In recent times, self-preferencing by digital platforms has come to the fore. While it is normal for companies to promote their own products over those of their competitors, and, in many cases, this can lead to efficiencies, a dominant company giving preferential treatment to its own products or services in downstream or related markets (self-preferencing) can raise concerns. Given its potential for pro-competitive effects, it remains challenging to identify the exact circumstances in which self-preferencing is anticompetitive. In addition, a number of the concerns raised by self-preferencing may not relate solely to their potential anticompetitive effects. This means that different tools – including, but going beyond competition law – may be used to address problematic self-preferencing practices.
By Pedro Caro de Sousa1
I. INTRODUCTION
It is normal for companies to promote their own products over those of their competitors, and, in many cases, this can lead to efficiencies. However, a dominant company giving preferential treatment to its own products or services in downstream or related markets (self-preferencing) can raise concerns. Despite its ingrained flexibility, competition law is limited to addressing behavior that proves anticompetitive in specific circumstances, and is unable to address many other concerns that self-preferencing may give rise to. Even when self-preferencing is anticompetitive, it is important to evaluate whether a regulatory reaction would be preferable, e.g. because it is faster, more effective and can foster legal certainty.
In recent times, self-preferencing by digital platforms has come to the fore. This note will address the question of how best to address self-preferencing in this sector.
The next section will look at the difficulties of determining when self-preferencing conduct is anticompetitive. Given that competition law is loath to impose a generic duty not to favor one’s own products in the digital sphere, the main challenge concerns the identification of limiting principles that can provide guidance as to when self-preferencing is anticompetitive. The difficulties of such an exercise are apparent in the reactions to the Google Shopping decision (the “Decision”) by the European Commission (“Commission”).2
A second section will look at other avenues to address problematic self-preferencing conduct, by reviewing suggestions on how to deal with such practices contained in studies on competition law and the digital economy commissioned by competition authorities and governments. These studies display widespread unease with simply relying on competition enforcement to deal with self-preferencing, and offer a plethora of alternatives – some suggesting changes to the competition regime, others advancing openly regulatory options, many threading a middle path. The direction of travel seems to be towards paying increasing attention to self-preferencing in the digital realm, but with different mechanisms being adopted around the world.
II. THE LIMITS OF ANTICOMPETITIVE SELF-PREFERENCING
A. The Google Decision
It is uncontroversial under EU law that abuses can be committed in a market other than where the infringing company is dominant.3 Acts whereby a company dominant in one market uses its market power to commit an abuse in another market typically fall under the wide umbrella of “leveraging” practices.
In Google Shopping, the Commission sanctioned Google for self-preferencing – i.e. granting more favorable positioning and display to Google’s comparison shopping service in its general search results pages when compared to competing comparison-shopping services.4 This led to decreased traffic for competing shopping services, while increasing traffic for Google’s own service.5 In the Commission’s view, the importance of traffic volume for comparison shopping services,6 and the fact that Google Search results accounted for a large proportion of their traffic and could not be effectively replaced by other sources,7 meant that these effects were anticompetitive. This was particularly so because the conduct made it more difficult for competing comparison shopping services to reach a critical mass of users that would allow them to compete against Google, which therefore foreclosed competitors in the market for comparison shopping services.8
The Decision also held that the Commission was under no duty to prove that anticompetitive effects occurred.9 Instead, it sufficed to demonstrate that the conduct was merely capable of having, or likely to have, foreclosure effects,10 regardless of its success in practice.11 In other words, the Commission merely had to demonstrate that the conduct created a possibility that Google’s competitors might cease to compete or would have reduced incentives to innovate.12 This occurs not only where market access is made impossible for competitors, but also where the conduct of the dominant undertaking is capable of making that access more difficult, thus interfering with the structure of competition on the market.13
B. The Debate
Following the adoption of the Decision, a debate sprung up about the exact circumstances under which self-preferencing is anticompetitive. Some argued that, in practice, the Decision imposed a duty on digital platforms not to discriminate in favor of its own products. Others argued that the Decision was strictly limited to the facts, which in this case led to anticompetitive effects. In between these extremes, many authors elaborated theories of harm that sought to explain why Google’s conduct might be abusive.
