The 2023 Merger Guidelines are ten months old, but they still need some unpacking. To help our readers think through the issues that the Guidelines raise, Antitrust magazine asked two experienced legal scholars—Professors Tim Wu and Herbert Hovenkamp—to discuss the Guidelines with Antitrust editorial board member James Keyte. This discussion took place on May 28, 2024. The transcript has been edited for length and readability.
James Keyte: I want to start with your views on the overarching framework for assessing harm to competition, and I will start with Herb. You have written about the need to focus on measurable metrics, such as price and output, more traditional aspects of the consumer welfare standard. Do I have that right, and why is that?
Herbert Hovenkamp: Actually, I have drifted a bit in the last four or five years. I am not nearly as enamored of the phrase “consumer welfare” as I used to be, mainly because it has been abused so much, a lot by people who do not really understand it. My view today is that we need to go back to the judicial decisions. Ever since the 1890s, without significant departure, the courts and the Supreme Court have focused on practices that tend to reduce market-wide output or increase price, with the exception of the Robinson-Patman Act. I think that is a metric that is hard to measure but nevertheless is capable of being measured and it is consistent with the case law.
Keyte: Tim, you have written and spoken about, if I have it right, that antitrust should focus on harm to the competitive process. Could you explain what you meant by that?
Tim Wu: I share some similarity to what Herb said in the sense that there is something attractive and alluring about trying to reduce hard decisions to a more quantitative measure—it is the scientific aspiration—but having been close to it in government and in cases, I have started to feel that, despite the ambition of the consumer welfare quantitative approach, it has had great problems dealing with the harm of exclusion and also capturing effects on innovation.
I think we should retreat to reaffirm an older standard, one that is still mentioned in many of the cases and that I think is still binding in a sense, Microsoft being an example, which is about harm to the competitive process. It is a slightly humbler approach. It looks at an ongoing process of competition, assuming something is going on, and tries to identify behavior that is sharply out of whack with what you would consider a more natural or permissible range of competitive tactics. It has a hint of the fair competition standard in the FTC Act, and it is easier for judges, lawyers, and even economists to deal with.
Keyte: Tim, recognizing that neither of you were involved in the drafting of the Merger Guidelines, do you think a focus on the competitive process had an influence on the drafting of the Guidelines?
Wu: To an extent, in that it is a proxy for the sense that the Merger Guidelines—whether 2010 or other versions—were overly optimistic about our ability to reach deterministic answers about some of the key questions involved.
But the broader impulse that drove the new Guidelines was a sense that merger policy was failing in the United States and that we were passing on too many of what turned out to be too many anticompetitive mergers or mergers that did not serve the goals of the antitrust laws of trying to increase output or any other metric. There is just dissatisfaction with a formula that, in almost all cases, lead to the same result—i.e., no merger challenge—particularly in new industries, like the tech industry, where many people think, something is not right here.
Keyte: Herb, do you think the Merger Guidelines, in their final form, are sufficiently linked to price and output?
Hovenkamp: No, but I think they are an improvement. I agree with the 2023 Guidelines’ changing the structural standard to an 1800 Herfindahl-Hirschman Index (HHI). I think the jury is still out on lowering the increase in the HHI from 200 to 100. That one has flown under the radar a little bit, but if you do the math, that is an incredibly big change. A merger of a 7 percent firm and an 8 percent firm would raise the HHI by more than 100, so that standard would really shift antitrust in a much more aggressive way. I am in favor of doing that, but I am not sure I am in favor of doing it quite that much.
Also, I have serious doubts that Congress is going to come up with the money to fund that level of enforcement. It is one thing to state very aggressive standards for going after mergers, but this recent National Bureau of Economic Research (NBER) study of about a hundred mergers—this was actually in reference to the 2010 Guidelines—ended up confirming that the Agencies should have been more aggressive, but then estimated that if the Agencies really wanted to enforce structural standards where they should be, the Agencies would need an enforcement budget roughly four times as big as their current budgets. Well, good luck with that. But I do agree with the basic proposition that enforcement against horizontal mergers needs to be more aggressive.
