Response to follow up questions and information on horizon scanning
We are grateful for the opportunity to have provided evidence to the Committee on 8 July 2020. Thank you for your letter of 21 July 2020 with the Committee’s follow-up questions. As well as our response to the questions in this letter, please find attached an annex with further information about the CMA’s ‘horizon scanning’ work and identified trends in digital markets, as promised during the session.
As we set out in the final report of the market study, as well as in the session with the Committee, the decision not to launch a market investigation was a balanced one.
As you helpfully recognised at the hearing, there needs to be a change of direction so that we move away from a ‘whack-a-mole’ approach to addressing these issues and take a forward-looking approach. For the same reason, our view is that, although a market investigation can produce the one-off interventions to remedy the specific problems identified in that market investigation, these are not adequate to achieve the lasting, flexible, long-term solutions to the complex range of problems we have identified, in a dynamic sector. Legislating to establish a new regulatory regime is needed.
In particular, a market investigation can only be based on the assessment of the problems in the forms that they take at the time of the investigation; and the remedies it implements can only be targeted at those problems. The experience of these sectors, however, is that business models and practices evolve and change at a rapid rate. As a result, a competition authority relying on existing powers that allow one-off interventions risks finding itself targeting concerns that are no longer the main forms of harm.
This is why the conclusion of our market study was to recommend the establishment of a new pro-competitive regulatory regime that is able to monitor these markets on an ongoing basis and agile enough to tackle the harms as they appear. A new Digital Markets Unit (DMU) of the form we have recommended would be in a position to establish up-front principles for the platforms to adhere to in conducting their business through an enforceable code of conduct to limit the exercise of their market power and a series of procompetitive interventions to address the sources of market power.
However, like the Committee, we also recognise that a new regulatory regime will take some time to be implemented, and there is the risk of continued harm in the interim. The acceptance by the Government of the Furman Review’s strategic recommendations, coupled with the commissioning of the Digital Markets Taskforce to take forward work now, has given us renewed confidence that the Government is committed to taking forward pro-competition reforms in this area that are consistent with the findings and aims of our market study. In particular, the Government confirmed it intends to set up a Digital Markets Unit that will enforce a code of conduct for platforms with Strategic Market Status. We do not consider it would be helpful to the process of designing such legislation if we were to run a market investigation in parallel.
How long would a market investigation into the open display advertising market take?
Although our completed market study would provide an important starting point for any market investigation, such a market investigation could not simply move quickly to implement the remedies that have been identified in our market study.
The Enterprise Act 2002 sets out the statutory timeframe for the CMA to undertake a market investigation. As I set out at the session, this is 18 months with a further six- month implementation window for any remedies. In addition, there is the possibility to extend the initial 18 months period by a further six months for special reasons. An appeal against the findings of a market investigation could add a further 1-2 years to the timeframe.
The CMA does its utmost to ensure that cases can be delivered as quickly as possible within the statutory timeframes. There are, however, some unavoidable constraints on this that result from the statutory framework.
Although the decision on whether to make a market investigation reference is one for the CMA Board, conduct of such an investigation, and the ultimate decision, is made by a group of independent panel members. The panel members would of course be considerably informed by the findings of the market study, but the law requires that they must reach their own decision on the nature of any problems – whether there is an ‘adverse effect on competition’ – and similarly to reach their own decisions on effective and proportionate remedies.
Furthermore, because market investigations can lead to interventionist measures being imposed on companies, they are subject to strict and demanding procedural requirements to ensure that decision-making is transparent and affected parties are fully consulted as thinking develops. In practice, this means that the process must be broken down into a number of stages with adequate opportunity for companies to provide their responses, and for those responses to be considered by the independent panel.
If remedies are being contemplated that are highly interventionist or require careful design, then the affected parties must be provided with adequate time to submit their views and respond to all arguments. Decisions within a market investigation are appealable to the specialist Competition Appeal Tribunal. A number of past judgments by the Tribunal have emphasised the importance of giving full opportunity to the parties to respond to all aspects of the case, and for the CMA to have considered each of these. If an appeal is successful, then the decision will be remitted back to the CMA for further consideration, increasing the time for any remedies to be implemented.
