Tom Lenard:
Hello, and welcome back to TPI’s podcast, Two Think Minimum. It’s Friday, November 19th, 2021, and I’m Tom Lenard, President Emeritus and Senior Fellow at the Technology Policy Institute. And I’m joined by Scott Wallsten, TPI’s President and Senior Fellow, and Senior Fellow Sarah Oh.
Today, we’re delighted to have our guest, Larry White. Larry is the Robert Kavesh Professor of Economics at the NYU Stern School. He’s also General Editor of the Review of Industrial Organization and the author of numerous articles and books on industrial organization, antitrust, general regulation, and financial & bank regulation. He has also held a number of senior government positions: Senior Staff Member of the Council of Economic Advisors, Member of the Federal Home Loan Bank Board, and Chief Economist at the Antitrust Division at the Department of Justice.
Welcome to Two Think Minimum, Larry. Delighted to have you here!
Larry White:
Thank you, Tom. I’m very pleased to be here.
Tom Lenard:
I want to discuss your recent paper on market delineation of monopolization cases and particularly in the context of the Google and the Facebook antitrust cases. But before getting to that, I’d like to get a few of your observations generally on trends in antitrust.
You became Chief Economist at the Antitrust Division in the early eighties at a time when the enforcement agencies, the FTC and the Department of Justice Antitrust Division, were adopting the economic approach of the Chicago School, for want of a better term.
It’s kind of an inexact term, but at any rate, it’s basically a microeconomics approach to antitrust, and now 40 years later, that approach is under attack, and there is somewhat of a counterrevolution. What will become of it remains to be seen. But is there something to be learned looking back at this 40-year arc?
Larry White:
Oh, good question. First, you know, I was fortunate to be there at the early stages. William F. Baxter, Professor of Law at Stanford University, was the Assistant Attorney General for antitrust at the time. I was fortunate to be able to be his Chief Economist, and I was willing to do it because Bill had, I thought, a very sensible way of thinking about the intersection between economics and antitrust.
And though he was not an economist, he had, I thought, the best appreciation for microeconomics that I’ve ever seen in an attorney. We were, again, I was fortunate to be there when the Horizontal Merger Guidelines of 1982 were promulgated in June of 1982, and that’s turned out to be an extremely successful intellectual step forward 39 years later. Some version of those guidelines is in place in virtually every antitrust agency that is serious about merger control. It’s been expanded, it’s been enriched, but the basic idea that we’d developed at the time has passed the market test. It’s still being used.
I think that is a sensible way, as a more general matter, to be thinking about how microeconomics can inform antitrust policy. And I am distressed, as many of my industrial organization colleagues are distressed, by the recent trend of moving away from the microeconomics approach and bringing in just a grab bag of other things. Of course, the other things are legitimate issues of government policy that we ought to be debating and talking about, but I think the right way for those issues to be brought in is as part of larger public policy decisions. I think it’s best to focus antitrust where antitrust is best applied, and using the tools that best allow antirust to do what it’s supposed to do.
Scott Wallsten:
So, it’s as if somebody who read your January 2021 paper on what not to do and decided that’s exactly what we should do. We use antitrust for everything.
Larry White:
Well, alas and yep. That’s right, Scott. And more is the pity that that’s where we are at as of November 2021.
Tom Lenard:
So, your mentioning of the merger guidelines is a good segue into the major topic that we want to talk about today. So, as you indicated, you had a major role in developing those guidelines in 1982 when you were at the DOJ, and those guidelines specified a methodology for defining markets in the context of a merger or a proposed horizontal merger, the hypothetical monopolist paradigm, as you indicated, it’s now routinely applied and well accepted.
But as you explained in this new paper, which has not yet been published, but it’s also something you’ve written about in several other papers over the years. I’ll give the title so people could read it when it comes out. It’s called “The Dead Hand of Cellophane and the Federal Google and Facebook Antitrust Cases.” Market delineation will be crucial but elusive, but as you explained in this paper and elsewhere, and I don’t think it’s as well understood as perhaps it should be or widely appreciated, the paradigm from the merger guidelines, the hypothetical monopolist paradigm, is not useful for defining markets in the case of the monopolist, and the problem of defining markets in those cases is as you explain, and hopefully you’ll expand on it here and explain it here, is not really resolved.
