Vertical Relations and Platforms: Updating for the Digital Dimension
16th CCP Annual Conference. (8-10 July).
A warm welcome to this year’s CCP conference—online! Our conference will be comprised of three two-hour sessions, one per day on 8/9/10 July. Sessions will be running 15:00 – 17:00 in London time, 16:00-18:00 in Brussels and 10:00-12:00 in Washington, DC. We will be blogging a summary for each session throughout the afternoons. Follow us and look out for updates on Twitter @ccp_uea and access the full conference agenda here. Registration remains open throughout the event and we will have a YouTube live stream running.
Our blogging team is constructed from our excellent research-active CCP doctoral students: Anush Ganesh, Selvin Thanacoody, Vasudha Wattal, Thanh Doan; and Research Associates James Craske and Bryn Enstone.
Introducing the topic of this year’s conference: vertical relations and platforms
Vertical relationships and platforms have always been a complex area within competition policy. Some aspects of agreements can be anti-competitive but new vertical structures can also have strong positive benefits, and authorities, therefore, need to strike a careful balance. The rapid growth of the digital economy, and especially the development of large online platforms, raises major new challenges for this assessment. The topics covered will relate to platforms, ecosystems and various types of agreements involving platforms and other businesses.
We hope our speakers (experts in their respective fields) will provide valuable insight into this panoply of issues, and related ones, from a variety of academic and policy-making perspectives. They will do so at a crucial time, given the impending review of the EU vertical agreements block exemption regulation and guidelines and the forthcoming CMA report on digital advertising.
Day 1 (8th July)
Keynote Speaker Pierre Régibeau
Pierre Régibeau (Chief Economist at DG Competition) began his Keynote by saying that DG Comp is in the middle of revising its guidelines on vertical restraints. He explained that for any reform to take place, it is important to see what has changed in markets and to read through the relevant academic literature. He stated that the current big issue seems to be the continued rise in strength and market power of digital platforms. When engaging in reform on guidelines for vertical restraints Régibeau made the point that it is important to try to consider the application of the rules to other markets and not restrict it only to platforms.
He felt the need to distinguish between different platform models such as Ad based platforms and App-based platforms and look at their way of monetising which would be through consumers in the former and through the device in the latter. He described how platforms use data differently; a firm like Google also has a vertical aspect and the issue begins when it has its own set of verticals. Market definition is one more area of consideration in platforms and he agreed that it might not be necessary for the firm to have high market share in all markets for it to have influence over several markets. He stated that control of one side gives power over other sides.
Régibeau made the point of considering the behavioural aspects of platforms concerning blocking access to the market and self-preferencing. He stated that these behaviours need to be detected when reforming the rules for platforms. He mentioned the importance of regulating the harmful behaviour of dominant platforms but was apprehensive of the use of regulation and break-up of firms: regulation is to be carried out very cautiously and only done when it is necessary and there is enough evidence that it works. He concluded his talk by saying that the need for regulation can only be ascertained through further research and market study.
Session 1 – Most Favoured Nation Clauses
(Chair: Sean Ennis. Panel: Thibaud Vergé, Dirk Auer, Nelson Jung)
Thibaud Vergé, Professor of economics at ENSAE Paris and visiting professor at the University of Bergen, delivered his presentation on ‘Platform MFN Clauses: economic theory and empirical evidence’. His talk focused on a discussion of existing theoretical literature as well as more recent empirical findings on the competitive effects of price parity clauses (PPCs). Thibaud opened the discussion outlining the reasons for the overall ambiguity in the effect of PPCs. On the one hand, as one platform increases prices, suppliers uniform all prices, which harms the consumer because of limited competition. On the other hand, Thibaud suggests, if inter-brand competition is fierce enough, then under parity, suppliers are tempted to drop out of a very expensive platform and have incentives to undercut their rivals on cheaper platforms. This may benefit platforms and consumers, but suppliers could be harmed. He further drew attention to the ambiguity in the ability of narrow PPCs to replicate the outcomes of wide PPCs.
