-specific and less than all-encompassing in scope, thus demonstrating the continuing difficulty of squaring this ostensibly technocratic, keenly focused field of law with more amorphous and disputed public interest concerns.

The purpose here is not to argue for or against greater incorporation of public interest considerations within the substantive antitrust analysis, although there is plenty of literature that does precisely this. ¹⁸ Rather, the emphasis is on the extent to which such concerns can and have been incorporated within the current EU legal framework, in particular in light of the nominal movement toward a “more economic approach.” The article is structured as follows. Section II sets out the wider background to the consideration of public interest within EU competition law. Section III outlines the established approach to accommodating such concerns within this context, including recognized derogations to the concept of economic activity, the extent to which public interest can be accounted for in substantive antitrust analysis or included within the exception rules available under Articles 101 and 102, and the possibility of justifying such conduct by reference to the State action defense. Section IV takes a different tack, departing from this conventional narrative to suggest the de facto recognition of public interest concerns in the Commission’s enforcement practice in three areas: public utilities sectors, pharmaceutical markets, and the digital economy. Both the achievements and the limitations of using antitrust law to pursue wider public interest values in these contexts are considered. Section V concludes briefly.

II. The Broader Context of Public Interest in EU Competition Law

EU competition law is a complex beast. The same rules apply across what has for most of the EU’s history been an ever-expanding range of national jurisdictions and disparate market circumstances. Additionally, the Court of Justice approved a plurality of goals from the outset, including the development of the internal market ¹⁹ and preservation of structural competition in a manner generally considered to reflect Ordoliberal influences. ²⁰ The question of the extent to which EU competition law can and does pursue a wider spectrum of public interest objectives is thus complicated by the complexity of the law itself. This section considers three overarching factors of relevance to the discussion to follow.

First, EU competition law has, for some time, been undergoing a process of modernization and readjustment. ²¹ The system long faced criticism that undue emphasis was placed on the protection of competitors at the expense of the wider competition process. Largely in response, the past decade or so witnessed a more fulsome embracing of the so-called more economic approach to antitrust enforcement. ²² The cornerstone of this development has been the Commission’s adoption of an “anticompetitive foreclosure” standard as the basis for intervention. ²³ While the Court has not (yet) endorsed unequivocally the more economic approach, much of its recent jurisprudence reflects a more nuanced understanding both of the operation of markets and the optimal role for competition law. In particular, the Court has focused its attention principally on behavior that harms competition. ²⁴ In bringing EU law more directly in line with the consumer welfare paradigm, however, the Commission has essentially disclaimed any role for broader public interest considerations. ²⁵

On the other hand, this question is being asked in the wider context of a rapidly shifting landscape for EU competition law more generally. Relevant reforms include greater “decentralization” of enforcement to national competition authorities ²⁶ alongside parallel efforts to encourage a culture of private enforcement; ²⁷ a widespread populist backlash against “neoliberal” Europe generally, with searching questions about “the future of Europe” ²⁸ as a result; and the increasing rhetorical prominence of the notion that competition law should pursue “fair” market outcomes, without much indication as to how this slogan fits with established and evolving antitrust practice. ²⁹ Each of these developments poses challenges to the more economic approach, and all have a potential role to play in determining the receptiveness of EU competition law to public interest concerns.

Second, the EU antitrust rules do not exist within a regulatory vacuum, and public interest considerations find purchase to a greater or lesser extent within the broader legal framework that underpins the single market. There is, first, the wider competition landscape, including the State aid rules which constrain the ability of Member States to intervene in domestic economies, albeit with limited derogations for public interest type interventions. ³⁰ The internal market rules, similarly, aim to ensure a level playing field for the free movement of goods, services, capital, and workers, with a default presumption that State-imposed barriers to “market access” contravene EU law. ³¹ The fundamental freedoms nonetheless admit of the possibility of departing from this approach to further public interest objectives, ³² an amazingly broad range of which have been recognized ³³ —provided, however, that the restriction of competition entailed is proportionate to the legitimate objective pursued. ³⁴

What is notable is the consistent manner in which public interest concerns are approached within this regulatory framework. Whether under the State aid rules or the fundamental freedoms, the underlying assumption is that the public interest, broadly construed, is best served through open and undistorted competition. To the extent that particular public interests may conflict with this free-marketeering vision, the latter may yield in defined circumstances. Yet, although EU law is not insensible to public interest considerations, the implication is that such concerns are extraordinary phenomena, a derogation from the norm of the open and competitive internal market. Accordingly, EU law typically responds to the challenge of countervailing public interest concerns, not by accommodating them within its regulatory framework, but by carving out space outside the rules, and imposing more or less demanding requirements on those which seek to rely upon such derogations. ³⁵

Finally, there is the difficulty of identifying any precise pan-EU “public” interest. Indeed, much of the historic rationale for approaching public interest considerations as derogations stemmed from the fact that, although EU law mandates the universal acceptance of unencumbered competition, there is incomplete harmonization of other regulatory rules. ³⁶ Member States thus continue to pursue a wide spectrum of differing values. Most basically, despite the rhetoric of “ever closer union,” in reality there is no single European “demos.” Instead, as the demoicracy concept reflects, EU law is essentially premised on the mutual recognition of discrete identities and not their merger. ³⁷ Yet EU competition law comprises a single set of rules which must be applied, equally and uniformly, across the Member States. To the extent that public interest considerations are allowed a role here, there accordingly may be considerable difficulty in isolating a single consistent understanding of what exactly is in the putative public’s interest, with significant scope for competing and conflicting visions. ³⁸ Moreover, and more skeptically, there is often a substantial gulf between what the European public, such as it exists, thinks that it wants, and what the Commission thinks that public should want.

III. The Orthodox Treatment of Public Interest in EU Competition Law

It is in the context of this multifaceted and somewhat fluid legal framework that our exploration of the orthodox approach to claimed public interest considerations begins. Although assessment under the EU antitrust rules is attuned to the particular “legal and economic context” of a restraint, ³⁹ the existing jurisprudence does not provide any direct avenue by which these rules may, in themselves, be deployed to pursue discrete public interest objectives. As this section explains, to the extent that the latter is pursued by economic operators in a manner that triggers antitrust scrutiny, it is therefore necessary to derogate from application of Articles 101 or 102, whether by classifying the conduct as falling outside the scope of “economic activity,” by reference to the specific exceptions provided by Article 101(3) or the objective justification concept under Article 102, or by overreaching the competition rules through the exercise of State action.

