Introduction and summary
I am most grateful for the invitation from Germany’s Federal Network Agency for Electricity, Gas, Telecommunications, Post and Railway (BNetZa) to renew contact with old friends at the Body of European Regulators of Electronic Communications (BEREC), and the opportunity to renew mental engagement with the tricky regulatory issues posed by the evolving digital communications market. I am especially pleased to have the chance to tackle this precise topic - ‘Platform regulation - antitrust law versus sector-specific legislation’ - speaking as it does to both my current role as head of a competition authority, and my previous one as a telecoms regulator. Indeed when I received the invitation I did wonder whether the Bundesnetzagentur was showing an unexpectedly masochistic streak: given my new responsibilities, might I not argue that ex post competition is where it’s ‘at’ today, and ex ante sector regulation should be retired from the scene?
Rest assured, there is no need for Annegret Groebel to put on the hair shirt, nor for President Jochen Herman to bring out his funeral suit. Because in fact we do not face a binary choice between antitrust and sectoral regulation. Indeed telecoms regulation has at its heart the concept from antitrust of recognising dominance, or ‘significant market power’, and then tackling this with remedies that aim to replicate the effects of competition - and wherever possible provide the conditions for competition. Notwithstanding all the advances in electronic communications, there remain natural monopolies and externalities such as interoperability, that merit continuing ex ante regulation. So antitrust law and sectoral regulation must work in unison.
But both need to recognise the nature of the epoch we are in and change accordingly. It seems appropriate here in Bonn to remember the work of Joseph Schumpeter, Professor at the university here during the late 1920s. Schumpeter showed that great historical waves of technological innovation clump together - canals, steam, steel and electricity, mass production and now - though he died too early to see it himself - Information Technology. In each epoch, the rules of the economy tend to be best adapted to the wave that has passed, not the wave that is breaking. And it is the breaking wave that will bring with it utterly new benefits to humanity - all the way from Big Data methods of improving public health through epidemic detection to cat videos - while, of course, also presenting us with risks that have never been faced before.
Today, as regulators, we have the responsibility but also the great historical privilege of playing an influential role in the deployment throughout the economy of the latest of these defining technological eras. As regulators, we must try to minimise the inevitable mismatch between how we’ve done things before and the opportunities and risks of the new. My goal today will be to describe how both antitrust and telecoms regulation ought to change, and why I believe this period requires more ex post and less ex ante regulation.
Overall I will focus on 3 general points:
First, blanket solutions should be avoided. Instead an evidence-based assessment of potential adverse effects of specific industry features or practices should be carried out before either ex ante regulatory or ex post enforcement tools are deployed. In either case this should be closely targeted to the specific harm identified, and every care given to avoid disproportionate actions and unwelcome side-effects. In that respect, online platforms and the digital economy do not differ from any other sector: there is no need to reinvent the regulatory wheel.
Secondly, the significant risks associated with premature, broad-brush ex ante legislation or rule-making point towards a need to shift away from sector-specific regulation to ex post antitrust enforcement, which is better adapted to the period we’re in, with its fast-changing technology and evolving market reactions.
Thirdly, as regulators, policymakers, businesses and consumers, we all need to adapt our practices to harvest the benefits of the new while containing its costs and risks.
1. Blanket solutions a poor fit for the still-evolving web
The diversity of ‘online platforms’
So let’s begin this assessment by considering online platforms in broad perspective.
First, is the size of the so-called digital giants a problem in itself? From a competition perspective, the answer is no. Success in winning customers is not cause for suspicion or condemnation. And size is not equivalent to dominance. Where there is dominance, such companies must show especial care not to abuse that position, and must expect especially watchful scrutiny by the authorities. But dominance itself is not illegal.
The second broad question: is there sufficient commonality in the evolving ecosystem of the web for us to judge that certain business models ought to be subject to regulation or enforcement action? I challenge you to tell me what characteristics the following online models uniquely share: communications and social media platforms; operating systems and app stores; audiovisual and music platforms; e-commerce platforms; content platforms (itself a diverse group); search engines; payment systems; sharing platforms … and the list could go on (1 - see footnotes at the end).