- A General Duty not to Self-Favor?
The Decision has been criticized for opening the door to a quasi-regulatory duty not to favor one’s own products. These criticisms build on claims that the Decision is based on “traditional theories of unlawful leveraging.”14 However, “leveraging abuse” is an umbrella term covering different types of unilateral practices that foreclose competition. The case law typically identifies specific features that distinguish the conduct at issue from competition on the merits, such as deteriorations in quality, margin squeezing, or a refusal to supply an indispensable input. Importantly, leveraging can produce a number of pro-competitive effects, flowing from efficiencies related to the one-monopoly-profit theorem, technological interdependence, protecting goodwill and reputation, or economies of joint production or sale.15 Absent the identification of specific features that make the conduct anticompetitive, it becomes very hard to distinguish between lawful and unlawful conduct, particularly since self-preferencing is a normally accepted business practice.16
Focusing solely on whether the practice amounts to self-preferencing leveraging could put any dominant company on notice about the antitrust risks of adopting a commercial strategy favoring its own goods and services. It has been said that this does not fit well with traditional categories of exclusionary abuse, in particular because it would amount, in practice, to a general duty on dominant companies not to favor their own goods and services. However, competition law remedies must be proportionate to the harm and address the anticompetitive effects of the practice under analysis.17 As such, the imposition of such a general duty, absent a finding that self-preferencing is always anticompetitive, is hard to square with competition law principles. Authors who read the Decision in this way often conclude that it is more akin to sectoral regulation than to a competition intervention.18
- A Pure Effects Analysis?
At the other extreme, the Decision has been defended as having correctly focused, as is often the case in unilateral conduct cases, on whether the individual (leveraging) practice foreclosed competition in this specific instance.19 However, while such an argument may succeed at framing the Decision within a traditional competition framework, it does so at the cost of raising a number of conceptual and practical issues.
Conceptually, and since foreclosure does not require the absolute exclusion of all competitors from the market, such an approach gives rise to threshold questions regarding the point at which business practices become unlawful. When does a reduction in the competitors’ ability to compete become serious enough to be anticompetitive? Does it suffice for conduct by a dominant company to make it harder for competitors to compete? Or must the conduct prevent competitors from competing effectively, or even excludes them from the market in practice?20 These are important questions that underpin wider debates about how to identify anticompetitive unilateral practices, e.g. the need to deploy the as efficient test in Article 102 TFEU cases.
A related practical implication of this approach is that focusing solely on foreclosure equates with pursuing a pure effects approach. The problem with this – and the reason why pure effects approaches are actually quite rare in practice, despite the rhetoric surrounding them – is that it becomes extremely hard to derive decision rules from an analysis that merely focuses on whether a discrete practice had foreclosure effects in specific circumstances. Such an approach requires a pure in-depth effects test in every case. A logical inference of such a minimalist interpretation would be that, while the Decision confirms that self-preferencing can be anticompetitive in some cases, it does not attempt to provide guidance about when this might be the case – which is particularly striking given the widespread alarm about how common self-preferencing by digital companies has become. This interpretation of the Decision has even led some authors to go as far as to hold that the Decision did not reveal a legal theory of harm explaining under which circumstances self-preferencing is abusive.21
- Just Another Theory of Harm?
