EUROPEAN COMMISSION

Brussels, 15.12.2020

SWD(2020) 363 final

COMMISSION STAFF WORKING DOCUMENT

IMPACT ASSESSMENT REPORT

Accompanying the document

Proposal for a

REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL

on contestable and fair markets in the digital sector (Digital Markets Act)

{COM(2020) 842 final} - {SEC(2020) 437 final} - {SWD(2020) 364 final}

GLOSSARY

CONTENTS

1.    INTRODUCTION: POLITICAL AND LEGAL CONTEXT    

1.1.    Political context    

1.2.    Field of intervention    

1.3.    Relationship of the initiative with other ongoing initiatives    

2.    PROBLEM DEFINITION    

2.1.    What are the problems?    

2.1.1.    Weak contestability of, and competition in, platform markets, or risk thereof    

2.1.2.    Unfair gatekeeper practices vis-à-vis business users    

2.1.3.    Legal uncertainty for market players    

2.2.    What is the size of the problem?    

2.3.    Problem drivers    

2.3.1.    Market failures    

2.3.2.    Fragmented regulation and oversight    

2.3.3.    Conclusion: problem drivers’ effects    

2.4.    How will the problem evolve?    

3.    WHY SHOULD THE EU ACT?    

3.1.    Legal basis    

3.2.    Subsidiarity: necessity and added value of EU action    

4.    OBJECTIVES: WHAT SHOULD BE ACHIEVED?    

4.1.    General objective    

4.2.    Specific objectives    

4.2.1.    Address market failures to ensure contestable and competitive digital markets for increased innovation and consumer choice    

4.2.2.    Address gatekeepers’ unfair conduct    

4.2.3.    Enhance coherence and legal certainty to preserve the internal market    

4.3.    How do the objectives link to the problems identified?    

5.    WHAT ARE THE AVAILABLE POLICY OPTIONS?    

5.1.    What is the baseline from which policy options are assessed?    

5.2.    What are the main parameters that determine the range of available policy options?    

5.2.1.    Scope: core platform services, gatekeepers, thresholds    

5.2.2.    Unfair practices    

5.2.3.    Speed and Flexibility    

5.2.4.    Enforcement framework    

5.2.5.    Summary and main trade-offs    

5.3.    Policy options    

5.3.1.    Option 1 – Pre-defined list of gatekeepers and immediately applicable obligations    

5.3.2.    Option 2 – Partially flexible framework of designation and updating of obligations, including regulatory dialogue for implementation of some    

5.3.3.    Option 3 - Flexible option based exclusively on qualitative scoping thresholds    

5.4.    Policy options discarded at an earlier stage    

5.4.1.    A broad scope across platforms    

5.4.2.    Information obligations without addressing unfair conduct    

5.4.3.    A broad scope across markets    

6.    WHAT ARE THE IMPACTS OF THE POLICY OPTIONS?    

6.1.    Internal market    

6.2.    Growth and productivity    

6.3.    Competition and Innovation    

6.4.    International trade    

6.5.    Employment    

6.6.    Businesses    

6.6.1.    Gatekeepers    

6.6.2.    SMEs    

6.7.    Consumers    

6.8.    Regulatory Authorities    

7.    HOW DO THE OPTIONS COMPARE?    

7.1.    Effectiveness    

7.2.    Efficiency    

7.3.    Coherence    

7.3.1.    Coherence with the Digital Strategy    

7.3.2.    Coherence with the DSA    

7.3.3.    Coherence with other instruments    

7.4.    Proportionality    

7.5.    Subsidiarity    

7.6.    Conclusion    

8.    PREFERRED OPTION    

8.1.    The main building blocks of the preferred option    

8.2.    The scope of application    

8.2.1.    Identification of core platform services    

8.2.2.    Designation of gatekeepers    

8.2.3.    Obligations applicable to gatekeepers’ core platform services    

8.3.    Enforcement framework and governance    

9.    HOW WILL ACTUAL IMPACTS BE MONITORED AND EVALUATED?    

1.INTRODUCTION: POLITICAL AND LEGAL CONTEXT

1.The digital transformation has profoundly changed the functioning of the global economy and society. The Covid-19 crisis and the increased importance and use of digital services has only further evidenced the importance of ensuring a borderless, fair, and contestable Single Market for digital services where companies can thrive and where citizens have genuine choices and control.

2.This Impact Assessment examines the possible policy options to ensure a competitive Single Market for digital services and in particular fair and contestable platform markets. It combines the assessment of two initiatives previously presented in separate Inception Impact Assessments: (i) the Digital Services Act (‘DSA’) package: ex ante regulatory instrument of very large online platforms acting as gatekeepers; 1 and (ii) the New Competition Tool. 2 These two initiatives have been subject to two parallel public consultations. 3  

3.Given the breadth of the topics covered, both Inception Impact Assessments – including their respective consultations – were initially published separately. However, since the outset, both consultations were aimed at complementary solutions by “ensur[ing] a joint analysis of the results”, “with a view to exploring synergies and ensuring consistency on the policy options pursued, in particular as regards possible remedies and enforcement”. 4 The holistic approach presented in this impact assessment is the result of such exercise.

Political context

4.Over the last years, a wide range of studies at international level as well as by National Competition Authorities (‘NCAs’) have brought to the fore the acute problems afflicting digital markets in terms of contestability as well as of the fact that a number of large platforms are taking advantage of their position to restrict competition including by means of the imposition of unfair conditions on their trading partners and on consumers.

5.In this respect, the Commission has initiated a reflection process about the role of competition policy in a fast-changing world, which included commissioning a report from a group of independent Special Advisers to Commissioner Vestager published in April 2019. Among other aspects, the report concluded that “the specificities of competition in the digital world […] make market power “sticky”, and there is legitimate fear that the market power [large platforms] have acquired will be hard to challenge. Furthermore, they have been able to build, on top of their core competencies, entire ecosystems which make it hard for new entrants to compete on the merit and which, many observers feel, face little competitive pressure”. 5  

6.A subset of issues pertaining to all digital platforms had previously been addressed through regulation in the so-called ‘Platform-to-Business Regulation’ (‘P2B Regulation’), aiming to increase transparency and fairness in platforms that can easily hold asymmetric bargaining power. To analyse further emerging issues addressed in the independent Special Adviser Report, the Commission also established the ‘EU Observatory on the Online Platform Economy’bservatory on the Online Platform Economy supported by an expert group 6 , to support the Commission in monitoring and analysing the developments in the online platform economy. Evidence gathered by this expert group further confirmed the findings of the previous reports.

7.Similar reflections are taking place in some of the EU’s major trading partners, including the United States of America (‘US’), Japan, the United Kingdom (the ‘UK’), Australia and China. These reflections include calls for a new regulatory framework for platforms with “significant and durable market power” (US House of Representatives Majority Staff report 7 ), “substantial market power” (ACCC report 8 ), “strategic market status” (Furman report 9 ) and “bottleneck power” (Stigler Center report 10 ). The US report notably concludes that each investigated platform now serves as main gateway to consumers and other businesses that each platform uses this role as major gateway to maintain its market power; and that the firms have abused their role as intermediaries to further entrench and expand their dominance.

8.The need to address these concerns in digital markets was expressed in Commission President von der Leyen’s mission letter for Executive Vice-President Vestager, where she stated that in “striving for digital leadership, we must focus on making markets work better for consumers, business and society”. The letter tasked Executive Vice-President Vestager with ensuring “that competition policy and rules are fit for the modern economy”. 11  

9.This objective was reiterated in the Commission’s Communication Shaping Europe's digital future, as “it is important that the competition rules remain fit for a world that is changing fast, is increasingly digital and must become greener”. 12 In the same Communication, the Commission further stated that “competition policy alone cannot address all the systemic problems that may arise in the platform economy”. Against this background, the Commission also announced that it “will further explore, in the context of the DSA package, ex ante rules to ensure that markets characterised by large platforms with significant network effects acting as gatekeepers, remain fair and contestable for innovators, businesses, and new market entrants”. 13

10.In the European Parliament, the Committee on the Internal Market and Consumer Protection (‘IMCO’), the Committee on Civil Liberties, Justice and Home Affairs (‘LIBE’) and the Legal Affairs Committee (‘JURI’) published draft reports in April and in May 2020, as legislative own-initiative reports. 14 The final IMCO and LIBE Committees reports were adopted in September 2020 15 and the draft JURI report in October 2020. 16 In parallel to these reports, the European Parliament also adopted a resolution on competition policy on 18 June 2020, where it “calls on the Commission to assess the possibility of imposing ex ante regulatory obligations where competition law is not enough to ensure contestability in these markets”. 17 The Digital Markets Act (‘DMA’), by proposing ex ante rules for certain large platforms and aiming at ensuring fair and contestable digital markets, responds to these calls for action.

11.The European Council confirmed the need to act in its New Strategic Agenda 2019-2024, by stating that “[w]e will continue to update our European competition framework to new technological and global market developments”. 18 The Council of the European Union (‘Council’) also “supports the Commission’s intention to collect evidence of the issue and further explore ex ante rules to ensure that markets characterised by large platforms with significant network effects, acting as gate-keepers, remain fair and contestable for innovators, businesses and new market entrants”. 19

12.Furthermore, the Council welcomed the public consultation on a ‘New Competition Tool to address structural competition problems across markets’ and expressed its willingness to discuss the Commission’s proposal for a DSA Package. 20 The Council underlined “that new policy approaches for the Single Market have to be fit for the digital age [and] able to cope with new and agile business models, especially in the digital economy”. 21 Finally, the Council reiterated the importance of swift action on the DSA package in its most recent conclusions, in which it “looks forward to the Commission’s proposal for a Digital Services Act by the end of this year”. 22  

Field of intervention

13.The feedback and evidence collected pointed to an urgent need to act in the digital sector, due to the particular features of digital markets. On that basis, the present impact assessment focuses on intervention options with regard to digital markets, with a focus on those markets characterised by the presence of large digital platforms where problems are most prominent, and action appeared most pressingly needed.

14.In the digital sector, there is a small number of online platforms – often embedded in their own ecosystems – which have come to play a crucial role in the lives of millions – if not, billions – of individuals and companies. They intermediate a significant portion of transactions between consumers and businesses, and have emerged as a key structuring element of today’s digital economy. As such, these platforms have a major impact on, control the access to, and are entrenched in digital markets, leading to extreme dependencies of many businesses on these important platforms. The evidence points to negative effects on effective competition and on the contestability of the markets concerned. Member States in the EU observing these tendencies have begun to take regulatory initiatives to address these effects, potentially fragmenting the Internal Market.

15.The scope of this initiative is limited to the digital sector. In fact, the market concentration tendencies and the underlying market dynamics in the digital sector, as well as other characteristics of digital markets, have contributed to several market failures in this area, which are likely to lead to inefficient market outcomes in terms of higher prices, lower quality, less choice and innovation to the detriment of European consumers (see Section 2.2).

16.Even though some of relevant market features are also observed to some extent in other markets, they are most prevalent in digital markets. They include market features such as extreme scale economies, often resulting from nearly zero marginal costs to add customers and business users –- in contrast to off-line business models where such upscaling would involve major investments – and strong network effects associated to the multi-sidedness of online platforms, as well as data driven-advantages that often fundamentally change the competitive process, leading to sudden and radical decreases in competition (see Section 2.3.1). The presence of large platforms, often vertically or horizontally integrated in large ecosystems, exacerbates the negative effects that these features can trigger, thus making it impossible for the markets to self-correct.

17.The problems in the digital sector are also most pressing from an internal market perspective. In fact, the OPC and the targeted consultation of NCAs have largely shown that the most salient examples of market failures today stem from the digital sector. 23 Consumer organisations like BEUC have also prominently flagged the particular concerns surrounding digital markets. 24 Likewise, digital markets featured prominently in the expert reports commissioned for the Impact Assessment. 25

18.On the basis of the available evidence, including the Commission’s regulatory and competition enforcement experience, the Commission has mapped a number of ‘core platform services’ which exhibit these features and where absent regulatory intervention the identified market failures would effectively remain un-addressed.

19.These ‘core platform services’ are those where the problems identified in Section 2.1 are most evident and prominent and where the presence of a limited number of large online platforms that serve as gateways for business users and customers has led or is likely to lead to weak contestability of markets. On the basis of the evidence collected for this Impact Assessment (see Section 5.2.1), the screening of problems led to the identification of the following core platform services: (i) online intermediation services (including marketplaces and app stores) (ii) online search engines, (iii) social networking (iv) video sharing platform services, (v) number-independent interpersonal electronic communication services, (vi) operating systems, (vii) cloud services and (viii) advertising services, including advertising intermediation services, provided by providers of one or more of the above.

20.Furthermore, as explained in Section 5.2.1, it should be possible to designate gatekeepers whenever it can be demonstrated that a provider of core platform services: (i) has a significant impact on the internal market, (ii) operates one or more important gateways to customers and (iii) enjoys or is expected to enjoy an entrenched and durable position in its operations.

21.While this Impact Assessment focuses on issues caused by gatekeepers operating in digital markets, it does also fully recognise the benefits that online platforms bring to the economy and society. The purpose of the DMA initiative is therefore to allow these platforms to unlock their full potential by addressing the most salient incidences of unfair practices and weak contestability so as to allow consumers and business users alike to reap the full benefits of the platform economy. 

Relationship of the initiative with other ongoing initiatives

22.In parallel to the present Impact Assessment, the Commission is also presenting an Impact Assessment for the Digital Services Act (‘DSA’), an initiative seeking to address primarily societal risks of digital markets and, more specifically, of very large platforms. The definition of ‘gatekeepers’ in the DMA is different in nature and scope from the definition of ‘very large platforms’ falling within the scope of the asymmetric obligations under the DSA. Whilst a handful of gatekeepers may be subject to both the DSA and the DMA, the risks addressed by the DSA and DMA are, however, very different. The DMA addresses risks to contestability and fairness in digital markets where gatekeepers as defined are present. The DSA addresses risks derived from the fact that very large platforms have become de facto public spaces, playing a systemic role for millions of citizens and businesses, creating a need for more accountability for the content which these providers distribute on their platforms. The different risks that both initiatives seek to tackle also translate in different obligations, the content and applicability of which is clearly distinguishable. 

23.The impact assessment of the DMA is also being conducted in parallel with a number of ongoing reviews of certain competition rules, most notably the review of the Block Exemption Regulations for horizontal and vertical agreements, including in the motor vehicle sector. 26 These reviews are without prejudice to the Impact Assessment of the DMA, an initiative of a regulatory nature that does not affect – and is not affected by – those reviews. The Block Exemption Regulations pursue a different objective. They apply Article 101(3) of the TFEU by regulation to certain categories of agreements falling within Article 101(1) TFEU, thereby block exempting them from the application of Article 101(1) TFEU.

24.Also ongoing is the review of the Market Definition Notice, which is without prejudice to the Impact Assessment of the DMA. 27 The Market Definition Notice pursues a different objective since it is a soft law instrument providing guidance as regards how the Commission “applies the concept of relevant product and geographic market” in its enforcement of EU competition law. 28

2.PROBLEM DEFINITION 

What are the problems?

25.Over the past decade, online platforms have established their presence as important economic players and boosting efficiency, as well as spurring innovation and the development of new business models. Online platforms play an important role in many industries, allowing buyers and sellers of goods and services to trade and communicate with each other. They increase consumer choice and convenience, improve efficiency and competitiveness of industry and can enhance civil participation in society. Online platforms are key drivers of innovation in the digital world and their success is closely tied to the success of a range of businesses that use platforms to reach customers. Platforms allow especially smaller businesses to extend their operations beyond their home state, catering for consumers across the entire Single Market. 29  

26.At the same time, they also raise new issues relating to fairness, transparency and market distortions. According to the evidence collected by the Commission, digital markets are particularly vulnerable to the following three problem clusters.

27.First, in many digital markets large digital providers have emerged as gatekeepers serving as gateways for their business users and consumers. Some of these gatekeepers exercise control over whole platform ecosystems that are essentially impossible to contest by existing or new market operators, irrespective of how innovative and efficient they may be (‘weak contestability of platform markets’). As a result of the weak competitive pressure experienced by these large players, the likelihood increases that these markets do not function well – or may soon fail to function well – and thus do not deliver the best outcome for consumers in terms of prices, quality, choice, and innovation (weak competition in digital markets, or risk thereof).

28.Second, many businesses are increasingly dependent on these gatekeepers, which in many cases leads to gross imbalances in bargaining power and, consequently, unfair practices resulting in conditions for business users that would not be achievable under normal circumstances (‘unfair business conditions for business users’). 30 Such imbalances in bargaining power, coupled with the economic dependency of many business users and costumers on gatekeepers, allow the latter to obtain conditions that they would not be able to obtain in case of well-functioning and competitive markets.

29.Third, digital players typically operate at a global scale and deploy global business models. As a result, different national legislations within the EU 31 may lead to increased regulatory fragmentation and increased compliance costs for these large market players and the business users that rely on them (‘fragmented regulation and oversight’). Smaller players and startups are also negatively affected by this situation, as it impedes them from scaling up easily within the Internal Market in order to grow into contenders vis-à-vis the large, established players in the market.

30.This section describes each of the three problem clusters above in more detail.

Weak contestability of, and competition in, platform markets, or risk thereof

31.The market features of digital markets tend to favour the emergence of a few large firms that have become gatekeepers for many digital products and services. These gatekeepers represent a key segment of the digital economy and play an important role in providing third parties with online access to a large number of European consumers. 32 Where such markets have not yet gravitated towards high concentration, they are at the risk of doing so in the near to medium term.

32.There is evidence for a trend of growing market concentration (and, relatedly, growing mark-ups) at the industry level, which has been documented both for the US and for the EU. 33 In digital markets in particular, the level of concentration of economic power is unprecedented: the top seven of the large platforms account for 69% of the total EUR 6 trillion valuation of the platform economy, as a result of vertical and horizontal integration. 34 Large online platforms intermediating between businesses and consumers are growing at an exponential pace. They have several hundreds of millions of users (both businesses and citizens/consumers). 35 Total net revenues of some of these platforms (of billions of euros) double and triple over a few years. Moreover, five out of the world’s ten largest companies by market capitalisation are digital conglomerates (see Figure 3 ).

33.Several respondents, including startups, research institutes and trade associations, point out the positive impact of platforms on startups: by lowering the barriers to entry and extending to companies of all sizes the advantages of cost and speed that can be gained from trading online, they stimulate innovation and the dissemination of new products and technologies. 36 Nevertheless, the large majority of respondents to the OPC 37 and NCAs 38 broadly agreed that “one or few large players on the market (i.e. concentrated market)” constitutes a very important or important source or part of the reasons for market failures. In certain markets, it may be challenging to maintain ‘competition in the market’, notably where having only one network may be the most beneficial outcome for consumers. However, in such a situation it is essential to keep ‘competition for the market’ open. Any successful attempt by a firm to lock in a group of consumers, so that the market is no longer contestable for a new entrant, will prevent such ‘competition for the market’, with possible adverse consequences for prices, quality, choice and innovation. 39  

34.Large gatekeepers benefit significantly from the entry barriers characterising digital markets. In this context, new market operators that may want to enter or expand in digital markets where a gatekeeper is present may find it extremely difficult to overcome some of the inherent barriers to entry or expansion without access to a sufficiently large user base. 40 For instance, a new entrant must convince a sufficient number of users (due to the importance of network effects) to coordinate their migration to a new service, taking e.g. part of the social network along, or other associated data assets such as purchase or preference histories, or ratings. This lack of contestability due to high barriers to entry is extensively echoed in the academic literature. 41

35.These gatekeepers therefore have an entrenched market position, which is hard to contest, and which they further expand through the creation of ecosystems. The largest platform companies are active across many different markets, creating extended data-driven ecosystems around their core activities, often cross-subsidising one service with data or revenues from another. In this regard, a large number of respondents identified online intermediation services, search engines, operating systems for smart devices, consumer reviews, network and/or data infrastructure/cloud services, digital identity services, online advertising intermediation services, payment services, fulfilment services and data management platforms as activities that can strengthen the gatekeeper role of such large online platforms when any or all of these are integrated within a single corporate structure. 42

36.It is sometimes argued that incumbent offer their services often for free and that competition is ‘just one click away’ or that it is vigorous in some segments. This is a too narrow and selective view of the overall dynamics of the digital platform economy. However, the entrenched position of gatekeepers has shown to be lasting and essentially unchallenged by competing platforms, thus leading to weak inter-platform competition.

Unfair gatekeeper practices vis-à-vis business users

37.Gatekeepers’ successful business models based on platform economy specificities have allowed them to gain strong market positions and economic power, enabling them to create ecosystems for which they set the rules by which other economic players should abide. If set in an unfair manner, these rules can be detrimental to business users, and limit Small and Medium Enterprises’ (‘SMEs) online visibility and associated sales.

38.The enforcement experience and input to the OPC show that unfair practices can take on different forms. They could relate to gatekeepers’ size, their capacity to acquire and monopolise data, imposition of contractual conditions or preferential treatment and/or the interplay between these elements. For a full overview of different unfair gatekeepers’ practices and the ongoing and closed investigations and antitrust cases please see Section 5.2.2 and Annex 5.6 to the Impact Assessment.

39.One of such unfair practices is the imposition on business users of ‘anti-steering’ provisions, by which gatekeepers prevent business users from directing acquired consumers to offers other than those provided on the platform, even though such alternative offers may be cheaper or otherwise potentially more attractive. 43 For instance, an app store that does not allow its business users to advertise alternative subscription options outside its platform to acquire customers. Also, cross-platform parity clauses, 44 i.e. clauses that oblige business users to offer the same or better retail conditions as those offered on other platforms to the contract party, tend to disincentivise competition between platforms. In particular, they prevent business users from ‘rewarding’ other platforms that may provide better or cheaper platform services, by offering better retail prices or conditions on those platforms.

40.Another example is the imposition of the platform’s ID services, which is a lock-in strategy where the user is required to sign up/register with an email service of the gatekeeper’s core platform services when using another of its products (e.g. an operating system, social network). The US House of Representatives Judiciary Committee also describes lock-in strategies including free tier offerings for cloud services. 45

41.The broad category of ‘self-preferencing’ refers to practices in which a usually vertically integrated gatekeeper acting in the dual role of providing core platform services to business users and at the same time competing with them when providing ancillary services applies more favourable conditions to its own services compared to the third-party services hosted on the gatekeepers’ platform. Self-preferencing occurs in many situations in the online and offline world (e.g. in supermarkets with own brands). Such behaviour may not be considered generally anti-competitive under the EU competition rules or unfair in all business relationships. However, certain forms of self-preferencing may amount to an unfair business practice. An important concern here is the fair balancing of interests, in this case those of the gatekeeper platforms versus that of their business users. 46 In particular, the special position of gatekeeper platforms that play a dual role and may engage in favouring their own services may lead to the exclusion of alternatives by business users that are largely dependent on these gatekeeper platforms to reach consumers, reducing choice for them, and potentially undermining the quality of service and increasing prices. 47  

42.One example is an app store, which markets a number of its own popular apps and at the same time maintains a marketplace for competitors, self-preferencing its own marketplace by applying more favourable policies for its own products and selectively drafting rules favouring its own products. Another example is a search engine or marketplace treating more favourably its own products and services in the results displayed to end users.

43.Feedback to the OPC shows that business users consider self-preferencing to be a very common practice deployed by large, vertically integrated platforms. Responses by business users suggest that search and ranking algorithms often give preference to the platform’s own services, but also that a platform often has an incentive to bias its recommendations towards the content provider charging a lower royalty. 48

44.Business users are faced with limited or no access to vast amount of data (e.g. app store limiting the information that third-party app providers receive about their subscribers) as well as lack of any or meaningful interoperability to access such data that may be collected by gatekeepers.

45.The Impact Assessment study and input to the OPC point to practices that prevent both consumers and business users from switching. In the digital sector, being able to port both historical and real-time data is an important precondition for both multi-homing and switching. Business users and consumers alike repeatedly raise the issue of not being able to use any other platform or service because the incumbent refuses to provide an enhanced and continuous real-time ability to port personal and non-personal data in interoperable format. These practices affect contestability, and limit business users’ possibilities to move to or rely on alternative platforms or services. As yet another example, an advertising intermediation services provider collecting multiple datasets from business users’ services which it uses for better targeting and attribution measurement, but does not share them with advertisers.

46.Another example of a data related practice that could be considered unfair, and has been raised by many stakeholders in the context of the OPC, is the situation where a gatekeeper restricts business users from accessing and using the data that they provide, receive from their customers or generate in the course of their use of the gatekeeper’s platform or service, as is the case as regards an app store limiting the information that third-party app providers receive about their subscribers or an online intermediation service restricting data generated in the curse of the use of its platform by third party sellers and their customers. Feedback to the OPC shows that business users are regularly confronted with the imposition by large platforms of proprietary services and an authentication through the platform even when third party services are used to create a direct link with customers to the detriment of third-party providers. Respondents suggest that gatekeepers exclude business users from access to user data and attempt to remove the direct link between the client and third party suppliers (so-called disintermediation). 49

47.Other examples include (i) gatekeepers that use certain data that they received from business users for a particular use, for instance advertising services, for other, unrelated purposes, (ii) gatekeepers operating a marketplace benefit from their dual role and ability to evaluate product, sales and customer data generated from the sales of goods provided by third party merchant business users on its marketplace, or (iii) gatekeepers operating a video sharing platform that has access to a rich set of (first party) data about its consumers, data that it can re-use to improve its own products, including in other areas, but restricts the access to this rich set of data to its competitors.

48.Thanks to their strong market position, gatekeepers, can either limit access to their platform or make such access conditional upon specific requirements. Gatekeepers often impose unfair terms of access to business users, for instance in relation to price for the services they offer or accepting specific bundles which do not allow the mix-and-match by customers (e.g. a provider of cloud services bundling this service with other services or a social network services provider applying terms and conditions, which make the use of its services conditional on the possibility to collect and combine user data from multiple sources).

49.Another example is gatekeepers limiting the access to or the interoperability of certain of their platform services/functionalities (e.g. operating system) with the services offered by business users, reserving those functionalities to their own services. For example, in certain circumstances third-party providers of payment wallets may require access to near-field-communication (‘NFC’) functionalities in the hardware and the gatekeeper exclusively reserves such functionality to its own services.

50.The work supporting this Impact Assessment shows that the above problematic practices are most prominent in relation to the following core platform services: e-commerce marketplaces, online search engines, app stores, social networks, video-sharing services, operating systems, cloud, number-independent messaging services, and online advertising. 50 Also the negative effects of the problem drivers are most severe in relation to these core platform services. These services present characteristics that have been identified as driving the problematic practices assessed in this document, i.e. they have a multi-sided character, which allows them to accumulate data on all sides of the market and thus benefit from strong indirect network effects (e.g. social network services, video sharing services); they have an important intermediation function (e.g. marketplaces, app stores) and/or playing important ‘visibility’ role (e.g. online search, operating systems), and are characterised by the presence of big ‘gatekeepers’ (e.g. number-independent messaging services) which often are vertically integrated and operate a large ecosystem (e.g. cloud services).

51.From the foregoing, it becomes apparent that there is substantive evidence concerning the urgent need to address unfair practices by gatekeepers. The fact that this evidence cuts across not only jurisdictions of both common and civil law, but also across numerous enforcement bodies of different kinds and with diverse mandates is yet another indication of the urgency underlying the intervention. In fact, as outlined by the evidence quoted in this impact assessment and the numerous reports, there are only few other questions currently triggering a similar level of consensus between enforcement authorities, judicial bodies and law makers around the world like the need to tackle problems related to digital gatekeepers.

Legal uncertainty for market players

52.While in many areas of the single market, the objective is to ensure further integration, the online platform market is naturally integrated (due to the intrinsically cross-border nature of the platform economy). However, there is an increasing regulatory fragmentation of the online platform space in the EU (see Annex 5.4 and 5.5 to the Impact Assessment). In addition, coordination among national legislators may be insufficient, leading to potentially heterogeneous responses across the EU.

53.Such fragmentation becomes problematic where it creates increased compliance costs for all market players. This is particularly harmful for smaller platforms and startups, potential entrants and smaller business users since it creates regulatory barriers to entry and limits their ease of scaling up across the Single Market. At the same time, diverging laws may also endanger the benefits stemming from large platforms’ activity; such costs may imply regulatory shopping, ultimately resulting in unequal impacts on EU consumers. If emerging platforms are unable to grow sufficiently in order to compete with gatekeepers, the latter would be able to further gain power, strengthening their ability to establish market rules and (potentially) behave unfairly. This would exacerbate the above described issues of weak market contestability and competition, as well as unfair business practices.

54.In the OPC, respondents from all categories mention that EU level rules would prevent further legal fragmentation across Member States, considering that several Member States have already started to introduce new rules to address concerns arising from the presence of gatekeepers. Stakeholders generally consider that an effective coordination between EU bodies and the relevant national regulatory authorities is needed, especially in light of the fact that issues related to gatekeepers are likely to have an important cross-border component. Platforms in particular point out the need to minimise fragmentation and allow for a pan-European approach. 51

What is the size of the problem?

