EU COMPETITION LAW AND THE USE OF “MFN” CLAUSES IN E-COMMERCE

(1)

UNIVERSITY OF AMSTERDAM FACULTY OF LAW

EU COMPETITION LAW AND THE USE OF “MFN” CLAUSES IN

E-COMMERCE

LLM Thesis

Study programme: EU Competition Law & Regulation Name: Alexander Navračič

Student number: 12351210

Thesis Supervisor: Dr. Jan Broulík

(2)

DECLARATION

I declare that this thesis contains no material that has been submitted and/or accepted for the award of any other degree in any University or Institution of Higher Education. To the best of my knowledge and belief, this thesis contains no material previously published or written by any other person, except where due reference is made in the text. I hereby certify that I am fully aware of the consequences for the academic offence of plagiarism, in accordance with the applicable Program and Exam Regulations of the University of Amsterdam.

Date: 26 July, 2019

ACKNOWLEDGEMENT

I would like to thank Mr Broulík for his professional guidance, time dedicated to reading my drafts, his honest feedback and critical remarks that helped me to finalise this paper into this final version. Also, I would like to express my utmost gratitude to the University of Amsterdam, Ms Cseres and the Competition Law Department in particular, for their professional and patient approach to education throughout this past year, which undoubtedly provided me with solid basis for writing an academic paper on such a comprehensive topic.

(3)

ABSTRACT

Suggested Citation: Navračič, Alexander, ‘EU Competition Law and the Use of MFN Clauses in E-Commerce’ (June 14, 2019) [Master Thesis]. University of Amsterdam, Faculty of Law. Supervisor: Dr. Jan Broulík. Amsterdam 2019, 41 pages.

The paper focuses on the competition law assessment of platform MFNs in commercial contracts between online platforms and suppliers. The main focus is on respective assessments of articles 101 and 102 TFEU in relation to the topic of platform MFNs. The arguments and conclusions put forward in this paper stem from the relevant case-law and competition law theory. The paper answers the question of why abuse of dominance is the correct theory of harm for regulating platform MFNs, and how can the competition law enforcement discrepancy that is currently existing in the EEA be remedied.

(4)

LIST OF ABBREVIATIONS

B2B Business to Business B2C Business to Client

CJEU Court of Justice of the European Union EEA European Economic Area

MFNs Most Favoured Nation Clauses MS Member State

NCA National Competition Authority OTAs Online Travel Agents

(6)

Introduction

E-Commerce is a phenomenon that dominates the digital markets as it provides for the commercial use of the internet. In other words, anything that is purchased or sold online is part of E-Commerce, and thus, one can say that it is a market of markets. Indeed, this description is accurate, especially if one considers the fact that E-Commerce bridges selling part of the market with its purchasing part, and thus does not only provide for B2C transactions but also B2B transactions. Consequently, it has a proclivity to produce positive network effects with significant economies of scale for an undertaking that manages to be a sort of digital “point of contact” for all online transactions. Such points of contact are online platforms like Amazon, Booking.com, HRS or even Apple with its iTunes and AppStore.

As the network effects grew, so was an economic attractiveness of each of these platforms. With a sufficient number of users on both sides of the platform, it has become a common custom that online platforms started to adopt Most Favoured Nation clauses (hereinafter referred to as “MFNs”) in their contracts with sellers using their platform. This way, they ensured preferential treatment regarding prices of products/services sold via their channels, which essentially has anti-competitive effects on inter-platform competition. As a result, MFNs became the subject of competition law investigations by National Competition Authorities (hereinafter referred to as “NCAs”) and the Commission, initially in E-Books Case and later in OTAs Cases. A commonality between these landmark MFN decisions is that the NCAs’ enforcement caused a discrepancy in the application of EU competition rules. Thus, before the Commission’s Amazon Decision in 2017, the predominant approach to platform MFNs was art. 101 TFEU, which as will be explained by this paper, is not as appropriate for regulating this issue as art. 102 TFEU.

Therefore, the following chapters will explain the correlation between MFNs and online platforms (Chapter 1) as well as Effects of MFNs on Competition (Chapter 2). Next, the decision-making patterns of NCAs and the Commission will be examined (Chapter 3) followed by the competition law assessments of platform MFNs under art. 101 and art. 102 TFEU (Chapter 4). At last, assessment criteria based on the priority of art. 102 TFEU vis-à -vis the platform MFNs will be presented (Chapter 5) as well as outlooks for future coherence and enforcement (Chapter 6).

Methodology and Research Substance

The central object of this research is to indicate a proper assessment of platform MFN clauses under EU competition law. In that regard, the paper aims to answer the following

(7)

research question: How should MFN clauses used by online platforms in their commercial contracts be assessed under EU competition law? Also, to provide a more in-depth examination of the principal research objective, the following sub-questions were construed: How are articles 101 & 102 TFEU interpreted concerning platform MFN clauses? Which approach – 101 or 102 – turns out to be more appropriate in terms of enforcement and legal certainty? How should be the correct future application secured?

In order to reach the research objective of the paper, one will use theoretical concepts stemming from the general competition law theory; mainly genuine agent theory and abuse of dominance theory.

Furthermore, the methodology used in writing this paper is classic legal research, where one tries to adopt a perspective of a judge and apply the legal provisions according to their purpose and spirit.1_{However, this paper does not restrict itself only to comment how the law is}

applied, but rather goes a step further and states how it should be applied. As a result, some parts of the presented legal assessment involve description and interpretation of the EU competition law from the viewpoint of the participants in the legal order.

1. MFNs and Online Platforms

Since, the tendency of incorporating MFN clauses in commercial contracts by online platform providers increased in the last decade, some NCAs took the initiative and started to apply their national competition rules to these undertakings.2_{Another reason for the proactive}

approach of the NCAs is that the Commission has been silent on regulating anti-competitive practices in the E-Commerce sector for quite some time. It was not until May 2015 when the Commission acted and initiated a sector inquiry in E-Commerce that pointed to the main challenges for competition law enforcement regarding digital markets.3_{Thus, with a lack of}

action from the Commission, the NCAs had to take the initiative and enforce their national rules on the online platform providers. Consequently, each jurisdiction chose a different path to

1_{The research and interpretation theory developed by Ronald Dworkin in his Law’s Empire. See Ronald Dworkin,}

Law’s Empire (Harvard University Press, 1986).

2_{Among the first NCAs to start investigations in the subject of MFN clauses laid down in commercial contracts} used by online platform providers were France, Sweden and Italy that reached commitments with Booking.com, thus limiting the use of such clauses on the OTAs market. See Italian Competition Authority Press Release, ‘Commitments Offered by Booking.com Closed the Investigation in Italy, France and Sweden’ (Apr. 21, 2015) <http://www.agcm.it/en/newsroom/ press-releases/2207-i779-commitments-offeresd-by-bookingcom-closed-the-investigation-in-italy- france-and-sweden.html> accessed 4 April 2019.

