Merger Policy with 2020 Foresight: Fireside Chat

Below we have provided the full text transcript from the Fireside Chat portion of our live-streamed conference, Merger Policy with 2020 Foresight, from June 6, 2020.

Steven Salop:

Hello, I’m Steve Salop. I teach economics and law at Georgetown university law center. I’m here with Frederic Jenny, who’s the chairman of the competition committee at the OECD. And we’re going to have a short fireside chat here in the middle of the summer on antitrust policy in the US and in Europe. So Frederic let’s start just by you framing the debate with respect to the merger policy in the US and Europe. Okay, So there have been different levels of discussions in the UN and the US about merger policy. And I think that one of the reasons is that in the US there has been quite a bit of concern about the increase in what was seen as an increase in aggregate concentration, increasing profit rates, increasing what was perceived to be a decrease in competition.

Frédéric Jenny:

And therefore, there’s been a questioning of merger control, and whether it had been effective enough, whether the courts work too tough, whether the agencies were not daring enough, et cetera, et cetera. In Europe. I mean, first of all, I’m a bit skeptical about those statistics that are used. We had a debate on this at the OECD about two years ago. What is clear is that in Europe, we haven’t had much of an increase in concentration. Well, we have had an increase in profit rates in some industries like banking or finance or pharmaceuticals, but it doesn’t seem to come from the fed that concentration has increased. The second thing is the fact that only in Europe the EU Commission has not been shy about merger enforcement, not only horizontal merger enforcement, but as well conglomerate, mergers, et cetera. So we don’t have the same debate in Europe, but we do have a debate about whether, in fact, there has been too much enforcement, and whether competition authorities, and the Commission in particular, have taken too narrow a view of markets and too short perspective on mergers. With the results that there’s been an increasingly intrusive, emergent control, which has prevented large firms from growing. And of course this was exemplified by the Alstom Siemens case. So there is criticism also in Europe, but it comes from a different angle than in the US.

Salop:

Okay. Well, I want to talk about this later, this latest merger decision by the General Court, but let’s talk more generally first. One thing that I very different about the US Europe is the fact that in the US the agencies need to go to court in order to stop a merger and in Europe the Commission can do it on its own, and that could make a big difference in the way in which outcomes play out. So let me give you some data from the US. In the US, despite the fact that there are more and more mergers over time, and they’re bigger, mergers agency budgets have basically stayed the same. And the number of second requests, number mergers that are actually intensely investigated every year remains pretty much, pretty much the same and of the ones that aren’t investigated, almost every deal is challenged. So it looks like the agencies are only picking the low hanging fruit that is the most egregious deals there are. And very few mergers that get second requests are cleared as is most of them either get consent decrees, or they’re abandoned, or they go to court. And when the agencies go to court, or in fact if they issue a complaint if the company does not agree to a consent decree, the government almost always wins. Since 2012, I think the US agencies have won more than 30 cases including ones where the firms abandoned the deal, and they’ve lost only four. So it looks like in the US they’re so constrained by lack of resources and risk averse with respect to going to court, that they only bring the very easiest cases, whereas in Europe it’s quite different. So I was wondering if you felt that that was a big difference between the US and Europe.

Jenny:

Well, I think there are several layers of differences. One of them, of course, is that the perspective on mergers in general and large firms is quite different in Europe, not only for the institutional reasons that you mentioned, but also because we are more sensitive to Type 2 errors than is true in the in the US and we are less concerned about the Type 1 errors. So, that leads to a higher level of enforcement in Europe then there is in the US for reasons which are not only the institutional reason, but based in history and different history of the way the US development and the way Europe developed. So I think that this is also quite an important factor. Now, the fact that as you said, we have an administrative system in Europe means also that the challenges to mergers are going to be very few and far between because mergers need to be solved fairly rapidly. The Commission takes quite a bit of time to but it has constraints on the time we can spend on looking at mergers, but the courts have very few opportunities to actually say something on mergers, which is by the way, one of the reasons why the O2/Three merger was quite interesting because this is a case that finally went to the court on non-coordinated effects of mergers. And this was the first time that the European court was able to say what they thought of the 2004 modification of the standard for merger review in European law. So less involvement of the court. The Commission always wins in that sense. And it’s a general idea that power – economic power – is somehow threatening and that therefore intervention should take place. So any merger that might either vertically, conglomerately, or horizontally be a threat or an increase in market power is more easily challenged in Europe.

