Examination of witnesses
Professor Pinar Akman, Dr Orla Lynskey and Dr Nicolo Zingales.
Q83 The Chairman: I welcome our witnesses to the House of Lords inquiry into regulation of the internet. Today’s session is being broadcast online and a transcript will be taken. Our witnesses today are competition law experts. We are very grateful to you for taking the time to give evidence to the Committee. Could you briefly introduce yourselves and tell us a bit about your background?
Professor Pinar Akman: My Lord Chairman, thank you for the invitation. I am honoured to be here. I am a professor of law specialising in competition law at the University of Leeds, where I am also director of the Centre for Business Law and Practice. I work in areas of digital technology and the application of competition law to digital platforms. My background is in the prohibition of abuse of dominance in particular, and I have authored several articles and a monograph on the topic. I have been awarded the Philip Leverhulme prize in law, which I am going to use to look further into questions raised by digital platforms and the application of competition law.
In the interests of full disclosure, in case it comes up later, in the past I conducted one piece of research commissioned by Google. It concerned the then ongoing investigation of the European Commission into Google’s practices, which culminated in an infringement decision.
Dr Orla Lynskey: Many thanks for inviting me here this afternoon. I am an assistant professor of law at the London School of Economics, and I work and research primarily in the area of data protection law. In particular, my main piece of research has focused on the limits of individual control over personal data. That has brought me to consider structural or holistic approaches to the effective protection of individual rights in the digital context. As a result, I have conducted research in recent years on the concept of digital dominance or data dominance, and looked at the interplay between data protection law and competition law, and in part consumer protection law, in the digital environment.
Dr Nicolo Zingales: I am very grateful for the invitation and am honoured to have the opportunity to give evidence on this important topic. I am a lecturer in competition law as well as in information law. I deal with online platforms from both perspectives, with regard both to the challenges they pose to traditional competition analysis and to their role and responsibility in ensuring the effective protection of user rights. In that regard, I am the co‑founder and co‑ordinator of the Dynamic Coalition on Platform Responsibility, which is a forum of individuals from different constituencies who discuss the role and responsibility of online platforms.
In the interests of disclosure, I am a co-founder of MyData, which is an initiative that ensures that individuals have more control over personal data, and can make more informed choices and derive knowledge.
Finally, on one occasion I too was funded by Google, but it was not for a specific paper; it was for a summer fellowship. I raise it in the interests of full disclosure.
Q84 The Chairman: Thank you to all our witnesses for introducing themselves.
The scale and dominance of the big platforms is the subject of a lot of media attention, public debate and policy-making. In the view of our witnesses, is the dominance of digital platforms a genuine issue for public concern? Perhaps we could have a brief perspective from each witness.
Professor Pinar Akman: As a competition lawyer, I do not think that dominance and market power on their own are a cause for concern. In the economic literature and numerous studies on this topic to date, what matters is the conduct adopted by companies that might earn market power. Pretty much all modern competition law around the world is based on that principle, so having market power, even at the level of dominance, is not on its own a cause for concern. What would be a cause for concern is if companies engage in conduct that is anti‑competitive, distorts competition and ultimately harms consumers.
Size on its own does not tell us much. We do not know whether the size is the result of superior efficiency and being better than one’s rivals or the result of anti-competitive conduct. Modern competition law takes the view that size on its own does not tell us anything about the outcome as such. It might be a sign or a result of superiority or efficiency, and as long as it is not the result of an anti-competitive practice we would not be concerned.
Other factors to take into account would be barriers to entry to a market and whether, for example, consumers are multi-homing; whether there are switching costs; and whether there is access to capital. Can new entrants come into the market and challenge the incumbent? In the digital context, a recent study conducted by BEIS found that effective entry does not appear less likely in concentrated markets in the digital world. It looked at five case studies from the digital economy and found that sometimes concentration made entry more likely, because the bounty at the end in being successful in this market was larger than the bounty would be if there was more competition in the market.
We do not know whether there is an ultimate amount of concentration in innovation, but, on the basis of at least one study conducted recently, we know that concentration on its own does not make entry—which means new competition—less likely to occur. That is where I would stand as a competition lawyer.
Dr Orla Lynskey: Competition law is relevant in so far as it is the primary legal instrument available to us to regulate and constrain private market power in any way. However, competition law is not designed with the intention of remedying human rights problems or other problems that fall outside the remit of the concept of consumer welfare.
One potential source of unease is that firms occupying a position of strong market power might not simply be engaged in harmful economic conduct, which would be effectively constrained by competition law provisions. It might equally impact on the effectiveness of rights in other ways. It might not be a competition problem, but human rights problems might flow from the dominance of certain firms. I will give two examples.
A digital giant such as Google might have a direct impact on fundamental rights through the way it processes personal information in the context of its many services. However, it also has an indirect impact on the level of rights protections offered throughout the digital environment, because, for instance, it has a chokehold, or it is a gatekeeper for access to its own platform.
