Tipping Markets

(1)

Department of Law

Spring Term 2021

Master’s Thesis in Competition Law

30 ECTS

Tipping Markets

An analysis of the Commission’s proposal for a Digital Markets

Act

Author: Hedvig Sylwan

(2)

(3)

2

Abstract

The particular characteristics and practices of digital platforms have given rise to the phenomenon of “tipping”, where one platform provider takes the whole market. There has been a growing concern among competition law regulators and enforcers that the traditional ex-post antitrust instruments cannot remedy the distortion of competition it causes in digital markets. In December 2020, the Commission published a proposal for a Digital Markets Act which aims to combat weak contestability and unfair practices in platform markets. The proposal includes a list of ex-ante rules, a merger information requirement and a market investigation tool that gives the Commission greater flexibility in designating the rules.

The research questions of this thesis concern the challenges of addressing tipping digital markets with existing competition rules and how the Digital Markets Act will address them. In answering the first question, it reviews the underlying factors leading to high market concentration and entry barriers, the implications of tipping markets for consumer welfare and innovation, as well as high-profile cases such as Microsoft, Google Shopping and Facebook/Whatsapp. The thesis finds that although monopoly-akin market structures may not always be detrimental to the objectives of competition law, the entrenched incumbent positions of companies such as Google, Apple, Facebook, Amazon and Microsoft are often a cause for anti-competitive concerns. Furthermore, the ex-post character of article 102 TFEU makes it ineffective to target tipping markets, and the merger rules are not adjusted to the dynamic and uncertain digital economy.

(4)

3

Abbreviations

CFI Court of First Instance

CJEU Court of Justice of the European Union DMA Digital Markets Act

DSA Digital Services Act

EC European Community

EU European Union

GAFAM Google, Apple, Facebook, Amazon, Microsoft GDPR General Data Protection Regulation

MFN Most Favoured Nation

NCT New Competition Tool

TFEU Treaty on the Functioning of the EU

(7)

6

1 Introduction

1.1 Background The 21st

century can be described as the digital era. During the last 20 years, large parts of the economy and society have moved online, where digital platforms have a significant influence on market forces. Despite the fact that the largest digital platforms are a heterogenous group – all focusing on different product and service areas – they share the trait that they are online spaces where buyers and sellers meet, with the platform controlling the relationship between them. For example, Facebook connects advertisers with the social media users; Amazon enables small scale businesses to reach their consumers; Apple displays developers’ apps for their iPhone users, and Google links search engine users with website owners. In other words, digital platforms facilitate access to critical services online. The platforms’ roles as intermediators between end-users and business end-users have led EU regulators and academics to refer to them as digital gatekeepers.1

Google, Apple, Facebook, Amazon and Microsoft (“GAFAM”) are in many ways akin to monopolies. Due to their unfettered dominance on their primary markets, they have been able to develop their services by integrating complementary products and services that eventually have become whole digital ecosystems. Undoubtedly, there are many advantages to digital ecosystems. Consumers may appreciate the simplicity and personalised service a one-stop shop concept can offer, and businesses may be able to reduce their prices and access detailed consumer data to improve their services. It could thus be argued that the rise of digital platforms has given many efficiencies in terms of technological advancements, products and new ways to reach consumers, friends and family. However, as the platforms have gained more ground, the criticism against them has also grown.

A complicating factor is that the structure of digital platform markets seems to lead to concentration naturally. Several interacting characteristics of digital platforms provide incentives to consumers and business users to join only one platform on the market, creating a “winner takes it all”-situation. This phenomenon is called market

1

(8)

7

tipping, as one market player with a gatekeeper position may cause the market to tip in their favour.2

A market that has tipped is challenging to reverse. The particular features contributing to the tipping process create high barriers to entry in markets that have tipped, predisposing long-lasting incumbents.3

A tipped market position allows a company to adopt practices capable of distorting the competition on the markets on which they operate. There is a broad consensus that the existing EU competition law acquis cannot effectively solve the structural problems in digital markets, leaving the gatekeepers with unchecked power.4

The Commission has announced a Digital Markets Act (“DMA”)5

to address these issues. The DMA is built on two pillars: the first one is a set of ex-ante rules applicable for digital gatekeepers that lists both prohibitions and positive obligations; the second is an enforcement mechanism called “market investigation” which would give the Commission further power to intervene in digital markets. The proposal has been met with both embrace and criticism, and if enacted, it will have an unprecedented impact on the digital markets.

1.2 Purpose and research questions

The tipping element of digital markets has redefined how competition works in the digital economy. Regulators and antitrust enforcers are required to learn and adapt to new competitive dynamics where the companies no longer compete in the market, but for the market. The DMA is a forceful response to this, as it establishes ex-ante rules and immensely increases the Commission’s powers to intervene in digital markets’ structures. This thesis therefore seeks to examine how the legislative proposal attempts to address both structural and strategic barriers to entry in digital markets and what the competition implications are. Accordingly, the questions which this thesis attempts to address are the following:

1. What challenges do tipping markets pose to existing EU competition law? a. Why do digital markets tend to tip?

2

Petit, “Schuman Shorts: What is a Tipping Market?”, October 28, 2020.

3

Ibid.

4

The Commission, Factual summary of the contributions received in the context of the open public consultation on the New Competition Tool, October 8, 2020.

5

(9)

8

b. What are the perceived inadequacies of article 102 of the Treaty of the Functioning of the European Union (“TFEU”)6

and the EU merger control concerning digital markets?

c. How can digital platforms’ leveraging and merger strategies hamper competition?

2. How does the DMA address tipping markets?

a. How does the market investigation tool allow the Commission to intervene earlier in digital markets?

b. What would be the DMA’s emerging gatekeeper rules’ effect on competition?

c. What would be the reviewed DMA rules’ effect on innovation?

1.3 Delimitations

Given that the subject of this thesis is defined as the DMA and tipping markets, only the DMA rules enabling the earliest intervention possible in digital markets will be examined (e.g. the rules applicable to emerging gatekeepers). Accordingly, the other rules exclusively applicable to companies qualifying as gatekeepers as per the DMA will not be reviewed. Neither will this thesis touch upon any issues relating to anti-competitive agreements or other concerted practices. As such, a presentation of article 101 TFEU is not included in the general overview of EU competition law in chapter 2.

