Oral evidence - Online Platforms and the EU Digital Single Market

Members present

Lord Whitty (Chairman)

Baroness Donaghy

Lord Freeman

Lord German

Lord Green of Hurstpierpoint

Lord Liddle

Baroness Noakes [1]

Lord Rees of Ludlow

Lord Wei

_______________________

Examination of Witnesses

David Evans, Lecturer in Law, Chicago Law School and Chairman of the Global Economics Group, and Ariel Ezrachi, Slaughter and May Professor of Competition Law, Oxford University

Q23 The Chairman: Good afternoon. Sorry to keep you waiting. Welcome, and thanks very much for being prepared to help us on this inquiry. There are just a few preliminary points. There is a note on the website of the interests of the members of this Committee that are relevant to the work of this inquiry, if you care to look at it. This is a formal session. A full note will be taken. You will get a copy of the transcript, and you can revise any minor errors in it. So the session is on the record, is being webcast and will subsequently be accessible via the website. The acoustics in this room are not perfect, so could members of the Committee and witnesses please speak up. Mr Evans and Professor Ezrachi, perhaps you could introduce yourselves, and then I will come to the first question.

David Evans: Thank you very much, Lord Chairman. My name is David Evans. I am an economist. I happen to teach at a couple of law schools, at the University of Chicago and University College London, but I am not, and never have been, a lawyer. I have been working on this subject since 1995, so I have been working for about 20 years on issues of online competition and so forth. Briefly, as a statement of interest, I have an ecumenical view in this area. I have worked in the past year with some large internet players, in particular Tencent in China, and Google, which you probably know of. I have also worked for companies that are averse to some of the large online players. Last year, I worked for Netflix in an effort to block the Comcast-Time Warner Cable merger in the US, for example. I am also an adviser to a number of firms in this space, which tend to be small firms, and I am an investor. I have a lot of interests here, but my comments today are entirely my own and do not represent anyone else.

Ariel Ezrachi: I am Ariel Ezrachi. I am the Slaughter and May Professor of Competition Law at the University of Oxford. I am the director for the Centre for Competition Law and Policy. I have been writing on this subject for a while now with my co-author Professor Maurice Stucke, focusing on the link between quality and competition in online platforms. In terms of interests, I should mention that I have been doing work for Slaughter and May on wide and narrow parity clauses, and that Slaughter and May acts for Booking.com.

Q24 The Chairman: Thank you. In addition to members of the Committee’s interests, or non-interests, I should point out that Richard Feasey is our special adviser. He is not here today, but I understand that Mr Evans is acquainted with him, so we should also put that on the record for propriety’s sake.

Perhaps I can kick off with the first general question. On competition policy, what specific difficulties do multisided online platforms pose for the authorities and for Governments? There appears to be dominance by the most successful online platforms, which is reflected in their market share, however you calculate that. How dominant are those big platforms? Do they, and how do they, abuse their market power, and what do you think Governments and regulators should do about it?

David Evans: There were many questions there, Lord Chairman, so if I do not get to all of them, please remind me. I am not sure how long you want me to talk for. I could probably speak for about four days on those questions, but I suspect—

The Chairman: A bit shorter than that, please.

David Evans: So I will not do that. Let me make a few comments, and then perhaps there will be some follow-ups. I am not sure that I would characterise the situation as difficulties. I guess I would characterise it as competition authorities being very sophisticated; they have dealt with a lot of complex industries over time. There are differences concerning online platforms and it is important that they keep that in mind, and many of them do. First, these are multisided platforms, which is a new concept, although multisided platforms are actually very old—many, many millennia old. They are a different kind of business from, say, automobile manufacturing or farming. In particular, they are intermediaries between several different kinds of customers, and it is very important that competition authorities recognise that: that there are a number of stakeholders to these platforms, and that the stakeholders are interdependent. That is an important difference.

The other thing that happens to be true with online platforms is that we now have a 20-year history of fairly intensive dynamic competition. At every point, in 1995, 2000, 2005, 2010 and 2015, this sector will have looked very different from the way it looked five years previously. Sitting here in 2015, it looks a lot different from the way it did in 2010. In 2010, Apple was kind of getting off the ground with the iPhone, it was two years before Facebook’s IPO, and only two years before that MySpace was the dominant social network; it was not Facebook at all. This is an area that changes very quickly, which is a challenge for competition authorities because they have to recognise that dynamism, and that any time they are thinking about intervention, things may change very quickly right after they adopt it.

I will finish with the dominance question. I positively would not want anything I say today to be construed as suggesting that any of these platforms should get a free pass or that vigilant competition policy and consumer protection law should not be applied to them. Having said that, we can easily overstate the dominance and power of these platforms, in particular by calculating market shares. Many of these platforms look as though they are doing something discrete, such as social networking, search, an online mobile software platform or an app store, but the reality is that they all compete with each other quite intensely. They act that way, they say they are, and that needs to be taken into account as well, because in this area there is a lot of competition across categories. Just this past week, Facebook announced that it had vastly expanded search across its platform, thereby increasing the intensity of competition with Google, for example. You see a lot of cross-platform competition here that needs to be taken into account. In a lot of cases, these platforms are, on the one side, trying to get all our eyeballs to come to them. Once they have our eyeballs there, they want to sell them to the advertisers who want to reach us. Those are the fundamental dynamics for a lot of these internet players; they are all competing in different ways to get our attention, and once they get our attention there is this advertising market that they can sell that attention on to. It is easy to overstate the dominance of some of these players. Again, I am not suggesting that there are not these big, important players that all deserve the same kind of scrutiny that any large firm in the economy should get.

