Debate: Panel 3, Competition and Regulation in the Data Economy, moderated by Laurent Vallée

Presented at the Legal Challenges of the Data Economy conference, March 22, 2019.


LAURENT VALLÉE: Thank you. Thank you very much because I think, and I'm sure that the audience will fully agree, that this panel is the epitome of sexiness in a way. Just to launch the, maybe not the debate, but one first question, even if we talk about deluge of data or yottabytes. I discovered yottabytes today. And going back to some of my initial questions, it seems to me that you would all agree that when it comes to antitrust or merger control, the answer about data economy doesn't lie in new tools or a new revolution of tools, but slightly adapt tools that exist.

And to base this adaptation on cautious analysis of case law, even if some regulators, as you said, Professor Picker, might have regrets when they think about today about the Facebook and WhatsApp decision. Just too much data. Would you all agree? Or would you have some comments on this idea?

RANDAL C. PICKER: I was delighted by what I saw as the consistency across the conversation. Yeah, exactly.

LAURENT VALLÉE: It's key for me, all your intervention.

RANDAL C. PICKER: I agree with that. That's not to say we can't learn things. And so I think that's different. But in the United States we talk about the common law method. We've been learning in antitrust since we passed the Sherman act in 1890. But that's a framework in which we, as you just said, sort of work with each situation and look at it and understand it. We bring to bear the tools we've been working. Sometimes we'll screw up but I'm very comfortable.

LAURENT VALLÉE: Yes, please.

ETIENNE PFISTER: Yes exactly. I think exactly the same. We do not need major new tools. Maybe some tools can be slightly adjusted. Like a notification threshold, for instance, which I mentioned earlier. The possibility of interim measures should also be explored, especially. Sometimes the complainant does not dare either to complain or to ask for an interim measure. So sometimes it is up to the competition authority to take this responsibility. And not only to the plaintiff, because nowadays it is only the plaintiff who can ask for a preliminary measure. But apart from that, it seems to me that we do have the tools.

ANTOINE CHAPSAL: Yeah, I do totally agree too. I mean, we clearly have the correct framework. And the framework is flexible enough to be adapted in a way that we can take all the issues that we discussed, actually. So I think it will be probably a major mistake to try to innovate, and try to tackle this issue without taking into account what we already know because these are too complex issues to have the opportunity to innovate.

ALEXANDRE MENAIS: Let me just to comment some. Fortunately, we do not have a European champion in that field. Fortunately, we go more and more in verticality in social business. Fortunately, we miss a bit at your prevail. And I'm not promulgating that the tax and the data will fall together. Let's be clear. So I'm not saying that we need to use competition automatically as a barrier, of course. It's very complex but don't get me wrong. The majority of the revenue of those companies are not enough. You know the thing that we have seen.

RANDAL C. PICKER: So I do want--

ALEXANDRE MENAIS: Do you understand my point? The majority of Amazon revenue is no longer in that field. You know that.

RANDAL C. PICKER: So I do want to say, I think there's a category question. And you talked about this a little bit. So is it a competition question? Is it a privacy question? Is it a sector question? So are these natural monopolies? And you could easily conclude that they are. And we go, well that's not an antitrust issue. But that's not to say the law is done, at all. Yeah. OK.

LAURENT VALLÉE: OK. Maybe there are already some questions in the audience.

AUDIENCE: Yeah. It's more a remark than a question. But something that you didn't address today, which is to me one future very interesting question, would be how to assess the value of the data. We know that the gaffers are working on that. They are working with people like Anton [INAUDIBLE] or the big four, big five. I don't know how many there are now. But this would impact, probably, the competition arena as well.

LAURENT VALLÉE: Someone? Anyone?

RANDAL C. PICKER: The only thing I'll say on that is, look, I think you're going to see on exactly on that point. And I think the California statute that I just blew right by does some of that. I think you're going to see, in the United States at least, and maybe in California in particular, an effort to figure out ways to extract more value from these companies.