Exploitative abuse in digital markets is a topic that has gained great attention in recent years, mainly thanks to the German Bundeskartellamt’s decision concerning Facebook’s use of user data.7

Some of the DMA’s ex-ante rules can indeed prevent digital platforms from engaging in exploitative abuse, but its emphasis is on exclusionary practices. This thesis will therefore only focus on the DMA from an exclusionary conduct point of view. To that end, platforms’ use of data will only be discussed to the extent it can have a foreclosing effect on competitors. The General Data Protection Regulation

6

Treaty on the Functioning of the European Union, OJ [2012] C 326/01.

7

(10)

9 (“GDPR”) 8

is in some respects relevant to the DMA and will be mentioned, but neither the GDPR nor the DMA will be discussed from a privacy perspective.

1.4 Methodology

1.4.1 The legal analytical method and the EU legal method

This thesis is not only devoted to shed light on the law as it is in its current state, but also to predict and evaluate how the EU competition law will be if the proposal is implemented. The methodology best suitable for these purposes is a combined form of the legal analytical method and the EU legal method. The legal analytical method is similar to the more commonly used legal dogmatic method in that it aims to establish where the law stands today, but it goes further by also analysing and criticising the law. A legal analytical analysis can also be based on material that does not in itself create applicable law.9

The method is thus appropriate for this thesis where the preparatory work for the DMA is fundamental and where any legal acts, documents, case-law and legal writing that are relevant will be used as sources. For instance, although US competition law differs from EU competition law in many aspects, literature on American antitrust and big tech will be included as it raises many of the same concerns as in the EU. Furthermore, news articles will be cited for general information about technology companies’ business developments and market shares.

The EU legal method is also similar to the legal dogmatic method but acknowledges that EU law is an autonomous legal system with its own methodology and norm hierarchy. The sources of EU law can be divided into primary law and secondary law. The main sources of primary law are the Treaty of the EU and the Treaty on the Functioning of the EU, whereas the secondary law involves regulations, directives, decisions, recommendations and opinions that derive from the Treaties.10

Preparatory work has admittedly low value as a source of law according to the EU legal method, but as the second research question concerns a legislative proposal, it is the best source available for understanding the legislator’s intentions. Moreover, the Commission’s

8

Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (GDPR) OJ [2016] L 119/1.

9

Sandgren, Rättsvetenskap för uppsatsförfattaren, 4th ed., Norstedts Juridik, 2018, p. 50.

10

(11)

10

central role in the field of competition law makes any Commission-issued soft-law instruments, such as communications or notices, valuable sources for this subject.

1.4.2 Economic theories relevant for the analysis

Competition law is inherently influenced by economic theory, making legal economic reasonings an inevitable element of this thesis. The below-mentioned theories will not be subject to any in-depth analysis but will be discussed on a more general level to point out the ideological dividing-lines and considerations actualised when regulating digital markets.

Two economic schools that will become relevant – and more elaborately introduced in section 2.1 – are the Ordoliberal school and the Chicago school. The Ordoliberalism is a German school of economic thought that advocates a state-regulated competitive process and a sceptic approach to accumulation of private economic power.11

The Chicago approach posits that economic efficiency will enhance the overall consumer welfare, a standard measured in metrics such as price, choice, quality and innovation.12

As the issue of regulating highly technological markets embodies a tension between innovation and competition, the analysis will also include considerations based on innovation economics. One of the most influential theories on innovation is Joseph Schumpeter’s hypothesis that the degree of innovation positively correlates with market power. Schumpeter posited that a market consisting of large, dominant firms is the most favourable structure in order to enhance innovation. However, he also claimed that monopolies can only be temporary as an incumbent can be displaced by any time by a more innovative firm. The idea is that smaller firms can completely eliminate an incumbent with the introduction of disruptive technology, thereby transforming the whole industry through an innovative paradigm shift called ‘creative destruction’.13

11

Ramírez Pérez, van de Scheur, The Evolution of the Law on Articles 85 and 86 EEC [Articles 101 and 102 TFEU]. The Historical Foundations of EU Competition Law, Kiran Klaus Patel and Heike

Schweitzer (ed.), 1–16.

12

See for example Bork, The Antitrust Paradox – A Policy at War with Itself, Maxwell Macmillian International 1993.

13

(12)

11 1.5 Outline

The essay is structured into six parts, beginning with a chapter providing an overview of the current EU competition law rules concerning abuse of dominance and merger control. The next part presents the particular characteristics of markets that tend to tip and demonstrates how digital firms can take advantage of those tipping-facilitating features. Thereafter, a discussion about the pro-competitive effects generated by tipping markets is presented in chapter 4. This involves a measure of the consumer welfare standard in the markets of GAFAM and an examination of the value of Schumpeter’s theory on innovation for today’s digital markets. Chapter 5 analyses the practical complications with tipping digital markets with the existing competition rules. The analysis touches upon issues relating to applying conduct-targeting rules on structural problems, lack of preventative enforcement measures and how digital incumbents can cement their entrenched position by aggressive merger and leveraging strategies. Next, chapter 6 evaluates the market investigation tool and the provisions applicable to emerging gatekeepers. The provisions reviewed relate to MFN clauses, restrictions on multi-homing interoperability, data portability, data access and merger information requirement. Finally, the thesis ends with conclusions.

2 EU competition law

This chapter provides an overview of current EU competition law. It reviews the objectives of EU competition policy and presents some classic abuse theories under article 102 TFEU. Further, it describes the jurisdictional merger rules and the potential concerns that may arise from horizontal, vertical and conglomerate mergers.

2.1 The objectives of EU competition law

(13)

pro-12

competitive, with the constant risk of competition law harming the very values it aims to protect. Therefore, it is vital to bear in mind the overarching objective of EU competition law when evaluating any enforcement action or legislation in this area of law. Nevertheless, there seems to be little consensus regarding the goal of EU competition policy, and the perception of what it is has evolved since the adoption of the EU treaties. This evolution is not completely linear, with the Commission and the Courts sometimes conveying inconsistent messages and with some objectives more prevalent in certain types of situations. That said, this thesis will now attempt to provide an overview of the multitude of objectives expressed in policy and case-law.