Ariel Ezrachi: I think I can highlight five areas where we can reflect on some of the difficulties that we have with multisided online platforms. The first is the fact that we have platforms that provide us with a free product, so one side is free, the other side is paid. Here, the first challenge that a competition agency may face is when price is not the issue. Then, the other dimension of competition that becomes central is quality. Quality is a challenging area to measure, to quantify. The question then is how exactly you measure the competitive process when price is taken out of the equation. The competition authorities have to grapple with a dimension that is not always easy to cope with. In our research, we have found that, at times, while those platforms police themselves and operate in a very competitive environment, they can actually degrade quality to some extent, because when they have to choose between the free side and the paid side—the side where they make the revenues from advertisements—their loyalty, or their interest, obviously lies with that side. That is one of the challenges that I would note.

Another challenge is the fact that in these markets, especially search engines, we have networks that are data-driven. That changes the dynamic of competition and makes it very different from what we are accustomed to. Data-driven network effects mean that as you get big, you can get bigger, so there are economies of scale that make you much stronger and may create a barrier to entry for newcomers. It does not mean that the market is not dynamic or that there are no risks and potential competitors; it just means that this is another challenge in that market.

A third challenge is the ability to engage in very individual targeting, tracking and reaction, which you will not get in traditional markets. We all know that if we go to the cinema we might get a reduced ticket if we are students, or we might have to pay slightly more if we do not benefit from a discount. In the online environment there is an ability for the operators to identify the individuals and therefore to engage in what we refer to as ‘almost-perfect-price-discrimination’. Price discrimination could be beneficial. It could also be harmful at times. The argument is that this ‘almost-perfect price discrimination’ is a challenging area for competition agencies. The transfer of welfare from the pockets of consumers to the pockets of operators is more likely. Again, I emphasise that I am mentioning the difficulties. I do not want to paint a very dark picture, but these are the difficulties.

Another difficulty is the perception of competition. We tend to view the online market as highly competitive. There is a lot of choice and a lot of options. There is different pricing. The question is whether behind the veil there is perhaps slightly less competition - smart algorithms that can dampen the level of competition and very swiftly bring the market to some sort of equilibrium that might be less competitive than we are used to.

Linked to that is the possibility that the big platforms that we come across might, even unwillingly, serve as a hub-and-spoke platform, which means that they, by their function, facilitate some sort of communication between other parties.

For example, if we have a platform that drivers use in order to receive business, and the platform determines the price, what might the impact be of such a platform on the price and on the market? If everyone uses a dominant algorithm in order to optimise their pricing, what will that result in? What we call almost perfect price discrimination, business will call price optimisation. If you go online, you will find plenty of companies that offer you the possibility to purchase services that will enable you to optimise—in other words, give you the ability to target and extract. These are the five difficulties that I would mention.

With respect to the second part of the question—how dominant are the big platforms?—I will make three relatively brief points. In our research, we distinguish between platforms and super-platforms. There are various differences between them; different platforms might have different powers at different times. It is, of course, a very dynamic market, which also makes it an extremely challenging area for intervention. Where we see a lot of interdependence, or dependence, between a platform and a super-platform is where we get the sense that there is a lot of market power for the super-platform. I might elaborate on that later.

There are two other areas. One is network effects and the fact that this is when big platforms might become truly dominant. The other is big analytics: the ability not only to harvest information but to process it and to get a very quick and real-time view of the market. Again, if we are focusing on what makes platforms dominant, I echo David’s words on the difficulty of assessing market power or dominance in markets that are so dynamic.

Q25 Baroness Donaghy: How do you assess when there is an abuse of market power?

David Evans: I do not think there is anything terribly unique about online platforms relative to other businesses. We can identify unique attributes of them, and we have just heard a set of differences, but competition authorities have been working on many different industries for over 100 years in the UK, the United States and other parts of the world. At this point, there is a standard toolkit for assessing the abuse of dominance. There is a set of practices, such as tying. There is the concept of essential facilities, predatory pricing and so forth, which are potential abuses and which we need to take a look at, and there is a set of tools which competition authorities apply for that. There is no single answer to your question. One of the reasons why competition authorities approach things the way they do with regard to the abuse of dominance is that every situation is different. There is a set of tools, which you customise to the set of facts that you have before you. There are some aspects in the case of online platforms; the fact that in many cases they have a free side and a paid side is certainly one of them, and that needs to be taken into account. But the analysis of the abuse of dominance is very fact-dependent.