And if you say, well, you might say, well just do that through normal tax policy. But if you think normal tax policy is going to apply to everybody and you're really looking for ways to isolate these companies and tax them somehow, then a valuation of data approach might be some way of doing it. I'm in this weird position where one of my law school classmates is currently running for president. Amy Klobuchar. And she is, I think, open to that idea. And that will take us down a valuation approach, I think.


AUDIENCE: I'm sorry. I have another question for you, Professor.

RANDAL C. PICKER: Maybe over here.

AUDIENCE: Yeah, well, you can of course comment. But this is an idea you've just shared with us. You mentioned the idea that we should maybe focus on users rather than data or services. And I think, first of all, it's a very good point, notably because, for example, in this case where Microsoft bought GitHub.

I don't know if any one of you know GitHub but it's actually one of the biggest platform for coders in the world. It's an open data platform. And in this case, the commission had to review the transaction and cleared it absolutely without any condition. On the same basis as the one used for Facebook, and other cases were mentioned today.

So but why Microsoft bought GitHub. It's basically because of the number of users of this platform. It didn't buy the codes, which were basically open to everyone. So I think it's a very interesting idea.

And actually I would like to ask you, what do you mean in how we could take user better into account? Is this something we can do with behavioral economics that we've been using several years to assess the behavior of consumers or clients in competition. No? I just would like to know more about this idea you just raised.

RANDAL C. PICKER: Well, so I-- Did somebody else want to jump in? No? OK then, I'll talk.

LAURENT VALLÉE: I think it was you, Professor.

RANDAL C. PICKER: Yeah. Yeah it was. Thank you very much. So look, I think it's I said what I said. And what I mean by that is-- So I think the user's point strikes me as right. So when you ask, why did Facebook pay so much for WhatsApp? I think that's in the UK report, says exactly this as well. I think it's a really good point. What are they seeing there?

And we should as regulators-- and I gather we have a genuine regulator here now. Is that right? OK. OK. [LAUGHS] --we should ask that question. What are they seeing that we're not seeing? And how do we assess that?

And, look, either they saw them as potential competition that they were trying to kill off. And we should be troubled by that. Or they saw it as a way to, maybe it's a data advantage story. That's your story, right? Maybe that. I don't know that that's necessarily a bad story. If I have this very valuable thing that actually lets me do advertising better, and they've clearly gone into the infrastructure business, maybe that's a good thing, not a bad thing.

The other thing I want to say is the fact that you've got a lot of users doesn't mean you're necessarily going to succeed. So watch Google's abysmal social strategy over time. The whole Google Buzz fiasco was because of that. Google Plus, nothing. So and Twitter seems to be just Twitter and nothing else.

So I don't think having users is sufficient. I do think having this powerful network of users is hard to do. And when you asked me why did they want to buy WhatsApp, that was my answer. And that they were looking down the road. And we as regulators should have been as smart as they were, I think. I don't know if that's helpful or not. Anybody else --there's a hand back there.

AUDIENCE: [INAUDIBLE] --the GDPR in Europe. On an economic and more competition standpoint, I would like to know if the authorities are currently thinking of a similar principle like we have for the free circulation of goods, services, and persons? For data? Your knowledge and what would you think of that? Thank you.

ETIENNE PFISTER: Well, obviously this is a question that should be handled at a supranational level because if only one country does it, it may not be of much use. That's the first point.

The second point is that we-- it's a difficult trade off. Really. It's a difficult trade off. When you're thinking about data like an essential facility, for instance, and saying OK, Google, Facebook keep all your data and so on. You may think. But it might be less risky than for other kinds of assets, for instance, data and nonrival. So the fact that Google gives its data to a competitor, it does not prevent Google from using the same data. And also these data may be used for innovation, which may not compete with Google's products. OK. So that's a good point.

Maybe the third point is that transmitting this data is zero marginal cost. Good thing. So that's a plus side, but there's also minus side. As you mentioned, Facebook value can be very high for consumers. But of course, if you force Facebook or any other to give its data, then Facebook is going to change its model. And it's going to change this model over more so when it has some market power.