The competition rules’ original aim was to achieve market integration and protect an effective and undistorted competition. The first goal is explicitly stated in article 3(1)(b) TFEU, stipulating that the Union has exclusive competence in “the establishing of the competition rules necessary for the functioning of the internal market”. The latter objective has usually been interpreted as protection of competition in the market, where enforcement has often showed a preference for preserving a pluralistic market structure rather than the outcome of certain practices.14

The focus on the protection of the competitive process is usually considered as a demonstration of how influential the Ordoliberalism has been for EU competition law. The Ordoliberal school places a strong emphasis on consumer choice, so rivalry is considered a goal in itself.15

In 1990, the Chicago school of thought became more instrumental for EU competition policy.16

The theory advocates that competition policy should only protect the competition as such and not the competitors. With the growing influence from the Chicago school, the Commission adopted its own version of the consumer welfare model and the focus of EU competition law shifted from process to results. Consequently, the maximisation of consumer welfare has the last 20 years been commonly referred to as the

14

Case C-6-72, Continental Can, ECLI:EU:C:1973:22, para. 26 (“[t]he provision is not only aimed at practices which may cause damage to consumers directly, but also at those which are detrimental to them through their impact on an effective competition structure”), Case C-23/14, Post Danmark II,

ECLI:EU:C:2015:651 para. 72 (“In addition, since the structure of competition on the market has already been weakened by the presence of the dominant undertaking, any further weakening of the structure of competition may constitute an abuse of a dominant position”).

15

Behrens, The “Consumer Choice” Paradigm in German Ordoliberalism and its Impact upon EU Competition Law, Discussion Paper No 1/14, Europa-Kolleg Hamburg, March 2014, p. 1.

16

(14)

13 ultimate goal of EU competition law.17

This was, for example, clearly indicated by the statement by former Vice-President and European Commissioner for Competition Joaquín Almunia Amann in 2010, where he said that “[a]ll of us here today know very well what our ultimate objective is: Competition policy is a tool at the service of consumers. Consumer welfare is at the heart of our policy and its achievement drives our priorities and guides our decisions”.18

Still, the EU Courts have not fully embraced the consumer welfare standard.19

For instance, in Intel from 2017, the General Court established that article 102 is not aimed only “at practices which may cause damage to consumers directly, but also at those which are detrimental to them through their impact on an effective competition structure”.20

Finally, the rise of large digital conglomerates in the last decade has prompted a discussion on whether the objectives of EU competition law should be broadened. Damien Gerard has pointed out how frequently Margrethe Vestager mentions ‘fairness’ in her speeches, highlighting it as a potential ancillary goal.21

The fairness approach encompasses a focus on market structure and power – both economic and politically – and how these factors can stifle the start-up industry. This new competition ideology, which shares the interest of protection of market structure with the Ordoliberal school, could be viewed as contravening the consumer welfare objective, which is largely driven by the argument that competition policy should not protect competitors.

2.2 Abuse of dominance

Article 102 TFEU stipulates a prohibition of abuse of dominance and sanctions the unilateral behaviour of an undertaking. The finding of a violation requires a three-step analysis. Firstly, the relevant market must be defined, and secondly, the undertaking’s dominance of that market must be established. Thirdly, it must be found that the undertaking has abused its dominance.

17

Ezrachi, EU Competition Law Goals and the Digital Economy, June 6, 2018. Oxford Legal Studies Research Paper No. 17/2018, p. 4.

18

Almunia, Competition and consumers: the future of EU competition policy, speech, May 12, 2010.

19

The CJEU first mentioned consumer welfare in Post Danmark, para 22.

20

Case T-286/09 Intel, ECLI:EU:T:2014:547, para. 105.

21

(15)

14

2.2.1 Defining the relevant market and assessing dominance

The relevant market is defined in a two-pronged test, identifying the product market in the first step and the geographical market in the second. In the Commission’s notice on how to define the relevant market, the relevant product market is said to comprise “all those products and/or services which are regarded as substitutable by the consumer, by reason of the product’s characteristics, their prices and their intended use”.22

Thus, the substitutability is mainly assessed based on demand, e.g. which products the consumers would view as alternatives to each other. Demand substitutability is usually measured with a so-called SSNIP test, which examines the consumers’ reaction to a small but permanent price increase by the undertaking. If the consumers’ response to the price increase would be to switch to another product, the two products belong to the same product market.23

Another factor to consider when determining the relevant product market is supply substitution, where the supply-side’s reaction to the undertaking’s price increase is evaluated. Other suppliers that could switch production to the relevant product within a short timeframe and without incurring high additional costs should also be included in the relevant product market.24

Dominance is, as stated in Hoffmann-La Roche, a position of “economic strength enjoyed by an undertaking which enables it to prevent the effective competition being maintained on the relevant market by affording it the power to behave to an appreciable extent independently of its competitors, its customers and ultimately of the consumers”.25

A strong indicator of dominance is the existence of substantial market shares. As a rule of thumb, the Commission has stated that it is unlikely that an undertaking is dominant if it has a market share below 40%.26

However, a combination of other factors could also result in a finding of dominance where a firm has a smaller market share. The structure of the market is, in such instances, crucial for the dominance assessment, involving

22

Commission Notice on the definition of relevant market for the purpose of Community competition law [1997] OJ C 372/5, para. 7.

23

Commission’s notice on the definition of relevant market, para. 15–17.

24

Commission’s notice on the definition of relevant market, para. 20.

25

Case 85/76, Hoffmann-La Roche, ECLI:EU:C:1979:36, para 38–9.

26

(16)

15

dynamics such as the number and strength of competitors and barriers to entry.27

Factors such as superior technology, efficiency and economies of scale have also been determinative for a firm’s degree of market power in case-law.28

2.2.2 Abusive practices

Dominance in itself is not a violation of article 102 TFEU. Dominant undertakings are allowed to compete on their own merits; however, they are subject to stricter standards of competitive behaviour than other companies. In fact, a conduct which would be deemed as abusive if exercised by a dominant undertaking, could be regarded as pro-competitive if it is exercised by a new entrant. Article 102 TFEU provides a non-exhaustive list of examples of abusive practices. As a background to later chapters’ discussion on how the traditional theories of harm have evolved in the digital economy, a short presentation of some classic abusive practices is provided below.

