Doing Business in a World of Goliaths: Power Imbalances and Economic Dependency in Platform-to-Business Relations

• 1 Introduction

  Within the context of a consultation on online platforms launched in 2016 by the European Commission, responding small- and medium-sized enterprises, despite acknowledging the increasingly essential role played by online platforms in the economy at large, raised several concerns on the fact that significant bargaining power imbalances between platforms and business users lead to unfair practices.¹x Commission’s Synopsis Report on the public consultation on the regulatory environment for platforms, online intermediaries and the collaborative economy, https://digital-strategy.ec.europa.eu/en/library/results-public-consultation-regulatory-environment-platforms-online-intermediaries-data-and-cloud (last visited 1 February 2024). The following year, based on a survey, the Commission published a report on business-to-business (B2B) relations in the online platform environment, identifying six critical areas where online platforms more frequently engage in unfair practices to the detriment of business users, namely the unilateral imposition of terms and conditions, flawed and untransparent search and ranking systems, access to the platform (including termination of an account), self-favouring behaviours by vertically integrated platforms, data access and portability, and lack of redress possibilities.²x European Commission, Directorate-General for Internal Market, Industry, Entrepreneurship and SMEs, ‘Business-to-business relations in the online platform environment – Final report’, Publications Office, at 24 (2017), https://data.europa.eu/doi/10.2873/713211 (last visited 1 February 2024).
  As these reports show, although online platforms have established themselves as essential intermediaries in today’s economy by facilitating interactions and transactions between different groups of users, they are affected by several sources of market power, such as indirect network effects, economies of scale and economies of scope. Due to the interplay between these sources of market power, online platforms tend to acquire dominant or quasi-dominant positions in the market(s) in which they operate, which translates into a stronger bargaining position towards their business users, as the digital platform often represents for the latter the only effective access to the intermediated market.
  As the consultation launched by the Commission shows, online platforms could abuse the superior bargaining power to engage in unfair practices, such as, for instance, imposing unfair terms and conditions or denying access to the platform, favouring their products vis-à-vis those of the business users, or, again, excluding business users by deactivating their account. These practices could, in some cases, have fatal consequences on the business user, as exclusion from the platform may also mean exclusion from the (online) intermediated market.
  Some of these practices are increasingly being scrutinised by courts and competition authorities, both at the national and European level.³x For an overview of online platforms unilateral practices, see H. Schweitzer and F. Gutmann, ‘Unilateral Practices in the Digital Market: An Overview of EU and National Case Law’, Concurrences 101045 (5 July 2021), https://www.concurrences.com/en/bulletin/special-issues/unilateral-practices-in-the-digital-market/new-article-no101045 (last visited 1 February 2024). For instance, with respect to the imposition of unfair terms and conditions, the Commission opened in 2020 antitrust proceedings against Apple for imposing music streaming app developers to use Apple’s proprietary in-app purchase system and limit their ability to inform users of alternative purchasing possibilities;⁴x The press release can be accessed at https://ec.europa.eu/commission/presscorner/detail/en/ip_21_2061 (last visited 1 February 2024). self-preferencing has been at the core of the widely debated Google Shopping case, in which Google was accused of having favoured its own vertical search engine by ranking it more prominently in the research results of its general search engine while demoting competitors;⁵x Commission decision of 27 June 2017, OJ 2018 C 9/11; Case T-612/17, Google LLC e Alphabet, Inc. v. Commission, ECLI:EU:T:2021:763. and refusal to access and account termination cases have been dealt with by national competition authorities, such as the several cases initiated by the French competition authorities against Google having engaged in such practices in the advertising market.⁶x For an overview of these cases, see Schweitzer and Gutmann, above n. 3, at 5.
  Nonetheless, competition enforcement in the field of online platforms has largely proved insufficient, including in platform-to-business (P2B) relations, due to online platforms’ unique features, urging the European legislator to act.⁷x M. Botta, ‘Sector Regulation of Digital Platforms in Europe: Uno, Nessuno e Centomila’, 12 Journal of European Competition Law & Practice 500 (2021). This resulted, among other things, in the enactment of the P2B regulation in 2019 and, more recently, the Digital Markets Act (DMA).
  These attempts, however, have not been backed by a robust scientific debate which, as also the Commission noticed,⁸x European Commission, Directorate-General for Internal Market, Industry, Entrepreneurship and SMEs, ‘Business-to-Business Relations in the Online Platform Environment – Final Report’, Publications Office, at 17 (2017), https://data.europa.eu/doi/10.2873/713211. has mostly neglected the issues arising in the context of P2B relations and failed to provide solid theoretical foundations, rather focusing on the capacity of antitrust law to contain the overwhelming power of digital platforms⁹x See, for instance, D.S. Evans, ‘The Antitrust Economics of Multi-Sided Platform Markets’, 20 Yale Journal on Regulation 58 (2003); J. Crémer, Y. de Montjoye & H. Schweitzer, ‘Competition Policy for the Digital Era’ (European Commission Directorate-General for Competition 2019), https://data.europa.eu/doi/10.2763/407537 (last visited 28 July 2023); H. Hovenkamp, ‘Antitrust and Platform Monopoly’, 13 Yale Law Journal 1952 (2021); F. Jenny, ‘Changing the Way We Think: Competition, Platforms and Ecosystems’, 9 Journal of Antitrust Enforcement 1 (2021); J.M. Newman, ‘Antitrust in Digital Markets’, 72 Vanderbilt Law Review 1497 (2019); J. Rochet and J. Tirole, ‘Platform Competition in Two-Sided Markets’, 1 Journal of the European Economic Association 990 (2003). or on the relationship between antitrust law and data¹⁰x See, inter alia, B. Mäihäniemi, Competition Law and Big Data : Imposing Access to Information in Digital Markets (2020); N. Dadson, I. Snoddy & J. White, ‘Access to Big Data as a Remedy in Big Tech’, 20 Competition Law Journal 1 (2021); G. Colangelo, ‘Big data, piattaforme digitali e antitrust’, 3 Mercato Concorrenza Regole 425 (2016); K.N. Hylton, ‘Digital Platforms and Antitrust Law’, 98 Nebraska Law Review 272 (2019); M. Maggiolino, I big data e il diritto antitrust (2018); D.D. Sokol and R.E. Comerford, ‘Antitrust and Regulating Big Data’, 23 George Mason Law Review 1129 (2016). or, again, on the protection of consumer users in platform-to-consumer (P2C) relations.¹¹x C. Cauffman and C. Goanta, ‘A New Order: The Digital Services Act and Consumer Protection’, 12 European Journal of Risk Regulation 758 (2021); A. Moskal, ‘Digital Markets Act (DMA): A Consumer Protection Perspective’, 7 European Papers – A Journal on Law and Integration 1113 (2023).
  Besides, in several instances such legal instruments do not appear to always provide effective protection to business users in P2B relationships, either because the remedies provided are not effective or because the scope of application of such regulations (or of competition law) is too narrow and does not include smaller, but still powerful platforms, or certain abusive practices. This is particularly true for non-dominant platforms in which only the business user’s side multihomes (i.e. uses more than one platform simultaneously). In these cases, despite being non-dominant, such platforms can exercise a monopoly-like power towards the business users, and the current EU legal framework provides little to no protection.
  Given this context and in order to contribute to close the highlighted gaps in the literature, the present analysis aims to investigate if and to what degree European law provides effective legal mechanisms for safeguarding business users against unfair practices engaged in by online platforms in P2B relations and to propose how to address the resulting gaps in protection.
  In order to properly address this question, the present article will first deal with its premises, thereby defining the concept of online platform and illustrating its main features and sources of market power that justify the claim that online platforms tend to possess market and bargaining power. Due to space constraints, it will also narrow down the scope of the analysis to a single type of online platforms, namely transaction platforms, where issues arising from P2B relations are more pronounced. It will then proceed to analyse the relevant provisions and regulations within the EU legal system that provide some form of protection to business users in P2B relations in order to highlight potential gaps in protection. Because this inquiry requires a domestic comparative analysis, the present article will adopt a functionalist approach by defining an appropriate tertium comparationis. Finally, based on the gaps in protection thus highlighted, the present article will attempt to show that business users are in fact economically dependent on the platform, thereby suggesting that a prohibition of abuse of economic dependence could more effectively address the identified gaps in protection. After showing that Member States’ regulations on the abuse of economic dependence provide different level of protections, the present article will present a reform proposal that could close the protection gaps identified in the first part of the analysis, namely an EU-level harmonisation of the prohibition of the abuse of economic dependence.

• 2 Concept and Taxonomies of Online Transaction Platforms

  As anticipated, the present section aims at illustrating, on one hand, the notion of an online transaction platform and, on the other, its main features and sources of market power.
  With respect to the first objective, because transaction platforms are a type of online platforms, Subsection 2.1 will clarify the overarching notion of online platforms, while Subsection 2.3 will discuss the notion of transaction platforms and other types of online platforms from a taxonomical perspective, highlighting their main differences.
  With respect to the second objective, Subsection 2.2 will illustrate the main sources of market power and economic features that characterise all types of online platforms, while clarifying how the former act differently depending on the type of platform considered.
  The concepts discussed in this section will not only serve to clarify the concept of a transaction platform but also the reasons behind the choice to focus on such type of platforms rather than others, a choice dictated by space constraints. In addition, the discussion on the different sources of market power that inherently affect online platforms will serve to justify the claim that transaction platforms tend to acquire market power and are thus generally the stronger party in P2B relations. This will further inform the argument put forward in the fourth section on the economic dependence of business users vis-à-vis transaction platforms in P2B relations.

  2.1 Notion of Online Platform

  Although there is no agreed-upon definition of online platform,¹²x F. Bostoen, ‘Online Platforms and Vertical Integration: The Return of Margin Squeeze?’, 6 Journal of Antitrust Enforcement 355, at 364 (2018). most of the authors in the economic and legal literature seem to agree that a digital platform is an information technology service that is offered through the Internet and that facilitates the interaction between two or more unilaterally or reciprocally interdependent groups of users.¹³x Cf. A. Canepa, I mercanti dell’era digitale. Un contributo allo studio delle piattaforme (2020), at 30; Hovenkamp, above n. 9, at 1957; OECD, An Introduction to Online Platforms and Their Role in the Digital Transformation (2019), at 21. The elements encapsulated in this definition clearly portray the twofold nature of digital platforms, namely that of a digital service and that of a platform service. Several unique features can be inferred from both the digital and the platform nature of online platforms.
  Regarding their digital nature, online platforms are, first of all, ubiquitous. Platform providers can distribute their services all over the world with virtually no costs, and users from all over the world can interact through the platform.¹⁴x A. Quarta and G. Smorto, Diritto Privato Dei Mercati Digitali (2020), at 116. Secondly, the platform service is mostly or entirely provided through the collection and elaboration of enormous quantities of data.¹⁵x N. Srnicek, Platform Capitalism (2016), at 39.
  With respect to their ‘platform’ nature, online platforms, as intermediaries between different groups of users, are multisided markets, and, because of that, they are characterised by strong indirect network externalities that can be either unilateral or bilateral. As shown in the third section, the multisided nature of online platforms has relevant consequences for the application of Article 102 TFEU, particularly with respect to market definition.