A. The Concept of an “Undertaking”

Articles 101 and 102 apply only to restrictive behavior by “undertakings.” Although the concept was left undefined in the TFEU, the ensuing case law adopted a so-called functional approach, ⁴⁰ which hinges on the identification of an economic activity carried on by the defendant, ⁴¹ namely the provision of goods or services on a relevant market. ⁴² Conversely, conduct that escapes the definition of economic activity also escapes scrutiny under the antitrust rules. The parameters of the concept of an undertaking thus provide the first potential reprieve for public interest-oriented market behavior.

At first glance, however, the definition of economic activity is remarkably broad, meaning that only a limited subset of conduct can potentially be exempted on this basis. Both the legal status of the entity and the way in which it is financed are, nominally, irrelevant to its classification as an undertaking. ⁴³ Although the crux of economic activity is the provision of goods or services, it is unnecessary for the defendant to be doing so on a for-profit basis: instead, “[t]he basic test is…whether the entity in question is engaged in an activity which could, at least in principle, be carried on by a private undertaking in order to make profits.” ⁴⁴ The upshot is that, in theory, a wide range of market actors may constitute undertakings and come within the purview of the competition rules, including nonprofit entities, ⁴⁵ entities created and governed by public law, ⁴⁶ state companies, ⁴⁷ and those encumbered with public service obligations that make their operations less competitive than comparable private operators. ⁴⁸

The case law, however, recognizes several broad exceptions to this otherwise all-encompassing definition. The first is where the underlying impetus for the activity is the pursuit of social solidarity, defined as “the inherently uncommercial act of involuntary subsidization of one social group by another.” ⁴⁹ Activities benefiting from this exception involve a degree of redistribution ⁵⁰ and typically relate to social welfare provision administered through quasi-private mechanisms. ⁵¹ The second exception arises where, although the activity nominally consists in the provision of (usually) a service, in substance it reflects “the exercise of official authority,” ⁵² and so is more properly understood as an application of public power. Activities potentially benefiting from this exception range from air-traffic control services ⁵³ to environmental monitoring checks ⁵⁴ to the regulation of certain sports. ⁵⁵ A third exception relates to the so-called false self-employed. ⁵⁶ Generally, independent contractors constitute undertakings in their own right under EU competition law. ⁵⁷ Where, however, those individuals are in positions of particular vulnerability and dependence in relation to the purchaser of their services, they may fall outside the definition of an undertaking so as to facilitate, for instance, collective bargaining in pursuit of better working conditions. ⁵⁸

Each of these exceptions evinces a distinct public interest character, whether it is to advance social solidarity, to further nonmarket goals such as public safety or environmental protection, or to safeguard vulnerable workers. In each instance, moreover, successful invocation of the relevant ground for exemption enables the public interest concern to prevail in the event of conflict with the competition rules. Yet in line with the discussion above, the preferred approach is not to seek to embed such considerations within the substantive antitrust analysis, but rather straightforwardly to exclude the latter if necessary, in order to further other, putatively incompatible public interests. In this manner, the competition rules recognize the importance of nonmarket concerns, but seek to protect these primarily through an approach of exclusion rather than accommodation.

B. Accommodating the Public Interest Within Substantive Antitrust Analysis

Within the realm of economic activity, the competition rules apply with full force. Under Articles 101 and 102, suspect conduct is subject to an objective assessment of its impact, likely or actual, upon the functioning of effective competition. Article 101 thus requires the identification either of instances of collusion that can be deemed, “by their very nature, harmful to the proper functioning of normal competition,” ⁵⁹ or of evidence that “competition has in fact been prevented, restricted or distorted to an appreciable extent” as a result. ⁶⁰ The concept of abuse under Article 102 hinges upon the ill-defined notion of “recourse to methods different from those which condition normal competition,” ⁶¹ but is similarly aimed, ultimately, at dominant firm behavior which may “impair genuine, undistorted competition.” ⁶²

Substantive antitrust assessment, accordingly, focuses squarely on competition problems that arise from firm behavior. Although recent case law emphasizes the nuanced, context-dependent nature of this exercise, it does not recognize the possibility of excusing restrictive practices on the basis that they nonetheless further other socially valuable goals ⁶³ —or, conversely, of condemning competitive behavior that nonetheless conflicts with such values. Even leaving aside the question of public interest considerations, the Court has been criticized for an almost pathological focus on “competition as such,” to the point where overall consumer welfare-enhancing behavior may nonetheless fall foul of the antitrust rules. ⁶⁴ Against such a background, the challenge of accommodating what are essentially nonmarket public values directly within the assessment framework is clear.

Under Articles 101 and 102, the subjective intention of defendants is not determinative of the objective characterization of their behavior. In particular, any argument that a defendant may be motivated by “good” intentions—in the sense that its conduct pursues an outcome directly aligned to broader public interest considerations—is not a factor of relevance when assessing likely or actual impact from a competition perspective. In BIDS, ⁶⁵ for instance, the fact that an industry-wide collective reduction in capacity was designed to further industrial policy objectives regarding the development of the Irish agricultural sector was immaterial to its characterization as a restriction of competition by object. ⁶⁶ The Court in Slovakian Banks similarly gave short shrift to claims that the impugned collusion was prompted by the necessity to frustrate an unlicensed (and thus presumably unscrupulous) competitor, insofar as this did not negate the arrangement’s inherently anticompetitive nature. ⁶⁷ In Deutsche Telekom, which concerned a margin squeeze in the telecommunications sector, the defendant made a loss on providing residential fixed-line access which it cross-subsidized through higher call costs. But so too did its competitors, a pricing structure that was directly attributable to governmental policy to promote cheap line access on social inclusion grounds. Yet, once again, the wider background was deemed irrelevant, so that the defendant was held to breach Article 102 by maintaining an “unfair spread” between wholesale and retail prices in the fixed-line market alone. ⁶⁸

The extent to which a parallel “bad” intention might equate to an anticompetitive one is more complicated. In certain circumstances, the concept of the “legal and economic context” of putatively restrictive behavior may be read sufficiently broadly to take account of breach of other social or legal norms. As cases like Telekomunikacja Polska and AstraZeneca demonstrate, ⁶⁹ the fact that a defendant has violated or acted contrary to the spirit of other regulatory obligations may be considered significant in determining whether the same behavior constitutes an antitrust violation. The ongoing investigation into alleged collusion between carmakers accused of restricting competition in the development of emission cleaning technology similarly suggests that congruent public interest concerns (here, environmental protection) can feed into the determination of whether more novel forms of potentially anticompetitive behavior deserve condemnation as “hard-core” restraints. ⁷⁰

Yet the competition jurisprudence suggests a skeptical view of parallel public interests which involve objectives that are clearly distinct from those of open and undistorted competition. Thus, in Asnef-Equifax, the Court held, uncompromisingly, that a claimed breach of the data protection rules—treated as a fundamental rights issue within EU law ⁷¹ —was irrelevant when assessing whether collusion contrary to Article 101 had occurred. ⁷² In Siemens/Areva, concerning a noncompete agreement in the nuclear technology sector, the Commission made zero reference to the much-disputed nature of the underlying activity, in a commitment decision that focused solely on ensuring the competitive health of the sector. ⁷³ And remarkably, across the various competition cases involving products that pose a threat to public health, such as alcohol ⁷⁴ or tobacco, ⁷⁵ there is little discussion of the fact that the most immediate benefit of greater competition—lower prices—is of dubious value insofar as it facilitates greater consumption. ⁷⁶ Conversely, this theme has been explored in detail in the free movement case law, most recently in Scotch Whisky. ⁷⁷ The latter provides an interesting comparator for the antitrust context, insofar as the State-imposed public health restriction at issue comprised minimum pricing rules for alcohol which, unlike an increase in excise duty, steers any inflated revenues to sellers rather than the exchequer.