2.2.2.1 Abuse by pricing

Article 102(a) states that abuse may arise by directly or indirectly imposing unfair purchase- or selling prices or other trading conditions. One exclusionary strategy prohibited by the provision is ‘predatory pricing’. Predatory pricing occurs when a dominant undertaking deliberately lowers its prices to a level where losses are incurred in the short term in an attempt to drive its competitors out of the market.29

The rationale behind the theory is that the long-term foreclosing effect enables the dominant undertaking to accrue a monopoly where it can comfortably raise its prices to a level that would harm consumers.30

Another type of pricing strategy that is of interest when assessing the DMA is price parity clauses, or Most Favoured Nation (“MFN”) clauses as they are more commonly known. An MFN clause refers to a contractual term where one party guarantees that the other party will enjoy the best price and conditions offered for the

27

Hoffman-La Rouche para 40.

28

Case 27/76, United Brands ECLI:EU:C:1978:22, para 122, Case T-65/89, BPB Industries plc, Commission’s Decision OJ [1989] L 10/50, para 116.

29

Guidance on the Commission’s enforcement priorities in applying article 82 EC, para. 63.

30

(17)

16

service or product in question. This means that if a supplier wants to sell its goods to a discounted price to a third party, the same deal must be given to the customer with whom the supplier has entered an MFN agreement.31

Price parity clauses have mainly been assessed under article 101 TFEU, but – as will be discussed later in this thesis – there are also cases related to digital markets in which the Commission and the national competition authorities have found them abusive under article 102 TFEU. MFN clauses can give rise to exclusionary effects by making it more difficult for new entrants to negotiate lower prices from suppliers. Thus, the MFN clause’s exclusion of price competition can raise high barriers to entry and reinforce the incumbent’s strong market position.32

The use of MFN clauses can be categorised as an article 102(a) abuse because it may lead to higher retail prices, since it limits the supplier’s discretion to offer discounts to certain customers. However, an MFN clause could also be regarded as abusive under article 102(b) as it could prevent the entry of new businesses on the market.33

2.2.2.2 Refusal to supply and refusal to license

Article 102(b) provides that it is abusive of a dominant undertaking to limit production, markets or technical development to the prejudice of consumers. The most common theories of abuse that fall under this provision are ‘refusal to supply’ and ‘refusal to license’, which both will be of relevance when discussing the conduct of digital platforms later in this thesis. An undertaking is generally free to choose its business partners under the principle of freedom of contract, meaning it has no obligation to accept a business offer. However, a dominant undertaking may have a duty to deal under article 102 TFEU if it is the only supplier. A refusal to supply may either be where an undertaking refuses to enter into a contract or when the undertaking quits supplying a business partner. The first EU case concerning a refusal to supply was Commercial Solvents.34

The Court found it abusive that Commercial Solvents, a producer of the raw material for manufacturing a particular drug, halted its supply to a trading partner as it had decided to start

31

Macrae, MFN Clauses, Fidelity Rebates, and Loyalty Programs in the Digital Era: Current Status, and a Legal Lacuna?, Competition Policy International, May 6, 2021.

32

Long, Retail MFNs and Online Platforms under EU Competition Law: A Practical Primer, Competition Policy International Antitrust Chronicle, September 2019, available at SSRN: https://ssrn.com/abstract=3549847, p. 8.

33

Ibid. p. 8.

34

(18)

17

manufacturing the drug itself. Commercial Solvent’s plan to enter the downstream market was not considered an acceptable justification. On the contrary, competition problems typically arise when the dominant undertaking is also active on the downstream market.35

Following Commercial Solvents, the Court has dealt with refusals to deal in several notable cases in which the criteria for finding a refusal abusive have been crystallised. In Magill, three television companies refused to license their weekly program listings in full to Magill, a publisher of a tv guide.36

Despite this, Magill included the whole week’s program listings, thereby violating the terms of the contract with the television companies. The case is of interest as it sheds light on the interplay between intellectual property law and competition law. The Court imposed a mandatory license of the weekly program listings despite the television companies’ copyright, by establishing that “the exercise of an exclusive right which, in principle, corresponds to the substance of the relevant intellectual property right may nevertheless be prohibited by Article [102] if it involves on the part of the undertaking holding a dominant position, certain abusive conduct”.37

The Court found that the refusal was anti-competitive as it prevented the appearance of a new product, there was no justification for such refusal and that the television companies excluded all competition on a secondary market since they denied access to material indispensable for carrying out business on that market.38

Bronner concerned whether the publisher of a daily newspaper could compel access to its competitor’s nation-wide home delivery scheme.39

The Court discussed the ‘indispensability’ criterion, stating that the output must be indispensable inasmuch as no actual or potential substitute exists.40

Further, such substitute does not need to be as advantageous as the requested facility.41

In order to be regarded as indispensable, the Court concluded that it was not enough to argue that it was not economically viable for a competitor to set up their own scheme by reason of the small circulation of their distribution, but that it must be shown that it was not economically viable to create a second home-delivery scheme for the distribution of a scale comparable to the dominant

35

Guidance on the Commission´s enforcement priorities in applying Article 82, para. 75.

36

Joined cases C-241/91 P and C-242/91 P, Magill, ECLI:EU:C:1995:98.

37

Magill para 101.

38

Magill para 54–56.

39

Case C-7/97, Bronner, ECLI:EU:C:1998:569.

40

Bronner para. 41.

41

(19)

18 undertaking’s.42

Hence, the ‘indispensability’ condition envisioned by the Court in Bronner had a high threshold.

Like Magill, IMS Health highlighted the tension between IP rights and competition.43

IMS Health was the biggest player on the German market for providing sales data on pharmaceuticals and had elaborated a certain brick structure for presenting the data. That brick structure had become widely adopted by pharmaceutical companies, pharmacies and doctors and was considered the industry standard.44

When IMS Health denied the competitor NDC Health a license to the brick structure, NDC Health filed a complaint to the Commission that the refusal constituted abuse under article 82 EC. In the following court case, the Court of Justice confirmed the case-law of Magill and Bronner.45

As regards the balancing of intellectual property rights versus free competition, the Court stated that the latter could only prevail where refusal to grant a license “prevents the development of the secondary market to the detriment of consumers”.46

The Court further concluded that to satisfy the condition that a refusal “prevents the appearance of a new product”, the undertaking which has requested a license will produce new goods or services and not only duplicate the ones already offered by the IP right holder on the secondary market.47

In conclusion, the Court has in the abovementioned cases established an ‘exceptional circumstances’ doctrine for assessing the anti-competitive effects of a refusal to supply: a refusal only constitutes abuse if (i) the requested output is indispensable for carrying on the business in question, (ii) the refusal prevents the appearance of a new product, (iii) the refusal is likely to eliminate all competition on a downstream market and (iv) objective considerations do not justify the refusal. Criterion (ii) only applies if the output is protected by an exclusive right.