In between these extremes, many authors proposed narrower theories of harm to frame the Decision – and identify its limiting principles. Some proposals were even adduced by interveners supporting the Decision before the General Court. For example, the German government argued that Google’s practice was misleading, and thus did not amount to competition on the merits and artificially prevented competition based on the quality of the algorithm used to carry out specialized shopping searches.22 Another intervener postulated that the problem with Google’s conduct was that it had no economic rationale other than to foreclose competition on the secondary market.23
Many other explanations have been suggested in academic writings. A first set of theories of harm focuses on refusals to supply and the “essential facility” doctrine. In the past, the European Court of Justice has identified a narrow set of circumstances under which the essential facility doctrine may be applicable, namely when the refusal to grant access to a facility or input is likely to prevent any competition in a downstream market, access is indispensable, and access is denied without any objective justification.24 Applied in the context of online search, this would require proof that: (a) achieving certain ranking results is objectively necessary in order for businesses to compete effectively; (b) Google’s strategy of favoring its own services through visual prominence was likely to eliminate effective competition in a downstream search market; (c) the refusal to treat all results without discriminating between them led to consumer harm.25 These suggestions have been criticized on the grounds that the Decision’s reasoning does not rely on these doctrines, and the Commission has repeatedly argued that this was not the theory of harm underpinning the Decision.26
Another suggestion is that the Decision adopts an ad hoc theory of harm concerning constructive refusal to feature rival services in Google’s search engine. This would be akin to a Commercial Solvents-type refusal to continue a course of dealing, which does not require that the facility be indispensable.27 This latter suggestion links to a second type of proposed theories of harm: unlawful discrimination.
A number of authors have argued that the Decision pursues a theory of discriminatory leveraging. Some argue that it is unlawful for an essential digital infrastructure to “leverage” its dominance into an adjacent segment by discriminating in favor of its own subsidiary products against other trading partners.28 Others suggest that the Decision reflects a replacement of the indispensability requirement for a heightened, and platform-specific, “special responsibility” not to discriminate in favor of its own products. This platform-specific duty is linked not only to the undertaking’s position as a dominant undertaking, but also to its “gatekeeper” role within the wider digital ecosystem.29
Other discrimination-based analyses move away from essentiality, refusal to supply and their analogues, and focus squarely on the Treaty’s prohibition on applying “dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage.”30 This prohibition has been predominantly applied in settings where vertically-integrated dominant firms seek to advantage their downstream operations at the expense of rivals. Several cases display a clear theory of abusive self-preferencing – even if thus far they were restricted to incumbents in liberalized sectors.31 There are also cases that could be said to identify discriminatory self-favoring with anticompetitive effects – even if those cases mostly rely on traditional bases for identifying exclusionary conduct, such as the existence of abusive exclusivity arrangements.32.
Approaches to self-preferencing based on discrimination have also attracted their share of concern. Attempting to evaluate discrimination in the results of a search page is said to be inherently fraught, given that a search algorithm is by definition a means to discriminate, rank and pick winners based on some selected metrics. Applying a purely discrimination-based approach to search pages has also been said to be akin to granting a very wide discretion to competition authorities in regulating self-preferencing in the digital realm.33 Such an approach raises concerns similar to those flowing from the creation of a general duty not to engage in self-preferencing by competition agency fiat, so it is unsurprising that some authors who interpreted the decision as adopting a discrimination-based theory of harm have also concluded that it implements a quasi-regulatory approach to self-preferencing discrimination by digital platforms.34
A third approach sees the Decision as a variation on more traditional exclusionary abuses such as margin squeeze or tying, which do not require the dominant undertaking to be an essential facility. Under margin squeeze, for example, the low ranking and consequent lack of attention given to a website may foreclose its access to consumers, increase its advertising costs and generally inflate its costs and prices, thereby reducing the competitive pressure on the platform’s own product or service. It has been suggested that the test under such a theory of harm should be whether a dominant platform could offer its own product and effectively compete for end-users if it had to pay the price that it charges as-efficient-rivals for prominence.35 As regards tying, it has been argued that Google is able to induce selection of its tied good (i.e. shopping services) merely by granting it visual prominence in its general search pages. If the effect of inducement is similar to coercion in such circumstances, and inducement can occur via the exploitation of status quo biases, Google’s conduct can be said to amount to tying leading to the foreclosure of national markets for comparison shopping services (and also for the general search market). This, in turn, has consequences such as a reduction of innovation, lower quality of search and reduced consumer access to relevant service.36
A variant of this approach suggests that Google’s self-preferencing amounts to unlawful platform envelopment. Under this novel exclusionary theory of harm, anticompetitive behavior can occur when the digital platform enters another platform market and combines its own functionality with those of companies in that market. The digital platform therefore creates a multi-platform bundle that leverages shared user relationships and/or common components. This can be anticompetitive in certain circumstances, including self-preferencing,37 particularly when a platform that is dominant in one market distorts the terms of trade in the target market by bending its rules in the origin market in favor of its own products or services in the target market. The main potential anticompetitive effects of such practices relate to market foreclosure of as-efficient companies operating only in the target market as a result of the anticompetitive leveraging of market power in the origin market.38
A common element to all these proposals is that they try to determine a set of circumstances in which self-preferencing will be unlawful under competition law. The sheer number of proposals indicates that there are many ways in which self-preferencing may prove to be anticompetitive. However, this raises a number of questions. First, are we only concerned with anticompetitive self-preferencing? If we are worried about other types of self-preferencing, how should we deal with them? Relatedly, given how widespread are concerns about digital self-preferencing, do we believe that antitrust enforcement is the best option to address such practices even when they are anticompetitive, or are there more effective intervention mechanisms that we should deploy?