I will add a couple of things in relation to that. I do not want to respond to everything the professor said, but I will pick up on one of them. The distinction between price and quality is true for all businesses. Whenever we look at a merger inquiry or the abuse of dominance, we should look at quality in addition to price, and quality is a dimension of competition for every business. It is certainly not unique to online platforms. It is something that Mercedes-Benz and VW compete on, in addition to price.

It is absolutely incorrect to say that online platforms place their interest on the paid side. In fact, they operate exactly the other way, because the only way you can get eyeballs in order to get to the paid side is to get the eyeballs. So unless they are competing intensely on quality and thinking of ways to get us all interested in coming to their sites, they have nothing to sell, so competition on quality on that side tends to be quite intense.

Finally, I want to emphasise that when competition authorities look at abuse of dominance, it is very important that they recognise that multiple stakeholders are involved. For example, if you do something that you think will benefit one side of the platform—let us say advertisers—you need, because of the interconnectedness between the stakeholders, to take into account the fact that that intervention could harm the other side of the platform, the consumers who are coming to the site. Many competition authorities are cognisant of that fact—the CMA certainly is. This interdependence between the stakeholders is very important in assessing abuse of dominance. You can think of these platforms as balloons: you press on one side, and it will have an effect on the other side. So long as competition authorities take that interdependence and the dynamics into account, they have the right tools for assessing abuse of dominance. But, again, it is fact-specific; sometimes these platforms do bad things, sometimes they do not.

Q26 Baroness Donaghy: Do some kinds of abuse worry you more than others? Are you confident that the competition authorities are going to see off the worst of these abuses?

David Evans: When I get involved in cases, there are certain kinds of things that I get excited about more than others. One of the concerns about online platforms involves the fact that in order to become successful you need to develop a critical mass. Think about Uber starting in a city: it needs to get drivers, because if it does not have them passengers are not going to be interested, and it needs to get passengers, because if it does not have people looking for rides the drivers are not going to be interested. That same concept is true for all these platforms; you need a critical mass on both sides in order for the platform to be successful. One of the things I tend to look for is whether there is evidence that the dominant platform is doing something to make it too difficult for entrants to get that critical mass. Is it entering into exclusive contracts or engaging in some other kind of behaviour that makes it very difficult for these online platforms to get enough customers on both sides of the platform to get the thing ignited? There may be perfectly valid reasons for these dominant platforms having exclusive contracts and so forth. The activities that we are more concerned about are the ones that make it difficult for competitors to get enough critical mass to get off the ground.

Ariel Ezrachi: On the question of abuse, it might be worth mentioning that under European law abuse is not a sealed concept. We do not have a clear indication of exactly what is in that box labelled “abuse”. We know from case law that abuse can take place when some damage is caused to consumers and that it does not have to be direct damage; it can also be damage to the competitive process. It can be either exploitative abuse, where there is direct damage to customers, or exclusionary abuse, where what you damage is possibly the competitive process: the assumption is that, following the exit of competitor, you will have higher prices, lower quality and worse services. These are the starting terms of the definition. How one might go about defining a violation, or what should be considered abuse, is closely linked to policy, and that might differ between jurisdictions, since they might have different ideas about what competition law is set to achieve. Some jurisdictions take a more economically oriented approach, while some jurisdictions around the world might take an approach that embodies other values. We might go into that later. It is not a question that can result in a very clear answer, because it is a dynamic area where the competition agency can explore an open list of possible abuses. Some things in the online environment give rise to some difficulties and types of abuse that might be slightly different.

I agree that from an economic perspective we are using the same tools, but we should be asking, “Despite us using the same tools, what might be different? What other things might be slightly more unique to this area?”.

I guess I am going back to the issue of quality. We have to appreciate quality. When I say that quality becomes an important parameter, of course it is not that quality is not an important parameter in other areas, but in an online environment where price is not a leading parameter, quality becomes one. The argument you will often hear from the leading platforms is that the competitive pressure will force them to self-regulate and police their quality. That is true, of course, because they are striking a very delicate balance: they want to attract, because the data and the individuals they attract are used as some sort of value for the other side of the market, the advertisers. What is often given as a possibility or a safeguard that will protect customers from a reduction in quality is the fact that users might engage in multihoming or sampling, which basically means that if you are not happy with your search results, you will just open a different web browser. However, research and empirical evidence from the US indicates that most users, when asked how they react if the search results are not exactly what they expected, say that they will then try to change the search query – not the search engine. That indicates that most users actually perceive the switching costs to be much higher than we perceive them to be, so there is a sense here of a default choice. That links to the question of dominance. It means that platforms might be more powerful than we think, despite the area being very competitive. It also means that there can be types of abuse that are very subtle. I am not suggesting that the competition agencies should act upon all of them—that might be controversial.

—but it is enough that a search engine, for example, could downgrade the search results of a certain business in order to increase the income from the sponsor ads. That can have two effects: as long as you as a user believe that you are still getting the result, there is some sort of effect on what you are getting yet the search engine would make more money. To the same extent, the search engine can basically hold up the seller and, by downgrading the results, force them to bid for search words in other to upgrade its results. So you can see possibilities in that dynamic. Whether these things should be called “abuse” is a different question involving the level of intervention. We have to be very careful not to intervene too much, but as to whether there is the potential to use market power—of course there is.