Larry White:
That’s right. I just want to add one thing to our discussion of current trends, which do distress me. Reasonable men and women can differ about whether enforcement of the antitrust laws has been vigorous enough. Unfortunately, the vigor sometimes gets measured in how many cases are being brought, and that’s exactly the wrong metric.
But one could argue as my good friend, co-author, and co-editor, John Kwoka has argued, I think convincingly, we’ve allowed in the past decade or two, too many mergers that turned out to be anti-competitive to get passed through. To me, the proper remedy is, “Okay, let’s go back. Let’s think about tightening the criteria, being a little more skeptical about claims that entry is going to be easy, being more careful about whether spun off assets are really going to be viable.” I’m all for that. I believe the reaction, oh, we’ve been too lax. We need to do a bunch of other things, bring in other criteria, is the wrong direction.
Scott Wallsten:
What is the right metric?
Larry White:
The right metric in merger cases is, I think, a relatively easy one. It’s the kind of retrospective analysis that Kwoka and others have done, which is you look at the close cases that were allowed to go through, and you ask, “Were there price increases that followed those close cases?” If there were, then we’ve been too lax. If there were not, then either we got the line right, or maybe even we could ease up a little bit. But it’s looking at the close cases. You can’t just count cases. If we believe in deterrence, then counting cases…
You know, let me go to the extreme. Suppose we had a police officer on every corner. There probably wouldn’t be a lot of street crime. There wouldn’t be a lot of arrests. There wouldn’t be a lot of convictions. Does that mean we’re not enforcing the law? No, we’ve got a lot of enforcers out there. There isn’t a lot of crime happening. You can’t count cases as an indicator of enforcement. It’s just the wrong metric. So, I needed to say…
Let me now get to the question you asked. We’ve got a successful market definition or delineation paradigm for mergers. It asks: If this merger is allowed to go through, will things get worse? Will prices go up, will quality deteriorate, will innovation suffer, will something get worse? And the way to address that question is to ask, “Well, is it likely that after the merger market power could be exercised?” Well, we got to figure out what the market is.
And the innovation in 1982, which turns out to have been previously enunciated by Professor Morris Adelman at MIT back in 1969, it was either 1959 or 1969. Anyway, Adelman had come up with the same idea, which we didn’t know about in 1982, but it is the hypothetical monopolist. Look at the smallest number of firms, which would include the two merging firms, and ask if they acted as a monopolist going forward, if there was this coordination among a group of firms, could they successfully sustain a small but significant non transitory increase in price? This is what’s come to be called the SSNIP test.
If they could, then that constitutes a market. And now let’s look at these two merging firms in the context of that market and ask, “Do we think post-merger either, you know, unilaterally, just the two firms alone or in, you know, coordination, not explicit collusion, just small numbers coordination, firms would succeed in a significant increase in price?” If the answer is yes, then this merger is a potential problem. And then we have to think about what to do about it.
But the whole idea, the idea of market shares being an important determinant in this process means you’ve got to have a market, and it is the SSNIP test, the hypothetical monopolist, that gives you the boundaries of that market. That works wonderfully for mergers because the merger question is a prospective, a looking forward question. Will this merger make things worse?
Most monopolization cases, however, rest on an assertion that the defendant, the firm that is being accused of monopolizing, already exercises market power. So it’s not: “Will this activity be a problem? It’s this activity is already a problem. That’s a very different question. Now let me immediately mention, there is one exception to what I just said about monopolization cases, and that would be in an instance where a plaintiff is trying to head off a prospective action that the plaintiff believes will cause monopolization.
Say a major firm is proposing that it will insist that going forward, all repairs to its products, all repair parts must come through the company itself, that there will no longer be independent third-party repair shops.
Tom Lenard:
Did you just pull that example out of the air?
Larry White:
Well, you know, I mean just every once in a while, I get these bursts of imagination. But suppose the company has announced, you know, starting six months from now, this is what we will do. And a third-party repair shop marches into court and says, this is going to not only disadvantage me, but it’s going to allow the repair market to be monopolized. That’s a prospective question, and merger guidelines type of market delineation questions can be asked about that. One case that one thinks about in the monopolization area that went to the Supreme Court. This is a 30 or so year old case, it’s called Aspen Skiing, was this kind of prospective issue of whether the operator of three of the four ski resorts at Aspen could refuse to have interchangeability of its tickets with the fourth operator. This was a prospective issue, and also the issue of Aspen skiing being a relevant market. It was not a hard one to get to given the particular circumstances.