Moreover, even if parallel wrongfulness may augment or confirm the anticompetitive nature of less clear-cut categories of restrictions, it cannot provide a substitute for anticompetitive behavior as such. This is exemplified by the limitations of Article 102. Although undertakings holding significant market power have a “special responsibility” to protect “genuine undistorted competition,” ⁷⁸ this notional duty cannot function to compel dominant undertakings to, for instance, switch to more efficient fuel sources, or promote diversity within their workforce. ⁷⁹ Indeed, it cannot even, in principle, attack “bigness as such,” a revitalized contemporary concern of so-called hipster antitrust. ⁸⁰ We see further examples of this crucial limitation—that is, the need to anchor competition enforcement within the established framework of the antitrust rules—in relation to the Servier and Facebook cases, discussed in Section IV.

Finally, it is important to clarify that even when assessment under the antitrust rules depends on a demonstration of anticompetitive impact, this is an ineffective conduit by which to introduce public interest considerations. In line with the movement toward a more economic approach, there has been a turn away from form-based prohibitions under Articles 101 and 102, with greater emphasis being placed on what the circumstances of a restraint may indicate about its likely or foreseeable effects. The crux of this development, however, is the extent to which detrimental market outcomes can be successfully inferred (or otherwise) merely from the form that allegedly anticompetitive behavior takes, or whether a more involved consideration of the legal and economic context is required. This is distinct from the question of whether (actual or presumed) nonmarket effects can be brought to bear in assessing the notional competitive impact of market behavior, which, as discussed above, is in principle impossible within the current framework.

C. Justifying Anticompetitive Behavior on a Public Interest Basis

Under Articles 101 and 102, identifying prima facie anticompetitive conduct is merely a first stage in the assessment of legality. Both prohibitions incorporate exception rules which enable defendants to justify their behavior in certain circumstances, and here too questions of broader public interest may be relevant. Article 101 expressly provides for this possibility in its third paragraph, which sets out a fourfold cumulative list of requirements to be satisfied in order for the first paragraph prohibition rule to be declared inapplicable. Although Article 102 provides no such textual basis, the prohibition on abuse of dominance has been interpreted to incorporate the possibility of “objective justification,” which requires either a demonstration of countervailing efficiencies stemming from the alleged restraint or its objective necessity. ⁸¹ Under both provisions, the burden of proof lies with the defendant. The key question is the extent to which these exception rules might be relied upon to exempt anticompetitive behavior on noncompetition-oriented grounds.

Starting with Article 101(3), the threshold requirement is that the prima facie restrictive coordination nonetheless “contributes to improving the production or distribution of goods or to promoting technical or economic progress.” Earlier case law read this criterion expansively, with the Court accepting that any “considerations connected with the pursuit of the public interest” were potentially applicable. ⁸² A myriad of noncompetition public interests are discernible in older Commission practice, including protection of employment ⁸³ and the environment, ⁸⁴ and furthering industrial policy ⁸⁵ and regional development. ⁸⁶ Yet in tandem with the decentralization of the Article 101(3) exception under Regulation 1/2003, the Commission in its accompanying Guidelines asserted that the first prong of that rule refers only to “objective economic benefits,” that is to say, “efficiency gains.” ⁸⁷ While both cost savings and qualitative efficiencies were deemed acceptable, ⁸⁸ this narrower reading appears to foreclose the possibility of relying upon Article 101(3) to pursue noneconomic public interests. Of course, the Commission does not “make” EU competition law as such, as the Court has pointed out often quite forcefully. ⁸⁹ Yet a comparative absence of case law from the Court on the application of the exception rule since the publication of the 2004 Guidelines ⁹⁰ means that the Commission’s somewhat leftfield approach has remained unchallenged and thus unconfirmed—or refuted. A recent survey of enforcement practice at national level revealed a rather mixed bag, whereby domestic competition authorities have adopted differing views on the extent to which they must and shall follow the Commission’s interpretation. ⁹¹ This correlates, arguably, with the dilemma highlighted in Section II regarding the practical impossibility of identifying a single European “public” interest here.

In any event, even if Article 101(3) continues to encompass noneconomic public interests in the more economic era, it requires more than merely identifying such a countervailing concern. Additionally, any consumers “directly or likely affected” ⁹² must receive a “fair share of the resulting benefit.” This criterion may be difficult to satisfy, in particular where certain consumers suffer the immediate detriment (through, for example, higher prices), but society as a whole sees the benefit (through, for example, a cleaner environment). The restraints must also be indispensable—meaning “reasonably necessary” ⁹³ —to the underlying legitimate objective and must not entail the risk of long-term harm to the overall competitive process. Again, the difficulty of applying these criteria in the context of noneconomic public interest is that it may require the balancing of “essentially incomparable” ⁹⁴ phenomena. Whereas one can, fairly readily, assess the overall impact on consumer welfare of behavior that both harms one dimension of efficiency and enhances another, it is more difficult to determine what degree of enhanced private market power is proportionate to secure cleaner air, for instance, or more humane farming practices.

Although the Court has repeatedly recognized the possibility of invoking objective justification under Article 102, the parameters of this exception are less well defined. It is not beyond the realm of possibility that a claim of objective necessity might be successfully invoked on a public interest basis: for instance, to justify a refusal to continue to supply access to an indispensable but highly polluting input controlled by a dominant firm. ⁹⁵ Such a scenario is perhaps farfetched, however, and the Commission’s guidance on Article 102 cautions that, normally, it is the task of public authorities to make public interest-oriented determinations of this sort. ⁹⁶ The second avenue by which a firm may potentially seek to objectively justify its behavior is by demonstrating that any anticompetitive effects “may be counterbalanced, or outweighed, by advantages in terms of efficiency which also benefit the consumer,” ⁹⁷ an approach that would appear to exclude noneconomic public interest concerns. ⁹⁸ A possible explanation for this limitation is that firms acting unilaterally are likely to have less need, or at least less plausible motivation, to depart from normal methods of competition to further nonprofit-oriented public interest objectives. As discussed in Section IV, however, the opposite is not the case, and it will be suggested below that public interest concerns may conversely function to flesh out the amorphous “special responsibility” of dominant firms in certain instances.