Different platform characteristics will give rise to different issues, and regulation must remain case-specific if we are to minimise the risk of applying the wrong rule to a novel situation.
To give an example: if a platform is processing consumer data, one would want to be confident that it is respecting its privacy obligations to consumers. One might also need to be watchful in case it acquired an unmatchable advantage over rivals through its exclusive control over such data. But not all platforms process consumer data; and most of those that do, do not have market power; and of these a smaller group still would have the ability and incentive to abuse that power. So the analysis is situation specific.
Given the significant differences between the business models of the main digital platforms, one must be sceptical a priori about the extent to which any type of broad-brush legislation or economic regulation could provide satisfactory outcomes across such a wide variety of different situations. When the wave of innovation has stopped breaking, perhaps. But we are far from that today.
This innate scepticism is reinforced for me by recent work our authority has undertaken in digital markets.
No ‘digital one-size-fits-all’ - the need for an evidence-based approach
The Competition and Markets Authority’s (CMA) recent report on the economics of open and closed systems, which we prepared together with France’s Autorité de la Concurrence, confirmed that there is no ‘digital one size fits all’. It showed that openness is not necessarily always good for competition, nor are closed systems always bad. For example, Apple’s AppStore ‘walled garden’ approach may reassure customers with quality and consistency, while Android’s more open approach could allow for more entry and experimentation.
In our competition work on vertical search engines - ie price comparison websites (PCWs) - we have also been careful to shape our interventions to reflect particular circumstances. Four recent or ongoing CMA market investigations illustrate this: in payday loans and retail banking, we have wanted to encourage their development; in energy we wanted to understand what could make them flourish; and in motor insurance we have wanted to curtail certain artificial contractual restrictions on competition that the platforms had developed.
These examples show that blanket solutions should generally be avoided. Intervening without evidence that specific industry features or practices cause harm is putting the cart before the horse. And that rarely results in moving in the right direction.
When might online platforms give rise to economic harms?
But still, you may ask, are there not some common platform characteristics that might cause harm and that therefore can be tackled through common rules?
Online platforms exhibit fast-paced innovation and high rates of investment. However, the presence of network effects often makes it more likely that the ‘winner takes all’. Once a market has ‘tipped’, the platforms may have market power which could be used to discriminate against competitors or to the detriment of consumers and innovation (2). Competition authorities should be concerned about the appearance of market power where it is sustained over a period of time and where there are significant barriers to customers switching or ‘multi-homing’ which deter entry from more innovative or better platforms (3).
In the world of online platforms, barriers to switching may arise from a number of factors:
They can arise from contractual restrictions imposed by the online platform. Examples include certain ‘Most Favoured Nation’ clauses, which reduce the scope for differentiated prices between platforms and therefore reduce incentives to switch - as in the PCWs used in the UK private motor insurance sector cited above - or tying and exclusivity provisions.
Barriers to switching can also arise from the inability of customers to transfer their reputation or profile to a competing platform, making consumers ‘invested’ in a particular platform.
It could also be the case that a dominant platform has access to proprietary data - perhaps personal data or transaction history - that is inaccessible to rivals and could in principle create an unmatchable advantage.
So there may well be specific instances where online platforms can raise legitimate competition concerns, in particular when consumers are locked into a single unavoidable system with very limited contestability from competing systems.
However, both ex ante regulatory and ex post enforcement tools are likely to result in disproportionate actions and unwelcome side-effects if they are not carefully targeted at specific harm. There is no need to discard the competition playbook simply because platforms in the digital economy operate ‘online’. After all, offline platform markets, like newspapers, show that network effects in two-sided platform markets do not necessarily result in dominant positions and are not necessarily cause for concern in themselves. Indeed, the presence of network effects has sometimes been found to protect consumers from price increases, for example in the newspaper industry where the need to attract a large circulation for advertisers has been found to constrain the potential for increases in cover prices (4).