  2.2 Indirect Network Effects and Other Online Platforms’ Sources of Market Power

  As highlighted in the previous subsection, online platforms are multisided markets and, as such, are inherently characterised by indirect network effects.
  Indirect network effects arise because users of one side value the platform more the more users of the other side are using it,¹⁶x B. Caillaud and B. Jullien, ‘Chicken & Egg: Competition among Intermediation Service Providers’, 34 The RAND Journal of Economics 309, at 310 (2003). and such can be either unilateral or bilateral.
  Unilateral indirect network effects have a bearing on only one group of users who are interested in an interaction with another group of users but not vice versa:¹⁷x OECD, above n. 13, at 21; BKartA, ‘Working Paper – Market Power of Platforms and Networks’, B6-113/15 2016, at 64 ff.; G. Luchetta, ‘Is the Google Platform a Two-Sided Market?’, 10 Journal of Competition Law and Economics 185, at 191 (2014) doubts that services affected by unilateral network effects constitute a platform; however, according to L. Filistrucchi, D. Geradin & E. van Damme, ‘Identifying Two-Sided Markets’, 36 World Competition 33, at 33 (2013) unilateral network effects can exist in two-sided markets. the more users use Google’s search engine, the more advertisers buy targeted advertising space.
  In contrast, bilateral network effects occur when one group of users values the platform more because a larger number of users from the other side use it and vice versa.¹⁸x OECD, above n. 13, at 21; BKartA, above n. 17, at 10. For instance, in Amazon, a larger number of buyers make the platform more attractive for sellers, because they can offer their goods to a larger pool of consumers. Conversely, the attractiveness of the platform for buyers increases as the number of sellers increases, since the former will be able to benefit from a wider choice as to the type of goods they intend to purchase, without having to surf the Internet in search of better offers. Consequently, in order to trigger bilateral network effects, the platform is faced with a dilemma: to convince one party to join, the platform must also convince the other party to join and vice versa.¹⁹x Caillaud and Jullien, above n. 16, at 310. However, once the online platform manages to get one of the sides on board, bilateral network effects trigger a virtuous circle, as more users of the other side will also join the platform, and so on.
  Combined with the digital dimension of the platform service, indirect network effects can also be a relevant source of market power and lead to highly concentrated markets, which translates into strong bargaining power in bilateral contractual relations.
  Indirect network effects also act as an important barrier to entry for new entrants, since the latter have to convince at least one side to join them in order to attract the other side.²⁰x C. Shapiro, ‘Antitrust in a Time of Populism’, 61 International Journal of Industrial Organization 714, at 184 (2018). While this could be somewhat easier with unilateral network effects, it is much more complex with bilateral network effects, as one group of users is willing to switch only if the other also switches and vice versa.
  In addition to network effects, online platform markets are characterised by high fixed costs and very low variable costs, which translates into strong economies of scale.²¹x I. Graef, ‘Market Definition and Market Power in Data: The Case of Online Platforms’, 38 World Competition 473, at 483 (2015). Digital platforms require large investments in their initial stages to develop the software, collect the necessary amount of data, convince at least one side to join the platform, and so on.²²x M. Noel and D.S. Evans, ‘Defining Antitrust Markets When Firms Operate Two-Sided Platforms’, 3 Columbia Business Law Review 102, at 121 (2005). However, once the online platform reaches the critical mass that triggers network effects,²³x N. Petit, Big Tech and the Digital Economy: The Moligopoly Scenario (2020), at 81. the variable costs – such as maintenance costs and costs to distribute a digital service to a large number of users – become quite low.²⁴x Hylton, above n. 10, at 275.
  In addition, digital platform markets tend to be affected by strong economies of scope, the intensity of which is greater than in brick-and-mortar industries.²⁵x D. Condorelli and J. Padilla, ‘Harnessing Platform Envelopment in the Digital World’, 16 Journal of Competition Law & Economics 143, at 154 (2020). Once a digital platform has triggered network effects, the costs to be incurred to expand onto other digital markets become very low due to the digital nature of the services provided, and the availability of users and data.²⁶x Crémer, de Montjoye & Schweitzer, above n. 9, at 33; Ibid., at 154.
  Notwithstanding the simultaneous presence of these sources of market power, the digital nature of online platforms could lead to the conclusion that their markets, although concentrated, are in fact contestable. The costs for a user to switch to another platform may appear – in the abstract – to be almost non-existent. This is well expressed by Google’s famous defence – that ‘competition is just a click away.’²⁷x A. Kovacevich, ‘Google’s Approach to Competition’ (Google Public Policy Blog, 2009), https://publicpolicy.googleblog.com/2009/05/googles-approach-to-competition.html (last visited 25 November 2022).
  However, switching costs in platform markets can be high. The incentive to switch only exists if it is shared, in the sense that a user has an incentive to switch to another platform only if most of the users of the other group also do so and vice versa.²⁸x These are commonly known as collective switching costs, see C. Shapiro and H.R. Varian, Information Rules: A Strategic Guide to the Network Economy (1999), at 184. Moreover, almost all platforms implement reputational systems to foster trust in transactions.²⁹x B.G. Edelman and D. Geradin, ‘Efficiencies and Regulatory Shortcuts: How Should We Regulate Companies like Airbnb and Uber?’, 19 Stanford Technology Law Review 293, at 300 (2016). Such reputational mechanisms may constitute significant switching costs for sellers, who would be disincentivised to switch to a new platform if they could not transfer the reputation capital accrued in the previous platform to the new platform.
  Nevertheless, it is not uncommon for at least one group of users to use more than one platform simultaneously. In the literature, this phenomenon is known as multihoming.³⁰x Rochet and Tirole, above n. 9, at 991 ff. In markets where both user groups engage in multihoming, platforms compete fiercely with each other, and none of them is able to exercise or gain market power. Generally, however, either no user group engages in multihoming or only one does but not the other.³¹x M. Armstrong, ‘Competition in Two-Sided Markets’, 37 The RAND Journal of Economics 668, at 669 (2006). In the first case, the platform will be able to take full advantage of the sources of market power described above by fully exercising the market power they confer on both user groups.³²x A. Gawer, ‘Digital Platforms and Ecosystems: Remarks on the Dominant Organizational Forms of the Digital Age’, 1 Innovation: Organization and Management 110, at 114 (2022). In the second case, which could be referred to here as ‘asymmetric multihoming’, the economic literature has shown that the platform competes fiercely with other platforms in the market for the singlehoming user group, while exercising monopoly-like power over the multihoming group.³³x R. O’Donoghue and A.J. Padilla, The Law and Economics of Article 102 TFEU (2020), at 178. Asymmetric multihoming is particularly problematic because, although the platform exercises a monopolist-like power over the multihoming user group, it does not hold a dominant position in the relevant market, as it competes with other platforms in the market for the non-multihoming user group. For this reason, as will be seen in the following section, transaction platforms affected by asymmetric multihoming would likely not meet the dominance condition, thereby hindering the application of Article 102 TFEU to these platforms.
  In summary, the interplay of these various sources of market power contributes to place digital platforms in the stronger position in P2B relationships. However, it is important to note that platforms come in different types, ranging from social networks and dating platforms to marketplaces. Not only the impact of the aforementioned sources of market power differs depending on the type of platform under consideration, but each type also possesses unique characteristics that resist a blanket classification. In the following subsection, the different types of platforms will be examined, and the analysis will be narrowed down to one type of platform, namely transaction platforms, where the need to address power imbalances in P2B relations is more urgent.

  2.3 A Brief Taxonomy of Online Platforms

  Based on the relevant literature, it could be distinguished between transaction platforms, non-transaction platforms,³⁴x L. Filistrucchi and others, ‘Market Definition in Two-Sided Markets: Theory and Practice’, 10 Journal of Competition Law and Economics 293, at 298 (2014); BKartA, above n. 17, at 18 ff. The distinction between transaction and non-transaction platforms is also somewhat echoed in Regulation (EU) 2019/1150, OJ 2019 L 186, which distinguishes between online intermediation services – whose definition corresponds to that of transaction platforms – and online search engines – that belong to non-transaction platforms. and a third hybrid category that could be referred to as ‘non-transaction matching platforms’.
  Transaction platforms are online platforms that facilitate a transaction between two or more groups of users, whereby transaction means any exchange of resources or services that has economic relevance.³⁵x Filistrucchi and others, above n. 34, at 298. Marketplaces and app stores are examples of transaction platforms, as they facilitate the commercial exchange of (digital) goods between two or more groups of users.
  By contrast, non-transaction platforms are online platforms that facilitate any unilateral interaction between two or more groups of users.³⁶x Ibid. Such interaction – being unilateral – cannot amount to a transaction, which is bilateral in nature. Examples of non-transaction platforms are advertising-funded social networks and search engines, as they facilitate the unilateral interaction between the advertisers and the users.
  Finally, non-transaction matching platforms could be defined as online platforms that allow for bilateral interactions between different groups of users, provided that this interaction is not a transaction in nature. Dating apps are an example of this kind of platform, where different user groups (men and women, for instance) can interact without engaging in any form of transaction.
  Because of the nature of interactions that these different types of online platforms facilitate, both transaction platforms and non-transaction matching are characterised by bilateral network effects, while unilateral network effects occur in non-transaction platforms.
  Generally, as anticipated, despite sharing some common traits, these three types of platforms appear to display several different features that require separate consideration. First, non-transaction matching platforms are not used by business users and fall therefore outside the scope of this analysis. Second, bargaining and market power imbalances might raise more serious issues for business users in transaction platforms than for advertisers in non-transaction platforms. As outlined above, transaction platform’s abusive conduct in P2B relations could have severe effects on the ability of business users to effectively carry out their activity, while non-transaction platforms’ abusive conduct towards advertisers is unlikely to have such serious effects. Because of that, and due to space constraint, the subsequent analysis will only focus on transaction platforms’ P2B relations.

• 3 Bargaining Power Imbalances in P2B Relations and Gaps in Protection in the EU Legal System

  In order to determine whether and to what extent EU law effectively protects business users from harmful conduct engaged in by transaction platforms in P2B relationships, all relevant provisions potentially addressing the issue should be considered. This means that, in order to conduct this analysis, all relevant provisions that could potentially address the issues under consideration should thus be taken into account, even if they pursue different objectives. This is, in essence, an exercise in domestic comparative analysis of all the provisions that serve the specific function of addressing power imbalances in P2B relations and will thus require adopting a functionalist approach. As for every comparative analysis (including domestic comparative analysis) that adopts a functionalist approach, a tertium comparationis should be set beforehand.³⁷x H. Kötz, ‘Comparative Law in Germany Today’, 51 Revue internationale de droit comparé 753, at 755 (1999). Because the present research is aimed, inter alia, at assessing if and to what extent the EU legal system provides adequate legal tools to address transaction platforms’ abuses in P2B relations, the tertium comparationis should be based primarily on a meta-legal definition of abusive or harmful conduct in P2B relations. As anticipated, such a benchmark would allow one to compare different provisions of the EU legal system that, despite belonging to different branches of the law, could potentially provide some form of protection to business users in P2B relations. In particular, such a definition of abusive conduct in P2B relations and the corresponding benchmark would allow, first, the identification of the provisions in the EU legal system that potentially address this kind of abusive conduct, and, second, a comparison between these legal tools that highlights whether and to what extent each of them tackles transaction platforms’ abusive conduct in P2B relations. Clearly, the adoption of a different benchmark could yield different results.

  3.1 Tertium Comparationis: Transaction Platforms’ Abusive or Harmful Conduct

  As anticipated, this subsection is devoted to the definition of abusive conduct a transaction platform could engage in in the context of P2B relations, that is, the tertium comparationis mentioned above. Such definition draws from the two closely related concepts of abuse of right, which includes the abuse in the exercise of contractual powers,³⁸x S. Rowan, ‘Abuse of Rights in English Contract Law: Hidden in Plain Sight?’, 84 The Modern Law Review 1066, at 1067 f. (2021). and the abuse of dominance, given that both relate to a person’s per se lawful conduct that has harmful effects towards third parties.
  An abuse of right occurs, among others, when the right is (unilaterally) exercised in a way that, although being per se lawful,³⁹x E. Friedler, ‘Moral Damages in Mexican Law: A Comparative Approach’, 8 Loyola of Los Angeles International and Comparative Law Journal 235, at 272 (1986). is contrary to its socioeconomic purpose.⁴⁰x A. di Robilant, ‘Abuse of Rights: The Continental Drug and the Common Law’, 61 Hastings Law Journal 687, at 691 ff. (2010); on the notion of abuse of right, see also, among others, H. Lauterpacht, The Function of Law in the International Community (2011), at 294; M. Byers, ‘Abuse of Rights: An Old Principle, A New Age’, 47 McGill Law Journal 389, at 406 (2002). It therefore occurs if a right exists and the right holder exercises such right in an abusive manner.⁴¹x A. Kiss, ‘Abuse of Rights’, in R. Bernhardt (ed.), Encyclopedia of Public International Law (1992), para. 1; G. Palombella. ‘The Abuse of Rights and the Rule of Law’, in A. Sajó (ed.), Abuse: The Dark Side of Fundamental Rights (2006) 5, at 9-10.
  Similarly, the prohibition of abuse of dominance concerns a per se lawful conduct that falls within the private autonomy of the undertaking but only becomes unlawful when both engaged in by a dominant undertaking and satisfying the other requirements set out in Article 102 TFEU.⁴²x O’Donoghue and Padilla, above n. 33, at 3.
  Since the present analysis focuses on P2B relations, that is, contractual relations, the right that a platform might abuse should be identified in its contractual autonomy. Due to the platform’s market and bargaining power, business users’ private autonomy is limited, and this could allow the platform to exercise its right in an unfair and detrimental manner.
  That stated, an abusive conduct could be here defined as

  any unilateral lawful conduct that the platform, as a consequence of its market/bargaining power, can arbitrarily engage in as long as such conduct remains within the scope of the platform’s private autonomy and can cause an economic harm to the business user.

