Dear readers, I am delighted to present you with this month’s guest article by Renato Nazzini, Professor of Law at King’s College London. All the best, Thibault Schrepel
Mergers in the Digital Economy
For a few years now, legal scholars and policy-makers have been questioning whether existing legal frameworks are fit to address perceived anti-competitive effects of mergers in the digital economy. The digitalisation of the economy is changing the ‘competition game’ due to the success of a new generation of firms that overcome traditional operational constraints by exploiting significant scale economies and demand-related efficiencies, cut across industries, drive new value, growth, and innovation. In recent years, digital platforms and AI-driven firms conquered and transformed entire markets, providing ecosystems in which traditional industries and new services became increasingly interconnected. It can be assumed, for the purposes of this article, that digital markets are concentrated1OECD Ecoscope, Competition in the digital age, 31st March 2019; F. Calvino, C. Criscuolo (2019), “Business dynamics and digitalisation,” OECD Science, Technology and Industry Policy Papers, No. 62; S. Calligaris, C. Criscuolo & L. Marcolin (2018), “Mark-ups in the digital era,” OECD Science, Technology and Industry Working Papers, No. 2018/10. and competition in the digital world may occur, in essence in three ways: (a) competition for the market, which is less frequent, e.g. Facebook overtaking Myspace as the leading social network in the first half of the noughties; (b) competition between ecosystems, e.g. competition between Google and Apple in smartphone ecosystems; (c) competition within ecosystems, e.g. competition between several payment methods on Amazon or eBay.
A particularly heated debate has developed in relation to mergers. There is now a clear trend in scholarship and policy thinking in competition authorities, not to mention political views and, to an extent, public opinion fuelled by media hype (and, some argue, their hostility towards big tech) and political agendas, that looks with great skepticism at any acquisition by big tech firms.2There is, of course, no definition of who “big tech” companies are. Google/Alphabet, Apple, Meta/Facebook, and Amazon are generally included, but Microsoft, eBay, and Netflix would also make excellent candidates. And the list could get longer. In this article, I have deliberately kept the definition open and the terminology suitably vague. As a result, proposals have emerged with the common denominator of making it easier for a competition authority to block a merger, when the acquirer is a big tech company, either by changing the substantive test or by reversing the burden of proof or by lowering the standard of proof.
In this article, I ask whether such proposals are a good idea and, assuming that there might be a problem in the digital economy, whether there are better alternatives.
2. Overview and criticism of the current proposals
The proposals that have been put forward can be distinguished into three categories: (1) proposals to change the substantive test; (2) proposals to reverse the burden of proof; (3) proposals to change the standard of proof.
Starting with proposals to change the substantive test, in the United States there have been calls for changing the threshold for challenging a merger in digital markets from a ‘substantial’ to a ‘material’ lessening of competition. A test of ‘material lessening of competition’ is not really about lowering the standard of proof, but it effectively changes the substantive test, requiring a lower threshold of anti-competitive effect. This ‘material’ anti-competitive effect would still have to be proven to some standard, for example balance of probabilities. The attack on legal certainty is less serious here than with standards such as ‘increasing the risk’ of harm to competition. But the problem remains the same: since ‘material’ would catch many more mergers than the current standard, are all these mergers going to be prohibited? Or would a competition authority, by interpreting the inherently vague ‘materiality’ test, be allowed to pick and choose which mergers to block, other than those that are clearly pro-competitive and harmless?
In the United Kingdom, the Furman Report recommended the substitution of the ‘balance of probabilities’ test with what it called the ‘balance of harms approach’. While under the ‘balance of probabilities test’, competition authorities must only look at the probability of harm occurring, under the (fortunately) now-defunct ‘balance of harms approach’, competition authorities must also look at the likely scale of such harm. The new test would therefore have boosted antitrust enforcement by leading competition authorities to intervene every time that the scale of the possible harm is high albeit the likelihood of such harm is low and lower than the likelihood required by the current ‘balance of probabilities’ test. Mergers more likely than not to be efficient, pro-competitive, or neutral could have been blocked under such a test. While presented as a reform of the standard of proof, this test would have altered the substantive test as it would have been, in reality, a form of vaguely structured discretion to block mergers unlikely to cause harm, if the harm was perceived to be high. Rightly, the CMA itself rejected this approach and current UK Government proposals do not envisage any changes to the merger regime.