An appeal to the Competition Appeal Tribunal, even if ultimately unsuccessful, also has the impact of delaying the implementation of any remedies. Past experience indicates that this can be expected to add some 1-2 years to the implementation timeframe. Other things being equal, there is therefore an incentive for parties to appeal against any decision arising out of a market investigation in relation to remedies that they are opposed to, if only to delay implementation of such remedies.
The CMA, however, also recognises the concerns about the time required for legislation and is committed to doing what we can to enable this to happen quickly.
The Digital Markets Taskforce was established in March 2020 and is led by the CMA with members from Ofcom and the ICO (Information Commissioner’s Office). The Taskforce will provide advice to the Government on the design and implementation of pro-competition measures in digital markets, to be overseen by the proposed DMU (Digital Markets Unit).
This advice will include the types of remedies which could be available to the DMU, and the basis on which the DMU could implement these. It will build on the recommendations of our market study in respect of Google and Facebook but will also consider broader digital markets outside of digital advertising. The Taskforce is focused on providing advice to the Government on the framework for regulation.
We are also taking forward important work on the interplay between competition policy and privacy, and in this context, we have launched the Digital Regulation Cooperation Forum together with Ofcom and the ICO. We also continue to work closely with other governmental and regulatory bodies, such as the Financial Conduct Authority, the Intellectual Property Office and the Centre for Data Ethics and Innovation. We are engaging closely and actively with our international counterparts to influence the debate and support progress on this matter on a global scale.
Importantly, the decision not to make a market investigation reference can be revisited. After the work of the Taskforce has concluded, we will be in a position to assess whether the actions being taken by the Government are sufficient to address the full range of issues we have identified, or whether further action by the CMA, such as for example in form of a market investigation, might still be required.
Would the CMA consider enforcement action in the open display advertising market and, if so, why is it not considering a market investigation?
Taking enforcement action and making a market investigation reference are two separate functions of the CMA that are independent of each other and have different statutory requirements attached to them.
Our market study report was absolutely clear that enforcement action, under the competition law prohibitions in the Competition Act 1998 (the prohibition against anti- competitive agreements and the prohibition against unilateral abuse of a dominant market position), has not been ruled out. Indeed, as the report set out, the CMA is actively considering possible enforcement cases in the digital sector, drawing on the work of the market study. The CMA stands ready to take possible enforcement action, where appropriate, where it identifies evidence of evidence of digital platforms exploiting their market power or otherwise engaging in anti-competitive practices.
The Committee asked about the CMA’s response to the suggestion that the CMA ‘is reluctant to take enforcement action because it is overly concerned about losing in court’. We believe our track record of work proves otherwise.
In relation to enforcement under competition law, o ve r the p ast financial ye a r a lon e the CMA h a s ta ke n action resulting in over £56 million of fines imposed and 10 director disqualifications. In recent years, we have taken enforcement action against large multinationals in the pharmaceutical sector, Pfizer and GlaxoSmithKline, in cases which have involved around £130 million of fines, and appeals up to the Court of Appeal and the EU Court of Justice. As at today, we are having to fight appeals in four separate cases; this belies any suggestion of timidity or a reluctance to risk court action on our part.
And in relation to consumer protection law enforcement, we have taken action in many areas including: tackling misleading online reviews on Facebook and eBay, the way information is presented on hotel booking sites such as Booking and Expedia, and the way that ticket resale websites such as viagogo operate.
The decision not to make a market investigation reference at the end of the market study is not one that is amenable to a quantified impact assessment. It is necessarily based on a judgement taking into account a wide range of factors and uncertainties both in terms of timescale and outcomes. As explained above, the relevant comparison is between (i) a market investigation and (ii) legislation to establish a new pro-competition regulatory regime. It is possible that the former could lead to certain measures being put in place more quickly than would be achievable through the legislative route. However, the difference in implementation timescales is unlikely to be as large as some observers have speculated. Moreover, measures that could be put in place following a market investigation would be more limited than those available through a new pro-competition regulatory framework underpinned by primary legislation.
For the reasons set out earlier, our view at the time of publishing the market study was that it was better to work with the Government to progress a new pro- competition regulatory framework. We nonetheless recognise there is inevitable uncertainty around timescales. For that reason, we have explicitly kept the option of a market investigation on the table in case things change and it were to look like legislation would take substantially longer than is currently planned.