III. NORMATIVE PROPOSALS IN AGENCY STUDIES INTO COMPETITION LAW IN THE DIGITAL AGE
The competition community has seen a flurry of studies on the digital economy, many published or commissioned by competition authorities themselves. A number of these studies discuss how best to address self-preferencing, and proposed solutions along a continuum from facilitating competition enforcement to adopting a purely regulatory approach. These solutions, most of which require legal changes, can be usefully categorized into three broad types.39
A. Making Competition Enforcement Easier
The most prominent example of a report suggesting the use of competition enforcement to address self-preferencing is the EU Experts Report, which proposes two main changes to make such – and other – competition enforcement more effective. First, within the existing framework, competition enforcement should recognize market power accruing to digital platforms that are “unavoidable trading partners” (also called “intermediation power”).40 The other change would be to reform the error-cost framework governing competition enforcement by making changes to the burden and standard of proof in the context of highly concentrated markets characterized by strong network effects and high barriers to entry. In particular, the Report suggests placing the burden of proof for showing pro-competitiveness on the incumbent in such circumstances.41 This has implications for enforcement against self-preferencing in the digital sphere. Competition law does not impose a general prohibition of self-preferencing on dominant firms. Self-preferencing must generally be shown to be anticompetitive under an effects test. However, the Report’s proposal mean that dominant digital platforms in markets with particularly high barriers to entry which serve as an intermediation infrastructure of particular relevance would bear the burden of proving that self-preferencing has no long-run exclusionary effects on product markets.42
Suggestions along these lines have been made elsewhere. The French competition authority has also expressed a preference for competition enforcement to deal with self-preferencing, while suggesting the adoption of measures that would make enforcement more effective in the digital sphere. Among the possibilities it listed to this end, the Autorité suggested that the essential facilities doctrine could be extended to markets where there are gatekeepers / where interoperability is important.43 The Autorité also suggested that competition authorities could be empowered to act specifically against “structuring” platforms, even when they are not dominant, particularly as regards certain “problematic” conduct, such as discriminating between services provided in the platform. A presumption of harm could also apply to such conduct.44 The BRICS report also suggested utilizing the principle of special responsibility for platforms to preserve competition by creating a level playing field for downstream and connected markets.45 It also suggests that many conducts, including failing to display similar business users on equal term in search results or on product/services comparison sites, could amount to per se abuses in certain circumstances.46
B. Quasi-Regulatory Competition Rules
This last report shows how small is the step from reversing the burden of proof to adopting quasi-regulatory provisions against self-preferencing. This is also apparent in a number of studies about how to adapt competition law to the digital age.