As regards the burden of proof, one of the proposals that has acquired momentum is to introduce a rebuttable presumption against acquisitions by firms having a ‘strategic market status or bottleneck power’. According to the proposal, where such companies engage in acquisitions, the merging parties bear the burden of providing evidence that either the merger does not raise any significant competitive issue, or that expected efficiencies justify the acquisition.3See M. Motta & M. Peitz, Big Tech Mergers, supra; the EU, Stigler, and ACCC Reports; and the French Senate, Proposition de loi visant à garantir le libre choix du consommateur dans le cyberspace. Where they are unable to do so, the acquisition should be blocked. This deceptively simple proposal would effectively result in wild and open-ended discretion by competition authorities to block any mergers. There is currently no guidance, case law, or established practice on how the merging parties can prove that a merger is pro-competitive or, at least, not anti-competitive. Efficiency arguments hardly ever succeed in merger control and it is difficult, and extremely uncertain, for the merging parties to anticipate competition authorities’ novel and untested theories of harm. And if the merging parties do not know the case against them and on what evidence it is based, how can they demonstrate that the merger is not harmful? In practice, such a proposal would force merging entities to fight against a ‘black hole’ where every theory is, in principle, possible. And it is important to remind ourselves that we are talking here about the complex mergers that fall into a grey area, not of the mergers that are clearly harmless or clearly anti-competitive. In this grey area, to require the parties, and not the prosecuting authority, to come up with theories of harm and evidence to disprove them risks open-ended, unstructured, uncertain merger reviews that could go either way at every turn and in which the competition authority would be allowed to sit on the fence and argue that the ‘required legal standard’, whatever it may be, has not been met with respect to any conceivable theory of harm. Nothing could be further away from the current understanding of legal certainty.
Turning now to the standard of proof, in the United States, several scholars4Joint Response to the House Judiciary Committee on the State of Antitrust Law and Implications for Protecting Competition in Digital Markets (Jul 2020). advised the US Congress to consider lessening the current legal standard (which requires, in general, a showing that competitive harm is more likely than not) that plaintiffs must meet to successfully challenge the acquisition of nascent competitors with a new one which simply requires “showing that the challenged conduct increases the risk of competitive harm”.5Ibid. In the United Kingdom, the Government, on the advice of the CMA, consulted on a proposal to lower the standard of proof to block an acquisition by an undertaking designated as having strategic market status in phase 2 from the balance of probabilities to a realistic prospect that the merger would result in substantial lessening of competition.6A new pro-competition regime for digital markets (CP 489, July 2021) paras 187 – 191, available at <https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1003913/Digital_Competition_Consultation_v2.pdf>. The proposal has, however, been abandoned and the Government no longer proposes to make any changes to the substantive test for intervention in merger control.7A new pro-competition regime for digital markets – Government response to consultation (CP 657, 6 May 2022) https://www.gov.uk/government/consultations/a-new-pro-competition-regime-for-digital-markets/outcome/a-new-pro-competition-regime-for-digital-markets-government-response-to-consultation#part-7-strategic-market-status-merger-reform.
Such new formulations of the ‘standard of proof’ are inherently uncertain because of the lack of precision of the language adopted. ‘Increasing the risk’ is a standard under which a competition authority could, in practice, block any merger as long as it shows that it may have a possible, however minimal, anti-competitive effect. The lack of certainty of this test lies in the wide spectrum of cases that it covers. Only mergers that are clearly pro-competitive or have no conceivable anti-competitive effects would be safe. All the others would be up for grabs. But which ones will be blocked? Surely, not all mergers that have a possible anti-competitive effect, however minimal and however unlikely. This would be absurd. A competition authority would, thus, have a wide discretion to pick and choose which mergers to block without any ex-ante certainty for merging firms. The now-abandoned proposal to lower the prohibition standard in the United Kingdom to a ‘realistic prospect’ of competitive harm would have caused exactly the same problems.
3. Principled objections to lowering the prohibition standards for big tech acquisitions
In addition to the legal certainty and administrability problems highlighted above, there are three fundamental problems with all the proposals discussed in the previous section. First, the even higher discretion of competition authorities to block mergers may have significant, unintended consequences. Secondly, the proposals in question would be an abandonment of the presumption that mergers that cannot be proven to be anti-competitive must be permitted, on which the current merger control regimes are based. Thirdly, the evidence that economic harm has been caused by big tech mergers and that the current legal framework is not suitable to deal with the problem is, at best, thin.