Exclusionary tactics are quite difficult when we think about the online environment. If you have a super-platform that is the gatekeeper to a certain market, and that platform decides to promote its own services, it can engage in exclusionary practices that would push out smaller platforms or applications. Of course there is nothing unique here: the refusal of access or refusal to supply can be brought in under a very general category that we use in competition law, but it is unique in the sense that this is an area where we have fewer and fewer super-gates, and if we accept that there is an issue here with limited switching and perceived difficulty in switching, that might be a possibility.

I shall give one last example: perfect price discrimination, the possibility for a platform or application to harvest information to the extent that they can basically identify something that is very close to your reservation price, which means that they are able to know how much you are likely to spend on a given product. If you are likely to spend more, you will just have to pay the display price, but if they know that you have some reservations—if your history, the data that were gathered on you, indicates not—as you see that, you will immediately also get a coupon. It does not work by increasing the price if you can pay more, because I think many of us would feel quite uncomfortable with that, but it can work from the other perspective of providing discounts. Whether this should be classified as abuse is a different question that might be answered differently by different people with different backgrounds and different political and social contexts, but there is certainly a range of behaviours that we can identify in this area of online commerce where you can find instances of someone able to use or leverage their market power.

Lord Rees of Ludlow: Surely it is an abuse if a search engine orders the results depending on who advertises on that search engine.

Ariel Ezrachi: Such degradation, when all it does is push someone down the rankings, is an easy example. I was thinking of something else. Companies are sophisticated and they know exactly where the threshold is. However, what about a different positioning of the search results, the ad results and the way in which you structure your index pages? Very subtle changes to the way they display the answers could have an effect.

I want to add something about the power of default. When we speak about human biases, I think that sometimes we do not appreciate them fully. In one of our meetings with one of the competition agencies, we were told that it had a list of more than 100 biases. One of them is basically the default—the fact that many times we will just not try to look for an alternative because we perceive the switching costs to be higher. This is why, for example, a lot of the search engines will pay a lot of money to super-platforms to be the default search engine. I might be wrong on the numbers, but according to what I heard last year Google paid around $1 billion to be the default on the Apple browser. That money reflects value, that value reflects bias, and I think that bias reflects that this is an area where it is easier than we sometimes think to abuse us. However, whether an act is legally an abuse is a different question.

Lord Green of Hurstpierpoint: Professor, when you talk, a lot of what you say makes this sound like a problem not a million miles from that of the supermarkets. There is a lot of competition among them, but they have tremendous buying power from their own suppliers. If there is activity that might be deemed abusive, it is almost certainly on the supplier side, not the consumer side. A lot of what you describe suggests that that is quite an interesting analogy, until you get to the point about perfect price discrimination, which in their wildest dreams supermarkets would not claim to have, and perhaps also the power of default. Those two aspects of what you have just been saying are more unique to large online platforms, while much of the rest of what you are saying would be characteristic of the supermarket industry. Is that fair or not?

Ariel Ezrachi: To some extent, yes. Buyer power in the retail industry is a big issue. We often speak about the way in which buyer power might impact on the upstream providers—not always as a competition problem, by the way. Competition law will have different thresholds for intervening in the case of buyer power. To that extent, I agree that there is possibly an issue with buyer power, but even the degrading of search results, if you accept that price is not the only parameter for welfare, could have welfare effects on customers if you think about allocating resources in the most efficient way. If a search engine can promote allocative efficiency by providing me with the thing that matches me most but I am encouraged one way or another to purchase something that satisfies my needs, yet it is not the item that fully addresses them, there could be a problem. Again, though, I highlight that that does not mean there is a legal problem; I am just highlighting a possible dynamic. Whether this should turn into a legal problem is a much more challenging question. The triggering event for intervention would be hard to establish, and in dynamic markets there is always a worry about whether we are chilling innovation and competition.

Q27 Lord German: I want to ask about data and competition. The European Commission says that sometimes data is the price that people pay in order to have free tools on online platforms. If you think of that as being where the debate regarding competition lies, are the online platforms sufficiently open and transparent about the way in which they collect and use data for competition agencies and Governments to be able to make very clear decisions about how they move forward in regulatory ways?

David Evans: That is a great question. I shall start with the first part of it concerning data. It is of course not correct to say that using any of these platforms is free. The transaction is that the consumer gets to see content—in effect, the content is sort of the bribe that is being paid to you, the consumer, to come to the platform, so in some sense you are getting paid to come to the platform with the content you get to see: search results, news, entertainment or whatever. In addition to that, for the advertising-type platforms, you are providing data to them. That may be to a company like Google or Bing if you are using it for search, or it could be to a company like Yahoo! or any of the online media. You are providing data to them; it is something that you are in effect giving them. That is sort of the price that they are getting out of you. So it is a complicated transaction, and data is certainly an element of that for many of these platforms.