The two empirical questions that he brought forward were whether commissions or prices are higher under PPCs; and whether commissions or prices are higher under wide rather than narrow PPCs? The empirical evidence on the competitive effects of PPCs across sectors points in the direction that removal of PPCs leads to decrease in prices. His concluding thoughts focussed on whether there is still need to be worried about PPCs and perhaps it should be considered whether OTAs can circumvent a ban on PPCs, referring to the introduction of Booking.com’s preferred partner program that emerged shortly after the ban. He also pointed to the emerging empirical evidence that suggests that lower prices on direct sales might be negatively affecting rankings on OTAs.
Dirk Auer, Senior Fellow in Law and Economics at the International Center for Law and Economics, discussed the legal and economic literature surrounding MFNs and their effect on competition. His presentation, ‘Platform MFN Clauses: Why should online firms want fair trade?’, began with an outline of three critical features of online retail (i.e., innovation, visibility and multiple distribution channels). On the one hand, retailers prepare their products in a manner that provides the in-store experience for their users at the point of purchase as well as after-sales services. But they are faced with a trade-off of whether to choose a platform to boost their visibility (though incur platform fees), or whether to consider direct sales channels at the risk of reliance on “loyal” customers. He addressed the considerable free-rider problems which undermine platform investment, such as when consumers may use a platform to browse through products, but purchase through direct channels or rival platform.
The key question that Dirk brought to the audience was whether MFN related competition enforcement, as well as literature from law and economics, accounts for these key features. Dirk finds that empirical studies do not say much about welfare effects since innovation, investment and entry are unexamined. In his concluding remarks, Broadly, Dirk considered that there is a need for empirical and theoretical literature to focus on the effects of these features. Competition authorities too should consider a few of these effects such as competition from other distribution channels, consumer loyalty concerning brand or platform, and the balance of bargaining power, among others.
Nelson Jung, a partner in Antitrust Practice at Clifford Chance, walked through his talk on the “Reflections on the 10th anniversary of the hotel online booking saga”. He began by highlighting that the key milestones in the EU and UK on MFNs span beyond the hotel online booking case, and considered whether, over the period, there has been more legal certainty on how MFNs are assessed. He went on to discuss the current legal framework and pointed out that there was a need for clear and narrow definitions such as those of, “hardcore” restrictions and who qualifies as a buyer. Nelson regards that even a decade since the hotel online booking and other cases, there has been a lack of consistency in the assessment of MFNs across jurisdictions. In today’s framework, wider MFNs are more likely to raise antitrust enforcement risks, but narrow MFNs have seen inconsistent approaches. His concluding thoughts were that there is a need for reform and greater legal clarity in this direction, and the EC/CMA has opportunities to devise an agreed framework to assess MFNs that account for economic evidences, strongly suggesting that MFNs are not “by their very nature injurious to competition”.
Session 2 – Bidding for Advertising
Chair: Sally Broughton Micova. Panel: Giuseppe Colangelo, Elias Deutscher
The session began with CCP’s Sally Broughton Micova introducing the concept of advertising in platforms and stressing on its importance in the digital economy. The first speaker was Giuseppe Colangelo who presented his paper titled, ‘Competing through Keyword Advertising’.
Giuseppe Colangelo began his presentation by explaining how keyword search advertising functions. He explained that search engines enable individuals to find websites and online content through the use of keywords. When displaying results, they display both natural results and those that are termed ‘sponsored results’ as a result of being paid (through a process of bidding for a higher place) to be placed alongside natural results. He argued that trademark holders complain that the use of a competitor’s mark to trigger the display of one’s own advertising as a sponsored link might have an adverse effect on the essential functions of the trademark.
