WIREDGOV

A speech delivered yesterday by Andrea Coscelli to the OECD/G7 conference on competition and the digital economy.

Like many agencies around the world, the CMA has had a growing focus on digital commerce, reflecting changes in the economy and consumer behaviour. Much of our recent casework has had a strong digital flavour, and our policy and advocacy work is increasing in this area too.

But the effectiveness of merger control in digital markets has been the subject of particular discussion and attention recently and I want to focus on that today. Over the last decade, Amazon, Apple, Facebook, Google, and Microsoft combined – the so called GAFAM quintet – have made over 400 acquisitions globally, with more than half of these – close to 250 – just in the most recent 5 years. Some of these acquisitions have had exceptionally high valuations.

However only a handful of these mergers have been scrutinised by competition authorities, and none have been blocked. As an economist, these statistics naturally lead me to question whether we as competition authorities have got the balance right. Is it right that across all 400 of these acquisitions, there has not been a single prohibition? On this basis, is it possible to argue that we’ve correctly balanced the risks of under- and over-enforcement?

Global attention

These are the questions that many Governments and competition experts across the world are trying to grapple with. And if there has been underenforcement – what is behind this? Is it a reflection of limitations or gaps in the toolbox of competition authorities or should we simply reassess how we use the tools we have?

In the UK, the Digital Competition Expert Panel, chaired by Professor Jason Furman, thought there had been underenforcement and considered whether changes to the competition framework are needed to face the economic challenges posed by digital markets. It made a number of recommendations, including to strengthen the mergers regime. In the EU, the report from the special advisers to Commissioner Vestager, which considered the future challenges of digitisation for competition policy, also suggested changes to the approach for substantive assessment of mergers. And in the US, the Committee for the Study of Digital Platforms, chaired by Professor Fiona Scott Morton recently recommended more scrutiny of mergers involving so-called ‘bottleneck firms.’

The work of these groups is an important contribution to our thinking (and our current approach) in this area.

The CMA’s review

At the CMA we have also been looking hard at these questions. Last year we commissioned an independent study aimed at evaluating past merger decisions in the digital sector taken by the CMA’s predecessor organisations, the Office of Fair Trading and the Competition Commission, including in cases such as Facebook/Instagram and Google/Waze.

The objective of the study was to review the theories of harm typically pursued by competition authorities in relation to these types of mergers and how these have been evaluated – as well as considering whether the decision the authority came to was reasonable, based on the evidence available at the time, and, given the market evolution following the merger, whether with the benefit of hindsight it has led to a detrimental outcome.

I’d like to take this opportunity to set out some of my key reflections, based not only on the outcome of this review but also on our experience in recent cases.

To begin with, I don’t think the evidence suggests, at least in the UK, that there is some fatal flaw or gap in the overarching regime. The CMA’s merger control tools are and remain, in the main, fit-for-purpose.

In relation to jurisdiction, we have, in the UK a threshold involving 2 alternative tests. If the turnover-based test is not met, we can also consider the parties’ combined share of supply, and exercise jurisdiction if this exceeds 25% and any kind of increment in share is brought about by the deal. This is a flexible test which, in practice, has meant that the CMA has consistently been able to exert jurisdiction over transactions in digital markets, for example where the turnover of the target was limited, but the value of the deal was high.

In relation to the substantive assessment, application of the current regime undoubtedly raises some challenges. Post-merger counterfactuals can be difficult to assess in new and rapidly-evolving markets. Similarly, some traditional forms of substantive analysis, focused on price effects, may fail to capture other metrics of competition effectively, including quality, privacy and innovation.

Nevertheless, my reflection, based on my initial consideration of some of the potential solutions tabled to date, is that there are no ‘neat fixes’ to these challenges. And there is a risk that reinventing our entire approach leads to more harm than good. But that doesn’t mean that there shouldn’t be some evolution in our approach. Today’s study of past digital mergers highlights that there are incremental steps competition authorities can take to improve our ability to assess these mergers.

The first area is in how we define the counterfactual - what we think would have happened, absent the merger occurring. The review we commissioned found the average age of acquisition targets is 4 years old or younger in nearly 60% of cases. The median age of Google’s targets is 4 years, and Facebook’s 2.5 years. There are considerable difficulties in understanding the implications for competition of the acquisition of such young firms, given it is very difficult to predict how they could develop.