3_{Commission Press Release, ‘Antitrust: Commission launches E-Commerce sector inquiry’ (May 6 2015)} <http://europa.eu/rapid/press-release_IP-15-4921_en.htm> accessed 4 April 2019.

(8)

regulate MFNs, which caused enforcement discrepancy that created legal uncertainty.4_An

example of such discrepancy can be a different approach to MFNs concerning narrow and wide MFN clauses. While some jurisdictions allow narrow MFNs and forbid only wide MFNs5_,

another jurisdiction banned MFNs in absolute terms6_{. The problem with discrepancy among}

MS is even aggravated by the fact that the rules that are being applied at the national level by NCAs are practically the same.7_{Moreover, another issue when it comes to enforcement is the}

substantive approach the NCAs take towards MFNs used by online platforms – some jurisdictions prefer art. 101, while others opt for art. 102 approach. This discrepancy does no good to legal certainty, and unclarity as to which article is going to be applied in proceedings has significant negative implications for a right of undertakings to conduct business, as theories of harm differ when it comes to art. 101 versus art. 102.

Further, another problem that arises due to the incorrect application of art. 101(1) TFEU to platform MFNs by NCAs is the denial of multilateral nature of this provision, which fundamentally contradicts the legal spirit of the provision. So far, most NCAs’ investigations to online platform agreements containing MFNs were based on art. 101 but the decision was directed against only one party to the agreement (the online platform), instead of all parties to the agreement.8_{As Akman points out in this regard, where the anti-competitive practice has}

nature of the “agreement”, investigating the conduct of and addressing the decision to only one of the parties to the agreement9_{implies a mismatch between the NCAs’ theory of harm and}

parties’ action.10_{Such mismatch raises doubts about whether the NCAs operate on the basis of}

the correct theory of harm. Clearly, it opposes the well-established concepts of the competition law theory – addressing the decision to only one of the cartelists suggests that the theory of harm is based on unilateral conduct, but then the legal action is being pursued under a provision prohibiting multilateral conduct.11_{As a result, the correct theory of harm that should be used in}

4_{Pinar Akman, ‘A Competition Law Assessment of Platform Most-Favored-Customer Clauses’ Journal of} Competition Law & Economics, Vol. 12(4) 781, 788.

5_{Sweden allows for narrow MFNs as was established in Booking.com commitments. France initially also accepted} the use of narrow MFNs but this changed with the adoption of “Macron Act” that banned all kinds of MFN clauses. It was adopted as a reaction to the Autorité de la concurrence investigation of Booking.com. For specific

provisions see

https://www.legifrance.gouv.fr/affichTexte.do;jsessionid=66289F27574A8C3B376FF9739E4D8D34.tpdila07v_ 2?cidTexte=JORFTEXT000030978561&dateTexte=&oldAction=rechJO&categorieLien=id&idJO=JORFCONT 000030978558.

6_{Germany rejects all forms of online platform MFN clauses – See HRS and Booking.com decisions.} 7_{Akman (n 4), 784.}

8_{HRS Decision of Bundeskartellamt.}

9_{As opposed to all parties to the agreement, which is the case in cartel decisions.} 10_{Akman (n 4), 784.}

(9)

enforcement proceedings relating to online platform MFNs is an abuse of dominance laid down in art. 102 TFEU. The reasons are twofold:

I. Online platforms are agents of the sellers with which they enter into contracts containing MFN clauses, and this qualifies art. 101(1) TFEU inapplicable as under the competition law theory there must be at least two separate undertakings12_;

II. In the absence of market power held by the online platform, such clauses would be unlikely to raise competition law concerns.13

In order to analyse the competition law issues stemming from the use of MFN clauses by online platforms, the following section will shed a bit of light on the basic terms that are necessary to be defined.

1.1. Online Platforms

Online platforms are business models present in the E-Commerce sector, whose primary economic purpose is to act as an intermediary and thus enable sellers and buyers to meet and enter into transactions between them.14_{Mostly, they serve as a meeting point for the supply and}

demand part of the particular market. This pattern already implies their inherent feature, which is that the online platforms are a two-sided market place. In other words, it concentrates transactions both downstream and upstream from the intermediary.15_{Thus online platforms, as}

opposed to the traditional offline retail model, do not only deal with downstream end-consumers but also with upstream suppliers. Essentially, the platform offers services to the two different consumer groups, the value of which, from the perspective of one of the groups, depends upon the number of members of the other group.16

Since the role of platforms is mostly related to connecting sellers and consumers, the success of their business model depends on the number of users they attract on both sides of the market – producing indirect network effects.17_{Thus, the online platforms have to develop}

price and product strategies that are capable of balancing their interests against those of other

12 Genuine Agent theory described in the Guidelines Vertical Restraints. Commission, ‘Guidelines on Vertical Restraints’ (Communication Document) OJ C 130, 19 May 2010 (hereinafter referred to as the ‘Guidelines’). 13_{As they would fall within the exemption thresholds of De Minimis Notice or VRBER.}

14_{Commission, ‘Communication on Online Platforms and the Digital Single Market’ COM (2016) 288 final, 1} <https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52016SC0172&from=EN> accessed 4 April 2019.

15_{David S. Evans and Richard Schmalensee, ‘Guide to the Vocabulary of the New Economics of Multisided} Platforms’ (2016) SSRN <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2793021> accessed 4 April 2019. 16_{Julian Wright, ‘One-Sided Logic in Two-Sided Markets’ (2004) 3(1) Review of Network Economics 44, 44.} Also see Roberto Roson, ‘Two-Sided Market: A tentative Survey’ (2005) 4(2) Review of Network Economics 142, 142.

17_{Renata B Hesse, ‘Two-Sided Platform Markets and the Application of the Traditional Antitrust Analytical} Framework’ (2007) 3(1) Competition Policy International 191, 191.

(10)

parties involved in online transactions.18_{The most common system that online platforms choose}

for creating such balance is an agency model.19_{In light of this model, the platforms enter in}

separate contracts with the sellers, and then the sellers enter into their own sales contracts with particular customers. Furthermore, the agency model enables the platforms to earn their profits by charging commission fee from commercial exchanges they facilitated20_{without requiring}

them to purchase any good/service.

In fact, the double-sided face of online platforms creates a sort of a digital eco-system that unifies supply and demand for particular goods or services; it can be said that they are controllers of crucial infrastructure that enables the relevant markets to exist digitally. This is to say especially if the platform is a “digital unicorn”21_{that managed to grasp the momentum}

at the very outset of the E-Commerce boom and now attracts many customers globally. In such a case, platform benefits from positive network effects as the overall value in a platform and its ability to generate consumer benefits grow as more users participate.22_{As a result, online}

platform providers create a circle of dependency for SMEs who engage in a borderless commercial exchange on the platform, which essentially grants the platform providers bargaining power towards sellers. In light of the above, online platform providers must to a certain extent be perceived as controllers of infrastructure and not mere agents that only charge their commission fee, as access to the particular platform with wide consumer reach becomes something as indispensable part of the business model for the sellers using it.