He described the existence of Non-brand bidding agreements in different jurisdictions. He explained that the EC carried out an E-commerce Sector Inquiry which reported that some retailers are limited in their ability to use or bid on the trademarks of certain manufacturers to get a preferential listing on search engines’ paid referencing service or are only allowed to bid on certain positions. He explained that the CMA encountered brand-bidding restrictions in the markets for broadband, credit cards, energy, flights and home insurance, while the Netherlands ACM found restrictions in the hotel sector, and that the US FTC has ruled that the largest online retailer of contact lenses unlawfully entered into a web of anticompetitive agreements with rivals, preventing them from bidding for search engine result advertisements that would inform consumers that identical products were available at lower prices.
He then explained the difference in position concerning Trademark Protection for Keyword Advertising in the US and the EU and said that in the EU, the legal assessment of keyword advertising is centred only on the question of whether competitors are confusing or informing customers while in the US, the expansive approach to trademark protection embraces an additional layer, broadening the class of entities which can be held liable for using keywords corresponding to trademarks. This was followed by looking at the results from empirical studies which concluded differentially on the capability of advertising on competitors’ keywords to divert significant shares of traffic but shared a common view about the limited risk of consumer confusion generated by competitive advertising on brand search.
He then went on to describe that the IP-Antitrust clash mainly relates to online advertising restrictions in selective distribution channels and threw some light on the EC’s E-Commerce Sector Inquiry of 2017, case of Guess, and the case of 1-800 Contacts. He made the point that RPMs are not a good fit for analyzing keyword advertising since the TM settlements fall within the scope of the trademark, unlike RPMs which are doubly suspect. He concluded that the European approach is better suited to ensuring a balance between the interests at stake by circumscribing the scope of trademark protection to confine its potential anticompetitive effects and that principles of the limited trademark protection in the keyword advertising scenario have been set by the CJEU in trademark litigations, rather than in antitrust investigations.
Response to Colangelo
Elias Deutscher began his presentation by commending Colangelo’s work due to its thoroughness and followed it with 3 questions. 1. Whether the focus on TM facilitates or complicates Antitrust analysis, 2. Whether vertical and horizontal non-brand bidding restraints should be analyzed the same way, and 3. Should vertical non-brand bidding restraints be analyzed ‘by-object’ or ‘by-effect’? He made the case for looking beyond the scope of the trademark to assess the competitive impact of vertical non-brand bidding restraints and stressed on the need for better understanding and empirical analysis of the economic rationale and welfare effects of vertical non-brand bidding restraints.
Deutscher then explained the importance of display advertising in making sure that search advertising is effective as it plays a role in making the consumer aware of the brand and leads to the formation of opinion while search advertising is usually effective only at the stage of the purchase. He then explained how vertical non-brand bidding can be used as a tool to recover costs through price discrimination by illustrating through an example with 2 customer group: 1. Wealthy but busy, and 2. With budgetary constraints but willing to search. He explained that instead of pricing high or low, customers can be filtered to extract surplus from both groups. He explained that vertical non-brand bidding restraints increase search cost and by filtering customers, wealthy customers will click on the first sponsored link in Google and will be channelled to the online store of brand owners (e.g. guess.eu) where they can be charged a premium while those constrained by budgetary concerns will be served by other retailers bidding on long-tail of other search terms as they will click down further on the search list.
He concluded by making a case an effects-based analysis of vertical non-brand bidding restraints and said that there are potential pro-competitive as well as anti-competitive effects of vertical non-brand bidding restraints. Colangelo then responded with admiration for Deutscher’s arguments regarding an effect-based approach and responded to his first 2 questions by saying that TM facilitates competition law analysis and that vertical and horizontal non-brand bidding restraints should be distinguished.
Sean Ennis finished the session by drawing parallels between the previous session and this one by noticing a free-rider problem in both.
Day 2 (9th July)
Welcome to the second day of CCP’s annual conference. Registration remains open throughout the event and we will have a YouTube live stream running. Our first session will be on “data and exclusives”.
Session 3 – Data and Exclusives
Chair: Wynne Lam. Panel: Daniele Condorelli, Simon McDougall
The role of data, and exclusive use of data as encouraged by privacy regulation, may create a possibility for some firms to expand into other domains, while firms in those other domains do not have a chance to expand their business in the reverse direction. This session will explore ways in which counter-balancing policy interests intersect and impact each other.