And as to the sustainability of perceived market power, I have previously spoken about the digital giants blowing a Schumpeterian gale through our economies - giants which, despite their size, are themselves not necessarily immune from being toppled over in this gale. MySpace and Bebo, if you remember them, serve as useful reminders of how short-lived perceived dominance can be (5).
2. Shifting emphasis of regulation from ex ante to ex post
What if, despite all the difficulties in identifying a group of businesses that can be usefully categorised and treated as ‘online platforms’, and challenges in identifying common and predictable patterns of harm, we nevertheless were to heed calls for economic regulation of digital platforms?
In terms of the timing of any intervention, I see 3 types of risk: (i) acting prematurely, (ii) inadvertently ossifying evolving market structures, and (iii) acting too late. The most significant risks, in my view, arise from premature broad-brush ex ante legislation or rule-making in markets that are still rapidly evolving.
Let me explain these risks and the policy implications flowing from them in a bit more detail.
The risk of acting prematurely
The potential costs of ex ante regulation should be carefully considered. For instance the cost of compliance with certain ex ante regulations can be substantial.
Premature ex ante regulation can not only impose substantial direct compliance costs, but can also reduce potential competition. This is of course something that many of the delegates at today’s conference - both regulators and regulated - will be very aware of in your ongoing efforts to optimise communications regulation.
So let me choose an example from outside of regulation - the European Court of Justice’s (ECJ) landmark ‘right to be forgotten’ ruling in May 2014. The ECJ found in the particular case that a person’s right to data protection could not be justified merely by the economic interest of the search engine. Since then Google has processed over 300,000 requests relating to over a million URLs. Now you may think that Google with all its financial strength can well afford the considerable additional resources needed to process all of these cases. But many smaller companies and potential entrants could not sustain these additional compliance costs, and may therefore be held back from mounting a competitive challenge.
Leaving aside costs of compliance, protecting consumers by virtue of ex ante regulation is inherently difficult in digital markets where consumer preferences evolve fast and in a less predictable manner.
It is important not to be overconfident in identifying the preferences of consumers and therefore specifying what is in their interest. For example, in the trade-off between security and convenience, most policymakers and regulators would tend to place a strong emphasis on the former, wishing to protect consumers from fraud and privacy abuses. But consumers themselves have consistently shown a strong preference for convenience. Sometimes just one less click has been enough to cause consumers to prefer one app over another, more secure one. And the analogue world can often offer a very imperfect guide to consumer preferences in digital markets, which continue to evolve apace. So even with the best intentions, the preferences that regulators ascribe to consumers may not necessarily be those consumers themselves will hold in the future (6).
The risk of ossifying evolving market structures through premature codification
If ex ante regulation is applied too early it risks protecting early innovators from a following wave of more welfare-enhancing disruption. To put it another way, if the new digital giants - once innovative firms - get entrenched in their positions and do not face credible threats, they will also tend to pass up opportunities to innovate and invest: today’s plucky innovators are tomorrow’s sleepy incumbents who’ll soon be calling for - or willingly succumbing to - regulation to protect their rents.
Where ex ante regulation is introduced, it therefore risks harming innovation by locking in existing standards and discouraging or preventing more disruptive innovations. The evolution of digital markets has been particularly difficult to predict. Recent changes include e-commerce morphing from auction-sites to more broadly embedded social media services; the rapid transformation of payment and communication systems; and a plethora of innovations in the incompletely-solved problem of search on mobile devices. But there’s one innovation we haven’t seen yet: the crystal ball informing us reliably of the impact of future innovations in digital markets. In its absence, we can’t know which ex ante interventions are free from the risk of inhibiting further welcome innovations.
Europe has adopted a particular approach to the issue of net neutrality. I respect that this received a great deal of debate from regulators and policymakers before being adopted last year (7). But even the strongest supporters of the net neutrality regime will allow that it will inhibit some business models, for example involving paid-for performance enhancement, from evolving. And some of these could have been welfare-enhancing.