A mobile phone has an operating system, which in the case of Google is the Android operating system. In order for apps to be available on the Android operating system, Google will have an influence over the terms and conditions offered by those apps. In that way, dominant digital firms have the potential for a particularly powerful influence over the effectiveness of all forms of rights in the digital environment—autonomy, data protection, freedom of expression, et cetera—and that is where the unease comes from.
Dr Nicolo Zingales: I agree with what has just been said about the important role platforms play in impacting fundamental rights. Specifically with regard to that, there is a distinction to be made between different types of platforms. When we talk about a notional platform, I would argue that some of these entities are a more critical architecture for the freedom of expression and interaction of users—for example, app stores and search engines come to mind as particularly important in that regard. The problem with the existence of a concentrated market is not so much the scale and dominant position of companies but the framework that we have for detecting possible infringement of the law. We do not have a co‑ordinated structure to deal with the range of issues that gives rise to.
I do not want to pre-empt my answer to other questions in that direction, but, as part of a team, in 2014 I conducted research into the terms of service of online platforms, which is one way to measure the exercise of market power by platforms. We found very problematic terms that kept users from an effective right to be heard in cases when content was removed. The terms deprived users of the right to access the courts and imposed a waiver on class actions, as well as mandatory jurisdiction in California on most occasions.
In general, the information that platforms provided about personal data they collected was quite insufficient. That is one example where there is no equal bargaining power between the two parties. As you know, users accept terms of service without reading them, and, even if they read them, they might not have the ability to understand them.
A further problem is that we cannot just rely on terms of service, as the platforms are constantly nudging us in one direction or another, and they effectively implement law through code. We need a system that is able to detect violations that occur through code, and allows users the opportunity to participate in the process and understand whether something that has happened to them is fair and legitimate, or is an abuse.
Lord Gordon of Strathblane: Is there something about the internet that has a built-in tendency to dominance—a virtual monopoly? You go from 51% market share to over 90% market share overnight.
Professor Pinar Akman: Indeed, and that is the result of network effects. On the internet, once somebody comes up with a product that becomes popular, there is almost a snowball effect; users attract more users. In an advertising-funded platform, users attract advertisers, so in a way success is exponential.
The study I mentioned earlier found that it also works in the other direction. If a platform starts losing users, it seems to lead to a rapid decline in some platforms that are no longer as popular as they used to be—for example, Yahoo in the search area, and Friendster in social networking. Success comes very quickly. An author has said that it has never been so easy to make a billion, but it has also never been so difficult to make a million. When you are successful, you are incredibly successful, but finding a product or a service seems to be the crucial thing.
Dr Orla Lynskey: Data has a role to play as well. There is a very lively debate at the moment among competition scholars and others about whether or not the possession of data on individuals, in particular the volume and variety of data, would lead to an advantage that could ultimately become a barrier to entry for potential competitors. That could be relevant in a single market. For instance, Google could use data in the context of Google Search to consolidate its position there, but it could also use that data in neighbouring or emerging markets. Competition authorities are struggling with how to treat that data.
Lord Gordon of Strathblane: What I am hearing suggests that, provided the barriers to entry are not impaired, you are not particularly worried about Google having, say, 94% market share in the UK.
Professor Pinar Akman: I would respectfully disagree with the point about data. Google itself had no data when it started, whereas Yahoo, the incumbent at the time, had loads of data. The same goes for the social networks that existed before Facebook. Google and Facebook were themselves new entrants in markets where there was an incumbent with data. It seems that data may not be the key factor that enables or prevents entry. In today’s world, there are institutions in the business of data. You can buy data. There are diminishing returns of scale with data. You need only so much data to work out what consumers prefer. In my opinion, as long as the entry barriers are not insurmountable, I would expect competition to do its work.
Dr Nicolo Zingales: The example of Google and Yahoo is particularly interesting. Yahoo was well ahead in the quality of its results. The big mistake it made was to sell its traffic to Google. By saying, “We are going to power our searches through Google”, it basically lost its competitive edge. This market is driven by scale, which is at the basis of the search industry. Now it has taken a further step, which is data, because searches are personalised. It is not only scale, which may indeed legitimate the presence of one or few players in the market because you need a lot of traffic to get accurate results; today, personalisation increasingly plays a role.
Lord Gordon of Strathblane: If a large dominant company buys a new entrant, is that automatically construed as anti-competitive, or just as a Christmas bonus for the new entrant?
Professor Pinar Akman: That is an important question. At the moment, we may not have the right competition tools to deal with it. We have examples where incumbents have bought new entrants—innovative companies that have something really different. Some of them have escaped scrutiny in competition law, because merger rules, as we have them at the moment, do not always have the capability to scrutinise such deals. It happens particularly when the two companies do not appear at face value to be competitors, but it might be an area that the incumbent may consider going into. That is important and it is something competition authorities will need to deal with.