18_{This is to emphasize the importance of the platform’s price structure for internalization of indirect network} effects, so that both sides of the market are inclined to use their service. In this regard see in more detail at: Jean-Charles Rochet and Jean Tirole, ‘Two-Sided Markets: A progress Report’ (2006) 37(3) RAND Journal of Economics 645, 646-647.

19_{Under the agency model the principals (suppliers) set the final prices. Other models that are also common are} wholesale model and merchant model. In terms of the former, the supplier sets the wholesale price and the platform marks up the retail price. Under the latter model, the supplier sells goods/services to the platform for wholesale price, and the platform then sets its own retail price. See further at: Justin P Johnson, ‘The Agency Model and MFN Clauses’ (2017) <https://ssrn.com/abstract=2217849> accessed 5 April 2019.

20_{Platform-Commission fee can either have a form of subscription fee or percentage from the value of the} transaction. The choice of model depends on the profitability assessed by screening the platform’s customer base and amount of commercial transactions per pre-determined period. Definition available at: Directorate for Financial and Enterprise Affairs - Competition Committee (OECD), ‘Hearing on across platform parity agreements’ (Note by France) 27-28 October 2015, page 2, <http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=DAF/COMP/WD(2015)64&doclang uage=en> Accessed 5 April, 2019.

21_{The term refers to a privately held start-up company with valuation over USD $ 1 billion. It refers to statistical} rarity of growth of such ventures over a short period of time. See further Cristea, I.A.; Cahan E.M. Ioannidis and John P.A. ‘Stealth research: Lack of peer‐reviewed evidence from healthcare unicorns’ (2019) European Journal of Clinical Investigation 49(4), 1.

22 _See _page ₇ _at: _{Rob Frieden, ‘The Internet of Platforms and Two-Sided Markets:} Legal and Regulatory Implications for Competition and Consumers’ (2017) <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3051766> accessed 5 April 2019.

(11)

1.2. Most Favoured Nation Clauses (MFNs) – General Perspective

Traditionally, an MFN clause is defined as a contractual provision that has the effect of binding one party to treat another party (e.g. the buyer) as favourably as that party treats its best customer.23_{In light of this definition, a buyer that has an MFN clause in his contract with the}

supplier has a guarantee that he will not pay a higher price than the supplier’s most-favoured-buyer.24_{Clearly, the MFN clause in traditional sense links prices between different customers}

of the same seller.25

Furthermore, with the rise of E-Commerce, the online platform providers have swiftly adapted their business models alongside the logic of MFN clauses.26_{When online platforms}

resort to MFN clauses in their contracts with sellers using their platform, their effect is that the sellers bound by such MFN clause cannot charge higher prices on the MFN-privileged platform than it does on another (rival) platform.27_{Therefore, suppliers are forced to equalise their prices}

on the MFN-privileged platform, with the lowest price they offer to other commercial partners, irrespective of the fact whether it is online or offline. Effectively, the MFN clause ties the consumers to the platform as it ensures the benefit of offering the lowest price for a particular good/service on the market. In other words, as opposed to MFNs in the traditional sense, the platform MFNs are “third party clauses” that connect the prices for the same customer for purchases from different platforms.28

1.2.1. Specific Types of MFN Clauses in E-Commerce: Narrow versus Wide

Platform parity clauses are, from the effects-oriented point of view, categorised in narrow and wide. In terms of narrow MFNs, their effect is that an online platform cannot be subjected to higher prices or lower conditions than those the seller provides on his own website.29_{Wide MFNs, on the other hand, have much stricter implications as to the commercial}

freedom of the seller. Such clauses require the seller to guarantee the platform that he will not afford higher prices and better standards not only on his own website but also on any other

23_{Jonathan B Baker, ‘Vertical Restraints with Horizontal Consequences: Competitive Effects of} Most-Favoured-Customer Clauses’ (1996) Antitrust Law Journal 517, 519.

24_{However, MFN clauses do not solely relate to price restrictions, they may also concern other terms and} conditions that are inherent to commercial partnership. In particular, Amazon Decision illustrates other limitations posed by MFN clauses – e.g. Competitors’ business model notification.

25_{Ingrid Vandenborre and Michael J Frese, ‘Most Favored Nation Clauses Revisited’ (2014) European} Competition Law Review 588, 588.

26_{This pattern has been predominantly present with online travel agents, price comparison websites and online} market places.

27_{Akman (n 4), 782.}

28_{Vandenborre and Frese (n 25), 588.}

29_{Ariel Ezrachi, ‘The competitive effects of parity clauses on online commerce’ (2015) European Competition} Journal 11(2-3) 488, 489.

(12)

competing platform.30_{Clearly, the effect of wide MFN is price parity as lowering prices on any}

sales channel automatically leads to lowering prices also on the MFN-privileged platform. Indeed, lower prices are generally considered as pro-competitive effect, however as will be explained in forthcoming chapters, online platforms market is very transparent, and thus the suppliers are constrained to sell through all incumbent platforms, ergo simultaneous decrease in price due to wide MFN essentially softens competition between the platforms, and results in price maintenance or even price increase charged to final consumers.31_{In addition, wide MFN}

clauses tend to incorporate obligations as to the notification of the business model of possible competitors to the platform.32

The next chapter will explain the effects of MFNs on competition in more detail.

2. Effects of MFNs on Competition

At first glance, the MFN clauses seem to generate potential benefits to consumers, especially with price transparency and reduction of transaction costs. However, they also raise competition concerns due to their ability to acquire or strengthen monopoly pricing.33_The

benefit or harm to consumers depends on the characteristics of the market affected, the specificities of the clause, and the nature of the seller or sellers who offer it.34_{As with any}

potentially anti-competitive conduct, positive and negative effects on competition must be considered.

2.1. Negative effects on competition

MFNs have far-reaching negative effects on all parameters of competition that can damage both competitors as well as consumers. The most common anti-competitive effects of MFNs will be discussed in the section below. Also, it must be pointed out that although platform MFNs stem from vertical relationships, their anti-competitive effects mainly strike on a horizontal level in the form of foreclosure and facilitation of collusion between platforms.35

30_{Amelia Fletcher and Morten Hviid, ‘Broad Retail Price MFNs: Are they RPM “at its worst”?’ (2016) 81(1)} Antitrust Law Journal 65, 71.

31_{Andre Boik and Kenneth S Corts, ‘The Effects of Platform Most Favored-Nation Clauses on Competition and} Entry’ (2016) 59 Journal of Law and Economics 105, 108.

32_{As was the case with E-Books and Amazon.}

33_{Margherita Colangelo, ‘Parity Clauses and Competition Law in Digital Marketplaces: The Case of Online Hotel} Booking’ (2017) 8(1) Journal of European Competition Law & Practice 3, 7.

