Self-Preferencing – Some Observations on the Push for Legislation at the National Level in Germany

While the debate on self-preferencing as a possible abuse of dominance under EU law is ongoing, various expert reports on antitrust enforcement in the digital economy identified self-preferencing as a competition issue to be tackled and provided several suggestions, including legislation. Germany has proposed draft national legislation, including a special provision prohibiting self-preferencing for companies with a paramount cross-market significance for competition. The proposal follows a hybrid approach, introducing aspects of ex ante regulation into competition law. The design raises some issues, notably that it is not limited to digital platforms and does not require any dominance. Due to its two-step approach, it may only render enforcement more effective in the mid-term. It does not resolve the issues of lack of resources at the competition authority or the difficulties to impose effective remedies to terminate infringements. From an EU perspective, a patchwork of different national rules on enforcement against abusive conduct in the digital economy.

By Silke Heinz¹

 

I. INTRODUCTION

Self-preferencing is a prominent topic in the area of abuse of dominance in the digital economy. It typically means a (vertically integrated) dominant platform giving its own products and services preferential treatment over those of its rivals on the platform. The European Commission first applied this theory of harm in Google Search, finding that Google positioned and displayed more favourably, in its general search results pages, its own comparison shopping service compared to competing comparison shopping services.² The concept has also featured in other proceedings, notably in the ongoing Amazon proceedings at Commission level³ and to some extent in the Amazon proceedings in Germany.⁴

Google Search has triggered a controversial debate in the competition community on how, whether, and when it should be considered abusive.⁵ The debate is not over yet: After almost seven years of proceedings the Commission issued the decision in 2017. Google’s appeal with the European Court is still pending. The oral arguments took place in February 2020, and there should be a ruling this year. However, the saga will presumably continue to the Court of Justice.

The reality of competition policy has overtaken this debate: various expert reports on competition policy and the digital economy published in 2018/019 have identified self-preferencing as a serious competition issue in the digital economy, notably in digital ecosystems, i.e. when operated by so-called digital gatekeepers. The reports view self-preferencing in certain scenarios as a possible abuse of dominance under Art. 10 2 TFEU⁶ or clearly identify self-preferencing as a competition issue to be tackled.⁷ They call for changes, either under the existing rules or by changing abuse of dominance rules or through a new regulatory regime (see below). In the meantime, Germany has published a draft legislative proposal on national competition law, including specific rules on self-preferencing (see below).

This article looks at the possible deficiencies in current antitrust enforcement identified in the reports that triggered legislative action, before presenting and analyzing the draft legislation in Germany, notably asking whether it will render antitrust enforcement against abusive self-preferencing more effective.

 

II. DEFICINCIES IN CURRENT ANTITRUST ENFORCEMENT?

The expert reports show that the initial experience with cases in complex digital platform markets left the impression that traditional antitrust enforcement might not be effective in the digital economy, in which network effects, huge returns to scale and big data lead to sometimes irreversible results.⁸

It should be noted that enforcement difficulties in the digital economy did not only occur at EU, but also at national level: for example, in Germany, the Düsseldorf Court of Appeals still found in 2015 that a platform side without any financial remuneration relation between platform and user could not be viewed as a market in terms of competition law. This ruled out treating a zero-price platform side as a separate market, including finding a dominant position therein.⁹ The situation only changed through legislation that entered into force in June 2017.¹⁰

This illustrates the initial learning curve for cases in multi-sided digital platform markets. Indeed, the Commission has become faster. It rendered a decision in Google Android after approximately three years, and in Google AdSense after less than three years.¹¹ The FCO took approximately three years to terminate the Facebook proceedings,¹² and appeal proceedings (interim relief) are still pending.¹³ The FCO’s closed proceedings against Amazon in record time after approximately eight months in 2019 – but the proceedings ended without a formal decision or formal commitments.¹⁴ The Commission Amazon case will be an important test on how efficient proceedings in a case involving self-preferencing can be under the current regime.