The risks of getting it wrong suggests that we need to begin to shift from broad-brush ex ante regulation to ex post antitrust enforcement, which is better adapted to responding to the rapidly changing innovative markets online platforms operate in. As Director-General Johannes Laitenburger has rightly pointed out, competition law focuses on ‘specific business practices’ in digital markets. Ex post tools have the inherent advantage of being more targeted and proportionate by examining the extent to which actual harm may have occurred based on empirical evidence on a case-by-case basis.
3. Evolving the ex post enforcement toolkit: the need to adapt our practices
However, relying that bit less on ex ante regulation, and that bit more on ex post enforcement, requires competition authorities, policymakers, businesses and consumers alike to adapt their practices in a collective effort to ensure the challenges brought by online platforms are addressed effectively.
As competition authorities, we need to:
First, ensure we do not act too late. Investigations, even where litigated through the courts, should not take 10 years to complete, and arrive only when the market has changed beyond recognition. This means considering opportunities for expedited action, including interim measures to prevent harm arising while we investigate, as well as means to achieve earlier outcomes through commitments or settlements.
Second, acknowledge that there are certain, familiar antitrust concepts which may not take sufficient account of the nature of digital markets and so should not be unthinkingly transposed across to them. The ‘essential facility’ doctrine, for instance, was developed in the context of infrastructure assets that are difficult to replicate. But concepts applying to ports cannot simply be copied and pasted into the digital world, where the potential source of market power for online platforms is generally not derived from big infrastructure requirements and high fixed costs, and nor do planning restrictions intrude. Equally, a focus on revenues can ignore the true nature of the economic value provided, where this lies in customer relationships or consumer data (8).
Third, look for opportunities to test and design remedies at an early stage to ensure they work in the real world and in a cost-effective manner. Online business models tend to be flexible, constantly evolving business models and have an abundance of data that can help improve remedy design.
Fourth, provide better protection for commercial complainants. It is vital that complainants are not afraid to come to the competition authority for fear of retaliation from a dominant platform.
Fifth, we need more and earlier international joint working. We must exploit synergies in very similar cases that are simultaneously being taken across multiple jurisdictions. For a recent example, consider the family of cases brought by multiple EU competition authorities in relation to hotel online booking. And we need to try to avoid a patchwork of potentially inconsistent regulatory or enforcement approaches.
Policymakers and regulators
There are also many useful steps that policymakers and regulators may wish to consider:
First, establish a minimum set of rights, including around privacy for consumers. Clear rules in such cases can avoid disputes and distrust. They also embed more fundamental rights, relating to citizenship, which do not lend themselves to ex post economic assessment in the way that questions of market power and commercial behaviour do.
Second, set clear standards around data protection by businesses. The ECJ’s recent ruling declaring the ‘safe harbour’ data agreement invalid, leaves many companies scrambling to overhaul their internet operations with no effective regime (9). Complex digital infrastructure decisions affecting thousands of businesses and millions of consumers should ideally not be left to judges.
Third, clarify responsibilities around consumer protection. The regulatory framework should ensure that online platforms provide clear information on how they operate (10), and what their responsibilities are, so consumers can make informed choices (11).
Fourth, ensure that market analysis is alive to the growing importance of content as a key differentiator, not only in the competitive battle between telecoms operators, fibre, cable and satellite companies, but increasingly as a key customer recruitment and retention tool for the internet-based digital platforms.
Fifth, seek out opportunities for removing obstacles to cross-border trade, particularly by standardising regulatory requirements. The EC sector inquiry into e-commerce may provide opportunities here.
Finally, consider deregulation. If policymakers were to seek to avoid every hypothetical consumer harm through pre-emptive ex ante regulation, they would likely prevent many best-case scenarios entailing significant consumer benefits from ever coming about (12). Policymakers and regulators should be open to the idea that a review of existing regulation and its suitability in the context of online platforms may in certain cases actually result in a withdrawal of such regulation - creating a reasonably level playing field by ‘levelling down’ as opposed to ‘levelling up’.