  The lawfulness requirement should be understood in an abstract manner, which means that the practice should not amount to a clear violation of any legal or contractual provision that applies to the platform but should rather constitute an exercise of the platforms’ private autonomy. This clearly follows from the general understanding of the concept of abuse, which presupposes the abusive practice to be – in the first place – a formally lawful exercise of a right.⁴³x Friedler, above n. 39, at 272; A.N. Yiannopoulos, ‘Civil Liability for Abuse of Right: Something Old, Something New…’, 54 Louisiana Law Review 1173, at 1195 (1994); however, according to some authors, the concept of abuse of right is a logomachy as the right should end when the abuse begins, see, for instance, di Robilant, above n. 40, at 83. Nonetheless, the lawfulness requirement should be understood without considering the provisions that will be analysed below, as these are to be compared with the benchmark.
  The market/contractual power requirement is based on the premises of the present analysis. Besides, a platform would not be able to engage in unfair conduct towards business users if it did not possess any market/bargaining power, as business users would be able to refuse the conditions unilaterally set by the platform or would not suffer any harm due the presence of viable alternatives.
  Based on the scope of the present research, abusive conduct should be narrowed down to only include P2B relations that have contractual nature, including the negotiation stage. This means that conduct is deemed abusive under the definition provided above only if it is engaged in by the platform at any stage (negotiation, stipulation, consummation, termination) of the contractual relationship between the platform and the business user.
  Finally, as the present analysis aims, among others, at examining potential gaps in business users’ protection from transaction platform’s conduct, such conduct should be capable of producing harmful effects to business users.⁴⁴x The requirement of harmful effects also derives from the notion of abuse, cf. E. Reid, ‘Abuse of Rights in Scots Law’, 2 Edinburgh Law Review 129, at 139 (1998).
  By echoing a distinction commonly adopted in competition law,⁴⁵x R. Whish and D. Bailey, Competition Law (2012), at 198. harmful conduct could be here divided into exploitative conduct and exclusionary conduct. In P2B relations, exploitative conduct, which generally indicates unfair or unreasonable conduct towards those who depend on the dominant undertaking for the supply of goods or services,⁴⁶x R. Thompson et al., ‘Article 102’, in D. Bailey and L. Elisabeth (eds.), Bellamy&Child European Union Law of Competition (2018) 859, at 903. could be identified in conduct that causes an increase in the business user’s costs through the imposition of excessive prices or unfair conditions.
  An example of exploitative conduct can be found in the antitrust proceedings initiated by the Commission against Apple for forcing music streaming app developers to use Apple’s proprietary in-app purchase system and limit their ability to inform users of alternative purchasing possibilities.⁴⁷x The case is still ongoing, and more details can be found at https://ec.europa.eu/commission/presscorner/detail/en/ip_23_1217 (last visited 20 January 2024).
  By contrast, exclusionary conduct, which generally indicates conduct that leads to the elimination, weakening or marginalisation of competition,⁴⁸x Thompson et al., above n. 46, at 902. would occur when business users suffer a reduction in their revenues due to their exclusion from or marginalisation within the platform. This could result, for instance, from a refusal to give access to the platform’s services, from the promotion of the platform’s own products in the intermediated market, to the deactivation of the business user’s account and so on.
  Examples of exclusionary conduct as defined here can be found both at the national level and European level. With respect to the refusal to deal, in 2020 the French Competition Authority deemed Google’s refusal to enter into negotiations with publishers on remuneration for the reuse of their protected content to be prima facie abusive.⁴⁹x Autorité de la Concurrence 9 April 2020 n. 20-MC-01, although the FCA qualified such conduct as imposition of unfair condition and discrimination. Cases of self-preferencing have been dealt with also by the Commission, for instance in the recent Google AdTech case, where Google is accused, among others, of having favoured its own display advertising technology (and more specifically its own ad exchange platform AdX) to the detriment of competitors.⁵⁰x For a more detailed description of the case, see https://ec.europa.eu/commission/presscorner/detail/en/ip_23_3207 (last viewed 1 February 2024). With respect to the deactivation of a business users’ account, a few cases have been scrutinised by the French competition authority that dealt with several cases involving Google in the ad market.⁵¹x For an overview of these cases, see Schweitzer and Gutmann, above n. 3, at 5.
  Every transaction platform is likely to have the incentive to engage in exploitative conduct both in the negotiation/stipulation phase and in the contract consummation phase (e.g. through a unilateral change of the contractual terms).
  By contrast, since the very existence of a transaction platform is based on the presence of at least two user groups, the implementation of exclusionary conduct might appear – at first sight – self-detrimental. Nevertheless, a distinction should be made between exclusionary conduct affecting all or most users of a certain user group and exclusionary conduct affecting only one or a few users. In the latter case, a transaction platform is not likely to suffer any (relevant) negative consequence from the exclusion of a small number of users. Moreover, many platform providers vertically integrate, meaning that they do not merely provide the digital intermediation service but also offer goods or services themselves in the intermediated market.⁵²x Bostoen, above n. 12, at 365; on the incentives for a platform to vertically integrate, see A. Hagiu and J. Wright, ‘Multi-sided Platforms’, 43 International Journal of Industrial Organization 162, at 164-70 (2015). A vertically integrated platform or a platform that intends to vertically integrate might have the incentive to exclude or at least marginalise business users operating on the platform and with whom the platform directly competes.⁵³x Cf. F. Bostoen and D. Mândrescu, ‘Assessing Abuse of Dominance in the Platform Economy: A Case Study of App Stores’, 16 European Competition Journal 431, at 435 (2020).
  Based on this benchmark, in the EU legal system there appear to be mainly three pieces of legislation that can potentially address the issues arising from abusive or harmful conduct in P2B relations as defined here, namely, competition law (more specifically, abuse of dominant position under Article 102 TFEU), Regulation (EU) 2019/1150 and the DMA. The following three subsections will discuss the effectiveness of each of these regulatory options to provide protection to business user in P2B relations.

  3.2 Abuse of Dominant Position under Article 102 TFEU

  Generally, Article 102 TFUE forbids a dominant undertaking to engage in abusive conduct. Its application to transaction platform’s P2B relations raises several issues both because its scope of application does not fully encompass all transaction platforms and because it is unable to address all abusive practices that a transaction platform may engage in.
  Particularly, the scrutiny of three key elements, which is indispensable – among other factors – for qualifying a practice as an abuse of dominance, presents notable challenges when applied to P2B relations. These elements encompass, notably, the definition of the relevant market in which the undertaking operates, the existence of a dominant position held by the undertaking within said market, and the identification of instances of abusive conduct engaged in by the undertaking.⁵⁴x O’Donoghue and Padilla, above n. 33, at 4 ff. These elements will be now briefly discussed.

  3.2.1 Market Definition

  Market definition in transaction platform markets is particularly complex, as transaction platforms are multisided markets.
  First, it is not clear whether one or two markets should be defined.⁵⁵x D. Mândrescu, The Application of EU Antitrust Law to (Dominant) Online Platforms (2022), at 93; R. Podszun, ‘The Pitfalls of Market Definition: Towards an Open and Evolutionary Concept’ in F. Di Porto and R. Podszun (eds.), Abusive Practices in Competition Law (2018) 68, at 72. In the economic literature, it has been proposed to define two different relevant markets in non-transaction platforms – one for each side – and only one in transaction platforms.⁵⁶x Filistrucchi and others, above n. 34, at 315; such a view, however, is not shared by all authors. G. Niels, ‘Transaction versus Non-Transaction Platforms: A False Dichotomy in Two-Sided Market Definition’, 15 Journal of Competition Law & Economics 327, at 330 (2019), for instance, argues that the distinction between transaction and non-transaction platforms is not relevant for the purposes of market definition, as in any case a single market should be defined. For the purposes of this article, however, which only focuses on transaction platforms, the literature agrees that, at least for transaction platforms, and in general terms, only one market should be defined. However, this conclusion fails, for instance, to take into account transaction platforms with asymmetrical multihoming, in which it could rather be appropriate to define different markets for each platform’s side.⁵⁷x M. Volmar, Digitale Marktmacht (2019), at 150; see also S. Wismer and A. Rasek, ‘Market Definition in Multi-Sided Markets’, Organisation of Economic Co-operation and Development 2017, at 10.
  Second, market definition can hardly be carried out through quantitative analysis.⁵⁸x Cf. G. Niels and H. Ralston, ‘Two-Sided Market Definition: Some Common Misunderstandings’, 17 European Competition Journal 118, at 119 (2021). The most widely employed quantitative test is the ‘small but significant non-transitory increase in price test’ (SSNIP test), which seeks to verify whether a 5% to 10% increase in price of certain goods or services by a hypothetical monopolist is profitable.⁵⁹x J. Faull and A. Nikpay (eds.), The EU Law of Competition (2014), para. 1147 ff.; M. Ferro, Market Definition in EU Competition Law (2019), at 116; O’Donoghue and Padilla, above n. 33, at 142. If it is, then the relevant market is defined; otherwise, the analysis must be redefined to include other goods or services that function as potential substitutes until the increase in price is profitable.
  However, transaction platforms, in order to overcome the ‘chicken-and-egg’ problem usually implement a non-neutral price scheme; that is, the platform charges a zero price on one side of the market (usually the consumer side) that is then subsidised by a higher price charged on the other side of the market (usually the business side).⁶⁰x Mândrescu, above n. 55, at 125.
  If the SSNIP test were to be applied for the definition of a transaction platform’s market that implement non-neutral price schemes, it should be determined, inter alia, on which of the two markets the price increase should be applied, how to take into account the effects of the price increase on the other side, how to increase a price equal to zero and so on.⁶¹x Cf. Niels and Ralston, above n. 58, at 119 ff.
  As of now, there are no clear-cut answers to those questions, and, although the SSNIP test, along with other quantitative tests, is usually regarded as delivering more reliable results than qualitative methods in market definition,⁶²x O’Donoghue and Padilla, above n. 33, at 152. only the latter appear to be usefully applicable to transaction platforms.⁶³x Cf. Noel and Evans, above n. 22, at 667 ff.; Wismer and Rasek, above n. 57, at 13 ff. In an attempt to overcome the arbitrariness associated to qualitative methods and the subsequent risk of under- or over-enforcement, the more recent case law and Commission’s decisional practice⁶⁴x Commission decision of 18 July 2018, OJ 2019 C 402/19. appear to lean towards new qualitative tests developed in the literature that imitate the structure of the SSNIP test, such as the small but significant non-transitory decrease in quality test (SSNDQ test).⁶⁵x OECD, ‘The Role and Measurement of Quality in Competition Analysis’ (2013), at 8 ff. Such a test, however, could lead to even more arbitrary results, as quality (similarly to almost every other qualitative variable) cannot be exactly measured; nonetheless, the analysis would focus on one single qualitative parameter, that is, quality, leaving out every other potentially relevant parameter and circumstance. These issues have been recently addressed in the new Market Definition Notice, in which the Commission adopts a rather flexible approach.⁶⁶x Commission Notice on the definition of the relevant market for the purposes of Union competition law of 22 February 2024, OJ 2024 C/2024/1645, at https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=OJ%3AC_202401645#ntc131-C_202401645EN.000101-E0131 (last visited 10/04/2024). As for the multi-sidedness issue, the Commission is aware of the fact that, depending on the circumstances, it could be necessary to define one or separate relevant markets, and that network effects should be taken into account in the assessment.⁶⁷x Ibid., paras 94-95. With respect to the SSNIP test issue, the Commission acknowledges the importance of considering non-price parameters and states that it may consider alternatives to the SSNIP test, such as the SSNDQ test.⁶⁸x Ibid., paras 96-98.

  3.2.2 Dominance

  The finding of dominance also raises a few issues in transaction platform markets. As dominance is a legal requirement, and not a purely economic fact, the existence of several sources of market power does not necessarily mean dominance pursuant Article 102 TFEU.
  First of all, digital markets – and therefore also transaction platform markets – are usually dynamic markets.⁶⁹x Cf. Newman, above n. 9, at 1520. Innovative services combined with network effects have the potential to quickly reverse an incumbent platform’s dominant position.⁷⁰x O’Donoghue and Padilla, above n. 33, at 196. Indeed, the existence of indirect network effects may result in a double-edged sword. Although, in general, they act as a significant barrier to entry, greatly favouring the incumbent and pushing it towards a monopolistic position, the dominant platform may lose its market position in a very short time, potentially being excluded from the market if the user groups manage to overcome the collective switching costs. MySpace’s demise following the birth of Facebook is an often-cited example of how quickly a platform can rise and fall.⁷¹x D. Coyle, ‘Practical Competition Policy Implications of Digital Platforms’, 82 Antitrust Law Journal 835, at 849 (2019). A similar fate befell eBay, which was pushed out of the market of handicraft products by Etsy in just a few months.⁷²x J.-U. Franck and M. Peitz, ‘Market Power of Digital Platforms’, 39 Oxford Review of Economic Policy 34, at 40 (2023). Nonetheless, it should be noted that, in its most recent decisional practice, the Commission has attempted to reduce the weight of this factor by stating that dynamic competition should not hamper effective antitrust enforcement, especially if the market does not display signs of instability for non-insignificant time periods.⁷³x Commission decision of 18 July 2018, OJ 2019 C 402/19, at 91.
  Second, the finding of dominance partly relies on presumptions deriving from market shares. For instance, a market share below 40% leads to a rebuttable presumption of non-dominance.⁷⁴x O’Donoghue and Padilla, above n. 33, at 193. In platform markets in which business users engage in asymmetrical multihoming, transaction platforms compete for the singlehoming side, while holding a monopolistic position vis-à-vis the multihoming side. In this case, provided that only one market for digital intermediation services has been defined, transaction platforms are likely to have smaller market shares capable of triggering a rebuttable presumption of non-dominance despite holding a (quasi-)monopolistic power on the multihoming side.
  As a consequence, the dominance requirement could not only hamper an effective enforcement in transaction platform markets but could also leave out transaction platforms that operate in markets where users engage in asymmetrical multihoming.