Notwithstanding this evolution, the expert reports still saw deficiencies and mostly pleaded for change, pointing out a variety of reasons. The EU expert report generally found the existing rules under Article 102 TFEU to be sufficiently flexible to deal with cases in the digital economy and argued for adapting the rules in practice (evolutionary approach). On self-preferencing, the report identified the current effects-based test as applied in practice as main obstacle to effective enforcement and suggested reversing the burden of proof in certain situations: i.e. for vertically integrated digital platforms in markets with high barriers to entry, where the platform serves as an intermediation structure of particular relevance and performs a regulatory function. In these cases, the platforms should bear the burden of proof that self-preferencing “has no long-run exclusionary effects on product markets.”¹⁵ The report also explicitly mentioned the challenges for adequate remedies in self-preferencing cases and floated the idea of restorative remedies.

The German report competition law 4.0 suggested going beyond the existing dominance rules and introducing a new EU platform regulation. It found that the questions around the theory of harm in Google Search were still unresolved and viewed the case law evolution as too lengthy. Moreover, the report did not consider a solution as feasible whereby the Commission simply changed the application of Article 102 TFEU without legislation.¹⁶ The report suggested the regulation should include a general prohibition for dominant digital platform companies to engage in self-preferencing of its own services compared to those of third parties.

The earlier German report on modernization of abuse control had noted that the existing (national and EU) abuse of dominance rules are designed for case-by-case enforcement and thus less suited for intervention against infringements that occur frequently and at a broader scale, due to the applicable high burden of proof, but stopped short of suggesting legislative change at national level.¹⁷

The Furman Report went further and saw the need for a regulatory approach at national level:

[…] antitrust enforcement, […], moves too slowly and, intentionally, resolves only issues narrowly focused on a specific case. In digital markets this has not established clear and generalisable rules and principles to give businesses certainty about the boundaries of acceptable competitive conduct.

The Furman Report suggested developing a new regime, i.e. a code of conduct applicable only to companies with “strategic market status,” providing these with clear general guidance for admissible conduct. In a first step, these companies would be identified as addressees of the code of conduct. In a second step, infringements of the principles of the code of conduct could be pursued.¹⁸ The report proposed that the code consider specific forms of self-preferencing unfair: an online marketplace excluding or suspending rival sellers from its platform to give its own product or service an advantage, or a platform that contains a search function giving an unfair advantage to its own services over its rivals in downstream markets through the ranking or presentation of results.¹⁹

Interestingly, none of these reports explicitly mentioned a possible lack of specialized resources at the authority level (national or EU)²⁰ – even though that is what practitioners often point out as another important aspect of lengthy proceedings. In addition, Google Search showed how difficult it may be in practice for a competition authority to effectively bring an infringement to an end and to impose a functioning remedy. Two years following the prohibition decision and the order for Google to provide equal treatment to rival comparison shopping services the Commission is still monitoring the implementation.²¹ In its Amazon proceedings, the FCO has also experienced some difficulties in the context of the (soft) commitments’ implementation and how complex the monitoring of this process can be.

 

III. PROPOSED LEGISLATION IN GERMANY

Following the various reports, Germany was first to pursue legislative changes. The Ministry for Economic Affairs and Energy published the official draft legislation to change German national competition law on January 24, 2020.²² The amendment is part of a general amendment of the Act Against Restraints of Competition (“ARC”), but includes new rules on abuse control with a view to digital markets.²³

1. Status Quo on Self-Preferencing Under National Abuse of Dominance Rules

The concept of self-preferencing is currently not explicitly mentioned in German competition law. The prohibition to abuse dominance (Section 19 ARC) inter alia prohibits a dominant undertaking to directly or indirectly impeding another undertaking in an unfair manner (unfair impediment) or directly or indirectly treating another undertaking differently from other undertakings without any objective justification (discrimination).