34_{See § 0.24 of LEAR Report for OFT, ‘Can “Fair” Prices be Unfair? A Review of Price Relationship Agreements’} <http://www.learlab.com/publication/1145/> accessed 6 April 2019.

(13)

2.1.1. Barriers to entry (foreclosure of markets) and price collusion

If a platform has a significant market share consisting of a wide customer base that the platform’s sellers can access, it can impede access to the E-Commerce market to potential competing platforms. The only way how new platforms can enter the market is via lowering the final prices for goods/services, which can be achieved by lower commission fees for the sellers.36_{However, sellers are discouraged to charge lower prices on the competing platform}

due to the wide MFN clause with the incumbent platform. Since a single sale of the same product, at a lower price, on the competing platform immediately decreases the price at the incumbent’s platform, the sellers are not eager to risk large-scale price decreases at their major sales point. Therefore, a potentially competing platform has not much space how to attract sellers and consumers, and thus create positive network effects.37_{Clearly, this is an}

exclusionary effect of MFN clauses as it discourages the incumbent’s sellers from entering in discussions with market entrants (e.g. start-up platforms) as they would risk losing access to the incumbent’s customer base, due to repercussions arising from non-compliance with the MFN clause.38_{Also, even if entry occurs, the MFN clause may have the effect of distorting the}

entrant’s choice of the business model towards a model more similar to that of the incumbent, hence limiting the price competition.39

Furthermore, across-platforms MFNs soften competition between platforms, thereby increasing the commission fees paid by the sellers and consequently, the prices charged by the sellers to the buyers.40_{Since platform markets are mostly oligopolistic, they are very}

transparent, and incumbents have explicit knowledge of each other’s pricing strategies.41_Also,

the suppliers are unlikely to lower their price on any platform as this would result in a price reduction on all platforms benefiting from wide MFNs.42_{Indeed, MFNs have also collusive}

effect on competition as any change in price on one platform results in across-platforms uniformity. Ergo consumers are deprived of the benefits of active price competition between undertakings.

36_{Opinion France/OECD.(2015), page 6-8.}

37_{This is especially in the case of efficient entrants that have capability to offer sellers lower charges for} transactions, and hence secure lower prices for goods/services on their new platform. See Hviid and Fletcher (n 30), 73-74.

38_{Jonathan B. Baker and Judith A. Chevalier, ‘The Competitive Consequences of Most-Favored-Nation} Provisions’ (2013) 27(2) Antitrust Magazine 20, 24.

39_{Boik and Corts (n 31), 122-123.}

40_{LEAR Report (n 34), § 6.51. Also, see Hviid and Fletcher (n 30), 73.}

41_{Bertin Martens, ‘An Economic Perspective on Online Platforms’ (2016) Joint Research Center of the European} Commission, 3 < https://ec.europa.eu/jrc/sites/jrcsh/files/JRC101501.pdf> accessed 6 April 2019.

(14)

Foreclosure and collusion should be seen as mutually reinforcing in the case of platform MFNs - the clauses may incentivise collusion, which then stiffens the pricing schemes of the incumbent platforms, and thus precludes the smaller competitors from competing and potential competitors from entering the market.43

2.2. Positive effects on competition

Online platforms can be beneficial for consumers as they allow them to easily compare prices across the internet, check the reviews and gain necessary information regarding particular product/service they wish to purchase.44_{However, in order for online platforms to provide for}

these benefits, they must also ensure reasonable turnover on their investment as well as limit the risk of freeriding.45

2.2.1. Freeriding and the protection of investment in the platform’s service

Perhaps the most significant efficiency of MFN clauses is protection against freeriding on the platform’s pre-purchase services to buyers (reviews, product information, advice, etc.).46

Having an MFN clause contained in an agreement with the seller allows the platform to defend its quality investments by preventing other platforms from freeriding on them.47_{Clearly, in the}

absence of such clause, consumers could use the platform only to the extent necessary for pre-purchase information gathering and then resort to a less efficient but cheaper competitor for the actual purchase.

Such efficiency is even more critical when it comes to platforms operating on the basis of the agency model, meaning that they serve as a point of contact between sellers and consumers. In such cases, platforms are mere intermediaries, and freeriding concerns are much higher since once the seller and consumer are in direct contact, they can circumvent the platform’s charges by finalising transaction externally.48_{Consequently, the whole business model of the platform}

would be compromised if it was not for the MFN clause.

43_{Morten Hviid, ‘Vertical Agreements between Suppliers and Retailers that specify a Relative Price Relationship} between Competing Products or Competing Retailers’ (2015) OECD Competition Committee Paper DAF/COMP(2015)6<http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=DAF/COMP(20 15)6&docLanguage=En> page 23, accessed 5 April 2019.

44_{OECD France opinion (n 36), 7.} 45_{Colangelo (n 33), 5.}

46_{LEAR Report (n 34), § 0.37 and § 6.47.} 47_{Ibid, § 6.71.}

48_{Paolo Buccirossi, ‘Parity Clauses: Economic Incentives, Theories of Harm and Efficiency Justifications’ (2015)} Competition Law & Policy Debate 43, 51.

(15)

3. MFNs in E-book and Hotel Booking Platforms Cases

3.1. E-Book cases

3.1.1. Apple/Penguin E-Books Decision (2012)

This is the first case where the Commission dealt with MFNs used by an online platform in its contracts with third-party sellers. Interesting about this decision is that the Commission had opted for art. 101(1) TFEU concerning this double-sided market competition law issue. It was held that the agreements between the parties were cartel agreements and the parties agreed to give up on the anti-competitive agenda that was codified in them.

In terms of the facts of the case, certain events must be mentioned as well in order to understand the bigger picture into which this case undoubtedly falls. From 2007-2010, publishers sold e-books to retailers mainly under a wholesale model; thus, the retailers were in charge of deciding the final prices.49_{However, most of the publishers were at unease about the}

growing power of Amazon, which disrupted the profitability of the wholesale model on the part of the publishers. Amazon’s retail price at 9.99 for bestselling e-books was below the list price as well as below the wholesale price set by publishers.50_{As a result, the Five Publishers that}

had been under the Commission’s investigation have decided to work on a plan that would prevent the decrease of e-books’ retail prices, and by the same token stop Amazon’s growth on global markets, including the EEA. The publishers were aware of the fact that only a common approach against Amazon could fall on fertile land, and discourage Amazon from its pricing policy. This possible collusion raised the Commission’s suspicion which instituted an investigation into the Five Publishers’ practices.