The Japanese Digital Platforms Reports suggests looking into whether competition law needs to be updated to reflect the fact that digital platforms are essential facilities to whom special duties should attach. Furthermore, the JFTC suggested that it would need to take a close look at a number of potentially problematic practices, including whether digital platform operators give themselves or their related companies preferential treatment by, for example, manipulating search algorithms. This last practice may be unlawful either as an exclusionary practice or as an unlawful interference in a competitor’s transactions.47
In Germany, an expert report proposed taking into account intermediary power when assessing relative dominance, and extending existing protections for small and medium enterprises (“SMEs”) against exclusionary practices by companies with significant market power to all companies.48 Another report suggested imposing some clear-cut prohibitions on dominant online platforms. This could be achieved through an EU Platform Regulation that both fleshes out and supplements competition law.49 This Regulation would apply to dominant online platforms with certain minimum revenues or users, and would specify rules of conduct inspired by competition law – including a prohibition on these platforms to favor their own services when compared to third-party providers, unless such preferencing is objectively justified.50 Draft amendments to Germany’s competition law similarly suggest empowering the Bundeskartellamt to prohibit undertakings with paramount significance for competition across markets (i.e. digital platforms) from treating competitors differently from their own services without objective justification.51
In the UK, the House of Lords Select Committee recommended imposing special obligations on online communications platforms that act as gatekeepers for the internet, to ensure that they act fairly.52 The Furman Report, prepared at the request of the Government, suggested subjecting companies with strategic market statuses53 to a code of competitive conduct. Such a code would regulate, for example, instances of platforms giving preferential treatment to their own upstream or downstream products and services, e.g. by means of a search function that gives their own services an unfair advantage over its rivals in downstream markets through the ranking or presentation of results. The code would require platform users to be provided with prominence, rankings and reviews on a fair, consistent, and transparent basis.54 The Furman Report further recommended the creation of a digital markets unit with an ongoing monitoring role and the power to enforce legally binding decisions and penalties for contraventions of the code55 – something that was accepted in principle by the Competition and Markets Authority (“CMA”).56
The UK Government has now set up a digital markets taskforce, housed in the CMA and headed by a senior CMA official, with the goal of advising the government on the practical application of measures set out in the Furman Report. This includes devising a potential methodology to designate digital platforms with “Strategic Market Status,” and advising on the form that a code of conduct to promote competition could take.57
The competition authorities of Belgium, Luxembourg and the Netherlands have suggested allowing competition authorities to impose non-punitive behavioral remedies against dominant digital platforms ex-ante, such as requiring the adoption of non-discriminatory rankings. These remedies would not be imposed solely in reaction to competition violations, even if they should follow and be closely inspired by competition law.58
C. Regulation
Many of the proposals above are clearly influenced by regulatory considerations, and reflect existing regulatory instruments that set out non-discrimination principles in areas touching the digital realm.59 However, these Reports shy away from openly calling for regulation or the creation of a specialized regulatory body. Instead, they suggest that these rules are natural extensions of competition law and that competition authorities can take over these competences. Occasionally, they also touch on the possibility of creating a specialized body in the future.
The main exception to this is the U.S. The Stigler Report not only recommended imposing special regulation on bottleneck digital companies, but also called for the setting up of a special regulator which would be responsible, inter alia, for imposing non-discrimination requirements and interoperability requirements on those companies.60 Similarly, the U.S. Congressional Research Service Report suggested the adoption of sectoral regulation prohibiting self-preferencing by digital platforms, or even prohibiting digital monopolists from entering adjacent markets.61
IV. CONCLUSIONS
Self-preferencing raises significant concerns in the digital sphere, particularly when pursued by large digital platforms. This note mapped out the ongoing debate about how best to address these concerns.
It is submitted that the analysis above revealed two main difficulties that these efforts face. First, given its potential for pro-competitive effects, it is challenging to identify the exact circumstances in which self-preferencing is anticompetitive. In addition, a number of the concerns raised by self-preferencing may not relate solely to their potential anticompetitive effects.