There are no final precedents on self-preferencing under German competition law yet. While it seems that self-preferencing could not qualify as a (pure) discrimination case,²⁴ it may qualify as an unfair impediment.²⁵ Indeed, the German report on modernization of abuse control expressed the view that a dominant vertically integrated platform’s self-preferencing in connection with information asymmetries and exclusionary effects on rivals on downstream markets can amount to an unfair impediment under existing rules.²⁶ In addition, the Federal Cartel Office (“FCO”) took issue with Amazon’s platform rules on externally generated product reviews, i.e. that Amazon prohibited their posting or removed these from the marketplace, unless they were generated by Amazon’s own external review program, but the FCO terminated proceedings without a final decision.²⁷

1. The Draft Proposal on Self-Preferencing

Draft Section 19a ARC contains special rules on abusive conduct for undertakings that are active in multi-sided markets and networks and that have a “paramount cross-market significance” (“pcms”) for competition.

1. The Draft Legal Provision

The provision follows a two-step approach: the FCO first needs to issue a decision finding that a specific undertaking has a pcms for competition. The criteria include: a dominant position in one or more markets; financial strength or access to other resources; vertical integration and activities in other related markets; access to competitively relevant data; the significance of its activities for the access of third parties to procurement and sales markets, as well as the influence caused by this on the business activities of third parties (draft Section 19a (1)). The criteria are not cumulative.

In a second step, the FCO can prohibit specific conduct for the undertaking identified as having a pcms for competition. The specific conduct is listed in five regulatory examples (draft Section 19a (2) no.1-5), which are exhaustive. The first regulatory example provides that it is prohibited for undertakings with a pcms for competition “when acting as an intermediary for access to procurement and sales markets, to treat the offers of competitors differently compared to its own offers.”²⁸ The prohibition decision can be combined with the first step in a single decision.

Section 19a (2) stipulates at the end that there is an exception if the conduct in question is objectively justified. The burden of proof for the justification is with the undertaking concerned. In addition, the rules on interim measures are applicable.²⁹

2. The Draft Proposal’s Explanatory Statement

The explanatory statement provides additional input.³⁰ It confirms that the list of pcms criteria is not exhaustive, and they do not need to be met cumulatively. The finding of pcms rather requires an overall assessment of all relevant aspects in a case.

On process: the statement explains the decision on a pcms is within the FCO’s discretion and would likely need to be limited in time. The statement considers a period of five to ten years as appropriate, because the duration should enable effective intervention against the undertakings concerned, possibly even in several proceedings pursuant to draft Section 19a (2), and undertakings with a pcms would likely have a very strong and permanent position.

The statement clarifies that draft Section 19a (2) no.1 indeed targets self-preferencing (even though the term is not mentioned) as a prohibited discrimination. The explanations refer to Google Search for the competitive harm self-preferencing can cause in digital markets (market foreclosure and restrictions for rivals to develop and market innovative offers based on competition on the merits). There is an assumption that self-preferencing by undertakings with a pcms, that are integrated in vertical or conglomerate markets, may add to further strengthening or expanding their cross-market power.

The high potential for competitive harm is also one of the reasons given for the reversal of the burden of proof. The explanations clarify that this seems to go beyond having to merely prove an objective justification: in essence the new rules set out the assumption that the conduct in question is abusive and prohibited – and while the addressees can rebut the assumption by demonstrating an objective justification, any non liquet will be to their detriment.³¹

In the second step, i.e. after finding a pcms for competition, the FCO cannot only prohibit specific conduct, but also order the addressee to terminate the infringement (including restorative measures) or accept remedies.³² These are administrative proceedings, i.e. they cannot lead to a fine. The FCO can only impose a fine if the addressee were to violate the FCO’s second-step decision (see draft Section 81 (2) no. 2 lit. a ARC).

1. Analysis and Comments

The following reviews first how the draft proposal combines suggestions from the various reports, before commenting on the provision’s design and the question whether it will render enforcement more effective.

1. Combination of Different Concepts

The proposed rules follow a patchwork approach – they are a combination of the concepts suggested in the various reports. The proposal combines competition law with special regulation, creating a hybrid set-up with elements of both an ex ante regulatory approach and of case-by-case enforcement. It is clear, though, that the FCO will be competent to enforce the new rules as part of its competition law powers.