It is possible that the rules may need to change so that such deals can be scrutinised so that they do not kill competition. We should remember that some new entrants may be innovating simply because they want to be acquired by Google. We need to think about the possibility that, if the deals are prevented, those innovations may not happen. I do not know whether we have enough evidence to work out which way it is at the moment, but it is definitely something we should be looking into.
Baroness Quin: I want to pick up a point Dr Lynskey made. Do you think the platforms recognise that they have an effect on human rights, or is it something they largely ignore?
To piggyback on my colleague’s question, how easy is it to scrutinise the kind of deal you mentioned? How easy is it to scrutinise something that on the face of it seems very opaque—at least to me?
Dr Nicolo Zingales: On the first point, which was directed to Dr Lynskey, I increasingly participate in conferences where representatives of platforms talk about what they are doing to give effective protection to fundamental rights, so they recognise that. Even at UN level, the special rapporteur on freedom of expression issued a report this year that said it was increasingly concerned about the lack of transparency.
Platforms have recently come up with a declaration about how they are moderating content. They are taking steps, but so far they are baby steps. There is increasing political pressure and recognition on their part, but it would be good if they had specific procedures in place to show that they have accountability by design.
On the second point, there is a very difficult question. To go back to one of the arguments that was made, data has a key role. It can easily move from one market to another; it can be used to build new services. Therefore, when there is an acquisition by a player that might not be in the same market but has many users and much information about what they are doing, that information can be used in another market. There needs to be more attention paid to those kinds of acquisitions. A good example is Facebook’s acquisition of WhatsApp.
More generally, a problem with the accumulation of data by platforms is that they have much more information about how the markets are going to evolve. Quite often, they see their competitors before those competitors or anyone else, even the competition authorities, realise they are a threat, and they buy them. I am not sure what the solution to that would be, other than to say that perhaps we should rely more on the knowledge that the platforms have. Rather than referring to general understanding in the industry, we should have a more subjective analysis of what they are doing. That answers only part of your question.
The Chairman: Dr Lynskey, do you want to deal with the issue of human rights?
Dr Orla Lynskey: If I may, I will go a bit further and deal with just one right: data protection or privacy. I recognise that we should not speak of platforms as a single entity because there is a lot of internal differentiation in their business models, how they are monetised and so on, and that merits consideration. However, in the sphere of data protection a platform would simply say, “This is a regulated area and, therefore, our scope for manoeuvre is quite limited”. That pushes the question back: if an area is not functioning effectively for individuals, it is because there is a market failure.
The rules introduced under the new general data protection regulation apply across the board, potentially to me as an individual processing personal data, in the same way as to Google. There are some differences as regards accountability, the need for documentation and so on, but in general the legislative framework that we have at present, in that sphere in particular, does not put special responsibility on firms that, because of their ubiquity or reach, might be particularly impactful on rights.
There is at least a case to consider a type of special responsibility for rights protection analogous to the special responsibility on dominant firms that we see in competition law, because everything they do has such a significant impact downstream, or on the market. You could make that analogy and say that that type of special responsibility should be extended across the board to the rights area, because the logic that everything they do has a more significant impact extends across the board to societal effects, including economic ones but going beyond them.
Professor Pinar Akman: On Baroness Quin’s question about how to scrutinise a merger when an incumbent buys an innovative firm, you are absolutely right that it is particularly difficult in dynamic markets to predict what is going to happen in the future and how competition will evolve. That was submitted by the Competition and Markets Authority in its evidence to this Committee. In a way, merger control is always about gazing into a crystal ball, because it is always prospective. Whenever a competition authority looks at a merger, it tries to predict what will happen in the future in that market.
One thing that authorities around the world are considering for this type of merger is transaction value. Currently, merger rules do not catch that aspect of the transaction; we look at the turnover of the parties, and in the UK we can also look at the share of supply of the parties, but we do not look at the transaction value.
A good question is why an incumbent is paying billions to purchase a new company that was established just last month. That suggests to us that the company thinks there is something to look at, and that is one thing for competition authorities to consider.
Viscount Colville of Culross: I do not know whether it is possible to do this briefly, but I will direct this question to you, Dr Zingales, because you do work that looks at the user-platform interface.
We are looking at the dominance of ownership of platforms. Is it possible for Governments to nudge the market so that more user or publicly or co‑operatively-owned platforms, where there is empowerment of workers, users and citizens, determine what happens on the internet? Is there anything that Governments can do and anything that we can suggest?
Dr Nicolo Zingales: In that respect, the Government have already tried to do something with regard to the so-called Midata initiative. They suggested that consumers should have access to their personal data. Going back to the initiative that I co-founded, MyData, the principle is that all data should not go exclusively to the platforms. An alternative solution is that platforms have access to data generated through them, but that data must be given to consumers as well, which would enable them to create decentralised structures that can move to other platforms to connect.