Another point that Antoine pointed out is that data is not the end of the story. There is data and there's eyeballs. The fact that I have many consumers watching us. So I'm going to make a bet on the fact that I'm going to make data circulate freely. But I'm not sure of the result. What if Facebook keeps its dominance? Then I'm going to have a Facebook which is more costly. And I may not have more competition anyway. So it's a very, very difficult trade off. It needs to be tackled and carefully, carefully examined but it's not an easy task.

AUDIENCE: Thank you very much. I guess my question is to all the panelists, I would say. And I feel like I'm mentioning the elephant in the room, which is the recent discourse on let's break up big tech companies because, I guess, there is this realization that maybe the mergers were-- that we just mentioned, Facebook, WhatsApp, Facebook, Instagram-- they shouldn't have been cleared in the first place.

And now there is this trend or, I don't know, hype around let's break them up. What do you think about that? I mean, I personally don't support that. But clearly there is something wrong going on. And disintegration, can this be thought of as a valid, I guess, remedy in the data economy? Thank you.

LAURENT VALLÉE: The idea being, I guess, that making competition against each other might offer me a service that is less privacy intrusive, that is--

RANDAL C. PICKER: Oh yeah. I'm not sure--

LAURENT VALLÉE: --the same point, I guess.

RANDAL C. PICKER: If someone else wants to jump in?

ETIENNE PFISTER: They're two different points. But to answer first the remark which was made. One of the remedies that we are thinking about is making some ex post decision on mergers. That is, because we are in a world of uncertainty when looking at these data related mergers, because we have so many things moving on very quickly, one solution would be, indeed, to maybe let the merger happen, especially if it is a small acquisition, for instance.

And look at the effects in, let us say one year, two years time. What's happening? And then, indeed, take the ex post measure which is deemed necessary to reintroduce competition, if competition has been abused. It's one kind of remedy that can be sought out. So this is one solution, which is on the table.

Again, I'm coming back to notification of threshold. The thing with Facebook, WhatsApp that has to be borne in mind is that actually Facebook lied twice, in a way. It lied the first time, and it has been condemned for that, when it said I cannot merge for three data sets. And then it lied to consumers when it said, I will not transfer your data to Facebook, and will not transfer. So this also has to be taken into account.

RANDAL C. PICKER: So as a lawyer we call it, potentially, a negligent misrepresentation, I think was the call. The commission did not call it a lie.

ETIENNE PFISTER: Absolutely. That's a good term. Thank you. Thank you, Professor.

RANDAL C. PICKER: Sure I'm happy to address the breakup question. But If someone else wanted-- No, no, go ahead. Yeah, yeah.

ANTOINE CHAPSAL: Just one comment on the ex post evaluation of merger. I think-- well I think this is really interesting. But it's a very tricky point because you can make the case that you have a merger between two companies. You wait and two things can happen. Either it works perfectly well or it's a disaster. And, well, if you see that it works perfectly well, you will see, oh and so we see huge increase in market share. And so on.

Well, you will say, well, this merger is anti-competitive. But it's not the merger that is anti-competitive. This is the fact that the people work to make the merger and the majority work well. You know what I mean? So my question is it will be very difficult to disentangle what is something that is caused simply by the merger and something that it is exogenous to this. And I would rather prefer, actually for instance, a turnover threshold which I think it's--

I'm not a big fan of the SEC's ante, ex post distinction. But, well, if we do ex post merger regulation, I think it's going to be tricky. But it's probably what we-- you are-- planning to do so.

ETIENNE PFISTER: You have several countries doing this kind of ex post examination. United States do it. United Kingdom do it. Sweden does it. So you have already many countries doing it.

It also depends on the kind of measure that you're taking to do your ex post measurement. The issue is not so much about market share because a company that has merged may gain some market share because it is more efficient, because it provides more services, because it attracts new consumers that were just not in the market before. But there are many other metrics that we can look at. For instance, if the margin on the advertising market, how did it evolve, for instance. This Is the kind of stuff that we can look at. How has the privacy evolved?