Draft Section 19a does not apply to dominant undertakings – as the German report competition law 4.0 suggested – but to undertakings that fall within a newly defined category of pcms for competition, which in principle seems similar to the concept of undertakings with “strategic market status” in the Furman Report. (There is, however, an important difference on whether the concept requires dominance, see below.) In addition, the draft proposal also follows the latter’s suggestion of a two-step procedural approach.³³ This “carve-out” from the general abuse of dominance control also marks a difference to the EU expert report (which a priori limited itself to looking at the existing framework of EU competition rules.)

On the other hand, the prohibition of self-preferencing is broad and of a general nature (and does not even mention the term self-preferencing), which is more in line with the German report competition law 4.0. The Furman Report was much more specific and detailed in outlining the type of self-preferencing conduct for platforms to be considered as unfair and generally prohibited (see above).

Regarding the reversal of the burden of proof: here the draft seems to implement the suggestion of the EU expert report, i.e. that the platform should bear the burden of proof that self-preferencing has no long-run exclusionary effects (see above). In fact, draft Section 19a (2) no.1 seems to go further: it reads like a per se prohibition of any different treatment of the platform’s offers and of those of competitors, without any link to a resulting impairment or exclusionary effect on competitors – subject to an objective justification. However, unlike enforcement of Article 102 TFEU, the FCO’s decision under draft Section 19a (2) cannot lead to a fine. (A sanction is only possible if the addressee were to violate the second-step decision.)

2. Comments on the Design of the New Provision

The draft proposal received many comments from the different stakeholders, of course with varying positions in substance. In practice, it seems that new rules, including on self-preferencing, will come, so questioning the wisdom of introducing legislation might almost seem obsolete. But the design has raised many questions. Some selected concerns are set out below.³⁴

One obvious question is why the draft provision does not mention “digital platforms.” Draft Section 19a only requires the addressee to have activities in “multi-sided markets and networks.” However, the deficiencies of traditional antitrust enforcement that triggered the proposed legislation relate to the particularities of the digital economy. It is questionable whether the new rules are really necessary for non-digital markets.

Another concern is that dominance is one of several possible conditions for a pcms, but it is not a mandatory requirement. This renders the possible scope of application of the new rules much broader than what the various reports identified as an increased risk for competition in the digital economy:³⁵ The current wording allows finding a pcms even if the undertaking has no dominant position in any market. And the new provision is not limited to preventing leveraging a dominant position into other markets, which seems to contradict the draft proposal’s own general explanations.³⁶ Ultimately, intervention in digital ecosystems below the level of dominance in any market is difficult as it might stifle innovation.

As mentioned, the prohibition of self-preferencing is designed as a per se prohibition of different treatment of an addressee’s own products with rivals’ product offers. In contrast, the other regulatory examples in draft Section 19a (2) no-2-5 all contain a reference to impeding competitors or adverse effects. The explanations do not offer any reasons for the stricter approach towards self-preferencing.

The very broad scope, combined with the far-reaching reversal of the burden of proof and the lack of a limitation for the validity of the decision on a pcms for competition in the law, may trigger questions as to the requisite legal clarity and appropriateness of the rules.

3. Will Antitrust Enforcement Against Self-Preferencing Become More Effective?

One major deficiency identified was the long duration of proceedings. As seen, lately the cases in the digital economy lasted about three years, without judicial review. It is unclear whether the new approach would be much quicker overall, at least in the near future. Finding a pcms for competition would likely be as complex as finding a dominant position (and establishing additional activities in downstream or adjacent markets relevant for self-preferencing) under existing rules. The second step, i.e. issuing a decision to prohibit self-preferencing, might indeed be quicker, given the reversal of the burden proof. In fact, the FCO has expressed the hope that under the new rules it would be easier and quicker to obtain relevant information from addressees than today.³⁷ And the FCO could base several prohibition decisions (step two) against one addressee based on the same single decision finding pcms, which would “streamline” enforcement.

However, the two-step concept means that addressees can presumably appeal the first step separately – which might delay the second step considerably. It can be expected that the initial decisions finding a pcms will be appealed up to the Federal State Court, which means a delay of at least three years or more.³⁸ Even if the FCO combines the first and second step, it is likely that the first final decisions will take several years.³⁹ That is of course to some extent normal if new legislation is introduced. But the various concerns raised by the broad approach (see above) may well render the cases unnecessarily more susceptible to lengthy judicial review. Ultimately, the new regime may thus only work effectively in the mid-term.