That alternative model is increasingly being advocated as part of data co‑operatives. A number of individuals establish the rules and try to protect certain fundamental values—for example, workers’ rights and personal data protection. It is a model that I would definitely recommend taking into consideration because it does not require the same level of intervention that is advocated against the dominance of platforms: it suggests that perhaps we can harness the power of the market and empower consumers to switch very easily, which is connected to your inquiry’s interest in data portability, and make more informed choices, as well as knowing the terms and conditions that will apply throughout their activity on the platform. At the moment, vague and sometimes quite unfair terms are being used against them, and we do not know how they are being implemented. If we switch to this model, users would be able to establish their own terms and conditions and then allow third-party platforms to connect to their ecosystems only if they fulfil those terms and conditions, so you flip the model around.
Viscount Colville of Culross: That is interesting.
The Chairman: That is a very interesting area. We may want to explore it a bit later, or on some other occasion.
Q85 Baroness Kidron: The other side of the equation is how the regulators see it. Would you say briefly what you think the position of the regulators is? We keep getting evidence about price, but perhaps they are not using some of the other qualities that are at their disposal. We would be very interested to hear from you on that.
More importantly perhaps, do you think the current system is suitable for the digital environment? We are thinking particularly about end-to-end services that do not necessarily look dominant but might dominate a user’s experience as well as a market.
There is also the question of jurisdiction. Professor Akman, you have already brought up the question of mergers. Could each of you look at the regulators’ current position and what tools may be missing from their toolbox?
Professor Pinar Akman: If I may, I will start with the Competition and Markets Authority as the competition regulator. The competition rules are flexible—perhaps too flexible—and could be applied to any set of circumstances. They are very short principles and rather vague. We may be lacking some of the tools with which to apply those rules to the digital markets that we have.
One particular example, which is quite fundamental, has to do with the point you made about price. Usual competition analysis is built on the idea of price and market power being defined as the ability profitably to increase price. That causes a serious problem for some of the online platforms that we are looking at today, because they are actually offering their services for free to the users. This is a serious problem for competition law assessment.
Any competition assessment, particularly on dominance, starts with defining the relevant markets: what is the product market that we are looking at? We cannot define dominance in the abstract; we have to establish what the product or service is. Because these markets are two-sided, and remuneration is coming only from one side, which in the case of, say, search engines is advertisers, there is difficulty in figuring out whether we look at the users’ side, at the users and the advertisers, or at the advertisers’ side because they are the ones paying for the service. For example, in the United States, court decisions have held that there cannot be an anti-trust market when the product is free, and that is the end of the inquiry. Obviously, courts and authorities in Europe have taken a different view, but it is a serious essential question to be dealt with at the beginning.
Unfortunately, I do not think that currently we have the tools to answer such questions with the required precision. That is one example where the competition authority, as the regulator, will be challenged when looking at the digital market. It will be similar with the role of data. Is big data a barrier to entry, or is it like any other input to a business? Because these firms are digital, their input is digital, in the form of data.
On efficiencies and the role of innovation, how much value does one put on innovation, which will obviously happen again in the future? There is competitive tension in the market, which is the usual way we think of competition; there are certain companies in the market and they compete in that market. There is also competition for the market, which is a far more dynamic perspective of competition, and that has to take innovation into account. How much emphasis to put on the role of innovation in competition law assessments is a challenge to the competition authorities at the moment.
Dr Orla Lynskey: Another challenge, in the context of a free market, is that, if we take the emphasis away from price, we are left to consider other parameters of competition: innovation, quality and choice. A lot of work has been done recently on how to measure improvements or disimprovements of quality. A digital product such as WhatsApp’s communication application is priced at zero, so how can we tell, after its acquisition by Facebook, whether it has improved or disimproved in quality? You might say that the quality of the data protection policy has lessened as a result of the transaction. However, it might have improved aspects such as data security, because it now benefits from Facebook’s data security infrastructure. Internally, when considering quality, there will be a lot of incommensurables or things that are very difficult to measure one against the other.
In some instances, competition analysis in digital markets might have blind spots simply because a lot of the focus is on the impact of a particular conduct or transaction on actual or potential competitors, whereas, for mergers, as has already been highlighted by the Committee, a lot of the acquisitions are taking place in parallel markets, and that falls into the blind spots of competition analysis.
If I think about that from my data protection perspective, I see the acquisition by big platforms of firms that are not direct or potential competitors, but are data processors in peripheral markets that are being gathered up slowly but surely. No one transaction is causing a big stir, but, when you look at the overall picture afterwards, you see large-scale data aggregation from a wide variety of sources.