Wynne Lam (CCP) introduced the session by providing an example of how personal data can be collected by firms when we engage in online activities and raised the question about how this data can be used by them, and whether it has any anticompetitive effect. The first speaker, Daniele Condorelli (University of Warwick), presented his recent research titled: “Harnessing platform envelopment through privacy policy tying”.
Condorelli began his presentation by explaining the key concept of platform envelopment and how platforms use privacy policy tying to collect data from online users in the context of advertising platform. Envelopment strategy is commonly used by platforms to enter the target market, which involves a firm offering its own functionalities and services with the target’s. For instance, Google entered the digital map market by bundling Google Map to its search platform, Google Search. Users usually get free access to online services only after agreeing to privacy policy tied by platforms, which allows platforms to extract their personal data. Envelopment to the new market enables platforms to gather richer data on the overlapping users, and privacy policy tying is the key tool to harness platform envelopment.
Following this, Condorelli explained that platforms have an incentive envelop to the target market, as richer data from enveloping can be used to improve the platform service quality in the primary market or exploited to raise advertising revenue. As a result, it would entrench the platform’s market position, and protect its monopoly profit from the threat of a more efficient entrant. The speaker emphasised that this would lead to less entry and consequently harm consumers. To regulate platform envelopment strategy, Condorelli proposed four possible policy remedies: i) Entry restrictions, ii) prohibition to data-tying, iii) allowing data sharing, and iv) allowing data portability. He argues that only the first remedy may lead to an increase in overall consumer welfare, whereas other remedies would suppress the data efficiency and economies of scope, which can actually harm consumers further. He concluded that privacy policy tying can leverage the platform’s dominant position, which may have foreclosure effects, and eventually harm consumers. Finally, he pointed out the threats of data-driven acquisition recently by big tech firms to the competition in the market.
Simon McDougall (Information Commissioner’s Office) took this opportunity to provide several practical perspectives on privacy policy and personal data protection. He began the talk by highlighting the importance of interaction between privacy policy and competition policy, which is shown by the recent collaboration between ICO, OFCOM and CMA to form Digital Regulation Cooperation Forum (DRCF). He then outlined the right and obligation of the data controller within the General Data Protection Regulation (GDPR), which involves sharing data within the internal organisation and with the third party.
He suggested several prospective areas that have not yet been explored; this needs academics and practitioners. The first issue is whether consumers should have the “right to be forgotten”, which is the right to request their personal data to be deleted by the entity. Second, McDougall addressed the key point of data portability and sharing that the fraction amount of data can be used by different parties without the disclosure of the underlying data. He suggested that this can be implemented by recent machine learning techniques, such as Federated Learning. Finally, Simon argues that although market power is driven by data, it may be misleading if we identify data itself as the only source of market power. For instance, many other factors can leverage platform market power, for example, computing power, or data processing skills. He concluded by encouraging more collaborative works and communications between people in the same or different organizations to improve our understanding of privacy, data protection in the fast-changing platform economy.
Session 4 – Differentiated Treatment
Chair: Bruce Lyons. Panel: Inge Graef, Ben Harries
This session focuses on developing analytical tools for assessing differentiated treatment, such as self-preferencing, secondary line differentiation, exploring the role of concepts like economic dependence and fairness.
CCP’s Bruce Lyons introduced the concept of differentiated treatment and invited Inge Graef to discuss her paper on ‘Differentiated treatment in platform-to-business relations’.
Inge Graef began her presentation by describing what is meant by differentiated treatment and defined it as discrimination or imposing dissimilar conditions ranging from more prominent placement of one’s own services in a ranking to preferential access to data and the favouring of businesses that pay higher levels of commission which would violate A. 102(c) TFEU. She mentioned landmark cases and investigations such as Google Shopping, Amazon (EU and Italy), Apple (Dutch ACM) and Apple App Store (EC). She noted that it is important to assess the legality of the practice which brought her to discussing her typology consisting of three categories: 1) pure self-preferencing, 2) pure secondary line differentiation, and 3) hybrid differentiation.