The broad and very general scope of the prohibition to engage in different treatment of third parties’ offers and the addressee’s own offers seems to fall short of providing clear-cut ex ante guidance on which concrete conduct is prohibited in practice for the companies concerned. There might be numerous differential treatments applied by potential addressees – willingly or unwillingly. It will thus depend on how the second step decision will detail each prohibition decision. That means enforcement of draft Section 19a (2) no. 1 will ultimately continue to be a case-by-case exercise, something which was perceived as a deficiency in some reports (see above) and which contains an element of ex post enforcement.

This leads to the issue of effective remedies: enforcement of the new rules will still face the challenge on how to effectively end possible infringements against the prohibition decisions taken in the second step. In practice it may already be a similar challenge to issue a decision in the second step, which clearly identifies the prohibited conduct in more details. The experience in Google Shopping shows that simply ordering equal treatment – which is the other side of the coin of the prohibition set out in draft Section 19a (2) no.1 – may not be sufficient to properly end the infringement (see above).

The issue is to some extent linked to the question of sufficient resources, and additional specialized resources for the digital economy. While this may be an annoying (because costly) implementation detail from the legislator’s perspective, it can have a massive impact in practice. The proposal does not really address this problem. (A heretic thought: maybe increasing the budget and employing more specialists would render antitrust enforcement more effective more swiftly and easily than the introduction of the new regime.) While the FCO says it has already employed specialists, it might still be necessary to substantially increase the resources. It may be helpful to create new specialized data science units – parallel to or as part of the chief economist units that were created when the more economic approach was implemented.

Some consider the draft legislation as insufficient in light of the large tech companies’ persisting market power and argue for “real” ex ante sector-specific regulation and a specialized administrative entity with effective oversight powers (like in telecoms), including on self-preferencing practices.⁴⁰ It is clear that the current approach aims at keeping self-preferencing conduct within the realm of competition law and under the control of the FCO – at least for now. The effectiveness of the new powers in practice will decide whether this attempt will be successful.

Finally, from a European perspective, the approach to push for national legislation would not seem to render enforcement more effective. While national abuse of dominance rules may be stricter than EU competition law (Article 3(2) Reg. 1/2002), in digital markets national boundaries lose their significance, and it would be better to have a level playing field in the EU, including on the rules on self-preferencing. If other Member States follow suit and introduce their own, slightly different national legislation, the enforcement of abuse of dominance in the digital economy will become even more of a patchwork. This is not desirable from a policy standpoint, nor for potential complainants or defendants.

One of the reasons for Germany to push forward may have been the expectation (and experience) that national legislation is easier to achieve than new rules at EU level with 27 Member States. The draft proposal adds political pressure on the EU and the Commission to move as well on possible new legislation. Commissioner Vestager indeed recently indicated that the Commission considers introducing ex ante regulation for digital platform gatekeepers, with a focus on preventing the tipping of markets.⁴¹ However, it is unclear whether the Commission would opt for a full regulatory approach in the end.⁴² It will be interesting to follow the development in this area, including during Germany’s presidency starting in mid-2020.

---

1 The author is a partner at law firm Heinz & Zagrosek mbB in Cologne, Germany. For transparency reasons, the author discloses that she has represented a complainant against Amazon in the German Amazon proceedings. The views expressed are her own.

2 European Commission, decision of June 27, 2017, Case AT.39740.

3 Regarding the selection of the winners of the Amazon “Buy Box” and the impact of Amazon’s potential use of competitively sensitive data collected from competing marketplace sellers for that, see Commission press release of July 17, 2019, https://ec.europa.eu/commission/presscorner/detail/en/IP_19_4291.

4 In the context of platform rules for user reviews of sellers and products, see Federal Cartel Office case B2–88/18, case summary of July 17, 2019, available in English at https://www.bundeskartellamt.de/SharedDocs/Meldung/EN/AktuelleMeldungen/2019/17_07_2019_Case_Summary_Amazon.html.