55.As explained in more detail in Section 2.1.1, the characteristics of digital markets often favour the lack of market contestability and the emergence of strong concentration, which tends to be accompanied by rising mark-ups and weaker competition. The trend of increasing industry concentration has been documented for both digital and non-digital industries alike. For instance, in 2014 the mean European high ‘digital intensity’ industry 52 had 4 percentage point higher sales concentration than in 2000. 53  

56.As regards trends in mark-ups, empirical studies suggest that company mark-ups have increased by 4% to 6% for the period 2001-2014, on average across country, 54 and that the result is mainly driven by the top of the mark-up distribution in the digital sector. 55 For the top 10% of the firms in the sample, the growth in mark-ups over the period 2001-2014 amounted to 20%, while the remaining firms in the sample exhibit a flat trend, i.e. mark-ups stayed roughly the same. 56 To the extent that this observed trend of increasing market power of this top 10% of firms is a sign of insufficient competitive constraints faced by these firms, increasing competition in these markets could contribute to slowing down the growth trend in mark-ups, decrease prices and increase choice, quality and innovation. For example, a recent study shows that more concentrated industries also feature a more negative relation between markups and investment and innovation. 57  

57.As regards the size of the problems related to unfair practices by gatekeepers (explained in Section 2.1.2), it is importance to note that the number of merchants and small businesses affected by gatekeepers’ conduct varies depending on the sector, but can be estimated to reach between one and four million. 58  

58.A good indicator of businesses’ dependence on platforms is turnover from sales and share of revenue via online platforms as a proportion of the company’s total revenue from e-commerce. According to the Observatory’s estimates, around half of enterprises derived more than 25% of their revenues from online platforms. For almost 10% of companies, online platform sales exceed 75% of all revenues; while according to Statista estimates, in 2017, 18% of company revenues across the EU-28 came from e-commerce, the highest proportion being 33%. 59

59.Another indicator of businesses’ dependence on platforms is the use of platforms to publish online advertising. Of SMEs in the EU that sell online, more than eight in ten rely on search engines as a mean of marketing their products or services. In 2018, an average of 26.2% of enterprises across the EU paid to advertise online. In northern European countries, such as Sweden and Denmark, this figure was over 44%. 60

60.The degree to which businesses have integrated into and depend on the platform economy is further illustrated by the fact that in some cases more than 50% of goods sold on a marketplace come from third-party sellers. There are over 26.4 million software developers in the world who depend entirely on large platforms providing the infrastructure and setting the rules for the distribution of their apps.

61.Gatekeepers’ unfair practices affecting businesses do not represent a one-off problem, but are systemic and recurrent. In the OPC on Ex Ante Rules, 88% of the businesses and business users that replied, encountered issues concerning trading conditions on large platforms. 61  According to Cullen International’s database, around 30 antitrust investigations concerning platforms have been formally opened in the EU (by DG Competition or NCAs) since 2015. 62 However, the prevalence of unfair practices by large gatekeeper platforms is evidenced not only in the number of cases that have been investigated by competition authorities, but also from the interviews and case studies run in the context of the support study for the Impact Assessment. 63

62.As regards the anti-competitive use of third party data, both the Commission and NCAs (Italy, the Netherlands and Germany) are running a number of investigations against four different large online platforms. In case of preferential treatment, except for the Commission cases, also at least two NCAs (Italy and the Netherlands) are running investigations concerning self-preferencing by online platforms. With respect to inadequate or late access to own (business-user related) data and lack of access to key functionality, the investigations were initiated in each case against three different online platforms in three different Member States, while anti-steering and MFN clauses were recently subject to investigations against two different online platforms.

63.The size of the market and the implications of the issues identified are also an angle of approaching the ‘the size of the problem’ question. The digital economy is estimated to account for between 4.5% to 15.5% of global Gross Domestic Product (‘GDP’) in 2019, depending on the definition. 64 The top 50 online platforms, representing an average of over 60% of traffic share 65 across the Member States, achieved worldwide revenues of almost EUR 276 billion in 2018, and employed almost 600 000 people. 66  Online platforms’ role is constantly increasing due to e-commerce upward trends; it has further strengthened with the widely introduced lockdowns due to the COVID 19 outbreak in 2020; consumers have shifted their habits more towards search engines, social media and online entertaining media. 67

64.The ongoing fragmentation in the Digital Single Market might also reverse the positive trends in cross-border online trade. Assuming a 10% decrease per year in online cross-border trade, the opportunity cost of the digital market fragmentation would be EUR 1.76 trillion after 10 years (see Annex 5.5 to the Impact Assessment).

65.The reduced contestability of digital markets in which gatekeepers operate seems to result in suboptimal innovation levels, with notably implications for societal welfare. 68 Relevant data supports the view that many markets are becoming more concentrated and display less competition. Profit margins are widening, with a few firms reaping a significant share. Innovation levels are also sub-optimal. 69  

66.Gatekeepers bring benefits for consumers in terms of convenience, increased choice of free of charge online products and services. However, there are also important adverse consequences for consumers, namely reduced choice in terms of number of competitive platforms 70 , insufficiently informed choice decisions 71 , and lack of data/privacy-friendly services 72 . Also, due to the distorted intra- and inter-platform competition described in the problem definition, consumers risk experiencing higher prices and/or less quality.

Problem drivers

67.This section describes the problem drivers for the problem clusters described in Section 2.1. These problem drivers can be grouped into two overarching categories, namely gatekeeper related market failures (Section 2.3.1), and fragmented regulation and oversight (Section 2.3.2).

Market failures

68.From the competition law perspective, the term ‘market failure’ indicates a situation in which a market does not allow consumers to benefit from the results of effective competition in terms of low prices, better quality, as well as more choice and innovation, while firms are able to earn supra-normal profits which are not competed away over time. 73  

69.While markets typically feature self-correcting mechanisms, there can be obstacles that prevent these mechanisms from operating, leading to non-transitory losses of economic value. 74 For instance, abnormally high profits in a market should in principle not be sustainable in the long run because they would attract new entry into this market. As the new competitors start offering the same or very similar products as the incumbent(s), they will steal market share, and hence profits, from them, until the abnormally high profits will gradually be competed away. However, this self-correcting mechanism may be impaired when there are, for instance, barriers to entry that make it very difficult or even impossible for potential competitors to enter the market and challenge the incumbents. Such barriers to entry are particularly salient in digital markets, because they do not allow entrants to be cost effective (because of scale and scope economies), to replicate the incumbent’s products or services (because of data dependency or vertical integration), or to induce consumers to switch away from the incumbent(s) (because of network effects, switching costs, or asymmetric information). Such barriers to entry therefore allow incumbents to sustain market power, which in turn leads to longer-term societal losses in terms of higher prices and less product variety for consumers, and less dynamic innovation. 75  

70.It is important to stress that the features of a market include both structural and behavioural ones and that demand-side considerations, in particular the behaviour of customers, play an equally important role in this regard. Therefore, in many cases, there is a combination of those elements leading to or constituting a market failure. 76

71.This section analyses a list of gatekeeper related market failures, notably: (i) entry barriers to gatekeeper markets (Section 2.3.1.1) and (ii) economic dependence and imbalanced bargaining power (Section 2.3.1.2).

Entry barriers to gatekeeper markets

72.Market players in the digital economy face important barriers to entry. This is due to the fact that digital market features can be exploited by gatekeeper platforms to strengthen their market position and prevent market entry.

73.There has been broad consensus among the NCAs 77 , as well as among the respondents to the OPC 78 that extreme economies of scale and scope, high start-up costs, high fixed operating costs, high degree of vertical integration, single-homing, switching costs, multi-sidedness, network effects, zero-pricing markets, information asymmetry, data dependency access to data, and behavioural bias are important or very important sources for market failures in digital markets. Moreover, according to an International Competition Network (‘ICN’) report, an important proportion of respondents indicated that most of these factors were playing an important role in digital markets’ power assessment in the competition enforcement cases that they have investigated. 79  

74.Regarding economies of scale, the Commission in Google Shopping, based its dominance assessment for the market for general search services among other things on the existence of barriers to expansion and entry, notably the significant investments in terms of time and resources required to establish a fully-fledged general search engine. 80 Likewise, in its Google Android decision, the Commission found that “developing a smart mobile OS [operating system] is a costly and time-consuming process”. 81 As regards economies of scope, the Commission made particular reference in its Preliminary Assessment in the Amazon e-book MFNs case to “[t]he ability of e-book readers to drive sales and lock-in customers: that with its Kindle e-book reader, Amazon operates a closed "ecosystem" (or "walled garden"). Customers who own a Kindle can use that e-book reader only for ebooks purchased in Amazon's Kindle store”. 82 In the same case, the Commission also found substantial economies of scale for e-book retailing, in particular because of the need to construct a sufficiently large catalogue of available titles (which requires agreements with a large number of E-book Suppliers), and because of the scale and scope of investments needed to set up a viable e-book distribution platform. The Special Advisers Report refers to “the presence of strong economies of scope favouring the development of ecosystems and giving incumbents a strong competitive advantage. Indeed, experience shows that large incumbent digital players are very difficult to dislodge”. 83  

75.Due to the two-sided nature of platform markets, once a gatekeeper managed to bring both sides of the market on board, it becomes very difficult for a new, emerging platform, to establish itself in the market, as it has to convince both users and developers simultaneously that it is a viable alternative to the already established platform. For instance, in establishing Google’s market power in the Google Android case, the Commission quoted Orange as saying that “[g]iven the two-sided character of this market (attracting enough developers requires having a large user base and users will reciprocally be attracted to shops offering many apps) it is indeed very difficult to offer an app shop in competition with Google Play given (i) its link with Android OS and (ii) its current size”. 84  

76.The problem of challenging a gatekeeper is often exacerbated in situations where at least one side of the market (typically the final users) is served at zero prices by the incumbent platform - with firms monetising their services through advertising and/or access to consumer data 85 - so that there is no room for the entrant platform to attract final users through aggressive pricing policies. 86  

77.The zero pricing strategies described above also explain why network effects tend to favour large incumbents preventing smaller rivals from effectively challenging incumbents and stealing market shares from them. The Commission has found in the Microsoft case, for instance, that network effects represented a relevant barrier to entry because “[a] media player would not meet with significant consumer demand if there was no or no significant amount of corresponding digital content which this player could play back”. 87 In Google Android, the Commission found that “network effects arise because, when deciding which licensable smart mobile OS to develop for, app developers consider the revenue potential of that OS and since they ‘earn their profits mainly by app downloads, mobile OSs with a large user base are considered more attractive by app developers”. 88  

78.Indirect network effects are particularly strong for large-scaled platforms also due to their unlimited capacity to expand data sets, i.e. data-driven network effects. In addition to this network amplification function, data is a major asset in the digital economy. It is particularly important for a business to have access to data related to its consumers and stemming from its activity on a platform since such data allow the business to adapt its market strategy. Business users’ dependency on data could be used to prevent them from competing effectively on the platform. This is particularly problematic when the business user is in direct competition with a gatekeeper who can use data generated by the business user’s activity to its own interest. Data can thus be used by gatekeepers as a barrier to entry, expansion and competition and is therefore an essential element for enabling market contestability. In the Google Shopping case, the Commission identified the availability of data in the form of user search queries, paired with users’ tendency to single-home on Google for their general searches, as an important barrier to entry 89 : “[B]ecause a general search service uses search data to refine the relevance of its general search results pages, it needs to receive a certain volume of queries in order to compete viably. The greater the number of queries a general search service receives, the quicker it is able to detect a change in user behaviour patterns and update and improve its relevance”. 90  

79.The presence of network effects and the multi-sidedness of certain markets imply that even markets where initially multiple competitors are active are particularly prone to tipping: once a firm has obtained a certain advantage over rivals in terms of market share, its position may become unassailable and the market may gravitate towards a situation of dominance or (quasi)-monopoly. This advantage can be due to its presence in other related services, access to data or simply because it is the first mover into the market. In these cases, markets may not yet have generated a gatekeeper, but show clear signs of increasing market power in the hands of one firm. Respondents to the OPC generally considered that important or very important market features of a tipping market are the following: (i) direct network effects; (ii) indirect network effects; (iii) users predominantly single-home and (iv) economies of scale. Respondents generally considered that tipping is common or to some extent common in digital markets. 91 When asked about the need for the Commission to be able to intervene early in cases of emerging gatekeepers to preserve/improve competition, the large majority of respondents agreed, including the majority of businesses and business associations, civil society organisations (i.e. consumer associations, NGOs and citizens) and public authorities (including NCAs). 92

80.Behavioural bias is another important feature of digital markets. This feature merits further attention in this section since it contributes to increasing switching costs and keeping users locked into the gatekeeper platform, i.e. leading to user lock-in, thus strengthening entry barriers. Platform companies routinely design services to optimise their users’ experience, often using advanced behavioural profiling and testing techniques, such as A/B testing 93 , or finely targeted personalisation of their service offerings. Gatekeepers use various techniques 94 (e.g. design of choices, misdirection, social pressure, sneaking items into the user’s shopping basket, and inciting a sense of urgency or scarcity) that ‘nudge’ users into certain decisions. A recent search on 11 000 shopping websites identified 1 818 patterns of practices used to incite users doing things they have not intended to do. 95

81.From the perspective of platform competition, research on the basis of ‘agent-based simulations’ also found evidence of biases that reinforce consumer lock in, such as ‘escalation of commitment’, and ‘availability bias’. 96 In ‘escalation of commitment’, users commit themselves to one platform, even when switching may provide higher user utility. Hence, those users never switch platforms. For instance, a consumer purchasing on a large e-commerce marketplace offering a range of products, would not switch to one or several other platforms even if the latter are specialised in the specific type of goods the consumer is interested in. Convenience and user habits would prevail over the benefit (e.g. higher quality) potentially resulting from the use of a more specialised platform. Users subject to an ‘availability bias’ may make platform choice decisions using a heuristic that relies on vivid or recent data. For example, users may easily recall a platform that has many users, as social media would be mentioning such a platform. Social norming (e.g. follow friends’ behaviour) may play an additional role for user lock-in and increase switching costs. Behavioural bias discourages switching to different alternatives (such as a different browser, different search engine, etc.) whenever certain software products come pre-installed on consumers’ devices, and therefore has similar adverse effects on competition as would limited information about the existence of these alternatives.

82.For instance, in the Google Android case, the Commission found that “users that find apps pre-installed and presented to them on their smart mobile devices are likely to ‘stick’ to those apps”. 97 In other words, users suffer from ‘default bias’ or ‘status quo bias’, which in turn makes pre-installation of operating systems, app stores, search engines, etc., a powerful tool to lock in users to these specific services: “In 2016, approximately 260 million smartphones were sold in Europe, of which approximately 197 million smartphones or 76% were Google Android devices. Practically all of these Google Android smartphones had the Google Search app pre-installed with the rest of the GMS bundle”. 98

83.The Commission’s enforcement practice under Article 102 TFEU has shown that that the presence of high switching costs makes it more difficult for entrants to contest the market position of firms that have already acquired a large customer base. For instance, in an internal document, Microsoft itself stated that “The Windows API […] is so deeply embedded in the source code of many Windows apps that there is a huge switching cost to using a different operating system instead”. 99 Switching costs are also relevant where customers are businesses, not final consumers. This is demonstrated by the Google Android case, where the Commission found that “OEMs wishing to switch to other licensable smart mobile OSs face switching costs. […] For example, Sony has estimated that the initial development cost ‘to implement the Android OS on our devices was approximately 50 million Euro, with lead time of 1.5-2 years’”. 100 One implication of high switching costs in the platform context is that either one (or both sides) of the platform tend to single-home for specific purposes, i.e. users only use one platform, rather than using several platforms simultaneously. 101 For instance, the vast majority of smartphone users owns either an iPhone or an Android phone, but not both at the same time, and they tend to be very loyal to their operating system.

84.The above entry barriers are gradually reinforcing each other due to the ‘winner–take-all’ dynamics in digital markets. The bigger the platform, the stronger the indirect network effects, the larger the amount of data and the higher its quality. This leads to increased insight into user profiles and preferences, allowing gatekeepers to offer them more personalised services and advertisements, thus attracting even more users and reinforcing consumer lock-in, favouring single-homing and rendering switching to alternative platforms more difficult.

Economic dependence and imbalanced bargaining power

85.Dependence and imbalanced bargaining power characterise business relations with all platforms 102 including small ones. What distinguishes however relations with gatekeepers, is the particularly strong level of dependency and the important scale of power imbalance, which together with unfair conduct engaged in by these gatekeepers can have serious harmful effects on the business users and customers.

86.First, gatekeepers have become a strategic business partner; an enterprise not present on these platforms would not reach a very significant number of consumers. Figures in some sectors illustrate well the strong degree of dependence. In 2024, consumers are projected to download 181 billion apps from biggest app stores. 103  Over 80% of social referrals to e-commerce sites come through the most used social platforms some of which having more than 2 billion monthly active users and 7 million active advertisers. 104 This explains the millions business pages and companies using these social media every month 105 to target this large audience.

87.Gatekeepers’ role as key trading partners is constantly strengthened due to e-commerce trends. The Business-to-consumers (‘B2C’) e-commerce turnover was growing at an average pace of 13% between 2014 and 2019 with turnover forecasted to hit EUR 621 billion in 2019. 106 On average, 16.2% of retail trade in 2020 in Europe takes place on line, almost double in comparison to 2018. The share of online shoppers in Europe making cross-border online purchase has also increased significantly over the past decade, nearing 50% in 2019. Cross-border B2C e-commerce sales in Europe are projected to grow at a double-digit rate at least through 2022. 107

88.Second, the incomparable economic strength of gatekeeper platforms show the extent to which their commercial relations with business users are imbalanced. In 2019, the biggest app stores generated over USD 83 in revenue. 108 In 2019, one of the most important social networks had over USD 70 billion in revenue. 109

Fragmented regulation and oversight

89.Various national rules in the EU are emerging in partial response to the problems identified. In addition, insufficient coordination among different national authorities setting rules vis-à-vis platforms may lead to potentially heterogeneous responses across the EU. Fragmentation already exists with regard to platform-specific regulation, as for example in the cases of transparency obligations and MFN clauses.

90.Furthermore, fragmentation results also from differing legislation relating to dependency situations in place in various Member States (Belgium, Bulgaria, France, Germany, Hungary and Italy) while in the remaining Member States there is no legislation addressing dependency in place. Regarding MFN clauses there are two different types of fragmentation: fragmentation due to the fact that some Member States imposed legislative bans and some Member States did not. Furthermore, fragmentation is observable due to differences in the MFN-legislations in those Member States where they are in place. For instance, in some Member States (Austria, Belgium, France, Italy) all types of price parity clauses are prohibited, i.e. also narrow MFN clauses, while in the remaining Member States MFN clauses may be challenged under EU competition law only.

Conclusion: problem drivers’ effects

91.The above-described problem drivers can lead to a number of issues. The economic dependence of business users on gatekeeper platforms and the imbalance of bargaining power - between these two types of players as well as between gatekeepers and smaller platforms - can result in important economic harm. Business users need to be present on gatekeeper platforms in order to reach consumers as the above analysis show, which allows gatekeepers to set the rules of access and use of their platforms in an unfair way. Gatekeepers can undermine the trading conditions for dependent business users by behaving unfairly, thus limiting (national and cross-border) sales and trust in the platform economy.

92.Digital markets’ features (see in particular discussion on market failures in Section 2.3.1) have mutually reinforcing effects which in the winner takes it all dynamics of these markets constitute unsurmountable entry barriers. The latter drive a number of issues related to weak market contestability and gatekeepers’ sustained market position leading to longer term societal losses in terms of products’ and services’ prices, consumer choice and suboptimal innovation opportunities (as illustrated in Section 2.2 and further described in Section 6). Gatekeepers control the conditions for innovation and entry by independent firms. An important effect of the exercise of control by gatekeepers is that they can inhibit innovation by potential alternative platforms or by applications providers operating on their platform. Potential competitors, which might offer an alternative route to customers, may find it challenging to gain a foothold in markets with gatekeepers.

93.The extensive nature of problems associated with gatekeepers has led a number of Member States to take or consider their own measures to address gatekeeper power. However, isolated and uncoordinated national approaches to addressing a problem which concerns cross-border platforms, and is hence pan-European in scope, risk creating different national rules which increase the compliance costs for platforms (and especially entrants or small scale platforms) operating cross border, and for business users, including SMEs, providing services across the EU. 110

How will the problem evolve?

94.Concentration and mark-ups in most digital markets have been increasing over the last years, and there is no indication that this trend will be inverted during the next years. In some cases, markets have already stabilised at a high concentration level and do not show any evidence of possible increase in competitiveness in the future. Data is also becoming more and more important, exacerbating the market failures associated with the control of data. 111

95.The COVID-19 crisis has dramatically increased the importance of e-commerce and trading via digital platforms in the EU’s economy. 112 This has only accelerated the dependency of users and businesses on the services provided by the few gatekeepers – as evidenced indirectly by the increase in stock market valuation of some of the largest platform companies.

96.The following graph illustrates the stock price development for five major big tech companies from 2014-2018. 113 To the extent that stock prices reflect market expectations of future profitability, this graph can be interpreted as measuring (future) profits of the respective companies. When comparing these figures to the S&P 500 index 114 , which grew by about 60-70% over the same period, this graph shows how the five digital operators – Alphabet (Google’s parent company), Apple, Facebook, Amazon, and Netflix – have consistently outperformed the market average.

Figure 1 - Stock price development for 5 big tech companies, 2014-2018

97.Absent any EU intervention, the economic drivers are likely to increase, exacerbating the observed problems. As an illustration, further development and use of voice assistants can also be expected to reinforce gatekeeper platforms’ position. Voice-activated services may create concerns in relation to search for online products/services/information. The provision of a single answer to a search request limits the possibility to access alternative results, thus reducing choice and limiting competition. 115  

98.Innovation would remain concentrated within a small number of gatekeepers, ultimately limiting consumers’ possibility to access innovation and data-friendly services provided by a larger number of platforms than gatekeepers.

99.Regarding fragmentation likely to occur in the near future, legislation to address imbalances in the relationships between digital platforms with economic power and their business users is currently in process to be adopted in several Member States (e.g. Germany, France and Romania). Some Member States (Belgium, Luxemburg and the Netherlands) are currently mainly supporting action at EU level. However, they would most likely take legislative action at national level in the absence of action at EU level. Those legislative projects already in the process of adoption and also those likely to be tabled in the absence of action taken at EU level demonstrate the likelihood of further fragmentation. 116

3.WHY SHOULD THE EU ACT?

Legal basis

100.Given the intrinsic cross-border nature of the services provided by gatekeeper platforms and the risk of further regulatory fragmentation regarding functioning of the Single Market for digital services, in particular in relation to gatekeeper platforms as well as functioning of digital markets, Article 114 TFEU is the relevant legal basis for this initiative.

101.As set out above, the current regulatory approaches at Member States level are a patchwork of existing or proposed regulatory solutions (see detailed description in Annex 5.4 to the Impact Assessment). This creates legal uncertainty for companies operating in the internal market, whether at national or on a pan-European basis and risks creating an appreciable distortion of competition in the internal market and undermine fundamental freedoms protected by the Treaty.

Subsidiarity: necessity and added value of EU action

102.The objectives of the intervention cannot be achieved by Member States acting alone, as the problems are of a cross-border nature, and not limited to single Member States or to a subset of Member States. The digital markets at stake (including those featuring gatekeeper platforms) are often of a cross-border nature, as is evidenced by the volume of cross-border trade, and the still untapped potential for future growth, as illustrated by the pattern and volume of cross-border trade intermediated by digital platforms. Almost 24% of total online trade in Europe is cross-border. It is estimated that by 2025 online marketplaces will represent 65% of cross-border online sales in Europe. 117

103.Even where these digital markets may be geographically defined as national in scope, the problems at stake nevertheless remain of a cross-EU nature for three main reasons. First, the goods and services offered by the market players concerned are typically of a cross-border nature. Second, digital players typically operate across several Member States, if not on an EU-wide basis, which is particularly the case for markets such as online advertising, social media, online retail, cloud services, e-commerce or online search. This is not to say that services such as online advertising and search do not have to be tailored to Member States’ languages - however, the overall business strategy will normally be EU-wide.

104.Accordingly, market failures in digital markets have Union relevance, as they can arise across borders and affect several Member States, thus not being limited to a specific national market of a Member State. 118  

105.As regards the particular case of unfair business practices, in the absence of an EU measure, there is a high risk that with national approaches, business users or application developers seeking to serve the internal market will need to understand a range of diverse rule-sets and pursue actions in multiple countries across the EU, which is likely to fragment the Single Market for digital services, create barriers to expansion and compliance costs, especially for start-ups and SMEs. A lack of harmonised rules in this space risks complicating the regulatory landscape faced by platforms operating on a pan-European or indeed global basis. An intervention at the EU level is therefore more efficient, insofar as it introduces a common set of rules across Member States to address in a consistent manner the same unfair business practices carried out by large digital gatekeepers across the Union.

106.Similarly, intervention by individual Member States or NCAs would be ineffective in tackling gatekeeper related market failures across the Union. Each Member State can only address market failures in its own territory, 119 imposing its own remedies, whereas market failures may affect the territory of several Member States because of the wider geographic scope of the relevant market concerned or the cross-border business activities of the market players concerned. Addressing market failures with a cross-border dimension at national level could also lead to inconsistencies in the remedies imposed, with the ensuing risk of fragmenting the Digital Single Market.

107.Therefore, by addressing market failures in respect of key digital markets, the functioning of the internal market will be improved through clear behavioural rules that give all stakeholders legal clarity and through an EU-wide intervention framework allowing to address market failures in a timely and effective manner.

4.OBJECTIVES: WHAT SHOULD BE ACHIEVED?

General objective

108.The general objective of this initiative is to ensure the proper functioning of the internal market by promoting effective competition in digital markets, in particular a fair and contestable online platform environment. This objective feeds into the strategic course set out in the Communication ‘Shaping Europe’s digital future’ as shown in Section 1.

Specific objectives

Address market failures to ensure contestable and competitive digital markets for increased innovation and consumer choice

109.As explained in Section 2.3.1.1, certain digital markets may not be functioning well and delivering competitive outcomes due to their particular features, in particular extreme scale (or scope) economies, and a high degree of vertical integration; direct or indirect network effects; multi-sidedness; data dependency; switching costs; asymmetric and limited information, and related biases in consumer behaviour as well as the conduct of gatekeepers. Therefore, a specific policy objective is to allow identifying and addressing such market failures in respect of key digital markets to ensure that these markets remain contestable and competitive. This will contribute to digital markets delivering low prices, better quality, as well as more choice and innovation to the benefit of EU consumers.

Address gatekeepers’ unfair conduct

110.As explained in Section 2.3.1.2, gatekeepers’ economic strength, their position of intermediaries between businesses and consumers together with markets dynamics fueling gatekeepers’ growth lead to an imbalance in power between gatekeepers and their business users. This enables gatekeepers to impose unfair commercial conditions on business users, thus hampering competition on the platform. Such unfair behaviour does also have a negative impact on (the emergence of) alternative platforms since it strengthens consumer lock-in thus preventing multi-homing. In light of this, a specific policy objective is to lay out a clearly-defined set of rules addressing identified gatekeepers’ unfair behaviour, thereby facilitating more balanced commercial relationship between gatekeepers and their business users, which would be also expected to create the right incentives for multi-homing.

Enhance coherence and legal certainty to preserve the internal market

111.The gatekeeper-related problems identified above are currently not (or not effectively) addressed by Member States in existing regulation. The national legislative initiatives may partially address problems identified but also lead to increased regulatory fragmentation. In addition, tackling issues identified through legislation at national level is suboptimal in light of the cross-border nature of the platform economy. Moreover, the systemic importance of gatekeepers for the internal market deserve a better coordinated and more effective intervention across the EU. As explained in Section 2.3.2, national laws lead to legal fragmentation of the platform space and increase compliance costs for all market players. Therefore a specific policy objective is to improve coherence and effectiveness of oversight and enforcement of measures vis-à-vis gatekeepers, thus contributing to increased legal certainty.

How do the objectives link to the problems identified?

112.The figure below shows how different objectives are linked with the problems and the underlying problem drivers. It also shows that the specific objectives, i.e. address market failures to ensure contestable and competitive digital markets for increased innovation and consumer choice, address gatekeepers’ unfair conduct and enhance coherence and legal certainty in the online platform environment for a preserved internal market, contribute to achieving the general objective of ensuring the proper functioning of the internal market (through effective competition in digital markets and through fair and contestable online platform environment).

Figure 2: Intervention logic tree – problem drivers, problems and objectives

5.WHAT ARE THE AVAILABLE POLICY OPTIONS? 

113.This section presents the three policy options retained (Section 5.3), the baseline scenario from which they will be assessed (Section 5.1), and the options that have been considered but discarded (Section 5.4).

What is the baseline from which policy options are assessed?

114.Gatekeepers are currently subject to two main sets of laws: first, competition laws which are applicable across all sectors of the economy, and, second, EU rules of a more sector-specific scope or with a less punitive nature including, among others, the P2B regulation, the General Data protection regulation (‘GDPR’) and the EU consumer law.

115.Broadly speaking, the EU competition policy toolbox includes rules on antitrust, merger control, State aid, and public undertakings and services. Generally, a distinction is made between competition rules allowing for an intervention ex post or ex ante. Antitrust enforcement – under Articles 101 and 102 TFEU and the accompanying implementing regulations – belongs to the first category, as it aims at detecting anti-competitive behaviour by companies that has the actual or likely effect of causing distortions of competition. Merger control and state aid rules aim at preventing anti-competitive outcomes by assessing ex ante whether a merger between undertakings or the granting of State aid would negatively affect competition. Intervention under the existing EU competition rules can therefore only occur if: (i) a company is dominant pursuant to Article 102 TFEU and abuses this position, (ii) there is an anticompetitive agreement or concerted practice between two or more undertakings covered by Article 101 TFEU, (iii) there is a merger/acquisition with EU dimension falling under the EU merger control rules 120 or (iv) a Member State grants aid falling under the EU State aid rules.

116.Under the baseline scenario, the Commission would not propose any changes to the current competition legal framework. This means that the Commission would continue to vigorously apply and enforce the existing competition law framework, in particular Articles 101 and 102 TFEU, against gatekeepers in digital markets, should the conditions for such intervention be met. Competition enforcement by the Commission would include making full use of the existing tools within this framework. The ongoing reviews of existing legislation (e.g. the Block Exemption Regulations for horizontal and vertical agreements) as well as of the Market Definition Notice would continue as planned.