Also, Apple Inc. planned to launch its new iBookstore app for which it needed to secure the best terms of sale in order to enter the e-books market successfully. Thus, in December 2009 Apple Inc. contacted the Five Publishers, each on an individual basis, regarding its intention to start selling e-books via its new platform and discussed possible sales models.51_The

Commission took the preliminary view that Apple at first considered entering the market under a wholesale model.52_{The reason for this assumption was that Apple has agreed to implement}

the agency model, only after such a proposition came from the Five Publishers, and was invited

49_{E-Books App no 39847 (Commission Decision, 12 December 2012), para 23. (hereinafter referred to as “Apple} E-Books Decision”)

50_{Ibid, para 26.} 51_{Ibid, para 32.}

52_{The very same sales model under which the Five Publishers had been operating with Amazon. See Apple} E-Books decision, para 34.

(16)

to propose retail prices for e-books.53_{Clearly, this model was beneficial for both sides – it}

secured Apple the elimination of meaningful retail price competition with Amazon, while the Publishers achieved their goal of raising retail prices above Amazon’s retail prices.54

Furthermore, Apple proposed draft agency agreements to the publishers that contained a retail price MFN clause.55_{In addition, the draft agency agreements contained maximum retail}

price grids that were set above the retail price charged by Amazon for new release e-books.56

These efforts to negotiate a priori the prices of future bestsellers indicate that Apple and the Publishers had an intention to collude on prices from the outset of their contractual relationship.57_{Also, the Commission found in its preliminary assessment that Apple has been}

facilitating the collusion between the Publishers by keeping the others informed about the negotiations.58_{This global strategy was introduced in EEA in 2010.}

Naturally, the Commission, in its legal assessment, opted for art. 101(1) TFEU as Apple’s and the Publishers’ conduct falls within its realm in the form of concerted practice. Indeed, the parties were directly and indirectly (via Apple) exchanging information regarding the initiation of the global strategy for the sale of e-books. This had the effect of the same prices and same commission fee the Apple Inc. would charge the publishers in future. When assessing whether a concerted practice is anti-competitive, regard must be paid in particular to the objectives which it is intended to attain and to its economic and legal context.59_{In addition,}

when deciding whether a concerted practice is prohibited by art. 101(1) TFEU, there is no need to consider its actual or potential effects once it is apparent that its object is to prevent, restrict or distort competition within the internal market.60_{Consequently, the Commission held that the}

objective of the concerted practice between Apple Inc. and the Five Publishers, in the economic context in which it had been pursued, had been to raise the retail prices of e-books in the EEA or prevent the emergence of lower retail prices for e-books in the EEA.61_{In light of the above,}

the Commission held that in order to achieve the common economic objective on a global basis,

53_{Apple E-Books Decision, para 34.} 54_Ibid.

55_{The retail price MFN clause provided that, in the event another retailer was to offer a lower price for a particular} e-book, including in situations where that retailer was operating under a wholesale model and thus was free to set retail prices, the publisher had to lower the retail price of that e-book in the iBookstore to match that other lower retail price. See Apple E-books Decision, para 35.

56_{Apple E-Books Decision, para 36.} 57_Ibid.

58_{Apple E-Books Decision, para 37.}

59_{Case C-8/08 T-Mobile Netherlands BV [2009] ECR I-04529, para 27.}

60_{Case C-105/04 P Nederlandse Federatieve Vereniging voor de Groothandel op Elektrotechnisch Gebied v}

Commission [2006] ECR I-08725, para 125; and Case C-209/07 Beef Industry Development Society and Barry Brothers [2008] ECR I-08637, para 16.

(17)

including the EEA, the parties to the agreement had to jointly convert the sale of e-books from wholesale model to an agency model with the same key terms62_{, ergo reshaping the whole}

industry. Clearly, this increased the publishers’ bargaining power vis-à-vis Amazon63_and

secured that lower retail prices would not be introduced in the EEA. Therefore, the Commission concluded that the concerted practice between Apple and the Five Publishers constituted a restriction of competition by object, which had been likely to have an appreciable effect on trade between MS within the meaning of art. 101(1) TFEU.64

In their commitments, the parties agreed to terminate the relevant agency agreements and abandon the use of retail price MFNs in their agreements.

Evaluation of the outcome of the case:

This case has rather unusual fact-pattern when it comes to MFN clauses used in the contested agreements. The clause constituted one component of a whole contractual complex, which as a whole, was designed to force a shift in the e-book industry to an agency model, thus putting the publishers in charge of retail prices.65_{Further, the purpose of the MFN clause, in}

this case, was quite unorthodox as it served not only the interests of the preferentially treated platform (Apple Inc.), but also the interests of the third-party sellers (the Five Publishers). In other words, the MFN clause acted as a “joint commitment device” – the Publishers could be sure that all of them would abide by the global strategy and pressure Amazon to switch to the agency model, ergo MFN clause was a tool to stabilise the cartel.

3.1.2. Amazon E-Book Decision (2017)

This is the second Commission decision addressing the e-books market and the use of MFN clauses. It differentiates from the Five Publishers case by the Commission’s resort to art. 102 TFEU approach in relation to MFN clauses articulated by Amazon. The case also ended up with Amazon’s commitments to cease its use of various types of MFNs in their contracts with the e-book suppliers as the Commission concluded their abusive effect.

As a result of the e-books market shift66_{from the wholesale model to the agency model,}

Amazon operates as an agent of the e-book suppliers and thus has no power over the prices of e-books sold by the third-party sellers. However, Amazon’s business model is not that straight

62_{Including the retail price MFN clause, the maximum retail pricing grids and the same 30% commission fee} payable to Apple Inc.

63_{See Apple E-Books Decision, para 91.} 64_{Ibid, para 98.}

65_{Daniel Zimmer and Martin Blaschczok, ‘Most-Favored-Customer Clauses and Two-Sided Platforms’ (2014)} 5(4) JECLAP 187, 198.

(18)

forward. It is a vertically integrated undertaking that is active upstream as a publisher, with its own imprints, and downstream as an e-book retailer for e-book suppliers, self-publishing authors and its own e-books.67_{Also, it manufactures its own e-book reading platform and tablet}

called Kindle.