Salop:

Yeah, I certainly agree that in the US we worry more about Type 1 errors, that is, false convictions, than we do about Type 2 errors, those false acquittals and under deterrence, and the institutional factors really reinforce that. You know, because the fact that you need to go to court that tends to lead to more Type 2 and fewer Type 1 errors, it sounds like in some sense you were saying before that maybe the right approach, the optimal approach would be somewhere in between the US and the European approach. Is that where you’re kind of coming out?

Jenny:

No, I think we have different traditions and different histories, and therefore there may not be a magic level of enforcement. I think that what happens is that in the US there is the feeling that there’s been under enforcement, given the fact that we are in a system, whether it’s concerned about Type 1 errors. In Europe, there may be the feeling that the Commission went overboard, given the fact that indeed there should be a more activist merger control because of our history, but that it still went way beyond what should have been done. So, I think it’s relative to where the cursor needs to be, but of course it does need to be same place in US and the EU.

Salop:

So, a lot of people think that the US and Europe should harmonize their policies. It sounds like you think that’s not really possible.

Jenny:

There are limits to what’s possible, you know. Yes, they shouldn’t harmonize, but the fact is that European competition law is an appendix to try to promote or to succeed in promoting the single market. So it’s not competition per se, which is pursuit. It’s just a tool to achieve something else. So there is a different perspective.

Salop:

Right, Well, it’s, we have, we have a recent case. That’s extremely interesting in this regard, that is the Sabre/Farelogix case, where the court in the US permitted the merger, whereas the CMA prohibited it. So that that’s, that’s a really, quite a striking example of the two approaches. Do you think the CMA got it right, or do you think the US court got it right?

Jenny:

Having been a judge myself in the past, I will not answer the question directly. I think that the courts have not intervened enough in Europe for the reasons that we’ve mentioned. This is why it’s extremely important when they do get the opportunity that they say something. And I think that they rose up to the challenge in the O2/Three merger, which I was referring to earlier on. But what’s interesting is that the court, in this case, the European court, went in the direction of what I would say is the traditional position of US courts. It basically said non-coordinated effects, to establish them, you have to show two things. One of them is that the merging firms were the closest competitors, and second, that the target was a very active firm, very active in the sense that it was very important to competition. So it’s set pretty high standard and it reproached the Commission to have used lower standards, which was basically – I mean, I’m going a bit overboard – but which was basically saying, well, a move from four to three is necessarily going to create non-coordinated effects.

Salop:

Okay. So, let me make sure I understand this: the two merging firms have to be the very closest competitors in order for there to be liability under what we call the US unilateral effects theory?

Jenny:

Yes. I mean, the court said it’s not enough to say that they were fairly close. You have to have a massive reduction in competition between those two firms. So that’s one of the two effects. And second, you also have to prove the disappearance of the target is going to have a significant effect on competition for the other players in the market.

Salop:

I see.

Jenny:

In both cases, it raised the bar.

Salop:

Okay. So, the closest competitor requirement, that’s actually something that was eliminated in the 2010 merger guidelines in the US. The 1992 merger guidelines in the US suggested – didn’t quite say it, it was a, awkward paragraph – but it seemed to suggest that unilateral effects would only count if the two firms were the closest competitors and the 2010 merger guidelines in the US, quite visibly and strongly walked away from that standard pointing out that even if the two firms were not the closest competitors, there could be a very substantial impact on competition. So the two jurisdictions have really reversed positions on this one.