First of all, the proposals under review would effectively (i) provide competition authorities with a wider and wild enforcement discretion, thus exposing them to political pressures and risks of ‘regulatory capture’; (ii) condemn, deter, or discourage socially desirable mergers8K.W. Wong-Ervin & J. Harkrider, Assessing The New DOJ, FTC Vertical Merger Guidelines (July 2020). since companies may decide to pursue those deals that enjoy greater political consensus instead of those driving positive synergies; (iii) serve as a tax on transactions;9K.W. Wong-Ervin & J. Moore, Acquisitions of Potential Competitors: The U.S. Approach and Calls for Reform, CLPD, 2020. (iv) harm competition by reducing or stalling the numbers of acquisitions companies may consider given the notable risks and expenditures that each M&A operation requires;10K.W.Wong-Ervin, et al., The Risks of Radicalism: Exacerbating Harms from Type I Errors, CPI (Apr. 2020). (iv) discourage innovation either by raising barriers to exit for nascent and venture capitalist firms or by short-circuiting the integration of pioneering new technologies into existing products; (v) distort international trade and the international market of capitals as powerful States or political blocs such as the EU may favour or discourage certain acquisitions, based on their national interests. Such consequences are heightened when competition authorities can block mergers without effective and timely judicial review or when their proceedings are regularly abandoned due to timing issues.
Secondly, there is an even more fundamental problem with these proposals. They would do away with the principle that mergers are lawful economic activity unless demonstrable harm can be established. The value of this principle is not dependent on the view one takes on neo-liberal economic theories or how concentrated this or that market is. Its value is much more fundamental in that it reflects one of the foundational, constitutionally protected, freedoms upon which most modern liberal-democratic economies are based: freedom of enterprise, enshrined, as “freedom to conduct business”, in Article 16 of the EU Charter, but very much part of the constitutional fabric of most democratic systems. Of course, the freedom to conduct business, as any other individual freedom, can and is significantly limited in post-modern democracies. However, the basic proposition remains that, except in defined circumstances in which harm to society can be established, people are free to go about their own business undisturbed, establish companies, joint ventures, and partnerships, buy and sell what is their property and create value for their customers and their shareholders, in compliance with generally applicable law. There are, of course, remarkably strict limits to this freedom already, and often for excellent reasons, but blanket rules that presume widespread and neutral economic activity such as transferring control or interests over businesses effectively unlawful would undermine the very foundation, not so much of our economy, but of our way of understanding freedom from, and within, the State.
Thirdly, and finally, the evidence that economic harm has been caused by mergers and that the current legal framework is not suitable to deal with the problem is, at best, rather thin and certainly insufficient to trigger the draconian reforms that are proposed. Evidence of increased concentration is not evidence that the current levels of concentration have been caused by allowing mergers that should have been blocked. It is quite possible, and, indeed, widely accepted, that the current level of concentration has been brought about by the very significant economies of scale and network effects that characterise the new digital space, so that the firms with the best business model, commercial foresight, or innovation strategy have emerged as dominant players in ecosystems that, almost invariably, they contributed to creating. Nor is it mere “acquisition counting” evidence of any problem. A study prepared by the economic consultancy Lear for the CMA reported that, based on information publicly available, between 2008 and 2018 Google acquired 168 companies, Facebook acquired 71 companies and Amazon acquired 60 companies. The study then analysed the transactions in question and reached three main findings, which are purely descriptive: (1) the volume of transactions is significant; (2) most transactions would seem to have a non-horizontal nature, though with notable exceptions; (3) targets are typically very young firms.11Lear, Ex-post Assessment of Merger Control Decisions in Digital Markets – Final Report (9 May 2019), paras 1.46 – 1.68, <https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/803576/CMA_past_digital_mergers_GOV.UK_version.pdf> While such findings are informative, there is today a tendency to jump from the descriptive data (number of transactions, type of target) to the conclusion that such acquisitions, or at least some of them, must have been anti-competitive. One does not need to be the most sophisticated practitioner of the Galilean method to understand the fallacy of this approach. Finally, the assessment of certain mergers such as Facebook/WhatsApp, Facebook/Instagram or Google/Double Click with the benefit of hindsight is frankly rather speculative and it seems hard to believe that if these mergers had been blocked, Facebook would no longer be the leading social media network or Google the leading general search engine. But let us assume, for the sake of the argument, that these mergers should have been blocked on the theory that the targets were going to become a credible competitive constraint on the acquirers. My contention here is that the law at the time was such that the mergers could have been blocked. At the risk of over-simplifying, US, UK, and EU law, at the material time, allowed blocking a merger if the acquisition would have resulted in the removal of a credible competitive constraint, even only a future constraint, provided, of course, that there was credible evidence to this effect. The evidence simply was not there. As regards Facebook/Instagram, it was later revealed, during a US House of Representatives’ Antitrust Subcommittee’s hearing on antitrust issues in tech in 2020, that there were internal documents, at the highest management level, suggesting that the acquisition of Instagram could have also been motivated by the objective of neutralising a nascent competitor. This evidence was never tested in the context of a merger review procedure or in court but, again, let us assume, for the sake of the argument, that it would have led to a prohibition in Facebook/Instagram. The solution to this problem, as will be discussed in due course, is not to change the substantive test, or the burden or standard of proof in mergers by big tech, but to strengthen the information-gathering powers of competition authorities and the sanctions for failure to disclose relevant information, including, of course, internal documents, which now feature very prominently in the assessment of mergers.