42

Bronner para. 46.

43

Case C-418/01, IMS Health, ECLI:EU:C:2004:257.

44

IMS Health para. 6.

45

46

47

(20)

19 2.2.2.3 Discriminatory pricing

Article 102(c) states that applying dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive disadvantage, may be an abusive conduct. Such discriminatory pricing can occur when an undertaking uses different tariffs for different customers or only offers selected customers discounts or rebates. An undertaking can legitimately set its prices in a discretionary manner. However, the conduct is unlawful if it has an exclusionary effect either on the undertaking’s rivals (“primary-line injury”) or the undertaking’s downstream customers (“secondary-line” injury).48

An example of conduct causing primary-line harm is where the firm price discriminates as a strategy to compensate for the losses incurred by its predatory pricing in one of its segments.49

Secondary-line price discrimination is similar to refusal to supply and refusal to license as it involves a strategy to exclude competitors on the downstream market, with the difference that only some competitors are affected by the strategy.

2.2.2.4 Tying and bundling

Article 102(d) provides that an abuse may consist in concluding contracts subject to acceptance by the other parties of supplementary obligations which, by their nature or according to commercial usage, have no connection with the subject of such contracts. Such practice is also known as tying and bundling, which means that two different products are sold together. Pure bundling is when two products are only available to purchase as a bundle. Tying, on the other hand, is when one product is sold conditional on the purchase of another product. The subtle difference between tying and bundling is that in a tying arrangement, the tied product is available as a stand-alone product, whereas the tying product is not available independently.50

Like refusal to deal and refusal to license, tying and bundling is a theory of harm that has received much attention in the antitrust debate about digital markets.

Tying is a standard business practice and may in many instances yield consumer efficiencies such as lower costs, better products or simply convenience. It would simply

48

Papandropoulos, How should price discrimination be dealt with by competition authorities? Concurrences No. 3–2007, 34–38.

49

Ibid p. 36.

50

(21)

20

not make sense to not sell some products together, such as shoes and shoelaces. Therefore, the assessment of a tying practice’s compatibility with article 102 TFEU involves a “distinct products” test. Two products are considered distinct if there is consumer demand for both the tied and the tying product as stand-alone products.51

Using the example of shoes and shoelaces again, there is arguably no demand for purchasing shoes without laces and accordingly, these products are not distinct. However, if the products are distinct, a further requirement for enforcement is that the tying practice is likely to lead to anti-competitive foreclosure in the tied or tying market.52

For a long time, the Commission and the Court seemed to hold the view that tying by a dominant firm is per se prohibited under article 102 TFEU. This presumption has later been replaced with an effects-based analysis where an overall assessment of the practice is required.53

A clear-cut example of tying is Tetra Pak II, where the packaging company had contractually tied the sale of cartons with the sale of carton-filling machines.54

The Court of First Instance rejected Tetra Pak’s argument that the machines were indivisible from the cartons since companies that manufactured cartons designed for use in machines manufactured by other companies existed.55

Therefore, the tied sale of machinery and cartons could not be considered to be per commercial usage and were not inseparable.

What tying and bundling and refusal to supply have in common is that they are both sub-categories to the more general theory of harm of leveraging. Leveraging refers to when an undertaking uses its power in one market to strengthen its position in another one. The traditional theory of harm is based on the assumption that a monopolist can establish dominance on another market by using its market force in the primary market and that two monopolies generate more economic damage than just one.56

Chicago scholars have for a long time objected to the view that there would arise any anti-competitive effects from a dominant firm’s leveraging. They posit that a monopolist cannot raise its prices in the complementary market without losing profits in the origin market, meaning the leveraging cannot cause any further economic harm than what is already occurring due to the monopolist’s position on the original market. The defensive

51

Guidance on the Commission´s enforcement priorities in applying Article 82, para. 51.

52

Guidance on the Commission´s enforcement priorities in applying Article 82, para 50.

53

The Computing Technology Industry Association, Competition, Competitors, and Consumer Welfare: Observations on DG Competition’s Discussion Paper on Article 82, February 2006, p. 18.

54

Case T-83/91, Tetra Pak II, ECLI:EU:T:1994:246.

55

Tetra Pak II, para. 82.

56

(22)

21

leveraging theory of harm offers a third view. The theory suggests that leveraging is not mainly a strategy to conquer a new market but to prevent erosion of the primary monopoly. Accordingly, despite the anticipated profit loss in the secondary market, the incumbent has incentives to leverage as it deters entry into the primary market and protects the incumbent’s overall profits.57

An example can illustrate the theory: imagine a cinema company, which fears that movie streaming will replace cinema in the future and therefore launches its own streaming service. Through the streaming service, the company promotes movies that are only shown in cinema by giving cinema discounts to its streaming customers. Thus, the company has secured its monopoly if the cinema market would disappear but has also managed to channel customers from the streaming market to the cinema market. The different theories about a firm’s motivation to engage in leveraging will be revisited in the analysis of the DMA.

2.2.3 Public enforcement

The presentation of article 102 TFEU will now be concluded with a few words about its public enforcement.58

The provision is enforced by the Commission and the Member States’ national competition authorities, with the Commission’s investigative and decisional powers set out in Regulation 1/2003.59

In abuse of dominance proceedings, the Commission is responsible for proving the infringement and to respect the fundamental procedural rights of the parties that are subject to a decision. These rights include, among others, the right to be heard and the right to careful and impartial examination.60

If the Commission ultimately decides that article 102 TFEU has been breached, it can impose remedies or fines. A fine can be up to 10% of the undertaking’s total turnover of the preceding business year.61

57

Ibid. p. 2083, 2092.

58

It should be noted that private enforcement is also possible under article 102 TFEU, but it has been left out from this section as it is not relevant for the purposes of this thesis.

59

Council Regulation (EC) No 1/2003 of 16 December 2002 on the implementation of the rules on competition laid down in Articles 81 and 82 of the Treaty, OJ [2003] L 4/1.