  3.2.3 Abusive Conduct

  When contrasted with the benchmark outlined above, the finding of an abuse could also be challenging in transaction platform markets, and this holds true both for the concept of abuse in general and for specific forms of abusive conduct.
  Generally, a practice is abusive⁷⁵x This mostly refers to exclusionary abuses. However, see below for exploitative abuses. if (1) it is likely to produce anticompetitive effects⁷⁶x Faull and Nikpay, above n. 59, para. 4.92. and (2) is carried out with methods contrary to those of competition on the merits.⁷⁷x Case 85/76, Hoffmann-La Roche & Co. AG v. Commission of the European Communities, [1979] ECR 00461, Rec 91.
  Regarding (1), anticompetitive effects generally occur if the practice can harm consumer welfare or if it is likely to adversely impact the ability or the incentives of equally efficient competitors to compete in market,⁷⁸x P. Ibáñez Colomo, ‘Anticompetitive Effects in EU Competition Law’, 17 Journal of Competition Law & Economics 309, at 337 (2021). which means that a platform’s practice that is harmful to business users is likely to be abusive under competition law only if it has a distortive effect on the market. Exclusionary or exploitative conduct aimed at individual or small groups of users, especially if less efficient, is unlikely to have anticompetitive effects and, therefore, is unlikely to trigger antitrust enforcement, despite being harmful to business users. Regarding (2), competition on the merits is usually – but not exclusively – assessed through the as-efficient-competitor principle,⁷⁹x Case C-377/20, Servizio Elettrico Nazionale SpA and Others v. Autorità Garante della Concorrenza e del Mercato and Others, ECLI:EU:C:2022:379, Rec 78. whereby a practice amounts to an abuse if a hypothetical non-dominant competitor as efficient as the dominant undertaking can continue to effectively compete in the market despite the practice adopted by the dominant undertaking.⁸⁰x Cf. Ibáñez Colomo, above n. 75, at 431 ff.; Faull and Nikpay, above n. 59, para. 4.265. Although other tests can be used to assess the ‘competition on the merits’ requirement (such as the ‘no-economic sense’ test),⁸¹x O’Donoghue and Padilla, above n. 33, at 282. the application of the ‘as-efficient-competitor’ principle and its related price-based ‘as-efficient-competitor’ test to some platform markets might hamper effective antitrust enforcement. This is because some platform markets – such as those with high switching costs – are winner-take-all markets, where no as-efficient competitor is likely to ever emerge. Similarly, for vertically integrated platforms that directly compete with their business users, the emergence of a competitor as efficient as the platform might be not only hypothetical but even impossible. This issue appears to have been acknowledged by the Court of Justice in Post Danmark II, in which it stated, at least with respect to pricing abuses, that no relevance should be given to the as-efficient-competitor test if ‘the structure of the market makes the emergence of an as-efficient competitor practically impossible.’⁸²x Case C-23/14, Post Danmark A/S v. Konkurrencerådet, [2015] CMLR 5, Rec 59.
  Besides, for some specific forms of abusive conduct, Article 102’s protection to business users could be even more limited due to transaction platforms’ unique features.
  For instance, while Article 102 TFEU might be effectively applied to non-price-based exploitative conduct, its relevance to the abuse of excessive pricing might instead be severely limited.
  In United Brands, the European Court of Justice elaborated a two-limbed test to assess whether a practice amounts to an abuse of excessive pricing.⁸³x Case 27/76, United Brands Company and United Brands Continentaal BV v. Commission of the European Communities, [1978] ECR 207, Rec 250 ff. The first limb of the test requires the European Commission to assess whether the price is excessive, generally by comparing the price charged with the costs of production.⁸⁴x Ibid., Rec 251. The price could thus be considered excessive if there is a significant difference with the price that the undertaking would have charged in a competitive market.⁸⁵x Case C-117/16, Biedrība ‘Autortiesību un komunicēšanās konsultāciju aģentūra – Latvijas Autoru apvienība’ v. Konkurences padome, Opinion of Advocate General Wahl, ECLI:EU:C:2017:286, paras. 17 ff. The second limb of the test requires the European Commission to prove that the price is unfair either in itself (e.g., by comparing prices over time) or in comparison with others.⁸⁶x Case 27/76, United Brands Company and United Brands Continentaal BV v. Commission of the European Communities, [1978] ECR 207, Rec 252.
  Although, as AG Wahl observed in its Opinion on the Latvian Copyright case, such a test works well in markets protected by high barriers to entry,⁸⁷x Case C-117/16, Biedrība ‘Autortiesību un komunicēšanās konsultāciju aģentūra – Latvijas Autoru apvienība’ v. Konkurences padome, Opinion of Advocate General Wahl, ECLI:EU:C:2017:286, para. 48. it is mostly based on short-run considerations and could thus lead to over- or under-enforcement if applied in markets such as platform markets, where dynamic competition, initial investments and, generally, innovation play a pivotal role.⁸⁸x O’Donoghue and Padilla, above n. 33, at 253. In addition, the United Brands test was developed for traditional one-sided markets, while digital platforms are multisided markets that usually implement non-neutral price schemes.⁸⁹x Caillaud and Jullien, above n. 16, at 310; R. Funta, ‘Economic and Legal Features of Digital Markets’, 10 Danube: Law, Economics and Social Issues Review 173, at 178 (2019). As mentioned above, transaction platforms usually offer their digital services for a zero price to one of the user groups (usually the consumer group) to trigger and maintain network effects. In order to recoup the loss, transaction platforms then charge a significantly higher price to the other group of users, which is nonetheless locked into the platforms due to network effects. Under the first limb of the United Brands test and absent any other criterion to take the platform’s entire price structure into account, the price charged to the latter user group could be easily qualified as excessive.⁹⁰x D. Mandrescu, ‘Abusive Pricing Practices by Online Platforms: A Framework Review of Article 102 TFEU for Future Cases’ 10 Journal of Antitrust Enforcement 469, at 496 (2022). As a consequence of that, the test cannot be usefully applied to transaction platform markets, because its application would lead to erroneous results and would thus need to be readapted in order for it to be usefully applied to transaction platform markets.
  Similarly, as for exclusionary abuses, some of the legal tests developed in Article 102 TFEU case law might provide even less protection to business users in P2B relations than what has already been observed with respect to the abuse in general.
  For instance, under the conditions set by the ECJ in the Bronner case,⁹¹x Case C-7/97, Bronner v. Mediaprint, [1998] ECR I-07791, Rec 40. an (outright) refusal to deal is deemed abusive if, inter alia, (1) the dominant undertaking is vertically integrated; (2) the input is indispensable to operate in the downstream market (i.e. there are no alternatives and the input cannot be replicated in an economically viable way); and (3) the refusal to deal has the effect of eliminating all competition in the downstream market.
  A refusal to deal by a transaction platform is unlikely to meet such a high threshold.
  First, all refusals to deal by non-vertically integrated transaction platforms would never be deemed abusive, although non-vertically integrated platforms that intend to vertically integrate could nonetheless be incentivised to engage in such exclusionary abuses.
  Second, transaction platforms are unlikely to constitute an indispensable input to the intermediated market, as there are always alternative (although often less economically advantageous) channels to enter that market, such as offline channels. App stores are among the few exceptions to that because the intermediated market for applications exists only in the digital dimension, and a transaction platform could thus constitute an indispensable input for app developers, especially if the app store belongs to a digital ecosystem, such as the Apple ecosystem.
  Third, in the unlikely event that the second requirement is met, only refusals that can eliminate all competitors in the downstream market are deemed abusive. Because the intermediated market can be (at least theoretically) accessed also through other channels, it is extremely unlikely that this requirement can be met.
  However, some forms of exclusionary conduct might not pose as significant challenges to transaction platforms as they do to non-transaction platforms. An example thereof is self-preferencing, possibly a new form of abusive conduct that has been (and still is) at the centre of a lively doctrinal debate.⁹²x See, for instance, R. Nazzini, ‘Google and the (Ever-Stretching) Boundaries of Article 102 TFUE’, 6 Journal of European Competition Law & Practice 301 (2015) and P. Akman, ‘The Theory of Abuse in Google Search: A Positive and Normative Assessment under EU Competition Law’, 301 Journal of Law, Technology & Policy 2 (2017), which concludes that self-preferencing does not easily fall within one of the traditional forms of abusive conduct; but see also F. Bostoen, ‘The General Court’s Google Shopping Judgment Finetuning the Legal Qualifications and Tests for Platform Abuse’, 13 Journal of European Competition Law & Practice 75 (2022) and L. Hornkohl, ‘Article 102 TFEU, Equal Treatment and Discrimination after Google Shopping’, 13 Journal of European Competition Law & Practice 99 (2022); both seem to argue that self-preferencing could fall within the scope of the abuse of discrimination. In 2017, the Commission sanctioned Google for having abused its dominant position in the general search engine market by demoting competitors’ specialised search engines while simultaneously securing the top spots for its own specialised search engine – Google Shopping – among the results of generic searches.⁹³x Commission decision of 27 June 2017, OJ 2018 C 9/11. The Commission, however, failed to explicitly classify such conduct within one of the well-established forms of abuse, and so did the General Court.⁹⁴x See Case T-612/17, Google LLC e Alphabet, Inc. v. Commission, ECLI:EU:T:2021:763. As shown in the literature, self-preferencing can indeed hardly fall within one of the traditional forms of abuse, mainly (although not only) because Google’s competitors neither have any trade relationship with Google nor have they ever requested access to Google’s first results page. Because there is no commercial relationship between Google and its competitors, it is hard to classify Google’s behaviour as discrimination, tying or constructive refusal to deal.⁹⁵x Cf. Nazzini, above, n. 89, at 307; however, according to F. Bostoen, ‘The General Court’s Google Shopping Judgement Finetuning the Legal Qualifications and Tests for Platform Abuse’, 13 Journal of European Competition Law & Practice 75, at 79 (2022), the General Court’s judgement would have categorised the abuse of self-preferencing under the umbrella of the abuse of discrimination. Nonetheless, in more than 80,000 words, the General Court mentioned only once Art. 102 (c) TFEU and never expressly verified this abuse’s legal test. Similarly, as there is no access request by Google’s competitors or a refusal by Google, Google’s conduct cannot be qualified as an outright refusal to deal either.⁹⁶x Ibid., and for a very thorough analysis of the possible theories of abuse in the Google Shopping case, see Akman, above n. 89. However, in transaction platform markets, business users either have a contractual relation with the platforms or have to obtain access to the platform in order to be able to operate in the intermediated market. As a result, self-preferencing conduct by a transaction platform is more likely to fall into one of the traditional categories of abuse.

  3.2.4 Remedies

  Despite its limitations, when it comes to the applicable remedies, antitrust law proves to be a very effective legal tool, as the Commission or any National Competition Authority (NCA) can impose not only fines but also any kind of behavioural (and structural) remedy.⁹⁷x Council Regulation 1/2003, OJ 2003 L 1, Art. 7(1).
  Besides, following the 2014 Damages Directive, which harmonised private enforcement across the EU,⁹⁸x Directive 2014/104/EU, OJ 2014 L 349/1. business users could seek compensation before the competent national court for the harm caused by a transaction platform’s anticompetitive conduct. Such claim requires the plaintiff to prove, inter alia, that the infringement resulted in actual anticompetitive effects.⁹⁹x O’Donoghue and Padilla, above n. 33, at 1215. However, in some cases, proving that platform’s abusive conduct produced anticompetitive effects may be challenging. For example, if a vertically integrated marketplace engages in self-preferencing by altering internal search engine results to favour its own products, it may be difficult for business users to demonstrate that such behaviour caused them harm.
  Besides, private enforcement does not include injunctive remedies, at least at EU level. As such, unless the jurisdiction in question provides for such a remedy, a business user could obtain compensation, but the platform could not be forced to end the anticompetitive practice, even if it is particularly harmful, such as, for instance, the deactivation of the business user’s account.