Before I answer the second part of the question, I want to return to the “free” thing. One needs to keep in mind that zero is a number. Free is a number; it is the price of zero. Some platforms actually have negative prices since they give rewards. Bing, for example, gives rewards if you search a lot on Bing, so in effect the price is negative. On some platforms it is zero. None of this is new; we have had free add-supported newspapers for 50, 100 or 200 years. Free is something that online platforms do, but so do shopping malls with shoppers, as do ad-supported newspapers and so forth. We should not get too hung up on free; it is a dimension of competition, and an important one. It does not make price irrelevant because the decision to make something free—or negative, or positive—is an important element of competition for these online platforms.

Then you come to the transparency of the online platforms. I think that varies. My assumption, any time I go online, is that it is quite transparent. If I go online, someone will take information from me and serve me an ad that I do not want to see, and I will probably say that it is intrusive. It is transparent; it may not be to our liking, but how it works is transparent to anyone who uses the internet on a regular basis. That does not mean that online companies do not do bad things with data; there are markets where consumer protection law and regulators ought to intervene, but by and large we all know the rules of the game. You go online, your data is fair game, you get served ads.

Lord German: But are those who collect the data sufficiently transparent about what they have collected in global terms in order for regulators and Governments to be able to say that this needs to be regulated in some way or it does not? Is there enough information available to Governments?

David Evans: I think so. What information is collected varies from platform to platform. For example, when you download an app on Facebook there is a clear list of things that you are agreeing to give to the app. Regarding other platforms, I do not necessarily know whether consumers have a firm list in mind when they sign on. Obviously, when cookies are inserted in your machine, there are issues as well. I am not sure that I am prepared to give a blanket answer for all platforms—some are more transparent and better behaved than others.

Lord German: The final question in that sequence is: if it is variable, how can Governments and regulators see the variants.

David Evans: When Governments have a cause for concern—for example, if there is evidence that a platform is doing something with data that seems inappropriate—they can open an investigation and make an inquiry into what data is being collected and for what purpose. I am not sure they can necessarily go online and come up with a comprehensive list of what information is being collected, but they certainly have the tools available to find out what is being collected and what is being done with it

Q28 Lord Green of Hurstpierpoint: If I may continue to worry away at this question of market power and how it gets used, one of the characteristics of some of the big platforms is that they are also providers of services on their own platform—in competition with others and providers of all goods, come to that. Amazon is one obvious example. Does that kind of situation where the platform is available to the consumer in two modes—as platform and then as supplier, alongside competing suppliers—pose special questions about abuse that we should be looking at?

David Evans: It poses special questions from a competition, law and economics standpoint, but as we have already heard it is familiar territory. That is what is known as the possibility of vertical leverage. In fact, I dealt with that in the US in the Comcast-Time Warner Cable case. On the one hand, Comcast was an ISP that provided broadband, but on the other hand it was also integrated into video programming. There were concerns as a result of that combination. That kind of vertical integration of businesses is pretty common. Tesco, like American supermarkets, has store brands, so in that sense it is competing with brands on the shelf. There are a lot of businesses where that situation exists. When that situation arises and the firm is dominant in one aspect of that business, it could use that market power to leverage into a position of dominance in the other area. That is a problem that we are attuned to in competition law. Over the years we have come to recognise that there are many pro-competitive efficiency reasons why these vertically integrated companies do various things. Competition authorities have become a little sceptical when they hear these leveraging stories. However, it can happen for online platforms just as for other businesses.

Lord Green of Hurstpierpoint: So you believe—I do not know whether Professor Ezrachi has any comments on this—that the existing toolkit for competition law is adequate to deal with that kind of situation, partly because it is not unique to the online world, which I accept, although it will clearly become more significant in the online world than it has been hitherto in the offline world, if that is the right phrase. Why should we be worried about perfect discriminatory pricing? A supplier and a buyer are in an arm-wrestling match, as it were, between the reservation price of the supplier and the reservation price of the buyer. The price will settle. Why should you suppose that there is potential abuse in that environment?

David Evans: Is the first part of that for me? I am not sure I agree with the proposition that vertical integration and the ability to leverage from one market to another are going to be more prevalent online than offline. They have been very prevalent in the physical world—in telecoms, energy and many areas—for a long time. The online size of the economy and the importance of online platforms will increase, but I am not sure whether I would necessarily characterise that as becoming more of a problem going forward.

Lord Green of Hurstpierpoint: In a number of the instances you have just cited, competition authorities, one way or another, forced unbundling as a way of dealing with some of the issues.

David Evans: In some cases, that is correct. Many of the mergers that come before the European Commission or CMA end up having some vertical component. It is more widespread once you dive into these offline businesses when it might immediately become apparent.

Lord Green of Hurstpierpoint: But since unbundling is often the response where there is a perceived abuse issue, does that mean we should be thinking about unbundling in the online world?

David Evans: I think it should be on a case-by-case basis. I am not sure there is a blanket answer to that. If there is a serious vertical abuse, unbundling is a remedy that one could consider. It has been employed in the past only in quite extreme circumstances; it is not something the competition authorities or regulators leap to quickly, but it is certainly something to be considered. I am not very exercised about price discrimination. It is a way that businesses make money. Some consumers end up paying more while others end up paying less. There is no consensus among economists or competition authorities that price discrimination is a problem, and that can have some benefits. It is not at the top of the list of things that the competition authorities worry about.

