Digital Markets: using our existing tools and emerging thoughts on a new regime
Andrea Coscelli, Chief Executive of the CMA, delivers a keynote speech to the annual Fordham Competition Law Institute conference on the CMA’s existing tools and a new regime to support competition and innovation in digital markets.
Good morning and it’s a pleasure to be with you this morning (albeit virtually from my office in London).
I plan to focus on the CMA’s work in relation to digital markets, and in particular, our emerging views in relation to the design of a new ex-ante pro-competition approach to address some of the harms in digital markets we see. Furthermore, I’ll talk about how the CMA is maximising the use of its existing toolkit to deal with these problems.
Why is competition in digital markets important now?
No-one could have anticipated the situation many Governments around the world now find themselves in as a result of this pandemic. And clearly economic recovery is now dominating the agenda in most jurisdictions.
Competition is vital to supporting economic growth. Competition spurs dynamic innovation, which is crucial in fuelling improvements in productivity and growth. Digital markets are widely recognised as being one of the most dynamic and innovative areas of most economies, with huge potential for value creation. It is imperative that we ensure these markets continue to drive innovation.
The need for a new approach to promote competition and innovation in digital markets
Earlier in the year we published the findings and recommendations arising from our year-long market study into online platforms and digital advertising. This piece of work considered how advertising revenue drives the business model of major platforms. Our work found that large multinational online platforms such as Google and Facebook now have a central role in the digital advertising ecosystem and have developed such unassailable market positions that rivals can no longer compete on equal terms. In particular, their large user base is a source of market power, leading to weak competition in search and social media.
This matters to consumers who receive reduced innovation and choice but also will be paying higher prices for goods and services when producers pass the high cost of advertising onto consumers. We found that Google’s prices are around 30% to 40% higher than Bing when comparing like-for-like search terms on desktop and mobile.
Furthermore, we are concerned the largest platforms are increasingly acting as a brake on innovation, setting the terms of competition in a way that tips the balance in their own favour and undermining the business models of new entrants and potential challengers alike.
Our key recommendation was that a new regulatory regime is required in the UK to ensure these markets continue to deliver benefits to consumers, businesses and the economy as a whole.
For me, the case for regulation is clearly made. We have firms with very substantial and enduring market power, protected by strong network effects, who are able to leverage into adjacent markets, and who can engage in envelopment strategies that further protect their core sources of market power. These firms are active across many markets and in many cases also act as an important access point to customers, giving them a strategic position. They can use this to exploit the many consumers and businesses who rely on them and act to exclude or quash innovative competitors. Existing tools are clearly not sufficient to address these potential harms. For me, regulation seems to be the absolute best way at this stage to ensure digital markets continue to thrive – and deliver the wider benefits we value so highly. Structural solutions might be needed in some cases if regulation is not effective - similar recommendations are included in the recent report by the US Judiciary Antitrust Subcommittee.
In the course of our work we heard from many companies who told us that the significant market power of some online platforms poses an existential threat to their businesses. We believe that, without reform, existing market dynamics in these industries will mean that the next great innovation cannot emerge to impact our lives in the way that previous advances in digital markets have done in the past.
As the Furman Review had done previously, we recommended that within the new regime a ‘Digital Markets Unit’ should be established with the ability to enforce a code of conduct to ensure that platforms with ‘Strategic Market Status (SMS)’, like Google and Facebook, do not engage in exploitative or exclusionary practices, or practices likely to reduce trust and transparency, and to impose fines if necessary.
The DMU would also have the ability to impose ‘pro-competition interventions’ to drive greater competition and innovation in digital advertising markets. These include requiring Google to open up its click and query data to rival search engines to allow them to improve their algorithms so they can properly compete. It would also include requiring Facebook to increase its interoperability with competing social media platforms.
The CMA is now building on these recommendations in its work leading a Digital Markets Taskforce, which was commissioned by the UK Government earlier this year to provide advice on digital regulation. Alongside the code of conduct and the pro-competition interventions, as part of our advice we are also considering a third pillar which would form part of the new SMS regime - a parallel merger regime for acquisitions by companies with Strategic Market Status. We are considering whether the evidence supports a policy justification for such a regime, based on the particular features of digital markets that increase the risks of consumer harm arising from acquisitions by particularly powerful companies, and the heightened risks of underenforcement. In particular, we are analysing the extent to which such concerns cannot be fully addressed under the standard mergers regime.
As I will mention later on in this speech, the CMA’s approach to digital mergers has already evolved considerably. It is against this backdrop that we are considering the merits and characteristics of a special parallel regime.