60

Although undertakings are not natural persons, the companies that are scrutinised still benefit from rights expressed in the Charter of the EU. The respondent’s right to be heard was, for example, confirmed in C-204/00 P, Aalborg Portland and Others v Commission, EU:C:2004:6 para. 66, and the right to careful and impartial examination in C-269/90, Technische Universität München, EU:C:1991:438, para. 14.

61

(23)

22

The combination of the high evidentiary requirements, the number of fundamental procedural rights and the sometimes intricate nature of infringements can, in some cases, require long investigation times. Traditionally, the Commission’s investigations have taken a few years. For instance, the Magill decision came three years after the initial complaint, and the IMS Health decision only one year after.62

In contrast, the investigation in Tetra Pak II took a full eight years, as it required an in-depth study of Tetra Pak’s milk-filling machines.63

Hence, the investigation time will mainly depend on the complexity of the case.

2.3 Merger Control

Corporate mergers and acquisitions – jointly referred to as concentrations – can entail many advantages, not only to the undertakings involved but also to the consumers. Combining the forces of two companies can streamline and speed up innovation, and the business may be able to reduce consumer prices as economies of scale generally have lower average costs. A concentration may however also raise anti-competitive concerns if it would lead to a substantial lessening of competition. The third pillar of EU competition law – with the Council Regulation (EC) No. 139/2004 (“the Merger Regulation”)64

as its main legislative text – therefore targets mergers that would impede competition.

2.3.1 Jurisdiction

Mergers which fall in the jurisdiction of the Merger Regulation are subject to a requirement of mandatory notification to the Commission, which has exclusive competence to review whether such mergers are incompatible with the common market. The Merger Regulation applies to concentrations that have a Union dimension, e.g. mergers between undertakings whose aggregate turnover reach the thresholds indicated in article 1. Article 1(2) establishes that a concentration has a Union dimension where (i) the combined aggregate worldwide turnover of all the undertakings concerned is more

62

Magill, Commission’s Decision OJ [1988] L 78/43, para. 5, IMS Health, Commission’s Decision OJ [2001] L 59/18, para. 5.

63

Tetra Pak II, Commission’s Decision OJ [1991] L 72/1, para. 5.

64

(24)

23

than EUR 5000 million, and (ii) the aggregate Union-wide turnover of each of at least two of the undertakings concerned is more than EUR 250 million. However, there is no Union dimension if each of the undertakings concerned achieve more than two-thirds of its aggregate Union turnover within one and the same Member State. Article 1(3) provides secondary thresholds for when a concentration is deemed to have a Union dimension, despite not satisfying the primary thresholds in article 1(2). The purpose of the thresholds is to allocate cases between the Commission and the national competition authorities in accordance with the principle of subsidiarity; unless the merger has a Union dimension, the national competition authorities are better suited to protect the competition interests of the Member States.65

According to article 4(5), the Commission may also review a merger without a Community dimension if the merger is notifiable under the national competition law of at least three Member States. The merger can consequently be referred to the Commission from the national competition authorities. A similar referral mechanism is provided by article 22, which states that one or more Member States may request the Commission to examine any concentration that does not have a Union dimension within the meaning of article 1 but affects trade between Member States and threatens to significantly affect competition in the Member State or Member States making the request. The provision is known as “the Dutch clause”, as it was initially introduced 1989 to address the lack of national merger control rules in the Netherlands and some other Member States. Since then, all Member States have implemented national merger control regimes, and the Commission has therefore historically discouraged article 22 referrals.66

However, for reasons that will be explained in chapter 6, the provision is likely to be invoked more often in the future.

2.3.2 Review of the merger’s effect on competition

Once Union jurisdiction has been established, the merger’s compatibility with the internal market must be assessed. A concentration is not compatible with the common market if it would significantly impede effective competition, in particular as a result of the creation or strengthening of a dominant position.67

The creation of dominance is not

65

Recital 11 of the Merger Regulation.

66

von Koppenfels, A Fresh Look at the EU Merger Regulation? The European Commission’s White Paper “Towards more Effective EU Merger Control”, Liverpool Law Rev, April 29, 2015, 7–31, p. 26.

67

(25)

24

anticompetitive as such, but by prohibiting mergers that would create a market configuration that may enable anticompetitive behaviour, the Commission can avoid the risk of competition impediments. Merger control thus has a forward-looking perspective, as the Commission must compare the competitive conditions that are expected to result from the merger with the conditions that would have been without the merger.68

The nature of the Commission’s appraisal is therefore of more abstract character than its examinations under article 101 or 102 TFEU, as it involves speculations about the merger’s potential effects on the market.

Horizontal and non-horizontal mergers give rise to different sorts of competitive concerns. Horizontal mergers, where two competitors become one entity, may impede the effective competition by removing some of the competitive pressure from the market. Without the competitive constraints exerted from rival firms on the market, a company can raise its prices without losing sales to competitors.69

Non-horizontal mergers include vertical mergers and conglomerate mergers. A vertical merger occurs when two companies on different levels in a production chain become one entity, whereas a conglomerate merger is a concentration of two companies that neither belong to the same market nor the same supply chain. Hence, conglomerate mergers are mergers between firms that have no existing or potential direct relationship. The markets may instead be indirectly connected, e.g. by their offering of complementary products or products which belong to the same product range but are not sufficiently substitutable to exist on the same relevant market.70

The potential anti-competitive effects of non-horizontal mergers are less straightforward than of horizontal mergers. As a general rule, anti-horizontal mergers mostly entail efficiencies. However, a vertical merger may result in foreclosure effects if it hampers competitors’ access to markets or supplies, resulting in raised costs for competitors in the downstream market or lower revenues for rival firms in the upstream market.71

The conglomerate theory of harm mainly concerns the risk that the merged entity will engage in abusive leveraging practices with market foreclosure as a result.72

68

Commission’s Guidelines on the assessment of horizontal mergers under the Council Regulation on the control of concentrations between undertakings [2004] OJ C 031/5, para. 9.

69

Commission’s Guidelines on the assessment of horizontal mergers, para. 24.

70

Neven, The analysis of conglomerate effects in EU merger control, December 2005, p. 5.