D. Overreaching Antitrust in the Public Interest: State Action and Beyond

Finally, even if behavior is classified as economic activity, is held to be restrictive of competition, and cannot be justified by the defendant(s), a further option is available by which, potentially, to reflect public interest concerns. It was noted in the preceding paragraph that the Commission is generally wary of claims that private undertakings can accurately isolate and enforce the notional “public” interest. ⁹⁹ Conversely, where ostensible private economic activity is in fact required by State regulation, then EU competition law recognizes a defense from liability for the undertakings concerned. ¹⁰⁰ Such an exception similarly arises where the extent of existing public regulation within the notional market renders independent competitive conduct impossible, thus similarly precluding anticompetitive activity from arising. ¹⁰¹ The “State action” defense thus provides potential cover for economic activity that restricts competition in furtherance of public interest objectives—but only where, crucially, the requisite degree of State involvement is demonstrated.

The difficulty of relying upon this exception in practice—again, notably, a derogation from the ordinary competition rules—is that the existing jurisprudence interprets the application of the State action defense in an incredibly narrow manner. In short, it is only if the existing regulatory framework removes all scope for freedom of market action from the defendant that the application of the competition rules is excluded. Thus, in BIDS, the fact that the collective scheme had been developed and implemented with the express support of the ministry for agriculture was irrelevant to its treatment under Article 101 since the undertakings had entered into it voluntarily. ¹⁰² More radically, in Deutsche Telekom, the wholesale and retail price levels that comprised the margin squeeze were both approved by the sector regulator. Yet the Court refused the application of the State action defense because the defendant nonetheless retained scope to lobby the regulator and seek a variation of the pricing spread. ¹⁰³ Thus, the fact that the public interest objectives pursued by private parties coincide with public interests recognized by the State, or even that the State approves of and encourages the pursuit of such objectives through private means, cannot in itself serve to immunize private conduct from antitrust scrutiny. There are good reasons for a rigid approach to the State action defense, including the need to avoid its strategic misuse. But the upshot is that it is difficult not only to carve space within the EU competition framework for public interest-oriented behavior of a wholly private nature but also for behavior that has a more plausibly public character.

A further derogation is found in Article 106(2) TFEU, which exempts from application of, inter alia, the competition rules the provision by undertakings of “services of general economic interest” (SGEIs) in certain circumstances. The SGEI concept corresponds broadly to most domestic notions of “public service,” ¹⁰⁴ and Article 106(2) exception enables Member States to opt-out of elements of the “open and competitive” internal market to achieve defined public interest goals at the national level. ¹⁰⁵ What is notable, for our purpose, is that as a derogation from the general application of EU law, Article 106(2) exception should in principle be strictly interpreted. Yet Member States have been given a fairly free hand in determining the categories of service that fall within the SGEI concept in their jurisdiction and are thus immunized from the application of the competition rules. ¹⁰⁶ This, arguably, reflects a rather pragmatic choice to accommodate domestic welfare provision with the strictures imposed by EU law, in order to avoid potentially destabilizing conflicts with national governments on this point. ¹⁰⁷

IV. The Public Interest in Practice in EU Competition Law

The preceding section described an apparent dichotomy in terms of the treatment of public interest considerations within EU competition law. The substantive rules are largely indifferent to such concerns, beyond the trite assumption that enhancing consumer welfare is in the public interest. Yet if interpreted broadly and without reference to the underlying context, competition law might obstruct large swathes of activity that the average European would consider to be entirely worthwhile. The competition framework deals with this dilemma, largely, by carving out such conduct from the substantive scope of the rules: whether ex ante by removing it entirely from the concept of economic activity or through the ex post deployment of an exception rule or the State action defense. The coverage of these derogations is less than complete, however, meaning that the competition rules still occasionally come into conflict with noneconomic public interest concerns: a position that contributes to the rather unfair characterization of the EU as a “liberalisation machine.” ¹⁰⁸ Yet, this article’s account has been straightforward: tightly focused competition concerns on the one side and public interest-oriented derogations on the other.

As this penultimate section demonstrates, however, the reality is more complex. Despite a veneer of technocracy that envelopes contemporary enforcement, antitrust remains a highly political field of law, ¹⁰⁹ and the Commission, moreover, an inescapably political entity. ¹¹⁰ Accordingly, even if EU competition law actively disclaims any direct role for public interest considerations within its substantive core, the competition rules can be—and not infrequently are—deployed in a manner that reflects a distinct understanding of how best to advance the public interest, broadly understood. Competition law has two main advantages within the wider framework of EU law in this regard. First, it can be enforced by the Commission acting alone, without positive approval from the Parliament and Council. It thus provides a more effective tool by which to address contentious issues that may struggle to reach the level of consensus required for direct legislative intervention. ¹¹¹ Second, the competition rules themselves are comparatively wide-ranging and amorphous; if not exactly a “blank cheque” ¹¹² for enforcers, then certainly amenable to progressive development and reinterpretation in light of changing market circumstances. Accordingly, even if it remains necessary, when applying competition law in furtherance of public interest objectives, to make a connection to the functioning of market competition, there is scope to bring the wider context to bear in doing so: what Ezrachi describes as antitrust’s “sponge-like” qualities. ¹¹³

There are two main ways in which broader public interest concerns are manifested in EU-level enforcement. First, there is the mere fact of concentrating recurrent activity in a particular area or in pursuit of particular objectives, which suggests the existence of a distinct “Union interest.” ¹¹⁴ Three areas of recent focus by the Commission are considered below: liberalizing utilities markets, pharmaceutical markets, and the digital economy. In addition, and more contestably, the theories of harm that have been developed in such cases often demonstrate a progressive, occasionally contentious, approach to the competition rules. As will be explored below, where obvious issues of public interest are at stake, this may provide an impetus toward expansion or innovation in the antitrust context. Thus, although EU competition law is not a vehicle by which the public interest can be pursued directly, public interest concerns may feed into the Commission’s approach to enforcement: resulting in more ambitious (and disputable) theories of harm and often marking a departure from the strict—and restrictive—logic of the “more economic approach.”

In the remainder of this section, we consider our three examples in detail. The purpose is more descriptive than normative. The aim is neither to suggest that competition law should be redirected toward the pursuit of noncompetition public interest goals nor that it invariably is. It will, however, be argued that the very particular public interest at stake in each example is apparent in the manner in which the Commission has approached its caseload in these areas, and in how the underlying legal rules are developed and expanded as a result. Accordingly, these examples provide a more granular picture of the relationship between competition enforcement and noncompetition public interests than the orthodoxy outlined in the preceding section otherwise suggests.