Our current thinking is that any special regime would have its own jurisdictional and substantive tests. In relation to the jurisdictional test, in contrast with the UK’s standard voluntary mergers regime, companies subject to the special regime could be required to notify all transactions to the CMA, subject to certain limited exemptions. In relation to the substantive assessment, competition concerns could be assessed under the standard ‘substantial lessening of competition’ test. However, the inherent uncertainty that often characterises developments in these digital markets, combined with the increased risks of consumer harm where the acquirer already has Strategic Market Status, may justify the use of a more cautious standard of proof than the ‘balance of probabilities’ threshold under the standard regime. The regime could also accommodate a separate assessment of non-competition concerns such as data protection.
Why regulation?
The origins of economic regulation in Europe arose in the regulation of previous state-owned monopolies like utility companies. Here regulation was less about competition but more about controlling outcomes, like prices, to ensure consumers received a fair deal. However, over the past 30 years or so, economic regulation has become far more focused on promoting competition and innovation through opening up access to markets. For example, in telecoms, Ofcom’s work to open up access to parts of BT’s existing network has enabled innovative competitors to provide more advanced broadband services, and is now helping to stimulate investment in new fibre networks from challenger firms.
There are also some brilliant examples of ‘regulation for innovation’ from the financial services sector in the UK. For example, the Financial Conduct Authority (FCA) has taken steps to try and ensure regulation does not act as a barrier to new innovation. It runs an ‘Innovation hub’ to support new and innovative businesses navigate regulatory requirements. It also runs a ‘sandbox’ to help businesses trial new products and services in a safe environment.
The new regime
Now this doesn’t mean that some of the existing fundamentals go out the window. The framework for antitrust is grounded in economic analysis, is well established and well understood. We believe that any new regime needs to be grounded in this framework.
For example, the notion of market power and the potential for abuse of this must still squarely factor in our consideration of where and when intervention is necessary.
Similarly, the existing case law around anti-competitive practices will still be important in guiding future consideration as to the effects of actions, such as self-preferencing, which we recognise, in some circumstances may have pro-competitive benefits.
But it does mean that we need to examine the accepted wisdom carefully, and not be afraid to change course, do things differently and try new approaches where necessary.
For example, it means looking hard at procedures and ensuring these strike the right balance between giving appropriate rights of defence to parties, without being exploited as a tool to frustrate or waylay. Similarly, we need to ensure we at the CMA act with appropriate evidence and due diligence, but equally recognise the pace many of these markets move at.
It also means looking hard at the skills and capability we need to appropriately monitor these markets, investing more in our ability to collect and interrogate data. At the CMA this is an area we have already been investing heavily in through our Data, Technology and Analytics Unit which specialises in data science, engineering, behavioural science and data and technology insights expertise.
It means not being afraid to try using new tools and approaches. Our work on Open Banking, for example, has demonstrated the potential that opening up access to data can have in driving innovation. Just last week we announced that users of products enabled by Open Banking topped 2 million – demonstrating clear demand for these services which have been enabled by this intervention.
Lastly, it means understanding that our work is likely to be far more wide-reaching than just competition. Digital markets are increasingly interconnected – action in relation to competition will never just occur in a vacuum, but increasingly have consequences for work in relation to privacy, online harms, intellectual property and consumer protection. We will need to work more closely than ever before with our partners in other agencies – both domestically and internationally – to tackle these problems together. This is something we are already doing in the UK, where the CMA is working with Ofcom (our communications regulator) and the Information Commissioner’s Office (our privacy regulator) through the Digital Regulation Cooperation Forum to support regulatory coordination in digital markets.
The path to this regime
In the UK, the path to establishing this new regulatory regime will likely still have some way to run beyond the delivery of our advice to Government at the end of the year. Clearly, we are keen to see progress in a timely manner and stand ready to assist in any way that we can.
However, in the meantime, the CMA is focused on using its existing powers to the maximum extent.
Using our existing tools: consumer enforcement
Over the past few years, a large part of our consumer protection work has been focused on building trust in online markets. We’ve examined the practices of the largest cloud storage providers, tackled unfair practices by online gambling firms and pursued social media influencers who’ve tried to conceal paid advertising.
We’ve also taken a number of enforcement cases in the online segments of the travel and entertainment sectors, helping to clean up secondary ticket websites, hotel booking sites and car hire booking platforms and intermediaries. A good example of our consumer enforcement in an online context is our work on fake reviews.
Last year we found evidence of a thriving online marketplace for fake and misleading online reviews. This sort of activity fundamentally distorts the competitive process. It is estimated that over three-quarters of UK internet users consider online reviews when choosing what to buy. Billions of pounds of people’s spending is influenced by reviews every year. Fake and misleading reviews not only lead to people making poorly informed choices and buying the wrong products, but more widely undermine consumers’ trust and confidence when shopping online.