Keyte: For both of you, the heads of both agencies and other commentators have maintained that the Guidelines reflect binding Supreme Court authority, and I know there has been some debate of that. Tim, do you agree with that?
Wu: I think that the answer is yes. As I said, I have a lot of respect for the drafters of the 2010 Guidelines, but there was a different orientation in that period that went in a more technocratic direction and did not include a lot of citations to binding precedent. I think this 2023 version represents a conspicuous effort to return to what I would call more “normal” methods of law enforcement, which are that you cannot ignore or regard as outdated a Supreme Court precedent that is a couple of decades old or that has economics maybe you don’t happen to agree with.
Maybe, also, it is important you cannot sidestep the intent of Congress. I think if you look at the 1950 Celler-Kefauver Act, Congress had pretty strong feelings about what it wanted. The Brown Shoe decision, which Herb and I have disagreed upon at times, strongly styled itself as an interpretation of the 1950 Act. I think there is a major corrective that goes back to more normalcy in terms of statutory interpretation and regard for precedent, and it is evidenced in the new Guidelines by the citation to Supreme Court precedent.
Keyte: What do you think of that, Herb?
Hovenkamp: I think a couple of things. Number one, there are two questions about precedent: one is, did the Guidelines use Supreme Court precedent correctly; and two, what about the balance between older Supreme Court cases and more recent circuit-level cases?
On that second one, the Guidelines read as though those old Supreme Court cases, including Brown Shoe, are like those mosquitos stuck in amber in Jurassic Park; they were just kind of frozen in time for 200 million years. But the Supreme Court got out of the business of substantive review of government merger cases in 1974. That was fifty years ago. There have been a few merger cases since then, but they mainly have been private actions and more tangential things.
I think the best way to read that is that the Supreme Court was basically inviting the lower courts to go on and make precedent and not be locked in, particularly on fact-intensive matters that need to be governed by economics and expert testimony. On the question of precedent as between the Supreme Court and the lower courts, that is pretty much my view.
With respect to the use of older Supreme Court precedent, the Guidelines tend to regard Brown Shoe as kind of a “super-precedent” and everything else as somewhat less important. That is the way I read it.
Let me give you an example. One year after Brown Shoe, in Philadelphia Bank, Justice Brennan—not an anti-enforcement guy; he was one of the liberals who voted with Warren most of the time, dismissed testimony by the smaller bank competitors saying, “Well, if the merger leads to higher prices, they are going to favor the merger; so what we really want to do is get some evidence from consumers.” That is what Philadelphia Bank says. That is not repeated in the Guidelines; that is, the new Guidelines pull back from Brown Shoe’s attitude toward price.
Then again, in 1964 in the El Paso Natural Gas case, the same thing. The only issue in that case was whether a merger that eliminated an aggressive price competitor ought to be illegal.
Then, flashing up a few years, we have Monfort v. Cargill—which, incomprehensibly to me, is not cited in the Guidelines. The theory of the case was that the merger would result in lower but nonpredatory prices, and the Supreme Court—once again writing through Justice Brennan—said it would be “inimical” to the goal of the antitrust laws to permit a merger to be challenged on the grounds that it creates lower prices that harm a competitor.
So, I think it is fine to look at precedent, but you’ve got to look at all of the precedent. I regard Brown Shoe as a serious outlier in Supreme Court merger precedents. The one that tracks it most closely is Von’s Grocery, but beyond that many of the Supreme Court decisions—including General Dynamics, Philadelphia National Bank—bend over backward to put distance between themselves and Brown Shoe, and the Guidelines simply don’t reflect that.
Keyte: Tim, what is your reaction to that? And do you think there is a risk given those views of the courts being somewhat confused or inconsistent in their application of the Guidelines, especially in relation to those Supreme Court cases and circuit Section 7 law that was developed in the 1990s and the 2000s?
Wu: I have a few views on this, and I think this gets more to the heart of where Herb and I may see things differently.
I don’t believe—and I don’t think Herb is really saying this—that the courts of appeals can overrule the Supreme Court. That is overstating it a little bit, but I don’t think the Supreme Court said, “We are out of this and therefore all of the stuff we said before is fossilized and does not exist.” I think the agencies and the courts need to obey the Supreme Court.