A. Competition Law and the Pursuit of Market Liberalization

The first area where a distinct “public interest” flavor to Commission enforcement is discernible is in support of the EU-level liberalization policies which have been pursued, most obviously, in telecommunications and energy markets in the past several decades. The desirability, indeed the inevitability, of liberalizing these typically State-owned monopoly sectors has long been a tenet of faith within the Commission, and competition enforcement commingles with increasingly demanding legislative efforts in this sphere. ¹¹⁵ The public interest here has several dimensions. There is, clearly, a public interest in securing cheaper better quality access to objectively necessary services like broadband, electricity, central heating, and public transport. This enhances consumer welfare in general but can also have a significant redistributive impact: lower energy prices can provide market-based solutions to alleviate fuel poverty, for instance, while increased competition in telecommunications may encourage more rapid broadband expansion in underserved geographic regions. Yet this strand of case law also aligns with more dogmatic assumptions regarding the optional development of the internal market: a perception that “there is no alternative to the liberalisation process.” ¹¹⁶ Accordingly, the focus on liberalized utilities markets also represents a more doctrinaire, disputed view about how such sectors ought to be structured and to function within the internal market. ¹¹⁷ The public interest at stake is thus not merely the desirability of securing cheaper and better services for (the most vulnerable) consumers; it moreover reflects the assumed societal value of the fully realized process of economic integration itself. ¹¹⁸

The first indicator of a distinct public interest is the sheer weight of cases that have been pursued within the telecommunications ¹¹⁹ and energy ¹²⁰ sectors. Yet the mere fact that the Commission brings many cases tells us little about its approach to the public interest issues at stake. To understand the latter, it is necessary to consider the sorts of cases being pursued, and how these map onto wider policy concerns. In short, the consistent theme is the extent to which antitrust enforcement has aligned with, and reinforced, the regulatory process of market liberalization. One sees this perhaps most obviously in the recurrent use of Article 9 case dispositions—which empower the Commission to accept behavioral or structural commitments, such as divestiture remedies, in lieu of findings of breach—in the energy sector, in a manner which arguably serves to compensate for the Commission’s comparative lack of legislative success here. ¹²¹ From a substantive antitrust perspective, of greatest interest is the extent to which regulatory concerns have fed into two of the most prominent theories of harm applied in these sectors, namely, refusal to deal and margin squeeze. The frequent application of these theories is unsurprising, given that the sectors under scrutiny involve vertically integrated network utilities with natural monopoly segments. What is more remarkable, however, has been the Commission’s approach to the interpretation and development of the case theories in this context.

Refusal to deal comprises the denial of access by dominant undertakings to strategically important market segments, in order to exclude competition in adjacent sectors. Under Article 102, claimed refusals are assessed under the Bronner criteria, which require a demonstration of “exceptional circumstances” to establish abuse. ¹²² What we see in the liberalization context, in particular, is a willingness to map the antitrust interpretation of exceptionality onto the parameters of the underlying sector-specific regulatory regime—which, as argued above, reflects an essentially ideological understanding of the optimal operation of such markets.

In Telekomunikacja Polska, for instance, the impugned refusal consisted of failure to comply effectively with access obligations under the EU framework for telecommunications liberalization. Although the conduct had been sanctioned by the domestic regulator, the Commission ¹²³ and the Court ¹²⁴ had little difficulty in construing the regulatory violation, additionally, as an antitrust offense. In ENI, the alleged refusal included failure to anticipate likely future requests for access and expand capacity accordingly, behavior that was conceptualized as “strategic underinvestment” in the defendant’s gas transport infrastructure. ¹²⁵ Such an approach similarly has a distinct regulatory flavor, suggesting that the special responsibility of dominant firms extends to forecasting and promoting future market development. More recently, in Slovak Telekom, ¹²⁶ the General Court confirmed one of the contentious “Telefonica exceptions,” ¹²⁷ to the effect that regulatory supply duties may function a proxy for indispensability, thus negating the need to consider nonduplicability from an antitrust perspective. ¹²⁸ Moreover, the Court “suggested” that Bronner is irrelevant in situations of constructive refusal. ¹²⁹ Such an approach makes little sense when measured against the cogent competition policy concerns that motivated the regulatory restraint implicit in Bronner ¹³⁰ and is at odds with that advocated in the Commission’s own Article 102 guidelines. ¹³¹ Yet it was vigorously advanced by the Commission in its earlier infringement decision, and defended by it on appeal. Each of these cases thus ably illustrates how the Commission’s enthusiasm for deploying antitrust to “correct” the process of liberalization at Member State level may prompt an expansion of the relevant competition provisions to account for the particularities of the underlying regulatory framework—creating, however, a risk of distortive effects for the competition rules longer term. ¹³²

Margin squeeze is the second theory of recurrent application. ¹³³ Given its conceptual proximity to the refusal to deal, this may be expected, and indeed the Commission’s Article 102 guidance assessed margin squeeze as a form of constructive refusal. ¹³⁴ Starting with Deutsche Telekom, ¹³⁵ however, the Commission has treated margin squeeze as a distinct category of abusive behavior, an approach subsequently granted the imprimatur of the Court. ¹³⁶ A logical but unsensible consequence, confirmed in the reference case of TeliaSonera, is that the Bronner criteria do not determine the existence of a margin squeeze. ¹³⁷ Conceivably, a dominant firm may have an absolute right to deny access to its infrastructure under Article 102; but if it chooses to deal, it becomes subject to a positive duty to ensure a “fair” spread between wholesale and retail prices. Notably, this unsatisfactory state of affairs was actively advocated by the Commission in its intervention in TeliaSonera, premised on the (patently spurious) argument that, otherwise, any application of Article 102 would be contingent upon identifying an indispensable input. ¹³⁸ This pared-down legal standard was then applied by the Commission in Slovak Telekom, where the muddled reasoning of TeliaSonera was again deployed to support the “suggestion” that constructive refusals escape the established Bronner conditions. ¹³⁹

What both strands of case law demonstrate is that, although the Commission may be happy in principle to tie its hands with more demanding assessment criteria in line with the “more economic approach,” when it comes to actually applying competition law to address anticompetitive behavior in liberalizing markets, it is almost indecently eager to escape these shackles. Moreover, the thrust of the Commission’s approach is that the competition rules coincide precisely with requirements under the liberalization framework. Yet where the Commission disagrees with how the national regulator has interpreted its task of liberalization, defendants receive little credit for having complied with the directions of the latter, as Deutsche Telekom illustrated. ¹⁴⁰ The most obvious takeaway is the extent to which (perceived) public interest considerations may influence both the intensity and the ambition of antitrust enforcement. Yet in the Commission’s eagerness to police the task of liberalization through competition law, there is little consideration of whether the rules emerging represent best practice for the latter discipline as a whole—and in particular, whether these interpretations continue to make sense once removed from the protection of the Commission, and applied by a diverse range of public and private enforcers within the increasingly decentralized EU system.