We have taken action in relation to this issue and secured commitments from Facebook, Instagram and eBay that they would take down the illegal content that we have identified and put in place systems to prevent this content reappearing.
We have taken this work into a second phase and are now investigating a number of major websites to consider whether they are doing enough to protect consumers from reading fake reviews on their sites. We’ve involved our expanding team of data and behavioural specialists in this work, helping to identify key patterns of behavior and likely indicators of fake content. We’ll be holding these businesses to account if we discover that they are not effectively policing their sites or addressing loopholes that allow fake reviews to appear.
Using our existing tools: competition enforcement
Competition enforcement is another key part of the CMA’s toolkit. This is an area we expect to be increasingly active in over the coming years, particularly pending a new regulatory regime like that described above.
While the UK was a member of the European Union, many of the biggest digital enforcement cases were undertaken on our behalf by the European Commission. From January, the CMA will be able to start to investigate the conduct that most affects UK consumers and we are actively considering potential enforcement cases in the digital sector. Given the cross-border nature of these markets, we are looking forward to working in close collaboration with our international partners.
Using our existing tools: mergers
Digital mergers are another key area of focus for the CMA and we have been working hard to develop our substantive assessment in light of our increased understanding of digital markets and the learnings from recent expert reports (such as the Lear report (PDF, 4MB) we commissioned on past mergers). Key elements of the developments in our substantive assessment include:
-
Analysis of a broad range of theories of harm, including those related to the loss of innovation and access to data (e.g. PayPal/iZettle and Google/Looker).
-
Consideration of dynamic counterfactuals such as the development of new products or services (e.g. Sabre/Farelogix and Amazon/Deliveroo).
-
Analysis of the valuation model and rationale for the merger to gain insights into the acquirer’s plans and expectations for the target (e.g. the Sabre/Farelogix and Google/Looker cases, as well as VISA/Plaid and Salesforce/Tableau).
-
Assessment of the impact of the merger on both sides of the market in digital platform mergers, taking account of the differences in customers’ options on each side of the market (e.g. viagogo/StubHub and Taboola/Outbrain).
We are continuing to monitor closely mergers in digital markets and to initiate investigations ourselves where appropriate when parties choose not to notify ahead of completion (e.g. Facebook/GIPHY).
We are also making full use of our evidence-gathering powers when assessing digital mergers. Internal documents are often a key source of evidence, as historic evidence such as market shares and switching data may be less informative of future competition in dynamic markets.
The potential importance of internal documents was brought into stark relief, for instance, by the recent disclosure by the US House Judiciary antitrust subcommittee of Mark Zuckerberg’s emails from 2012, which highlighted that neutralising the competitive threat was a key driver for Facebook’s acquisition of Instagram. If our predecessor agency, the Office of Fair Trading, had had access to these emails at the time of its review, it is not clear that this merger would have been cleared at Phase 1 without an in-depth investigation.
Our document review capabilities have been significantly enhanced since then and we may now require the production of a large volume of documents in appropriate cases, even at Phase 1. We are also increasingly considering making use of our compulsory information gathering powers to hold witness interviews (e.g. Amazon/Deliveroo).
We are also carrying out a major update to our substantive Merger Assessment Guidelines to reflect our current approach to merger review. These updates are broader than digital mergers, but digital markets are one of the most significant areas of development since the last review of the Guidelines in 2010.
Our DaTA unit
Lastly, earlier on I mentioned the work of our Data, Technology and Analytics team. Compared to staff members we have traditionally hired, this team has markedly different qualifications – e.g. PhDs in applied mathematics or physics or, in our new Behavioural Hub, psychology – as well as a range of backgrounds.
One area this team are increasingly focused on is scrutinising how digital businesses use algorithms and how this could negatively impact competition and consumers – something that will become increasingly important with the ever-increasing availability of large datasets and, given cloud computing, the ever increasing use of machine learning and artificial intelligence algorithms. We believe it is not acceptable for firms not to be able to explain the outputs of their algorithms. We plan to publish a paper on potential harms arising from algorithms in the coming months and to invite collaboration with firms, researchers and stakeholders on methods for authorities to investigate, mitigate, and remedy any harms. As part of this work we will be considering how requirements for auditability and explainability of algorithms might work in practice.
The need for continued international cooperation
I wanted to end with a word on international cooperation. The international nature of these ‘borderless markets’ means that it is essential for competition authorities to work with each other to share knowledge and expertise, intelligence and where possible, to tackle these problems together. Many of the problems we observe are common across jurisdictions. The more we can come to a common view of these problems and work together to deal with them, the more efficient and effective this is likely to be.
Closing remarks
Thank you again for listening. I look forward to working with many of you as our work in relation to digital markets continues and as we progress towards a new approach.