Jenny:

Well, on this one, yes. What had happened in the O2/Three merger was the fact that the four firms, which were in the market, really were as close competitors as you- I mean, there was really not much difference between the four. And so, the Commission decided that therefore going from four to three would have a severe effect. And that’s when the court said, “I’m not satisfied that this is enough. Just because you acquire a competitor, it’s not enough for me to think.” And the reason why the court said that was because it said in 2004, we kept the dominance idea, but we added the possibility of a reduction, a restriction of competition. So we have to think about this restriction of competition as a standard for merger control in view of the facts as if we were talking with dominance. Okay, so I need to have a fairly high bar to get there.

Salop:

So it’s the gap that you don’t actually have a prohibition; something like substantial lessening of competition is not within your statutes.

Jenny:

Well, it is within the statute, but it’s ambiguously within the statute. And the statute, the court said is still dominated by this idea of dominance. So you have to have a high bar if you want to target uncoordinated effects.

Salop:

Right. Okay. So, I take it that the Commission introduced upward pricing pressure indices, which would be pretty big in a standard four to three with a decent margin.

Jenny:

The court has some criticism of this. I mean, does recognize- I mean, it basically says that UPP would not be sufficiently precise to determine that there would be a major effect. But recognizes that the Commission use could be as well, so it doesn’t criticize the Commission on that ground. It does say that UPPs are a qualitative indication, it doesn’t tell us a whole lot. One has to go further. In general, the second point of this judgment is the fact that the court clearly has some qualms with economic models and basically says where the models have gotten more and more sophisticated, but as they are more and more sophisticated, they are less and less certain or predictable. And therefore, I need to have as evidence, a higher level of satisfaction or security than I used to, because of the sophistication of those models. I need very convincing evidence that it’s not only possible, but it is likely that there is going to be this restriction of competition that you play.

Salop:

Okay. So, should I interpret that to mean that if the Commission offers models, offers quantification of the effects, there’s a high bar to use the quantified evidence? Was therefore the court looking for non-quantifiable evidence? Non-quantified evidence?

Jenny:

The court doesn’t say that it. It has a general position that you should take into consideration everything which is relevant, but that doesn’t tell us much, really, but it does somewhere between the lines when you read the judgment, suggest that incredibly sophisticated models cannot be relied on entirely. And if you are going to, in this case, by the way, the Commission had used three different theories of harm to try to nail the case. But the court found them a bit speculative and couldn’t find evidence that would make those theories of harm believable, or a high degree of probability.

Salop:

Okay. So they didn’t have supporting evidence outside the model to create external validity for the model itself. Well, that certainly is a way to create more Type 2 errors, and fewer Type 1 errors. I mean, in the US you know, what I’ve been saying in my writing is that the increased use of quantification in the US has led to more Type 2 errors. That we are overly occupied with avoiding Type 1 errors than we are with realizing that that’s going to lead to more Type 2 errors. So here Europe is really going very strongly agreeing with that, but saying Type 2 errors are not so bad.

Jenny:

Well, yes and no, because there’s an institutional reaction has been going on in Europe for a while. Which is that the competition authorities try to bypass this by asking for powers that do not require it to have a reasoned decision. And for example, that’s why the market investigations of the UK time are so popular because you can, and there’s been, for example, in France there’s been a strong push that the competition authority could, on the basis of a market investigation, order structural remedies. Well, that’s a way to avoid having to justify to the courts. Its position was at the same time being able to act on the structure. So, I mean, I don’t personally like this evolution, but I recognize what you say. And I think that competition authorities are reacting by asking for other tools.

Salop:

Yeah. Well, I mean, I think typical unilateral effects case in the US, we have the benefit of the structural presumption. So sort of four to three merger that would trigger the structural presumption that the merger was anti- competitive, and then the burden would shift to the merging parties to justify the merger. So the, the econometrics, the simulation model could be put in by the firms, not by the government, and therefore the fact that the models aren’t reliable would make it tougher on the merging parties. In Europe it seems to, you’re saying, you know, there’s no structural presumption, four to three’s not enough, so I think it’s going to be really hard for the Europeans to win a unilateral effects case, unless they’re definitely the closest competitors. Even there, if the argument isn’t four to three, why would the other two continue to compete as hard since they’ve eliminated one of their competitors? And why would the merging firms compete as hard given that there’s one less competitor? So if you have to prove that, I’m not sure exactly what you would do beyond showing the logic and the fact that they were competing before, you know, with a unilateral effects model emerging firm was not a Maverick.