In my research, I have argued that we should be considering whether or not to use tools that are parallel or complementary to competition tools, such as the public interest test in the context of mergers, to assess that type of transaction. That type of test is currently used primarily in the context of media mergers, but we might be able to make some sort of analogy with the data protection context and say that the economic outcome of the transaction is not the sole consideration. We might also consider the broader societal impact of the transaction, and whether there might be implications down the line for individual autonomy, freedom of expression, data protection or choice, simply in the digital environment. That might merit further consideration.
The Chairman: Do you have anything to add, Dr Zingales?
Dr Nicolo Zingales: I agree with everything that has been said, but I have two further points. I agree that it is difficult to measure the impact on innovation and quality; it is much more difficult than with prices. If we think about predatory pricing, there is a clear rule; if you price your product on a measure below your costs, you basically aim to exclude your competitor from the market. It is difficult to pursue that kind of specific reasoning with regard to quality, and for example with regard to data protection, which is increasingly a measure of quality.
In that respect, the GDPR will bring more clarity—for example, through codes of conduct[1] that could set out different levels of protection for personal data in different circumstances. That could be a way to measure quality.
Another purely competition issue is that often the benefit brought by certain restrictions of competition in one market flows into another market. If we think about the free market that users get, there might be some restriction on the advertising side that enables the provision of the service for free. Or there’s the classic case of credit card companies, which impose a certain restriction on their merchants in order to offer their service uniformly. That is a big question for competition law, and there was a decision in the Supreme Court yesterday on that very issue. It is not exactly clear how we evaluate cross-market efficiencies. Currently, EU competition law tends to be quite focused on one market and does not allow one to take efficiencies in another market into account as long as the consumers in the two markets are substantially different. They need substantial commonality of consumers to take into account the benefit in the other market.
Baroness Kidron: I note that in Australia competition law sits with the consumer and competition regulator; they are one and the same. That may provide a more holistic approach to some of these issues. Are you nodding in agreement or recognition?
Professor Pinar Akman: Yes, both recognition and agreement with the principle that it might be far more useful to have broad powers within the same authority.
The Chairman: Are the other witnesses nodding in agreement?
Dr Orla Lynskey: Yes.
Dr Nicolo Zingales: Yes.
Q86 Baroness Chisholm of Owlpen: We have already heard a bit about conduct, dominance, transaction values and empowering consumers to move. I would like to take that a bit further. What do you feel about competition law assessments? Do they strike the right balance between short-term efficiencies and innovation?
Professor Pinar Akman: This is a very difficult question, because it requires us to know quite a bit about the counterfactual world on which we almost always have no information. Ex post assessment of competition interventions is very rare. In cases where there has been an infringement decision and a company has, let us say, stopped whatever the infringing conduct was, and in cases where there has been a non‑infringement decision, we never know what the actual impact has been on long-term innovation, so it is very difficult to answer.
At the moment, the authorities I am familiar with, such as the European Commission, are more concentrated on short-term efficiencies, and the longer in the future the alleged efficiencies are, the less weight they are likely to be given in traditional competition law assessment. One example is the Microsoft case, which found an infringement against Microsoft in the EU some years ago. Microsoft took great pains to argue that, if the EU proceeded in the way it was proceeding, there would be a detrimental effect on Microsoft’s incentives to innovate. Its rivals might innovate, but what about its incentive to innovate as the current dominant company? That was not well received.
Where innovation is taken into account, it is usually in favour of the competitors of the incumbent, as opposed to looking at the innovation that might come from the incumbent itself. Several commentators noted that, looking at the Microsoft investigations in the US and in the EU, if those interventions into Microsoft’s conduct had not taken place, Microsoft today might have been a far more serious competitor to Google, for example. It is very difficult to know.
Dr Nicolo Zingales: I broadly agree. We need to recognise, as was also mentioned in the previous question, the concept of system competition. You need to take into account that the incentive set up by having a company take part in multiple lines of businesses may be undermined if you focus on a very narrow market.
I also believe that the concept of nudging is powerful. The regulator should not be too strict on a company favouring one product or service over another, which was complained about in the Google case. There is some merit in scrutinising those kinds of practices—preferential placement—but, on the other hand, attention needs to be paid to the fact that algorithms constantly nudge us in all sorts of directions and that it is the role of the algorithm to detect what goes first and what goes second. You cannot adopt too mechanistic an approach to scrutinising that kind of practice. I do not know whether that is clear to the panel. I think it is one of the major challenges. I am arguing that it would undermine innovation to apply it too mechanistically without a de minimis exception, for example. Perhaps I can comment on the framework that needs to be set out to address that kind of practice in later questions.
Q87 Viscount Colville of Culross: You have explained some of the problems in dealing with competition law in the digital markets. I would very much like you to address what can be done. What reforms could be made to existing competition legislation to enhance the competitive process?