B. Competition Law and the Pharmaceutical Sector

The second area of concentrated enforcement is the pharmaceutical sector. ¹⁴¹ Multiple competing public interests loom large here, summarized by former Competition Commissioner Kroes as impacting both “the health and finances of Europe’s citizens and governments.” ¹⁴² Securing access to necessary medicine for patients has an inescapable public health dimension. Yet in Europe, where the final payer is often the State, access may be less a question of availability and more of cost to straitened public purses. Merely pushing down prices cannot be the sole objective, however. Not only is the pharmaceutical sector highly dependent on costly R&D efforts to generate a pipeline of innovative new products that deliver improvements for patient health. Moreover, the health of the pharmaceutical sector itself raises industrial policy considerations in light of its key role within the EU economy. Thus, there are few obvious answers to competition problems in these markets, as the Commission’s Pharmaceutical Sector Inquiry, concluded in 2009, acknowledged. ¹⁴³

The preceding subsection described how public interest considerations underpinned a markedly expansive approach to an established theory of harm (a refusal to deal) in liberalizing market segments. In the pharmaceutical sector, by contrast, the public interests at stake have prompted significant legal innovation, resulting in the recognition of more novel theories that capture particular types of strategic behavior. This subsection considers two: the treatment of “pay-for-delay” agreements and “regulatory gaming” practices. Notably, both police the switch from patent protection to generic competition, and thus align with the balancing of multiple competing public interests in pharmaceutical cases: the end of patent protection marks the end of the period during which originator companies can plausibly claim a legitimate “monopoly reward” for successful innovation, while the introduction of generic competition usually generates significant downward pricing pressure.

The term “pay-for-delay” refers to a broad category of agreements pursuant to which generic companies consent to delay or even abandon market entry in exchange for payment from the incumbent originator. ¹⁴⁴ Such agreements tend to emerge shortly before successful proprietary medicines come off patent, as originators seek to extend de facto the duration of patent protection—and attendant inflated profits. In Lundbeck, decided in 2013, the Commission held for the first time that “pay-for-delay” agreements may violate Article 101. ¹⁴⁵ The anticompetitive potential of such arrangements is relatively uncontroversial, ¹⁴⁶ and, that same year, the U.S. Supreme Court similarly agreed that “pay-for-delay” may violate Sherman Act §1, following a “rule of reason” assessment. ¹⁴⁷ What was more unusual about Lundbeck was that the Commission had proceeded on the basis that relevant restraints comprised object restrictions under Article 101. Yet the “by object” category is, in principle, reserved for only the most obviously, inherently harmful forms of coordination. ¹⁴⁸ Such a characterization is arguably less clear-cut in the context of “pay-to-delay,” where excluded rivals constitute merely potential competitors at the time of the agreement, where successful entry even absent the arrangement is not a given, where there are (sometimes dubious) intellectual property rights at stake, and where condemnation of the agreement may conflict with the right to effective judicial protection.

Yet the approach in Lundbeck is easier to support to the extent that the relevant “legal and economic context” of the restraint is read broadly to encompass wider public interest dimensions. This certainly seems to have been the approach of the General Court in Lundbeck in upholding the object characterization, ¹⁴⁹ in which it acknowledged and defended the role of generic entry in prompting often-dramatic regulated price cuts in the pharmaceutical sector. ¹⁵⁰ It also expressed skepticism about the robustness of certain domestic patents, primly repeating that “it is in the public interest to eliminate any obstacle to economic activity which may arise where a patent was granted in error.” ¹⁵¹ Accordingly, contra the argument that Lundbeck is incompatible with the restrictive understanding of the object category endorsed in Cartes Bancaires, ¹⁵² the case can instead be seen as a nuanced exposition of the latter approach. Thus, even if, strictly speaking, Article 101 cannot police the protection of human health, the latter may shift our perception of the underlying legal context or render the anticipated economic harm more compelling. While Lundbeck is on further appeal, ¹⁵³ the object treatment of “pay-for-delay” has been confirmed by the General Court in Servier, ¹⁵⁴ in which the Commission notably hedged its bets by treating the relevant restraints as restrictions both by object and effect. ¹⁵⁵

The second area of legal innovation involves so-called regulatory gaming. This refers to the exploitation of pro-competitive or neutral governmental regulations, which are misused by private undertakings for exclusionary purposes. ¹⁵⁶ The challenge from an antitrust perspective is that, frequently, the defendant’s behavior complies with the letter of the underlying regime, though not its spirit. Thus, regulatory gaming raises questions of the extent to which competition law should second-guess the approach of other regulatory frameworks or attempt to correct putatively “broken” regimes.

The most prominent example in EU law arose precisely in the pharmaceutical sector, where AstraZeneca was held to have violated Article 102 through conduct intended to prolong the duration of patent protection for its blockbuster ulcer medicine, Losec. ¹⁵⁷ The abusive behavior comprised two prongs. First, the defendant made misrepresentations to national patent authorities, which enabled it to obtain “supplementary protection certificates” (SPCs) which extended patent life. Second, it strategically withdrew marketing authorization for Losec capsules, thus denying generic entrants a piggyback licensing procedure; and in tandem sought to migrate existing patient users to tablet form, which retained patent protection. The difficulty was that such behavior was either permitted by the underlying regulatory regime ¹⁵⁸ or implicitly facilitated. ¹⁵⁹ Thus, the putative actus reus in each instance was conduct that represented “normal” business behavior within the market concerned.

What is notable about AstraZeneca, therefore, and where the underlying public interest considerations came to the fore, is the extent to which the existence of the abuse hinged on the defendant’s anticompetitive intention (or “mens rea”). ¹⁶⁰ As explained in Section III, in principle, the antitrust rules depend on an objective determination of the impact on competition, actual or anticipated, of the relevant conduct. Yet although AstraZeneca reaffirmed the orthodoxy that anticompetitive intent is not a necessary requirement under Article 102, here the defendant’s perceived objectives proved determinative of its abusive behavior. So, for instance, the fact that AstraZeneca exercised its otherwise lawful right to withdraw its marketing authorizations precisely to frustrate generic competition served to transform outwardly ordinary and permissible behavior into an antitrust violation. Yet every firm wishes ultimately to exclude competitors, in the sense of increasing its own market share, hence the limited utility of mens rea-esque concepts in the antitrust sphere. What thus distinguished AstraZeneca, arguably, was the wider context and more profound ramifications of excluding generic competition: in particular, the harmful consequences for national health systems, a concern that was considered in detail in the Commission infringement decision. ¹⁶¹ Much as in the context of “pay-for-delay,” the underlying public interest considerations did not dispense with the need for antitrust assessment, but rather were brought to bear in shaping and justifying an ambitious application of the competition rules.