Jenny:

No, it seems that with the court, I mean, this is a personal interpretation, but it seems to me that what the court had in mind was the fact that the target has to be a particularly active competitor. Four a good justification to block the merger. So it is a very high bar, but I think that I’m a bit uneasy sometimes with the structural presumption, because I think that particularly in the high tech sector or in the digital sector, it is not entirely obvious that we understand where the business models, the underlying business models are. And that, when I see some proposals in the US to say, well, it’s all very complicated. We’ve got those super powerful firms who are buying up all kinds of small firms, competition authorities, and don’t have the means to intervene, so what we should have is a structural presumption, and then to reverse the proof. I think that the real issue is that we don’t understand so well how competition works in industries where you have a lot of scope economies; you have technologies that you can adapt to different markets simultaneously, and you can benefit from both economies of scale as well as economies through ecosystems. And I think that it is useful. As a matter of fact, we’re going to try to do that at OECD to look at the competitive implications of ecosystems in those industries to try to have better tools to analyze those mergers. And I think that these may be a more promising way than just creating structural presumptions, which may or may not be justified, but that’s, again, that’s because I was in the court and I don’t like to reverse the burden of proof. I do think that economics is a bit different in those sectors and that therefore we have to sharpen our tools.

Salop:

Yeah, well, I mean, if we take Sabre/Farelogix as an example. That’s Sabre as a digital platform, for sure, so it’s a two-sided market, and Farelogix was entering on one side of the market. They were providing software and services that could be used by the airlines to disintermediate Sabre, so that the airlines could reach the travel agents themselves without having to go through Sabre. And in that case, you know, the government presented evidence that Sabre was worried about Farelogix that they wanted to take it out; they were concerned about the competition. So that would be the type of evidence that I think that you’re looking for in order to stop the deal. Whereas by contrast, well, let me ask you specifically, I mean, if you’re willing to opine on, if you don’t know facts, in the US the poster child for the structural presumption, I think for potential competition mergers, would be Facebook’s acquisition of Instagram and WhatsApp. I do know the facts of the Instagram case, or they were very small company at that point. They had like 40 million subscribers, fewer than 20 employees, and according to Facebook, they made Instagram into a major feature, into a major competitor and Instagram couldn’t have done it on its own. Whereas if the government wants to go against that case, they’re going to have to say something like, well, why couldn’t Facebook do it on its own? And gee, why did they pay a billion dollars for this company if it wasn’t because it was a potential competitor? What other evidence would you like to bring to bear on the case?

Jenny:

I don’t know. I was at the OFT at a time when there was the Instagram merger and the OFT looked at it and the OFT asked everybody around, everybody in the profession, is this a competitor or a potential competitor? Everybody said, absolutely not, this is ridiculous, this is a photo app, it has nothing to do. Within a year and a half, I mean, reality had changed completely, but it seems to me that I don’t see why. Or how on the basis of that kind of evidence, which was shared by the professionals themselves, a competition authority would have been legitimate in blocking this.

Salop:

Well, that’s why they need our presumption, so that they didn’t have rely on observers who obviously were wrong.

Jenny:

Yes. But when the observers are the players themselves, you know, you can make mistakes with presumptions as well. So I’m not sure it solved the problem it did. It makes for a more rapid determination of the case, but I think that unless the presumption is based on very well established, scientific facts, which is exactly what we’re lacking, because we don’t fully understand the way competition there, which is the reason why we’re looking for something in the first place. I think it’s very dangerous to base oneself on presumptions.