Professor Akman and Dr Lynskey, you talked about the difficulties of competition authorities looking at the ecosystems in too narrow a way, and often there is a parallel purchase and a parallel market that did not understand the effect. You explained earlier that a smaller company might be acquired outside the existing ecosystem and could have an unexpected effect.
Dr Zingales, you said that companies had more data than the regulators could have, so they understood better where the market was going. How could we nudge and change the legislation to help the regulators?
Dr Orla Lynskey: Competition law is the only legal tool we have, aside from economic regulation, to deal with dominance. In some ways, we are putting a lot of emphasis on competition law to expand and face a lot of problems within its remit. Competition law will deal with these issues case by case, so there is always the risk, which Professor Akman alluded to, that if you intervene at one level, for instance in the way I advocated in the merger context by having a public interest assessment of data-driven mergers, you might have unintended consequences from that type of intervention.
For instance, if you intervened in the data-driven context, you would say that aggregation by a dominant firm such as Facebook of entities such as Instagram and WhatsApp will lead to an aggregated dataset that is too large, reveals too much and gives too many insights about individuals. At the same time, competitors might say, as was argued in the US in the AT&T merger, “We need to merge in order to have exactly that kind of dataset at our disposal. We need that competitive insight in order to compete”. By disallowing that transaction in the UK or the EU, you might have a situation where it is allowed in the US. There would be competing claims that in my opinion would be impossible for any one competition authority to deal with, and to foresee the broader consequences for the market.
What does that point to as a response? There could be more collaboration between competition authorities, with an overall structural assessment of how the internet will function as a market now and in the future. That is obviously a huge exercise, but you can see from a lot of the economic literature that we are plagued by uncertainty.
The big question is whether we believe there will be another Schumpeter wave of destruction that sees existing firms wiped out of the market, or whether we will have entrenched dominance for the foreseeable future. We need more research on that question.
More immediately, we could see more collaboration between different regulators already operating in the digital sphere, so that there is no duplication of effort and everyone is on the same page about where their competencies overlap. In the UK, that could be the Information Commissioner, Ofcom, the CMA and various others. There are many instances where competencies overlap, and they are increasing. I can give some examples if that is helpful.
Viscount Colville of Culross: Are you suggesting that there should be a super-regulator that covers all those things? Would that solve the problem?
Dr Orla Lynskey: I do not know whether you would need an entirely new regulatory body. Dr Zingales has been involved in the idea of a digital clearing house, which has been proposed at EU level. In that instance, it was foreseen that there would be co‑operation between competition, consumer protection and data protection policies, in a very loose co‑operation mechanism, just to see where their areas of interest overlap.
Professor Akman has already mentioned that competition authorities might consider the question of a dominant firm extracting a lot of data from individuals. That might be problematic from a competition perspective, but it could also be a data protection question, or a consumer protection question. The very same legal problems would be analysed through different lenses. There is scope to join up thinking on how to deal with that type of issue, which is central to how we as individuals experience the internet.
The Chairman: Can I ask the witnesses to be reasonably brief? We are quite pressed for time. You have given us lots of evidence and there is more to come.
Professor Pinar Akman: Some of the issues we have mentioned with the application of the rules in digital markets will be resolved only through enforcement and case law. While that is happening, economics can catch up by offering us economic analysis of what is or can be happening in some of those markets, because economics underlies pretty much all competition law and enforcement. The problem is that economics and the law are lagging behind actual technological developments. The problem with enforcement is that it takes years, and by the time we have a decision the market has completely changed.
We should also bear in mind that it is possible that there is no competition law problem. It could be that competition is working effectively. Several authorities around the world are looking hard to find problems in these markets. If there are problems, I am certain that they will find them. The fact that we have not had much enforcement might be an indication that there is no competition law problem.
As to what authorities can do to innovate to stay more in line with the innovative companies they are looking at, they could set up their own data units, as the CMA is doing. They could use algorithms to catch conduct that is taking place through the use of algorithms in the markets, which the EU Commission has alluded to. There are ways in which the authorities can innovate to stay in tune with the markets.
Dr Nicolo Zingales: To add briefly to Dr Lynskey’s point, the digital clearing house is an important example of collaboration, but the authorities need to co‑operate not only at the level of exchanging possible ideas, theories about market definition and so on; they should collaborate on specific cases. Quite often, a data protection issue comes up in a competition law investigation, and the authority is not sure how to deal with it. It might not have the power to pass that information to other authorities, so that might be where we need a specific co‑operation agreement between different authorities. The authority might not have the ability to assess the issue, even within competition analysis, and give it the appropriate qualification from a data protection perspective, so that calls for a more integrated approach to specific cases.
Q88 Lord Goodlad: I have two questions. First, in your view, what principles or criteria should determine users’ right to data portability and platform account deletion? Secondly, could greater data portability and interoperability mitigate the control that the dominant platforms currently exercise over personal data?