She then went on to explain the three categories followed by discussing how to deal with them. The first category was said to include vertically integrated platforms that favour their own affiliated services with an exclusionary motive. She gave the examples of preferential place in ranking and preferential access to data. The second category was said to include non-vertically integrated platforms that give more favourable treatment to one or more non-affiliated businesses over others with an exploitative motive attached. The examples given were of ranking in hotel booking platforms such as Expedia. The third category was said to include platforms that differentiate among non-affiliated businesses in order to favour its own services elsewhere outside the relevant market such as in the case of Unlockd versus Google where the issue related to using non-Google supported ads in apps developed for Google’s device. She said that this has a mixture of both exclusionary and exploitative harm.
Graef then discussed a way forward and suggested that competition law can do more to prevent such abuses and employ a more dynamic interpretation of ‘as efficient competitors’ in addition to considering the role of economic dependence. She then talked about platform-specific regulatory intervention where the regulator can step in to assess practices on the notion of fairness. She concluded her presentation by suggesting that a general unfairness clause can be devised as a catch-all provision and made a recommendation to refer to Vertical Block exemption under A. 101 which are relevant in targeting anticompetitive effects of restrictive practices between these undertakings.
Ben Harries followed Graef’s presentation by discussing parallels between the telecom sector and platforms. He started his presentation by highlighting that concerns about differentiated treatment by vertically integrated firms are commonplace in telecom regulation and that an ex-ante approach is utilised while regulating the telecom sector. He took the example of Telecom provider, BT and said that it is characterized by enduring market power as it has its own vertical chain at different levels. He said that Ofcom’s role is to make sure that it does not prefer its own vertical chain so that competition exists in different levels of the telecom sector such as in the supply of ducts and poles and supply of wholesale broadband.
Harries then explained why a level playing field matters and said that without Ofcom engaging in ex-ante regulation, issues such as foreclosure, negating of a rival’s competitive advantage, expropriation of benefits of a rival’s investment and entry deterrence may occur. He then described the different elements of Ofcom’s ex-ante regulation which are 1) non-discrimination remedies, 2) vertical separation commitments, and 3) price regulation. He concluded his presentation by saying that this type of regulation might not be helpful in the case of platforms since economic dependence is not accompanied by market power in their case, unlike the telecom sector where it is accompanied by market power.
In pulling together thoughts on both talks, Lyons raised questions regarding the fairness test and the EC’s new market tool to Graef and a question on vertical separation to Harries. Graef responded by saying that in practice, a proportionality test would be able to meet the notion of fairness and that differentiated treatment can be addressed in the new competition tool. Harries responded by saying that vertical separation allows downstream businesses to choose between different options.
Session 5 – Agency
Chair: Morten Hviid. Panel: Pinar Akman, Kai-Uwe Kuhn
Our final session today explores the extent to which online platforms are agents of their suppliers, and the extent to which competition law should apply to agreements between platforms and their suppliers, will be explored, with a suggestion for the reinterpretation of the single economic entity doctrine.
Pinar Akman discussed platforms from an agency perspective. Platforms such as Amazon Marketplace, Apple App Store, Uber, eBay, Booking.com, and Airbnb are agents to their suppliers. Agency is defined as a fiduciary relationship between two parties where one acts on behalf of the other. However, agreements between agent and principal are not covered by Article 101 – TFEU and thus creates a platform gap in the application of competition law in digital markets. This gap can be filled when the platform competes with the suppliers, through competitive neutrality, wherein the commercial and competitive interests of the principal and the agent are aligned. In the platform context, two markets are under scrutiny: the market on which the agent offers agency services, and the relevant market on which the agent offers the principal’s products and services to customers.
Akman acknowledges that the risk the agent is bearing is key to address the issue of platforms under competition law. Yet, under agency law, there is no assumption that control lies with the principal, although platforms neither set prices of services nor do they engage in the (consumer) results. Moreover, Akman pointed out that platforms are remunerated by commission per successful contract despite performance related risk not being borne by platforms. Finally, she concluded that competition authorities and courts should acknowledge the platform gap.