Q29 Lord Liddle: At one of our sessions, Annabelle Gawer talked about how small businesses sold their services and products through platforms but at the same time were also consumers of the platforms in one way or another. Ordinary consumers get these services free from platforms, but businesses may have to pay. Is this a problem that we should be worrying ourselves about? Is this an issue of excessive pricing in these multisided markets?

Ariel Ezrachi: I heard her comment, but as far as I am concerned it is not an issue. There is an issue with the possible abuse of the platform’s position if it is a gatekeeper and is able to force or incentivise someone to pay more, but I am not sure I would define that as excessive. There are instances where you can see some sort of bottleneck and a platform becomes an essential platform that can charge more from others in order to display their goods or services, so in that instance it could be the case. In a market that was working well, you would expect a new entry or some dynamic process to work. In a market where you have economies of scale and network effects that have already created the giant that everyone relies on, maybe there is room for intervention. However, I do not see a specific problem with the price unless the platform becomes an essential gate into the market. If I may, I would like to make some comments on some other items that have been raised.

David Evans: Can I just respond to that first? The notion of excessive pricing on one side of the platform is fundamentally misguided. I shall give you an example. I shall move away from online platforms and talk about something that we are equally familiar with: shopping malls. Suppose that you have good competition between shopping mall developers. The shopping mall model is that you make it free to the shopper; the shopper can come to the mall for free. The way that shopping malls make money is that they rent space to the stores. Well, the shopper is getting it for free and the retail stores are paying rent. Do you conclude from that that there is excessive pricing? If there is competition in shopping malls, and stores have the opportunity to go to different malls and different places, overall the prices for the mall are competitive. What is happening here is that the mall is getting a fair price overall. That just happens to be the case with shopping malls, but it is also true for any platform that is charging a price of zero on one side; it is making money from one side but not from the other, so one side is in effect getting subsidised while the other side is paying the money. To one side the price is always going to seem high and to the other it will seem low, but that is not an exercise of market power. Overall, when you take both prices into account, the platform is not charging excessive prices. You need to take the platform as a whole, with both its sides, into account in order to assess whether there is any problem with the prices and whether they are above a competitive level.

Ariel Ezrachi: I would agree with that so long as the shopping mall is not the only shopping mall. As long as you do not have the only shopping mall in the country and no ability to sell any other way. To go back to the online reality, as long as you do not have one platform that you absolutely must sell through, that would be true. I should also mention that Europe takes a slightly different approach to excessive pricing from the US, so there is also a philosophical difference between the two jurisdictions.

I have a comment about the question on data that you asked. Most people would agree that data has become a currency that we are using. Commissioner Vestager mentioned it several times. That assumption is common. On the question of whether we have transparency of data, my answer would be slightly different; it would be no, there is not. Are we moving in the right direction? Yes, we are. Do we really know where we stand and what is being tracked? No, we do not. I will give you several examples. First, Google argues—this links to quality as well—that it has to maintain quality because it is such a competitive market that it will lose its position, yet there have been cases where it has had to pay fines in the US for misleading users over privacy and use of data. That did not affect its dominance. Unfortunately, we are not very sophisticated users when we use those websites; we just click something and assume that everything will be okay. There is a long list of applications that would track you without informing you.

There is an interesting case at the moment in Europe regarding an application that was basically kicked out of Google. Google argues that the application, called Disconnect, violated the rules. Disconnect argues that Google kicked it out because Disconnect offered users the ability to stop being tracked by other applications. For example, if you go on to a newspaper website, according to Disconnect 17 different applications will immediately track your activities there. Disconnect gave you the option to basically stop that. This also meant that Disconnect undermined the business model that the whole ecosystem is based on, which is, “Harvest as much data as you can, because it is are the currency that we used to generate income”. I do not know who is right and who is wrong. We might know in a few months when we hear about what happened with the case. Disconnect might be the villain, or it might be the angel that tried to help us. What is obvious is that there is an issue.

I contrast that with another example, Brightest Flashlight. With that application you just get a flashlight, but when you used it you thought you were getting something for free, which was lovely. What you did not know was that that application was tracking your location all the time, even when you were not using it, and that information was sold to third parties as part of harvesting. That was Brightest Flashlight business model. Again, with regard to the question of whether there is anything we can do, the FTC, with relatively limited tools to deal with that, eventually reached a settlement with that application. The application changed its privacy policy and it is now more obvious what it is doing, yet I believe the default is still that you allow it to track you. Many applications, even if they are used for a single purpose, when you click “OK” will track you way beyond what you expect, because this is the currency. We might have issues with that not just from a competition perspective but from a completely different perspective of consumer welfare and protection, but the engine, the fire at the heart of this market, is definitely data.

The Chairman: Thank you. We are aiming to finish at about 5.30 pm, so we need to hurry up a bit.