  3.3 EU Regulation 1150/2019 on Promoting Fairness and Transparency for Business Users of Online Intermediation Services

  Regulation (EU) 1150/2019¹⁰⁰x Regulation (EU) 2019/1150, OJ 2019 L 186. is intended to provide a certain level of protection to business users by directly addressing power imbalances in P2B relationships.¹⁰¹x M.L. Chiarella, ‘Platform Contracts: Legal Framework and User Protection’, 8 Athens Journal of Law 49, at 56 (2022). However, while its subjective scope of application expressly includes transaction platforms as defined here, it only covers transaction platforms in which one of the user groups consists of end-consumers.¹⁰²x P. Iamiceli, ‘Online Platforms and the Digital Turn in EU Contract Law: Unfair Practices, Transparency and the (Pierced) Veil of Digital Immunity’, 15 European Review of Contract Law 392, at 402 (2019). Therefore, the Regulation does not apply to all transaction platforms that facilitate B2B transaction – such as, for instance, agribusiness platforms.
  Furthermore, this Regulation fails to adequately address abusive behaviour that the platform may engage in.
  It does impose a few transparency obligations both ‘general’,¹⁰³x See, for instance, Regulation (EU) 2019/1150, OJ 2019 L 186, Art. 3. which apply to all contractual conditions regardless of their specific content, such as the obligation of clarity, and, ‘specific’,¹⁰⁴x Ibid., Arts. 5(1) and (3). which concern specific types of information or clauses, such as the obligation to disclose the main ranking parameters.¹⁰⁵x For the distinction between general and specific transparency obligations, cf. A. Palmieri, Profili Giuridici Delle Piattaforme Digitali. La Tutela Degli Utenti Commerciali e Dei Titolari Di Siti Web Aziendali (2019), at 52. However, the usefulness of such obligations is quite limited, given that enhanced transparency benefits more competitive markets, but less concentrated markets – such as most transaction platform markets – where the supply side enjoys a (quasi-)monopolistic position.
  Nonetheless, the Regulation contains a few substantive and procedural obligations that can to some extent limit exploitative conduct by the platform. For instance, under Article 8(a), the platform provider cannot impose retroactive changes to terms and conditions, unless they are required by law or are beneficial to business users. Similarly, under Article 3(2), the platform provider is obliged to notify business users of any unilateral changes to the terms and conditions, giving the business users a reasonable period of at least 15 days to either terminate the contract or adapt to the new conditions. This latter procedural obligation, however, proves to be of little usefulness in preventing exploitative conduct by the transaction platform, as it does not in any way alleviate the ‘take it or leave it’ situation in which business users are placed and merely delays it.
  As for exclusionary conduct, the Regulation does not provide any specific protection for exclusionary behaviour during the contract formation phase (i.e. refusal to deal). However, it contains some procedural obligation for exclusionary conduct in the phase of contract termination, as it requires the platform to provide the business user with a statement of reasons for a decision that restricts, suspends or terminates the provision of the platform’s services. Besides, such statement of reason must be sent 30 days before the implementation of that decision, in order to allow the business user to clarify the facts and circumstances. Similarly to what has been observed above, however, such procedural obligation does not in any way prevent the platform from engaging in exclusionary conduct, but merely delays the effects of such conduct. Nonetheless, if the platform fails to comply with such procedural obligations when deactivating the account of a business user, it could be forced to reactivate it, as shown in a recent German case involving Amazon abruptly terminating the account of a seller. Although the Court of Appeal of Munich later acknowledged that Amazon acted lawfully, as the business user was a serial infringer and therefore aware of the reasons behind the termination of the account, the Court had imposed interim measures on Amazon, ordering it to reactivate the business user’s account.¹⁰⁶x See for more details, Schweitzer and Gutmann, above n. 3, at 2.
  Besides sanctioning with nullity the contractual clauses that do not comply with certain transparency obligations, this Regulation does not contain specific remedies, but rather delegates the Member States to ensure adequate and effective enforcement of the obligations provided therein. In addition, the nullity remedy could not only prove to be ineffective in many cases but could also harm the business user if the nullity affected the whole contract,¹⁰⁷x Cf. Ibid., at 86 and 93. as the business user would need to renegotiate access to the platform.
  In conclusion, the Regulation only applies to B2C transaction platforms, thus excluding B2B transaction platforms from its subjective scope of application. Besides some limited obligations on exploitative conduct, the Regulation is largely ineffective in preventing both exploitative and exclusionary conduct that the platform could engage in and, in most cases, merely delays it.

  3.4 The Digital Markets Act

  The recently introduced Digital Markets Act (DMA)¹⁰⁸x Regulation (EU) 2022/1925, OJ 2022 L 265. aims at ensuring both contestability and fairness in the digital sector,¹⁰⁹x Ibid., Rec 4. whereby the fairness objective specifically concerns P2B relations and the need to address bargaining power imbalances.¹¹⁰x P. Ibáñez Colomo, ‘The Draft Digital Markets Act: A Legal and Institutional Analysis’, 12 Journal of European Competition Law & Practice 561, at 563 (2021).
  Besides non-transaction platforms and other digital services, the DMA applies to both B2C and B2B transaction platforms.¹¹¹x Cf. Regulation (EU) 2022/1925, OJ 2022 L 265, Arts. 2(21) and 2(20). However, its scope of application is limited to gatekeepers. A transaction platform can be qualified as a gatekeeper if it is designated as such by the Commission, whereby only very large transaction platforms that have (i) a significant impact on the internal market; (ii) are important gateways for business users to reach end-users; and (iii) enjoy entrenched and durable positions and, thus, are likely to be designated as gatekeepers.¹¹²x Ibid., Art. 3. On 6 September 2023, the European Commission designated for the first time six gatekeepers, namely Alphabet, Amazon, Apple, ByteDance, Meta, Microsoft. For more details, see https://digital-markets-act.ec.europa.eu/gatekeepers_en (last visited 7 February 2024). Therefore, the DMA does not apply to digital platforms that do not reach the high thresholds set in this Regulation even if they are dominant in niche markets or to very large platforms that have not been yet designated as gatekeepers. Similarly, transaction platforms that operate in markets where business users engage in asymmetrical multihoming are very unlikely to be designated as gatekeepers.
  Once a platform has been designated as a gatekeeper, it is subject to a long (and exhaustive) list of obligations contained in Articles 5, 6 and 7.
  By considering only the obligations that apply to transaction platforms and that are aimed at pursuing the fairness objective, it can be concluded that the DMA only partially prevents transaction platforms’ abusive conduct.
  As for exploitative conduct, it appears that the DMA provides full protection to business users from app stores’ exploitative conduct, as Article 6(12) imposes gatekeepers to apply fair terms to users to access app stores, search engines and social networks. However, such general prohibition of imposing unfair terms does not apply to other transaction platforms, which are only subject to a limited list of specific obligations and prohibitions, such as the prohibition to impose Most-Favoured-Nation clauses under Article 5(3) and (4), which are in any case mostly directed towards the contestability objective rather than the fairness objective.
  Similarly, exclusionary conduct is also only partially covered by the DMA. For instance, exclusionary abuses in the phase of contract formation (i.e. refusal to deal) are prohibited only with respect to app stores, thus leaving out all other transaction platforms. However, the DMA contains a comprehensive prohibition of the abuse of self-preferencing and of the anticompetitive use of business users’ data, but it does not grant any form of protection to business users for exclusionary conduct in the phase of contract termination, such as the arbitrary deactivation of the user’s account.
  With respect to remedies, the European Commission can only issue cease-and-desist orders for non-systematic violations of the DMA’s obligations, but it cannot impose any other behavioural or structural remedy, unless the gatekeeper engages in systematic noncompliance. The DMA does not contain any provision for private enforcement, although it has been observed that Articles 5 and 6 are sufficiently precise and unconditional to allow private litigation.¹¹³x A. de Streel and others, ‘Effective and Proportionate Implementation of the DMA’, Cerre (Centre on Regulation in Europe) 2023, at 182.
  All in all, the DMA provides protection only to business users of very large transaction platforms that have been designated as gatekeepers by the European Commission, and only partially covers transaction platforms’ abusive conduct, as legal protection varies based on the type of the transaction platform and the type of the abusive practice.

• 4 A Proposal for an EU-Level Prohibition of Abuse of Economic Dependence in P2B Relations

  The previous section showed that EU law only partially addresses power imbalances in P2B relations.
  First, non-dominant transaction platforms and transaction platforms that operate in markets where users engage in asymmetrical multihoming are only covered by Regulation (EU) 1150/2019, which, however, provides an overall ineffective protection to business users and its scope of application is limited to B2C transaction platforms.
  Second, comprehensive protection from exploitative conduct is granted only to business users of gatekeeping app stores; otherwise, there is no effective protection for pricing abuses, but rather only for non-price-based exploitative abuses, provided that the practice is likely to have anticompetitive effects. For exclusionary abuses, the DMA provides limited protection for specific abuses, while Article 102 TFEU ensures a more comprehensive protection provided that a dominant platform has engaged in such conduct and that such conduct is likely to have anticompetitive effects.
  Generally, it appears that EU law does not address abusive conduct directed at individual business users or less efficient ones.
  In this section, it will be argued that business users in P2B relationships are economically dependent on the transaction platform. The existence of economic dependence in P2B relations could suggest that a prohibition abuse of economic dependence may prove a suitable tool for addressing the gaps in protection outlined above. Because of that, in Subsection 4.2 the concept of abuse of economic dependence will be examined more thoroughly, also considering examples of some Member States’ regulation on the abuse of economic dependence. Drawing from those insights, in the final subsection, it will be proposed that the adoption of a prohibition of abuse of economic dependence at EU level could fill the gaps in protection highlighted in the previous section, and the main features that such a regulation should possess in order to effectively tackle these issues will be highlighted.

  4.1 Economic Dependence of Business Users in P2B Relations

  The economic dependence of one company on another could be broadly defined as the lack of sufficient and reasonable alternatives possibilities for the former to switch to other undertakings,¹¹⁴x T. Tombal, ‘Economic Dependence and Data Access’, 51 IIC – International Review of Intellectual Property and Competition Law 70, at 74; P. Bougette, O. Budzinski and F. Marty, ‘Exploitative Abuse and Abuse of Economic Dependence: What Can We Learn from an Industrial Organization Approach?’, 129 Revue d’économie politique 261, at 271 (2019); P. Këllezi, ‘Abuse below the Threshold of Dominance? Market Power, Market Dominance, and Abuse of Economic Dependence’, in M.-O. Mackenrodt, B. Conde Gallego & S. Enchelmaier (eds.), Abuse of Dominant Position: New Interpretation, New Enforcement Mechanisms?, (2008) 55, at 69. whereby ‘sufficiency of alternatives’ should be assessed objectively and ‘reasonableness’ of such alternatives should take into consideration the possibilities concretely available to the weaker undertaking.¹¹⁵x L. Feteira, The interplay between European and national competition law after regulation 1/2003 (2016), at 150. The presence of market power is not a precondition for economic dependence to occur, meaning that economic dependence can arise even if the stronger undertaking is not dominant in the relevant market.¹¹⁶x S. Scalzini, ‘Economic Dependence in Digital Markets: EU Remedies and Tools’, 5 Market and Competition Law Review 81, at 86 (2021).
  Based on this definition, business users in P2B relationships could appear to be economically dependent on transaction platforms, as they generally lack sufficient and reasonable alternatives.
  Indeed, as mentioned above, switching costs in transaction platforms tend to be quite high, because of bilateral network effects and practical lack of data portability.
  Bilateral network effects result in high collective switching costs because a member of one side will switch only if most of the members of the other side will also switch, and vice versa, while a lack of data portability can result in high switching costs for business users.¹¹⁷x V. Kathuria and J.C. Lai, ‘User Review Portability: Why and How?’, 34 Computer Law & Security Review 1291, at 1297 (2018). Restoring the reputational capital is one of them.
  As mentioned above, transaction platforms generally implement reputation systems that are based on user reviews.¹¹⁸x C. Dellarocas, ‘The Digitization of Word of Mouth: Promise and Challenges of Online Feedback Mechanisms’, 49 Management Science 1407, at 1407 (2003). Reputation systems play a very important role in transaction platforms, as they significantly reduce transaction costs and information asymmetries, and are so relevant in the digital economy that business users with many bad reviews or lack thereof could hardly sell any of their products or services on that platform (or on other sales channels). Because of that, it is unlikely that business users would switch to another platform if they cannot simultaneously transfer their digital reputation capital to the new platform. Indeed, if they nonetheless switched to another platform, they would have to rebuild their reputation from scratch.
  Although it could be argued that transaction platforms merely represent one of many possible sales channels, as business users will always have – at least – non-platform alternatives to reach consumers, non-platform sales channels are not always equivalent to transaction platforms. Three arguments might support this view.
  First, while large business users with a strong brand reputation could easily rely on other non-digital sales channels, and even their own online shop, this is not the case for small/medium business users or for business users that do not have a strong brand reputation. On one hand, these business users could not rely on their online shop, and, on the other, they would have to bear much higher costs (advertising, brick-and-mortar retailers, facilities etc.) to sell their goods or services. This would likely drive out of the market many smaller businesses and new entrants that cannot yet bear these costs, thus eliminating efficiencies that digital platforms bring about.
  Second, business users that sell digital products would have no alternative sales channels, even if they have some degree of brand reputation.
  Third, the growing importance of the digital economy makes transaction platforms almost an essential infrastructure for reaching end-users, as consumers increasingly prefer to buy products and services online than in brick-and-mortar facilities.
  Due to these reasons, transaction platforms are not equivalent to any other sales channels, meaning that, although other sales channels could theoretically constitute alternatives to transaction platforms, they are usually not sufficient or reasonable alternatives.