4. Conclusions and recommendations
While the reforms proposals discussed above are, in my view, poorly evidenced and highly problematic from a legal and policy perspective, I do not deny that there may be difficulties in assessing mergers in digital markets given the increased complexity of reconstructing the counterfactual in new and rapidly evolving markets to the current legal standards. However, if these standards are difficult to meet, ‘it is for good reason’.12K.W. Wong-Ervin & J. Moore, supra. Solutions should go in the direction of supporting competition authorities in meeting such standards, rather than replacing them with others that allow blocking mergers without sufficient evidence of competitive harm.
4.1. Enhanced in-house expertise
One of the major factors affecting competition authorities’ effectiveness has been their lack of expertise to understand the complexity and dynamics of digital markets.13See Lear, Ex-post Assessment of Merger Control Decisions in Digital Markets, supra; W. Kovacic, “Roads Not Taken: The Federal Trade Commission and Google”, Le Concurrentialiste, 2020. Similarly, the CMA itself in its response to the Furman Review, explicitly acknowledged that “decisions by our predecessor organisations did not, in some cases (such as Facebook’s acquisition of Instagram), fully consider important evidence that could have provided greater insight into how the markets at issue were likely to evolve in future” (CMA, The CMA’s response to the Digital Competition Expert Panel Final Report). Therefore, the recent trend according to which some competition authorities are building in-house expertise to help them navigate high-tech markets and understand their nature and dynamics is to be welcomed.
The CMA, for example, seems to be going in the right direction with the creation of a new DaTa Unit, where it has hired tech engineers, data scientists, behavioral scientists, and technology strategic insight advisers. These new experts will help shaping the new enforcement approaches needed by the CMA to deal with digital competition issues.
Similarly, the Australian ‘Digital Platforms Inquiry’ endorses the creation of a “specialist digital platforms branch” within the Australian Competition and Consumer Commission accountable for, among other things, “proactively monitoring and investigating instances of potentially anti-competitive conduct and conduct causing consumer harm by digital platforms, which impact consumers, advertisers or other business users (including news media businesses).”14See ACCC Australian Competition and Consumer Commission, Digital Platforms Inquiry, supra, p. 31.
4.2. Strengthened evidence-gathering powers
Since meeting existing legal standards might be difficult, it is important to ensure that competition authorities have adequate evidence-gathering powers. While requests for information or questionnaires are usually an effective modality to collect evidence, it could be worth considering, as suggested by the Lear Report, to equip competition authorities with dawn raids and document seizure powers also in merger review, given that careful acquirers may rarely leave smoking-gun evidence stated in written documents.15Lear Report, supra. Such powers, additionally, could also be used to gather information from potential alternative acquirers where appropriate.
Furthermore, certain competition authorities do not have the power to issue mandatory requests for information to the parties or third parties. It goes without saying that it would be important to give such powers to competition authorities so that they could gather reliable information across the board from all useful sources.
Along the same lines, competition authorities could be given the power to carry out mandatory interviews of the authors of internal documents or even any person who may have relevant information about the merger.
Finally, effective and deterrent sanctions should be available, and applied in practice, if corporations and individuals responsible to provide information to competition authorities provide false or misleading information or fail to provide information in the context of a merger review, while ensuring, of course, that the required safeguards in terms of legal certainty and due process in the enactment and enforcement of criminal or quasi-criminal offences are met.
4.3. Monitoring and ongoing analysis
While there is currently, in my view, not sufficient evidence of generalised harm caused by any acquisitions by big tech firms, this is not to say that more evidence and rigorous analysis would not be welcome. Therefore, the proposals in the EU Digital Markets Act and in the UK Government’s consultation response on a pro-competition regime for digital markets to require designated firms to report certain mergers to competition authorities even if such mergers do not meet the notification thresholds are, in principle, welcome. However, it is of paramount importance, as recognised by the UK Government’s response, that only transactions that have a nexus to the jurisdiction and meet certain minimum size-related criteria should be reportable, to avoid disproportionate burdens on businesses. It is also of utmost importance that the data so gathered, over time, is rigorously reviewed not only by competition authorities and their consultants but also, with the customary confidentiality safeguards, by the academic community, to allow for a more informed, evidence-based debate as to whether harm caused by acquisitions by big tech is really such that the current legal framework is incapable of preventing.
Citation: Renato Nazzini, Mergers in the Digital Economy, Network Law Rev., June 6, 2022