117.The majority of the respondents to the OPC indicated in their replies that, while some of the issues connected to gatekeeper powers could potentially be addressed by competition law enforcement through procedural and/or organisational changes, there were restrictions that could not be overcome with competition law enforcement. Respondents argued that the main challenges with regard to the enforcement of Article 102 TFEU relate to situations where dominance does not exist, and the difficulties with remedying a conduct found to be anti-competitive in an appropriate and effective manner, notably once the damage has already occurred. They considered that these challenges also have a negative effect on the duration of antitrust investigations and the ability of the existing competition law framework to ensure the contestability of the markets concerned. Respondents also highlighted the need for a regulatory solution regarding conduct recurrently showing negative effects on competition, as well as the need to pursue more exploitative cases and to take non-economic objectives into account in the competitive assessment. 121 A minority of respondents to the OPCs argued that Articles 101 and 102 TFEU are suitable and sufficiently effective in addressing market failures, and were primarily satisfied with the level of enforcement of the existing competition rules. These respondents pointed towards interim measures, sector inquiries, merger control and deadlines as potential ways to tackle any shortcomings that competition law may have. Some respondents also pointed to the ongoing reviews of competition legislation such as the Market Definition Notice, as well as the rules applicable to vertical and horizontal agreements as other ways to improve the existing competition law framework. Some respondents also argued for a broader use of sector inquiries and a review of the EU merger regulation. 122  

118.The Commission considers that the current legal framework would not allow it to address the market failures described in Section 2.3.1 for the following reasons.

119.First, existing EU competition rules cannot conceptually deal with market failures resulting from the behaviour of gatekeepers in the absence of some preconditions, such as the existence of an anticompetitive agreement in the case of Article 101 TFEU or of a dominant position in the case of Article 102 TFEU. In addition, in some instances, existing EU competition rules may be able to prevent or address a market failure, but not in the most effective manner. The Commission’s enforcement of Articles 101 and 102 TFEU rules can only take place ex post, i.e. after a competition problem has emerged. As a recent report by the European Court of Auditors (‘ECA’) also indicates “particularly in the digital economy, this may be too late to tackle a competition problem”. The ECA report also flags that “the Commission has currently no tools in its hands that would allow it to intervene ex ante i.e. before competition problems would occur”. 123 Moreover, – even when using interim measures, explained below – competition law enforcement requires a detailed economic and legal analysis which, jointly with the procedural safeguards, bring the duration of the investigations to at least around two years and usually more than that. In markets characterised by powerful network effects and economies of scope, competition law interventions may mean not only delays in the interventions but also that irreparable effects such as tipping may no longer be reversible.

120.Second, market failures associated to tipping markets cannot be tackled on the basis of the existing competition rules, notably where market tipping is triggered primarily by the market structure, and not (or only to a lesser extent) by any specific conduct.

121.Third, the existing EU competition rules do not necessarily capture all unfair business practices by large digital gatekeepers. This is because these practices do not necessarily have an anticompetitive object or effect under Article 101 TFEU, or may not be captured by Article 102 TFEU, if there is no effect on competition on clearly identifiable relevant markets. 124

122.Fourth, specific competition tools cannot address the gatekeeper related market failures:

·Prohibition (fines) and commitments decisions (Articles 7 and 9 of Regulation 1/2003 125 ) are decisions addressed to individual companies for a breach of the EU competition rules, and not suitable for addressing market failures that are not, or not exclusively, caused by such breach of the EU competition rules.

·Sector inquiries (Article 17 of Regulation 1/2003) are investigations that the Commission carries out when it suspects possible breaches of the competition rules in specific sectors of the economy. There is, however, no possibility to impose remedies following a sector inquiry.

·Interim measures (Article 8 of Regulation 1/2003) are a tool allowing the Commission to intervene in “cases of urgency due to the risk of serious and irreparable damage to competition” where a ‘prima facie’ infringement of the EU competition rules can be shown. Interim measures, however, would not allow the Commission to tackle the problems explained in this Impact Assessment for two main reasons: first, interim measures can only be imposed where a prima facie infringement of Articles 101 or 102 TFEU can be shown, and second, interim measures are founded on a very specific test requiring the finding of ‘urgency’ as well as ‘serious and irreparable damage’. Interim measures have only been used twice in the last nineteen years.

123.Fifth, the ongoing reviews of existing legislation (e.g. Block Exemption Regulations for horizontal and vertical agreements) as well as of the Market Definition Notice will also not tackle or address the problem drivers:

·The ongoing reviews of Block Exemption Regulations cannot tackle or address the problem drivers. The Block Exemption Regulations pursue a different objective than the DMA. They are by their very nature not aimed at addressing specific competition issues and/or market failures, but at block exempting agreements that are on balance efficiency enhancing, thus helping companies to self-assess compliance of their agreements with Article 101 TFEU.

·The ongoing evaluation of the Market Definition Notice cannot, by its very nature, tackle or address the problem drivers as it is a soft law document to provide guidance on the definition on the relevant market and on the relevant parameters to be taken into account for it when using Articles 101 and 102 TFEU or the EU Merger Regulation (Regulation (EC) 139/2004). The Market Definition Notice cannot therefore address competition issues and/or market failures in digital markets.

124.Some respondents to the OPC – across different stakeholder categories – considered that there was no need for the DMA and that the Commission should rather reassess the situation after the P2B Regulation had shown its effects. A minority of respondents, mainly several large platforms and their trade associations, and some research institutes and academics, disagreed with the need for the proposal of new ex ante rules as they consider that the risks posed by gatekeepers can be addressed with existing regulation. Some platforms, trade associations and national authorities emphasised the need to focus the regulatory attention towards specific actions and perceived market failures. 126

125.Under the baseline scenario, the Commission would continue to apply and enforce the existing more sector-specific EU rules including, among others, the P2B regulation, the GDPR and EU consumer law. The Commission considers however that the current regulations will also not tackle or address the problem drivers described in Section 2.3 for the following reasons:

·Regulation (EU) 2019/1150 on fairness and transparency for business users of online intermediation services (the ‘P2B Regulation’) entered into force on 12 July 2020. It is the first EU-level legislation specifically targeted at commercial issues engaged in by online platforms, or online intermediation services, as well as by online search engines. It applies to more than 10 000 platforms in Europe and reflects the fact that a certain dependency of professionals, or business users, is inherent in any successful online platform. This feature of online platforms means that the fairness, transparency and redress rights and obligations that the P2B Regulation provides are necessarily high-level and principles-based. Since this legal framework establishes a general ‘safety net’ for all professionals active in the online platform economy, it does not address issues deriving from the concentration of economic power and unfair business practices of a limited number of very large gatekeeper platforms.

·EU data protection legislation 127 specifies the fundamental right to the protection of personal data. It therefore covers business-to-citizen and government-to-citizen interactions, rather than commercial and competition-related issues. Article 20 of the GDPR provides a limited right to data portability 128 , though it is broadly considered that there are still many implementation challenges and that this right is at present insufficient to significantly lower entry barriers and to facilitate the contestability of markets. 129  

·EU consumer law does address a range of potentially harmful practices, at EU level notably through the Unfair Commercial Practices Directive (‘UCPD’) 130 and the Unfair Contract Terms Directive (UCTD). 131 While these directives define a number of relevant concepts, such as 'professional diligence' and 'good faith', their scope is explicitly limited to business-to-consumer transactions. Conversely, the Misleading and Comparative Advertising Directive (MCAD) 132 covers certain Business to Business (‘B2B’) relations. However, the provisions set forth in the MCAD are limited to a narrow subset of advertising practices, which are not specific to online platforms or digital markets, and do not deal with the unfair business practices carried out by large gatekeeper platforms.

126.Finally, in the absence of further EU legislation, and subject to enforcement of the existing legal framework, the legal fragmentation is likely to further increase as Member States are likely to continue to adopt horizontal or sector specific national measures against gatekeepers, as shown in Annex 5.4 to the Impact Assessment.

What are the main parameters that determine the range of available policy options?

127.The problems and drivers map on to a set of parameters that characterise the range of available policy options. These parameters include (a) the scope of the intervention; (b) the range of unfair practices at stake; (c) the speed of the instrument and the degree of flexibility offered; and (d) the investigative and enforcement framework available and appropriate. Before presenting the options, this section explains these main parameters and highlights some of the inherent trade-offs.

Scope: core platform services, gatekeepers, thresholds

128.As highlighted in Section 2.1, the scope of the intervention is characterised by two particular concepts – the nature of the ‘core platform services’ where problems arise, and the notion of ‘gatekeepers’, i.e. companies that offer one or more core platform services. Core Platform Services and Gatekeepers can be identified robustly by a series of criteria, as set out below.

129.The table below summarises the main features and practices in relation to the core platform services retained. The details about the evidence supporting those elements are provided in Sections 2.3.1 and 5.2.2. There is no consistent publicly available data about the mark-ups in each of these core platform services. In addition, some of these services do not generate direct revenue as prices are set at zero (e.g. online search services, social network services, number-independent messaging services), but are monetised via advertising services on the other side of the platform. Nevertheless, the players active on those services are some of the most profitable companies in the world (see Section 6.6.1).

130.There is also not much research developed about the impact of concentration in the innovation efforts in relation to these services. According to the Stigler report “Disruptive innovation in markets that are characterized by high concentration levels and network effects is likely to be reduced compared to a competitive market. […] Entrepreneurs may expect a low payoff to developing a free-standing product because of entry barriers and exclusionary conduct by the incumbent platform.” 138 This report further concludes that “The incipient but growing technical research supports a concern for the impact of big tech on innovation” based on some studies on the innovation in social platforms, internet software (e.g. operating systems and other related apps) and internet retail.

Table Error! No sequence specified.1: List of core platform services

131.As mentioned above, the identification of core platform services is relevant for both the designation procedure for gatekeepers, and the identification of those services provided by the designated gatekeeper. The evidence gathered during the OPCs indicated two guiding principles.

132.First, there is a wide agreement that gatekeepers exist and that their role and practices are increasingly determining the conditions of market participation on the consumer side as well as conditions of individual commercial relationships they engage in on the business user side. In circumstances where such practices are unfair, this may undermine balanced commercial relationships and the contestability of platform markets.

133.Second, while there is no consensus as to a single means to identify gatekeepers in the digital sector, to the extent that different sources refer to qualitative criteria as a means of identifying gatekeepers there is relatively broad consensus that such criteria need to take into account factors such as access to data, network effects and ability of these providers to leverage their economic power to multiple services. Similar qualitative criteria (e.g. access to data, financial resources, level of vertical integration) are also promoted by a recent initiative on the update of the German competition law. The OPC provided mixed views on this, with a majority of stakeholders calling for a combination of quantitative and qualitative criteria to identify gatekeepers and only very small number of respondents promoting use of qualitative criteria only. 180  

134.This designation of gatekeepers in this way leads to another important design parameter, namely on the choice of thresholds. Stakeholders generally support a mix of quantitative and qualitative parameters (see paragraph 133).

135.Suitable quantitative thresholds can be constructed from indicators for size (such as turnover and presence in various Member States) and for economic dependency (such as the number of business users and end users served on the platform). The weak inter-platform competition that results from such gatekeepers’ services having become entrenched would be captured by measures of persistence (such as the number of core platform services offered by the same group of undertakings, and the number of years this group has held its position). Depending on the precise indicators used and the level at which these are set, the addressable population can obviously be larger or smaller. Importantly, the combination of these quantitative parameters would in all cases only include a very limited number of companies and would exclude a very large set of providers of digital services. It would therefore not cause any undue lack of legal certainty, and the option of using higher or lower thresholds is a legitimate political choice, presenting some of the trade-offs described in this section.

136.Subject to the overarching criteria of size, dependency and persistence, a wide range of different indicators have also been investigated during the Impact Assessment. 182

137.A reliable set of specific indicators has been used to identify the effects of fixing the thresholds at higher or lower levels in terms of the number and identity of the undertakings that could potentially be brought within the scope of the different options. This exercise confirmed the accuracy of the methodology.

138.Alternative methods for designating gatekeepers were analysed but discarded, notably those that are directly inspired by notions of ‘significant market power’ as it is used for example in the regulatory framework for telecommunications services. While such models can serve as inspiration, and do capture notions of concentration, they are difficult to transpose directly as the diversity of platform business models is poorly captured by such a market-based analysis, and the relevant bottleneck power at stake in this Impact Assessment also derives from the tendency of gatekeeper companies to diversity.

139.The below paragraph specifies the quantitative parameters used to identify gatekeeper companies and their core platform services. 

140.For this exercise, the following were selected as proxies for the main three criteria:

141.For Size & internal market impact, the proxy selected is the EEA annual turnover of the group ( > EUR X billion) or the average market capitalisation 183 or the equivalent fair market value of the group ( > EUR X billion) in combination with its presence in more than three EU countries.

142.A significant turnover in the Union or the market capitalisation and the provision of a core platform service in at least three Member States is a good indicator that the provider of that service has a significant impact on the internal market. In other words, a provider of a core platform service should be deemed to have a significant impact on the internal market where it provides a core platform service in at least three Member States and where either its group turnover realised in the Union is equal to or exceeds a specific, high threshold or the market capitalisation of the group is equal to or exceeds a certain high absolute value. Both parameters are reflective of gatekeepers’ ability to monetise their users and of their financial capacity, including the ability to leverage access to financial markets to reinforce their already strong position.

143.For Dependency: the proxy selected concerns the number of users ( > X million EU users) in combination with the number of business users ( > X EU business users) across all core platforms services, as this is a proxy for the bottleneck power that characterises the problem.

144.A very high number of business users that depend on a core platform service to reach end users and a very high number of monthly active end users are indicative of the provider’s role as an important gateway. 184  

145.For Persistence, the proxy used relates to the number of large core platform services showing dependency, as well as the number of recent years during which the size and dependency criteria have been met.

146.Likelihood of an entrenched and durable position or the foreseeability of achieving such a position in future is very high where the contestability of the gatekeeper’s position is limited. This is likely to be the case where that provider has provided a core platform service in at least three Member States to a very high number of business users and end users during at least three years.

147.Other proxies, such as criteria estimating the degree of multi-homing or the rate of innovative entry have been suggested in the literature, but were discarded for the purpose of finding objective quantitative criteria that also have a relationship with the impact of the gatekeepers in the EU.

148.The above quantitative parameters could be combined in different ways with a view to providing consistent thresholds. 185 Two meaningful thresholds would be the following:

·Low threshold: Fixing the EEA annual group turnover threshold at the level of EUR 6.5-7.5 billion and the required number of core platform services showing dependencies at 30-45 million end users and 10 000 business users during a number of years at a single one. This threshold would result in 10 to 15 providers of core platform services.

·High threshold: Fixing the threshold at an EEA annual group turnover of EUR 5-6 billion but also including a minimum of two core platform services, with at least one showing dependencies at 30-45 million end users and 10 000 business users would reduce the group of providers captured to an estimated number of five to seven companies.

149.Other possible combinations of parameters (e.g. market capitalisation, required turnover, number of business users or end users, number of core platform services; EBITDA) could be used for defining plausible quantitative thresholds capturing the type of gatekeepers targeted by the initiative under this assessment, i.e. gatekeepers engaging in the problematic practices described in Section 5.2.2. The above two combinations of parameters for a low and a high threshold have been selected as representative examples for the purpose of providing clarity in assessing impacts and trade-offs while comparing options, without prejudging other plausible use of parameters and their combinations. Small variations around those numbers do not produce significant changes in the number of possible gatekeepers. At the same time, some combinations have been considered but judged inappropriate. For instance, fixing the EEA annual group turnover at EUR 1 billion and the required number of core platform services showing a dependency (during a number of years) at a single one would mean that more than 25 providers of core platform services would be captured. 186 It was concluded that such thresholds would give rise to many false positives.

150.An alternative to quantitative criteria is to use qualitative criteria. Such a designation based on qualitative criteria would require a market investigation, which would assess whether the provider of core platform services has a significant impact on the internal market, operates a core platform service which serves as an important gateway for business users to customers, and enjoys (or is expected to enjoy) an entrenched and durable position in its operations. This could be the case where the contestability of the core platform service is affected, or risks being affected, on a lasting basis, due to the position of the gatekeeper, entry barriers derived from network effects, in particular in relation to its access to and collection of personal and non-personal data or analytics capabilities, scale and scope effects, customer bias or other structural market characteristics, with the effect of impeding innovation, high quality of digital products and services, fair and competitive prices and choice for business customers and users.

151.In summary the scope of the intervention is characterised by the main parameters of choice of core platform service where the data-driven advantages and network effects are strong drivers, as well as the criteria that determine the gatekeeper companies in scope. While there is a good consensus that the criteria should focus on size, intermediation power, and entrenched position 187 , there are choices available on the quantitative thresholds, or the use of qualitative thresholds.

Unfair practices

152.The public consultation offered strong support for an intervention tackling gatekeepers’ unfair practices. In fact, the large majority of the respondents to the OPCs and to the NCA questionnaire agreed that the Commission should be able to intervene in markets where gatekeepers are present, including a large majority of businesses and businesses associations, all civil society organisations (including consumer associations, NGOs and trade unions) and all public authorities. 188 Those respondents considered that this would both create the right innovation incentives across the market, and contribute to increased consumer choice paving the way for new platforms and innovative and privacy-friendly services.

153.Those disagreeing refer to the fact that the concept of a gatekeeper is too broad and should instead be assessed on a case-by-case basis and that the Commission can already intervene in the case of gatekeeper’s conduct using Article 102 TFEU. As explained in Section 5.1, the Commission considers that Article 102 is not sufficient to deal with all the problems associated with gatekeepers given that a gatekeeper may not necessarily be a dominant player, and its practices may not be captured by Article 102 TFEU if there is no demonstrable effect on competition. Moreover, Article 102 does not always allow intervening with the speed that is necessary to address these pressing practices in the most timely and thus most effective manner.

154.The table below summarises the list of obligations retained associated to the key unfair practices identifed against the evidence for each practice. A key element for such a list of practices is the applicability to a gatekeeper’s core platform services. In some cases, the practices at stake target specific core platform services.

155.The evidence included in the table consists of (i) antitrust decisions adopted by the Commission and other competition authorities, (ii) examples of complaints and investigations compiled by an external contractor in the context of the Impact Assessment study, (iii) evidence from the study supporting the Observatory for the Online Platform Economy, (iv) evidence from the reports by the expert group for the Observatory on the Online Platform Economy, complemented with stakeholder input, (v) studies conducted on digital sectors by other public authorities, and (vi) other Impact Assessments and regulations. 189 For some of the practices listed below there is no decision or judgment confirming its effects on the market. Nevertheless, the multiple complaints, investigations and studies raising awareness, and suggesting solutions, to those practices are a strong indication of their relevance and of their negative impact on the internal market.

156.Beyond the practices identified in the table, a range of other practices were examined. For example, economic regulation of costs and prices was examined and discarded. Other examples of specific practices examined but not retained include generic and broad rules on fair access conditions to any core platform service, or rules which would allow any provider of ancillary service to get full access and interoperability with a gatekeeper’s service. Other practices considered and frequently proposed in the literature – like for example banning the pre-installation of software – were replaced by more proportionate obligations – in this case, the possibility to give customers the possibility to always un-install applications – or dropped completely – for example, a general ban on tying.

157.Aside specific practices examined and discarded, a category of other obligations was also considered but rejected, namely broadly formulated or generic practices (such as self-preferencing in general) that would require an in-depth competition like analysis to be carried out.

Table 2: List of obligations with examples and underlying evidence

158.Finally, it is also worth noting that gatekeepers frequently raise arguments concerning the efficiencies that their practices bring about as a way to counterbalance and justify their potential negative effects. These arguments – raised not only in the OPC but also in numerous past and ongoing investigations (in fields such as antitrust, consumer protection or privacy) – are often one-sided and do not seem to match the evidence underlying this Impact Assessment including the calls for regulation raised by an overwhelming majority of respondents to the OPCs. Such efficiency-related defenses have also been rejected by the Courts as being unfounded. 204

Speed and Flexibility

159.Another important design parameter for the available policy options concerns the architecture of the intervention, notably speed of intervention, and the degree of flexibility concerning the main elements of the intervention.

160.Concerning the speed of application, the main choices are between (a) model of immediately applicable obligations, (b) a model where a degree of appreciation is necessary, notably as regards the implementation of a given obligation, and (c) a fully flexible model, where obligations or remedies are only imposed subsequent to an investigation carried out by an authority.

161.Such models have regulatory precedents or analogues in other acts of Union law. For instance, the Unfair Commercial Practices Directive (‘UCPD’) 205  and the Directive on Unfair trading practices in the agricultural and food supply chain 206 have models of immediately applicable black-lists of practices that are essentially lists of prohibited conduct.

162.In contrast, the European Electronics Communication Code offers a more flexible, case-by-case regulatory framework, whereby a regulator can impose a set of remedies following an analysis of problems in the internal market related to the telecommunications sector. 207

163.For the flexibility element, there are essentially three parameters that could characterise the main choices. First, a flexibility on the designation of a gatekeeper, e.g. by using only qualitative thresholds, or by updating the thresholds in light of market developments. Second, a flexibility on the list of practices that should be subject to a remedy and to their implementation. For instance, the list of practices could be left fully open, subject to an update, or selected after a market investigation or an analysis from a pre-determined list of practices. A third source of flexibility concerns the type of core services in scope, as discussed in the previous section. It is easily conceivable that future technological developments require different core services to be in scope of the instrument.

164.Stakeholders in the OPC have generally supported a mix of approaches, combining the speed of immediately applicable obligations and a more flexible approach. 208

Enforcement framework

165.A final important design parameter is linked to the enforcement powers necessary and available to ensure that the rules concerned are not undermined, ineffective or absent enforcement. This concerns individual cases of non-compliance as well as cases of more systematic non-compliance by gatekeepers. It is relevant to all options, and also linked to the monitoring of the implementation.

166.In order to ensure effective compliance with the ex ante rules, in any enforcement framework the Commission should have investigative and enforcement powers to allow it to investigate, enforce and monitor the ex ante rules, while at the same time ensuring the respect of the fundamental right to be heard and to have access to the file in the context of the enforcement proceedings. In particular, the Commission should have access to any relevant documents, data and information necessary to open and conduct investigations and to monitor the compliance with the obligations addressed to designated gatekeepers, irrespective of who possesses the documents, data or information in question, and regardless of their form or format, their storage medium, or the place where they are stored.

167.In order to have sufficient deterrence powers it is necessary that after a due process the Commission shall be able to impose fines and periodic penalties or take the necessary measures to restore compliance. In this respect the Regulation 1/2003 offers a well- known and legally sound model that can be replicated in Options presented in this Impact Assessment.

168.In the extreme case, where an investigation shows that a gatekeeper has systematically infringed the obligations and has further strengthened or extended its gatekeeper position, the Commission should be able as it is the case of Article 7 209 of Regulation 1/2003 to impose the structural remedies necessary to guarantee that market participants are not irreversibly harmed by this repeated and illicit behaviour. This consideration applies to all categories of intervention in terms of effectiveness.

169.However, to guarantee proportionality of the intervention the Commission should only impose structural remedies either where there is no equally effective behavioural remedy or where any equally effective behavioural remedy would be more burdensome for the gatekeeper concerned than a structural remedy.

170.In addition, the notion of systematic non-compliance should be linked with repeated infringements. Gatekeepers shall be deemed to have engaged in a systematic non-compliance in cases where the Commission issued at least two or more non-compliance or fining decisions. The Commission shall have the duty to explain whether and which remedy or remedies it preliminarily considers necessary and proportionate. At any point in time during the proceedings gatekeepers shall be put in condition to offer commitments that if accepted would terminate the infringement.

171.The above-described system is construed to safeguard proportionality of the intervention. It is based on existing instruments.

172.Structural measures are to be seen an ultima ratio measure for repeated infringement and to be taken into account where every other possibility has failed. To be noted that in the application of Regulation 1/2003 this circumstance has never occurred.

Summary and main trade-offs

173.In summary, the key parameters that determine the choice of options are related to the scope, the set of obligations related to unfair trading practices, the flexibility of the architecture and the scope of enforcement powers. These parameters are linked via a set of trade-offs that are set out below, mainly in terms of considerations of overall proportionality and effectiveness.

174.The first trade-off is linked to the gatekeepers in scope and the intensity of obligations. A wider scope of gatekeeper platforms corresponds with a lower intensity of the obligations linked to unfair practices and vice-versa, as the degree of harm caused by unfair practices is linked to the strength of the gatekeeper power.

175.The second trade-off is linked to the catalogue of obligations and the flexibility of the instrument. The more flexible the intervention, the less prescriptive the obligations need to be and vice-versa.

176.A third trade-off is linked to the speed and flexibility of the intervention. An intervention that requires a detailed analysis for each case will necessarily be slower in effect than an intervention based on a list of immediately applicable prohibitions.

177.The fourth trade-off is linked to the scope of the remedies in case of systematic non-compliance. An intervention that requires a structural remedy will necessarily be slower since it will require several enforcement steps before the effective compliance will be ensured. However, at the same time, it is an ultima ratio remedy that will only be considered if all other measures do not produce the required result.

178.While the set of parameters and trade-offs in this section theoretically allow many different combinations, not all of them are credible or meaningful.

Policy options

179.Based on the analysis of the main parameters that characterise the problem, and the trade-offs presented in the previous section, three distinct policy options have emerged from the available matrix of combinations as plausible ones and will therefore be fully assessed.

180.Option 1 is a non-dynamic option with a set of immediately applicable obligations addressing clearly defined unfair practices by gatekeepers designated solely on quantitative thresholds in specific core platform services. This option contains no dynamic elements, but is presented with distinct two sub-options on scope as distinct alternatives, on the basis of different thresholds. Sub-option A is presented as a sub-option with a small number of gatekeeper companies in scope (some 5-7 companies) while sub-option B contains a wider scope of gatekeeper companies (some 10-15 gatekeepers), based on a lower quantitative threshold.

181.Option 2 is a semi-flexible option, combining a set of immediately applicable obligations with some degree of flexibility, notably through a dialogue on some of the obligations and a mechanism for updating the practices and obligations. It further comprises a mechanism designating gatekeepers based on a combination of quantitative and qualitative thresholds and including the designation of emerging gatekeepers. Again, this semi-flexible option is presented with two sub-options that reflect alternatives on the scope. Sub-options A and B are sub-options on this semi-flexible option, following the same distinction on the quantitative threshold as Option 1.

182.Option 3 is a fully flexible option providing for a dialogue on all the obligations listed and a dynamic updating mechanism allowing for the inclusion of additional core platform services and of additional obligations where following a market investigation such an inclusion is considered appropriate and justified. Further, the designation of gatekeepers is based only on qualitative thresholds.

183.These options represent distinct alternatives based on the inherent trade-offs that underline the problem definition. They are distinguished by the architectural element relating to the flexible or dynamic character of the proposed intervention.

184.The table below gives an overview of these policy options.

Table 3: Parameters of policy options

Option 1 – Pre-defined list of gatekeepers and immediately applicable obligations

Summary of option 1

185.Option 1 would consist of the following elements:

(a)a closed list of core platform services identified in Section 5.2.1;

(b)designation of providers of core platform services as gatekeepers based solely on the quantitative thresholds; and

(c)the whole list of obligations identified in Section 5.2.2 would be immediately applicable without any ability of a regulatory dialogue.

Identification of core platform services

186.Option 1 would provide for a new targeted ex ante regulatory framework, which would apply to identified ‘core platform services’ (see Section 5.2.1) provided by designated gatekeepers. These core platform services define the perimeter both for the designation of gatekeepers as such, and for the identification of those individual core platform services provided by the designated gatekeeper which would have to comply with the clearly defined closed list of obligations identified in Section 5.2.2.

Designation of gatekeepers

187.Under Option 1, providers of core platform services would be designated as gatekeepers based on pre-defined quantitative thresholds. These quantitative thresholds would provide a high degree of legal certainty for market operators. Conversely, Option 1 would not allow any flexibility to identify gatekeepers on the basis of criteria other than the quantitative ones and would also not enable the providers of core platform services to, in exceptional circumstances, demonstrate based on serious and substantiated arguments that they do not meet the conditions in paragraph 133. This would also exclude any type of case-by-case analysis following an in-depth market investigation.

188.In view of this, and as explained in paragraph 148, two sub-options could be considered:

·Sub-option 1-A, i.e. high threshold implying the designation of five to seven gatekeepers.

·Sub-option 1-B, i.e. low threshold implying the designation of 10 to 15 gatekeepers.

189.Under Option 1, the fixed quantitative criteria would also exclude any possibility to identify and designate those providers of core platform services that are expected to enjoy an entrenched and durable position in the near future.

Identification of obligations applicable to gatekeepers’ core platform services

190.Under Option 1, a closed list of obligations that the designated gatekeepers would have to comply with would be defined in the rules themselves. These obligations would be set on the basis of the criteria presented in Section 5.2.2 and are identified in the Table 2 .

191.The designated gatekeepers would be required to comply with all the obligations laid down in the rules. Under Option 1 the gatekeepers concerned could not engage in a dialogue with the regulator about the measures they intend to take or have taken in order to comply with these obligations (i.e., the whole set of obligations would be immediately applicable).

Enforcement framework

192.Option 1 foresees implementation, supervision and enforcement at the EU level by the Commission as the competent regulatory body. Given the pan-European reach of the targeted companies, a decentralised enforcement model does not seem to be a conceivable option, including in light of the fragmentation that the initiative is supposed to address, nor would it be proportionate given the limited number of gatekeepers that would be in scope of the proposed framework. However, to integrate the national expertise in the platform economy, the initiative would envisage that the Commission consults a ‘network of regulators’ before taking decisions that could be considered under Option 1 (e.g. designation of gatekeepers; non-compliance; fines; period penalty payments; remedies decisions in case of systematic non-compliance).

193.To be able to effectively carry out its work, the Commission would enjoy clearly-defined and circumscribed procedural powers, which would include:

·The power to request information from the gatekeepers or third parties to determine compliance with the rules;

·The power to conduct on-site inspections to collect any information that may be necessary to establish such a non-compliance;

·The ability to adopt interim measures in case of a risk of serious and irreparable damage for business users or end users of gatekeepers, where there are strong indications of a prima facie finding of infringement of obligations addressed to gatekeepers;

·The ability to make legally binding voluntary measures that the gatekeepers may offer in the context of the non-compliance procedure to ensure the effective implementation and compliance with their obligations; and

·The ability to adopt non-compliance decisions, including fines and period penalty payments where necessary and justified.