As I flagged at the hearing, there is also a risk that a market investigation could lead to further delay on a legislative framework being put in place. The risk here is not a legal one, but rather, is due to the inherent challenges of running two processes looking at the same things in parallel, and the pressures to wait until a market investigation has reached its conclusions before starting the legislative process.
Although the decision on a market investigation reference was not meaningfully quantifiable, the market study includes a detailed examination of the scale of detriment arising from current practices in the markets and attempts to quantify this where it is feasible to do so meaningfully.
I apologise if I gave the Committee the impression that we were not concerned with harmful impacts on consumers resulting from the platforms’ impact on publishers. That was not my intention at all. As I said at the hearing, our starting point is to focus on the consumer impact, in line with our statutory duty. As we set out in Chapter 6 of the report, the detriment to consumers arising from these platforms takes a number of forms, and the impact on publishers is an important element of it.
Indeed, we felt it was particularly important for the market study to shed light on the scale of the impacts on publishers in this opaque market. Specifically, we were able to quantify the main factors determining Google’s and Facebook’s impact on publishers’ ability to monetise their services, including: the extent to which publishers are dependent on Google and Facebook for traffic to their sites (on average 38% in 2019); and the proportion of advertising revenues taken by Google and other ad tech intermediaries through which publishers sell advertising (at least 35%). We also showed that publishers have much less access to data than Google and Facebook have, and that the ability to target advertising significantly increases its value to advertisers (by around 70%).
Together, these all indicate a significant detriment. As set out in the market study report, we expect that stronger competition would drive greater innovation and put downward pressure on fees, which would directly benefit publishers.
This assessment is of course based on the most recent available figures. The digital advertising market has evolved very rapidly over recent years, in terms of both scale and how it functions. It is expected to continue to go through significant change; and the practices of the major platforms are also expected to change. For example, during the market study Google announced its intention to limit third-party access to cookies from its Chrome browser. To our minds, this reinforces the points made above that a new regulatory regime is needed in order to move away from a backward-looking ‘whack-a-mole’ approach, and towards one that is by construct forward-looking, flexible, and able to deal swiftly with concerns as they arise.
The competition landscape is not currently equipped to deal adequately with the number and complexity of the issues now arising in these markets. A one-off intervention from a market investigation would not necessarily address all issues and, for this reason, our view – which is consistent with the Furman recommendations – is that an ongoing regulatory remit is required that can monitor the market on an ongoing basis which is agile enough to identify and tackle the harms as they appear. An ex ante pro-competition regulatory regime to regulate the activities of online platforms is better geared towards addressing the range of concerns that we found and preventing future harm from occurring than our current powers that apply retrospectively. This new regulatory approach, implemented through legislation, is needed to produce the best outcomes for consumers, publishers, advertisers and other market participants.
I hope you find this letter provides full responses to the questions that you have raised.
ANNEX: FOLLOW UP TO ORAL EVIDENCE
Importance of horizon-scanning
The CMA’s Digital Ma rkets S trate g y emphasised the importance of understanding and keeping pace with digital market transformation as well as having systems and structures in place to anticipate potential future market failures. The CMA has invested significantly to ensure it has the necessary capability to do so. In 2019, as part of its digital strategy, the CMA established its Data, Technology and Analytics Unit (DaTA Unit). This CMA DaTA Unit is a mixture of data engineers, data scientists, technology experts, and behavioural scientists. Within the CMA’s DaTA Unit, we have a dedicated team focused on understanding and explaining data and technology markets. Part of this role includes developing a systematic function to scan the horizon for new and emerging innovations (such as new products, processes, business models and services) and for data and technology trends. The CMA’s DaTA Unit will use ‘traditional’ scanning methods, such as manual data gathering and expert interviews, but also explore the use of data science tools for automated scanning. The Unit has recently undertaken a short horizon-scanning project as part of the work of the CMA’s COVID-19 Taskforce, the high-level findings of which are below.
The CMA’s DaTA Unit has also developed close links with other organisations working in these areas. These include the newly formed Regulatory Horizons Council, the Centre for Data Ethics and Innovation, UK Research and Innovation, the Government Office for Science, and other regulatory horizon-scanning functions.