I think in particular there is a danger, and something that was manifest in the 2010 Guidelines, that economics and notions of efficiency were taking on a Constitution-like force where it doesn’t matter what the statute said; it doesn’t matter what the Supreme Court might have said in the 1960s— if it does not make economic sense, then it is not good law. You see some of this—many people may disagree—in the Robinson-Patman Act, but it does not mean that it is no longer good law. The sense that there is almost a constitutional weight to efficiency concerns is one I think that needs to be challenged.
On Brown Shoe in particular, I think if you look at recent merger decisions, the courts are constantly citing Brown Shoe. It does have a seminal quality, and I think that is because it did mark the interpretation of the 1950 Act and so has this particular status in statutory interpretation as the Court’s first interpretation of the relevant statute. That is why, as long as courts keep following it, it makes a lot of sense for the Justice Department and the FTC to key their arguments toward it as significant precedent.
I don’t know why they didn’t choose to cite some of the other cases—they had more cases cited and I think pared back on them—but I think evidently there is a strong disagreement as to how much force Brown Shoe still holds.
Keyte: Some have viewed the 2023 Merger Guidelines as reflecting a diminished role of economic analysis, especially as it relates to effects. Do either of you take that view and, if so, is that a good thing or not a good thing?
Hovenkamp: Yes, I lean toward that view. On market definition, I think what makes me a little bit uncomfortable is the Guidelines’ willingness to deviate from the hypothetical monopolist test a little bit too rapidly. I think there already is a disjunction emerging between the 2023 Guidelines and the Agencies—because the agencies know that to win a case, they have to use the hypothetical monopolist test for market definition because it tends to define narrower markets.
I think the Guidelines would have been better if they had said: “When the data are available”—that is the most important limitation on the hypothetical monopolist test—“we will use the hypothetical monopolist test to define markets. When they are not available, then we will have to look at alternatives,” as they did in the recent Illumina/GRAIL case, because there were no sales from which you could use the hypothetical monopolist test.
The other place where I think the Guidelines depart from serious economics in a harmful way is in the modification of the significant and non-transitory increase in price (SSNIP) test, which is part of unilateral effects and also a part of market definition, by changing it to “prices and terms” (SSNIPT). I don’t think that is going to last. It represents a non-economist’s view of what is capable of being measured. Everybody knows that competition exists on both price and nonprice avenues. However, that is really not the question. The question is being able to come up with models that have predictive power, and the way you can do that is by focusing on price. So all the tools that we have—things like elasticities and upward-pricing pressure—are tools that depend on responses to price changes.
Now you’ve got this test that says, “We are going to look at things like working conditions or product quality.” I think the people who drafted that did not have a very good conception of how impossible it will be to model merger analysis using that kind of test, and I think that is going to hurt enforcement more than help it.
Unilateral effects mergers have risen dramatically in the last ten or fifteen years. Today, more than half of the merger cases assessed by the two agencies are under unilateral effects analysis. One of the reasons it has triumphed is because it is economically manageable, and I would expect to see that trend continue; but if these econometricians both for and against are having to incorporate things like better working conditions into these tests, it is going to complicate them so severely that I think that kind of merger analysis is going to be very strongly limited. Do I think that will actually happen? Probably not. What I think will happen is that economists will simply ignore it and keep right on using price, or otherwise they are going to try to figure out some way to convert these other terms to their cash value.
Keyte: Tim, let me reorient the question a little bit for you toward the Neo-Brandeisian school of thought that I think emphasizes structure more than a technocratic assessment of effects case by case. Do you think that is accurate, and do you think that is a good thing about the Guidelines? Does it reflect an emphasis more on structure than a kind of hyper-technocratic assessment of effects?
Wu: I wouldn’t say the new Guidelines are Neo-Brandeisian guidelines. As Herb said, I think there is still a very strong and central role of economic analysis in these Guidelines. But, from an agency perspective, there was something wrong here. Something was going on with the overly technocratic analysis where there was just too much being missed.