194.In order to ensure the effectiveness and speed of intervention – as well as a way to ensure legal certainty and to replicate the obligatory nature that data-gathering powers would have on gatekeepers – proceedings would be subject to binding legal deadlines. Respondents to the OPCs and NCAs, generally argued in favour of binding deadlines for both the Commission and the businesses concerned in order to ensure expediency and legal certainty. 210 Respondents also added that deadlines would ensure a swifter outcome, which is all the more necessary, in particular in digital sectors, both for a swift resolution of the case and for providing sufficient legal certainty to the market. As regards binding deadlines for the businesses concerned, respondents argued that this would avoid risks of certain businesses slowing down the process with dilatory conducts, and that these deadlines should be coupled with the possibility of imposing fines for non-compliance to ensure speed and effectiveness.

195.In addition, in order to ensure due process and protection of rights of the parties to the procedure, it is important that any addressee of the decision has the opportunity of being heard on the final decision considered and that all decisions taken are subject to judicial review.

196.Finally, under this option the Commission would also have the power to ensure that in case of a systematic non-compliance further appropriate and proportionate measures are taken to ensure that objectives of the ex ante rules are not undermined. The exact scope of such behavioural or structural measure should be proportionate to the infringement committed and necessary to ensure compliance with the ex ante rules (see Section 5.2.4).

197.The enforcement powers required to enforce the prohibitions and obligations under this option are not new creations but on the contrary largely reproduce existing powers that the Commission has under the competition and regulatory frameworks 211 .

198.As regards the powers inspired from the EU competition acquis, the main source of reference would be Regulation 1/2003, the legal text governing the conduct of competition investigations by the European Commission. In this respect, Regulation 1/2003 contains also tools which largely mimic the five investigative measures referenced in paragraph 193 as being necessary for the enforcement of this option. Some differences between the investigative measures under Regulation 1/2003 and this option would, however, need to be included. Most notably, this would involve the inclusion of an explicit power to access databases, algorithms and other technical elements that are characteristic to the digital economy. While under Regulation 1/2003 the Commission also has the ability to access and conduct searches into these elements, the nature of the tool under the current impact assessment justify the inclusion of an explicit power in this respect. Remedies sanctioning a refusal to access those should also be provided for. Another difference would concern the fact that – at least for the time being – the ability to inspect other premises under Article 21 of Regulation 1/2003 would not seem necessary for the purposes of the current instrument.

199.In a similar vein, powers to request information from companies or the ability to propose commitments to the regulators are also commonplace in regulatory systems. One such example can be found, for example, in Article 78 of the EU’s telecommunications framework. 212  

Option 2 – Partially flexible framework of designation and updating of obligations, including regulatory dialogue for implementation of some

Summary of option 2

200.Option 2 would consist of the following elements:

(a)a closed list of core platform services identified in Section 5.2.1;

(b)a combination of quantitative and qualitative criteria to designate providers of core platform services as gatekeepers;

(c)the obligations identified in Section 5.2.2 would consist of immediately applicable obligations including some obligations where regulatory dialogue may facilitate their effective implementation; and

(d)new practices may be added on the basis of a market investigation.

Identification of core platform services

201.Option 2 would also provide for a new targeted ex ante regulatory framework, which would apply to identified ‘core platform services’ (see Section 5.2.1) provided by designated gatekeepers. These core platform services would again serve as the perimeter for the designation of gatekeepers as such, and for the identification of those individual core platform services provided by the designated gatekeeper which would have to comply with the clearly defined obligations as set out in Section 5.2.2.

202.Similarly to Option 1, once this list of core platform services would be identified, there would be no possibility to update such list beyond the revision of the rules themselves and there would be no ability to update them by means of tools provided in the rules themselves.

Designation of gatekeepers

203.Under Option 2, providers of core platform services would be designated as gatekeepers based on the combination of pre-defined quantitative thresholds but also following a case-by-case assessment in the context of a market investigation.

204.Like under Option 1, also under Option 2, two quantitative thresholds could be considered as defined in paragraph 148:

·Sub-option 2-A, i.e. high threshold implying the designation of five to seven gatekeepers.

·Sub-option 2-B, i.e. low threshold implying the designation of 10 to 15 gatekeepers.

205.Because of the combination of quantitative and qualitative thresholds, Option 2 would include a certain degree of flexibility, which would allow to capture two important dynamic elements of the platform ecosystem.

206.First, Option 2 would have the ability to designate gatekeepers not yet enjoying an entrenched and durable position, but which are expected to enjoy such a position in their operations in the near future. Such a designation would prevent core platform services, where these emerging gatekeepers operate, to tip because of weak contestability of the market concerned. The phenomenon of tipping – an irreversible loss of competition in a given market that occurs in a sudden manner – is further explained in Section 2.3.1.1 above.

207.Second, Option 2 would also foresee that the designation of the gatekeeper should be regularly reviewed where there would be a material change in any of the facts on which the designation decision was based, and where the decision was based on incomplete, incorrect or misleading information provided by the undertakings.

208.Finally, under Option 2, in view of the dynamic element and combination with the qualitative assessment of the gatekeeper status, the provider of core platform services would be able to present, in exceptional circumstances, serious and substantiated arguments to demonstrate that, in circumstances in which the relevant core platform service operates, it does not fulfil the objective requirements for a gatekeeper (see paragraph 133) and should therefore not be designated directly based on the application of quantitative thresholds, but only subject to a further investigation. The purpose of such an implementation of a legal presumption would not be to demonstrate, on pure economic grounds, efficiencies deriving from a specific type of behaviour by the provider of core platform services since this is not relevant to designation of such a provider as a gatekeeper.

Identification of obligations applicable to gatekeepers’ core platform services

209.As under Option 1, a closed list of obligations that the designated gatekeepers would have to comply with would be defined in the rules themselves. These obligations would be set on the basis of the criteria presented in Section 5.2.2 and are identified in the Table 2 .

210.The designated gatekeepers would be required to comply with all the obligations laid down in the rules. However, for some obligations, Option 2 would provide the gatekeeper with the possibility to discuss with the Commission the measures it intends to take or has taken in order to ensure their effectiveness. This would provide additional flexibility in tailoring the implementing measures by the gatekeepers to the given obligation and circumstances of each gatekeeper.

211.Option 2 would also include a flexible element by allowing to update the list of obligations whenever new unfair practices would be determined following a market investigation. Consideration was given to the question whether an additional flexibility in relation to the possibility to also update the list of core platform services should not be included in this option. This possibility has not be retained since it would bring Option 2 too close to Option 3; while the options’ design was aimed at ensuring a set of options that are not only plausible but also well distinct from each other as regards the relevant trade-offs, thus providing a wider array of policy choices.

Market investigation framework

212.Option 2 would envisage a possibility for the Commission to carry out a market investigation in the following types of situations.

213.First, the Commission would carry out a market investigation in order to designate on a case-by-case basis a provider of core platform services that meet the conditions referred to Section 5.2.1 as a gatekeeper. In doing so, the Commission would take into account a number of elements, such as the size, operations, the number of business users depending on the core platform service to reach end users and the number of end users, entry barriers derived from network effects and data driven advantages, in particular in relation to provider’s access to and collection of personal and non-personal data or analytics capabilities or scale and scope effects the provider benefits from including with regard to data.

214.In the context of the market investigation, the Commission could identify as a gatekeeper not only a provider of core platform services that already enjoys an entrenched and durable position, but also those providers that are expected to enjoy such an entrenched and durable position in the near future.

215.Second, the Commission could initiate a market investigation to identify possible new practices. A report summarising such a market investigation could serve as a basis for a possible revision of the Regulation, either based on a dedicated empowerment for the Commission enabling it to update the obligations in the rules themselves or by means of a full review of the rules. Under a market investigation additional digital services could also be assessed. However, in order to maintain some legal certainty, this could only result in an update of the list of core platform services in the context of the review of the Regulation, which could possibly take place every three years.

216.Third, the Commission would carry out a market investigation when there is the suspicion that a gatekeeper has systematically infringed the obligations laid down and has further strengthened or extended its gatekeeper position.

217.The different procedures to be followed by the market investigation framework would be uniquely designed for this option but would nevertheless have some similarities with the frameworks under EU competition law and sector-specific regulation.

218.In this respect, market investigations – like competition investigations – would be initiated by opening decisions and gatekeepers which are the object of them would have the opportunity to be heard in relation to the allegations raised by the Commission. Unlike in procedures under Regulation 1/2003, the use of the market investigation framework would also in some cases also be subject to timeframes.

219.In a similar vein, the procedures governing the market investigation framework would also have parallelisms with regulatory systems. One such example would, for example, be Article 63 of the EU’s telecommunications framework allowing for the designation of ‘undertakings with significant market power’. 213

Enforcement framework

220.Similarly to Option 1, also Option 2 foresees implementation, supervision and enforcement at the EU level by the Commission as the competent regulatory body for the reasons explained in Section 5.3.1.5.

221.Furthermore, the new rules under Option 2 would envisage the same set of enforcement powers for the Commission as envisaged under Option 1.

222.Finally, under this option the Commission would also have the power to ensure that in case of a systematic non-compliance further appropriate and proportionate behavioural or structural measures are taken to ensure that objectives of the ex ante rules are not undermined (see Section 5.2.4).

Option 3 - Flexible option based exclusively on qualitative scoping thresholds

Summary of Option 3

223.Option 3 would consist of the following elements:

(a)a closed list of core platform services identified in Section 5.2.1;

(b)designation of providers of core platform services as gatekeepers following a pure qualitative assessment;

(c)the obligations identified in Section 5.2.2 would all be subject to a regulatory dialogue; and

(d)new practices and new core services may be added on the basis of a market investigation.

224.Consideration was given to the question whether some quantifiable elements should not be included in this option. This possibility has not be retained since it would bring Option 3 too close to Option 2; while the options’ design was aimed at ensuring a set of options that are not only plausible but also well distinct from each other as regards the relevant trade-offs, thus providing a wider array of policy choices.

Identification of core platform services

225.Option 3 would also provide for a new targeted ex ante regulatory framework, which would apply to identified ‘core platform services’ (see Section 5.2.1) provided by designated gatekeepers. These core platform services would again serve as the perimeter for the designation of gatekeepers as such, and for the identification of those individual core platform services provided by the designated gatekeeper which would have to comply with the clearly defined obligations as set out in Section 5.2.2.

226.While Option 3 would start by focusing on a list of core platform services, the flexibility of the tool would mean that it should also be able to assess whether new digital services need to be incorporated. Where such a need is proven after a market investigation, the new digital service in question would be added to the scope of the rules.

Designation of gatekeepers

227.Option 3 does not include any quantitative thresholds and is solely based on case-by-case qualitative assessments. As such, these qualitative assessments would be the only way to determine which gatekeepers would fall under the scope of the rules and would seek to determine whether the provider of the core platform service has a significant impact on the internal market, operates a core platform service which serves as an important gateway for business users to customers, and enjoys (or is expected to enjoy) an entrenched and durable position in its operations.

228.Similarly to Option 2, Option 3 would have the ability to designate gatekeepers not yet enjoying an entrenched and durable position, but which are expected to enjoy such a position in their operations in the near future. However, contrary to Option 2, Option 3 would not be bound by any quantitative elements. While the inclusion of quantitative elements as guidance could in principle be implemented while preserving the flexibility of the tool, such elements would also defeat the purpose of Option 3 as they would still constrain its ability to look at any company, regardless of its size or position in the market. The inclusion of such thresholds would also undermine the flexibility of this option as it would require regular reviews of the legislation in order to update the (indicative) thresholds in a manner that does not constrain its flexibility.

229.As compared to Option 2, however, Option 3 could result in the designation of a higher number of gatekeepers, namely those active in the new digital services added to the scope of the rules after a market investigation.

230.Option 3 would also foresee that the designation of the gatekeeper should be regularly reviewed where there would be a material change in any of the facts on which the designation decision was based, and where the decision was based on incomplete, incorrect or misleading information provided by the undertakings.

Identification of obligations applicable to gatekeepers’ core platform services

231.As under Option 1 and 2, a closed list of obligations that the designated gatekeepers would have to comply with would be defined in the rules themselves. These obligations would be set on the basis of the criteria presented in Section 5.2.2 and are identified in Table 2 . Similarly to Option 2, Option 3 would include a flexible element by allowing to update the list of obligations whenever new unfair practices would be determined following a market investigation. Option 3 could comprise more obligations than Option 2 in case additional core platform services were to be included in the scope and new practices as regards those additional services were determined to be unfair.

232.Option 3 entails a maximum degree of flexibility amplifying the level of dialogue to the point of allowing potential gatekeepers to present allegations as regards all obligations.

Market investigation framework

233.Option 3 would envisage a possibility for the Commission to carry out a market investigation in the following types of situations.

234.First, by not including any list of quantitative criteria, the Commission would always need to engage in a market investigation to determine which providers of core platforms services should be considered as gatekeepers. In order to do this, the Commission would pay attention to the conditions prevailing in the market as well as to the position of the provider of services.

235.Second, a market investigation would also be used by the Commission to specify the services in which the provider would be behaving as a gatekeeper. In case the service in question is not pre-defined in the list of core platform services, the Commission could update the list prior to imposing any obligations.

236.Third, the Commission could also use a market investigation to update the list of practices that are unfair or are contributing to a lessening of the contestability of the market. Once this practice or practices are identified, possible obligations could be added to the list of obligations imposed.

237.Fourth, the Commission would carry out a market investigation when there is the suspicion that a gatekeeper has systematically infringed the obligations laid down and has further strengthened or extended its gatekeeper position.

238.The procedures to be followed by the market investigation framework would be similar to the ones described in Option 2.

Enforcement framework

239.Similarly to Option 1 and 2, also Option 3 foresees implementation, supervision and enforcement at the EU level by the Commission as the competent regulatory body for the reasons explained in Section 5.3.1.5. Furthermore, the new rules under Option 3 would envisage same set of enforcement powers for the Commission as envisaged under Option 1 and 2.

240.Finally, under this option the Commission would also have the power to ensure that in case of a systematic non-compliance appropriate and proportionate behavioural or structural measures are taken to ensure that objectives of the ex ante rules are not undermined (see Section 5.2.4).

Policy options discarded at an earlier stage

A broad scope across platforms

241.As indicated in the Inception Impact Assessment, an option of amending the P2B Regulation was considered. Further horizontal rules could be established for all 10 000 online intermediation services and search engines that are currently falling within the scope of the P2B Regulation. This could cover prescriptive rules on different specific practices that are currently addressed by transparency obligations and beyond.

242.The impact of this option on the internal market would be further harmonisation in a wider range of areas, but without necessarily addressing the issues at stake. As such, the option would include a risk for these issues to be addressed at national level, resulting in no substantial improvement in the functioning of the Single Digital Market.

243.Imposing stringent measures horizontally would risk being disproportionate and have a negative impact on innovation and competition in the online platform economy. Stricter rules under this option would be intrusive for many of the 10 000 entities currently falling within the scope of the P2B Regulation, but could be especially harmful for smaller platforms, possibly also limiting their growth. The option would be expected to increase innovation for platforms’ business users but the extra burden for smaller platforms would stifle their potential to invest in innovation. In the same way, this option would risk to negate any positive impact on the ability of business users to compete by the negative impact on the ability of smaller platforms to compete. Given the large number of platforms covered, and in order to avoid a disproportionately negative effect on smaller players, there is also a risk that the rules have to be toned down with the result that problems relating to gatekeepers would not be addressed in the most adequate and vigorous manner.

244.As any stricter rules would apply to all platforms, compliance costs would be more burdensome for smaller platforms with limited resources. As such these rules would have a limited impact on gatekeepers, and possible a negative impact on smaller platforms due to the disproportionate regulatory burden on them. Because of its wider scope encompassing all platforms, this option would also lead to high enforcement and coordination costs for authorities.

245.The impact on business users and SMEs would be dependent on whether they would qualify as platforms themselves. For non-platforms, the stricter rules would provide benefits across all platforms. For platforms, the impact would be beneficial if the costs they incur for complying with new rules are lower than the benefit from a fairer behaviour by the platform they use and/or compete with (and vice versa).

246.The impact on consumers would be limited, as the mitigated impact on the platform economy (due to its scope regulating also smaller platforms) would also limit this option’s positive impact on consumers.

247.Therefore, an option based on the P2B Regulation, targeting not only gatekeepers but all platforms, was discarded as this would constitute a mismatch with the problems and their drivers as identified in this Impact Assessment.

248.The reversed scenario – i.e. changing the scope of P2B Regulation to gatekeepers only – would not be a conceivable way forward as it would eliminate the beneficial impact of its fairness and transparency rules addressed to non-gatekeeper platforms.

Information obligations without addressing unfair conduct

249.Furthermore, at the Inception Impact Assessment stage, an option was considered which would empower a regulatory body to collect information from large online platforms acting as gatekeepers. These data-gathering powers would be supported by enforcement powers in case of refusal to supply this information. The purpose would be to better inform the implementation of the existing legal framework by gaining, for example, further insights into gatekeepers’ business practices and their impact on these platforms’ users and consumers, the scope of gatekeepers’ data gathering, treatment of their own downstream operations compared with those of third parties and indicators of the outcomes resulting from these practices.

250.The impact of this option on the internal market would consist of a better knowledge of platform ecosystems but it would not lead to any improvement of the internal market functioning as it would not set any behaviour-changing rules. Furthermore, by leaving the problems unaddressed the option would include a risk for the identified issues to be addressed at national level, resulting in legal fragmentation limiting the digital market functioning.

251.In the same way, the impact of this option on growth, innovation and competition would be limited, as it would only increase regulators’ understanding of gatekeepers’ trading practices and business models but not foresee any regulatory measures. There might be a reputational effect associated with more transparency but this would not affect competition (substantially).

252.The impact on platforms would be limited and focused on gatekeeper platforms as other platforms would not be subject to legal obligations and information requests would be proportionate. There would be a benefit to smaller platforms in the long run due to regulators’ better understanding of the issues and their effective redress. Enforcement costs would be incurred by the Commission, but this would be in the public interest since it would allow a better understanding of the platform economy and hence increase EU public administration capacity to tackle related issues.

253.The impact on business users and SME’s would also be limited. Irrespective of the type of business user concerned, i.e. platform or not, the information-gathering mechanism would only allow a better understanding of issues at hand but would not change competing business users’ situation vis-à-vis gatekeepers. While SME platforms and business users could be requested to provide information to enforcement authorities, this would not be a legal obligation for them. The authorities’ increased insight into gatekeepers’ practices would enable better regulation and enforcement and thus have a positive impact on all SMEs. This impact would, however, be delayed in time given the time-frames applicable to any legislative process.

254.There would not be a concrete impact on consumers in the short term, as this option would only provide further insight in how gatekeepers treat consumers and the role played by consumers in business models.

255.Therefore, an option limited to improving access to information on the issues at stake was discarded as it would be insufficient to address the problems identified in this Impact Assessment or affect their drivers.

A broad scope across markets

256.Finally, as indicated in the Inception Impact Assessment, an option of having a market investigation regime with a horizontal scope (i.e. extending to all markets) as opposed to just a digital scope was considered. Under this option, market investigations would also be applicable to non-digital markets, which are not directly part of the objectives of this initiative. In fact, respondents to the OPC indicated that market failures occur in all sectors and markets and highlighted that no sector is immune to (potential) market failures. 214 At the same time, a high number of respondents who indicated that market failures can occur in all sectors and markets mainly pointed to digital examples in their replies. Respondents indicating that market failures mainly or solely occur in digital sectors/markets argued that the characteristics of the digital sector (e.g. economies of scale and scope, data accumulation and dependency, network effects, lock-in, zero pricing) make digital markets particularly prone to the emergence of quasi-monopolistic market structures.

257.BEUC pointed out that competition law enforcement in digital markets, though important, has not been effective enough in dealing with all problems in these markets and consequently not been able to remedy, let alone prevent, harm to consumers in a timely manner. 215 Indeed, there is an extensive economic literature and numerous reports as explained in Section 2 describing the growth of digital markets and their particular characteristics that makes them prone to market failures, and where resources would be better focused, at least at the initial stage of any new investigation regime.

258.This option would not allow an immediate and continuous response to the most pressing instances of gatekeeper related market failures in respect of key digital markets. In fact, it would divert the focus of investigations to other markets which are not in the scope of this initiative and would therefore not be very effective in addressing the market failures listed in Section 2.3.1. In the particular case of the problem driver related to the fragmented regulation of digital markets, this option would not be effective as it would not set common rules to address unfair practices and increase market contestability in digital markets and would therefore not lead to a more coherent regulatory approach across the EU.

259.In terms of efficiency, this option would result in a large amount of resources being dedicated to other investigations, and thus limit those available to tackle the pressing problems in digital markets. The costs for the regulatory authority would be necessarily higher than all the alternatives considered with no additional benefits in terms of its ability to achieve the objectives listed in Section 4.

260.Therefore, an option to have a market investigation regime extended to all markets was discarded as it would be out of the scope of this initiative and not focused in addressing the problems identified in this Impact Assessment or affect their drivers.

6.WHAT ARE THE IMPACTS OF THE POLICY OPTIONS?

261.This section presents the main impacts of the three options described in Section 5.3 compared to the baseline scenario.

262.The categories of stakeholders which would be affected, directly or indirectly, by the retained policy options are: platforms (gatekeepers and non-gatekeeper platforms), business users depending on platforms (e.g. hotels, sellers in marketplaces, app developers, banks) and possibly competing with the gatekeeper, competitors (e.g. innovative entrants), consumers, and regulatory authorities. Impacts for these stakeholder categories have been assessed in the following sub-sections covering the internal market (Section 6.1), growth and productivity (Section 6.2), competition and innovation (Section 6.3), international trade (Section 6.4), employment (Section 6.5), businesses – i.e. gatekeepers and SMEs in their role both as competitors and business users – (Section 6.6), consumers (Section 6.7) and regulatory authorities (Section 6.8).

263.By way of background, among the respondents who replied to the relevant question in the OPC, 91% agree that there is a need to consider dedicated regulatory rules to address negative societal and economic effects of gatekeeper platforms. 216 This view is supported by many targeted submissions by different groups of stakeholders, such as small and medium platforms and their associations, telecom operators and their associations as well by national regulatory authorities in different sectors (e.g. electronic communication services).

264.For the impacts developed in this section see also Annex 3 to the Impact Assessment, which specifies in detail who would be affected by the preferred option and how.

265.The problems and their underlying drivers as identified in Section 2 can lead, individually and jointly, to a number of negative outcomes specified where relevant in this section.

Internal market

266. Preventing fragmentation of the internal market is one of the most important policy objectives enshrined in the Treaties of the EU, and preserving the cross-border nature of the platform economy contributes to this objective. A 2016 European Commission Communication on the opportunities and challenges of online platforms for the Digital Single Market stressed the pivotal role of online platforms in the European single market. Services and products such as search engines, price comparison websites, online marketplaces and creative content outlets offer strong links to the rest of the economy. 217 A conservative estimate at the time of the Communication put the number of EU companies in Europe ´heavily´ using online platforms to trade goods and services at one million, with more than 50% of these being SMEs.

267.A study requested by the IMCO committee of the European Parliament concludes that interventions aiming at increasing the contestability of the digital sector would have a significant positive and growing contribution to achieve all of the potential benefits of a Digital Single Market, also resulting in lower prices and greater consumer choice, productivity gains and innovation. 218

268.Christensen et al (2018) estimated, using the RHOMOLO model 219 , that implementing the third pillar of the Investment Plan for Europe, including efficiency gains from the Digital Single Market, would contribute to a 1.5% increase in GDP per year until 2030 and create between 1 and 1.4 million jobs. 220 In particular, the impact of a more efficient Digital Single Market ranges from 0.44 to 0.82% changes in GDP and between 307 and 561 thousand additional full-time equivalents (‘FTEs’).

269.As explained in Section 2, national legislations have started appearing or are under consideration in different Member States, which drives fragmentation of the Digital Single Market in the platform space. One of the main objectives of this initiative is therefore precisely to prevent the fragmentation of digital markets. Option 1 would already allow some quick alignment of platform-related rules across the EU through horizontal measures by relying on automatic quantitative criteria to identify gatekeepers and implementing immediately all obligations. However, given its static nature, it would leave scope for some market fragmentation. Under sub-option 1-A, the obligations would be applied to a smaller number of gatekeepers. This could create some fragmentation resulting from the different treatment given to the largest (and thus captured by the quantitative designation process) gatekeepers, as platforms exhibiting similar features and characteristics as the largest gatekeepers would be subject to no obligations. Under sub-option 1-B, this possible difference in treatment between gatekeepers is less likely but there is a risk that some platforms which do not exhibit gatekeepers features and characteristics are forced to comply with the obligations, reducing their ability to compete with the gatekeepers.

270.Option 2 would capture a broad scope of unfair practices (via a dynamic updating mechanism) and gatekeepers (including those platforms falling only under the qualitative criteria as well as those platforms that are expected to enjoy a gatekeeper position in the near future). Gatekeepers would thus be treated in a harmonised way across the EU. It would still imply some delay in enforcing the obligations for gatekeepers designated on the basis of the qualitative criteria. Under sub-option 2-A, and as compared to sub-option 2-B, these delays would be more predominant, as a higher number of gatekeepers would have to be designated via a market investigation. In any case, the possible fragmentation associated to the unequal treatment of gatekeepers discussed under Option 1 would be less likely and only occur temporarily.

271.Option 3 would, on the one hand, add to Option 2 by tackling a broader scope of unfair practices by gatekeepers in additional digital services, thus allowing full harmonisation across the EU. On the other hand, this option could result in a staggered implementation of the measures because of the delays associated to the need for the regulator (i) to conduct market investigations to designate every gatekeeper and (ii) to engage in dialogue with each gatekeeper for the implementation of all obligations.

Growth and productivity

272.The platform economy contributes heavily to the European economy as revealed by its size and is expected to continue to grow steadily. The digital economy was estimated to account for between 4.5% to 15.5% of global GDP in 2019, depending on the definition. 221 Traffic share is one of the most important proxies of the sector. The top 50 online platforms represent 60% of the traffic share in Europe reaching revenues for about EUR 276 billion in 2018 and employing almost 600 000 people.

273.The European market of online platforms makes a significant contribution to GDP and the European economy as a whole. Revenues of the sector in Germany for instance reached EUR 33 billion in 2015. Cross-border ecommerce in Europe was worth EUR 143 billion in 2019 (without travel), and 59% of this market, EUR 84 billion, is generated by online marketplaces. Marketplaces with European capital represent 11% of the market, an increase of 17% compared to one year before. 222  

274.As already explained in the recent P2B Regulation impact assessment, trade intermediated through online platforms is expected to follow an upward trend as most consumers opt for platforms when purchasing goods and services online. 223 B2C e-commerce turnover was growing at an average pace of 13% between 2014 and 2019 with turnover forecasted to hit EUR 621 billion in 2019 and is set to be worth EUR 717 billion in 2020. This sector is expected to increase in value by around 14% per year. 224 In addition, the COVID-19 crisis accelerated the shift to online retail at an unprecedented pace pointing to the importance of the online platform economy. Usage of digital devices has increased significantly during the COVID pandemic, which is likely to increase the relative importance of online platforms compared with the off-line world. Specifically, following the lockdown, one global survey found that consumers spent more time on social media and mobile applications (by 47% and 36% respectively). 225

275.More generally, several empirical studies confirm that more competition on markets results in higher productivity in affected industries, which translates into economic growth. 226 Other studies also confirm the positive effects of competition on the productive efficiency of companies due to (i) ‘between-firms’ effect, by which better companies succeed while the worst ones fail and leave the market, and (ii) a ‘within-firm’ effect by which companies in competitive environments are better managed. 227

276.Addressing gatekeepers’ unfair business practices would have a positive impact on the online platform economy in general. The envisaged measures would limit the chilling effects unfair conduct has on sales. Since gatekeepers are such an important channel to reach markets and consumers, business users argue that unfair practices (e.g. pretended privacy considerations, limitation to data access, etc.) would lead to up to 15% loss in their sales. 228 Businesses, especially smaller ones, would be more confident in engaging with gatekeepers if the latter (are obliged to) comply with clear fairness rules. From that perspective, a regulatory action would be expected to result not only in more sales through smaller platform but also to have a positive impact on market growth.

277.Importantly, competitive entrants contribute to growth in the digital sphere; the obligations considered – e.g. data access or interoperability - can allow entrants to grow and compete effectively. While digital market features exacerbating the problem drivers (such as e.g. network effects) cannot be changed, the rewarding effect on gatekeepers’ further expansion - when that expansion is due to unfair business conduct - can be attenuated by the measures proposed. The measures would thus create fairer and more equitable conditions for all market players, allowing them to take greater advantage of the growth potential of the platform economy.

278.All three options considered would reinforce trust in the platform business environment. Option 1 would do it in a quick but relatively static way, and could originate some frictions in the business environment by leaving no room for any implementation dialogue. In terms of sub-options, under sub-option 1-A those frictions would only affect the largest gatekeepers, while under sub-option 1-B, there could be a larger number of gatekeepers impacted. Option 2 would be less immediate but particularly effective in that it foresees an adaptable framework, based both on a clear set of immediately applicable obligations and a flexible list of obligations subject to an assessment of the applicability of the conducts to the specific case. It would also allow tackling practices in markets where there is a risk of tipping, and contribute to a more competitive platform ecosystem. Option 3 would additionally allow a dynamic updating of the list of core platform services, thus tackling a potentially larger set of digital services than Options 1 and 2. However, it would have the drawback of delaying (i) the implementation of obligations which could be made immediately aplicable under Option 2 and (ii) the designation of the largest gatekeepers that could be quickly identified on basis of quantitative criteria under both Options 1 and 2. By giving too much discretionarity power to the regulator, Option 3 could also give rise to a risk of a lower level of legal certainty, which would impact negatively the business environment.