But for the most part, and this is certainly true of the current Google and Facebook cases. In the accusations, the federal government — in the case of Google, it’s the US Department of Justice Antitrust Division; in the case of Facebook, it’s the US Federal Trade Commission — is claiming that these companies already have market power, have monopolized, and that there are particular actions that they have engaged in that have allowed them to buttress to maintain and enhance their market power. So, it’s you know, resting on the idea that these defendants already have market power.
Well, the merger guidelines paradigm simply will not work for that kind of issue because, and this goes back to the title of my paper, the Cellophane Fallacy, you can’t ask the question, “Well, could these firms raise their price by a small but significant non-transitory amount?” Every enterprise in the land, whether they are a true monopolist, or a true firm with market power, or any seller of a commodity in a highly competitive market, they should all, if they are profit maximizing firms, they should all answer the question, “No, I can’t raise my price any higher. Look, if I did, I would lose too many customers to somebody else who would sell them something else.” So, asking the merger guidelines type hypothetical monopolist, SSNIP test question is simply the wrong one to ask. All it can tell you is whether the firm is maximizing its profits at current levels of prices. It can’t address the question, does this firm have market power?
Unfortunately, this is the question that the Supreme Court in 1956 asked in a case that the US Department of Justice had brought against DuPont, claiming that DuPont had market power and had monopolized the market for cellophane. And the Justice Department said cellophane is a relevant market. I wasn’t in the courtroom in 1956, but I’m guessing that the attorney for the Department of Justice probably brought into the courtroom to show to the judge, a roll of cellophane. And he probably would have showed, and you know, unfolded it a little bit and said, “Look, your honor, this is cellophane. It’s special. It’s unique. It’s transparent. It’s flexible. It’s special. It’s a relevant market.”
Well, the defendants DuPont came in and said, “Look, there are other flexible wrapping materials. There’s waxed paper, there’s aluminum foil, there’s parchment paper, there’s craft brown paper.” And there were probably three or four other things. And those all do things that are similar to what cellophane does. And DuPont had their sales personnel testify that “Geez, we cannot raise our prices. Our customers tell us they would go buy other flexible wrapping materials if we tried to charge a higher price.”
First, the District Court judge sort of latched onto that argument and said, “Well, geez. DuPont can’t… the relevant market has to be flexible, wrapping materials, not cellophane. And in that flexible wrapping materials market, DuPont has only a 20% share. And so they can’t be a monopoly. They can’t have been monopolizing, case dismissed.” The case was appealed to the Supreme Court and the Supreme Court majority, in a 4-3 decision came to the same conclusion.
That is the cellophane fallacy. The idea that if you ask the firm, can you raise your price? They’re always going to say no if they are profit-maximizing firm. That doesn’t tell you anything about whether there is really market power being exercised. Instead, there needs to be some kind of backward engineering that is at least conceptualized. But that, alas, has not been carefully enunciated in any way that the antitrust bar, the IO economics community, has been able to get their hands around. And the latest manifestation of that is the current DOJ case against Google and the current FTC case against Facebook.
Tom Lenard:
The problem basically is if you can’t delineate the relevant market, you can’t demonstrate a monopoly power because you have no market…
Larry White:
What do market shares mean if you can’t delineate a market?
Tom Lenard:
Talk about how the DOJ and the FTC respectively have approached this issue in their complaints.
Larry White:
The DOJ is suing Google saying they have a dominant position in general search. That’s their online general search. That’s the market on both Google and Facebook are platforms. Of course, they’re two-sided or maybe more than two sided, but let’s just stay with two-sided markets. And the DOJ says, Google dominates general search. That’s one side of their platform, and that’s a relevant market. And the other side is advertising that’s aimed at general search users. And what does the DOJ use as its support for the claim that those two sides are relevant market? Well, general search is unique. It’s special. They wave their hands a lot, and similarly on the advertising aimed at general search users, they wave their hands. There’s just a lot of hand-waving at this stage. Maybe when this case goes to trial in two years, there will be a greater specificity, but at the moment, it’s a lot of hand-waving.
