Sensible Tech Regulation in the EU
This blog discusses the ongoing deliberations surrounding the EU Digital Services Act ('DSA') and Digital Markets Act ('DMA') and the implications for innovation following the initiatives from the European Commission (the 'Commission'). This goes to the core of the tensions we are trying to address in thinking about a regulatory system. These regulatory initiatives and impact on innovation were debated during two round table discussion hosted on 13 October and 5 November 2020 with speakers from the Commission, economic consultancies, academics, competition lawyers, data protection and privacy lawyers and independent researchers. Another initiative from the Commission is the New Competition Tool ('NCT'), which, if it comes into force would be used on a case-by-case basis, similar to the CMA's market investigation powers - but very different in that it would apply to 'digital markets' only. It is understood that the NCT is supposed to address similar problems as the Prohibitions and Obligations in the proposed DMA. Given this overlap, this blog does not consider this tool in any detail.
The digital economy and the impact on competition
The discussion of the need for regulation followed a global debate on the digital economy and the impact on competition. This debate was ignited due to a general belief of systemic falling across a number of markets and the acknowledgement that antitrust enforcement can only do so much. There is a conception (rightly or wrongly) that antitrust enforcement is not delivering in terms of speed, remedies and a reluctance to focus on consumer exploitation. Antitrust cases can take up to 10 years, including court proceedings, which appears rather slow for dynamic and fast moving technology markets. While interim measures could form part of the solution, interim measures do not - like enforcement action - look at the root cause like contestability of the market and interoperability measures. Interim measures are a useful tool, but not the panacea, as they focus on specific conduct rather than the market structure. The remedies imposed following antitrust enforcement in Europe have made no difference in the market. For example, in the Android case, a choice remedy was imposed to tackle default settings. The practicality of the remedy was to provide a choice screen assuming that an alternative was able to appear. To appear on the screen, it was necessary to participate in an auction conducted by Google and it would be incumbent on consumers to open the choice screen. One interpretation is that the Commission appears to be too timid to mandate remedies that would work. For example in the markets that have already tipped like in social media, it may be necessary to consider structural remedies or other remedies like interoperability. But it may be necessary to go further like unwinding the merger with Instagram to create two big effective competitors in the market place. Moreover, a structural remedy may be reasonable around adtech where Google has a lot of the market power. There are other remedies for example (i) cloning of software and technology; (ii) licencing to create more competitors; and (iii) bans on certain contract terms. In the US there appears to be some consensus that structural remedies are a good way forward due to the proven difficulty in overseeing conduct remedies. For example, in big vertical transactions such as Ticketmaster/Live Nation, there were a lot of antitrust concerns at the time, and the conduct remedy has not worked. Thus, the DOJ have had to go back to investigate the very tactics that parties predicted they were going to engage in at the time.
A number of reports have been written on the digital economy and the impact on competition. In the UK, there has been the CMA's report on 'Online platforms and digital advertising market study' and the report 'Unlocking Digital Competition' delivered by an expert panel chaired by Professor Furman. In the EU, the report entitled 'Competition policy for the digital era' drafted by a panel of experts at the request of the Executive Vice President of the European Commission Margrethe Vestager. In the US, the report entitled 'Investigation of Competition in Digital Markets' delivered by the Subcommittee on Antitrust of the House Judiciary Committee and the report prepared by the Stigler Committee at University of Chicago Booth School of Business on 'Digital Platforms'. In Australia, the Australian Competition & Consumer Commission issued its report on 'Digital platforms inquiry'. The consensus from the majority of the reports is that the broader structural reform cannot be delivered by antitrust enforcement only. There is a need for ex ante regulation to promote competition in certain sectors dominated by few big technology platforms. Thus, we are no longer talking about whether we should have regulation but rather what form it should take. The debate has refocused with the Commission's announcement of the forthcoming ex ante instruments, which will complement current ex post antitrust enforcement. Ex ante regulation is not a solution to all problems in the market, so the Commission will continue to enforce existing antitrust law. Unlike many of the reports mentioned, the US House Judiciary Committee does not suggest regulation, but focuses more on resuscitating antitrust enforcement. There appears to be opposition to regulation in the US, as the conception is that regulation has never done anyone any good. There is ideological baggage against regulated industries in the US. In the EU the position is different, but it is worth asking whether regulation will achieve the stated objective of creating new EU competitors and achieving digital 'tech sovereignty' of which the Commission speaks.
While there is a feverish impulse to push forward with regulatory initiatives in Europe, the appetite for regulation in Europe is not universal and there appear to be some scepticism around a regulatory approach from some Member States. Thus, it may not be an easy process to enact regulation, as it needs to be debated and agreed by the European Coucil, European Parliament and in trilogue before it is agreed. Moreover, there is still a lot of uncertainty about what it is going to look like. Would it be like telecoms regulation, which is implemented by the national regulator? While telecoms relies on data and network like technology platforms, the telecoms sector is very different in that it is static and coherent, whereas technology platforms operate within a non-coherent and dynamic sector. Moreover, technology platforms have vastly different formats and business models. Regulation needs to be designed to take this account, so it will not be possible to have a single rule that fits all akin to utility regulation. Regulation would need to be sensitive towards these differences. When regulating the digital sector, it is important to decide whom to regulate. There are two options: (i) regulate a large number and a broad variety of companies that are active with a certain business model in a certain sector; or (ii) asymmetric regulation which target a few companies. The latter option requires you to define those whom you want to regulate, so there is an addressee of the law.