In addition, the CMA undertakes a significant amount of intelligence gathering and horizon-scanning across its range of functions as part of its normal work. This is done through a variety of means including our extensive contact with stakeholders across industry, government and academia, our engagement in broader arenas such as conferences and meetings, and intelligence collection through our investigations. The market study involved extremely extensive engagement and consultation both nationally and internationally with a huge range of different stakeholders. This form of wide engagement enables the CMA to be on the front foot in identifying competition and consumer issues early to reduce potential harms and maximise the undoubted opportunities that digital markets present.
The Digital Markets Taskforce was commissioned by the Government in March 2020 to consider the practical application of the potential pro-competitive measures set out by the Furman review (the Digital Competition Expert Panel). It is led by the CMA with members from Ofcom and the ICO (Information Commissioner’s Office), and is also considering future trends and risks as part of its remit. The CMA’s DaTA Unit is a core contributor to this work. In addition to horizon-scanning, the CMA’s DaTA Unit is applying broader futures tools, such as scenario planning, to the Digital Markets Taskforce programme. Scenario planning generates multiple versions of the future using projected trends. These can then be used to stress test policy or interventions against various future outcomes. We expect to have further insights from this later in the year. This work will not only explore further the future developments in digital markets but also ensure that the Digital Markets Taskforce provides recommendations and advice to the Government that are as future-proof as possible.
Finally, the recently established Dig ita l Reg u latio n Coo p e ratio n Fo rum builds on the CMA’s existing close links with Ofcom and the ICO and includes as one of its objectives the anticipation of future developments by developing a shared understanding of emerging digital trends. An important aspect of this is to share knowledge and expertise across authorities, recognising that we are together stronger in terms of pooled resources and expertise. This is a key driver of this initiative.
Key public policy issues for digital technology
At the evidence session we identified a number of high-level public policy concerns based on the CMA’s horizon scanning so far. These included the development of AI and algorithmic systems which may lead to the risks of collusion and also the risks of distorting consumer choice through forms of self-preferencing, price- or product- steering and the exploitation of consumer biases. We also identified the relationship between privacy and competition policy as an important public policy issue which is likely to become increasingly significant with the development of the Internet of Things and an increasingly networked environment.
At a broader thematic level, as mentioned, the CMA has undertaken a specific internal horizon-scanning project as part of the work of our COVID-19 Taskforce which looked at the potential long-term implications of the pandemic. High-level trends for digital markets which have emerged from this exercise include the following:
• The risks, as more businesses move online, that those without online access may face higher prices and less choice in bricks and mortar retail. New technologies and new business models may maximise economic benefit for online businesses but could potentially leave consumers vulnerable, for example as to the awareness and exercise of online rights. Algorithmic pricing concerns may be exacerbated.
• Digital markets may tend to greater concentration as large businesses acquire small firms or small firms fail. Companies will also continue to invest in new technologies or bolster their positions in existing and emerging markets such as cloud computing. Big technology companies may also extend their reach into supply chain ecosystems. This may result in an increase in the power and reach of existing digital technology companies.
• Approaches to regulation may need to be re-examined to balance innovation and business while protecting consumers. This may include the use of innovative methods and technologies such as data science techniques for spotting emerging market issues. It may also include the use of technology to improve regulatory outcomes.
• Government is likely to increasingly partner with technology firms to provide services to citizens, such as tax updates via social media, or e-commerce companies delivering medical supplies.
• Greater international engagement and cooperation is likely to be needed, both as a result of the international nature of online trade but also wider issues like climate change.
It is worth noting that the trends above are taken from the specific context of the CMA’s COVID-19 work rather than from a broad scanning process; the CMA has a much wider work programme on digital markets and technologies. Moreover, these trends are not firm predictions but set out some ways in which the future might evolve in what is an uncertain and complex environment.
 We note that this timetable has already been condensed to the extent possible by the Enterprise and Regulatory Reform Act 2013. Prior to that legislative reform, the initial period was 24 months.
 This includes competition and consumer protection enforcement, merger control, market studies and investigation references, policy and advocacy work and research and other work undertaken by the Office of the Chief Economic Adviser.
 https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/ file/896827/Digital_Regulation_Cooperation_Forum.pdf