Kai-Uwe Kuhn raised concerns regarding communication problems between law and economics when competition practitioners try to understand legal arguments using economics. According to law, the distinction between institutions is based on observable characteristics whereas in economics there are no distinction between institutions if they generate the same incentives. Thus, the main question that the agency problem poses in the context of platforms is whether all incentive contracts are written to align the incentives of the different parties.
Therefore, the conflict of interest is central to the agency problem, and in any vertical arrangement, there is an economic conflict of interest. The idea of aligning incentives is to allocate risk between the parties (i.e. liquidity risk, risk in trade credit, product liability).
Kuhn stressed that fiduciary duty is relevant when not all contingencies can be specified as it completes the contract and can thereby maintain higher-powered incentives. Lastly, the modern theory of the firm is also relevant is addressing agency concerns. The firm is not seen as a residual profit stream but as an allocation of residual control rights over assets. In this new configuration, the main question is to define where competition policy stands.
And that ends a rich and packed day 2 of CCP’s conference. We look forward to you joining us for the final day tomorrow afternoon.
Day 3 (10th July)
Welcome to the final day of CCP’s annual conference. Our agenda is packed today and includes our Keynote, Mike Walker (Chief Economic Advisor, CMA) as well as three panels. Speakers will offer insight into the CMA’s Digital Advertising Report, Bundling and Content, Exclusives and Superstars. Registration remains open throughout the event and we will have a YouTube live stream running.
Keynote: Mike Walker
Mike Walker used his talk to highlight several important points in the recent CMA market study report on online platforms and digital advertising, which focuses on the cases of two big tech platforms: Google and Facebook. Google and Facebook have brought huge benefits to consumers by providing free online services like search, content, or social network. However, when these two platforms become bigger and more dominant in the market, it raises concerns about the issues of market power and lack of competition. Therefore, the interesting question to ask is whether consumer welfare is higher and online advertising price is lower in the presence of more competition. After the brief introduction, Walker discussed the first important area of the report, which is the competition in the search market. He shared a striking fact that despite being dominant in the market with 90% market share, Google still pays £1.2 billion to be the default search in Apple devices of UK users. This type of side payment helps Google maintain its market position by reducing competition in mobile search, which account for over two-thirds of general search. Walker also made the point to distinguished between general search and vertical search, which the latter only provide search results for a specific topic, for example, Booking.com in hotel search or Amazon in retail search. He then claimed that vertical search platforms have limited competitive constraint on the dominant search platform, Google.
The speaker then moved on to the second area, which is the online advertising market. The key concern here is that content providers earn a lower proportion of advertising revenue than media platforms, which may lower the content quality provided to online users. The third area is about social media market, where Facebook is the dominant players. Walker explained that although online users multi-home to different social network sites, it is likely that they always use Facebook, which shows the evidence of its market-dominant position. Facebook also has great market power as it earns significantly higher margins of online advertising revenue than other competitors. The main concern of Facebook dominance is that the strong network effects in the social media market and the lack of interoperability may have foreclosure effects on competitors and prevent potential entry. The speaker then mentioned another concern about the envelopment strategy of platform ecosystem as this would expand the market power of the big platform, which is potentially harmful to entry and innovation.
Finally, he concluded by suggesting several solutions to both current harms, and long-term harms caused by big platform dominance. To mitigate current harms, the competition authority can enforce the codes of conduct to regulate platforms with strategic market power behaviours, which is structured around three main objectives: “open choices”, “fair trading”, and “trust and transparency”. To reduce long term harm, several tools can be used as policy remedies, for instance, enforcing data portability and interoperability, or tighter merger control for big tech acquisitions.