The other day, to take a random example, I was looking at maybe buying a place here in New York, and I was looking essentially at two main search engines, StreetEasy and Zillow, and then I suddenly realized—I was wondering why StreetEasy is so bad—that the FTC managed to allow Zillow to buy StreetEasy somewhere in the 2010s.
There are just so many examples, particularly in tech, which was like a national experiment for the 2010 approach, where blatantly obviously anticompetitive mergers went ahead. I can think of the acquisition of Waze by Google and the acquisition of Instagram by Facebook. Some of these strike me as so blatantly anticompetitive that there needed to be some kind of—I don’t know whether it is structuralism or just telling the agencies to change course, but there was something going wrong here.
Hovenkamp: I agree with everything Tim said. I just want to add one thing. I think there is a question when we talk about under-enforcement about whether it is a mistake in the previous Guidelines or about whether the Agencies have not been following them, because if you look at the record of merger enforcement, most of the challenged mergers are on HHI numbers that are way above the threshold. You can write the threshold wherever you want it, but if you don’t pay any attention to it in your enforcement policy, it is not going to make a whole lot of difference.
As I was saying a little bit earlier, I completely agree with these Guidelines’ reduction in the HHI standard, but they are going to have to come up with the resources or the mechanisms to enforce those standards, because that is where the real question about outcomes of merger enforcement really becomes relevant.
Keyte: Let’s turn to textualism debate. Certainly, Chair Khan and AAG Kanter have asserted that the Merger Guidelines are in part strictly following the text of Section 7, and when we look at the text of Section 7, we see it is really prohibitory in language about the effects of acquisitions that may lessen competition or tend to create a monopoly. But some have observed that the Merger Guidelines seek just to make markets optimally competitive and, in a sense, open up markets for rivals almost in a regulatory sense.
Do you think these Guidelines are really sticking to a prohibitory view of Section 7 or are they doing more than that?
Hovenkamp: I think in some areas, yes. I think there are too many areas where the Guidelines are using what you might think of as “scare tactics” rather than balanced analysis, particularly in the ancillary areas, like mergers of networks, vertical mergers, and potential competition mergers, where the only things they see are harm, and they don’t even talk about offsetting benefits the way that earlier Guidelines have.
Also, I think one of the problems with the “substantially lessen competition” standard is that it needs some kind of metric. You have different metrics out there. The Neo-Brandeisian view—I may be paraphrasing it—is that if you are reducing the number of players in the market that is lessening competition. I don’t believe that, and I don’t think too many serious people believe that, certainly not when you have ten or twelve or more firms in the market.
I prefer “lessening competition” to be assessed in terms of reduced output and higher prices, an economist’s definition of those terms. By the way, that is the way it was understood under the Sherman Act during the first half of the twentieth century. So, it is kind of a debate about whose dictionary you look at. If the phrase “lessened competition” means increased price or reduced output, then I think it is pretty clear. If it means something else, then is it going to be anybody’s guess as to what it should mean?
Keyte: Tim, your thoughts?
Wu: First of all, it is interesting going back to the text, and I think it is consistent with the broader sense that we need to return antitrust to something more like you see in other areas of law enforcement and law in general.
I think that this debate of what we have heard about “lessened competition” is an example of something that is right at the margins but has been misused. I would agree that there are situations where you may have a merger that may reduce the number of firms in an industry—to think about Herb’s example, say twelve to eleven, two weaklings come together and somehow produce a stronger company—and therefore you have an increase in concentration, but also an increase in competition.
But I think that kind of possibility and the insistence that we have to not see an increase in concentration as a decrease in competition did get abused quite extensively in the 2010s during the “antitrust winter,” when people would insist that for a four-to-three merger you need to look really carefully and understand that this is actually increasing competition. So there is such a thing as a correct idea taken way too far.
Going back to the track record, I think it proved—I don’t want to say “too easy” because it was expensive—too possible for companies and defendants to seize on that point, which is correct, that you could have reduction in competitors and more competition and use it to justify mergers that I think in retrospect were clearly anticompetitive.
Keyte: Let me focus on the “create a monopoly” textual language. Some read the Guidelines to say that if you are viewed as a monopolist, there is not much you can do in the merger area at all, for example as it relates to potentially raising rivals’ costs, changing other firms’ bargaining positions, and so on. It seems fairly broad relative to the text of Section 7 that focuses on “creating” a monopoly. What is your reaction to that, Herb?