71

72

Tetra Laval/Sidel (COMP/M.2416) Commission’s Decision OJ [2001] L 43/13, para. 364–365,

(26)

25

A particular challenge in appraising a merger’s competition implications is that one must take both actual and potential competition into account. The term potential competition refers to undertakings that are not currently on the market but are likely to launch a product or service that will compete sometime in the foreseeable future. Potential competition is relevant both when assessing whether the merging companies may be potential competitors and in terms of barriers to entry created by the merger.73

In its guidelines on horizontal mergers, the Commission states that a merger between potential competitors can only have anti-competitive effects where two conditions are fulfilled: first, the potential competitor must already exert significant competitive constraints or there must be a high probability that it will become an effective competitive force. Second, there must not be a sufficient number of other potential competitors on the market.74

3 Tipping-facilitating features of digital markets

This chapter explores the definition of tipping and the characteristics of digital markets that make them prone to tip. It establishes that these certain features are inherent in the digital platform business model but that digital platforms can also actively induce them, making the barriers to enter the market both structural and strategic. Moreover, it concludes that it is difficult to identify which markets will tip.

3.1 Definition of tipping markets

Tipping is not a new concept. In the Commission’s guidance on exclusionary abuse from 2009, tipping is mentioned as a relevant factor to consider when assessing anti-competitive foreclosure.75

As a concept, tipping is quite self-explanatory – it describes the dynamic process of how a market is tilting more and more to one platform’s end until it finally tips over in favour of that platform. Still, there is no clear consensus on what the concept entails and how to identify a tipping market. The tipping effect is usually described as a result of network effects, but there are many other concurrent elements that

73

Council Regulation (EC) No 139/2004 of 20 January 2004 on the control of concentrations between undertakings (the Merger Regulation) OJ [2004] L 24/1, para. 49.

74

75

(27)

26

can contribute to the process that will be presented in this chapter. Accordingly, there is a bundle of reasons for why markets tip, and it is not clear which factors have more weight than others. Hence, the competition law community lacks a simple and commonly accepted test for measuring tipping.76

What constitutes a tipped market may be an even more difficult question. A general answer would probably be that a tipped market is a monopoly-like market with such high barriers that it is unlikely to self-correct.77

If one were to articulate a legal test for ascertaining when a market has tipped, concentration would of course be a necessary prerequisite, but that does not paint the whole picture. Additionally, a tipped market would exhibit the characteristics that will be introduced below as making a platform prone to tip, such as strong network effects, lack of multi-homing and lock-in effects. However, these factors do not help in distinguishing a tipped market from a tipping market. Perhaps the concept of a tipped market is most easily understood by looking at examples of tipped markets. Assumingly, Google, Apple, Facebook and Amazon are considered prime examples of companies that have tipped at least their origin markets, as they have all dominated their markets for a long time and undoubtedly have a significant impact on the internal market in the EU. At the same time, it is well-known that Facebook is facing a serious threat from the relatively new social media app TikTok, especially in the younger consumer segments.78

Possibly, this is illustrative of how widespread the misconception of what a tipped market is, where big is commonly confused with insurmountable entry barriers.

3.2 Multi-sided platforms

A prominent feature of digital gatekeepers is that they are multi-sided platforms, or, as also commonly described, platforms operating on multi-sided markets. The concept of multi-sided platforms and markets does not have a clear and uniform definition. Therefore, two different definitions that both can contribute to a deeper understanding of this distinct business model are presented in the following.

76

Petit, Moreno Belloso, A Simple Way to Measure Tipping in Digital Markets, Promarket, April 6, 2021.

77

Tierney, Tipping in digital platform markets, Concurrences (webinar) September 21, 2020.

78

(28)

27

The first definition is provided by the economist Mark Armstrong, who describes two-sided markets as markets where ‘two or more groups of agents interact via intermediaries or “platforms” and where one group’s benefit from joining a platform depends on the size of the other group that joins the platform’.79

His definition focuses on the existence of indirect network effects, which will be further described below.

The second definition is provided by Rochet and Tirole, who make a further distinction than just network effects. According to them, “A market with network externalities is a two-sided market if platforms can effectively cross-subsidise between different categories of end-users that are parties to a transaction”.80

Hence, what distinguishes a two-sided platform from a one-sided platform is the skewed allocation of transaction costs, as one side accounts for a larger share of the total costs. In multi-sided markets, one side of the market is generally given preferential treatment at the expense of the other side. For instance, Facebook’s advertising revenues from business users enable it to offer zero-price to the social media users, and Google offers its search engine service for free as it makes money on advertisement and data analytics. The zero-price charge on one side of the platform has a significant effect on the competition as it precludes the possibility for new entrants to implement an aggressive pricing strategy to establish itself on the market.

The above definitions highlight that the business model of two-sided platforms has certain traits, such as network effects and zero-pricing, that often result in high entry cost. Thus, two-sided markets have high barriers to entry per definition.

3.3 Network effects

Network effects are generally referred to as the single most important reason why markets tip. The effect can be direct or indirect, where direct network effects occur when the convenience of using a product or service increases with the numbers of users that adopt it. Direct network effects are most apparent in social media platforms or other communication mediums, such as direct messaging apps.81

Indirect network effects are

79

Armstrong, Competition in two-sided markets, RAND Journal of Economics vol 37 no 3 2006 pp 668-691.

80

Rochet, Tirole, Platform Competition in Two-Sided Markets, European Economic Association 2003, 990– 1030, p. 1017 f.

81

(29)

28

often exhibited in two-sided markets, where one side of the market becomes more valuable with the increase of users participating on the other side. Consumers are, for instance, more likely to purchase a smartphone that offers the most apps, and businesses will focus on developing apps that are compatible with the most popular smartphone. Platforms that have built entire digital ecosystems thus benefit from strong indirect network effects.82

A concept related to network effects that is usually used in the tech start-up world is ‘critical mass’, which is the term for the level of users required to sustain growth. Only after critical mass has been achieved can a platform’s viability be established, making it a crucial turning point for business models relying on network effects. However, there is no clear method for measuring critical mass, and the number of users needed depends on the type of network.83

The strength of a network cannot be solely measured based on user numbers. Instead, the most important factor seems to be the level of connectedness between the users. A dating app will, for instance, have poor network externalities if the users are spread out over the world but will thrive if the network is developed into local clusters. However, a platform such as Airbnb will gain more on having a global cluster since consumers usually travel outside their cities.84

This illustrates that an analysis of the power of a network must be qualitative and take into account the particular business model at issue.