The benefits brought about by certain online platforms may provide good arguments for pursuing a deregulatory approach. For instance, they can expand the range of options and information available to consumers, by facilitating reputational feedback mechanisms, thereby potentially reducing the problem of asymmetric information between producers and consumers. This could lower or even remove the need for regulation, allowing more scope for market competition to fix problems.
This holds for all of the Digital Single Market agenda. But as today’s conference agenda is mainly about electronic communications, may I say that I sincerely hope both the European Commission and BEREC will keep this deregulatory opportunity firmly in mind in the forthcoming review of the EC electronic communications framework (13). The bar for introducing new forms of communications regulation - such as has been proposed to deal with oligopoly industry structures, also to cover ‘Over The Top’ players - must be a high one (14).
At the same time, businesses also need to play their part in ensuring ex post enforcement works better than ex ante regulation. In particular, they should:
First, improve the transparency of the information available on how they operate. The amount of public information about the workings of digital platforms is low. Is it any surprise, then, that these platforms continue to arouse suspicion?
Second, take more responsibility for satisfying consumers that their legitimate concerns about privacy and data protection are being fully respected.
Thirdly, look for opportunities to grow by innovation. Much of the focus of our biggest telecoms operators has been on managing the cashflows that sustain their debts acquired through spectrum purchases and M&A (mergers and acquisitions). A lot of attention has been given to pricing structures, consolidation and cost control, and regulatory bargaining, rather than break-through technologies and services. Let me ask the distinguished panel of industry speakers who speak after me this morning - should we expect more innovation from Europe’s telcos?
Last but not least, consumers will also need to continue to adapt:
First, by being engaged and proactive. Consumers need to recognise the benefits of switching between platforms or search engines. Keeping providers on their toes can be a powerful tool we all have at our disposal as consumers. This also includes switching to a provider who is transparent and reassuring about the use of personal data - the ‘currency’ consumers use to pay for apparently free services.
Second, by providing effective feedback. A recent petition in London relating to transport regulation is an interesting example of proactive consumer engagement: more than 130,000 people signed up within a matter of days. We need consumers to continue to help guide policymakers, regulators and competition authorities on where they see their best interests.
Conclusion on antitrust law versus sector-specific legislation
To conclude. We do not face a binary choice between antitrust and sectoral regulation; instead they must complement each other. Competition authorities and communications regulators must work together to update and adapt our practice to tackle the challenges we face effectively. The costs of premature, unmeritorious interventions are likely to be very high in this sector, given the positive impact of welfare-enhancing innovations. A necessary shift towards the use of reinvigorated, ex post tools will allow for more evidence-based, and therefore more targeted and proportionate, enforcement. The digital platforms should be judged and treated according to how they behave, and how this affects consumers (15).
A Digital Single Market Strategy for Europe COM(2015).
- For example the Commission is currently carrying out investigations into Google regarding alleged abuse of dominance. It is the Commission’s preliminary view that Google is abusing a dominant position, in breach of EU antitrust rules, by systematically favouring its own comparison shopping product in its general search results pages in the European Economic Area. The Commission is concerned that users do not necessarily see the most relevant results in response to queries, to the detriment of consumers and rival comparison shopping services, as well as stifling innovation. The statement of objections alleges that Google treats and has treated more favourably, in its general search results pages, Google’s own comparison shopping service ‘Google Shopping’ and its predecessor service ‘Google Product Search’ compared to rival comparison shopping services. Google has responded to the Commission’s statement of objections.
- Barriers to customer switching or multi-homing may make it more difficult for platforms to compete on an ongoing basis. Concerns may arise in particular when barriers to switching and multi-homing have led to the emergence of a dominant platform as they may in some cases be sufficient to deter entry even from more innovative or better platforms.