Q30 Lord Wei: We have received evidence from some industries, such as the hospitality industry, about the pricing policies used by bottleneck platforms that for example allow suppliers on them to offer a lower price for other channels. We have also been referred to examples of platforms using their own terms and conditions for example to remove suppliers from their channels without any explanation or dispute settlement mechanism. Is there a case in these examples for requiring more transparency and accountability on the B2B—that is, the business-to-business—side?

Ariel Ezrachi: I think so. You mentioned the hospitality sector, and I assume that you are referring to parity. Very briefly, because it is one of those big areas where there is a lot of economics and policy, the problem with vertical relations, when one entity is giving services to another, is that there is some room for externalities, and in order to facilitate the services and to create an environment where one entity will provide services, you somehow have to limit competition to some extent - to limit vertical externalities, horizontal externalities and the hold-up problem so that there will be an incentive to invest. That is a typical vertical situation. To resolve that, parity clauses are often introduced, which ensure that no one will free-ride on the investment that you as an operator have made in demand-enhancing features. The question—and this is where there could be criticism—is whether the restrictions that you put on competition are proportionate, or whether they go beyond what is necessary. The debate these days is between wide and narrow parity. If we think about the differences between them, and I am not sure how much time I have on this—

Lord Liddle: Not a lot.

Ariel Ezrachi: I will take 30 seconds then. The question is whether it is necessary for you, as the company that invested in buying the search words on the super-platform, to ensure that no one else sells at a lower price than you, or whether it is enough that you ensure that the seller of the product gives you an assurance that it will not sell it at a lower price than you. So the question is: how wide should this protective net be? At the moment, the consensus around the world seems to be that wide parity is problematic and should be distinguished from narrow parity, where you create only a very specific link. There are theories of harm for both of them. I am oversimplifying it, but if the end result is that in vertical relations there will always be some justification for some limitations, the question—and we have to review this on a case-by-case basis—is whether that limitation is adequate and does not go beyond what is necessary in order to facilitate the competitive process and the welfare-enhancing features that we, as customers, benefit from.

David Evans: I will just take the second half of the question, in the interests of time. The terms and conditions issue is really interesting. These platforms all basically have their own governance systems. You call it “terms and conditions”, but it is actually more elaborate than that. Back in 2009, a third of the workforce at Facebook basically policed the website, because the terms and conditions were, “Don’t have hate speech”, “Don’t have pornography”, and things like that. A large portion of the workforce was basically going over the site and making sure that it was clean, and when there were issues the violators got bounced, in that case from Facebook. You can think of most of these platforms as operating very large communities, where different customer groups come together on the platform, generally in some kind of virtual sense, and whenever you have a large community, people do bad stuff. The platforms recognised that, and almost all of them had not only terms and conditions but penalties, so that when people and businesses behaved badly they could bounce you off the platform, penalise you and so forth. What do we think of that? In general, we should think well of it, because as a participant in these platforms you generally want them to have your interests at heart, and you would prefer other participants on the platform not to do bad things. There is a recent example of this. TripAdvisor was criticised for not policing its platform well enough. One of its problems was that people would post positive reviews of their own properties, or get their friends to do it, in order to get higher in the TripAdvisor rankings. That is kind of an obvious thing to do, and it was fairly rife behaviour. It was criticised for not policing that enough.

So when I think about these platforms, I think of a set of things that you want to encourage them to do to make things work well. There is a possibility that they could abuse it; they could see a competitor and use the governance mechanism or the terms and conditions to eject it from the platform, so you have to worry about that. By and large, however, we would like these platforms to operate an efficient governance mechanism to make sure that the platform is convenient. I would not be enthusiastic about making that job too difficult by layering in too many restrictions on the platform. I guess the analogy I would give you is a bouncer in a nightclub. A nightclub is a multisided platform that facilitates men and women getting together, for example. Typically it has a bouncer outside the door, and when men get rowdy—obviously the women do not—the bouncer can eject them. The last thing you would want is a regulation that said that when the nightclub did that, the man could go through an appeals process in order to determine whether he really should have been bounced from the platform. That would be inefficient. Abuse of the terms and conditions is certainly possible, just as abuse of anything is possible, but by and large as a consumer of these platforms I personally would like to encourage to do more of that.

Q31 Lord Green of Hurstpierpoint: I wonder if I could put both of you on the spot and ask whether there are, let us say, three specific measures which the Commission ought to promote in order to enhance or protect competition in particular in the online world.

David Evans: I will start, if I may, with a perhaps facetious answer to that, but perhaps not.

The Chairman: Feel free.

David Evans: The first thing I would do is prevent American lobbyists and lawyers crossing the Atlantic in order to persuade the Commission to engage in investigations and other activities to constrain competitors. Honestly, an awful lot of what you see going on is in fact two kinds of things. One is that it is other internet players—some large, some small—trying to get everyone excited in order to restrain their competitors down the road. There has been an awful lot of that over the last 10 or 15 years in Europe. Most recently, Microsoft made a very large investment in lobbying and so forth. Again, I am not prejudging any issues, but there is an awful lot of that going on. That was my partly facetious answer.