Hovenkamp: A couple of things. Number one, there is an interesting paper—I think it is coming out in the Texas Law Review—written by one of my former students plus a coauthor, John Newman, who was working at the FTC. The paper argues that we have kind of locked the “tend to create a monopoly” language in the closet and do not pay any attention to it. It is an interesting paper. I think it is pretty much wrong.
The fact is that, under the structural standards that the 2023 Guidelines impose, a merger that creates a 30% firm will be prohibited in virtually any case except the acquisition of a rival that has a market share of, say, 1 percent or less. For example, the merger of a 28 percent firm and a 2 percent firm would trip the trigger under the Philadelphia Bank element of the Merger Guidelines.
Under the structural standard—1800 HHI plus 100 increase—a merger between a firm with a 7% market share and an 8 percent market share would increase the HHI by 112—if that merger were in a concentrated market, that would trip the trigger.
We are currently condemning mergers where the post-merger firms are in the 15%-to-30% range if we are following the Guidelines. I am not talking about what we are actually doing, but what the Guidelines say. In that case, I don’t think the phrase “tend to create a monopoly” adds very much. What merger is out there that tends to create a monopoly but does not trip one of those structural triggers? That is my first point.
My second point comes back to what Tim said a minute ago, because I agree with him, and that is the question of what about the firm that is already a monopolist, or let’s say one of the Big Tech companies that gobbles up small rivals; should those be condemned?
My answer is yes, but we have to distinguish between acquisitions of competitors, people, or firms that pose a genuine threat of entry into competition from acquisitions of complements. If you start looking at the lists of firms that have been acquired by the Big Tech platforms—Wikipedia, by the way, maintains ongoing and I think up-to-date lists of Big Tech acquisitions of small firms, and they number in the hundreds—the great majority of them involve acquisitions of complements and are being used by the dominant firms to improve their own technology.
For example, Google acquired this tiny, tiny firm called Viewdle. One of the things Viewdle does is it enables searches of digital photographs, so it greatly enhanced Google Search’s power to do searches of images as opposed to searches of words. That is not a merger that is eliminating competition; that is a merger that is improving the quality of Google’s search engine.
So one of the things we have to do carefully is have strong rules against acquisitions by dominant firms, and I agree with Tim that Facebook/Instagram falls into that classification. In fact, the FTC’s complaint is crystal-clear on that topic, that Zuckerberg was afraid that Instagram would mature into a substantial Facebook rival. That is a very good reason for challenging that merger.
But when the merger is one of a complementary product that does not pose a significant threat to lessening of competition, then I think we are using merger law the wrong way.
Keyte: Let me expand on that a bit. The Guidelines seem to suggest that even if a dominant firm, Big Tech or otherwise, acquires a complement, the Agencies will consider whether it would be better for competition if the target were in the hands of a different rival or was left to grow organically. Tim, do you agree that this is reflected in the Guidelines, and what is your view in general of that complementarity acquisition by dominant firms?
Wu: I think the 2023 Guidelines reflect the agencies looking back at the period of about 2008 to about 2020 during which—I did a count myself once—the major platforms acquired over 1000 smaller companies. That includes Microsoft and some of the other ones. There was a grand total of one challenge that led to a consent decree, and that was over the Sabre/Farelogix merger. So, if you have an algorithm that does not fire on 1000 acquisitions during a time in which companies seem to be building their market power and seem to be becoming further entrenched, maybe you need a different tool.
The concept of nascent competition is very important in this area. The Microsoft court said it well. I am going to paraphrase, but they said roughly: “You can’t just allow a dominant monopolist to stamp out its potential or nascent competitors at will. The fact that they have not developed into actual competitors cannot be an excuse for just allowing a monopolist to vacuum up any potential threats in their nascent stages.”
I think the Guidelines are trying to deal with that better, and so is the FTC lawsuit against Facebook and Instagram—which is a little bit like a mea culpa, I guess, for passing on those.