  4.2 The Prohibition of Abuse of Economic Dependence in France, Italy and Germany

  The previous subsection showed that P2B relations are generally characterised by a state of economic dependence. This may suggest that a prohibition of abuse of economic dependence could address the issues arising in P2B relations more effectively than the other legal instrument analysed in the previous section. Because, at the national level, several EU Member States have adopted regulations prohibiting the abuse of economic dependence,¹¹⁹x F. Cafaggi and others, ‘Study on the Legal Framework Covering Business-to-Business Unfair Trading Practices in the Retail Supply Chain: Final Report’, European Commission Directorate-General for Internal Market 2014, at 43. this section aims at analysing some of these regulations in order to verify, on one hand, whether such regulations provide a high and uniform level of protection in P2B relations and, on the other, to highlight the main features that a prohibition of abuse of economic dependence generally possesses, at least at the national level.
  For these purposes, the Italian, French and German regulations on the abuse of economic dependence will be briefly considered.
  In all these systems, an abuse of economic dependence occurs, first, if a state of economic dependence exists and, second, if the stronger undertaking has abused such state of economic dependence.¹²⁰x Tombal, above n. 111, at 74.
  As for the first requirement, in all the three legal systems, the lack of sufficient and reasonable alternatives as defined in the previous subsection is a relevant factor to determine the existence of economic dependence.¹²¹x Këllezi, above n. 111, at 61 ff.; cf. also Cafaggi and others, above n. 116, at 51 ff.
  However, in France, despite Article L. 420-2, paragraph 2 of the French Code de Commerce does not mention it, the lack of alternatives has been deemed in the case law to constitute only one of four cumulative conditions that must be met for a finding of economic dependence – namely, in addition to the lack alternatives, brand reputation, significant market shares and a significant turnover achieved by the claimant due to the relationship with the undertaking.¹²²x H. Friederiszick and S. Reinhold, ‘The Economics of Dependence: A Theory of Relativity’, ESMT Working Paper, 2021:21-02, at 5. In Germany, by contrast, the lack of reasonable alternatives, which should result in a clear power imbalance, is assessed in the light of the relevant market defined according to the criteria set out for the abuse of dominant position, but not as strictly.¹²³x K. Markert, ‘GWB § 20 Verbotenes Verhalten von Unternehmen Mit Relativer Oder Überlegener Marktmacht’, in U. Immenga and E.-J. Mestmäcker (eds.), Wettbewerbsrecht (2020), para. 16; Friederiszick and Reinhold, above n. 119, at 6. Such assessment has the purpose of ascertaining the sufficiency requirement.¹²⁴x Markert, above n. 120, para. 14. In addition, recently the German prohibition of abuse of economic dependence – contained in Section 20 of the Gesetz gegen Wettbewerbsbeschränkungen (GWB) – has been expanded to include, on one hand, undertakings that operate in multisided markets with respect to the companies that are dependent on those intermediary services to access supply and sales markets and, on the other, to specify that a state of economic dependence can arise from an undertaking’s dependence on data controlled by another undertaking.
  Finally, in Italy, the lack of sufficient and reasonable alternatives features merely as one of the factors that should be assessed for a finding of economic dependence, which is instead more generally defined as the ability to impose excessively imbalanced conditions.¹²⁵x Scalzini, above n. 113, at 91. Nonetheless, the lack of sufficient and reasonable alternatives is employed in practice as the main criterion to determine the existence of a state of economic dependence, given the mere descriptive nature of its legal definition.¹²⁶x Ibid. Similar to Germany, the Italian lawmaker, too, has recently introduced a rebuttable presumption of economic dependence for intermediation platforms that are necessary to reach end-users or suppliers, also considering network effects or control over data.¹²⁷x M. Maugeri, ‘DdL concorrenza e piattaforme digitali’, Persona e Mercato 1 (2022), http://www.personaemercato.it/ddl-concorrenza-e-piattaforme-digitali-brevi-considerazioni-sulla-proposta-di-modifica-della-disciplina-sullabuso-di-dipendenza-economica-di-marisaria-maugeri/ (last visited 1 February 2024).
  Regarding the second requirement, namely the notion of abuse, it is generally satisfied if (i) the stronger undertaking has unreasonably and excessively exercised its economic freedom and (ii) the undertaking would have not been able to engage in such conduct absent the state of economic dependence.¹²⁸x Feteira, above n. 112, at 151-8 and 170-1. All three jurisdictions considered accompany a general prohibition of abuse of economic dependence with a non-exhaustive list of abusive practices, such as refusal to supply or imposition of unfair conditions. With respect to that, both the Italian and the French systems explicitly qualify the refusal to supply as a standalone abuse, while the German regulation does not do so. It, however, refers to the non-exhaustive list of abuses provided under the abuse of dominant position, which includes the refusal to supply.¹²⁹x Cf. Markert, above n. 120, para. 1. In addition, both the German and the Italian legislators recently expanded such non-exhaustive lists by prohibiting specific abuses for online platforms, such as the refusal to supply data or the exploitation of network effects.¹³⁰x Cf. Scalzini, above n. 113, at 83.
  Despite the several similarities between the three regulations, there appears to be less homogeneity regarding whether a practice by a stronger undertaking needs to produce effects on the market in order for it to be considered an abuse.
  On this point, the French system requires the abusive practice to be capable of affecting the functioning or the structure of market competition,¹³¹x C. Grynfogel, ‘Droit Français des Abus de Domination – Article L. 420-2 du Code de Commerce’, JurisClasseur Commercial, para. 181 (2017). thereby setting a condition that could limit the scope of application of such prohibition.¹³²x I. Graef, ‘Differentiated Treatment in Platform-to-Business Relations: EU Competition Law and Economic Dependence’, 38 Yearbook of European Law 448, at 490-1 (2019).
  Under the Italian system, abuses that do not have an impact on the market only give rise to private sanction, while abuses affecting the market trigger public enforcement of the Italian Competition Authority.¹³³x M. Libertini, ‘La responsabilità per abuso di dipendenza economica: La fattispecie’, 29 Contratto e Impresa 1, at 14 (2013).
  Similar to Italy, the German regulation, too, does not require an abusive practice to have an impact on competition,¹³⁴x Graef, above n. 129, at 491. and, indeed, Section 20(1) GWB explicitly applies to small or medium enterprises.
  Overall, while the Italian and German regulations on the abuse of economic dependency show remarkable similarities and a fair degree of flexibility, the French system appears more limited in its scope, both with respect to the first requirement (i.e. the state of economic dependence) and the second requirement (i.e. the abuse of economic dependence). With respect to the first requirement, the lack of sufficient and reasonable alternatives is only one of four requirements, while, with regards to the second requirement, a practice is deemed abusive only if it has an impact on the relevant market. In addition, unlike the German and the Italian systems, the French system’s prohibition on the abuse of economic dependence has not been adapted to better cover abusive practice in P2B relations.

  4.3 A Proposal for an EU-Level Prohibition of the Abuse of Economic Dependence in P2B Relations

  The previous subsection showed that national prohibitions on the abuse of economic dependence provide different levels of protection for business users in P2B relations. This could support a call for an EU-level regulation on the abuse of economic dependence, which, through harmonisation, would level these differences and provide a uniform degree of protection for business users across the EU.
  Nonetheless, besides the need to harmonise such matters, there are other reasons why such a solution would be preferable to others.
  First, most of the issues that arise in the application of Article 102 TFUE to P2B relations – such as the definition of the relevant market, the assessment of dominance, and the necessity to ascertain the existence of anticompetitive effects – would not necessarily concern a prohibition of abuse of economic dependence, as it is generally sufficient to demonstrate that one of the parties of a contractual relationship is economically dependent on the other and it is not necessary for the abusive conduct to be able to affect the market.
  It is true that some of the issues that emerge in the application of antitrust law to P2B relations – such as the assessment of dominance, the United Brands test or the Bronner test – can be overcome through a readaptation of the respective conceptual tools. However, some other issues – such as those deriving from the concept of abuse in general – would require a radical reconfiguration of antitrust law and of its goals. Since power imbalances only arise in certain sectors of the economy, it may be excessive to make radical changes to antitrust laws to address issues that are specific to those sectors. By contrast, the abuse of economic dependence can pursue objectives that are different from those of competition law, such as that of contractual fairness, similarly to Regulation EU 2019/1150 and the DMA, while leaving antitrust law untouched.
  Second, ex ante regulation of the kind of consumer regulation might be too rigid, as business users would always qualify as the weaker parties in P2B relations. However, unlike consumer regulation, where consumers are always natural persons, business users can be both natural and legal persons, including large companies with a strong brand reputation. As seen, such undertakings are likely to enjoy a strong countervailing power in P2B relations and are thus less worth of protection. By contrast, if the scope of application of such an ex ante regulation were limited to small- and medium-sized companies only, it might not cover larger businesses that, for some reasons, are indeed the weaker party in P2B relations, such as app developers. Similarly, such ex ante regulation might not take into consideration those cases in which the platform, due to low switching costs on both sides, does not have market power. Because a prohibition of abuse of economic dependence always allows for a case-specific analysis aimed at assessing whether economic dependence actually occurs, such issues are likely to be taken into account, thus making a prohibition of abuse of economic dependence a more suitable instrument than consumer-like ex ante regulation.
  All this considered, and based on the outcome of the previous sections, an EU prohibition of abuse of economic dependence should present the following features.
  First, the existence of an economic dependence should not require a prior definition of the relevant market as strictly as set out in antitrust law. Indeed, as illustrated above, none of the jurisdictions considered above requires a prior definition of the relevant market comparable to that required in antitrust law, including the German system, in which the definition of the relevant market is necessary to assess the existence of sufficient alternatives.¹³⁵x Markert, above n. 120, para. 14.
  Second, a finding of economic dependence should not require the transaction platform to be dominant on the relevant market and should rather be solely based on the lack of sufficient and reasonable alternatives as defined above. This could, on one hand, make such solution applicable also to smaller transaction platforms affected by asymmetric multihoming and, on the other, address one of the main shortcomings of the French system, which, as seen, requires four cumulative requirements to be satisfied for a finding of economic dependence.¹³⁶x Grynfogel, above n. 128, para. 169. These requirements set quite a high threshold that partly resembles that of Article 102 TFEU. In addition, a rebuttable presumption of economic dependence in transaction platform markets could also be reasonable because transaction platforms usually have market power for the reasons above mentioned and business users would have to prove a negative fact, that is, that there are no reasonable alternatives. Nonetheless, it should be noted that such presumption could increase the risk of false-positive, especially when the platform is not capable of rebutting the presumption, which could then lead to higher entry costs for new platforms.
  Third, the EU-level regulation should contain both a general prohibition of abuse of economic dependence and a non-exhaustive list of abuse, which could balance flexibility and legal certainty, in a manner similar to the national legislations considered above. The non-exhaustive list of abuses should also include platform-specific conduct, in a way similar to that of the German and the Italian provisions, in order to avoid ambiguities and facilitate the enforcement in the digital sector.
  Fourth, enforcement should be both public and private. Public enforcement could enable NCAs (or any other competent authority) to prevent abuses that affect large numbers of business users, while private enforcement could ensure adequate protection for practices that affect one or few business users. However, judges should have the power to impose both cease-and-desist orders and other behavioural remedies, such as obligations to enter into agreements with the business user, besides compensation.
  Provided that an EU-level regulation on the abuse of economic dependence has these features, an adequate legal basis for such a set of provisions could be found in Article 26 and Article 114 TFEU. More specifically, such a legislative intervention could fall into the Digital Single Market Strategy, regarding the development of trans-European networks, where the EU has a shared competence with Member States.
  Finally, an EU-level prohibition of abuse of economic dependence should harmonise the laws of EU Member States only with respect to P2B relations, while leaving untouched the prohibitions on the abuse of economic dependence in other industrial sectors. A harmonisation of the Member States’ laws on abuse of economic dependence in other industrial sectors would, indeed, require further analysis in order to avoid overregulation, while a harmonisation limited to P2B relations would be unlikely to raise any issue related to the subsidiarity principle.
  Although the enactment of an EU-level prohibition of abuse of economic dependence in P2B relations could prove to be an effective tool to close the gaps in protection identified in the third section, its usefulness should be evaluated in the light of possible shortcomings, in order to assess whether the latter could offset such benefits. In particular, such reform proposal would have the effect of further complicating an already fragmented set of regulations, which would render it more difficult for incumbent and emerging platforms to comply. In addition, since the proposed legal tool is built on a specific theoretical background, it would contribute to further fragment the theoretical foundations on the regulatory landscape on online platforms.