Furthermore, the Commission had expressed concerns over the parity clauses in contractual agreements with independent E-book suppliers whose effect may constitute an abuse of dominant position, ergo breach of art. 102 TFEU. The contested MFNs were: Business Model Parity Clause, Selection Parity Clauses, Agency Price Parity Clause, Promotion Parity Clause, Discount Pool Provision and Notification Provisions.68

As the Commission has written in its decision, Amazon had created a contractual template that includes the above-mentioned parity clauses for its agency agreements, which serves as the basis for negotiations with e-book suppliers.69_{This suggests that Amazon was}

offering ready-made contracts, and MFN clauses had non-negotiable status. However, this would only be possible for Amazon, if it held a significant market share that would grant it a possibility of such dominance vis-à-vis the suppliers. The Commission, in its preliminary assessment, found that Amazon held a constant market share for retail distribution of English language e-books in the English-speaking region of 70-90% in the period 2010-2015.70

Similarly, the market share for the distribution of English language e-books, between 2011-2015, was 80-90% in the whole EEA.71_{Moreover, the Commission came to the same results}

when it examined the situation on the distribution market for German-language e-books in the EEA. It found that Amazon held a market share of 50-70% in 2015.72_{Also, Amazon’s potential}

market power was confirmed by the fact that from 2011 to 2015, most of the e-book sales from the suppliers in the EEA went through the Amazon platform.73_{In light of these facts, Amazon}

was, without a doubt far ahead of its closest competitor with its several times larger net turnover.74

Also, the Commission stated that Amazon’s strong bargaining position towards suppliers as well as its retail-level competitors was amplified by its “Kindle Ecosystem” through which it managed to secure customer lock-up as the consumers who own Kindle can

67_{E-book MFNs and related matters (Amazon) App no 40153 (Commission Decision, 4 May 2017), para 17.} (hereinafter referred to as “Amazon Decision”)

68_{For the Commission’s elaboration on each of these MFNs see Amazon Decision, paras 24-35.} 69_{Amazon decision, para 38.}

70_{Ibid, para 58.} 71_Ibid. 72_{Ibid, para 61.} 73_{Ibid, para 64.}

(19)

purchase e-books solely from Amazon’s Kindle store. This results in a situation where consumers who had already purchased Kindle would face costs in switching to another e-book reading platform.75_{Therefore, Amazon manages to maintain a large consumer base at e-books’}

market, which is an essential asset when it comes to concluding contracts with the suppliers. The Commission concluded that Amazon is an unavoidable trading partner for e-book suppliers, and was prepared to use that position to obtain terms and conditions, which its direct competitors were not able to obtain.76

Thus, the MFN clauses used by Amazon in contracts with the suppliers are an indication of Amazon’s exploitation of its superior position on the EEA’s e-book market. Firstly, the Business Model Parity clause has significant foreclosure effects at the e-book distribution level, which reduces the competitiveness of e-book retailers and harms the consumers as the less competition may result in increased prices and less choice. This clause allows Amazon to freeride on the alternative business models that have been proposed to e-book suppliers by Amazon’s competitors, which essentially discourages them from entering the market. Also, the e-book suppliers were unwilling to license their catalogue to star-ups and innovative entrants as they did not want to risk Amazon’s retaliation for breaching their contractual commitments under this clause.77_{Secondly, the Selection Parity Clause was held to reduce the innovation,}

quality and choice as it disincentivizes the book suppliers and retailers from developing e-books that are not primarily text-oriented. Essentially, it posed an obstacle for retailers to enter the market or expand their market share. Thirdly, the Agency Price and the Promotion Parity clauses and the Discount Pool Provision (together referred to as “Retail Price Parity Provisions”) were held to deter the expansion/entry of competing e-book retailers as they allowed Amazon to obtain higher commissions from the suppliers. Essentially, they strip the competing e-book retailers of the possibility to increase its market share by charging a lower commission to retailers and thus allowing the suppliers to attract consumers on lower prices.78

Fourthly, the Notification Provisions serve as a sort of a security mechanism that ensures practical fulfillment of the obligations arising from the previously mentioned MFN clauses. In other words, the suppliers are expected to notify Amazon every time there is a situation where their competitor is entitled to more favorable treatment. In this regard, the Commission has found indications that Amazon threatens to punish the suppliers who do not notify.79

75_{Ibid, para 65(2).} 76_{Ibid, para 67.} 77_{Ibid, para 89.} 78_{Ibid, para 122.} 79_{Ibid, para 138.}

(20)

Finally, the Commission concluded that the MFN clauses used by Amazon vis-à -vis the e-book suppliers have the potential effect of foreclosing competitors by preventing them from developing alternative business models and pricing systems in the internal market.80

Consequently, Amazon offered binding commitments and promised to terminate the disputed MFN clauses in its contracts with the suppliers.

3.2. Hotel Booking Platforms Cases (OTAs Cases):

3.2.1. Germany – B9 – 66/10 HRS Case (2013)

The HRS case was the first of its kind regarding the platform MFN clauses in hotel booking sector, and the only decision finding an infringement.81_{The Bundeskartellamt}

prohibited HRS from enforcing its MFN price parity clauses and ordered it to erase them from their contractual agreements and general terms vis-à-vis hotels by 1 March 2014.82_{The relevant}

market under consideration was an online hotel booking market in Germany. The main competitive concerns arising from MFN clauses were – reduction of competition between platforms, foreclosure of potential entrants to the online platform market and adverse effects on the competition between hotels.

An interesting aspect of this case is the way how sec. 1 GWB / art. 101(1) TFEU was applied. The German NCA held that MFN clauses used by HRS and its hotel partners are agreements between undertakings, which can impact trade between MS.83_{These MFN clauses}

obliged HRS’ hotel partners to grant HRS the best conditions in comparison to other online hotel booking platforms with respect to price, room availability and cancellation conditions.84

Next, Bundeskartellamt has examined the “genuine agent” doctrine that would deem art. 101(1) inapplicable. However, this doctrine was rejected as it was held that HRS is not a dependant agent of the hotels. The NCA’s line of argumentation was based on para. 18 of Vertical Guidelines according to which an agent has to bear the economic risk of his own for the application of art. 101 TFEU. Clearly, this was the situation in HRS as it bears its own financial and economic risk stemming from investments in advertising their brand,

80_{Ibid, para 157.}

81_{All decisions that came after were ended by the commitments.}

82_{B 9-66/10 HRS Bundeskartellamt Decision (3 Spetember 2013), para 13. (hereinafter referred to as “HRS} Decision”)

(21)

establishment of technical refinement or establishment of contractual network with a large number of hotels.85

Furthermore, Bundeskartellamt states that HRS cannot be a genuine agent because the MFNs do not produce restriction of competition emerging from the principal (the hotels) but rather from the agent (HRS itself). Also, the NCA explicitly states that it was the HRS that unilaterally imposed the general terms and conditions covering MFNs on the hotels and that hotels do not exert any influence on the activities of HRS.86_{Indeed, this line of argumentation}

is quite unorthodox under art. 101 substantive analysis. Using words such as “unilaterally” or that partners do not “exert influence” implies the existence of the platform’s market power and lack of bargaining power on the part of the hotels. Consequently, if hotels cannot express their demands, there is a lack of consensus, which is condition sine qua non for the conclusion of an agreement between undertakings. The Court has explicitly stated this in Bayer Case: “the proof of an agreement within art. 101(1) must be based upon the direct or indirect finding of the existence of a subjective element that characterises its very concept – a meeting of minds between operators.”87_{In other words, the Bundeskartellamt seems to use language that is more}

common to the art. 102 analysis.

Moreover, the economic analysis of the German market showed that the hotel booking sector is highly concentrated. HRS, Expedia and Booking hold together 90% market share, and all of them resort to similar MFN clauses.88_{Hence, it is strange why Bundeskartellamt had not}

adapted its substantive approach along with the examined market characteristics indicating market power concentration among the three undertakings.