The Federal Trade Commission in its pleadings, and it turns out just the way the cases are structured, there have been a lot more legal documents in the FTC case than in the Google case. In fact, just on Wednesday, two days ago, the FTC filed another 55 page brief supporting its claims and refuting the latest Facebook claims. But the Federal Trade Commission is saying that in the platform context of Facebook, it is the social networking services that is the relevant market. What is curious about the Federal Trade Commission case is, at least I haven’t looked yet at their Wednesday filing, but prior to Wednesday, they talked about harm to advertising, but they never labeled advertising on social network platforms as a relevant market. Not sure why, but how does the Federal Trade Commission support its claim that social networking is a relevant market? Well, again, there’s just a lot of, “This is unique. This is special.” There’s a lot of hand waving.
It won’t surprise you, on the other side, Google says, “Oh no, we’re not special. There’s lots of other kinds of searches that people do. If they’re looking for a specific products, they may go directly to a place like Amazon. If they are looking for travel arrangements, they may go directly to Expedia or Travelocity or someplace like that. We’re not special and advertising, oh, that’s fiercely competitive. We have to go up against all those other offers of advertising space every day.”
What is Facebook saying? “Oh, no, there are lots of other ways that online people interact. We’re not special, and we’re fiercely competitive in advertising.” One sort of thinks, at this stage at any rate, it’s a lot of, he said, she said, types of arguments. No one has quite yet said, “Liar, liar, pants on fire!” Maybe you can’t say that in court, but geez, it’s feeling a lot like that.
There is no paradigm. There is no way. The Justice Department, the Federal Trade Commission has not been able to come in and say, “Here’s a methodology, your honor. Here’s the way to think about what the relevant market is.” We could do that in mergers starting in 1982. We still, 65 years after the cellophane case, we still can’t do that in monopolization. And you know what that also impedes is that neither the Department of Justice or the Federal Trade Commission has in any of their documents said, “Your honor, the defendants are going to tell you that they are in fiercely competitive markets and that they can’t raise their price, and so they can’t be monopolizing. Your honor, don’t listen to them. That’s the cellophane fallacy (with, of course, an explanatory footnote), and instead, your honor, we are going to explain that dah, dah, dah.”
They haven’t said that. Why haven’t they said that? Because they don’t have that last sentence in there. “Instead, your honor, we will explain…” They don’t have it. And since the burden is on the plaintiffs, the Department of Justice, the Federal Trade Commission, in this kind of case, the fact that they don’t have a paradigm, they don’t have a market delineation process, I think is going to be a real problem for the plaintiffs in this case. But, Tom, if you’d like to know how I really feel…
Tom Lenard:
So does that mean, you know, they just really can’t bring them a monopolization case because they don’t have…
Larry White:
Well, you know, they can bring a case and maybe, in the Federal Trade Commission case, the Federal District Court in the District of Columbia, the judge who’s hearing the case, issued an initial opinion back in June of this year where he refused the Federal Trade Commission’s request for an injunction. That’s what they’re asking for right now. But he gave the Federal Trade Commission the opportunity to refile an amended complaint.
In his opinion, he says, at least at this stage, “Well, maybe the federal trade commission has delineated a market.” I’m willing to sort of (I forget his exact language, but…) entertain the idea. I’m paraphrasing here. But then he turned the Commission’s complaint back to them saying, “You’ve offered almost no evidence to support the fact that Facebook is a dominant firm in your delineated market.”
In fact, it was a bit as an outsider, the Federal Trade Commission in a parenthetical phrase had said, Facebook has over a 60% share of this delineated market and offered no documentary support for that claim. You know, it was in parentheses, nothing, nothing to support it.
And of course, the judge landed on that and said, “You know, you’ve got to do better. In principle, you can do better. So, that’s why I’m going to dismiss the complaint, but give you the opportunity to refile and do better.”
Scott Wallsten:
But the general issue in monopolization cases, you’re saying is, I mean, it’s common to all monopolization cases or at least most of them, right? So, what’s the track record of when agencies bring these cases, and given this problem, why did they and bring them?
Larry White:
That’s a good question. All right. Look, sometimes in a sense, the circumstances are so overwhelming. Now, the Microsoft case of the late 1990s was a monopolization case. The basic argument is that Microsoft was making life difficult for potential challengers to its operating system. That the operating system was the center of the monopolization case, and there were some really silly things that got said in the context of whether operating systems were a relevant market or not. There were even some economists who floated, you know, SSNIP-type tests in the context of that.