The other thing that you see going on is entrenched incumbents in the physical world. Physical newspapers and publishing properties are obviously very threatened by what they see going on in the online world, and they are looking in many cases for some restraints on the online players just because they see their industries being challenged. Quite seriously, I think that Governments need to be very, very wary of efforts to impose restraints on competition in this space, while at the same time fully recognising that there are some large, powerful players that could do bad things. I totally agree with that, but one needs to be very careful about encouraging companies to file complaints just to shackle their competitors. That is point one, and the longer point.

Point two—I do not think I will get to three, so I will just stick to point two—is that the problem in this area is that competition authorities have slowly acquired the right toolkit and fully recognised that these multisided platforms, which most of the online players are, behave differently from traditional firms, and we need to customise the economic analysis to hold them to account. Courts and the competition authorities have been slow to recognise that there are multiple stakeholders here, so I would encourage competition authorities and legislatures to recognise that there are multiple interdependent stakeholders and that you need to take all their interests into account. If you do something that is going to benefit one group of stakeholders, you need to be sure that the benefits that they are getting are more than outweighed by the harm that other stakeholders—they could be consumers or small businesses—are incurring.

Ariel Ezrachi: I agree that one of the risks that we face is getting it wrong when we intervene. That is something that both the Commission and the UK competition authority are aware of. I believe that competition officials are aware of the dynamics of the market and the need not to send a signal that would chill competition, although there might be some differences in approach between Europe and the US in the level of intervention and the way in which they perceive consumer harm.

In terms of what could be done better or differently, the role of data is something that we still need to learn about and understand a bit more. I am not sure that we yet understand the way in which the data affects this ecosystem. We might at times give it too much weight and at other times not enough weight either in the analysis of mergers or the analysis of certain behaviours on the market. This is a relatively complex area. I think that competition agencies are aware of it, but more understanding is required of the unique dynamics that are developing in the online market.

The Chairman: We are now running over time, but I have two questions, one from Lord German on the competence of the authorities and one from Lord Rees on the general philosophy of competition. Perhaps we can take both of those together and have your final answers.

Q32 Lord German: You have probably answered the first part about whether competition authorities need more powers, skills, expertise and so forth, although you may want to add to that, but do you feel that in Europe we are spending a lot of time trying to analyse what competition means but not actually doing enough about it?

Lord Rees of Ludlow: What are the main goals of competition law? Is it just about establishing whether there has been economic harm to the consumer or are there other goals that need to be taken into account?

The Chairman: Quick answers please.

David Evans: Thank you for those succinct questions. I think that the competition authorities are fully set up to deal with this sector. They have been doing it for years and they know how to do it. I made some comments earlier about improvements that I would like to see, but I think that the competition authorities are very well situated to handle the online world. They are a much, much better alternative than considering regulation. They are much more flexible and they know how to deal with these businesses—they have a proven track record in that. Are we spending too much time on this? Yes, I think that we are.

I am not going to answer all the questions; I will just end on this point. I could have had most of this conversation with you in 1995, 2000, 2005 and 2010. Every time I have had these conversations in the past, there has always been some parade of horribles: some company is going to dominate everything and the world will be horrible. Then you look five years forward and everything has changed. Just think about where we are today. Seven years ago, Apple was not a significant player in this space. Today it is enormous. Two-thirds of activity that takes place on smart mobile phones takes place on the Apple platform. That was not true seven years ago. Apple is a big player in this space now. Seven years ago, Facebook was behind MySpace. Now it is an enormous, interesting closed platform. Seven years ago, Microsoft was at the top of the hill. It was an extraordinarily important player in this space. Today, it has been gutted. It is obviously a very profitable company, but its position in PCs is declining because everyone is switching to smartphones and mobile devices. It is not nearly as significant a player in this space now as it was seven years ago. Things change very rapidly. That is a real caution to intervention, because how do you intervene when things are changing so quickly and when you have a history over 20 years of things changing in massive ways every few years?

The Chairman: Professor, last words?

Ariel Ezrachi: Last words. As to whether we spend too much time, I am not sure that I agree. There has been a change and I absolutely agree that we need to be extremely careful not to overdo it and not to intervene too much, but I do not think that there is a real discussion at the moment of over-intervening; the discussion is about what the new forms of harm are, and where we are headed. So the discussion is forward-looking. How are the dynamics changing and what are the risks? I do not think that anyone is standing there assuming that we are going to become extremely interventionist—at least, I hope not.

That question links to the other question of what competition law is about. Competition law at the core is very similar around the world, although it can be rather different in different jurisdictions. It is sometimes referred to as a sponge—it can absorb different social and political goals in different jurisdictions. To some extent, the way you understand the role of the competition agency and competition law in a given jurisdiction will impact on what you think is the right answer to the question of how much we should look into that. I guess the question of what competition law is about is very complicated, as is how that answer might impact on our decision about how much we should think about it and whether we should intervene. I will stop with that.

The Chairman: Thank you both very much. You have given us quite a lot to think about. You have answered some questions but you have raised a lot more. That is much appreciated. Thank you also for being prepared to overrun your time.

[1] Baroness Noakes was present in the public gallery as she has recused herself from the inquiry.