Competition and Innovation

279.Weak market contestability and lack of competition - driven by gatekeepers’ strong bargaining power and market features leading to entry barriers - are among the problems identified in this Impact Assessment. There are strong links between patterns of innovation and competition. When businesses compete more fairly on their merits, this incentivises them to innovate and offer a better range of higher quality products and services that meet consumers' expectations. Greater competition also drives efficiency in processes, technology and service. According to Federico et al. (2019), a significant amount of innovation is driven by disruptive firms. By making its offer to customers attractive in a new way, a disruptive firm can destroy a great deal of incumbent profit while creating a large amount of consumer surplus. Competition enforcement precisely seeks to protect the competitive process by which disruptive firms challenge the status quo. 229 Several empirical studies confirm that an increase in competition leads to a significant increase in R&D investment by neck-and-neck firms. 230  Conversely, the view according to which market concentration or large firm size is associated with a higher level of innovation is not supported by empirical evidence. 231 Shapiro (2012) highlights the considerable empirical evidence that greater competition spurs innovation. 232

280.Innovation patterns in the online platform economy are characterised by the following trends. On the one hand, online platforms drive innovation, driven by a competitive strategy. On the other hand, network effects drive higher concentration which may hinder innovation because it remains concentrated among a reduced number of players. At the same time, gatekeepers – due to their impact on the entire ecosystem - are able to set innovation trends for their sector and even beyond (i.e. to non-platform companies). This has the double effect of spreading gatekeepers’ innovative solutions to smaller players but could also limit the emergence of other types of innovation.

281.Although the online platform sector invests heavily in innovation, smaller companies that depend on gatekeepers are discouraged from innovating so as not to compete with the gatekeeper. 233 Preventing patents or pre-emptive activities, for instance, is one way to gain monopoly power and to increase barriers to entry. If this pattern is dominant, the pace of innovation in the long run slows down. 234 Acquisition of startups is another way for gatekeepers to cement their market power. 235 While acquisitions may have a positive effect for entrepreneurship and innovation, in the long-run they may result in higher market concentration and insufficient diffusion of innovation. 236

282.The evidence shows the concentration of R&D investment among few dominant firms, and with a sustained trend. The trends in the investment in R&D depicted in our Impact Assessment study suggest a cluster of high volumes of investment among big five companies; and a widening gap across time between large and small companies. The study shows that financial resources that could be invested in R&D are diverted to mergers and acquisitions, which results in higher market concentration instead of increase in the quality and quantity of products and services for consumers. The pattern of innovation dedicated to competing 'for the market' has a detrimental effect on consumer choice and surplus. 237  

283.Moreover, market concentration results in accumulation of cash-flow that is available for R&D investment and innovation or mergers and acquisitions. The Impact Assessment study illustrates the concentration of liquidity among the top five companies, each of them ranging between 10% up to 30%, while the remaining 17 companies are on average below 1%. Five companies accumulate 90% of total free cash-flow that could be distributed among all 22 companies. This suggests that smaller companies may face some financial constraints, failing to attract venture capital to finance R&D projects, while large firms have enough own funds to embark on innovation.

284.Furthermore, Carayannis et al (2014) shows that innovation and productivity are important drivers for competitiveness. 238 Autor et al (2020) and Decker et al (2018) show that a growing productivity gap between very big firms and the rest may result in lower business dynamism and lower productivity growth. 239  A more efficient Digital Single Market with the right incentives to innovate should contribute to a more competitive EU digital economy. The measures under consideration are the most effective in increasing market contestability and can be expected to contribute to lower prices for business users due to increased competitive pressure. For instance, promoting switching through e.g. rules against the misuse of data, self-preferencing, or lack of inter-operability can enhance competition and contribute to dynamic patterns of innovation.

285.Option 1 is expected to have a positive and quick impact on overall innovation and competitiveness since it would immediately create a fairer and more balanced business environment for business users and platforms; the largest gatekeepers’ compliance costs may decrease to certain extent gatekeepers’ innovation ability but given those gatekeepers’ financing capabilities, the regulation would not substantially affect their innovation capacity. The lack of flexibility of Option 1 in relation to the implementation of all the obligations could however have a negative impact on the innovation efforts of those companies. Depending on the sub-option considered, the potential effects would be different: in case of sub-option 1-A, this lack of flexibility would affect only very large platforms (which are nonetheless the ones with the highest financial capabilities), while under sub-option 1-B, it would affect many more platforms (including some that may not exhibit the features and characteristics of a gatekeeper).

286.Options 2 and 3 would in principle affect more platforms, including those that are expected to enjoy an entrenched gatekeeper position in the near future. In theory this could have a direct negative impact on the innovation incentives of some smaller gatekeepers. However, by more broadly and flexibly addressing the issues encountered by gatekeepers’ business users and creating more competitiveness opportunities, these two options would allow the creation of a healthier business environment for other platforms contributing to restoring and/or installing competitive dynamics in the platform economy. Alternative platforms are currently facing a number of challenges e.g. for developing compelling offers (lack of data and consumers due to strong network effects), for accessing venture capital for competing services, portability, risk of leverage, etc. Also, business users (e.g. e-commerce merchants, service providers and application developers) face issues such as dependency, unfair contractual relations, unequal distribution of revenues/profits and exclusion. In light of this, the expectations for Options 2 and 3 are to spur overall technological innovation in the digital markets (concentrated so far within a limited number of gatekeepers) to other market players, thus creating more competition and innovation to the ultimate benefit of consumers. Both options are in this respect estimated to yield direct benefits of many billions of euros annually, in addition to improved innovation levels and entrepreneurship, which are complex to quantify in precise terms but likely equally if not more important in size and impact. Option 3 could have a broader impact than Option 2 by potentially affecting companies in a larger set of digital markets, but would also give rise to a lower level of legal certainty for gatekeepers and business users as a result of the excessive discretionary powers attributed to the regulator, and thus potentially risk some of these companies’ innovation efforts.

International trade

287.The promotion of higher competitiveness of digital markets is of particular importance in increasing trade and investment flows. According to an United Nations Conference on Trade and Development (‘UNCTAD’) report, 240 digitalisation contributes significantly to increasing the scale, scope and speed of trade. 'Information and Communication Technologies (‘ICT’) products are already a significant part of the global trade (in 2017 they are estimated to have reached USD 530 billion, representing 10% of total global trade in services).

288.All three options are designed in such a way as to target any gatekeeper platform in an objective and non-discriminatory manner (see Section 5.2.1). The objective scoping criteria applicable to all options target EU presence and do not take into account the location of the corporate headquarters of the company in question. In doing so, the options would be future-proof and consistent with the EU’s international obligations, including non-discrimination under the General Agreement on Tariffs and Trade and the World Trade Organisation. The EU is a major market which will remain open for business but competition in the EU should remain fair and markets contestable.

289.In addition, as mentioned in Section 1.1, an intense debate is on-going about the need to regulate gatekeepers in most jurisdictions around the world such as Japan 241 , Australia 242 , US 243 or China pointing to the global consensus on the need to complement competition policy with ex ante measures (see also Annex 5.3 to the Impact Assessment). This debate has included deep reflection processes in most of the EU’s main trading partners, many of which are considering options similar to the ones presented in this Impact Assessment.

290.For example in the US, antitrust hearings of Amazon, Apple, Facebook, and Google have taken place in the US Congress House of Representatives and before the Federal Trade Commission. 244 In October 2020, the House of Representatives’ Committee on the Judiciary issued a Majority Staff Report in which a broad range of significant remedies are proposed, following a detailed assessment of the effects of a number of unfair and anticompetitive practices by these platforms, in order to restore competition in digital markets. 245 These remedies notably include structural separation, line of business restrictions as well as non-discrimination rules for dominant platforms including on access and pricing.

291.Finally, the present initiative would establish a proportionate regulatory framework promoting a fair and contestable online platform environment in the EU, one in which new platforms can emerge and scale-up, to the benefit of users around the globe, not just in the EU.

Employment

292.An overview by the OECD of the main literature covering the links between competition and employment confirms that market competition stimulates employment growth in the long term. 246 The aggregate effect mainly results from a positive impact on productivity growth, which increases labour demand, and through aggregate demand, given that more competition lowers prices and therefore tends to increase real wages. This generates a virtuous circle of output and demand growth in the long run. 247  

293.In the short run, the response to increased competition can lead to an increase in unemployment, e.g. through process innovation that replaces labour intensive machinery with new machines to increase productivity at the cost of labour. However, econometric simulations of the effect of increased competition leading to redundancies in an industry demonstrate a return to a steady growth path with rising employment after two to three years. 248  

294.According to an UNCTAD report, 249 digital transformation has strongly contributed to job creation across the G20. Between 2006 and 2016, total employment in the G20 grew by 13%, a net gain of almost 127 million jobs with highly digital-intensive sectors contributing with 43% of these net job gains. Jobs in the ICT sector comprised 11.8% of total employment of the G20 countries, in 2017. The Covid-19 crisis called for the adoption of new labour regulations favouring teleworking regimes. Digital services are of extreme importance as tools enabling teleworking regimes. Therefore, making these services more accessible is even more important today for a functional labour market.

295.One of the studies carried out by the Commission 250 suggests the possible creation of thousands of additional jobs in case of regulatory corrective measures (see Annex 3 to the Impact Assessment). Even under the assumption that no additional jobs would be created, given the millions of people employed in the sector and the millions of SMEs depending on online platforms to reach their customers, taking adequate measures to ensure the proper functioning of the platform economy would safeguard these millions of jobs.

Businesses 251

Gatekeepers

296.First and foremost, it should be stressed that there is broad consensus across various firms of different sizes and business models in the tech community that there is a need for rules addressing the detrimental impact of gatekeeper practices and conduct. 252 Most respondents to the OPC, including businesses and business associations, consumer associations and NGOs, also agreed with the possibility to adopt a combination of policy options to address concerns in digital markets. 253

297.Second, as explained in Section 5.2.1, this initiative also foresees a mechanism for the designation of gatekeepers subject to the list of obligations that would capture those that are effectively gatekeepers prone to engaging in unfair conduct and/or reducing the contestability and competition in digital markets.

298.Third, the targeted scope of options imposing rules only on the largest platforms, or on undertakings contributing to a market failure, strongly contributes to the proportionality of any potentially resulting compliance costs.

299.Evidence shows gatekeepers’ increasing supra-normal profits as well as their ability to (i) obtain conditions that would not be possible under normal market circumstances, and (ii) act independently from competitors, business users and consumers. For instance, the graph below shows that some of the largest companies by market capitalisation included in the S&P 500 index in November 2020 are the companies running some of the most important digital platforms. 254 Also the multiple antitrust investigations on abuse of dominance against many gatekeepers provide relevant evidence about the unfair conditions imposed on business users by those gatekeepers (see Section 5.2.2).

Figure 3 – Market capitalisation of largest companies included in the S&P 500 index, Nov 2020

300.Compliance costs under all three options would largely substitute for the already high costs large platforms incur for complying with divergent regulatory measures gradually put in place in different Member States. Such costs would imply some additional legal compliance officers to check company policies against the new rules; some employees to interface with the regulator and respond to requests for information. These would be higher the longer the list of obligations and the broader the digital services in scope. Compliance costs would thus be the highest under Option 3 and Option 2 as compared to Option 1 as it would include potentially more practices as a result of the updating mechanism and the need to reply to more requests for information in the context of market investigations. On the other hand, the fact that Option 1 would not allow for any dialogue for the implementation of the obligations, would give rise to additional compliance costs as compared to the other two options. In terms of the sub-options, sub-option A would result in lower total costs than sub-option B as it would affect a smaller number of gatekeepers, although the cost per gatekeeper should be the same for both sub-options.

301.Compliance costs under all options would be miniscule as compared to the gatekeepers revenues and could be absorbed by gatekeepers with little incentive for them to pass on costs to business users (e.g. by limiting their access to the gatekeeper platform) or to consumers (see Section 6.1.8).

302.Indirect (other than compliance) costs may be higher, as proposed measures are expected to have impact on gatekeepers’ business models and potentially reduce their supra-normal profits. The impact of such changes is difficult to quantify. While some loss of revenue for gatekeeper is expected, there are no indications that this would result in significantly higher fees and/or reduced quality for businesses and consumers. Consumers are at the core of platforms’ business strategy and, due to the relevance of indirect network effects and economies of scale, gatekeepers need to attract an important number of consumers in order to be able to (i) attract businesses (and vice versa) thus allowing online matching of offer and demand, and (ii) benefit from the virtual growth cycle characterising the platform economy.

303.All three options would moreover not be geared towards eliminating legitimate monetisation opportunities. They would aim at eliminating unfair behaviour towards business users and other market failures, thus rather enhancing trust in the platform business model. A set of measures that contribute to a more dynamic online platform economy and more contestable markets would particularly benefit smaller competitors who would face lower barriers when entering the market. It can therefore be expected that an increased market contestability would continue to incentivise gatekeepers to bring innovative products to the market and compete for consumers and business users; this even in case gatekeepers’ business models are impacted by the regulatory measures.

304.Fourth, all options are designed in a targeted way, taking into account the currently available experience and evidence about the impact of specific unfair practices by gatekeepers on their business users and customers as well as on the contestability of digital markets.

305.Given that the rules only aim to prevent unfair and harmful conduct, they should not hamper market entry (even) by gatekeepers if the latter is based on fair means of competition. As far as they do not use their market positon in an abusive way their ‘first mover advantage’ could be preserved.

SMEs

306.SMEs would not be targeted by the list of obligations as they are very unlikely to qualify as gatekeepers. On the contrary, the adoption of rules levelling the playing field would allow SMEs (including business users competing with gatekeepers) to grow throughout the internal market.

307.All three options foresee a comprehensive form of regulatory oversight and SMEs would benefit from a more innovative and competitive business environment incentivising them to seize the digital single market opportunities and grow (see Annex 3 to the Impact Assessment).

308.Competitors and new entrants would benefit from the levelling of the playing field and from enhanced opportunities to scale up and compete with these gatekeepers as a result of the removal of important barriers to entry and expansion. Measures preventing unfair self-preferencing and limitations in interoperability would give them the ability to compete on the merits (e.g. develop their own distribution channels or their own ID services). Data-related rules, including data portability, which would facilitate switching and multi-homing and thereby increase potential user base would allow them to bring innovative solutions to the market. Measures promoting multi-homing and user switching would give competitors and new entrants a real chance to capture a new stream of demand, propose competitive offers and grow. The increase transparency would give them opportunities to compete more equally with the gatekeepers. Rules on MFN clauses would increase incentives for competitors to develop alternative (to the gatekeepers’ ones) distribution channels, since they could expect that better service or lower price can be awarded by the business user with better commercial conditions.

309.Given that measures envisaged under all three options are aimed at increasing market contestability, it could be expected that they would result in more competition for business users. Business users would have more confidence in selling online, as they would be protected from unfair practices. Measures against data misuse would prevent that their data are exploited for the only benefit of the platform. Access to data generated by business users’ activity on the platform would allow them to adjust their business model to demand and better meet customers’ expectations. Business users would have the ability and incentive to choose among different platforms where to offer their service/product. Business users would have an increased possibility to multi-home and switch thus benefitting from increased choice of services and the ability to combine services according to their actual needs and interests (due to the obligation for gatekeepers not to make the use of the core service conditional upon the use of ancillary services). They could as a result benefit from lower prices for intermediation services and reduced distribution costs.

310.Finally, access to digital markets allow SMEs to increase their productivity and reduce their costs. According to a study from OECD countries, in 2015 only 20% of SMEs engaged in sales through e-commerce, against 40% of large firms. This digital gap slows productivity growth and widens inequalities. More competitive digital markets resulting in more affordable services would allow SMEs an easier access to digital technologies. Ultimately, given that SMEs are the bulk of many national economies, a massive adoption of digital technologies by them would generate a shift of aggregate productivity and welfare. 255 Since several of the business models of the gatekeepers are extensively benefiting from network effects and thereby large number of business users or end users, it is not expected that the obligations introduced by a new framework would result in gatekeepers terminating provision of services to SMEs that are often dependent on these gatekeepers and their core platform services. Not only would this remove many of the benefits that gatekeepers enjoy due to their unique position, but could further accelerate switching by both business users and end users to alternative providers of same or similar core platform services.

311.Option 2 and 3 as compared to Option 1 would allow the Commission to address the issues SMEs face in the dynamic digital markets in a more agile way, including issues associated to markets that risk tipping in the absence of an intervention and new unfair practices by gatekeepers. By foreseeing a dialogue between the regulator and gatekeepers for (some of) the obligations, they would allow a more flexible implementation of those obligations that would disrupt less the commercial relationship between gatekeepers and their business users. By foreseeing the possibility to resolve problems stemming from additional core platform services, Option 3 would additionally allow capturing the fast changing character of digital markets, but at the same time would give rise to a risk of a lower level of legal certainty to gatekeepers and business users as a result of the discretionary power given to the regulator. On the other hand, Option 1 could have an immediate effect on SMEs by allowing the automatic identification of all gatekeepers under scope and the immediately implementation of all obligations. In terms of sub-options, sub-option A would benefit more the business users of the largest platforms that would be captured by a high quantitative threshold, while sub-option B would benefit also business users of smaller platforms.

Consumers

312.Digital markets are becoming more and more relevant for consumers. According to the Digital Economy and Society Index (‘DESI’) 2020’, internet use has continued to increase year-on-year with 85% of Europeans surfing the internet at least once per week. 256 Using the internet for listening to music, playing games or watching videos is still the most common activity (81% of individuals). Reading news online is the second most popular activity (72% of individuals), followed by e-commerce (71%), bank online (66%) and social networks (65%). According to Eurostat figures, more than six out of 10 consumers from the EU28 made online purchases in 2019, the highest proportion made purchases three to five times in a period of three months and bought goods or services for a total of between EUR 100 to EUR 499. 257 Improving competition enforcement in digital markets is thus particularly relevant for the protection of European consumers.

313.Even though the digital sector and the companies offering digital services contribute strongly to consumer surplus, the increased market concentration in digital markets does not allow consumers to enjoy the full potential of these dynamic markets. In fact, the high concentration level is detrimental for consumer surplus as it results mainly in lower choice and higher prices/costs. Although data to estimate the loss in consumer surplus is limited, there is some illustrative evidence. For example, if commission fees in large app stores were to be reduced from 30% to 15%, the average prices of apps and digital content acquired through these apps would fall, which would increase consumer surplus by up to EUR 490 million in the EU per year based on Statista data. 258  

314.The choices for consumers are limited by lock-in effects and lack of innovative alternatives that are restricted by gatekeepers’ unfair business practices and more generally by the market failures in digital markets. In the longer run, consumers risk experiencing lower quality and/or less innovative services and/or higher prices. This initiative aims at addressing these concerns with a view to ensuring optimal and secure consumer experience online.

315.A regime that protects EU consumers from business practices that keep the prices of goods and services artificially high would ensure that consumers have access to better quality, wider choice and innovative goods and services at affordable prices. Numerous studies confirm the benefits of competitive markets for consumers. 259 More competitive digital markets will allow consumers to multi-home among alternative platforms offering differentiated commercial propositions. In addition, some of the measures considered under the options aim at reducing the search and switching costs associated with mulit-homing, for instance by allowing portability of data, creating conditions for interoperability, increasing transparency in the market, etc.

316.All options would indirectly contribute to safeguarding value added for consumers and to ensuring greater respect of privacy and consumer interests. 260 This would be achieved by contributing to (i) fairer competition on gatekeeper platforms (intra-platform competition) and among platforms (inter-platform competition), (ii) stronger contestability of the markets where gatekeepers are present, and (iii) better functioning of the internal market through enhanced regulatory oversight at EU level.

317.It is also important to notice that although interventions foreseen by this initiative may require changes to the existing business models, this does not risk harming consumers. The need for gatekeepers to maintain a large user base in order to optimise indirect network effects and higher level of contestability and competition in which the assessed measures would result, will rather increase gatekeepers’ incentive to innovate and offer lower prices.

318.It could also be argued that the rules under assessment would lead to curtailing the size of network effects and economies of scale thus reducing associated advantages for consumers. It is important to note that the objectives behind the measures considered aim at allowing also non gatekeeper platforms benefitting from such advantages. This would contribute to a competitive dynamics that benefits consumers who will be able to benefit from larger choice while gatekeepers can continue enjoying important network effects and economies of scale and scope.

319.It should be stressed that the obligations under this initiative will neither ban specific monetisation models (such as ad-based models) nor prevent the uptake of new services by gatekeepers - they prevent them from acting unfairly in their operations and reduce competition in the markets where they are present. The obligations envisaged would also not prevent an emerging gatekeeper from enjoying network effects and economies of scale and scope, instead they would ensure that other market players can also benefit from those features and would thus be able to compete under fair terms and innovate.

320.Even in those cases where, due to the multi-sided character of platform markets, there is a cross-subsidisation between the different sides with consumers benefitting from zero prices, additional regulation-compliance costs for gatekeepers cannot be expected to translate into ‘higher’ prices for consumers. This is so since, as explained in Section 6.6, consumers are at the core of platforms’ business strategy of indirect network effects and feedback loops. This is evidenced by the fact that in services like general search and social networks, even smaller platforms offer their services to consumers for free while obtaining their revenues via advertisers. In addition, gatekeepers would not risk losing consumers by setting prices for services which are currently free of charge. Consumers would expect zero-priced services to remain free of charge. Setting a price for gatekeepers’ services that are currently free would be perceived by consumers differently as compared to increases in already existing monetary prices (i.e. not zero). Consequently, any attempt by gatekeepers to make users pay for services that were previously offered for free would imply the risk for them of reducing the attractiveness of their services and of encouraging users to switch to other platforms continuing to offer their services free of charge. 261

321.Instead of increase in consumer prices, the expected increase in market contestability and competition would increase the diversity of offers available to consumers and would reduce the prices for advertisers, which would then indirectly translate in lower prices charged by those advertisers when selling their products and services to consumers. 262 , 263 Current excessive expenditures on advertising per user are driven by high market concentration, and could be a proxy for consumer detriment. For example, in 2019 a total of EUR 55.4 billion was spent in digital ads in 21 countries of the EU (including the UK) 264 , corresponding to around EUR 110 on advertising per user per year. A more contestable and competitive market would reduce those costs significantly (e.g. a reduction of 10% in ads expenditure would already generate gains of more than EUR 5 billion per year). In addition, for many digital markets where consumers are offered services ‘free of charge’, in practice they receive the service in exchange for their attention and their data, which can then be monetised through digital advertising. In a more contestable and competitive market, it would be clear to consumers what data is collected about them and how it is used and, crucially, consumers would have more control over their data. 265

322.Moreover, the Impact Assessment study for this initiative estimates that gatekeepers’ financial resources that could be invested in R&D are currently diverted to mergers and acquisitions, which results in higher market concentration instead of increase in the quality and quantity of consumer products and services. This pattern of innovation dedicated to competing 'for the market' has a detrimental effect on consumer choice and surplus. In addition, the positive impact on innovation stemming from higher market contestability is not limited only to diversion of money from mergers and acquisitions to R&D. Other expected indirect effects include an increase in entrepreneurship and creation of new products and solutions meeting consumers' needs rather than focused on exploiting a gatekeeping position. This may have a multiplicative effect increasing the size of the European single market, and hence, GDP and online cross-border trade. All options are estimated to allow to recover to a large extent this opportunity cost. All options would thus have a clearly positive effect on overall welfare.

323.As shown in Section 6.6.2, all three options would lead to positive implications for business users who would benefit from reduced prices for intermediation services. This in turn would allow business users to lower prices for consumers and offer them higher quality of service. Consequently, consumers would benefit from increased choice of products and services, better tailored to their needs (since they would e.g. be able to have a direct contact with businesses), and offered by different business users possibly through a larger number of platforms. This could lead to higher search costs but consumers would still have the possibility to use the gatekeeper services, if they find it preferable to use a single platform; their choice would however not be limited to offers provided through/on the gatekeeper platform. They would also benefit from lower prices for intermediation services which would be passed down to consumers in the form of lower prices for goods and services, thus generating cost savings to consumers.

324.Following the above, all three options would generate high benefits for consumers. Option 1 would generate immediate benefits but to a smaller extent given the risk associated the static nature of its application. Option 2 would be more flexible, thus favouring a more effective implementation of obligations and designation of gatekeepers. This would create more competitiveness opportunities. It would also allow tackling practices in markets where there is a risk of tipping as well as new unfair practices by gatekeepers, and thus generate more consumer benefits. Option 3 would on the one hand generate benefits for consumers of a broader range of services, and on the other hand, risk some innovation efforts by businesses (because of the low levels of legal certainty) with the consequent negative impact on consumers. In terms of sub-options, sub-option A would benefit more the consumers of the largest platforms and of the business users of those platforms, while sub-option B would benefit also the consumers of smaller platforms and respective business users. A possible drawback of sub-option B would be the risk of preventing platforms that are wrongly designated as gatekeepers from competing intensively with those gatekeepers, which would thus reduce their ability to innovate and launch new services in the market.

325.Detailed overview of the implications for gatekeepers, competitors, business users and consumers is presented in Annex 3 to the Impact Assessment.

Regulatory Authorities

326.All three options imply enforcement costs to be essentially incurred by the EU Commission, with some administrative burden for national authorities. This includes the costs with preparing and processing information requests as well as the preparation of guidelines, designation of gatekeepers, enforcement of the general obligations, including the specification of some of the obligations. Annex 3 to the Impact Assessment provides a qualitative and quantitative overview of these costs. As compared to Option 1, Option 2 and Option 3 imply additional resource-related costs both for the Commission and for national authorities. However, it can be objectively considered that this higher administrative burden would be largely outbalanced by the benefits of reducing the impact of practices which severely undermine the trading conditions for millions of business users and further entrench gatekeepers’ incontestable positions.

7.HOW DO THE OPTIONS COMPARE?

327.This section assesses the effectiveness, efficiency, coherence, proportionality and subsidiarity of the different policy options as compared to the baseline scenario and among each other.

Effectiveness

328.Three parameters appear essential for assessing the effectiveness of each option: legal certainty, speed of intervention and flexibility of the approach. The relative importance given to each of these three parameters is specified in the following three paragraphs.

329.Speed of intervention is essential in digital markets where, due to the market specificities explained in Section 2, the larger the gatekeeper the greater and quasi-automatic its capacity to gain power and strengthen its position, further reinforcing its ability to engage in unfair practices. In particular, in the digital sector it is common to observe markets tipping quickly in favour of one gatekeeper once that gatekeeper has obtained a certain advantage over rivals. The unfair practices identified is this Impact assessment are harmful and action is required in the most efficient manner possible. They affect negatively SME business users and small scale platforms, which may force the latter to exit the market, thus further weakening market contestability and strengthening legal fragmentation issues. Such negative effects of the problems identified could not be easily reversed and should therefore be addressed in a timely manner preventing their further proliferation and irreversibility. Speed is therefore given important consideration in the options’ comparison in this Section and in Table 4 .

330.Legal certainty is important for meeting expectations of all economic actors interacting in a given ecosystem. Together with regulatory predictability, legal certainty guarantees business trust and creates the right incentives for investment and innovation, for both gatekeepers and SMEs. It is therefore given important weight in Table 4 and in the overall assessment in this section.

331.Flexibility is an important criterion with a view to guaranteeing that a system is future-proof and agile. In the present case flexibility could be introduced at several levels of the options, namely, in relation to the designation of gatekeepers, the implementation of the obligations, the update of the list of obligations and of the list of core platform services. Some of these elements are more relevant than others and an excess of flexibility may not always be desirable as it creates negative externalities on other parameters. In fact, the optimal level of flexibility needs to strike the right balance between a regulation being agile and providing for a solid regulatory intervention setting a stable and clear framework. Flexibility has therefore been given less weight when comparing options’ effectiveness.

332.By including a set of obligations on gatekeepers’ behaviour the three options would contribute to both objectives of addressing unfair practices by gatekeepers and facilitating further contestability of the platform markets concerned. Measures related to data portability as well as interoperability and self-preferencing are particularly important for the objective of addressing unfair practices by gatekeepers. Such measures allow business users to benefit from fairer business conditions in relation to gatekeepers’ core platform services, thus also contributing to a level playing field. Gatekeepers would no longer be able to benefit from preferential treatment that derives from unfair behaviour, e.g. in terms of display/ranking or conditions of data access, portability, interoperability, which would also facilitate further contestability of the platform markets concerned. Rules set for anti-steering, side-loading as well as obligations concerning other unfair practices address the issue of unfair platform-to-business practices in specific contexts (i.e. economic dependence of one of the parties; imbalance in commercial relationship) thus contributing to more balanced P2B relations and acting on the imbalance of bargaining power (i.e. one of the drivers behind the fairness concerns). At the same time, such rules allow to address the weak contestability on digital markets since they would contribute to business users’ and consumers’ ability to use alternative services to those offered, or in certain circumstances even imposed by gatekeepers, thus attenuating both entry barriers (driving weak market contestability), and consequently businesses’ economic dependence on gatekeepers (the other driver behind unfairness). Consumer choice – which is closely related to competition and hence to market contestability - could also increase directly, notably requiring consumer portability provisions for gatekeepers, making it easier for consumers to switch (thus acting on the entry barriers driving the weak contestability problem). Indirectly, consumer benefits would also derive from lower prices for gatekeepers, although rules would need to be designed to avoid adverse effects on security and privacy, for instance.

333.Option 1 would contribute to the objectives of addressing gatekeepers’ unfair conduct and ensuring contestable and competitive digital markets, by allowing to tackle those gatekeepers’ unfair practices on the basis of a list of obligations. Given that all obligations under this option would be immediately applicable, they would have direct quick effects.