The Chairman: Can we start with Dr Lynskey? You have addressed some of these issues partially, but please add to them.
Dr Orla Lynskey: Data portability is underpinned in principle by the idea of individual control over personal data. The new Data Protection Act in the UK, which reflects the general data protection regulation, seeks to give individuals more effective control over their personal data, which is where initiatives such as MyData come into play, because they allow individuals to do something with that data.
Will that be effective, or lead to interoperability? Data portability is one thing, but it does not necessarily mean interoperability. Interoperability means that, if I have my data on one platform and you have yours on another, we can interact with it; for instance, I could be on Facebook and Dr Zingales could be on a different social networking service, and we could communicate. Although that might unlock markets, it could be problematic from a data protection perspective, because it leads to data duplication. There are questions about whose privacy policy and whose terms of use prevail in those circumstances. That might be an example where you need joined-up thinking about what would be the best way to proceed and whether or not you force interoperability.
Dr Nicolo Zingales: I agree. Data portability and control is not a panacea, much like transparency, because users often do not have the ability to appreciate all the circumstances, and they might not exercise their control adequately. The problem with data portability is that quite often data relates to other individuals, so we might want to make sure that users use such data appropriately.
As regards affecting dominance and improving the situation in the market, one needs to bear in mind that there is always the possibility that, if the labour that has been put into a platform is taken to another platform, it might undermine the incentive to produce that labour in the first place, which is why the general data protection regulation does not apply to so-called inferred data; it applies only to data provided by the user or observed on the platform by users of the platform. It would probably be far-fetched to require the platform to give all the structures it has for the user to be able to port them anywhere else.
With regard to the difference between portability and interoperability, interoperability is a different concept that requires much more effort; it requires collaboration, because the application programme interfaces are constantly updated. A basic standard needs to be set, probably across the industry, and updated with technological development. It is much more difficult to achieve that. Portability goes some way to improve the situation, but it does not resolve the problems of lock-in and switching costs in moving to another platform.
Q89 Lord Goodlad: How do you think GDPR will impact, if at all, on competition law or competitive assessments?
Professor Pinar Akman: My colleagues will be able to add more on the details of the regulation, but, if I may, I will make a small point. GDPR may have an adverse effect on competition in the market, because it is a very extensive and detailed regulation, which obviously will come at great cost to small businesses, and potentially new entrants, in the market where it applies.
Regulation itself can create entry barriers for new entrants, and GDPR may be an example; you may have come across websites that literally stopped operating in the EU when GDPR entered into force, because they were not in compliance. For companies such as Google and Facebook it is going to be far easier to comply with GDPR, but for new and small companies, which could challenge those large companies, compliance will come at a greater cost.
Dr Orla Lynskey: I may differ slightly on that point. GDPR may increase the costs of doing business in so far as any regulation, such as labour law regulation, increases the cost of doing business. GDPR scales the obligations it imposes depending on the scale of the data processing operation. If you have a company with five people but you are doing large-scale data analytics, the accountability obligations and other features that are imposed on you will be significant, but rightly so, I believe.
As to how it might influence the competitive environment, it is something that competition authorities need to be aware of when assessing competition in the market. We may see the possibility of companies starting to compete effectively with one another on the basis of the data protection conditions they offer users, because to date we have not seen that type of competition emerging. For instance, now we are seeing models where businesses are saying, “You can have less analysis of your data by making a micropayment for this service, or full-scale data analytics and the service for free”. There could be other issues with that, because in essence we are pushing the responsibility back on individuals to pay for their data protection, so minimum standards would still need to be in place.
Dr Nicolo Zingales: I do not have much to add, other than a point on the complaints that GDPR is affecting competition in the market because small companies cannot cope with all the regulations. The regulation is indeed imposing some obligations across the board, which are essential for it to be effective, and then it imposes an asymmetric type of regulation where there is a lot more emphasis on the capabilities and impact of data controllers. They will need to exercise a greater level of care, which I think favours the small players, who will simply have to comply with minimum requirements.
Q90 Baroness Quin: Given that most competition regulation is currently carried out at EU level, will there be problems for the UK post Brexit in relation to enforcement of competition law, or, given what you have said about the importance of co‑operation between authorities, is it likely that we will be very closely aligned to the EU in any case?
Secondly, in the new post-Brexit world, will a UK regulator be able to take effective enforcement action against US-based companies?
Professor Pinar Akman: That is another great question. I had the honour of giving evidence to the EU Sub-Committee on the Internal Market on the issue of Brexit and competition. Two things came out of that process.
First, resources are very important. Unless the CMA’s resources are increased proportionately to the expected increase in its workload, it simply will not be able to cope with the workload. There will be a whole set of mergers that the CMA will need to look at, which currently it does not because of the one-stop-shop system whereby the EU Commission deals with cross-border mergers.