Second, it is still unclear how best to address problematic self-preferencing practices – even those that are anticompetitive. The studies discussed above make suggestions that go from merely trying to ensure that competition law is enforced in a timely manner, to prohibiting all self-preferencing by certain companies. A related question concerns which agency should enforce the rules that might be adopted as regards self-preferencing. These different suggestions reflect local traditions and concerns about the effectiveness of enforcement given local institutional constraints. It is also possible, and perhaps likely, that they reflect diversity in local objectives when dealing with digital self-preferencing which may go beyond the difference in goals of competition regimes around the world.62 This, however, is a matter for a different paper.
1 OECD – Competition Division (Paris). The present note expresses the author’s personal opinion. It does not reflect the position of the OECD, or any of its Members.
2 Case 39740 Google Search (Shopping).
3 Case C-333/94 P, Tetra Pak v. Commission, EU:C:1996:436, paragraph 25; Case C-52/09, Konkurrensverket v. TeliaSonera Sverige AB, EU:C:2011:83, paragraph 85.
4 Andrea Amelio, Thomas Buettner, Cyril Hariton, Gábor Koltay, Penelope Papandropoulos, Geza Sapi, Tommaso Vallett & Hans Zenger “Recent Developments at DG Competition: 2017/2018,” (2018) Review of Industrial Organization 1, p. 6; General Court, Report for the Hearing in Google Shopping, para. 59.
5 Id. para. 67
6 Id. para. 66.
7 Id. para. 68.
8 Id. paras.641- 642; Andrea Amelio, Thomas Buettner, Cyril Hariton, Gábor Koltay, Penelope Papandropoulos, Geza Sapi, Tommaso Vallett & Hans Zenger “Recent Developments at DG Competition: 2017/2018,” (2018) Review of Industrial Organization 1, p. 7.
9 Andres Caro “Leveraging market power online: the Google Shopping case,” (2019) Competition Law Journal 17(1) 49, p. 50.
10 Decision, para. 601, for the need to demonstrate anticompetitive foreclosure effects; and para. 606, where it is held that “the Commission is not required to prove that the Conduct has the actual effect of decreasing traffic to competing comparison shopping services and increasing traffic to Google’s comparison shopping service. Rather, it is sufficient for the Commission to demonstrate that the Conduct is capable of having, or likely to have, such effects.”
11 Case T-321/05, AstraZeneca, EU:T:2010:266, paragraph 347, confirmed on appeal in Case C-457/10 P, AstraZeneca v. Commission, EU:C:2012:770, paragraphs 109 and 111. See also Case T-286/09, Intel v. Commission, EU:T:2014:547, paragraph 186 (and case law cited therein).
12 Decision, paras. 594-595.
13 Decision, para.339, referring to Case C-52/09 Konkurrensverket v. TeliaSonera Sverige AB, EU:C:2011:83, paragraph 63; Case T-286/09 Intel v. Commission, EU:T:2014:547, paragraphs 88, 149 and 201.
14 Concurrences (2019, hors de serie) “The Google Shopping Decision,” 155, p. 155. According to this paper, at p. 158, the Decision invokes leveraging at para. 649.
15 Louis Kaplow “Extension of Monopoly Power through Leverage,” (1985) Colum. L. Rev. 85 515, p. 517-519.; Ward S. Bowman “Tying Arrangements and the Leverage Problem,” (1957-1958) Yale L.J. 67 19, p. 24-29; Michael D. Whinston “Tying, Foreclosure, and Exclusion,” (1990) Am. Econ. Rev. 80 837; Christian Ahlborn, David Evans, Jorge Padilla “The antitrust economics of tying: a farewell to per se illegality,” (2004) Antitrust Bulletin; Spring (49) 287. See also the European Commission’s Report on Competition Policy for the Digital Era, p. 6.
16 This concern is compounded by the fact that, as was acknowledged by the Commission’s own economists, “the Decision does not discuss directly Google’s incentive to foreclose.” See Andrea Amelio, Thomas Buettner, Cyril Hariton, Gábor Koltay, Penelope Papandropoulos, Geza Sapi, Tommaso Vallett and Hans Zenger “Recent Developments at DG Competition: 2017/2018,” (2018) Review of Industrial Organization 1, p. 10.