334.However, given that the designation of gatekeepers is based only on quantitative criteria, it could lead to type I errors (false positives) or type II errors (false negatives) depending on whether the threshold would be set at a low level (as per sub-option 1-B) or at a high level (as per sub-option 1-A). In fact, in the case of sub-option 1-A there would be the risk of failing to identify gatekeepers that, similar to the designated gatekeepers, may equally have an important internal market impact, operate an important gateway to end users and have an entrenched position, but which are relatively smaller. This would result in the unfair practices by those gatekeepers not being tackled. In the case of sub-option 1-B, there would be the risk of designating an excessive number of platforms as gatekeepers, including those that are not engaging in unfair practices but that, since they would be above the thresholds, would have to comply with the obligations.

335.Option 1 would fail to include in scope emerging gatekeepers whose position is likely to become entrenched in the near future as well as any new unfair practices by gatekeepers not part of the initial list of obligations. It would accordingly be a very static approach to deal with the dynamics of digital markets and, where potentially equally harmful gatekeeper behaviour would not be adequately addressed.

336.With respect to the objective of enhanced coherence and legal certainty, Option 1 allows an immediate aligning of platform-related rules across the EU through horizontal measures by relying on automatic quantitative criteria to identify gatekeepers and implementing immediately all obligations. These rules would preclude Member States from legislating in the areas covered by the new framework. Option 1 provides for effective and coherent EU-wide oversight through the establishment of a single regulator at the EU level, in cooperation with a network of national authorities. By basing the gatekeepers’ designation process on pure quantitative elements, this would provide a high degree of legal certainty through a clear signaling effect to the market. However, given its static nature, Option 1 would leave margin for some market fragmentation to remain. Under sub-option 1-A, only very large gatekeepers would be in scope. This could create some fragmentation resulting from the different treatment given to the largest gatekeepers (and thus captured by the identification process), as platforms exhibiting similar features and characteristics as the largest gatekeepers would not be subject to the obligations. Under sub-option 1-B, this possible difference in treatment between gatekeepers is less likely but there is a risk that some platforms which do not exhibit gatekeepers features and characteristics are forced to comply with the obligations, reducing their ability to compete with the real gatekeepers.

337.Option 2 would be effective in curtaining a wider range of unfair practices and increasing contestability in digital markets in a flexible way. For some of the practices a dialogue between the competent regulatory body and the gatekeepers concerned may be required to ensure that measures considered or implemented by the gatekeepers better achieve its goals. By introducing the possibility for such a dialogue, Option 2 can be expected to be more effective in addressing unfair practices hampering market contestability and competition. It will, at the same time, be proportionate for the gatekeepers concerned, since they would have certain margin of appreciation in implementing measures that effectively ensure compliance with the identified obligations. Therefore, it would be legitimate to expect that it would both create the right innovation incentives across the market, and contribute to increased consumer choice in terms of number of platforms proposing innovative and privacy-friendly services. By comprising a dynamic updating mechanism, Option 2 would also allow tackling new unfair practices. It would also allow tackling market failures related to gatekeepers that are expected to have an entrenched position in the near future. Option 2 is therefore also more effective in fulfilling the specific objective of addressing weak market contestability and competition than Option 1.

338.A drawback of Option 2 as compared to Option 1 is the fact that, by being based on market investigations to designate additional gatekeepers and by foreseeing a dialogue between the competent regulatory body and gatekeepers for some of the obligations, it could generate some delays in the implementation of those obligations and for those gatekeepers.

339.Similarly to Option 1, the sub-options in Option 2 would be subject to a trade-off. In the case of sub-option 2-A, the impact of the intervention would be less immediate given that a higher number of gatekeepers would have to be designated via a market investigation, and thus unfair practices by those gatekeepers would not be tackled for a period of time. In the case of sub-option 2-B, more gatekeepers would be automatically designated by means of the quantitative criteria but this would entail the risk of designating platforms that do not qualify as gatekeepers and that as a result of that type I error, these platforms would have a reduced capacity to compete with the gatekeepers.

340.As regards the objective of enhanced coherence and legal certainty, similarly to Option 1, Option 2 provides for effective and coherent EU-wide oversight through the establishment of a single regulator at the EU level, in cooperation with a network of national authorities (same regulatory design as under Option 1), thus contributing to legal certainty. The flexibility of tackling new unfair practices by gatekeepers and including gatekeepers that are expected to have an entrenched position in the near future could be expected to further reduce regulatory interventions at national level, thus contributing to the extent possible to preserving the digital single market in the online platform space. Given the combination of quantitative and qualitative criteria, type II errors would be less likely and thus gatekeepers would be treated in a more harmonised way across the EU than under Option 1. The possibility for the provider of core platform services to present, in exceptional circumstances, serious and substantiated arguments to demonstrate that it does not fulfil the objective requirements for a gatekeeper and should therefore not be designated directly based on the application of quantitative thresholds, but only subject to a further investigation, allows to address most of the concerns related to the possible lack of reliability and robustness of the quantitative thresholds set.

341.Option 2 would however imply some delay in enforcing the obligations for gatekeepers designated on the basis of the qualitative criteria, which could result in a temporary fragmentation of the market. Under Sub-option 2-A, these delays would be more relevant, as a higher number of gatekeepers would have to be designated via a market investigation, but the possible fragmentation associated to the unequal treatment given to gatekeepers discussed under Option 1 would be less likely.

342.Like under Option 1, to integrate national expertise in the platform economy, this option would also envisage that the Commission consults a ‘network of regulators’ before taking decisions. Option 2 would thus contribute to both addressing legal uncertainty (the problem identified) and to reducing fragmentation of regulatory approaches across the EU (the driver) in relation to a defined list of practices within a closed list of core services.

343.Option 3 provides for a fully flexible approach in achieving the specific objective of addressing gatekeepers’ unfair conduct and ensuring contestability of digital markets, including the possibility to include in scope new unfair practices by gatekeepers (as in Option 2) and additional digital services. 266 The inclusion of new digital services in scope is however not very likely as the core platform services listed in Section 5.2.1 are precisely the ones for which there is strong evidence that market failures are present and need to be addressed. In addition, it would always be possible under Options 1 and 2 to also include new digital services in scope during the review of the Regulation, which could possibly take place every three years.

344.Option 3 does however have the drawback of being slow in effectively addressing the problems compared to Options 1 and 2. In fact, given the need for a market investigation to designate all gatekeepers, and the possibility for a dialogue with the latter to determine the implementation of all the obligations, Option 3 would not allow an immediate response to the most pressing instances of gatekeeper related market failures. Given the importance of a timely regulatory response to the issues identified (explained above), this is an important drawback.

345.Regarding the objective of enhancing coherence and legal certainty, Option 3 provides for an effective and coherent EU-wide oversight through the establishment of a single regulator at the EU level, in cooperation with a network of national authorities. Option 3 would, on the one hand, add to Option 2 by tackling a broader scope of unfair practices by gatekeepers in additional digital services, thus reducing even more the need for intervention at national level and contributing to a more homogeneous approach to the issues at stake. However, the regulatory powers would be less circumscribed and may create a lower level of legal certainty. In addition, the need for the regulator (i) to conduct market investigations to designate all gatekeepers, and (ii) to engage in a dialogue with each gatekeeper for the implementation of all obligations would result in a staggered implementation of the measures. This may in turn create chilling effect and counteract the positive impact of the effectively addressed unfair behaviour on market contestability, which could also negatively affect innovation and consumer choice. Consequently, Option 3 would have a mitigated impact on the objective of ensuring market contestability and competition.

346.In light of the above, Option 2 – offering a reasonable trade-off between speed, legal certainty and flexibility, appears as meeting the overall general objective of improving the internal market functioning the most effectively since it allows for a more adaptive solution, which is a more appropriate way to tackle issues arising in the fast-changing platform environment.

347.It is essential to stress that the effectiveness of the measures (including their underlying remedies) under all options depend on their enforcement. Experience and evidence show that sanctions such as fines would not be sufficient to incentivise the gatekeepers targeted in this Impact Assessment to stop unfair practices they are engaging in. This is an important consideration to make when assessing the proportionality of the regulatory rules considered.

348.Table 5 below compares the effectiveness of the three options in achieving each specific objective pursued taking into consideration their respective speed of intervention, flexibility and legal certainty.

Table 4: Comparison of options in terms of effectiveness

Efficiency

349.The efficiency comparison is based on a benefit estimate for the preferred option and on a cost comparison between the options (both detailed in Annex 3 to the Impact Assessment). A robust quantitative cost-benefit comparison between the different options proves however difficult for the initiative under consideration given the absence of reliable data, and thus estimates of costs and benefits are only provided for the preferred option, which serves as a reference for the magnitude of the remaining options. The approach taken for estimating costs and benefits under each of the options is as follows: assumptions in relation to costs are overestimated in order to ensure that costs stemming from the measures assessed are not minimized; assumptions in relation to benefits are conservative to guarantee that benefits are not overestimated.

350.All options imply (a) regulatory costs at EU and national levels, i.e. for the Commission as the competent regulatory body at EU level to ensure implementation, supervision and information gathering, and for the network of national regulators to respond to eventual consultations from the Commission; (b) compliance costs for gatekeepers to deal with new rules and to respond to the regulator’s requests for information; and (c) (minimal) costs for the business users and platforms not subject to the obligations but which may at times need to respond to the regulator’s requests for information. 267

351.It needs to be noted that while all obligations would be legally applicable to all designated gatekeepers, not all obligations would be relevant for each gatekeeper, since not every gatekeeper would be engaging in all unfair practices targeted by the initiative. If a gatekeeper were engaging in one or more of the unfair practices, this would require changes in its behaviour but would not necessarily translate in direct costs. Compliance costs have been estimated at EUR 1.41 million per year and per platform. This regulatory burden should be weighed in light of the economic power of gatekeepers in scope and against the fact that they would have already existing internal services to comply with other pieces of EU legislation (e.g. EU Merger Regulation; Consumer protection cooperation (‘CPC’) Regulation). This possible synergy in terms of compliance would further reduce the impact of additional costs which is marginal as compared to the enormous revenues earned by gatekeepers. In addition, the regulatory dialogue foreseen under Options 2 and 3 for obligations requiring further specification would reduce the burden on gatekeepers since it would allow tailoring the specific obligation to the particular situation of the gatekeeper concerned, which can be expected to reduce the overall compliance cost. Benefits stemming from the initiative would amount to billions (see following paragraphs) and lead to greater innovation potential amongst smaller businesses as well as improved quality of service, with associated increases in consumer welfare (as specified below).

352.Under Option 1, five to seven platforms would be covered under sub-option 1-A and 10 to 15 platforms under sub-option 1-B, which implies an overall annual cost for platforms’ compliance with rules ranging between EUR 9.87 million and EUR 21.15 million. This calculation is based on the estimate of the compliance costs per platform, i.e. EUR 1.41 million per year. Based on benchmarks of similar practices within the Commission, networks and national authorities, enforcement cost for the Commission can be estimated at between EUR 6.4 million (under sub-option 1-A) and EUR 10.5 million (under sub-option 1-B) while for national authorities, these are estimated at EUR 4.3 million per year when summing the costs of all 27 Member States.

353.The estimates on the number of platforms under Option 2 are based on screening of the quantitative criteria and an assumption on the number of additional gatekeepers designated via a market investigation. In relation to the latter, this is very difficult to estimate upfront given that only after a market investigation it would be possible to determine whether a given provider of core platform services meets the criteria. In this context, and for the purpose of these calculations, it is assumed that Option 2 would cover up to a maximum of between 15 gatekeepers (sub-option 1-A) and 20 gatekeepers (sub-option 1-B). 268 This implies an overall compliance cost ranging between EUR 21.15 million and EUR 28.2 million per year. The administrative costs for the EU Commission are estimated at EUR 16.7 million per year. Costs for national authorities (for all 27 Member States) are estimated at EUR 6 million per year.

354.While a precise assessment of the number of gatekeepers targeted under Option 3 would be misleading (since their designation would be based on qualitative criteria only), it is assumed to be 25 (as an upper bond) for the purpose of quantification and options’ comparison. The number of gatekeepers designated under Option 3 could theoretically be higher than the number of gatekeepers designated under Option 2, given that it could also include gatekeepers active in core platform services that could be added after a market investigation. The above upper bond assumption has been made in order to provide a cost estimate in the worst case scenario, i.e. reflecting the highest possible costs that the measures could generate under this option. Under this assumption, Option 3 would imply a compliance cost per year for gatekeepers of around EUR 35.25 million. The administrative costs for the EU Commission are estimated at EUR 18.2 million per year, and those for national authorities (for all 27 Member States), at EUR 6 million per year.

355.In terms of benefits, while a quantified comparison of the different options 269 proves difficult to establish, it is an objective qualitative assessment to consider that the impact of putting in place an effective and proportionate regulation addressing dysfunctions in the platform economy in an effective and flexible way (as foreseen under Option 2) would lead to a different profit distribution, i.e. with a greater societal benefit. Concerns about excessive gatekeeper profits that could be extracted based on their grossly imbalanced bargaining power would be effectively addressed and any such profits would be distributed to business users and consumers; hence, the more appropriate (effective but also proportionate) the regulatory measures, the more optimal the re-distribution of profit.

356.If we assume that the measures foreseen under Option 2 contribute to preserving the internal market in the platform space - thus allowing cross-border trade projections by 2025 to be maintained - this would lead to EUR 92.8 billion benefits. 270 The benefits can be expected to lead to greater innovation potential amongst smaller businesses as well as improved quality of service, with associated increases in consumer welfare. Assuming that interventions foreseen would reduce competitive asymmetries between gatekeepers and other platforms, a consumer surplus of the preferred option could be estimated to increase by EUR 13 billion, i.e. around 6% increase as compared to the baseline. 271 Impact on economic growth is estimated to range between EUR 12 billion and EUR 23 billion. 272 Benefits would be similar under sub-options 2-A and 2-B, with the former having the advantage of not incurring in a type II error, and the latter having the advantage of implementing the obligations on gatekeepers quicker than if they would only be captured after a market investigation under sub-option 2-A.

357.The benefits under Option 1 would be lower, they would cover a more limited set of market failures given that it would not be possible to add new unfair practices to the scope neither to tackle market failures in tipping markets. The benefits stemming from Option 3 could potentially be higher than under Option 2 given that additional services could be added to the scope of the obligations. However, as explained in Section 7.1, the inclusion of additional services in scope is not very likely to take place given that the core platform services listed in Section 5.2.1 are precisely the ones for which there is strong evidence that market failures are present. 273 The potentially higher benefits under Option 3 could therefore not materialise. Furthermore, Option 3 would also originate a lower level of legal certainty, which would result in a chilling effect and counteract the positive impact sought on market contestability.

Coherence

358.An assessment was carried out of the various policy options’ coherence with (i) the Commission’s digital strategy, (ii) the DSA and (iii) other regulatory instruments.

Coherence with the Digital Strategy

359.All three options are coherent with the Commission’s digital strategy in their contribution to ensuring a fair and competitive digital economy, one of the three main pillars of the policy orientation and objectives announced in the Communication Shaping Europe's digital future. They would constitute a coherent, effective and proportionate framework to address problems in the digital economy that currently cannot be tackled or cannot be tackled effectively.

Coherence with the DSA

360.All three options are coherent with and complementary to the proposal for the update of the e-Commerce Directive (‘ECD’) under the DSA. While the DSA is a horizontal initiative focusing on issues such as liability of online intermediaries for third party content, safety of users online or asymmetric due diligence obligations for different providers of information society services depending on the nature of the societal risks such services represent, the present options are concerned with economic imbalances, unfair business practices by gatekeepers and their negative consequences, such as weakened contestability of platform markets. To the extent that the DSA contemplates an asymmetric approach which may impose stronger due diligence obligations on very large platforms, consistency will be ensured in defining the relevant criteria, while taking into account the different objectives of the initiatives.

Coherence with other instruments

361.All three options align with other EU instruments, including with the EU Charter of Fundamental Rights and the European Convention on Human Rights (‘ECHR’), the GDPR, the EU’s consumer law acquis and the P2B regulation.

362.The definitions to be used under all options are coherent with the definitions used in EU existing legislation, in particular the definitions of ‘online intermediation services’ and ‘online search engines’ used in the P2B Regulation. With their scope targeted to gatekeepers, the options complement well the horizontal obligations for all online platforms under the P2B Regulation. All options also complement existing EU competition law by addressing ex ante unfair practices by gatekeepers that either fall outside the existing EU competition rules, or cannot be addressed in the most effective manner by these rules.

363.All options complement the data protection laws. Transparency obligations on deep consumer profiling will actually help inform GDPR enforcement, whereas mandatory opt-out for data combination across core platform services goes beyond GDPR protections. Anti-circumvention clauses will clarify that compliance with obligations in this initiative may require consent under GDPR. This is also without prejudice to data minimisation principle, including using anonymised data where possible. The introduction of the dynamic updating of core platform services and gatekeepers’ practices would be subject to full respect of the fundamental rights to fair proceedings and good administration as enshrined in the ECHR, which are binding on the EU institutions. Given that the mechanism for the imposition of remedies is administrative in nature and not criminal or quasi-criminal, the fundamental rights of the Charter enjoyed in the case of criminal proceedings would not apply. 274 However, when acting under the new framework in general and market investigation regime in particular, the Commission’s investigation powers would be counterbalanced by ensuring that undertakings involved enjoy effective fair process rights such as the right to be heard, the right to a reasoned decision and access to judicial review, including the possibility to challenge enforcement measures. These rights apply in case of administrative proceedings. 275 This design to preserve fundamental rights is also consistent with – if not superior to – the safeguards applicable similar investigation regimes elsewhere in the world. 276

364.All options leverage existing platform regulation, without conflicting with it, while providing for an effective and proportionate enforcement mechanism that matches the need to strictly enforce the targeted obligations vis-a-vis a limited number of very large cross-border providers.

365.Different from the P2B Regulation 277 , all options foresee EU-level enforcement of a narrow set of very precise unfair practices engaged in by a restricted group of large, cross-border gatekeepers. This EU-level enforcement mechanism is consistent with the enforcement of the P2B Regulation. Gatekeepers are likely to exist in respect of several cross-border core platform services, and a central EU-level regulator with strong investigatory powers is required both to prevent fragmented outcomes as well as to prevent circumvention of the new rules. To this end, the new EU-level regulator can leverage the transparency that each of the online intermediation services and online search engines have to provide under the P2B Regulation on practices that could precisely be illegal under the list of obligations – if engaged in by gatekeepers.

366.All three options would - while recognising the differences – align with the experiences from the targeted and tailor-made ex ante regulation of specific sectors, including the rules applicable to electronic communication services or short-selling. All options would be coherent with existing initiatives targeting harmful trading practices in the offline world. These existing initiatives are designed to tackle practices relevant to and specific for the offline sector or context in which they arise. 278 They do not overlap with the unfair practices described under Section 2.1.2 which, together with the different business models by which gatekeepers operate, are very different and warrant separate treatment.

Proportionality

367.Option 1 would be targeted to gatekeepers above a pure quantitative threshold. This would create legal certainty for gatekeepers. The obligations it would cover would be identified based on objective criteria supported by the evidence gathered. However, the absence of dialogue between the regulator and gatekeepers would preclude the possibility to exchange with the regulator to specify the application of certain obligations.

368.Option 1-A would leave uncovered some gatekeepers that would be similar to the designated gatekeepers, in the sense that they may equally have an important internal market impact, operate an important gateway to end users and have an entrenched position, but which are relatively smaller. This would call for a potential intervention by the Member States or competition law to tackle those. This would thus leave more room for action at national level but at the same time raise a concern of further regulatory fragmentation. Option 1-B would cover a larger number of platforms, thus reducing the risk of regulatory fragmentation as the obligations are applied to a larger set of platforms. However, it would risk capturing under scope providers of core platform services that would be above the thresholds but that do not present features and characteristics similar to gatekeepers. There is thus a risk of disproportionality in relation to the application of the obligations to those platforms.

369.Option 2 allows achieving the objectives effectively since it sets a comprehensive ex ante framework providing for a list of obligations that are clearly identified and circumscribed. Option 2-A would also result in a uniform application of obligations to designated gatekeepers (including those that are expected to enjoy an entrenched position in the near future) that although not falling under the high quantitative threshold criteria would nonetheless operate core platform services that exhibit similar features and characteristics to those of gatekeepers corresponding to the quantitative criteria. The main drawback of sub-option 2-A would be the possible temporary fragmentation resulting from having gatekeepers automatically designated and others designated via longer market investigations. Sub-option 2-B would, similarly to the case of sub-option 1-B, provide for a harmonised application of the obligations given that a larger set of platforms would be in scope, but still risking to capture providers of core platform services that would be above the thresholds but do not present features and characteristics similar to gatekeepers. In both cases compliance costs for gatekeepers are reasonable thus allowing to safeguard the benefits they create for the internal market.

370.Option 2 is proportionate since it would be also addressing the wider possible range of unfair practices at EU-wide level identified on the basis of similar criteria as Option 1, while at the same time providing for a regulatory dialogue in relation to the application of some of the obligations, where necessary and justified. Similarly to Option 1, Option 2 foresees cooperation with NCAs and with sectorial bodies.

371.Option 3 leaves strong discretionary power to the regulator in terms of both designation of gatekeepers (based on qualitative criteria only assessment) and scope of intervention (given the flexibility left to the regulator to include additional digital services and practices in the scope). Option 3 would from that perspective lead to a lower level of legal certainty, which is essential for a thriving business environment. Option 3 allows however for extensive regulatory dialogues, which possibly makes it less burdensome than regulatory measures not allowing for such dialogues. At the same time the longer procedures which would stem from the market investigation nature of this option limit its potential for addressing the problems identified in a timely manner.

372.For all options, and as explained in Section 5.2.1, the core platform services in scope are only those where there is strong evidence of (i) high concentration, where usually one or very few large digital platforms set the commercial conditions with considerable autonomy from their competitors, customers or consumers; (ii) few large digital platforms acting as gateways for business users to reach their customers and vice-versa; and (iii) gatekeeper power often misused by means of unfair behaviour vis-à-vis economically dependent business users and customers. Option 3 provides for the possibility of other digital services being added to the list after a market investigation and based on an empowerment given to the Commission. This would allow to cover in a flexible way all the digital services where there is weak contestability and gatekeepers engage in unfair behaviour, but would create lower legal certainty.

373.As explained in Section 5.2.1, the gatekeepers in scope under Option 1 would only be those that meet the quantitative criteria that serve as a proxy for the features characterising gatekeeper status. This could risk including too few providers of core platform services in scope if thresholds are set at a too high level (false negatives) or too many platforms in case lower thresholds are fixed (false positives). Under Options 2 and 3, gatekeepers in scope are more likely to correspond to those which (i) have a significant impact on the internal market; (ii) operate a core platform service which serves as an important gateway for business users to customers; and (iii) enjoy an entrenched and durable position in their operations or are expected to enjoy such a position in the near future.

374.Finally, and as explained in Section 5.2.2, the list of obligations foreseen under all options is justified as the obligations have been limited to those practices that (i) are of egregious nature, (ii) can be identified in a clear and unambiguous manner to provide the necessary legal certainty for gatekeepers, and (iii) for which there is sufficient experience with the harmful effects. Option 2 and 3 provide for the possibility of a regular dialogue between the Commission and the gatekeepers concerned, as well as for the addition of other practices if deemed unfair following a market investigation and based on an empowerment by the Commission.

375.In order to ensure the effectiveness of the tool, all options could include a series of remedies to ensure that designated gatekeepers comply with the obligations. This would include initially fines and penalty payments in case gatekeepers do not comply with the obligations. As a last resort, and in case of systematic failure to comply with the obligations, even after the imposition of fines and penalty payments, other types of measures could be envisaged under specific conditions and circumstances (see Section 5.2.4). These would however be applied in extreme cases, i.e. once all other means to ensure fair behaviour have proven insufficient, and thus not put at risk the proportionality of the measures, under all three options.

Subsidiarity

376.All three options respect the subsidiarity principle. The intrinsic cross-border nature of the digital economy and of the provision of core platform services provided by gatekeepers, suggests that the objectives pursued cannot be effectively reached by Member States alone. Rather to the contrary, as shown in Annex 5.4 to the Impact Assessment, regulatory initiatives by Member States lead to divergent regulatory solutions and regulatory fragmentation. EU action would avoid further fragmentation of the single market into different, potentially contradictory frameworks – including the resulting jurisdictional issues. This is expected to decrease gatekeepers’ incentives to develop unfair practices in relation to new core platform services or expand further unfair behaviour related to existing practices.

377.Furthermore, while all three options foresee enforcement and strong coordination at EU level, they also envisage the involvement of national authorities in the decision making process to ensure that Member States’ expertise is taken into account. The new ex ante framework would harmonise rules in the areas targeted by these rules, without prejudice to the right of Member States to consider further measures in order to improve contestability of markets or to fight against acts of unfair competition that are unrelated to the presence of gatekeepers within the meaning of the new framework and where such obligations would be compatible with EU law.

Conclusion

Table 5: Policy options - comparison

378.Effectiveness. Compared to Option 1, Options 2 and 3 would allow tackling new unfair practices and market failures related to gatekeepers that are expected to have an entrenched position in the near future. Option 1 would be particularly effective in quickly dealing with the market failures in digital markets. The combination of immediately applicable obligations and of the possibility for a regulatory dialogue with gatekeepers for some of the obligations would make Option 2 more effective than Options 1 and 3. The latter two would be: too static (in the case of Option 1) or too flexible (in the case of Option 3). Option 1-A could give rise to type II errors (false negatives) while Option 1-B could result in type I errors (false positives). Under Option 2-A, the risk of false negatives would be minimised by the possibility of designating gatekeepers also on the basis of qualitative criteria. In that case, the drawback in comparison with sub-option 2-B would be the delays associated to the need of conducting market investigations to designate some of the gatekeepers. 279 Option 3 is future-proof and would reduce the risk of type I or type II errors by basing designation on a pure qualitative test. At the same time, it does not allow addressing the problems in a timely manner, implies less legal certainty than the other two options and would result in a staggered approach to the market.

379.Efficiency. Compared to Option 1, Options 2 and 3 would generate higher benefits because they would allow tackling a higher number of market failures. Option 3 could in theory target even more market failures given the possibility to add other digital services to the scope of the regulation, following a market investigation. This is though unlikely to take place. However, as a consequence of discretionary power given to the regulator, Option 3 would also create a lower level of legal certainty which would result in a chilling effect, thus counteracting the positive impact (that measures effectively addressing unfair behaviour would have) on market contestability. As regards sub-options A and B, magnitude of benefits would be similar, with the former having the advantage of not incurring in type II errors and the latter of implementing quicker the obligations to gatekeepers (that would otherwise be captured after a market investigation under sub-option 2-A). In relation to compliance costs, they would be miniscule as compared to the profits of gatekeepers. They would also be insignificant as compared to the range of benefits resulting from tackling unfair practices in digital markets. Nevertheless, Option 1 would result in lower compliance costs as gatekeepers would not be subject to any market investigation and less practices would be tackled. Option 3 would imply the highest compliance costs given the need to conduct a market investigation for any designation and the possibility to designate gatekeepers from services other than the current core services. As for sub-options 2-A and 2-B, the former would result in lower compliance costs since it would imply designating a lower number of gatekeepers, although the higher number of market investigations associated to sub-option 2-A could imply higher costs of replying to requests for information.

380.Coherence: All three options are coherent with other EU instruments and EU international commitments.

381.Proportionality. Option 1 would be too static, in the designation process and the implementation of the obligations. This could originate a disproportional implementation of obligations. Options 2 and 3, by allowing a more flexible approach would be more proportionate. Sub-option 2-A could lead to temporary fragmentation resulting from some gatekeepers being automatically designated while others via longer market investigations; under sub-option 2-B there would be the risk of capturing platforms that do not present features and characteristics similar to gatekeepers’ ones. Option 3 would lead to a lower level of legal certainty which is essential for a thriving business environment.

382.All three options respect the subsidiarity principle.

383.In light of the above, political choice is needed on whether Option 2-A or Option 2-B would better address the issues at stake and achieve the policy objectives pursued. Option 2 allows for timely intervention for the most egregious practices and more gradual approach for measures needing further tailoring and specification. The comparison between Options 2-A and 2-B leads to a trade-off to be made between the speed of the regulatory intervention and the scope of the target population. As specified above, in the case of sub-option 2-A, the impact of the intervention would be less immediate (given that a higher number of gatekeepers would have to be designated via a market investigation), and thus the unfair practices by those gatekeepers would not be tackled for a period of time. In the case of sub-option 2-B, more gatekeepers would be automatically designated but with the risk of designating platforms that do not qualify as gatekeepers. Both sub-options (i) address unfair behaviour, including new unfair practices and tipping markets, leading to positive impact on market contestability, innovation and consumer choice, (ii) address most of the regulatory fragmentation problems as obligations are applied to both gatekeepers designated on the basis of quantitative and qualitative criteria, including emerging gatekeepers, and (iii) create legal certainty for those.

8.PREFERRED OPTION

The main building blocks of the preferred option

384.The core substantive elements of the preferred option are explained in Section 5.3.2. The preferred option provides for a new ex ante regulatory framework built around the following elements:

(a)First, the ex ante framework would only apply to clearly identified and closed list of core platform services (i.e. a numerous clausus of core platform services) that are most broadly used by business users and end users and where, based on the evidence collected and presented in this Impact Assessment, more apparent and urgent concerns about weak contestability and unfair practices by gatekeepers arise;

(b)Second, only providers of core platform services that meet the specific conditions analysed could be designated as gatekeepers. Such designation would take place by applying mix of quantitative and qualitative thresholds;

(c)Third, designated gatekeepers would be required to comply with the set of clearly defined obligations in order to address a negative impact of unfair practices discussed in Section 5.2.2 on fairness in commercial relationship between these gatekeepers and their business users and contestability of platform markets. Such obligations would encompass (i) immediately applicable obligations and (ii) obligations where a degree of appreciation would be required in view of the implementation of a given obligation.

The scope of application

Identification of core platform services

385.The preferred option foresees up-front a set of clearly identified core platform services that feature number of specific characteristics discussed in Section 5.2.1.