Session 6 – Panel Discussion: The CMA Digital Advertising Report
Chair: Francine Lafontaine. Panel: Cristina Caffarra, Fabien Curto Millet, Pablo Ibanez Colomo, Gábor Koltay
Following the keynote by Mike Walker about key findings of the CMA advertising report, Francine Lafontaine introduced the panel and allotted one area of the report to be introduced by each speaker for further discussion. The 4 areas that were covered were- 1) Search advertising (introduced by Gábor Koltay), 2) Self-preferencing (introduced byFabien Curto Millet), 3) Data as an antitrust concern (introduced by Cristina Caffarra), and 4) Legal and institutional dimensions (introduced by Pablo Ibanez Colomo).
- Search advertising
Koltay began by providing admiration for the CMA Report’s broad ambit into search services and commented that in the past, such core concepts have not been so thoroughly discussed. He explained how data collection directly helps highly targeted advertising due to the exactness and precision that data provides. He emphasised that there are services for which consumers or businesses can’t pay such as in the case of Google Search or Gmail and that these markets need to be analysed as the firm has an incentive to foreclose the market to others. Caffarra agreed with Koltay that the depth of the CMA report is unprecedented but noted that other jurisdictions such as Australia have also engaged on these issues previously. Millet argued that Search is a query and location game and that search advertising has far more scope than display advertising which requires far more data to be successful. He emphasized on the need to assess what sort of data is relevant as it varies widely
2. Self-preferencing
Millet began by saying that self-preferencing is considered as a possible factor of harm in the CMA Report that it is important to discuss both how to tackle it and also if we need to tackle at all. He was apprehensive of outright bans and mentioned that even Commissioner Vestager had a cautious attitude towards self-preferencing with no approach seeming generalisable and requiring case-specific analysis. He mentioned the English Court decision of Streetmaps/ Google where there was a clear finding that search results in question were pro-competitive and that many other courts have found similar results in many jurisdictions such as Germany, Brazil and South Korea to name a few. He said that one shouldn’t take a rigid stance on this and everything needs to be judged on the basis of evidence alone. Koltay jumped in to say that self-preferencing can be tackled through competition or regulation and that vertical efficiencies and pro-competitive effects should be balanced with actions that lead to foreclosure and other anti-competitive effects. Caffarra generally agreed with Millet and opined that a blanket ban or regulation on self-preferencing needs to be designed very carefully by taking into consideration the differences in the way different platforms function. Millet disagreed on the point of business models and said that all platforms look at the same incentive and that it needn’t be a matter of consideration. Colomo emphasized on the need to look at the institutional framework to make regulation or a blanket ban effective and said that a ‘by object’ blanket ban would be inappropriate. He also added that regulation in this field provides an enormous challenge. Mike Walker opined that self-preferencing is not a bad thing and that it can lead to competition difficulties which need to be assessed carefully.
3. Data as an antitrust concern
Caffarra began by commending the depth of the CMA Report as it emphasizes on the role of data and providing consumers control of their data. She said that the Report takes note of all asymmetries and suggests to provide consumers with more control. She emphasized on the need to provide consumers with the ability to choose to provide their data in return for the platform’s service and suggested changing the policy to a default opt-out instead of the current default opt-in. She said that the CMA Report is in line with the movement happening elsewhere and contextualized it by talking about the Bundesgerichtshof’s preliminary ruling on Facebook and said that it was an extraordinary occasion as a judge supported an abuse on the basis of limiting consumer choice. She said that the exploitative concerns voiced in the CMA’s Report are not just a matter of consumer and data protection, but also include competition issues as it allows the platform to entrench competition over rivals. She said that there are issues regarding non-application of exploitative abuse principles to data markets due to the lack of a benchmark. She said that this is the very problem due to the monopoly of the platforms and the need is to create a benchmark. Koltay noted the difference in advertising before and after the emergence of data and opined that it is important to allow consumers to choose to provide data instead of taking it by force as many consumers would pay for the service instead of providing their data. Millet was apprehensive of mixing competition law with aspects outside of competition law such as consumer choice and said that there should be clear reasons why this