3.4 High economies of scale and scope

Due to the critical value of a large user base, most digital businesses prioritise growth over profit initially. Therefore, it is common for digital platforms to make losses for a long period in the high-growth phase. The sunk costs of entry are accordingly high, but the average costs reduce significantly as the company grows. The non-linear revenue curve is due to the digital platform being an asset-light business model where the cost of offering digital services does not increase proportionally to the number of users.85

82

Ibid. p. 1.

83

Petit, Are “FANGs” monopolies? A theory of competition under uncertainty, Working Paper, October 2019, p. 20-22.

84

Iansiti, Zhu, Why Some Platforms Thrive and Others Don’t, Harvard Business Review, January-February 2019.

85

(30)

29

Businesses in traditional markets may also experience cost advantages when scaling their operations – the cost per produced book is generally cheaper if a company produces a million copies than a thousand. However, as the marginal cost of a new customer usually approaches zero after hitting the critical value for a digital platform, the high returns on scale is a typical trait for digital markets. To challenge a firm that benefits from high economies of scale, a new entrant must often enter the market on a comparable high scale from the outset. The high economies of scale combined with indirect network effects reinforce each other and create strong incumbents that are difficult to displace, making companies less inclined to enter the market.86

3.5 Data

Control over data is key to success in digital markets. By analysing the data, platforms can improve their services and target the consumers with personalised advertising. The platform can further monetise the data by selling it as customer insights to business users on the other side of the market.87

User behaviour analytics and predictive analytics based on enormous data sets are referred to as Big Data analytics. Big Data is subject to increasing returns to scale: network effects combined with control over data create a positive feedback loop where the more users a platform has, the more data it can collect, and the more data obtained, the more it can cement its dominant position.88

Google’s search engine is an illustrative example of this reinforcing effect of data. The search algorithm provides more accurate results for every additional search made, thus attracting new users to the search engine over time. Consequently, as proven in a study conducted by the economists Prufer and Schottmüller, data-driven markets will almost always tip eventually.89

It should be noted that the digital businesses’ use of data is to some extent regulated by the GDPR. The GDPR is a privacy regulation that protects individuals’ right to their personal data, but it does not directly address data-related competition issues.

86

Hacaup, A German Approach to Antitrust for Digital Platforms, Digital Platforms and Concentration, Second Annual Antitrust and Competition Conference, Stigler Center, 2018, p. 8-10.

87

The Furman report p. 23.

88

Competition policy for the digital era, p. 106.

89

(31)

30

However, as will appear in this paper’s analysis of the DMA, some of the principles set out in the GDPR can also counteract barriers to enter data-driven markets.

3.6 Lock-in effects

Network effects and a platform’s access to data can lead to lock-in effects. Where network effects are prevalent, the individual incentives for a user to switch platform are low –for example, there is no point in using a new social media app if no one else uses it. Hence, an entrant must convince a larger masse to move to its new platform.90

Furthermore, the platform provider’s access to data can prevent users from migrating to a new platform since it would entail a loss of personal data or business data.91

A business user may, for instance, lose all its reviews or customer insights if it would switch platform. As a result, a company can maintain its monopoly power despite providing an inferior product. An entrant may offer a better technology or price, but the users will not switch to it as long as the incumbent has the data and a larger network. Thus, the switching costs make it more difficult for new firms to challenge the dominant company on the market.

3.7 Lack of multi-homing

When several providers offer similar two-sided platforms, the users on each side may choose to only use one platform (“single-homing”) or several (“multi-homing”). A good example of platforms with multi-homing on both sides is hotel booking platforms, where both hotels and travellers tend to use multiple reservation sites. Conversely, an example of platforms where one side usually single-homes is web search engines. Generally, two-sided platforms with multi-homing on each side do not give rise to any competition concerns. If the users on each side can switch platform, the platforms must compete to attract the users and, consequently, offer competitive terms. The platforms will accordingly have to co-exist and there will be less risk of the market tipping. However, a two-sided platform where one side multi-homes while the other side single-homes may become a “competitive bottleneck”, as described by Armstrong.92

In such situations, the multi-homing agents must use the platform as it is the only way to reach the

single-90

Competition policy for the digital era, p 23.

91

Furman report p 36.

92

(32)

31

homing users. The platform can then charge the multi-homing side higher prices, knowing the multi-homing agents will continue to use the platform because of its access to the single-homing users. One side’s single-homing will thus influence the other side, with the consequence that the multi-homing side is exploited, and the single-homing side is aggressively targeted.

What factors determine if the users will single-home or multi-home? Usually, this will depend on how costly it is for users to multi-home, meaning lock-in effects have great significance in this aspect.93

In the market for mobile applications, the consumer side rarely multi-homes as most people only have one telephone and one computer, and consumers tend to stick to the same operating system for all their devices. In comparison, the business users may be more likely to multi-home, although the interoperability necessarily requires an alteration of the actual product. Hence, even if most applications work with both Apple and Android today, developers would likely hesitate to convert an app to a new operability system with few users on the consumer side. Sometimes, the lack of multi-homing can simply be explained by the end-user’s convenience. Web search engines are an example of this, as the cost of switching platforms is minimal – switching to another search engine than the default choice does not take more than an extra five seconds and changing the default setting less than a minute. Therefore, the switching costs must be assessed in light of the relevant business model.

3.8 Lack of differentiation

Differentiation in the products or services offered on the market can make it less prone to tip. A differentiated market is more resilient to tipping since the consumers are not interested in the biggest network, but the network best suited to their preferences.94

Revisiting the example of online dating services, this is also illustrative of a highly differentiated market. Some online dating services differentiate themselves from the rest of the actors on the market by targeting a narrower customer base, such as Jdate, a website exclusive for Jews. Others have successfully distinguished their services by tweaking the dating premises, like how Bumble only lets women make first contact.95

Accordingly, a

93

Duch-Brown, The Competitive Landscape of Online Platforms, JRC Digital Economy Working Paper April 2017, p. 6.

94

Parker et al., Tipping Digital Markets, Frontier Economics, 2021, p. 3.

95

Tipping Markets

Department of Law

Spring Term 2021

Master’s Thesis in Competition Law

30 ECTS

Tipping Markets

An analysis of the Commission’s proposal for a Digital Markets

Act

Author: Hedvig Sylwan

Abstract

Table of Contents

Abbreviations

1 Introduction

2 EU competition law

3 Tipping-facilitating features of digital markets