Yet the pharmaceutical sector also provides an important reminder of the continuing need to root even a purposive interpretation of the competition rules within the existing legal framework. This is seen in the fate of Servier, the second decision in which the Commission condemned regulatory gaming. ¹⁶² While, as noted, the General Court confirmed the Article 101 aspects, the Article 102 claim fell at the first hurdle, as the Court was unpersuaded of the robustness of the Commission’s market definition exercise. Absent dominance, Servier’s putatively anticompetitive unilateral behavior was immune from scrutiny, regardless of the public interests at stake. Thus, the soft influence of the public interest has unavoidable limits in the antitrust sphere.

C. Competition Law and the Digital Economy

The third and final example is a more recent focus on competition problems within the digital economy. The inescapable rise of Big Tech, coupled with the comparative absence of effective regulation within the digital sphere, has led to urgent calls (often from outside the antitrust community) for more aggressive enforcement against the digital giants. ¹⁶³ In the EU, these challenges are inextricably tied up with the “fairness mantra” ¹⁶⁴ associated with the tenure of Commissioner Vestager at DG Competition. In embracing the argument that competition law and policy are directed toward securing “a Europe that gives everyone a fair chance,” ¹⁶⁵ Commissioner Vestager has not only aligned this area of EU law with wider policy concerns about the optimal future development of the social market economy. ¹⁶⁶ Moreover, the sheer breadth of the fairness concept—including, inter alia, the right of “all companies—large and small—[to] a fair, fighting chance to succeed on their merits,” ¹⁶⁷ “fair pay” ¹⁶⁸ for workers, and a duty on undertakings to pay “their fair share of tax” ¹⁶⁹ —inevitably calls into question the Commission’s commitment to a “more economic approach,” with its ostensible single-minded focus on behavior that results in anticompetitive foreclosure. ¹⁷⁰

What is interesting about the themes highlighted by Commissioner Vestager is how these transcend even a “big is bad” understanding of competition policy. Instead, she has drawn upon a broad range of public interest concerns associated with the emergence of hugely powerful, increasingly ubiquitous, and apparently unregulatable digital platforms. These include data privacy, ¹⁷¹ the increasing pervasiveness of algorithmic decision-making, ¹⁷² the problems of adequately taxing digital services, ¹⁷³ the labor rights of vulnerable workers who provide services through digital platforms, ¹⁷⁴ and, ultimately, the threats to the fabric of our democracy that may be posed by the shift from public to private control of large areas of the economy. ¹⁷⁵ The fact that such themes plainly go beyond her remit as Competition Commissioner received tacit acknowledgment upon her re-nomination for a second term in September 2019, when Commissioner Vestager was given the additional mandate of securing “a Europe fit for the digital age.” ¹⁷⁶ Yet the rhetoric of fairness, even at its most ambitious, is by no means decoupled from the bread and butter of her antitrust enforcer role. Commissioner Vestager has explained the logic as such:

[I]f we value an open economy, and a liberal society, we have to show that those values benefit everyone, not just the select few[.] So if we want to show that our society treats everyone fairly, however big or small, we need to prove it in the market. ¹⁷⁷

It is in recent enforcement efforts within the digital economy that, arguably, we see these ideas play out most clearly.

The Commission, of course, has a rich history of robust enforcement within the “new” economy, including infringement decisions against Microsoft in 2004 ¹⁷⁸ and Intel in 2009. ¹⁷⁹ Both involved the application of well-established rules to more novel market situations—tying and refusal to deal in Microsoft, exclusive dealing in Intel—and both culminated in what were, at the time, record-breaking fines for violations of Article 102. These cases may thus be viewed as “setting the scene” for the trio of infringement decisions pursued against Google—Shopping, ¹⁸⁰ Android, ¹⁸¹ and AdSense ¹⁸² —since 2017. Yet these later cases mark a more muscular development of the antitrust rules in several ways, which, we suggest, is inspired in part by a perceived need to address broader public interest concerns.

First, the most immediately striking feature is the size of the fines imposed: more than €8 billion in total across the three Google decisions, including a record-breaking fine of €4.34 billion for Android alone. ¹⁸³ It is difficult to avoid the impression that these very high fines are a deliberate policy choice, set at such an elevated level precisely to demonstrate the willingness of the Commission to intervene against market actors perceived to pose a threat to the public interest. ¹⁸⁴ Such high fines were, moreover, enabled by the policy choice to pursue the cases as Article 7 infringement decisions rather than Article 9 commitments. ¹⁸⁵ This was despite overtures toward settlement in Android, ¹⁸⁶ and more remarkably, despite three rounds of ultimately unsuccessful market-testing of proposed commitments in Shopping (which, however, preceded Commissioner Vestager’s arrival at DG Competition). ¹⁸⁷ A similar “public interest” dimension is discernible in the Commission’s 2016 determination that Apple was required to repay €13 billion in uncollected taxes to Ireland for breach of the State aid rules, ¹⁸⁸ a decision the Commissioner admitted was motivated by a perceived need to “do something” within the limited powers available to the Commission to address public anger over tax injustice. ¹⁸⁹ Even if the subject matter of these decisions is often far removed from the focus of public anxiety regarding the digital economy, such cases nonetheless perform an important demonstrative function, setting the EU up as “the determinator,” ¹⁹⁰ prepared to rein in Big Tech in the public interest.