Next, the decision was addressed only to HRS and not also to the hotel partners as is often the case with cartel agreements, nor was the decision addressed to Booking and Expedia whom the NCA found to use the same MFN clauses. Oddly enough, HRS is the only one that is punished under the provision that is designed to penalise the collective anti-competitive behaviour.

Clearly, the particularities of this case indicate that Bundeskartellamt chose quite an unfortunate route to rule on this case. The factual background as well as the market characteristics of German hotel booking sector indicate that art. 102 TFEU might have been more effective way to tackle MFN clauses. Also, it would save time and resources as, under

87_{Joined Cases C-2/01 P and C-3/01 P Commission v Bayer [2004] ECR I-00023, para 27 (hereinafter referred to} as “Bayer Case”).

(22)

art. 102, there would be only one proceeding involving all three major platforms, instead of three separate proceedings. The situation culminated in the 2015 Booking.com decision where Bundeskartellamt rejected the platform’s EEA uniform commitments to apply only narrow MFN clauses89_{, and thus prohibiting all forms of platform MFN clauses on the German}

market.90

3.2.2. France, Italy, Sweden – Booking.com Case (2015)

Booking.com was subject to simultaneous investigations in these three MS regarding the use of wide MFN clauses in contracts with their hotel partners. Due to a significant cross-border nature of this case, the Commission coordinated the proceedings, which eventually culminated in accepting the Booking’s commitment to applying only narrow MFN clauses. As a result, Booking.com had changed their contractual policy in whole EEA to involve only narrow MFNs.

Similarly, as Bundeskartellamt in HRS decision, the NCAs have based their decision on the substantive assessment under art. 101(1) TFEU. The main competition concerns they raised in their assessment were horizontal foreclosure on the online platform level and price restriction vis-à-vis the hotel partners.91_{Moreover, the NCAs approach is very much consumer-oriented,}

which eventually results in their acceptance of freeriding justification for narrow MFN clauses. They point out that hotels do not pay anything for the service until the consumer makes a booking; hence it would be detrimental to the platforms if they could not equalise prices that the same hotels offer at competing platforms. Clearly, this is in contrast with HRS decision where the freeriding justification for MFNs was rejected as unnecessary and unable to increase the quality of service.

Moreover, the NCAs had assessed competition law concerns on a horizontal level between the platforms but not that much in relation to vertical restraints towards hotels. It had been acknowledged that MFNs restrict their freedom to compete on prices to a certain extent, but if this issue is to be assessed under 101, it is expected that some assessment on that part would be provided, too. After all, the NCAs claim that there is a vertical agreement between Booking.com and the hotels.

89_{Resulting from the French, Italian and Swedish decision.}

90 _See _{Bundeskartellamt’s} _press _release _from ₂₃ _December ₂₀₁₅ _available _at <https://www.bundeskartellamt.de/SharedDocs/Meldung/EN/Pressemitteilungen/2015/23_12_2015_Booking.co m.html > accessed 25 April 2019.

91_{596/2013 Booking.com Sverige AB Konkurransverket Decision (15 April 2015), paras 21-23. (hereinafter} referred to as “Booking.com Sverige”)

(23)

Finally, the result of this case caused more discrepancy than uniformity concerning the competition law assessment of platform MFNs. The NCAs involved accepted the use of narrow MFNs as they provide for tangible advantages to consumers stemming from increased quality of service and supply of comparison services free of charge.92_{However, shortly after}

Booking.com implemented its commitments, the French legislator broke them by outlawing platform MFNs in all their forms, with Italians following the suit in 2017.93

4. EU Competition Law and Platform MFNs

After reviewing the landmark cases that indicate the application practice of competition rules towards platform MFNs, it is vital to assess the substantive differences between art. 101(1) and art. 102 TFEU. Clearly, in light of the case-law mentioned above the predominant theory of harm on which the competition authorities operate in relation to MFNs has been foreclosure94_{, rather than collusion, which is a predominant theory of harm when it comes to}

art. 101(1) TFEU. Thus, the competition authorities tend to argue foreclosure as the main competition concern but then base their decision on art. 101(1) TFEU that substantively revolves around collusion. It must be pointed out that foreclosure is mostly related to the abuse of dominance, as only undertakings with significant market power are capable of affecting market structure to such extent as foreclosure of competitors would occur. This is not to exclude the possibility of cartel’s collusive behaviour having foreclosure effects on the market, however, in such case, the sole foreclosure must be seen as a mere side-effect of the concerted practice, which in such situations has a different objective. Also, if the foreclosure is the central objective of the concerted practice, the undertakings need to jointly hold significant market power to bring about such result, and then one deals with abuse of collective dominance rather than a cartel.95

In light of the above, the following section will explain why art. 102 TFEU, along with its national equivalents, is more suited to regulate online platform MFN clauses.

92_{See in more detail at: European Competition Network, ‘The French, Italian and Swedish Competition Authorities} Accept the Commitments Offered by Booking.com’ <https://webgate.ec.europa.eu/multisite/ecn-brief/en/content/french-italian-and-swedish-competition-authorities-accept-commitments-offered-bookingcom> accessed 27 April 2019.

93_{See further at: Osborne Clarke, ‘Hotel Price Parity Clauses in Italy’ (12 October 2017)} <https://www.osborneclarke.com/insights/hotel-price-parity-clauses-in-italy/> accessed 27 April 2017.

94_{Particularly Amazon Decision and hotel booking decisions (HRS Decision and Booking.com Sverige).} 95_{Akman (n 4), 804. Also see HRS and Booking.com Cases.}

(24)

4.1. Substantive Assessment Under Art. 101 TFEU and its Inapplicability

When assessing platform MFN clauses under art. 101, one has to establish whether they fall within its scope. As has been established, the relationship between the sellers and online platforms is based on the agency model, which then legally means that agreements between them do not fall within the art. 101 realm, as they are not agreements between “two or more undertakings”.96

4.1.1. Scope Limitation – Agency Model

As has been established, the dynamics on which the online platforms cooperate with the sellers is the agency model. According to the Guidelines on Vertical Restraints, an “agent” is a “legal or physical person vested with the power to negotiate and/or conclude contracts on behalf of another person (the principal), either in the agent’s own name or in the name of the principal” for the purchase/sale of goods/services supplied by the principal.97_{The importance of realising}

that most of platform MFNs operate on agency model is crucial for the correct substantive assessment on the part of competition authorities, as disregarding this borders with ignorance of the provision’s spirit of the law. Consequently, the status of a platform as an “agent” signifies that art. 101 shall not apply to the platform MFNs since this article is inapplicable to agreements between undertakings that form a single economic unit.98_{The exact wording of art. 101 requires}

that there is an agreement between separate undertakings.99_{Therefore, if the platform and the}

sellers are to be considered as a single economic unit, art. 101(1) TFEU does not apply.100_In

addition, to fulfil the necessary element of “separation between the undertakings”, such separation shall not merely be formal. The examination of this requirement is assessed, as is the case in EU competition law, by the effects-based approach; thus, the decisive factor taken in consideration is the unity of their conduct on the market.101

Furthermore, the agreement between the online platform and the sellers is vertical by nature; hence, it must be assessed in light of the Guidelines on Vertical Restraints. Although Guidelines are soft law, they codify the Commission’s decision-making practice in relation to

96_{Art. 101(1) of Consolidated version of the Treaty on the Functioning of the European Union [2012] OJ C 326/47.} 97_{Guidelines (n 12), para 12. In addition, it is not material how the parties or national law qualify the agreement} in question for determining whether an intermediary is an agent; see Guidelines (n 12), para 13.