5 Se e.g. Bo Vesterdorf, Theories of Self-Preferencing and Duty to Deal – Two Sides of the Same Coin, 1(1) Competition Law & Policy Debate, 4 (2015); Nicolas Petit, Theories of Self-Preferencing under Article 102 TFEU: A Reply to Bo Vestedorf, 2015, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2592253; Pinar Akman, The Theory of Abuse in Google Search: A Positive and Normative Assessment under EU Competition Law, Journal of Law, Technology and Policy, 2017, 301; Thomas Höppner, Duty to Treat Downstream Rivals Equally: (Merely) a Natural Remedy to Google’s Monopoly Leveraging Abuse, 1 CoRe (2017) (3), 208; Inge Graef, Differentiated Treatment in Platform-to-Business Relations: EU Competition Law and Economic Dependence, Yearbook of European Law, (38) 2019, 448.

6 See expert report for the European Commission by Crémer/de Montjoye/Schweitzer, Competition policy for the digital era, April 4, 2019 (“EU expert report”), p. 7; and report for the German Ministry for Economic Affairs and Energy by Schweitzer/Haucap/Kerber/ Welker, Modernisierung der Missbrauchsaufsicht für marktmächtige Unternehmen, August 28, 2018 (“German report on modernization of abuse control”) p. 102/103, in particular combined with the exploitation of information asymmetries.

7 See Furman/Coyle/Fletcher/McAuley/Marsden, Unlocking digital competition, Report of the Digital Competition Expert Panel, March 13, 2019 (“Furman report”) p. 58, 60; ACCCC, Digital Platforms Inquiry, final report, June 2019, p. 12; report of Kommission Wettbewerbsrecht 4.0 for the German Ministry of Economic Affairs and Energy on the development of EU competition law, Ein neuer Wettbewerbsrahmen für die Digitalwirtschaft, September 9, 2019 (“German report competition law 4.0”), p. 53.

8 See for example German report competition law 4.0, p. 50.

9 As a result, when the Federal Cartel Office (“FCO”) reviewed a complaint by German publishers against Google, it could not conclude that there was a general search engine market, see FCO decision of September 8, 2015, B6–126/14, Google/VG Media. The FCO had to revert to the online advertising market instead. Google Search could thus not have been brought in Germany.

10 New Section 18 (2a) Act Against Restraints of Competition.

11 Commission decision of July 18, 2018, Case AT 40099 – Google Android. The Commission issued the decision in Google Adsense on March 20, 2019, see Commission press release https://ec.europa.eu/commission/presscorner/detail/en/IP_19_1770. Even though these cases do not strictly concern self-preferencing as a theory of harm but more traditional approaches like tying and exclusivity agreements.

12 FCO, decision of February 6, 2019, B6–22/16.

13 The FCO lost in interim relief proceedings (against its decision ordering Facebook terminating the infringement) in 2019, and the Federal Court is supposed to hear that case (still interim relief) in June 2020.

14 See supra note 3.

15 See p.7.

16 Publishing new guidance would not bind the courts, and just having one case as a basis for new guidance might not be sufficient, see p. 54.

17 It pondered the possibility of introducing a regulatory example in the national abuse of dominance rules on self-preferencing (in the context of abuse of information asymmetries by digital platforms) for clarification purposes, but ultimately left the question open, see p. 111.

18 Annex D, Strategic recommendations A and C.

19 P. 60/61.

20 The Furman Report suggested to create a new digital market unit in order to enforce the new rules, which may be an indirect acknowledgement that the current resources at the level of the CMA would not be sufficient. Of course, this may also be influenced by Brexit and the general need to increase resources in this context.

21 See E-003869/2019, Answer given by Executive Vice-President Vestager (EP Parliamentary Question), February 10, 2020.

22 https://www.bmwi.de/Redaktion/DE/Downloads/G/gwb-digitalisierungsgesetz-referentenentwurf.pdf?__blob=publicationFile&v=10.

23 See also Thomas Weck, The New Abuse Rules in the German Competition Act – What’s in it for the EU?, CPI column, April 14, 2020, https://www.competitionpolicyinternational.com/the-new-abuse-rules-in-the-german-competition-act-whats-in-it-for-the-eu/.