17 Article 7(1) of Regulation 1/2003. The difficulty in identifying a competition remedy that would work in the context of the Decision has been remarked upon by many – see German Report “A new competition framework for the digital economy,” p. 74.
18 See the literature mentioned in OECD (2018) Implications of E-commerce for Competition Policy DAF/COMP(2018)3, p. 38; and, more recently, Christian Bergqvist “Discrimination and Self-favoring in the Digital Economy,” arguing that the case implements a rule akin to Regulation 2019/1150 on promoting fairness and transparency for business users of online intermediation services.
19 Andrea Amelio, Thomas Buettner, Cyril Hariton, Gábor Koltay, Penelope Papandropoulos, Geza Sapi, Tommaso Vallett & Hans Zenger “Recent Developments at DG Competition: 2017/2018,” (2018) Review of Industrial Organization 1, p. 9, 12.
20 I believe that Niamh Dunne “Dispensing with Indispensability,” (2020) Journal of Competition Law & Economics 16(1) 74, p. 100, raises a similar point as regards the Decision in terms of uncertainty regarding the nature of the advantages required for the conduct to be anticompetitive.
21 Edward Iacobucci & Francesco Ducci “The Google Search Case in Europe: Tying and the Single Monopoly Profit Theorem in Two-Sided Markets,” (2018) European Journal of Law and Economics 47 15, p. 18. At p, 19, the authors note that “In contrast, the Commission has made explicit a theory of harm based on tying between Android devices and Google search and Chrome browser in the Android decision.”
22 General Court, Report for the Hearing, para. 323-324.
23 Id. para. 326.
24 Case C-7/97 Oscar Bronner GmbH & Co KG v. Mediaprint Zeitungs- und Zeitschriftenverlag GmbH & Co KG, ECLI:EU:C:1998:569, para 41. Bo Vesterdorf, “Theories of Self-Preferencing and Duty to Deal – Two Sides of the Same Coin,” (2015) Competition Law & Policy Debate 1(1) 4, argued that this was the only circumstance under which self-preferencing could infringe Art 102 TFEU.
25 Robert Bork, George Sidak “What does the Chicago school teach about internet search and the antitrust treatment of Google?” (2012) Journal of Competition Law and Economics 4 663; M. Lao, “Search, essential facilities and the antitrust duty to deal,” (2013) Northwestern Journal of Technology and Intellectual Property 11 275.
26 Decision, para. 651; General Court, Report for the Hearing, paras. 300, 303.
27 Pablo Ibáñez Colomo “Indispensability and Abuse of Dominance: From Commercial Solvents to Slovak Telekom and Google Shopping,” (2019) Journal of European Competition Law & Practice, 10(9) 532, p. 541, 548. In short, the author argues that refusal to supply cases such as Oscar Bronner merely involve a “passive refusal” to deal with a rival. Google Shopping is instead more like cases of active termination of a course of dealing, such as Commercial Solvents and CBEM-Télémarketing.
28 OECD (2018) Implications of E-commerce for Competition Policy DAF/COMP(2018)3, p. 37.
29 Niamh Dunne “Dispensing with Indispensability,” (2020) Journal of Competition Law & Economics 16(1) 74, 99-102.
30 Apparently adopting this approach, see Thomas Hoppner, Felicitas Schaper, and Philipp Westerhoff “Google Search (Shopping) as a Precedent for Disintermediation in Other Sectors – The Example of Google for Jobs,” (2018) Journal of European Competition Law & Practice 9(10) 627, p. 629-630, 638.
31 See Nicolas Petit “Theories of Self-Preferencing under Article 102 TFEU: A Reply To Bo Vesterdorf,” p. 3-4 – listing Case T-229/94 Deutsche Bahn AG v. Commission ECR [1997] II-1689; Case C-242/95 GT-Link A/S and De Danske Statsbaner (DSB) ECR [1997] ECRI–4449; Commission Decision COMP/39.388 German Wholesale Electricity Market (E.ON).