My sense of what carried the day was, geez, everybody knows Microsoft is making — what’s the technical term here? — boatloads of money, off its operating system. It must be a relevant market. They do have a dominant position here. So, the monopolization cases can be won, but without a paradigm, it feels like we’re going to get erratic results. Does the judge happen to believe the story that the plaintiff is telling, or does he or she believe that what the defendants are saying? I’d be much happier if we had a paradigm
Tom Lenard:
If you were Assistant Attorney General for Antitrust or Chairman of the FTC, what would you try to do about this problem?
Larry White:
Good question. First, I hold no fantasies about ever being in either position. I would turn four or five of my smartest… You know the metaphor: Lock them up in a room for nine months, four or five of my smartest economists and see what they can come out with. Because I don’t have a paradigm. The closest I can suggest is something that Phil Nelson and I proposed, I think, close to 20 years ago, which was in a sense to reverse-engineer the process.
If you think there are a set of activities in the Google case, the Justice Department is claiming that the exclusionary, the anti-competitive actions were that Google paid large sums of money to Apple, to the manufacturers of other smartphone and computer hardware, except where the Microsoft operating system was going to be installed, but paid large sums to get the Google search program installed as the default program on those devices.
By being the default, that made it much, much harder for alternative search engines to gain a foothold in this search market. So, if you think that’s the case, then you ought to reverse engineer that and have a decent model that could say, but for those actions, what would the search market look like? And what might pricing or quality in general search look like? Maybe, right now, as we know the price is zero. It doesn’t have to be zero. Under some circumstances, I can imagine that a search engine charges its users for the search privilege, or if advertising is so lucrative for the platform, it could pay individuals to come search on the platform. Remember these are two-sided markets, and the pricing on the two sides is very idiosyncratic and will depend on where the money — you know, follow the money — where the money is. If the advertising is so lucrative, you may want to pay individuals to come and search on your platform.
Scott Wallsten:
So, a few places have tried that actually.
Larry White:
I believe Bing at times has tried that. It may not work. It may be a stigma that says, “Gee, this is not a very good search engine, and that’s why we have to pay people”.
I don’t know, but you’d need a model. Nobody has a model. We need a reverse-engineering model. This is what Phil Nelson and I proposed like 20 years ago, we failed a market test. Nobody has showed a pick-up on that idea. I don’t know what at this point. I don’t know. Other than back to your original question, get four or five really smart economists, lock them up in a room for nine months and see what they come out with.
Tom Lenard:
If four or five smart economists in a room for nine months could… if you think they could solve this problem, then it’s obviously a solvable problem. That’s not a huge cost, right?
Larry White:
That’s right. I don’t know whether they could, and if those four or five came out and they sort of scratch their heads and they say, “Sorry, Larry. We tried. We can’t do it.” Then, well, I don’t know.
Tom Lenard:
Well, let me ask one final question and then, because I think we’re running out of time. It seems to me, and I’m not as much of a sort of certainly an antitrust guru as the other people here. But it seems to me that this issue is not occupied other leading antitrust economists to a great extent. First of all, is that true? And second of all, why is that?
Larry White:
Well, I think that’s right. I’ve been distressed at how little attention leading antitrust economists, IO economists, have paid to this question. And I’m not sure why. I know that anytime, you know, every once in a while I get a call from an attorney who wants to bring some kind of exclusionary case. And I right off the bat say, “You may have a good story here, but I’m not the right guy for you because I don’t know how to define or delineate a relevant market in your kind of circumstance. If you approach me on a merger case, I can help you. On this one, I don’t know how to do this.” Other economists don’t seem to be so worried, but I don’t understand why that is so.
Tom Lenard:
Maybe because it’s a hard problem.
Larry White:
It’s a hard problem. It’s a hard problem, but not enough people seem even to be paying attention to it. And that’s what frustrating.
Tom Lenard:
Well, I found your arguments persuasive as to why it is a problem.
Larry White:
Why, thank you, Tom.
Tom Lenard:
We were delighted to have you here today and thanks very much.
Larry White:
Thank you. Thank you for the opportunity. I appreciate this, and let’s hope something can happen.