Another really important point that came out of that process was that co‑operation will be essential. The CMA will need co‑operation agreements with key stakeholders such as the EU, the US and other jurisdictions, because several of the issues will be cross-border. How will the CMA go into another jurisdiction to collect evidence of an infringement? If it takes a decision, how will it enforce that decision in a foreign jurisdiction?
Those things can be resolved only through co‑operation agreements between the CMA and the other stakeholders. When the UK leaves the EU, presumably the CMA will no longer be a member of the European competition network, which is currently the network that enables such information sharing and co‑operation with member state competition authorities. I only hope that there will be some alternative arrangement that will put the CMA on that footing after the UK leaves the EU.
In principle, the rules can be effectively enforced against US‑based companies. In practice, it will depend a lot on the resources that the CMA has and its ability to enforce its decisions in foreign jurisdictions and collect evidence in foreign jurisdictions.
The Chairman: We have a couple of questions left. The Lord Bishop will ask questions and then Lord Allen will ask a further question. I will ask you to respond to Lord Allen’s question in writing, but I would like to put it on record.
Q91 Lord Bishop of Chelmsford: You have already touched on some of these issues, but you may wish to add something. Do you think competition law could alleviate the need for other forms of internet regulation? I am very mindful of your brief but affirmative answer to Baroness Kidron earlier, which may be relevant here. Could the consumer welfare standard be amended to encompass the non-economic concerns that you also mentioned earlier? Obviously, this is not just about price.
Dr Nicolo Zingales: There is increasing understanding that consumer welfare might encompass some non-economic objective, but what is often missed in the debate is the nuance of when the consumer welfare standard becomes relevant. The two key moments during competition analysis when it can be a factor are at the stage of justification of conduct and in the imposition of remedies. Why do I say that?
Often, a company might be pursuing one of the fundamental rights objectives or might be mandated by law to undertake a certain action. That action might be justified on the basis of something that is not strictly economic. For example, if companies agree that they might set standards of access to their platforms, on the basis of privacy, to preserve that kind of value, it might be a legitimate justification, or what is called an objective justification in competition law.
The other element where it is particularly relevant to look at other areas is that, when the competition authority imposes a remedy, it has the duty to make sure that it is not infringing other rights. When it imposes the remedy, it should bear in mind that it has to craft the remedy in such a way that it preserves, for example, the data protection and intellectual property not only of the company in question but also of third parties.
The problem with the discussion is that often it simply refers to the fact that the standard by which we should judge conduct is not an economic one, and we should pursue all sorts of values as the ultimate objective of competition law. The objective of competition law is primarily economic. At EU competition law level, we also have market integration objectives, and the query is what we will do after Brexit in that regard. However, I would caution against putting it all together in one basket, specifically because it would create an unpredictable standard and some policy leverage that is also very inappropriate for the allocation of competencies, the rule of law and legal certainty.
Professor Pinar Akman: In my opinion, the consumer welfare standard should not be amended to encompass any non-economic concerns. That is not because the consumer welfare standard is perfect; it is far from perfect, but, of the other options we have, it is the most concrete, if that is the right expression. If we include other concerns that might be more political or might have to do with issues that the competition authority cannot really deal with in its assessment, we turn the business environment into a very uncertain one, which will put off businesses from investment and innovation.
According to the Enterprise and Regulatory Reform Act, which established the Competition and Markets Authority, “The CMA must seek to promote competition, both within and outside the United Kingdom, for the benefit of consumers”. That is the guiding principle in the legislation. I think it is the right one, and we should stick to it in future.
The Chairman: Dr Lynskey, you have the last word.
Dr Orla Lynskey: I simply concur with what has just been said.
The Chairman: That is very precise. Thank you very much. You have given us a lot of evidence. One of the things you talked about earlier was the need for co‑ordination of regulations. I am going to ask Lord Allen to put a question on the record. Given that time is pressing, I will then ask you to respond to it briefly in writing.
Q92 Lord Allen of Kensington: As we know, there is a plethora of regulators on the internet. I would like your views on two things. First, do you think there is a need for an ombudsman for consumer complaints? Would that be helpful?
Secondly, there seems to be quite a gap where we are playing catch‑up all the time as new platforms and models are coming out. We had examples about blockchain changing the world. Is there a need for a regulator with a role in working with the industry in horizon planning? If you think there is, would that be best served through a new regulatory body, giving existing regulators more authority, and, as we have just heard, a co‑ordinating role to bring things together, so there is not the level of conflict we currently see with regulation across the internet?
The Chairman: If you would be so kind as to respond to us on that in writing, we would be very grateful. If there is anything else that you think we might have asked but did not, or anything you might have said if I had not been pressing you so hard to speed up on occasions, we would welcome it. The evidence has been very useful and informative for the Committee. Thank you very much for giving us your time today.