- See for instance the Office of Fair Trading’s (OFT) decision on the anticipated acquisition by Northcliffe Media Limited of Topper Newspapers Limited of 16 July 2012, where the presence of network effects in two-sided market was a contributing factor to a clearance decision, in particular paragraphs 126 and 127: ‘Were Northcliffe to raise the cover price of The Post, it is likely that its circulation would fall resulting in both its fixed costs having to be met by the remaining readers or by advertisers […] due to the presence of indirect network externalities, declining circulation would make it harder for Northcliffe to raise revenue from advertisers. As a result, Northcliffe would not have the ability or incentive to reduce profitably the output of this title.’
- Alex Chisholm, ‘Giants of digital: separating the signal from the noise and the sound from the fury’, CRA Competition Conference, Brussels, 10 December 2014.
- For further discussion, see The commercial use of consumer data: report on the CMA’s call for information, CMA , June 2015.
- See the European Commission’s press release on the agreement to enshrine the principle of net neutrality into EU law.
- This is illustrated by Google paying nearly $1 billion for the crowd-sourced travel-mapping app Waze, which generated little revenue at the time. See the OFT’s merger control decision of 11 November 2013 regarding the completed acquisition by Motorola Mobility Holding (Google, Inc.) of Waze Mobile Limited, ME/6167/13. Note that in turnover-based merger control the focus on revenues means that some potentially anti-competitive mergers could fall below the regulatory radar, whereas the UK is able to examine such transactions under its share of supply jurisdictional test. A further illustration is the EC investigation into the Facebook/WhatsApp merger, where the Commission analysed to what extent potential data concentration issues could hamper competition in the online advertising market. Case No COMP/M.7217 - Facebook/ WhatsApp, 3 October 2014.
- See this month’s ruling of the European Court of Justice regarding the transatlantic Safe Harbour agreement in Maximillian Schrems v Data Protection Commissioner (Case C-362/14), see also Financial Times, 7 October 2015, ‘Europe’s costly judgment on data protection’ (£).
- For instance, the law requires consumers to be given ‘material’ information (information they require to make an informed transactional decision). Failure to do so could be a misleading omission contrary to Regulation 6 of the Consumer Protection from Unfair Trading Regulations 2008.
- The speed at which technology is changing - and the rapid emergence of novel online business models - present challenges to the application of many existing and traditional regulatory frameworks. Issues relating to online platforms encompass not just competition law, but many other regulatory regimes, including intellectual property rights, taxation, employment, cybersecurity, consumer protection and data protection.
- Adam Thierer (2014). Permissionless Innovation: The Continuing Case for Comprehensive Technological Freedom. Virginia: Mercatus Center ; Koopman, Mitchell, Thierer, ‘The Sharing Economy and Consumer Protection Regulation: The Case for Policy Change’, December 2014, Mercatus Research; see also MaryAnne M. Gobble, ‘Regulating Innovation in the New Economy’ in Research-Technology Management, March/April 2015.
Consultation on the evaluation and the review of the regulatory framework for electronic communications networks and services.
- This refers to how a number of electronic communications markets have evolved so that only a small number of multi-functional operators are now competing, offering broadband, TV, fixed and mobile telephony packages, based on some combination of fibre, cable, copper and wireless networks. Some commentators have questioned whether new ex ante powers may be needed to deal with such concentrations. However, both economic literature and empirical observations show how oligopolistic competition can be effective - one reason why there is no presumption against this form of competition in the rest of the economy. BEREC’s recent report on oligopoly analysis and regulation found that not all oligopolies raise competition issues and therefore oligopolies are not necessarily always problematic such that they require regulatory action. The report notes that oligopolistic market settings are only of concern when they contribute to a non-competitive market outcome, resulting in significant consumer harm/welfare loss, thus requiring regulatory action to address evident or potential market failures. In any event, in assessing the case for new regulation, we should not downplay the potential of existing competition tools to tackle such issues.
- I am grateful to a number of my colleagues at the CMA for contributing to this speech, in particular Nelson Jung, Paul Tregear, Robert Ryan, Tony Curzon-Price, Mike Walker and Simon Constantine.