98_{Single economic unit doctrine was established in cases: Case T-11/89 Shell v Commission [1992] ECR II-757,} para 31; Case 73/95 P Viho Europe BV [1996] ECR I-5482, para 51 (hereinafter referred to as “Viho Europe”). Furthermore, in CEPSA Judgment it was upheld by the Court, that vertical agreements fall under the scope of art. 101(1) TFEU only if the operator/agent is regarded as an independent economic operator, which then implies an agreement between two undertakings. Case C-217/05 CEPSA [2006] ECR I-11987, para 38 (hereinafter referred to as “CEPSA”).

99_{This is also supported by the CJEU’s jurisprudence, see: Case 32/65 Italy v Commission [1966] ECR 389, 408.} 100_{Alison Jones and Brenda Suffrin, EU Competition Law: Texts, Cases and Materials (5}th_{ed, Oxford 2014), 773.} 101_{CEPSA (n 98), para 41.}

(25)

vertical agreements and thus clarify which agreements are not covered by the 101 prohibition.102_{In particular, para 18 of the Guidelines states that in case of agency agreements,}

“all obligations imposed on the agent in relation to the contracts concluded and/or negotiated on behalf of the principal fall outside art. 101(1) TFEU.”103_{Hence, all agreements between}

online platforms and sellers that establish agency model between them fall outside of art. 101 TFEU.104

In order to establish whether the agreement in question is categorised as an “agency”, the Commission looks at who carries the financial or commercial risk.105_{To phrase it another}

way around, an intermediary (agent) loses its agency status and art. 101 applies only when each of the undertakings bears their own commercial risk linked to the performance arising out of their agreement.106_{However, the CJEU has stated that even if the intermediary (agent) bears}

only a negligible risk, it does not render art. 101 applicable.107

As could be seen in the cases discussed in the preceding section, the online platforms predominantly act as intermediaries that earn their revenue via commission fee per each transaction. For example, in Booking.com/Expedia case, the OFT108_{noted that OTAs do not}

take the title of hold inventory to hotel accommodation, which suggests their position as agents.109_{Furthermore, some may posit that the platforms are not “genuine agents” because}

they bear their own financial and economic risk by investing in their brand advertising, establishing supply networks, investing in IT support for their functioning, etc.110_However,

Akman argues to the contrary by saying that – “such implication confuses the fact that the agency business has its own costs and risks itself as a business with the decision on whether the business acts as an agent of another party in a given transaction with a third party.”111_Hence,

the mere fact that the agency business has costs of its own cannot disqualify the business from being an agency.112_{Next, another mistake that occurs in decisional practice is immediate}

disregard of the agency model on the basis of a fact that the platform does not act in the interest

102_{Guidelines (n 12), paras 8-22.} 103_{Ibid, para 18.}

104_{Jones & Suffrin (n 100), 776.} 105_{Guidelines (n 12), para 13.}

106_{Examples of commercial and financial risk (on part of the agent) include: distribution costs, costs related to the} promotion stock maintenance, etc.

107_{See Guidelines (n 12) para 125, which explicitly states that the intermediary may bear some insignificant risks} but still be considered an “agent”.

108_{Predecessor of the CMA in the UK.}

109_{OFT1514dec OTAs Case OFT Decision (31 January 2014), para 4.3.}

110_{This was vocally argued by the Bundeskartellamt in para 148 of HRS Decision.} 111_{Akman (n 4), 808.}

112_{See Guidelines (n 12), para 15. The last sentence states: ‘stipulating that the risks related to the activity providing} agency services in general are not material for the assessment of “genuine agent”’.

(26)

of a single principal, but sources products/services of multiple principals at the same time, and thus must be regarded as an “independent agent who provides services on an entirely independent basis.”113_{Clearly, this is not the correct interpretation as the Vertical Guidelines}

stipulate that it is not material for the assessment whether the agent acts for one or several principals.114

Furthermore, the online platform and the third-party sellers indeed have a separate legal personality, and in most of the cases are not legally affiliated. In light of this argument, it could be claimed that art. 101 is applicable as we have two separate undertakings. However, such argumentation must be rejected as there is more to it than a mere first glance assessment of the legal relationship between the undertakings concerned. The CJEU held in Daimler Chrysler, that where an agent who has separate legal personality, does not independently determine his own conduct on the market and merely carries out the instructions of his principal, art. 101 TFEU does not apply to such a relationship.115_{Obviously, this is the case with the online}

platform sales – the suppliers source their products/services online and remain in charge of the development of transaction from the beginning until the end, and it is only upon the successful sale is made between the supplier and end consumer when the platform takes its share. For example, the Booking.com and Amazon do not bear any financial, nor commercial risks associated with the transaction – the principal (seller) decides all “parameters of the transaction”116_{with the consumer. Therefore, in light of the Guidelines and CJEU’s}

argumentation, the online platforms are agents of their sellers, for the purposes of competition law, in transactions with consumers as they do not bear any risk other than not receiving a commission.

However, even if online platforms are genuine agents, the Guidelines foresee two exceptions based on which the art. 101 could nevertheless be applicable – i) provisions that regulate the relationship between the principal and agent; and ii) agency agreement that facilitates collusion.117_{The first exception does not apply to the agency agreements examined}

by this paper as MFNs by their nature do not relate to the relationship between a platform and

113_{Case C-311/85 VVR v Social Dienst [1987] ECR 0381, para 20. Probably, similar logic was followed by} Bundeskartellamt in the HRS Decision, where they considered the fact that HRS has multiple principals as an evidence of independence, hence clear way for art. 101 TFEU. Indeed, it must be pointed out that this clash between jurisprudence and Guidelines does not enhance legal certainty.

114_{Guidelines (n 12), paras 13 and 15.}

115_{Case T-325/01 Daimler Chrysler AG v Commission [2005] ECR II-03319, para 88.} 116_{These are: price setting, delivery arrangements, cancellation on part of the consumer, etc.} 117_{Guidelines (n 12), para 16.}

EU COMPETITION LAW AND THE USE OF “MFN” CLAUSES IN E-COMMERCE