Salop:

Well, I think we’re back to where we started. It all has to do with Type 1 and Type 2 errors. If you think the cost of the Type 1 error is smaller than the cost of the Type 2 error, that’s what justifies a presumption as a general matter. It’s a matter of decision theory. So that’s the scientific basis. And so, you know, I think that in this case, as I’m hearing you the fear of the Type 1 error would the cost of the Type 1 error would be the fact that Facebook was able to make Instagram into a much better service than it would have been otherwise, and so there were consumer benefits from having Instagram grow in a way that Facebook was able to implement. Whereas the Type 2 error is the fact that Facebook is a durable monopoly, protected by network barriers to entry and economies of scale, and keeping the market open so that a potential competitor could come in and get over the network barriers to entry, that’s worth a lot, because moving from one to two is where you get big benefits from moving from one to two. And the question would be, how can we try to quantify, or at least informally, we economists can formally quantify those in order to form the presumption, or more technically quantify in order to analyze the case.

Jenny:

Yes. But in the way you put it, you solve the issue by saying it’s a question of one to two, which assumes that there was some substitutability there, which was not obvious. I mean, again, I fully recognize that there is no perfect no magical solution, but I’m a bit wary of competition authorities establishing presumptions, which are not things on which everybody agrees. And I think that there has been quite a bit of this in Europe, in other areas of competition law enforcement, and this has led the Europeans to have questionable decisions. I would say, if you think, for example, about conditional discounts and things like this there were presumptions there, which were completely unjustified economically.

Salop:

Which way do you see the presumptions in terms of conditional discounts?

Jenny:

Well, the presumption was that conditional discounts were anti-competitive.

Salop:

Not in the US.

Jenny:

Until the court reversed the Intel decision. But up to that time, it was a given that conditional discount was necessarily an anti-competitive scheme, which I don’t think it was economically adjusted. It can be anti-competitive scheme, but it’s not necessarily one. So I think that when you look at many of the past decisions, the missionary decisions, I mean, they were just misguided. Why? Because there was this presumption.

Salop:

Yeah, well, I testified for Intel, so I should not opine on this. Though, I would say, I think, you know, my article on traditional discount says that they are a significant threat. So I wasn’t say that there should be a presumption, but only that one should worry about them and not rely on what a price cost test, which is what the issue was Europe. We’re almost out of time, so, let me see if I can try to press you with it since John’s book is about US enforcement. How do you come out on US enforcement? Do you think we’re doing a good job? Do you think we’re erring on the side of false negatives, or do you think we’re erring on the side of false positives? And what about our analysis? Do you think it’s well grounded?

Jenny:

Well, I think the difference between the US and the EU is that you have a much more active intellectual debate in the US and we have much more action in Europe. And in a sense, it’s odd because given the level of activity that exists in Europe, we should probably have less action, but I think that we’ve already said why this is the case. So, in terms of sophistication, the debate among economists is very strong in the US. The debate between competition authorities on the one hand and politicians in Europe is much stronger, with the politicians taking a dim view competition, the way competition law enforcement is conducted, asking to be able to review the decisions of and possibly to overturn the decisions of competition authorities, because they are misguided, because they have too narrow a view of what relevant markets are, they don’t realize or they don’t factor in the unfairness of international competition and so on and so forth. But this debate is not really a debate – an intellectual debate – it’s a debate on policy led by politicians, whereas in the US you have much higher level of sophistication of arguments, but at the same time, for the reasons that we’ve mentioned and possibly because the courts are resistant, you have sure, I mean, you said at the beginning that you don’t have a lower level of enforcement, but whether it is really because the Justice Department and the FTC don’t have the resources, I’m not sure that’s why you don’t have more enforcement. I’m not able to judge here. But we certainly have in Europe more daring solutions, if you think about, I don’t know, the Facebook case in Germany, for example, there you have a decisions which are not grounded in theory, they are grounded in an attempt by competition authorities to remain relevant and therefore to act and in this case, of course the court said no, and I can’t understand why. So you probably wouldn’t have this in the US, but this is fairly common in Europe.

Salop:

Well, we have to end, I think if I can summarize where I think the US and Europe are at this point, it’s really a role reversal that Europe is “shoot first ask questions later,” and the US is “big talk, no action.” I don’t think most people think about US versus Europe that way, so it’s a very interesting conclusion for our discussion. Thank you very much.

Jenny:

Thank you very much. Bye bye.