mixing is happening and why consumers having choice may make markets work better. Walker noted that it is important to take forward the debate about privacy and competition and that privacy regulation should not be seen as damaging to competition. He said that consumer data is used to reduce competition and that firms use privacy regulation as a shield from competition. Colomo made the point that re-allocation of rents is the key concern in cases of exploitative abuse in data markets. He said that he agrees with the CMA Report that competition law can be used for this but that it puts a lot of pressure on competition authorities. He mentioned Google Shopping and Android as good examples of the difficulties faced by competition authorities as there were criticisms of the remedies and said that it will become the bread and butter of competition law with experience
4. Legal and institutional dimensions
Colomo began by saying that it is important to focus on the choice made towards a regulatory regime and expose matters with which competition law struggles the most. He mentioned that the CMA Report does a commendable job of exposing these difficulties that competition law faces in achieving these goals and also talked about the pro-competition regulatory regime mentioned in the Report. He emphasized on the need for crafting a legal test which comprises of a greater focus on remedies and remedy design with a focus on how data plays a role in the inner working of companies shaping of their business models and products. Koltay re-emphasized on the need for an institutional framework to make things work and that it is important to resolve a conflict of interest which arises as a result of competition law. Walker made the suggestion these issues cannot be left to competition law alone and that a regulatory approach is necessary. Caffarra remarked, ‘competition law is not a religion or the Catholic Church’ and that it needs to evolve as a result of novel circumstances in response to Colomo’s cautious opening statement and said that his position was not encouraging. Colomo responded by saying that institutions matter and so do the rationale behind legal doctrines. He re-emphasized the need for a legal regime that can be predicted and closed his argument by saying that even Phillip Areeda believed in competition law but did not believe in remedies that aren’t effectively handled. He finished by saying, ‘dreaming is not sufficient’. Caffarra made the final statement of the day by saying that it is a matter of trade-offs, nothing is set in stone and that everything can change
Lafontaine then brought the session to an end.
Session 7 – Bundling and Content
Chair: Amelia Fletcher. Panel: Greg Taylor, Mike Cragg
Session 8 – Exclusives and Superstars
Chair: Damien Geradin. Panel: Leonardo Madio, Eliana Garces
In joint work with Elias Carroni and Shiva Shekhar, Leonardo Madio provided insights into the effects of exclusive deals on welfare. Agents entering a platform are not homogenous. They can be identified as either atomistic or superstar. Superstars are the ones that are highly valued such as popular artists (i.e. Beyonce, Taylor Swift), popular apps (WhatsApp, Instagram) or popular broadcasts (Real Madrid, Juventus) and they have a strong bargaining power vis a vis the platform. In other words, they can make a take-it or leave-it exclusivity offer to the platform. The results show two main trade-offs: exclusivity allows more rent extraction whereas non-exclusivity enables higher audience reach. Crucially, the presence of network externalities is at the heart dynamics of the analysis.
Network externalities drive superstars to sign exclusive contracts with the platform, hinder competition and make exclusivity less likely under vertical integration. Put differently, they show that if competition is intense, consumers are more grouped through network externalities, which allows for more surplus extraction. From a welfare perspective, this encourages entry of small firms. However, exclusivity is reduced under a vertical integration between the platform and the superstar, which in turn reduces entry by small firms. Importantly, network externalities increase consumer surplus under exclusivity.
Among the policy lessons to take away, there is a pressure on price under vertical integration; banning exclusive deals can have unintended effects and the theory of harm needs to be reassessed.
In her talk, Eliana Garces highlighted the main contributions of Madio’s work. However, she pointed out that models are often calibrated in a way that reaches the intended results and that the assumptions underlying the models should be carefully justified. Garces added that, when studying firms’ entry, economic analysis’ central focus is on network externalities but entry depends significantly on consumer heterogeneity and platforms differentiation. In the context of platforms, many questions remain unanswered. For instance, it is not clear whether higher market power is associated with less quality. There is thus a gap to be filled in the literature, with the idea of creating more value without destroying the value that is already there as we are moving towards a world of data.
Competition Policy blog 