386.The analysis underpinning the Impact Assessment shows that there are number of core platform services that meet these characteristics, notably:

(a)Online intermediation services, such as online marketplaces and software application stores which enable business users to reach and contact end users, to provide or offer services or products to the latter. They can become a key access point for business users to reach end users.

(b)Online search engines that can significantly affect the commercial success of business users and therefore unfair practices carried out by a gatekeeper providing such online search engine services have the capacity to affect a large number of end users and businesses alike.

(c)Operating systems are at the heart of devices and ecosystems and are characterised in particular by economies of scale and high switching costs, and benefit from network effects. A gatekeeper can use its control over the operating system to engage in unfair practices limiting the contestability of the services concerned.

(d)Online social networking services are characterised in particular by strong network effects, data driven advantages and high switching costs. When controlled by a gatekeeper, online social networking services represent an important gateway not only for end users but also increasingly for business users.

(e)Video-sharing platform services that can become the default or at least a preeminent platform to consume and to share video content online. When operated by a gatekeeper, they are thus a very important access point for video content providers and offer significant audiences for advertisers.

(f)Number-independent interpersonal communication services are services for which network effects are particularly strong when they are run by a gatekeeper, the risk of unfair business practices and a lack of contestability is particularly strong.

(g)Cloud computing services provide infrastructure to support and enable functionality in digital services offered by others and at the same time offer a range of products. The vertical integration by a gatekeeper of a large cloud computing services provider can lead to unfair business conditions, for instance unjustified limitations to interoperability and data portability.

(h)Online advertising services are often related to other core platform services, such as online search engines and online social networking services. Gatekeepers operating the latter types of core platform services very often also provide online advertising services and may engage in unfair practices, which goes in particular to the detriment of their business users that is, advertisers and publishers and also further limit contestability.

Designation of gatekeepers

387.The addressees of the preferred option would be those providers of core platform services that meet the following conditions: (i) have a significant impact on the single market, (ii) operate one or more important gateways for business users to reach end users, and (iii) they enjoy or are expected to enjoy an entrenched and durable position in their operations.

388.Under the preferred option, these providers of core platform services would be designated based on combined application of quantitative and qualitative criteria. Where a provider of core platform services would meet cumulatively a set of quantitative thresholds established in the regulation it would be automatically designated as a gatekeeper by the Commission. This allows a fast and effective protection of the interest of all the business users affected by the gatekeepers’ unfair behaviour. Two specific combinations of parameters, for a low and a high threshold respectively, have been selected for the purpose of providing clarity in assessing impacts and trade-offs for the options’ comparison. As explained in Section 5.2.1, other plausible policy options exist as to the use of the economic parameters or their possible combinations. These thresholds act as quantifiable proxies of the qualitative criteria in terms of size and reach of the business, number of business users and duration in time of the market position. Such a quantitative threshold may be set at a high level (sub-option 2-A) or low level (sub-option 2-B). In choosing between sub-options 2-A or 2-B it is preferable that these thresholds should be set at a sufficiently high level. This is to ensure that only very large systemic players with a significant internal market presence and which are gateways to a large number of end users, clearly holding an entrenched and durable position, should be deemed to be a gatekeeper on the basis of quantitative criteria. Such undertakings should be subject to a fast designation process which is limited to verifying whether the quantitative criteria are met. 

389.While considering the high probative value of the considered quantitative threshold, it cannot be completely excluded that in very exceptional circumstances a provider of core platform services that meets these quantitative thresholds nonetheless does not act as a gateway for its business users and end users. To ensure necessary proportionality in such exceptional circumstances, the provider of core platform services should have the opportunity to present serious and substantiated arguments in order to demonstrate that, in the circumstances in which the relevant core platform service operates, and taking into account other relevant elements 280 , the provider does not meet the conditions discussed in paragraph 387. It is important however to note that the purpose of the possibility to rebut the legal presumption is not to demonstrate, on pure economic grounds, efficiencies deriving from a specific type of behaviour by the provider of core platform services since this is not relevant to designation of such a provider as a gatekeeper.

390.Furthermore, even if a provider of core platform services does not meet the quantitative thresholds that does not in itself mean that it may not constitute a gatekeeper. In fact, the preferred option would envisage a possibility to designate the provider of core platform services as a gatekeeper following a market investigation, which would have to show that the provider of core platform services meets the conditions discussed in paragraph 139.

391.This would also ensure the regulatory playing field and regulatory symmetry of the obligations laid down in the rules. This also represents the optimal trade-off in terms of necessity to assess the market conditions and time that would take to designate gatekeepers and thereby time within which the problems identified would be effectively addressed. In this respect it is to be noted that quantitative thresholds set at a high level would take comparatively more time and resources than a more straightforward designation based on lower quantitative thresholds. Conversely, while relatively low quantitative threshold level would allow immediately capturing the majority of gatekeepers, it would entail the risk of extending disproportionately gatekeeping obligations to a large number of platforms.

392.The designation of a provider of core platform services as a gatekeeper following a mechanism that combines quantitative and qualitative indicators features strong support by stakeholders and is considered as an appropriate mix of ensuring flexibility, speed and legal certainty.

393.Under the preferred option, the qualitative criteria would allow designating not only providers that are already enjoying an entrenched and durable position in their operations, but also those for which this is not yet the case, but which are rapidly acquiring market strength and building towards becoming gateway due to specific market features and their capacity to put competitors at a disadvantage in the market, i.e. emerging gatekeepers.

394.The preferred option would provide for regular review of the gatekeepers status, a possibility which seems particularly important in such a dynamic market environment. Such a regular review of the gatekeeper status would in principle have to be carried out on regular intervals of two years. 

395.The designation decision addressed to providers of core platform services that meet the conditions would be subject to judicial review and would, beyond the regular review, also foresee reassessment at the request of the affected firm in case of material changes concerning the designation conditions.

Obligations applicable to gatekeepers’ core platform services

396.Under the preferred option, once a provider of core platform services is designated as a gatekeeper, all its core platform services that individually meet the conditions of being an important gateway for business users to reach end users would have to comply with a clearly defined set of obligations relating to a clearly identified set of unfair practices.

397.The obligations under the preferred option would address the unfair practices by gatekeepers that weaken market contestability (Section 2.1.1) and undermine the fairness of commercial relationship of gatekeepers towards their business users or in some cases towards third parties (Section 2.1.2).

398.The obligations under the preferred option would be either immediately applicable or would in certain cases envisage the possibility of a regulatory dialogue between the Commission and the gatekeeper concerned in view of ensuring that the measures gatekeepers intend to implement ensure effective compliance with the obligations. The set of obligations that would be included in the preferred option is explained in details in Section 5.2.2.

399.The distinction between immediately applicable obligations and obligations subject to dialogue is based on the analysis of the measures in question. The obligations relative to transparency and non-discrimination are self-evident. Obligations which require evaluation of interoperability conditions or customisation considering the specific nature of the core platform service offered are subject to a dialogue. Such a dialogue can be launched by the Commission, either upon request of the gatekeeper concerned, or where the Commission finds on its own initiative that any measures that the gatekeeper has already implemented or still intends to implement are likely to fall short of what is required to ensure compliance with the obligations concerned. This possibility of a regulatory dialogue should facilitate compliance by gatekeepers and allow them to signal any circumstances.

400.As an additional element to ensure proportionality, gatekeepers should be given an opportunity to request the suspension of a specific obligation in exceptional circumstances that lie beyond the control of the gatekeeper. Where compliance with a specific obligation is shown by the gatekeeper to endanger the broader economic viability of the EU operations of the gatekeeper concerned, for example because an unforeseen external shock has temporarily eliminated a significant part of end user demand for the relevant core platform service, it would ultimately harm innovation and welfare if that core platform service were unable to continue its operations once the exceptional circumstances would cease to apply. Similarly, in exceptional circumstances solely justified on the limited grounds of public morality, public health or public security, and based on a reasoned request by the gatekeeper, the Commission could decide that the obligation concerned does not apply to a specific core platform service.

401.The combination of a regulatory dialogue to facilitate compliance with limited exemption possibilities will ensure the proportionality of the obligations without undermining the intended ex ante effects on fairness and contestability.

Enforcement framework and governance

402.Under the preferred option implementation, supervision and enforcement would be carried out at the EU level by the Commission as the competent regulatory body.

403.The preferred option is built on the clearly identified behavioural measures, which will be laid down in the new rules. These rules will be based on a legal presumption that, to meet the objectives of safeguarding contestability of core platform services and fairness of their commercial relationships, gatekeepers need to comply with such regulatory behavioural measures, i.e. obligations. This will ensure the necessary legal certainty and predictability of the rules as well as ensure that the rules apply only where this is necessary and justified, i.e. are proportionate to the objective sought.

404.The preferred option would also lay down adequate and proportionate enforcement powers of the Commission with clearly defined procedural enforcement framework and clearly set deadlines that the Commission would need to respect. The enforcement powers and processes applicable would be unique to the preferred option. However, as explained in Sections 5.3.1.5 and 5.3.2.5 they would nevertheless be able to learn in some of their aspects from existing regulatory and competition law powers. In addition, the procedural framework would also lay down clear rules on redress available to gatekeepers or other concerned parties, including access to judicial remedies.

405.The ex ante rules under the preferred option will be complemented by the possibility for the Commission to launch a market investigation in a limited and well identified number of cases:

(a)to designate gatekeepers that meet the conditions laid down in Section 5.2.1, or may meet them in near future;

(b)to update the list of unfair practices and corresponding obligations laid down in the rules; and

(c)to serve as a basis for further remedial action if the behavioural measures clearly prescribed by the rules are systematically infringed by the designated gatekeepers.

406.It is worth recalling that the possibility of updating the core platform services by means of empowerment by the Commission following a market investigation has been excluded from the preferred option. The reason is linked to legal limits to the market investigation powers under the chosen ex ante legal instrument that cannot include implicit powers to adapt the scope. Instead, the Commission should propose the necessary legislative adaptations by including in the evaluation of the effectiveness of the regulation the regular review of the list of core platform services in view of ensuring that digital markets across the EU are contestable and fair.

407.Under the preferred option, further remedies (see Section 5.2.4) would be envisaged for the purpose of ensuring effective remedies against systematic non-compliance.

408.Such remedies should be modelled on the well-established precedent of Regulation 1/2003, and offer a graduated, step-by-step process of increasing sanctions, with due process rights at each step. This enforcement framework reflects the potentially limited dissuasive power of monetary fines alone, and may, as explained in Section 5.2.4., contain as ultima ratio behavioural or structural remedies after all other avenues have been exhausted, noting that even in established legislation such a tool has never been used.

409.In the preferred option the Commission will ensure close cooperation with and between the competent independent authorities of the Member States, with a view to informing its implementation and to building out the Union´s expertise in tackling fairness and contestability issues in the digital sector. In this context, the Commission will establish an information exchange and consultation network consisting of relevant independent authorities of the Member States, which shall also deliver opinions on the individual decisions of the Commission.

410.As regards the possible integration of the new Commission powers under the Digital Markets Act and responsibilities of the Board envisaged under the DSA it is important to note the very different objectives of the two sets of rules and corresponding expertise and competences that may be required from the competent enforcement bodies to ensure compliance with the respective rules.

411.The Board under the DSA, including the participation of the national Digital Services Coordinators, enhances the cooperation system, particularly necessary for ensuring the supervised risk management approach for regulating the due diligence of very large platforms. This system ensures in particular that primarily systemic societal, and not economic, concerns brought by those platforms with an EU-wide impact are appropriately addressed through cooperation at the EU level supported by the activities of the Board, thereby ensuring sufficient expertise and appropriate competencies. However, the main regulatory compliance activities continue to be carried out by the competent national Digital Services Coordinators.

412.Furthermore, contrary to the decentralised approach under the DSA, where the focus of regulatory compliance activities is on the competent national Digital Services Coordinators, the implementation and enforcement of harmonised rules under the DMA is to be ensured at the EU level by the Commission who has the necessary means and expertise, without any decentralised competences.

413.In view of this, it could not be considered under the preferred option that any of the investigation and enforcement competences and powers could effectively be carried out by the Board whose tasks relate to facilitating implementation, cooperation and enforcement of very different rules as those envisaged by the DMA.

9.HOW WILL ACTUAL IMPACTS BE MONITORED AND EVALUATED?

414.Given the dynamic nature of online platforms, monitoring and evaluation of impacts needs to constitute an important part of the proposal. It also responds to explicit demands by stakeholders, including Member States (e.g. France), for a dedicated monitoring function, and reflects the self-standing monitoring option considered in the Inception Impact Assessment. The monitoring therefore will be divided into two parts: (i) continuous monitoring which will report on the latest developments in the market every second year potentially involving the EU Observatory of the Online Platform Economy, and (ii) operational objectives and specific indicators to measure them.

415.Regular and continuous monitoring will cover the following main aspects:

a)Monitoring scope-related issues (e.g. indicators for the designation of gatekeepers, range of designated gatekeepers and its evolution, use of the margin of appreciation in the designation);

b)Monitoring unfair practices (compliance, enforcement patterns, evolution); and 

c)Monitoring as a trigger for launch of a market investigation.

416.The following indicators would be potentially used:

Table 6: Measuring indicators

417.The monitoring will also take due account of the conceptual work of the Expert Group of the Online Platform economy under its work stream on Measurement and Economic Indicators. 281

418.Of particular importance in the monitoring framework is the evolution of the market, in terms of new unfair practices, additional core platform services, and new gatekeepers. The monitoring framework, including through the dedicated Observatory of the Expert Group on the Online Platform Economy as well as through the market investigation part of the preferred option, will continuously monitor the evolution of these factors. This is a core part of keeping the regulation future proof.

419.To this end, the preferred option contains a specific obligation on the Commission to review whether designated gatekeepers continue to meet the scope of the obligation. This monitoring function is essentially part of the preferred option to keep the rules in line with market developments.

420.Furthermore, specifically the legislation proposed should be reviewed at least every three years, to ensure that other elements, notably other scope related issues (such as new services) require adjustments.

421.Finally, the monitoring framework also needs to monitor compliance with the regulation, and the effectiveness of the enforcement framework, including to which extent the range of available remedies were actually used, the effectiveness of the implementation dialogues, and the responsiveness of the companies in scope to the obligations.

(1)

      Inception Impact Assessment for the Digital Services Act package .

(2)

      Inception Impact Assessment of the New Competition Tool .

(3)

      Open Public Consultation on Single Market - new complementary tool to strengthen competition enforcement and Open Public Consultation on Digital Services Act package – ex ante regulatory instrument of very large online platforms acting as gatekeepers .

(4)

      Inception Impact Assessment of the New Competition Tool , at page 3; and Inception Impact Assessment for the Digital Services Act package , at page 4.

(5)

     J. Crémer, Y.-A. de Montjoye & H. Schweitzer (2018), Digital policy for the digital era , page 70.

(6)

      EU Observatory on the Online Platform Economy .

(7)

         US House of Representatives Majority Staff Report, Investigation of Competition in Digital Markets , October 2020.

(8)

         ACCC report, Digital Platforms Inquiry, Final Report , June 2019.

(9)

     Furman report,  Unlocking digital competition, Report of the Digital Competition Expert Panel , March 2019.

(10)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(11)

      Mission letter to Executive Vice-President Vestager , 10 September 2019.

(12)

     European Commission Communication, Shaping Europe’s digital future , 19 February 2020, at page 5.

(13)

     Ibid, at page 5.

(14)

     Pursuant to Article 225 of the Treaty on the Functioning of the European Union (‘TFEU’).

(15)

      https://emeeting.europarl.europa.eu/emeeting/committee/en/agenda/202009/IMCO .

(16)

     In its draft report, the IMCO considered that “by reducing barriers to market entry and by regulating large platforms, an internal market instrument imposing ex-ante regulatory remedies on these large platforms has the potential to open up markets to new entrants, including SMEs and start-ups, thereby promoting consumer choice and driving innovation beyond what can be achieved by competition law enforcement alone”. The draft report of JURI considers that “the acquisition of significant market power by dominant platforms has led to a situation in which “the winner takes it all”, and the market is composed of a small number of players each exerting market dominance over their competitors and imposing their business practices on users”. It further “calls on the Commission to assess the possibility of defining fair contractual conditions to facilitate data sharing with the aim of addressing imbalances in market power; suggests, to this end, to explore options to facilitate the interoperability and portability of data”.

(17)

      European Parliament resolution of 18 June 2020 on competition policy – annual report 2019 (2019/2131(INI)) , 18 June 2020.

(18)

     European Council, A new strategic agenda 2019-2024 , 20 June 2019. See also European Council Conclusions of 22 March 2019 .

(19)

      Council of the European Union Conclusions of 9 June 2020 , responding to the Commission’s Communication Shaping Europe’s Digital Future.

(20)

     Council of the European Union, Conclusions on a deepened Single Market for a strong recovery and a competitive, sustainable Europe , 11 September 2020.

(21)

     Ibid. 

(22)

     Council of the European Union, Special Meeting of the European Council– conclusions , 2 October 2020.

(23)

      Summary of the Stakeholder Consultation on the New Competition Tool and Summary of the contributions of the NCAs to the impact assessment of the new competition tool .

(24)

     For example, BEUC’s reply to the OPC states that the “challenges posed in particular by large players in digital markets require new instruments in addition to traditional competition law enforcement in order to protect consumers’ interests in an effective and timely manner.”

(25)

     See Annex 5.3 to the Impact Assessment.

(26)

     See the dedicated webpages on DG Competition’s website at: https://ec.europa.eu/competition/consultations/2018_vber/index_en.html (VBER), https://ec.europa.eu/competition/consultations/2019_hbers/index_en.html (HBERs), https://ec.europa.eu/competition/sectors/motor_vehicles/legislation/mvber_review.html (MVBER).

(27)

     See the dedicated webpages on DG Competition’s website at: https://ec.europa.eu/competition/consultations/2020_market_definition_notice/index_en.html .

(28)

      Commission Notice on the definition of relevant market for the purposes of Community competition law , OJ 97/C 372/03, at point 2.

(29)

     See brochure How do online platforms shape our lives and businesses?  

(30)

     In this Impact Assessment the notion of unfair business practices refers to both terms and conditions as well as the actual business practices of gatekeepers.

(31)

     Where appropriate in this Impact Assessment, references to EU should be understood as comprising the EEA.

(32)

     In the OPC on Ex Ante Rules a majority of stakeholders note that “certain platforms and their ecosystems have become unavoidable to access a large variety of contents and services on the internet. Those structuring platforms have become gatekeepers not only within their services, but for the internet at large.” See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(33)

     M. Bajgar, G. Berlingieri, S. Calligaris, C. Criscuolo & J. Timmisn (2019), Industry Concentration in Europe and North America , OECD Productivity Working Paper; G. Grullon, Y. Larkin, & R. Michaely (2019), Are US Industries Becoming More Concentrated?, Review of Finance, Volume 23(4), pages 697–743; Market Concentration - Note by Jason Furman, Hearing on Market Concentration , 7 June 2018; G. Gutiérrez & Th. Philippon (2017), Declining Competition and Investment in the U.S., ; G. Gutiérrez & Th. Philippon (2018), How European Markets Became Free: A Study of Institutional Drift .

(34)

     Source: R. Fijneman, K. Kuperus, J. Pasman (2018), Unlocking the value of the platform economy , KPMG report for the Dutch Transformation Forum 

(35)

      Observatory expert group Progress report on the Measurement of the Online Platform Economy .

(36)

     See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(37)

     See Summary of the Stakeholder Consultation on the New Competition Tool .

(38)

     See Summary of the contributions of the NCAs to the impact assessment of the new competition tool .

(39)

     See M. Motta & M. Peitz (2020), Intervention trigger and underlying theories of harm - Expert advice for the Impact Assessment of a New Competition Tool , Chapter 2, and G. S. Crawford, P. Rey, & M. Schnitzer (2020), An Economic Evaluation of the EC’s Proposed “New Competition Tool” , Section V.C.

(40)

     In the OPC on Ex Ante Rules, respondents in general consider that unfair practices by gatekeepers have a concerning impact on competition, innovation and consumer choice. Competition is hampered when gatekeepers create barriers for new market operators to enter the market, thereby resulting in reduction of investments and innovation and consumer choice stifling. Unfair practices are considered to be the means by which digital platforms increase the cost of switching or multi-homing for users, thereby limiting market contestability and preserving their market power. See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(41)

     G. Biglaiser, E. Calvano & J. Crémer (2019), Incumbency advantage and its value, Journal of Economics and Management Strategy, volume 28(1), pages 41-48; A. Afilipoaie, K. Donders & P. Ballon (2019), What Are the Pro- and Anti-Competitive Claims Driving the European Commission’s Platform Policies? A Case Study Based Analysis of the European Commission’s Take on Platform Cases .

(42)

     In the OPC on Ex Ante Rules, several hundreds of respondents identified each of these activities. See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(43)

     Several stakeholders, such as media publishers or game developers, raised concerns about this specific issue in the OPC on Ex Ante Rules as well as through different legal actions taken both in the Europe and the US.

(44)

     Often called Wide Most Favoured Nation (‘MFN’) clauses. These clauses generally also apply to the business user’s direct sales channels, however this element of such clauses would not be affected by the options presented in this impact assessment.

(45)

      https://judiciary.house.gov/calendar/eventsingle.aspx?EventID=3113 .

(46)

     See I. Graef (2018), Algorithms and Fairness: What Role for Competition Law in Targeting Price Discrimination towards End Consumers?, Columbia Journal of European Law, volume 541, pages 546–8. and I. Graef (2019), Differentiated Treatment in Platform-to-Business Relations: EU Competition Law and Economic Dependence, Yearbook of European Law, Volume 38, pages 448–499.

(47)

      Observatory expert group report on differentiated treatment .  

(48)

     See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(49)

     See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules. 

(50)

     For a more detailed overview see IA support study .

(51)

     See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(52)

     This notion of ‘digital intensity’ is rather broad and encompasses all firms with relatively high exposure to ‘Information and Communication Technologies’ (in terms of their investments, or input purchases), as well as firms reporting online sales. For the definition of ‘digital intensity’, see F Calvino, C Criscuolo, L Marcolin, & M Squicciarini (2018), A taxonomy of digital intensive sectors, OECD Science, Technology and Industry Working Papers 2018/14.

(53)

     M. Bajgar, G. Berlingieri, S. Calligaris, C. Criscuolo & J. Timmis, (2019), Industry Concentration in Europe and North America, OECD Productivity Working Papers, 2019-18, OECD Publishing, Paris.

(54)

     The study used firm-level data sourced from the commercial dataset Orbis® by Bureau van Dijk (BVD). It provides information on firms' localisation, annual balance sheet and income statements, although the number of observations per country can vary significantly. It covers the period 2001-2014 for 26 countries: Australia, Austria, Belgium, Bulgaria, Denmark, Estonia, France, Finland, Hungary, Germany, Indonesia, India, Ireland, Italy, Japan, Republic of Korea, Luxembourg, the Netherlands, Portugal, Romania, Slovenia, Spain, Sweden, Turkey, the UK, US. See also J. Federico, D. Leigh & S. Tambunlertchai (2018), Global Market Power and its Macroeconomic Implications, IMF Working Paper WP/18/137.

(55)

     See S. Calligaris, C. Criscuolo, & L. Marcolin (2018), Mark-ups in the Digital Era, OECD Science, Technology and Industry Working Papers 2018/10. See also https://www.oecd-ilibrary.org/docserver/4efe2d25-en.pdf?expires=1601160653&id J. De Loecker, & J. Eeckhout (2017), The Rise of Market Power and the Macroeconomic Implications .

(56)

     Ibid.

(57)

     See J. Federico, D. Leigh & S. Tambunlertchai (2018), Global Market Power and its Macroeconomic Implications, IMF Working Paper WP/18/137.

(58)

     According respectively to estimates from the P2B Impact Assessment and the Online Platform Economy Observatory.

(59)

      https://platformobservatory.eu/state-of-play/power-over-users/ .

(60)

      https://platformobservatory.eu/state-of-play/power-over-users/ .  

(61)

     In general, most of the issues presented by the users are due to a perceived imbalance in bargaining power between platforms and business users, which hampers competition, fosters uncertainty in relation to contractual terms and also results in lock-in of consumers. See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(62)

     See IA support study .

(63)

     See IA support study .

(64)

     UNCTAD (2019), Digital economy report .

(65)

     According to the Online Platform Economy Observatory, traffic share is the most revealing indicator of the economic significance of online platforms.

(66)

      https://platformobservatory.eu/news/covid-19-and-online-platform-economy/ .  

(67)

      Ibid . 

(68)

     A. Ezrachi & M. Stucke, Digitalisation and its impact on innovation, R&I Paper Series, Working Paper 2020/07, European Union 2020, page 34.

(69)

     Ibid., page 22.

(70)

     The choice for consumers is limited by lock-in effects and lack of innovative alternatives blocked out of the gatekeeper platform(s).

(71)

     The Online Observatory report on differentiated treatment shows that they may be subject to search diversion, i.e. platforms may have incentives to a biased order of products/services presentation, which would divert consumers from products/services they initially intended to buy, pushing them to purchase more and/or more expensive products/services.

(72)

     Gatekeepers’ extraction of information leads to consumer profiling, unwanted advertisement targeting and privacy concerns.

(73)

     See for instance J. De Loecker, J. Eeckhout & G. Unger (2020), The Rise of Market Power and the Macroeconomic Implications, The Quarterly Journal of Economics, Volume 135(2), Pages 561–644; J. De Loecker & J. Eeckhout (2018), Global Market Power, NBER Working Paper No. 24768; S. Barkai (2020), Declining Labor and Capital Shares, Journal of Finance, Volume 75, Issue 5; S. Calligaris, C. Criscuolo, & L. Marcolin (2018), Mark-ups in the Digital Era, OECD Science, Technology and Industry Working Papers 2018/10, on trends in firm-level mark-ups across 26 countries for the period 2001-14. They find that average mark-ups are higher and have grown more in ‘digital intensive’ sectors than in less ‘digital intensive’ sectors over the 2001-2014 period.

(74)

     See M. Motta & M. Peitz (2020), Intervention trigger and underlying theories of harm - Expert advice for the Impact Assessment of a New Competition Tool , Chapter 2.

(75)

     See M. Motta (2014), Competition Policy -Theory and Practice, Cambridge University Press.

(76)

     See R. Whish (2020), The New Competition Tool: Legal comparative study of existing competition tools aimed at addressing structural competition problems, with a particular focus on the UK’s market investigation tool .

(77)

     See Summary of the contributions of the NCAs to the impact assessment of the new competition tool . Some NCAs indicated that some of the questions in the questionnaire did not apply to them, because they did not have come across this particular feature or scenario in their recent case-work. When reporting on the views expressed by NCAs on particular issues, this Impact Assessment only reflects the views of those NCAs that did in fact express such a view.

(78)

     See Summary of the Stakeholder Consultation on the New Competition Tool .

(79)

     77% for network effects, 51% for economies of scale, 49% for data, 44% for consumer bias 41% for switching costs. See ICN ‘Report on results of the ICN survey dominance/substantial market power in digital markets’ (‘ ICN Report on digital markets’ ).

(80)

     Case AT.39740 Google Search (Shopping), Commission Decision of 27 June 2017, paragraph 272.

(81)

     Case AT.40099 Google Android, Commission Decision of 18 July 2018, paragraph 462.

(82)

     Case AT.40153 E-book MFNs and related matters (Amazon), Article 9 Decision of 4 May 2017, paragraph 65.

(83)

     J. Crémer, Y.-A. de Montjoye & H. Schweitzer (2018), Digital policy for the digital era , page 70.

(84)

     Case AT.40099 Google Android, Commission Decision of 18 July 2018, paragraph 600.

(85)

     See M. Motta & M. Peitz (2020), Intervention trigger and underlying theories of harm - Expert advice for the Impact Assessment of a New Competition Tool , Chapter 2.

(86)

     See OECD Policy Roundtable on Two-Sided Markets , 2009, in particular the contribution by the European Commission, pages 157-186.

(87)

     Case COMP/C-3/37.792 Microsoft, Commission Decision of 24 March 2004, paragraph 420.

(88)

     Case AT.40099 Google Android, Commission Decision of 18 July 2018, paragraph 464.

(89)

     In addition, in Apple/Shazam (Case M. M.8788 Apple/Shazam, Commission Decision of 6 September 2018, paragraphs 221 ff.), the Commission found that the merger would give Apple access to Shazam’s consumer data, which would give it the “[a]bility to use the Customer Information to put competitors at a competitive disadvantage”, while the evidence on Apple’s incentives to do so was mixed.

(90)

     Case AT.39740 Google Search (Shopping), Commission Decision of 27 June 2017, paragraph 287.

(91)

     See Summary of the Stakeholder Consultation on the New Competition Tool .

(92)

     See Summary of the Stakeholder Consultation on the New Competition Tool and Summary of the contributions of the NCAs to the impact assessment of the new competition tool .

(93)

     A/B testing (also known as split testing) is a process of showing two variants of the same web page to different segments of visitors at the same time and comparing which variant drives more conversions. A/B testing is one of the most important ways to optimise a website's funnel in digital marketing.

(94)

     A recent JRC report - Technology and Democracy: Understanding the influence of online technologies on political behaviour and decision-making - describes such techniques as “design choices that benefit an online service by coercing, steering, or deceiving users into making unintended and potentially harmful decisions”. There are patterns used in websites and apps that make users do things that they didn't mean to, like buying or signing up for something (see https://darkpatterns.org/ ). To explain such coercive and manipulative techniques, the JRC report refers to the “roach motel” example, i.e. it is easy for users to get into a certain situation, but difficult to get out. For instance, creating an account would require just a few clicks, but deleting it would involve more than 10 steps that are difficult to achieve without instructions.

(95)

     A. Mathur, G. Acar, M. J. Friedman, E. Lucherini, J. Mayer, M. Chetty, & A. Narayanan (2019), Dark patterns at scale, Proceedings of the ACM on Human-Computer Interaction, volume 3, pages1–32.