24 The notion of “different treatment” requires that the relevant comparison covers similar undertakings. However, pursuant to precedents there is no similarity if the comparison involves another undertaking on one side and a group company/affiliate on the other. Thus, merely treating group companies at downstream level advantageously compared to third parties (self-preferencing) is not considered as a discrimination under German law, see Markert in Immenga/Mestmäcker, Wettbewerbsrecht GWB, 5. Auflage 2014, § 19 Rn. 121, German report on modernization of abuse control, p. 99, with further references.

25 Impediment is broadly defined as any conduct that objectively impairs a third party’s opportunities to compete. This requires an actual and not only potential adverse effect, but in practice the bar has typically not been overly high, see Bechtold/Bosch, GWB, Kommentar, 9. Auflage 2018, § 18, Rn. 8. The focus of the review in practice is often the required balancing of interests, in order to determine whether the impediment is indeed unfair.

26 P. 111.

27 The FCO was concerned that because Amazon’s own review program was only available for Amazon retail, this would put independent sellers at a disadvantage through rerouting the supply flow towards Amazon’s own retail activities. Amazon justified the conduct with consumer protection concerns (fake reviews). Amazon undertook to gradually open its own review program also for certain sellers, see FCO case summary, supra note 3.

28 Roughly, the other examples prohibit to impede competition in markets where it is not yet dominant; use data for making market entry more difficult; demand terms and conditions that allow the use of data of others; make portability of data more difficult; withhold information from other companies about their success in markets, see for the summary and overall analysis Podszun/Brauckmann, Germany’s Pressing Ahead: The Proposal for a Reformed Competition Act, CPI column of November 6, 2019, https://www.competitionpolicyinternational.com/germanys-pressing-ahead-the-proposal-for-a-reformed-competition-act/.

29 The draft legislation also proposed to lower the conditions for imposing interim measures.

30 P. 78 of the draft proposal. It is rather short on self-preferencing, with just one paragraph.

31 P. 78 of the draft proposal. It is rather short on self-preferencing, with just one paragraph.

32 P. 76.

33 Something which the German report on EU competition law explicitly rejected.

34 For a more comprehensive set of comments, see for example the submission of Studienvereinigung Kartellrecht to the Ministry for Economic Affairs and Energy: https://www.studienvereinigung-kartellrecht.de/sites/default/files/stellungnahmen/73a15d4c84c8c2c30822454ada3ed1fa/200214_stuv_stellungnahme_digitalisierung_missbrauchsaufsicht_10_gwb_novelle_bmwi.pdf.

35 The German report on EU competition law suggested to introduce a prohibition of self-referencing for dominant companies (p.54); the concept of strategic market position in the Furman Report is linked to dominance, see p. 42, 59.

36 Which say that the new rules target companies that do not only have a dominant position in individual platform or network markets, but additionally have resources and a strategic position, enabling them to exercise significant influence over third parties’ business activities and to continuously expand their own activities into new markets and industries (p.75).

37 See FCO comments on draft proposal, published February 25, 2020, available at https://www.bundeskartellamt.de/SharedDocs/Publikation/DE/Stellungnahmen/Referentenentwurf_10_GWB_Novelle.html.

38 If by the time the first step has been finally decided, the validity of the decision on a pcms would have expired, this could lead to an ongoing game of first step and judicial review.

39 In particular including interim relief proceedings.

40 See for example Thomas Höppner/Jan Markus Weber, Die Modernisierung der Missbrauchskontrolle nach dem Referentenentwurf für eine 10. GWB-Novelle, K&R 2020, 24.

41 See for example the coverage in Global Competition Review on April 24, 2020, on the Commissioner’s comments in the context of the ABA Spring Meeting, Vestager: EU may introduce competition rules for “digital gatekeepers.”

42 It seems that the Commission will now first outsource a study on the gatekeeping power of digital platforms, including on self-preferencing practices, and the question whether ex ante regulation is required to complement competition law, see for example coverage in PaRR on May 11, 2020, EC launches digital platforms ‘gatekeeping’ assessment tender.