Even though I have to mention, but there is a really tricky trade off, again, between privacy and better services. When we look at cases, we always look at the worst scenarios. What is a negative story that could happen? And it is only if we think that the negative story may happen that we begin to look at the positive side effects of a story that looks negative.

When I talk about the data advantage scenario, for instance, it's all about a competitor being more efficient. So we have to first be wondering will it become so efficient that it may drive out some rivals? And first we look at this side of the scenario. And afterwards we look at, wait, it may drive out some rivals, but it may also bring some more efficient products and so on. So, both sides. But very often, only one side is enough to make a conclusion about a case.

ALEXANDRE MENAIS: For me, just again, that's usage. This is really the topic. And it's nobody. And when you set up a platform and-- you said that perfectly-- most of those platforms before-- I used to work for eBay-- I can tell you that it was impossible, at one moment, really, to anticipate what would be the usage by the user of the platform. So we need, to the point and to the point, that probably user would be the key driver. I'm not able to explain why, in fact, my friend is going to Instagram and me, I stick with Facebook and things like that. This is usage, usage, usage, and, in fact, technology, the majority of the time. And in particular, you know the way the trend is, I'm sure that would be probably one of the key driver, to assess usage of the platform.

LAURENT VALLÉE: Maybe one or two last questions.

AUDIENCE: Thank you. My understanding, but I am not completely sure, about what the panel, is that they'd say, competition is policy. It's good enough with some few modification to deal with the digital issues. And regulation is not the solution. It's my understanding but maybe I'm wrong. If it's that, if I understand, what do you think about the solution proposed by the Furman Report?

RANDAL C. PICKER: So yes. So I agree. I want to say, and I'll just try to be clear about it. I thought what I heard said here was is that when we are working within antitrust, we have the right set of tools to do that. But that-- it's what I'm trying to say. It's a category question. Maybe antitrust is not the right lens to be looking at these markets.

And that's what the Furman Report says. I mean the Furman report, says exactly what we just said, which is they like traditional antitrust, but. And the but is they say, we need more procompetitive tools. I had a slide-- Though I had 44 slides so how could you remember them all? I had a slide in which I quoted from that report. And that report calls for, I'll call it data sharing, as a tool for doing that.

And so that is very much a kind of access regulation that we associate with regulated industries. And I think what I tried to suggest is I completely understand why you might think that's the right approach. I do not think that will be easy to do. It's not to say it's the wrong undertaking. We just ought to to be prepared as to how hard that is.

And I do think the Cambridge Analytica situation and the fact that Facebook set up this platform API and then the consequences associated with that. That's what we ought to be aware of. Now, I don't think Cambridge Analytica is a big deal. I just want to be clear. I just do not. So that doesn't scare me, but I don't know that that's the majority view.

LAURENT VALLÉE: Maybe one last question.

AUDIENCE: Thank you. Just one quick last question. So we say that data is nonrivalous goods. But yet, under GDPR, when data is collected by a company to be processed, it can be processed only by this company. How do you articulate that because this company can not transfer data to another one. It cannot be used for another purpose. So we kind of have a weird articulation between the fact that the data has to be processed only for one thing, for one purpose. And the fact that we assume that it can be reused by another company.

ANTOINE CHAPSAL: Yeah, probably, just one kind of economic answer to your question. I mean, actually, it is welfare increasing to share nonrival goods. So-- but here we have an issue, which is privacy. So there is a trade off between the fact that data is nonrival and could be shared. And the fact that everybody decides to share data will be, probably, welfare increasing because this is big results from the economic theory.

And on the other hand, we have this privacy issue. And I think this is one economic lecture of the regulation. The regulation has to deal with these two kind of opposite objectives. One, ensure welfare and ensure that you can share your data because this is welfare increasing. And the other one is-- I don't know if it's consumer protection. And it's two different kinds of law. I mean it's not my problem, if I may say. But the thing is, I clearly understand that we have two opposite objectives and the regulation is trying to deal with these two opposite objectives.

LAURENT VALLÉE: OK. Well, let me say thank you very much to the four of you and thank you to the audience. Thank you very much.

Big data