The review suggested that the time-frame of 2 years typically used by us to assess the counterfactual may be somewhat limiting and should be extended, as even in the fast-moving digital world, becoming successful is likely to take somewhat longer. The review also suggested that, if the target is also of interest to entities other than the acquirer, an alternative transaction could become the relevant counterfactual.

More broadly, the review also put forward that competition authorities should be willing to accept more uncertainty in their assessment of the counterfactual. Any assessment of future developments, no matter how well informed, is likely to contain a significant degree of uncertainty. We should therefore be cautious in concluding that that the absence of the same kind of compelling evidence we might have about near-term market developments should necessarily provide grounds for clearance.

The second step we can take is to enrich the information set we rely on when dealing with mergers in digital markets.

The review suggested the use of dawn raids to uncover valuable evidence such as the future plans of the target and whether the acquirer perceived the target as a threat. This is not entirely unprecedented, with the European Commission having carried out dawn raids at the premises of merging parties in 2 cases to date, albeit predominantly where gun-jumping was suspected. While there are questions around whether dawn raids of this type should ever be necessary, this does underline the ever-increasing importance of the merging businesses’ internal documents to our assessments, and the importance of taking strong enforcement action where merging parties provide incomplete or misleading information in response to our requests.

The Lear report also suggested placing more focus on the transaction value to better understand the relevance of the transaction – if the price paid by the acquirer seems hard to explain based on current or likely future earnings, we should scrutinise the rationale for the acquisition with particular rigour and consider, in particular, whether the purchase price could reflect the benefit of killing off emerging competition.

The third step highlighted by the review is to gain a better baseline understanding of key markets in the digital sector, and in particular the likely entry strategies. This is key to understanding when acquisitions may reflect the result of a ‘succumb or be quashed’ threat, offer the most promising path to the commercial development and use of the target’s technology, or an essential exit strategy for investors. This will require competition authorities to understand better the investors’ perspective on these deals.

I mentioned earlier my view that CMA’s merger control tools are and remain, in the main, fit-for-purpose. Now I don’t say this blithely. It is something I, and many of my colleagues at the CMA, have spent a great deal of time considering. And I am not so naïve as to think this will necessarily remain the case. Clearly, we will need to continue to monitor this area closely and consider whether we have all the tools that we need to protect consumers effectively.

In particular, I think it will be important to consider whether there might need to be some form of closer scrutiny for acquisitions by particularly powerful companies. In concentrated markets with very significant barriers to entry and expansion, where competition largely exists ‘for’ rather than ‘within’ the market, the elimination of even a very small or nascent competitor could remove an important source of competition. In such markets, it could be that any entrant with a credible strategy and route to funding is worth protecting.

Putting this into practice

At the CMA we have already begun applying some of these steps in our casework. In particular, in several recent cases we have considered the need to use a dynamic counterfactual, considering not only what would have happened absent the merger occurring based on the current state of competition, but also based on how the market is likely to evolve.

We recently conducted an in-depth investigation of the acquisition by PayPal of the FinTech company iZettle. The payment services industry in which these parties are active is a fast-moving and dynamic market, and iZettle was a relatively new and innovative player. We therefore assessed the merger against a dynamic counterfactual considering the likely future situation of PayPal and iZettle absent the merger occurring.

PayPal submitted that the rationale for the merger was to combine 2 complementary product offerings, in particular PayPal’s online payment service solutions with iZettle’s in-store/offline product offerings to create a multi-channel solution. In considering the counterfactual, we concluded that, absent the merger, PayPal would likely have been a stronger competitor than was currently the case, having a strong incentive to broaden its offline offering and a range of options at its disposal to do so. We also found, however, that iZettle would likely have focused on developing its existing offline services, rather than developing the online presence it would have needed to become a strong multi-channel competitor.

Having conducted a detailed investigation, we have provisionally found that the merger did not raise competition concerns. We provisionally concluded that, notwithstanding a more strongly competing PayPal in the counterfactual, sufficient competitive constraints will remain on the merged company from other developing providers of offline payment services, such as SumUp and Square, as well as traditional providers of payment services such as Barclaycard and Worldpay. The merger was also unlikely to lead to the elimination of a nascent competitor in the multi-channel space, as iZettle was, taking into account the merger counterfactual, ultimately unlikely to significantly challenge Paypal here.