(96)

     This section quotes E. Katsamakas & H. Madany (2019), Effects of user cognitive biases on platform competition, Journal of Decision Systems, volume 28(2), pages 138-161.

(97)

     Case AT.40099 Google Android, Commission Decision of 18 July 2018, paragraph 781.

(98)

     Case AT.40099 Google Android, Commission Decision of 18 July 2018, paragraph 783.

(99)

     Case COMP/C-3/37.792 Microsoft, Commission Decision of 24 March 2004, paragraph 463.

(100)

     Case AT.40099 Google Android, Commission Decision of 18 July 2018, paragraph 470.

(101)

     See Support study to the Observatory for the Online Platform Economy, Report on the main obstacles and opportunities for multihoming, https://platformobservatory.eu/research/ .  

(102)

     See Regulation (EU) 2019/1150 of the European Parliament and of the Council of 20 June 2019 on promoting fairness and transparency for business users of online intermediation services .

(103)

     https://www.statista.com/statistics/1010716/apple-app-store-google-play-app-downloads-forecast/  Statista

(104)

      https://www.businessofapps.com/data/facebook-statistics/ .

(105)

      https://www.businessofapps.com/data/facebook-statistics/ .

(106)

      Ecommerce report 2019 .

(107)

     Report European Cross Border B2C E-Commerce Market 2020 Double Digit Growth Expected after 2020.

(108)

     Source: Statista .

(109)

      https://sproutsocial.com/insights/facebook-stats-for-marketers/ .

(110)

See IA support study .

(111)

The amount of data created each year in the digital economy is growing at an exponential rate. In 2020, it is estimated to reach 47 zettabytes at worldwide level compared to 12 zettabytes in 2015. Forecasts point to 142 zettabytes in 2035. Source: Statista (2019), Digital Economy Compass .

(112)

      https://platformobservatory.eu/news/covid-19-and-online-platform-economy/ .  

(113)

     Stock prices of each company are normalised to 100 in 2014, i.e. they are expressed relative to their respective value in 2014. This graph therefore allows to compare the development of stock prices across different companies, but not their absolute level.

(114)

     The S&P 500 is a stock market index that measures the stock performance of 500 large companies listed on stock exchanges in the United States. It is one of the most commonly followed equity indices.

(115)

Competition in the voice-assistant markets will become more and more difficult as the algorithms underpinning the assistants benefit from the concentration of access to customers’ and users’ accumulated data. Incumbent platforms benefitting from large volume and variety of datasets will be able to provide more refined search results through their own assistants. This is an important competitive advantage vis-à-vis smaller and/or start-up platforms.

(116)

The likelihood of forthcoming legislative action is further supported by numerous reports of influential national authorities.

(117)

     See IA support study .

(118)

The replies of citizens and stakeholders to the Commission’s OPC and the feedback of the NCAs replying to the Commission’s questionnaire indicate that market failures appear to be widespread across the Union, in particular in digital markets of cross-border nature. See Summary of the Stakeholder Consultation on the New Competition Tool and Summary of the contributions of the NCAs to the impact assessment of the new competition tool . While respondents indicated that market failures may occur in all industry sectors, several respondents emphasised that they are particularly prominent in the digital sphere.

(119)

     In addition, Member States may not have the means to adopt appropriate measures to tackle market failures. Only some NCAs of Member States have instruments that enable them to tackle, to a certain extent, market failures, such as Greece and Romania. It can be expected that further Member States will adopt such national tools. Eight NCAs signalled that the competition rules applicable in their respective Member States have been amended in order to deal with market failures or that there are plans for doing so, namely Belgium, Bulgaria, Austria, Romania, Lithuania, Iceland, Germany and Greece. See Summary of the contributions of the NCAs to the impact assessment of the new competition tool .

(120)

     A merger or acquisition will be of an ‘EU dimension’ where the aggregate turnover of the undertakings concerned exceeds given thresholds; irrespective of whether or not the undertakings effecting the concentration have their seat or their principal fields of activity in the EU, provided they have substantial operations there. Council Regulation (EC) No 139/2004 of 20 January 2004 on the control of concentrations between undertakings (the EC Merger Regulation ), OJ L 24, 29.1.2004, pages 1–22, Article 1(2).

(121)

See Summary of the Stakeholder Consultation on the New Competition Tool and Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(122)

     See Summary of the Stakeholder Consultation on the New Competition Tool .

(123)

European Court of Auditors, Special Report 24/2020: EU audit report: merger control and antitrust proceedings , 19 November 2020, at paragraph 59.

(124)

     While certain forms of unfair business practices can be abusive under Article 102(a) TFEU, finding such an abuse not only requires a dominant undertaking but generally also an effect on competition. If an undertaking imposes on its trading partners or obtains from them terms and conditions that are unjustified, disproportionate or without consideration but without affecting competition on the market, competition law generally does not apply (See recital 9 of Regulation (EC) No 1/2003. Some national competition laws also prohibit the abuse of economic dependence). Such behaviour resulting from imbalances in bargaining power that do not affect competition is usually the domain of unfair trading laws.

(125)

  Council Regulation (EC) No 1/2003 of 16 December 2002 on the implementation of the rules on competition laid down in Articles 81 and 82 of the Treaty , OJ L 001 4.1.2003, page 1.

(126)

     See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(127)

     Regulation (EU) 2016/679 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC ( General Data Protection Regulation ), OJ/L 119/1 (2016).

(128)

     While some voluntary efforts for data portability by some platforms have been underway since 2017 in the ‘Data Transfer Project’, the project described itself still as ‘early stage’ and activity peaked in 2018 on the project. It should not be underestimated that this ‘Data Transfer Project’ is at present limited to only several large online platforms, which means that actual or potential competitors do not (yet) benefit from this project.

(129)

     See for example O. de Hert, V. Papakonstantinou & G. Malgieri (2018), The right to data portability in the GDPR: Towards user-centric interoperability of digital services, Computer Law & Security Review, volume 34(2), pages 193-203.

(130)

      https://eur-lex.europa.eu/legal-content/EN/TXT/?qid=1601555056590&uri=CELEX:32005L0029 .  

(131)

      https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A31993L0013 .  

(132)

      https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32006L0114 .  

(133)

     Sources supporting the assessment include the existing enforcement experience under competition rules, both within the EU and beyond, and other areas of law (e.g. protection of personal data); the numerous expert studies and reports both within the EU as well as internationally; complaints from business users and customers of gatekeepers as well as several on-going regulatory (e.g. Australia; Japan) or enforcement interventions (e.g. US); the reports and support studies for the Observatory on the Online Platform Economy drawn up by the independent Observatory Expert Group and external contractors respectively; the IA support study, which provided both quantitative (e.g. data analysis; case studies) and qualitative input; and the broad consultation across stakeholder groups.

(134)

The activity undertaken by the firm has been also considered in a recent advice by the CMA on the Digital Markets Unit. CMA is recommending Digital Markets Taskforce (‘DMU’) initially prioritising digital firms active in particular activities (e.g. online marketplaces, app stores, social networks, web browsers, online search engines, operating systems and cloud computing services).

(135)

     The IA support study contains an analysis of business areas including mobile operating systems, app stores, desktop operating systems, search, social media, advertising (incl. search, display & video), e-commerce and cloud services. See also Annex 5.6 to the Impact Assessment.

(136)

Video streaming or video-on-demand services are currently characterised by less pronounced network effects (given the high costs of producing the distributed content) and switching costs are also not particularly high as subscriptions can be easily cancelled. In addition, there is no evidence of the presence of the problems described in Section 2.1 in video streaming or video-on-demand services.

(137)

In industrial B2B platforms the clients tend to be big, sophisticated companies which are not easily swayed by the platforms choice of ranking. They do not exhibit a similar dependency of the provider-side of the market at present. Switching costs are significantly lower: if products or services are delisted, corporate clients can insist that they be reinstated. Corporate clients have more leverage over platform decision. Support study to the Observatory for the Online Platform Economy, Developments concerning B2B platforms and emerging issues.

(138)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(139)

See for instance, Case AT.40099 Google Android, Commission Decision of 18 July 2018, section 9.4.

(140)

ACM, Market study into mobile app stores , 11 April 2019.

(141)

     US House of Representatives Majority Staff Report, Investigation of Competition in Digital Markets , October 2020.

(142)

     Furman report,  Unlocking digital competition, Report of the Digital Competition Expert Panel , March 2019.

(143)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(144)

      https://platformobservatory.eu/app/uploads/2020/09/Analytical-Paper-5-Business-user-and-third-party-access-to-data_final.pdf .

(145)

  https://platformobservatory.eu/app/uploads/2020/09/Analytical-paper-1-Platform-data-access-and-secondary-data-sources_final.pdf .

(146)

      https://platformobservatory.eu/app/uploads/2020/09/Analytical-Paper-2-Differentiated-treatment_final.pdf .

(147)

   As defined in Directive (EU) 2016/1148.

(148)

     Source: Statcounter .

(149)

     CMA report on Online platforms and digital advertising .

(150)

     US House of Representatives Majority Staff Report, Investigation of Competition in Digital Markets , October 2020.

(151)

     Furman report,  Unlocking digital competition, Report of the Digital Competition Expert Panel , March 2019.

(152)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(153)

     ACCC report, Digital Platforms Inquiry, Final Report , June 2019.

(154)

     Source: Statcounter .

(155)

     CMA report on Online platforms and digital advertising .

(156)

     US House of Representatives Majority Staff Report, Investigation of Competition in Digital Markets , October 2020.

(157)

     Furman report,  Unlocking digital competition, Report of the Digital Competition Expert Panel , March 2019.

(158)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(159)

     ACCC report, Digital Platforms Inquiry, Final Report , June 2019.

(160)

     See CMA report on Online platforms and digital advertising , page 119-123.

(161)

     CMA report on Online platforms and digital advertising .

(162)

     US House of Representatives Majority Staff Report, Investigation of Competition in Digital Markets , October 2020.

(163)

     Furman report,  Unlocking digital competition, Report of the Digital Competition Expert Panel , March 2019.

(164)

     See for instance data at worldwide level (including platforms mostly active in China) in Statista .

(165)

     Furman report,  Unlocking digital competition, Report of the Digital Competition Expert Panel , March 2019.

(166)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(167)

Source: Statcounter .

(168)

     US House of Representatives Majority Staff Report, Investigation of Competition in Digital Markets , October 2020.

(169)

     Furman report,  Unlocking digital competition, Report of the Digital Competition Expert Panel , March 2019.

(170)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(171)

     See for instance shares reported by Statista at a worldwide level.

(172)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(173)

     See CMA report on Online platforms and digital advertising , section 5.

(174)

     CMA report on Online platforms and digital advertising .

(175)

     US House of Representatives Majority Staff Report, Investigation of Competition in Digital Markets , October 2020.

(176)

     Furman report,  Unlocking digital competition, Report of the Digital Competition Expert Panel , March 2019.

(177)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(178)

     ACCC report, Digital Platforms Inquiry, Final Report , June 2019.

(179)

      https://platformobservatory.eu/research/ .

(180)

See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(181)

     Very similar criteria have been considered by the CMA in its recent Advice on the DMU (see in particular point 4.19 of the Advice; available here: https://assets.publishing.service.gov.uk/media/5fce7567e90e07562f98286c/Digital_Taskforce_-_Advice_--.pdf ).

(182)

     See IA support study , indicator and cluster analysis.

(183)

     Market capitalisation is the aggregate valuation of the company based on its current share price and the total number of outstanding stocks. It is calculated by multiplying the current market price of the company's share with the total outstanding shares of the company and reflects the relatively significant potential to monetise these users in the near future.

(184)

The respective relevant levels should be set at a level reflecting a substantive percentage of the entire Union population when it comes to end users and of the entire population of businesses using platforms to determine the threshold for business users.

(185)

     Very similar parameters pointing to digital firms with the likely Strategic Market Status (‘SMS’) have recently been considered and proposed by the CMA in its Advice of the DMU. For example, CMA proposes the use of the firm’s revenue as a criteria for the focus of the DMU’s enforcement activities, which should be on firms with annual UK revenue in excess of GBP 1 billion, and particularly those which also have annual global revenue in excess of GBP 25 billion.

(186)

     This estimate as well as other estimates as to the number of gatekeepers to be covered under different options is largely based on publicly available data, supplemented with data on end user numbers sourced from different external providers.

(187)

See notably CERRE Report, Digital Markets Act: Making economic regulation of platforms fit for the digital age , 24 November 2020.

(188)

     See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules, Summary of the Stakeholder Consultation on the New Competition Tool and Summary of the contributions of the NCAs to the impact assessment of the new competition tool .

(189)

See also Annexes 5.3 and 5.6 to the Impact Assessment.

(190)

  https://www.bundeskartellamt.de/SharedDocs/Entscheidung/DE/Entscheidungen/Missbrauchsaufsicht/2019/B6-22-16.pdf?__blob=publicationFile&v=8 .

(191)

  https://en.agcm.it/en/media/press-releases/2017/5/alias-2380 .

(192)

See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(193)

See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(194)

  https://www.jftc.go.jp/en/pressreleases/yearly-2019/October/191031.html .

(195)

     Case AT.40099 Google Android, Commission Decision of 18 July 2018.

(196)

     Case AT. 39530 Microsoft (Tying), Commission Decision of 16 December 2009.

(197)

See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(198)

      https://www.ftc.go.kr/solution/skin/doc.html?fn=508d97db636c2f7f0961bf6361cfd44f09977d1a7a06f4dd5603f17c11d61013&rs=/fileupload/data/result/BBSMSTR_ 000000002402/ .

(199)

  https://www.agcm.it/dettaglio?db=41256297003874BD&uid=C66CBC6B87379163C125840000581AE0&view=vw0301&title=A529-GOOGLE/COMPATIBILITÀ%20APP%20ENEL%20X%20ITALIA%20CON%20SISTEMA%20ANDROID%20AUTO&fs=%20%2082_CE/102_CE-Abuso%20di%20posizione%20dominante .

(200)

  https://www.acm.nl/en/publications/big-tech-and-dutch-payment-market-tightening-rules-needed-maintain-level-playing-field .  

(201)

See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(202)

  https://www.justice.gov/opa/pr/justice-department-sues-monopolist-google-violating-antitrust-laws .

(203)

  https://ec.europa.eu/digital-single-market/en/news/study-contractual-relations-between-online-platforms-and-their-professional-users .

(204)

Csse T-201/04 Microsoft v Commission, judgment of 17 September 2007, at paragraphs 1091 ff.

(205)

  Directive 2005/29/EC of the European Parliament and of the Council of 11 May 2005 concerning unfair business-to-consumer commercial practices in the internal market and amending Council Directive 84/450/EEC, Directives 97/7/EC, 98/27/EC and 2002/65/EC of the European Parliament and of the Council and Regulation (EC) No 2006/2004 of the European Parliament and of the Council .    

(206)

  Directive (EU) 2019/633 of the European Parliament and of the Council of 17 April 2019 on unfair trading practices in business-to-business relationships in the agricultural and food supply chain .

(207)

  Directive (EU) 2018/1972 of the European Parliament and of the Council of 11 December 2018 establishing the European Electronic Communications Code (Recast) .

(208)

     See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(209)

Article 77 of the Code establishes that where “the national regulatory authority concludes that the appropriate obligations […] have failed to achieve effective competition and that there are important and persisting competition problems or market failures identified in relation to the wholesale provision of certain access product markets, it may, on an exceptional basis […] impose an obligation on vertically integrated undertakings to place activities related to the wholesale provision of relevant access products in a business entity operating independently”. Article 77 as well as its accompanying recitals establish a series of proportionality criteria that must be fulfilled prior to the imposition of such a requirement.

(210)

     See Summary of the Stakeholder Consultation on the New Competition Tool and Summary of the contributions of the NCAs to the impact assessment of the new competition tool .

(211)

     See for example also Article 9 of Regulation (EU) 2017/2394 of the European Parliament and of the Council of 12 December 2017 on cooperation between national authorities responsible for the enforcement of consumer protection laws and repealing Regulation (EC) No 2006/2004 (Text with EEA relevance), OJ L 345, 27.12.2017, page 1.

(212)

  Directive (EU) 2018/1972 of the European Parliament and of the Council of 11 December 2018 establishing the European Electronic Communications Code (Recast) .

(213)

  Directive (EU) 2018/1972 of the European Parliament and of the Council of 11 December 2018 establishing the European Electronic Communications Code (Recast) .

(214)

     See Summary of the Stakeholder Consultation on the New Competition Tool .

(215)

See workshop with BEUC members on the Impact Assessment for a possible New Competition Tool, 1 October 2020. See also BEUC’s response to the OPC on the NCT: “challenges posed in particular by large players in digital markets require new instruments in addition to traditional competition law enforcement in order to protect consumers’ interests in an effective and timely manner.”

(216)

See Annex 2.1: Synopsis Report Open Public Consultation Ex Ante Rules.

(217)

      https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52016DC0288&from=EN .

(218)

      Contribution to Growth: European Digital Single Market: Delivering economic benefits to citizens and businesses (2019) .

(219)

     For more information see: https://ec.europa.eu/jrc/en/rhomolo .

(220)

     M. Christensen, A. Conte, F. Di Pietro, P. Lecca, G. Mandras, & S. Salotti (2018), The third pillar of the Investment Plan for Europe: An impact assessment using the RHOMOLO model (No. 02/2018). JRC Working Papers on Territorial Modelling and Analysis.

(221)

     UNCTAD (2019), Digital economy report .

(222)

      Cross-border Europe, annual analysis of the best global cross-border platforms operating in Europe , EU 28.

(223)

      Impact Assessment Annexes, SWD(2018) 138 final : 71% of consumers would have preferred platforms for their purchases. This figure is an underestimate given the COVID epidemics but provides already an idea of the important use of platforms by consumers.

(224)

      European Ecommerce Report 2017 . While the causal link between GDP growth and the economy of online platforms is difficult to demonstrate, considering these figures, it is reasonable to expect a relatively significant positive and growing contribution of the platform economy to the digital internal market and economic growth.

(225)

    Hootsuite Digital 2020 global statshot report .

(226)

     Ahn (2002) similarly concluded that “[a] large number of empirical studies confirm that the link between product market competition and productivity growth is positive and robust. […] Empirical findings from various kinds of policy changes […] also confirm that competition brings about productivity gains, consumers’ welfare gains and long-run economic growth". S. Ahn (2002), Competition, Innovation and Productivity Growth: A Review of Theory and Evidence, OECD Economics Working Paper No. 317. See also S. Nickell (1996), Competition and Corporate Performance, Journal of Political Economy, volume 104(4), pages 724-746, which found that the most competitive firms experienced productivity growth rates 3.8-4.6% higher than the least competitive. See also R. Disney, J. Haskel & Y. Heden (2003), Restructuring and productivity growth in UK manufacturing, The Economic Journal, volume 113(489), pages 666-694; R. Blundell, R. Griffith & J. van Reenen (1999), Market Share, Market Value and Innovation in a Panel of British Manufacturing Firms, Review of Economic Studies, volume 66(3), pages 529-54; S.I. Januszewski, J. Köke & J.K. Winter (2002), Product market competition, corporate governance and firm performance: an empirical analysis for Germany, Research in Economics, volume 56(3), pages 299-332. 

(227)

     J.M. Arnold, G. Nicoletti & S. Scarpetta (2011), Regulation, Resource Reallocation and Productivity Growth, European Investment Bank Papers, volume 16(1), pages 90-115; OECD’s project of 2013 on Supporting Investment in Knowledge Capital, Growth and Innovation .

(228)

Commission services’ meetings with stakeholders.

(229)

     G. Federico, F. Scott Morton & C. Shapiro (2019), Antitrust and Innovation: Welcoming and Protecting Disruption, NBER Working Paper No. 26005.

(230)

     P. Aghion, N. Bloom, R. Blundell, R. Griffith & P. Howitt (2005), Competition and Innovation: an Inverted-U Relationship, The Quarterly Journal of Economics, volume 120(2), pages 701-728. On empirical work, see P. Aghion, S. Bechtold, L. Cassar & H. Herz (2014), The causal effects of competition on innovation: Experimental evidence, Journal of Law, Economics, and Organization, volume 34 (2), pages 162-195.

(231)

     S. Ahn (2002), Competition, Innovation and Productivity Growth: A Review of Theory and Evidence, OECD Economics Working Paper No. 317; R. Gilbert (2007), Competition and innovation, Competition Policy Centre, UC Berkeley.

(232)

     C. Shapiro (2012), Competition and innovation. Did Arrow hit the bull’s eye?, chapter 7 of Josh Lerner and Scott Stern (eds.), The Rate and Direction of Inventive Activity Revisited, pages 361-404.

(233)

      https://ec.europa.eu/digital-single-market/en/news/impact-assessment-proposal--promoting-fairness-transparency-online-platforms .

(234)

     R. Gilbert & D. Newbery (1982), Preemptive patenting and the persistence of monopoly. The American Economic Review, volume 74(1), pages 514-526.

(235)

 The magnitude of online firms' acquisition is on the rise as highlighted by the Furman report, Unlocking digital competition, Report of the Digital Competition Expert Panel , reporting that the top 5 larger online platforms have carried out more 400 acquisitions worldwide in the last 10 years.

(236)

K.A. Bryan & E. Hovenkamp (2020), Antitrust limits on startup acquisitions, Review of Industrial Organization, volume 56, pages 615–636.

(237)

     See IA support study .

(238)

E. Carayannis & E. Grigoroudis (2014), Linking innovation, productivity, and competitiveness: implications for policy and practice, The Journal of Technology Transfer, volume 39(2), pages 199-218.

(239)

     D. Autor, D. Dorn, L.F. Katz, C. Patterson, & J. Van Reenen (2020), The fall of the labor share and the rise of superstar firms, The Quarterly Journal of Economics, volume 135(2), pages 645-709 and R.A. Decker, J.C. Haltiwanger, R.S. Jarmin & J. Miranda (2018), Changing business dynamism and productivity: Shocks vs. responsiveness. Technical report, NBER.

(240)

     UNCTAD (2019), Digital economy report .

(241)

  Japanese Fair Trade Commission Report regarding trading practices on digital platforms , October 2019.

(242)

     ACCC report, Digital Platforms Inquiry, Final Report , June 2019; Japanese Fair Trade Commission Report regarding trading practices on digital platforms , October 2019.

(243)

     Stigler Center report, Committee for the Study of Digital Platforms Market Structure and Antitrust Subcommittee Report , July 2019.

(244)

      https://judiciary.house.gov/calendar/eventsingle.aspx?EventID=3113 ; https://on.ft.com/33Bnq6T ; https://www.ftc.gov/policy/hearings-competition-consumer-protection

(245)

     US House of Representatives Majority Staff Report, , October 2020 .

(246)

     OECD (2015), Does competition kill or create jobs?, OECD Global Forum on Competition, DAF/COMP/GF(2015)9.

(247)

     See also A. Dierx, J. Heikkonen, F. Ilzkovitz, B. Pataracchia, M. Ratto, A. Thum-Thysen & J. Varga (2015), Distributional macroeconomic effects of EU competition policy – A general equilibrium analysis, paper to be published in a World Bank-OECD publication on Competition Policy, Shared Prosperity and Inclusive Growth, who estimate that enforcement of the EU competition rules by the European Commission has a sizeable impact on the creation of new jobs (they estimate around 650 000 after 10 years).

(248)

     OECD (2015), Does competition kill or create jobs?, OECD Global Forum on Competition, DAF/COMP/GF(2015)9, paragraph 78.

(249)

     UNCTAD (2019), Digital economy report .

(250)

See IA support study .

(251)

Detailed overview of the preferred option’s implications for gatekeepers (but also for competitors, business users and consumers) is presented in Annex 3 to the Impact Assessment.

(252)

OPC, direct submissions.

(253)

     See Summary of the Stakeholder Consultation on the New Competition Tool .

(254)

     Source: Statista .

(255)

     See https://www.oecd.org/going-digital/sme/resources/D4SME-Brochure.pdf .

(256)

     See https://ec.europa.eu/digital-single-market/en/use-internet .

(257)

     See https://ec.europa.eu/eurostat/statistics-explained/pdfscache/46776.pdf .

(258)

See IA support study .

(259)

     See for instance S. Ahn (2002), Competition, Innovation and Productivity Growth: A Review of Theory and Evidence, OECD Economics Working Paper No. 317. See also for example, a study by the European Commission (2015) on The Economic Impact of enforcement of competition policies in the functioning of EU energy markets , which found that the Commission's decision finding an abuse of dominance by E.ON lead to a reduction in prices for both wholesalers and retailers to the benefit of consumers. See also the Note by the UNCTAD Secretariat (2014), The benefits of competition policy for consumers .

(260)

Online platforms benefit from asymmetry of information (they dispose of large data sets compared to consumers). Platforms’ analytical capacity gives them the possibility to use advanced algorithms and machine learning techniques to facilitate targeting, discriminatory practices and behavioural manipulation. BEUC considers that such practices can have  an impact on demand and distribution of wealth  – “the most vulnerable consumers might end up paying higher prices than under a competitive price scenario (when personalisation is combined with commercial practices seeking to increase the individual consumer’s willingness to pay). They may also be used to target biases and reinforce existing or desired viewpoints with the aim of keeping users engaged with the firm’s platform so as to generate advertising revenues .” BEUC (2019), The Role of Competition Policy in Protecting Consumers’ Well-being in the Digital Era .

(261)

Cf. General Court in Cisco and Messagenet.

(262)

In fact, higher advertising prices represent increased costs to the companies producing goods and services which are purchased by consumers. These costs are expected to be passed through to consumers in terms of higher prices for goods and services, even if the downstream market is highly competitive.

(263)

     See Section 6 of CMA report on Online platforms and digital advertising .

(264)

See IAB Europe AdEx Benchmark 2019 Report .

(265)

See Section 6 of CMA report on Online platforms and digital advertising .

(266)

As explained in Sections 5.3.3.3 and 5.3.3.4, the inclusion of additional core platform services in the scope could also imply the inclusion of additional obligations that would relate to those services and additional gatekeepers that would be active in those services.

(267)

     See detailed explanation in Annex 3 to the Impact Assessment.

(268)

     For sub-option 2-A, five to seven out of the 15 gatekeepers would be identified on the basis of the quantitative threshold; the corresponding number would be 10 to 15 for sub-option 1-B.

(269)

Annex 3 to the Impact Assessment aims at quantifying benefits for the preferred option. It seems difficult to distinguish however which benefits could be attributed to each of the options. It appears therefore that a quantitative comparison of benefits would not be sufficiently reliable.

(270)

     Cross-border e-commerce in Europe was worth EUR 143 billion in 2019, with 59% of this market being generated by online marketplaces. This is projected to increase to 65% in 2025 ( Ecommerce News Europe (2020 )).

(271)

     In line with the impact assessment requirements an attempt was made by JRC (see Annex 4.2 to the Impact Assessment) to quantify consumer surplus which would stem if the measures foreseen under the preferred option were to be adopted and implemented. It is important to stress however, that such quantification remains a highly theoretical exercise; this is the reason why the qualitative assessment of implications for consumers should be considered as a more reliable analysis of the impacts of a regulatory intervention.

(272)

 Higher investment in R&D in the ICT sector in EU27 leads to an overall increase in the EU27 income between 0.09% to 0.17% of 2014 EU GDP, this is between EUR 12 billion and EUR 23 billion; input-output micro-econometric modelling, See Annex 3 to the Impact Assessment.

(273)

In addition, it would always be possible under Option 2 to also include new digital services in scope during the review of the Regulation, which could possibly take place every three years.

(274)

     See H. Schweitzer (2020), The New Competition Tool: Its institutional set up and procedural design , Chapter II.

(275)

     Ibid, at Chapter V and Chapter X.

(276)

See R. Whish (2020), The New Competition Tool: Legal comparative study of existing competition tools aimed at addressing structural competition problems, with a particular focus on the UK’s market investigation tool , at Chapter 4 and Chapter 7.

(277)

The P2B Regulation applies to all online intermediation services and all online search engines, regardless of their size. Given the relatively wide scope of the P2B Regulation, which includes many services that are provided locally, the regulation is enforced at the level of Member States. Given that the Regulation mainly requires the relevant providers to engage in one-off actions that simultaneously benefit their entire user bases (e.g. providing transparency in general terms and conditions, or putting in place an internal complaint-handling mechanism), the regulation’s dual private and public enforcement mechanism is geared towards finding ‘systemic’ breaches of the regulation, which generally will not require in-depth investigations and economic analyses by regulators. For example, a regulator or commercial court will likely only find that an online intermediation services provider breached the obligation to provide internal complaint-handling if they can establish that it is entirely absent or that a pattern exists of unfairly rejected complaints. 

(278)

See for example Directive (EU) 2019/633 of the European Parliament and of the Council of 17 April 2019 on unfair trading practices in business-to-business relationships in the agricultural and food supply chain . To improve farmers’ and small and medium sized businesses’ position in the food supply chain, the EU adopted this legislation banning certain unfair trading practices. These include (but are not limited to): late payments for perishable food products, last minute order cancellations, unilateral changes to contracts, refusal to enter into a written contract, returning unsold or wasted products or payment for buyer’s marketing.

(279)

     The possibility of false positives under Option 2 is partially addressed by the fact that the provider of core platform services would be able to present, in exceptional circumstances, serious and substantiated arguments to demonstrate that it does not fulfil the objective requirements for a gatekeeper.

(280)

     Such relevant elements would include: (i) the size, including turnover and market capitalisation, operations and position of the provider of core platform services; (ii) the number of business users depending on the core platform service to reach end users and the number of end users; (iii) entry barriers derived from network effects and data driven advantages, in particular in relation to the provider’s access to and collection of personal and non-personal data or analytics capabilities; (iv) scale and scope effects the provider benefits from, including with regard to data; (v) business user or end user lock-in; and (vi) other structural market characteristics.

(281)

      https://platformobservatory.eu/app/uploads/2020/07/ProgressReport_Workstream_on_Measurement_and_ Economic_Indicators_2020.pdf .