The leaked DMA impact assessment indicates the plan is to focus on the following services: (i) online intermediation services; (ii) online search engines; (iii) operating systems and (iv) cloud services. Hopefully, the regulation will be drafted in a more nuanced way than presented in the leaked documents. For example in relation to online search engines, it would be relevant to ask what search is, as you can break it down crawling, indexing and querying. You need to ask what the key features of each of those represent and where the barriers to entry are. Relying on the chosen services from the leaked document, one could reasonably ask whether these are the services where consumers are directly harmed in terms of higher prices, lack of innovation and lower quality products. The objective of competition law is consumer welfare, but consumers are not really seen as relevant actors in this sphere. This is one of the blind spots in competition law. In drafting the regulation it is important to deal with this blind spot of competition law - the consumer must be brought into centre stage. Antitrust enforcement is reluctant to tackle exploitative abuse where consumers are directly harmed, which is extremely relevant in two sided markets. Regulators have been reluctant to go after this kind of exploitation. Thus far, enforcement action under Article 102(a) TFEU has mainly been confined to excessive pricing, but this interpretation of exploitation is too narrow. Inequality of bargaining power and the ability of some technology platforms to create 'give or take' options, so consumers have no choice, but to accept T&C without knowing what is happening to their data is an example of exploitation. Any practice where the platform deprives consumers from having a choice could potentially be seen as exploitative without the need to define a benchmark. In situations where the platform do not give a choice to consumers or if it forces consumers to give their data to use the services, as opposed to giving consumers a choice of either giving their data or pay to use the service. Unfair contract terms could be considered both under Articles 101 and 102(a) TFEU. There is a lot of talk about consumer welfare, but very little about specific consumers in individual cases. A recent exemption is the German Facebook ruling where the Federal Supreme Court emphasised the need to take the consumer onboard as a sovereign market actor.
While none of the regulatory instruments has been published yet, it is understood that the regulation is going to be asymmetric by applying to gatekeeping platforms only and contain a set of prohibitions and obligations around (i) data practices; (ii) self-preferencing and (iii) tying and bundling. Behaviour can be ambivalent, so it will not be easy to set out a clear list of prohibitions ex ante. Self-preferencing by a dominant company can be problematic, but how is self-preferencing defined? It can be a very subtle e.g. in the Google shopping case even the shade of a colour was discussed. Moreover, it will be necessary to define gatekeeper. Will it be a narrow definition of a few companies within the services mentioned or a broader definition considering gatekeepers across the digital economy? If a company is defined as a gatekeeper, how often will this gatekeeping status be reviewed? Would the gatekeeping status cover the entire company including an autonomous subsidiary? What would be the process to identify the services with gatekeeper status? When would the gatekeeper status end? It is important in designing these instruments that we preserve companies' incentive to innovate.
Another blind spot in competition law is the role of data. We have a data driven economy, but what role does data play? It is clear that data is important for privacy and data protection, but less clear what role it plays in competition law. Is the speed at which data is collected important? What is the marginal value of data? What is the data in question? How is it collected? How is it used? Is there a particular characteristic of data that makes it more valuable? Is it disaggregated? Does the value of data decline over time? Is there a quality element to the data? If so, what does high quality data looks like? The volume and complexity of data and the resources needed to process the data. Are there any pro-competitive justifications for exclusive access to the data? Is it possible to replicate the data? Are there any important features of the data? Some argue that data would need to be essential to be a barrier to entry - this is not clear, as data might contribute to market power already below the high threshold of being essential. Does the platform have a certain data advantage that lead to certain competitive edge? Positive network effects due to data can lead to better product features, which can create an additional reinforcement of a data feedback loop on top of the feedback loop created by network effects. Competitive relevant data is an important aspect in the market power assessment. Many of the features that create competition problems in terms of barriers to entry and expansion and network effects also create great efficiencies.
A platform can be present in more than one market i.e. across a number of markets. A presence across complementary markets enables platforms to build valuable datasets, which enables them - amongst other things - to offer highly targeted advertising services. So platforms that are present in multiple markets are able to leverage their strength from one market to another and thereby force new entrants to enter in several markets simultaneously, which makes entry more costly and risky and subsequently make entry less likely to occur. Being present in multiple markets will also give rise to efficiency such as those arising from vertical integration or from increased competition where the digital platform is the new entrant challenging the incumbent. These correspondent efficiencies must not be ignored. While it is important not to stifle innovation, it is equally important to look at the market conditions we are faced with at the moment. We have huge companies with very stable market share over a long period of time. Same side network effects that arise from the accumulation of user data by an incumbent search engine can act as a barrier to entry and expansion in the market, but also improve the relevance of the search results - the greater the number of search queries received the quicker it is able to update and improve the relevance of the search results. Thus, the greater the number of users using a search engine, the more data the search engine has to improve the relevance for the search results or the greater the ability to improve the quality of its service. Thus, the more valuable it is likely to be to a given user. Newcomers, who do not have this data, will have difficulty providing quality search from the beginning and put competitive pressure on Google.
Only time will tell what these regulatory initiatives will look like, but if the leaked documents are anything to go by then the suggested regulatory initiatives go well beyond what the Furman expert panel suggested in their report 'Unlocking Digital Competition'. It is also worth a reminder that the regulatory initiatives must be consistent with the rest of the European competition law framework.
Dr Liza Lovdahl Gormsen
Senior Research Fellow in Competition Law and Director of the Competition Law Forum, BIICL
Join the conversation
No comments have been added to this blog entry.