• 5 Conclusions

  The emergence of digital platforms has posed, and continues to pose, countless challenges. European law, despite making significant strides, needs to adjust to these evolving circumstances. The considerable bargaining and market power that digital platforms tend to acquire due to their unique characteristics puts them in a superior and stronger position vis-à-vis professional users, so that, unlike normal B2B relations, P2B relations can be assumed to be structurally unbalanced to the advantage of platforms. Although of great social and economic importance, the problem of the protection of professional users in P2B relationships has been rather neglected in the literature. This contribution aimed to advance the debate on this issue and to propose a possible solution, being the introduction of an EU prohibition of abuse of economic dependency in P2B relations. Specifically, the intention is to formulate a proposal that, on one hand, would not leave any gaps in protection and, on the other, would remain within the limits of what is strictly necessary. To this end, the proposal formulation was preceded by a domestic comparative analysis of the EU legal system in order to precisely identify the cases in which business users enjoy sufficient protection and those in which they do not. The comparative analysis of the EU legal system required both the elaboration of a tertium comparationis – achieved through the definition of the concept of platform and of abusive conduct – and to focus the analysis on transaction platforms only, because other types of platforms present different problems and characteristics.
  Based on this analysis, it became feasible to pinpoint the specific gaps in protection that currently exist in the EU legal system. Building upon this foundation, the prohibition of economic dependency abuse appeared to be the most fitting mechanism to address these gaps. However, the effectiveness of this approach relies on incorporating in such proposal particular characteristics that can be drawn from the existing prohibitions of economic dependency abuse in a select few Member States.

Noten

• 1 Commission’s Synopsis Report on the public consultation on the regulatory environment for platforms, online intermediaries and the collaborative economy, https://digital-strategy.ec.europa.eu/en/library/results-public-consultation-regulatory-environment-platforms-online-intermediaries-data-and-cloud (last visited 1 February 2024).

• 2 European Commission, Directorate-General for Internal Market, Industry, Entrepreneurship and SMEs, ‘Business-to-business relations in the online platform environment – Final report’, Publications Office, at 24 (2017), https://data.europa.eu/doi/10.2873/713211 (last visited 1 February 2024).

• 3 For an overview of online platforms unilateral practices, see H. Schweitzer and F. Gutmann, ‘Unilateral Practices in the Digital Market: An Overview of EU and National Case Law’, Concurrences 101045 (5 July 2021), https://www.concurrences.com/en/bulletin/special-issues/unilateral-practices-in-the-digital-market/new-article-no101045 (last visited 1 February 2024).

• 4 The press release can be accessed at https://ec.europa.eu/commission/presscorner/detail/en/ip_21_2061 (last visited 1 February 2024).

• 5 Commission decision of 27 June 2017, OJ 2018 C 9/11; Case T-612/17, Google LLC e Alphabet, Inc. v. Commission, ECLI:EU:T:2021:763.

• 6 For an overview of these cases, see Schweitzer and Gutmann, above n. 3, at 5.

• 7 M. Botta, ‘Sector Regulation of Digital Platforms in Europe: Uno, Nessuno e Centomila’, 12 Journal of European Competition Law & Practice 500 (2021).

• 8 European Commission, Directorate-General for Internal Market, Industry, Entrepreneurship and SMEs, ‘Business-to-Business Relations in the Online Platform Environment – Final Report’, Publications Office, at 17 (2017), https://data.europa.eu/doi/10.2873/713211.

• 9 See, for instance, D.S. Evans, ‘The Antitrust Economics of Multi-Sided Platform Markets’, 20 Yale Journal on Regulation 58 (2003); J. Crémer, Y. de Montjoye & H. Schweitzer, ‘Competition Policy for the Digital Era’ (European Commission Directorate-General for Competition 2019), https://data.europa.eu/doi/10.2763/407537 (last visited 28 July 2023); H. Hovenkamp, ‘Antitrust and Platform Monopoly’, 13 Yale Law Journal 1952 (2021); F. Jenny, ‘Changing the Way We Think: Competition, Platforms and Ecosystems’, 9 Journal of Antitrust Enforcement 1 (2021); J.M. Newman, ‘Antitrust in Digital Markets’, 72 Vanderbilt Law Review 1497 (2019); J. Rochet and J. Tirole, ‘Platform Competition in Two-Sided Markets’, 1 Journal of the European Economic Association 990 (2003).

• 10 See, inter alia, B. Mäihäniemi, Competition Law and Big Data : Imposing Access to Information in Digital Markets (2020); N. Dadson, I. Snoddy & J. White, ‘Access to Big Data as a Remedy in Big Tech’, 20 Competition Law Journal 1 (2021); G. Colangelo, ‘Big data, piattaforme digitali e antitrust’, 3 Mercato Concorrenza Regole 425 (2016); K.N. Hylton, ‘Digital Platforms and Antitrust Law’, 98 Nebraska Law Review 272 (2019); M. Maggiolino, I big data e il diritto antitrust (2018); D.D. Sokol and R.E. Comerford, ‘Antitrust and Regulating Big Data’, 23 George Mason Law Review 1129 (2016).

• 11 C. Cauffman and C. Goanta, ‘A New Order: The Digital Services Act and Consumer Protection’, 12 European Journal of Risk Regulation 758 (2021); A. Moskal, ‘Digital Markets Act (DMA): A Consumer Protection Perspective’, 7 European Papers – A Journal on Law and Integration 1113 (2023).

• 12 F. Bostoen, ‘Online Platforms and Vertical Integration: The Return of Margin Squeeze?’, 6 Journal of Antitrust Enforcement 355, at 364 (2018).

• 13 Cf. A. Canepa, I mercanti dell’era digitale. Un contributo allo studio delle piattaforme (2020), at 30; Hovenkamp, above n. 9, at 1957; OECD, An Introduction to Online Platforms and Their Role in the Digital Transformation (2019), at 21.

• 14 A. Quarta and G. Smorto, Diritto Privato Dei Mercati Digitali (2020), at 116.

• 15 N. Srnicek, Platform Capitalism (2016), at 39.

• 16 B. Caillaud and B. Jullien, ‘Chicken & Egg: Competition among Intermediation Service Providers’, 34 The RAND Journal of Economics 309, at 310 (2003).

• 17 OECD, above n. 13, at 21; BKartA, ‘Working Paper – Market Power of Platforms and Networks’, B6-113/15 2016, at 64 ff.; G. Luchetta, ‘Is the Google Platform a Two-Sided Market?’, 10 Journal of Competition Law and Economics 185, at 191 (2014) doubts that services affected by unilateral network effects constitute a platform; however, according to L. Filistrucchi, D. Geradin & E. van Damme, ‘Identifying Two-Sided Markets’, 36 World Competition 33, at 33 (2013) unilateral network effects can exist in two-sided markets.

• 18 OECD, above n. 13, at 21; BKartA, above n. 17, at 10.

• 19 Caillaud and Jullien, above n. 16, at 310.

• 20 C. Shapiro, ‘Antitrust in a Time of Populism’, 61 International Journal of Industrial Organization 714, at 184 (2018).

• 21 I. Graef, ‘Market Definition and Market Power in Data: The Case of Online Platforms’, 38 World Competition 473, at 483 (2015).

• 22 M. Noel and D.S. Evans, ‘Defining Antitrust Markets When Firms Operate Two-Sided Platforms’, 3 Columbia Business Law Review 102, at 121 (2005).

• 23 N. Petit, Big Tech and the Digital Economy: The Moligopoly Scenario (2020), at 81.

• 24 Hylton, above n. 10, at 275.

• 25 D. Condorelli and J. Padilla, ‘Harnessing Platform Envelopment in the Digital World’, 16 Journal of Competition Law & Economics 143, at 154 (2020).

• 26 Crémer, de Montjoye & Schweitzer, above n. 9, at 33; Ibid., at 154.

• 27 A. Kovacevich, ‘Google’s Approach to Competition’ (Google Public Policy Blog, 2009), https://publicpolicy.googleblog.com/2009/05/googles-approach-to-competition.html (last visited 25 November 2022).

• 28 These are commonly known as collective switching costs, see C. Shapiro and H.R. Varian, Information Rules: A Strategic Guide to the Network Economy (1999), at 184.

• 29 B.G. Edelman and D. Geradin, ‘Efficiencies and Regulatory Shortcuts: How Should We Regulate Companies like Airbnb and Uber?’, 19 Stanford Technology Law Review 293, at 300 (2016).

• 30 Rochet and Tirole, above n. 9, at 991 ff.

• 31 M. Armstrong, ‘Competition in Two-Sided Markets’, 37 The RAND Journal of Economics 668, at 669 (2006).

• 32 A. Gawer, ‘Digital Platforms and Ecosystems: Remarks on the Dominant Organizational Forms of the Digital Age’, 1 Innovation: Organization and Management 110, at 114 (2022).

• 33 R. O’Donoghue and A.J. Padilla, The Law and Economics of Article 102 TFEU (2020), at 178.

• 34 L. Filistrucchi and others, ‘Market Definition in Two-Sided Markets: Theory and Practice’, 10 Journal of Competition Law and Economics 293, at 298 (2014); BKartA, above n. 17, at 18 ff. The distinction between transaction and non-transaction platforms is also somewhat echoed in Regulation (EU) 2019/1150, OJ 2019 L 186, which distinguishes between online intermediation services – whose definition corresponds to that of transaction platforms – and online search engines – that belong to non-transaction platforms.

• 35 Filistrucchi and others, above n. 34, at 298.

• 36 Ibid.

• 37 H. Kötz, ‘Comparative Law in Germany Today’, 51 Revue internationale de droit comparé 753, at 755 (1999).

• 38 S. Rowan, ‘Abuse of Rights in English Contract Law: Hidden in Plain Sight?’, 84 The Modern Law Review 1066, at 1067 f. (2021).

• 39 E. Friedler, ‘Moral Damages in Mexican Law: A Comparative Approach’, 8 Loyola of Los Angeles International and Comparative Law Journal 235, at 272 (1986).

• 40 A. di Robilant, ‘Abuse of Rights: The Continental Drug and the Common Law’, 61 Hastings Law Journal 687, at 691 ff. (2010); on the notion of abuse of right, see also, among others, H. Lauterpacht, The Function of Law in the International Community (2011), at 294; M. Byers, ‘Abuse of Rights: An Old Principle, A New Age’, 47 McGill Law Journal 389, at 406 (2002).

• 41 A. Kiss, ‘Abuse of Rights’, in R. Bernhardt (ed.), Encyclopedia of Public International Law (1992), para. 1; G. Palombella. ‘The Abuse of Rights and the Rule of Law’, in A. Sajó (ed.), Abuse: The Dark Side of Fundamental Rights (2006) 5, at 9-10.

• 42 O’Donoghue and Padilla, above n. 33, at 3.

• 43 Friedler, above n. 39, at 272; A.N. Yiannopoulos, ‘Civil Liability for Abuse of Right: Something Old, Something New…’, 54 Louisiana Law Review 1173, at 1195 (1994); however, according to some authors, the concept of abuse of right is a logomachy as the right should end when the abuse begins, see, for instance, di Robilant, above n. 40, at 83.

• 44 The requirement of harmful effects also derives from the notion of abuse, cf. E. Reid, ‘Abuse of Rights in Scots Law’, 2 Edinburgh Law Review 129, at 139 (1998).

• 45 R. Whish and D. Bailey, Competition Law (2012), at 198.

• 46 R. Thompson et al., ‘Article 102’, in D. Bailey and L. Elisabeth (eds.), Bellamy&Child European Union Law of Competition (2018) 859, at 903.

• 47 The case is still ongoing, and more details can be found at https://ec.europa.eu/commission/presscorner/detail/en/ip_23_1217 (last visited 20 January 2024).

• 48 Thompson et al., above n. 46, at 902.

• 49 Autorité de la Concurrence 9 April 2020 n. 20-MC-01, although the FCA qualified such conduct as imposition of unfair condition and discrimination.

• 50 For a more detailed description of the case, see https://ec.europa.eu/commission/presscorner/detail/en/ip_23_3207 (last viewed 1 February 2024).

• 51 For an overview of these cases, see Schweitzer and Gutmann, above n. 3, at 5.

• 52 Bostoen, above n. 12, at 365; on the incentives for a platform to vertically integrate, see A. Hagiu and J. Wright, ‘Multi-sided Platforms’, 43 International Journal of Industrial Organization 162, at 164-70 (2015).

• 53 Cf. F. Bostoen and D. Mândrescu, ‘Assessing Abuse of Dominance in the Platform Economy: A Case Study of App Stores’, 16 European Competition Journal 431, at 435 (2020).

• 54 O’Donoghue and Padilla, above n. 33, at 4 ff.