In our recent in-depth investigation of the intended acquisition of Clearscore, a FinTech firm, by Experian, we used a range of evidence to assess the likely competitive impact of the merger. Both parties provide free credit scores to users and match users to credit products via their digital platforms. Our approach was significantly influenced by the dynamic nature of the market.

As part of our assessment we reviewed internal documents, examined the expansion plans of the merger parties and their rivals, sent questionnaires and spoke to third parties, assessed evidence on user behaviour, and considered evidence as to the likely impact of technological and regulatory developments. Our merger review in this case covered a number of theories of harm including whether, after the merger, the quality of the parties’ free products would be worse compared to what they would be if the merger didn’t take place. This could be due to less investment in new product features and functionalities, a slower pace of development, or less effort to improve customers’ journeys and experiences. After we provisionally prohibited the merger, the parties decided to abandon the transaction.

The merger demonstrates the CMA’s ability to assess a non-price theory of harm in a dynamic, digital market using current laws and its usual substantive assessment frameworks.

Evolution not revolution

The key point I would like to make today it is that we should be thinking evolution, not revolution, of merger assessment tools to deal effectively with mergers in the digital economy.

With this in mind, alongside the publication of the study today, we are also launching a call for views, with the intention of informing an update of our Merger Assessment Guidelines. These Guidelines have not been updated since 2010 and there has been substantial evolution and growth in digital markets since this time. The call for views seeks input on specific areas which could be updated to reflect the lessons of the most recent economic literature, as well as to provide greater clarity about how the CMA will deal with particular issues arising in digital markets.

We are keen to engage with fellow competition authorities as well as those in the wider competition and indeed financial, investment and venture capitalist communities, as part of this process.

Focus on the digital economy

As I said earlier, this work on mergers is part of the CMA’s broader portfolio of work on digital commerce – which is a key theme for the CMA, as set out in our recent annual plan. Alongside mergers, we are also focusing our attention on developing our thinking on the need for reform in other areas, for example in relation to our competition and consumer tools. Many of the reforms we believe are needed here are unique to the UK regime, however others, such as strengthening our information gathering powers are likely to be of interest more broadly.

On the subject of reform, we are also working to develop our views on the design of ex-ante regulation of digital platforms, building on the various ideas put forward in some of the reports I mentioned earlier, most notably Professor Jason Furman and the UK Digital Competition Expert Panel. We are keen to ensure that any regulation is informed by evidence so that it is well-designed and targets the specific problems identified, as well as works in tandem with existing competition and consumer tools.

However, the majority of our focus continues to remain on taking effective action on substantive concerns in digital markets using our existing powers and tools. Over the past few years the CMA has undertaken significant consumer enforcement work focused on digital markets. This includes: our investigation into secondary ticketing websites to ensure that better information is given to consumers about tickets being resold on their platform – we continue to vigorously pursue this with viagogo; our action against several of the biggest online hotel booking sites about practices that might mislead consumers and prevent them from getting a fair deal, which lead to 6 of the biggest operators giving formal commitments to change their business practices; and our investigation into social media stars not declaring their commercial relationships resulting in 16 influential celebrities, with millions of followers between them, committing to say clearly if they have been paid or received any gifts or loans of products which they endorse.

And to support our ongoing work, we are also investing in building-up our own internal expertise. Our Data, Technology and Analytics unit is now up and running. This unit of around 15 people, comprised mainly of data scientists, is key to strengthening the CMA’s ability to understand how firms are using data, what their machine learning and artificial intelligence algorithms are doing, the consequences of this and what action we need to take.

We also plan to engage more closely over the year ahead with digital businesses, investors and venture capitals to develop our understanding of key business models, strategies and practices in the digital economy, and what this means for competition and consumer policy.

Lastly, across all areas of the CMA’s work, co-ordination and collaboration internationally is essential. The digital economy is global, so we will continue to work with our international counterparts through forums such as today, to share best practice and develop a common approach to issues in relation to digital markets.

Thank you for your time this morning and I look forward to continuing this debate and discussion through the course of the day.