The influence of the public interest is also evident in the emergence of self-preferencing as the most prominent theory of harm in the digital sphere. Self-preferencing refers to situations whereby a digital platform gives “preferential treatment to one’s own products or services when they are in competition with products and services provided by other entities using the platform.” ¹⁹¹ It was the crux of the prohibited behavior in Shopping, can arguably be seen as the underlying concern in Android, ¹⁹² and is the subject of on-going investigations involving Amazon Marketplace ¹⁹³ and Apple iTunes. ¹⁹⁴ While in Shopping, the Commission disclaimed any novelty for self-preferencing as a theory of harm, ¹⁹⁵ the case law cited to support its established status fell considerably short. ¹⁹⁶ A cynic might also observe that novelty would have been an inconvenient obstacle insofar as the Commission was determined to impose a (then) record-breaking fine. ¹⁹⁷ In their report on Competition Policy for the Digital Era, the Commission’s expert panel on digital markets went even further, arguing for a reverse burden of proof in certain instances whereby the platform would “bear the burden of proving that self-preferencing has no long-run exclusionary effects on product markets.” ¹⁹⁸

Yet self-preferencing is of interest beyond the mere fact that its novelty demonstrates legal innovation in response to broader public concerns about the digital economy. The theory has a distinct regulatory quality, which chimes with calls for the aggressive application of competition law to provide an ad hoc tool for regulating the digital economy. The Commission’s expert panel thus identified a category of platforms that perform “a rule-setting function,” whereby platforms “act as regulators, setting up the rules and institutions through which their users interact.” ¹⁹⁹ In such circumstances, the authors proposed an overarching duty on dominant platforms to ensure that “competition on their platforms is fair, unbiased, and pro-users,” ²⁰⁰ a specific application of which is the reverse burden of proof for self-preferencing set out above. ²⁰¹ The notion that dominant platforms have a positive duty to ensure “fair” outcomes for rivals, through for instance modifying the manner in which they display competing services, has inescapable parallels to more traditional forms of utilities regulation. Yet whereas the latter has typically been contingent upon the existence of bottleneck monopolies that render effective competition otherwise impossible, both the Commission and its expert panel expressly disclaimed any necessary connection between the theory of self-preferencing and the case law on refusal to supply. ²⁰² The focus instead on “strategic market status” ²⁰³ thus aligns with broader academic concerns about the role that digital platforms play as gatekeepers within contemporary markets, and indeed contemporary society more broadly. ²⁰⁴ Similarly, the suggestion to relax the conventional antitrust concern with error cost by erring on the side of prohibiting potentially anticompetitive conduct in concentrated digital markets, ²⁰⁵ is consistent with broader public skepticism about digital market power.

Accordingly, what we see in the digital economy is how broader public interest concerns feed into discussion of the optimal approach to competition policy, with, moreover, some evidence from recent enforcement practice of the redirection of the latter to provide a more muscular tool to regulate digital dominance. There are, once again, inherent limits to such an approach, typified by the “serious doubts” expressed by the Dusseldorf Higher Regional Court ²⁰⁶ when considering the Bundeskartellamt’s much-debated decision to sanction alleged data protection breaches by Facebook as a violation of the German rules on abuse of market conduct. ²⁰⁷ That is, competition law remains competition law, and thus provides an inapt tool to regulate, for instance, fake news or discriminatory algorithmic decision-making, not merely from a legitimacy but also an effectiveness perspective. But it is indisputable that the arguments being raised transcend purely technical questions of antitrust policy—and it is increasingly difficult to maintain that such concerns are not reflected in at least the tenor of policy debate with respect to the digital economy, if not also the application of the antitrust rules in practice.

V. Conclusions

Debates regarding the treatment of public interest under the EU competition rules are caught between several contrasting impetuses. Competition law is focused on ensuring effective market competition, and in the EU, protecting “competition as such.” Yet most people would accept that, in the final analysis, markets should serve the wider public interest and not vice versa. It is, accordingly, difficult to maintain that competition law should be applied and developed with its eyes closed to noneconomic public interest concerns. Yet there is often an acute difficulty, at least in the EU, in identifying a single set of legitimate or approved public interests that merit accommodation within the competition framework. Moreover, while this article has focused primarily on the “centralized” application of the competition rules by the Commission, within an increasingly decentralized enforcement landscape, such public interest claims may not only take on greater urgency but also have greater disruptive or distortive potential.

This article has argued that it may be inferred from recent enforcement in three distinct segments that the considerations at issue transcended an inflexible concern with “competition as such.” The examples surveyed, however, do not tell a clear and consistent story about when and how public interest concerns can and should be accommodated. In the context of the network industries, for example, the Commission has used antitrust enforcement to pursue a rather singular vision of how such markets should be structured and ought to operate. Yet the available empirical evidence suggests that not all consumers are likely to benefit ²⁰⁸ and that the advisability of unbundling policies may vary significantly between different national markets. ²⁰⁹ In the pharmaceutical sector, it is difficult to avoid the suspicion that the root cause of the competition problems discussed above is, in many instances, the poor practices of national patent offices and other domestic regulatory agencies. Thus, competition enforcement, instead, became a kind of “proxy war” in such cases, fought to tackle market problems outside the purview of the Commission’s regulatory jurisdiction. An inescapable paradox of calls for greater intervention in the digital economy, moreover, is the underlying assumption that the public has a set of concerns about firm size which are not expressed in its market behavior: ²¹⁰ that is, if we the public are so concerned about the omnipresence of the tech giants, then why do we keep using these services when other options are available?

At its core, EU competition law reflects a persistent, though by no means uncontentious, belief that open and competitive markets, integrated across the Member States, are the best means to further public interest objectives more broadly. But that does not mean, however, that EU competition law is immune to wider public interest concerns. Noncompetition considerations can affect our understanding of how markets work or shape our approach to the derogations that exist within the competition system. Moreover, and perhaps most importantly in practice, the fact that compelling parallel public interests may be at stake can influence the underlying calculus of whether antitrust intervention is merited in a particular instance and how vigorously it should be pursued.

Yet competition law is not a particularly nimble tool by which to pursue such interests more directly as a standalone concern. Competition law can break up cartels that seek to delay green innovation, for instance, or punish dominant drug makers that block access to cheap generic versions of blockbuster medicines. But, conversely, it is an inapt mechanism by which to require firms to invest in sustainability, and there are cogent policy reasons why antitrust is generally reluctant to force generic-level pricing on originator pharma companies during the period of patent protection. Thus, competition law can and should be used to further the public interest, broadly construed, where the latter coincides with its overarching task of promoting beneficial social outcomes through the use of open and effective competition to control private power. But it is not a panacea for all contemporary ills, nor a lazy fallback that can be assumed to absolve other legal disciplines of the need to figure out how best to regulate problems that fall more squarely within their jurisdiction.

Competition law and the public interest go hand in hand, as the quote that opened this article recognizes. But competition law alone cannot be enough to protect the range of interests that have been invoked in this context. There are, famously, “limits of antitrust,” and the extent to which these rules can be deployed in furtherance of noncompetition-oriented yet socially valuable objectives is undoubtedly one such limit. Claims to the public interest are clearly cognizable within the EU competition framework, albeit typically through an approach of carving out such concerns from the application of the rules rather than accounting for noncompetition interests within the substance of the antitrust provisions. Whether the latter would be a good idea, or rather a slippery slope, is a provocative question largely beyond the scope of this article. What is clear, however, is that competition law can be part of the solution when it comes to addressing broader societal problems that extend beyond competition policy as such, but it would be naive to think that it is the only or optimal solution here.