32 E.g. Case T-65/98 Van den Bergh Foods Ltd v. Commission [2003] ECR II-465. See Christian Bergqvist “Discrimination and Self-favoring in the Digital Economy,” and Nicolas Petit “Theories of Self-Preferencing under Article 102 TFEU: A Reply to Bo Vesterdorf,” p. 7.
33 Edward Iacobucci & Francesco Ducci “The Google Search Case in Europe: Tying and the Single Monopoly Profit Theorem in Two-Sided Markets,” (2018) European Journal of Law and Economics 47 15, p. 20-21.
34 See footnote 18 above.
35 Friso Bostoen “Online platforms and vertical integration: the return of margin squeeze?” (2018) Journal of Antitrust Enforcement 6 355, p. 374; OECD (2020) Line of Business Restrictions DAF/COMP/WP2(2020)1.
36 Edward Iacobucci & Francesco Ducci “The Google Search Case in Europe: Tying and the Single Monopoly Profit Theorem in Two-Sided Markets,” (2018) European Journal of Law and Economics 47 15. More generically, see Nicolas Petit “Theories of Self-Preferencing under Article 102 TFEU: A Reply to Bo Vesterdorf,” p. 5-7.
37 Other types of platform envelopment include bundling and virtual bundling.
38 Daniele Condorelli and Jorge Padilla on “Harnessing Platform Envelopment through Privacy Policy Tying”; OECD (2020) Conglomerate Effects in Mergers DAF/COMP(2020)2.
39 Some reports did not address the issue, or merely noted that self-preferencing can amount to an abusive practice: see Autoridade da Concorrência “Digital ecosystems, Big Data and Algorithms,” p. 69-70.
40 European Commission’s Report on Competition Policy for the Digital Era, p. 49.
41 Id. p. 51.
42 Id. p. 61.
43 Contribution de l’Autorité de la concurrence au débat sur la politique de concurrence et les enjeux numériques, p. 5.
44 Id. p. 7-9.
45 Digital Era Competition: A BRICs Review, p. 43, 548-549.
46 Id. p. 624-625.
47 Japan Fair Trade Commission “Digital Platforms Report,” Chapter 2.
48 Report on Modernising the law on abuse of market power (English Summary), paras. 3-4, 6.
49 Report “A new competition framework for the digital economy,” p. 49.
50 Id. p. 50-51.
51 Article 19A (2) of Germany’s Ministry of Economics draft for the 10th amendment to the German competition act.
52 House of Lords “Regulating in a digital world,” para. 171.
53 I.e. companies with enduring market power over strategic bottleneck markets.
54 Report of the Digital Competition Expert Panel “Unlocking Digital Competition,” p. 55-62.
55 Id. p. 62-63.
56 CMA’s Digital Markets Strategy, p. 11.
57 https://www.gov.uk/government/publications/digital-markets-taskforce-terms-of-reference/digital-markets-taskforce-terms-of-reference–3.
58 Joint Memorandum of the Belgian, Dutch and Luxembourg competition authorities on challenges faced by competition authorities in a digital work, p. 5-6.
59 In Europe, this includes Regulation (EU) 2015/2120 of the European Parliament and of the Council of 25 November 2015 laying down measures concerning open internet access, [2015] OJ L310/1 (Net Neutrality); the European Electronic Communications Code (EU) 2018/1972; and Regulation (EU) 2019/1150 of the European Parliament and of the Council of 20 June 2019 on promoting fairness and transparency for business users of online intermediation services OJ L 186.
60 Final Report of Stigler Committee on Digital Platforms, p. 32-33.
61 U.S. Congressional Research Service “Antitrust and Big Tech,” p. 35-36.
62 ICN (2007), Report on the Objectives of Unilateral Conduct Laws, Assessment of Dominance/Substantial Market Power, and State-Created Monopolies; OECD “Public Interest Considerations in Merger Control,” DAF/COMP/WP3(2016)3.