• 55 D. Mândrescu, The Application of EU Antitrust Law to (Dominant) Online Platforms (2022), at 93; R. Podszun, ‘The Pitfalls of Market Definition: Towards an Open and Evolutionary Concept’ in F. Di Porto and R. Podszun (eds.), Abusive Practices in Competition Law (2018) 68, at 72.

• 56 Filistrucchi and others, above n. 34, at 315; such a view, however, is not shared by all authors. G. Niels, ‘Transaction versus Non-Transaction Platforms: A False Dichotomy in Two-Sided Market Definition’, 15 Journal of Competition Law & Economics 327, at 330 (2019), for instance, argues that the distinction between transaction and non-transaction platforms is not relevant for the purposes of market definition, as in any case a single market should be defined. For the purposes of this article, however, which only focuses on transaction platforms, the literature agrees that, at least for transaction platforms, and in general terms, only one market should be defined.

• 57 M. Volmar, Digitale Marktmacht (2019), at 150; see also S. Wismer and A. Rasek, ‘Market Definition in Multi-Sided Markets’, Organisation of Economic Co-operation and Development 2017, at 10.

• 58 Cf. G. Niels and H. Ralston, ‘Two-Sided Market Definition: Some Common Misunderstandings’, 17 European Competition Journal 118, at 119 (2021).

• 59 J. Faull and A. Nikpay (eds.), The EU Law of Competition (2014), para. 1147 ff.; M. Ferro, Market Definition in EU Competition Law (2019), at 116; O’Donoghue and Padilla, above n. 33, at 142.

• 60 Mândrescu, above n. 55, at 125.

• 61 Cf. Niels and Ralston, above n. 58, at 119 ff.

• 62 O’Donoghue and Padilla, above n. 33, at 152.

• 63 Cf. Noel and Evans, above n. 22, at 667 ff.; Wismer and Rasek, above n. 57, at 13 ff.

• 64 Commission decision of 18 July 2018, OJ 2019 C 402/19.

• 65 OECD, ‘The Role and Measurement of Quality in Competition Analysis’ (2013), at 8 ff.

• 66 Commission Notice on the definition of the relevant market for the purposes of Union competition law of 22 February 2024, OJ 2024 C/2024/1645, at https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=OJ%3AC_202401645#ntc131-C_202401645EN.000101-E0131 (last visited 10/04/2024).

• 67 Ibid., paras 94-95.

• 68 Ibid., paras 96-98.

• 69 Cf. Newman, above n. 9, at 1520.

• 70 O’Donoghue and Padilla, above n. 33, at 196.

• 71 D. Coyle, ‘Practical Competition Policy Implications of Digital Platforms’, 82 Antitrust Law Journal 835, at 849 (2019).

• 72 J.-U. Franck and M. Peitz, ‘Market Power of Digital Platforms’, 39 Oxford Review of Economic Policy 34, at 40 (2023).

• 73 Commission decision of 18 July 2018, OJ 2019 C 402/19, at 91.

• 74 O’Donoghue and Padilla, above n. 33, at 193.

• 75 This mostly refers to exclusionary abuses. However, see below for exploitative abuses.

• 76 Faull and Nikpay, above n. 59, para. 4.92.

• 77 Case 85/76, Hoffmann-La Roche & Co. AG v. Commission of the European Communities, [1979] ECR 00461, Rec 91.

• 78 P. Ibáñez Colomo, ‘Anticompetitive Effects in EU Competition Law’, 17 Journal of Competition Law & Economics 309, at 337 (2021).

• 79 Case C-377/20, Servizio Elettrico Nazionale SpA and Others v. Autorità Garante della Concorrenza e del Mercato and Others, ECLI:EU:C:2022:379, Rec 78.

• 80 Cf. Ibáñez Colomo, above n. 75, at 431 ff.; Faull and Nikpay, above n. 59, para. 4.265.

• 81 O’Donoghue and Padilla, above n. 33, at 282.

• 82 Case C-23/14, Post Danmark A/S v. Konkurrencerådet, [2015] CMLR 5, Rec 59.

• 83 Case 27/76, United Brands Company and United Brands Continentaal BV v. Commission of the European Communities, [1978] ECR 207, Rec 250 ff.

• 84 Ibid., Rec 251.

• 85 Case C-117/16, Biedrība ‘Autortiesību un komunicēšanās konsultāciju aģentūra – Latvijas Autoru apvienība’ v. Konkurences padome, Opinion of Advocate General Wahl, ECLI:EU:C:2017:286, paras. 17 ff.

• 86 Case 27/76, United Brands Company and United Brands Continentaal BV v. Commission of the European Communities, [1978] ECR 207, Rec 252.

• 87 Case C-117/16, Biedrība ‘Autortiesību un komunicēšanās konsultāciju aģentūra – Latvijas Autoru apvienība’ v. Konkurences padome, Opinion of Advocate General Wahl, ECLI:EU:C:2017:286, para. 48.

• 88 O’Donoghue and Padilla, above n. 33, at 253.

• 89 Caillaud and Jullien, above n. 16, at 310; R. Funta, ‘Economic and Legal Features of Digital Markets’, 10 Danube: Law, Economics and Social Issues Review 173, at 178 (2019).

• 90 D. Mandrescu, ‘Abusive Pricing Practices by Online Platforms: A Framework Review of Article 102 TFEU for Future Cases’ 10 Journal of Antitrust Enforcement 469, at 496 (2022).

• 91 Case C-7/97, Bronner v. Mediaprint, [1998] ECR I-07791, Rec 40.

• 92 See, for instance, R. Nazzini, ‘Google and the (Ever-Stretching) Boundaries of Article 102 TFUE’, 6 Journal of European Competition Law & Practice 301 (2015) and P. Akman, ‘The Theory of Abuse in Google Search: A Positive and Normative Assessment under EU Competition Law’, 301 Journal of Law, Technology & Policy 2 (2017), which concludes that self-preferencing does not easily fall within one of the traditional forms of abusive conduct; but see also F. Bostoen, ‘The General Court’s Google Shopping Judgment Finetuning the Legal Qualifications and Tests for Platform Abuse’, 13 Journal of European Competition Law & Practice 75 (2022) and L. Hornkohl, ‘Article 102 TFEU, Equal Treatment and Discrimination after Google Shopping’, 13 Journal of European Competition Law & Practice 99 (2022); both seem to argue that self-preferencing could fall within the scope of the abuse of discrimination.

• 93 Commission decision of 27 June 2017, OJ 2018 C 9/11.

• 94 See Case T-612/17, Google LLC e Alphabet, Inc. v. Commission, ECLI:EU:T:2021:763.

• 95 Cf. Nazzini, above, n. 89, at 307; however, according to F. Bostoen, ‘The General Court’s Google Shopping Judgement Finetuning the Legal Qualifications and Tests for Platform Abuse’, 13 Journal of European Competition Law & Practice 75, at 79 (2022), the General Court’s judgement would have categorised the abuse of self-preferencing under the umbrella of the abuse of discrimination. Nonetheless, in more than 80,000 words, the General Court mentioned only once Art. 102 (c) TFEU and never expressly verified this abuse’s legal test.

• 96 Ibid., and for a very thorough analysis of the possible theories of abuse in the Google Shopping case, see Akman, above n. 89.

• 97 Council Regulation 1/2003, OJ 2003 L 1, Art. 7(1).

• 98 Directive 2014/104/EU, OJ 2014 L 349/1.

• 99 O’Donoghue and Padilla, above n. 33, at 1215.

• 100 Regulation (EU) 2019/1150, OJ 2019 L 186.

• 101 M.L. Chiarella, ‘Platform Contracts: Legal Framework and User Protection’, 8 Athens Journal of Law 49, at 56 (2022).

• 102 P. Iamiceli, ‘Online Platforms and the Digital Turn in EU Contract Law: Unfair Practices, Transparency and the (Pierced) Veil of Digital Immunity’, 15 European Review of Contract Law 392, at 402 (2019).

• 103 See, for instance, Regulation (EU) 2019/1150, OJ 2019 L 186, Art. 3.

• 104 Ibid., Arts. 5(1) and (3).

• 105 For the distinction between general and specific transparency obligations, cf. A. Palmieri, Profili Giuridici Delle Piattaforme Digitali. La Tutela Degli Utenti Commerciali e Dei Titolari Di Siti Web Aziendali (2019), at 52.

• 106 See for more details, Schweitzer and Gutmann, above n. 3, at 2.

• 107 Cf. Ibid., at 86 and 93.

• 108 Regulation (EU) 2022/1925, OJ 2022 L 265.

• 109 Ibid., Rec 4.

• 110 P. Ibáñez Colomo, ‘The Draft Digital Markets Act: A Legal and Institutional Analysis’, 12 Journal of European Competition Law & Practice 561, at 563 (2021).

• 111 Cf. Regulation (EU) 2022/1925, OJ 2022 L 265, Arts. 2(21) and 2(20).

• 112 Ibid., Art. 3. On 6 September 2023, the European Commission designated for the first time six gatekeepers, namely Alphabet, Amazon, Apple, ByteDance, Meta, Microsoft. For more details, see https://digital-markets-act.ec.europa.eu/gatekeepers_en (last visited 7 February 2024).

• 113 A. de Streel and others, ‘Effective and Proportionate Implementation of the DMA’, Cerre (Centre on Regulation in Europe) 2023, at 182.

• 114 T. Tombal, ‘Economic Dependence and Data Access’, 51 IIC – International Review of Intellectual Property and Competition Law 70, at 74; P. Bougette, O. Budzinski and F. Marty, ‘Exploitative Abuse and Abuse of Economic Dependence: What Can We Learn from an Industrial Organization Approach?’, 129 Revue d’économie politique 261, at 271 (2019); P. Këllezi, ‘Abuse below the Threshold of Dominance? Market Power, Market Dominance, and Abuse of Economic Dependence’, in M.-O. Mackenrodt, B. Conde Gallego & S. Enchelmaier (eds.), Abuse of Dominant Position: New Interpretation, New Enforcement Mechanisms?, (2008) 55, at 69.

• 115 L. Feteira, The interplay between European and national competition law after regulation 1/2003 (2016), at 150.

• 116 S. Scalzini, ‘Economic Dependence in Digital Markets: EU Remedies and Tools’, 5 Market and Competition Law Review 81, at 86 (2021).

• 117 V. Kathuria and J.C. Lai, ‘User Review Portability: Why and How?’, 34 Computer Law & Security Review 1291, at 1297 (2018).

• 118 C. Dellarocas, ‘The Digitization of Word of Mouth: Promise and Challenges of Online Feedback Mechanisms’, 49 Management Science 1407, at 1407 (2003).

• 119 F. Cafaggi and others, ‘Study on the Legal Framework Covering Business-to-Business Unfair Trading Practices in the Retail Supply Chain: Final Report’, European Commission Directorate-General for Internal Market 2014, at 43.

• 120 Tombal, above n. 111, at 74.

• 121 Këllezi, above n. 111, at 61 ff.; cf. also Cafaggi and others, above n. 116, at 51 ff.

• 122 H. Friederiszick and S. Reinhold, ‘The Economics of Dependence: A Theory of Relativity’, ESMT Working Paper, 2021:21-02, at 5.

• 123 K. Markert, ‘GWB § 20 Verbotenes Verhalten von Unternehmen Mit Relativer Oder Überlegener Marktmacht’, in U. Immenga and E.-J. Mestmäcker (eds.), Wettbewerbsrecht (2020), para. 16; Friederiszick and Reinhold, above n. 119, at 6.

• 124 Markert, above n. 120, para. 14.

• 125 Scalzini, above n. 113, at 91.

• 126 Ibid.

• 127 M. Maugeri, ‘DdL concorrenza e piattaforme digitali’, Persona e Mercato 1 (2022), http://www.personaemercato.it/ddl-concorrenza-e-piattaforme-digitali-brevi-considerazioni-sulla-proposta-di-modifica-della-disciplina-sullabuso-di-dipendenza-economica-di-marisaria-maugeri/ (last visited 1 February 2024).

• 128 Feteira, above n. 112, at 151-8 and 170-1.

• 129 Cf. Markert, above n. 120, para. 1.

• 130 Cf. Scalzini, above n. 113, at 83.

• 131 C. Grynfogel, ‘Droit Français des Abus de Domination – Article L. 420-2 du Code de Commerce’, JurisClasseur Commercial, para. 181 (2017).

• 132 I. Graef, ‘Differentiated Treatment in Platform-to-Business Relations: EU Competition Law and Economic Dependence’, 38 Yearbook of European Law 448, at 490-1 (2019).

• 133 M. Libertini, ‘La responsabilità per abuso di dipendenza economica: La fattispecie’, 29 Contratto e Impresa 1, at 14 (2013).

• 134 Graef, above n. 129, at 491.

• 135 Markert, above n. 120, para. 14.

• 136 Grynfogel, above n. 128, para. 169.