This is slightly anecdotal, but one of the reasons why in the Waze/Google Maps merger the FTC decided to do a second request was the theory that Google Maps was for finding out where you are and Waze is for finding out where you want to go. The ability to find almost anything a complement—okay, maybe they were complements; many people have both on their phone—but I think they also kind of compete. The amazing ability of talented defense lawyers, even people inside the agencies, to think of anything as a complement did not have a good effect.
I think the Guidelines are trying hard to make sure that we do not let that playbook be repeated because I think it was a playbook. I think everyone in the tech world knew it. The business press just said: “Oh, yeah, that was brilliant. Facebook got rid of all of these dangerous competitors when they bought up WhatsApp and Instagram. That is great strategy.” It is a great strategy if you don’t care about competition and you only care about monopoly maintenance.
Keyte: Herb, any concern with that approach?
Hovenkamp: First of all, the standard articulated in Section 7 has been interpreted even by Brown Shoe to say “probabilities, not possibilities.” So I want to know how likely it is that this acquired complement would have developed into a competitor.
I would prefer to take a more aggressive standard when the acquiring firm is a dominant firm, and here I think there is one thing that needs to be done with respect to merger remedies, which the 2023 Guidelines don’t discuss, which is this is a good place to use nonexclusive licenses as a merger remedy. That is, going back to my Google/Viewdle case, if we are really concerned that the acquisition of Viewdle and this capability of searching images is potentially anticompetitive, then the fix for that is to say to Google, “Okay, you can acquire a nonexclusive license to Viewdle that will give you all of the technology you need to improve your own product, but it will not give you the right to exclude others from adopting the same technology for themselves.”
I think that is a perfectly good remedy that has been undervalued in merger cases, particularly when you consider the fact that if you look at these dominant platform acquisitions—setting aside a couple, such as Whole Foods—these dominant platforms are not after the buildings, machinery, or physical equipment. The main acquired assets are intellectual property and sometimes human capital, but mostly it is intellectual property, and the nice thing about intellectual property is that nonexclusive licenses will frequently work as a remedy, and it should be very high on the list of remedies when the government looks at these types of acquisitions.
Keyte: Tim, I have one more follow-up, as this is a very important area. Herb’s hypothetical was a Big Tech firm or very large firm buying a true complement to better their product. Would you agree under those assumptions that Big Tech or other large firms in a marketplace should not be prohibited from seeking to acquire or merge with true complements, or do you think all of them in a sense have this future potential to be competitors along the lines of Microsoft in the Section 2 sense?
Wu: I think the question may overestimate the degree to which you know whether something is truly a complement or not, so I would not be satisfied with a green light for like, “Well, that’s a complement,” which is an economic concept—and as the Google Maps/Waze example shows, it is susceptible to being stretched pretty far. A clever person can describe almost anything as a complement because almost everything is in some degree complementary.
That is particularly true when you are talking about firms that are in an early stage of their development. WhatsApp was a complement to Facebook; but when you looked at the documents, Facebook was worried that WhatsApp would develop a social network on top of its messaging service. So when you just ask, “Right now”—frozen in amber—“is this firm a complement?,” that is forgetting the question of what a firm might become, so I think it is too static a way to answer the question.
This is part of the move to lawyerly methods. I think the mistake the FTC made in the Facebook case was ignoring the email chain and the evidence. Sometimes companies straight-out say: “Well, this is why we’re buying them: we think they could become something dangerous to us.” In a situation like that, I don’t care if you can pay someone to say they are a complement; you should listen to what the heads of the firms say they are buying the company for. It is another way of saying that we should put some weight on intent evidence.
Keyte: Now, let’s start going through the Guidelines themselves. Herb has already talked about Guidelines 1 and 3, about lowering HHIs, potential coordination effects, and things like that. Tim, do you agree that that delta maybe makes the net too wide in terms of picking up mergers as well as the 30 percent presumption, even with a very small HHI change?
Wu: No, I don’t. I think that the structure of the Guidelines is to set out principles for when the Agencies should consider taking action. I think the problem with the old Guidelines is that there were too many safe harbors to hide in.
But this is not the end of it. The structure of the 2023 Guidelines, unlike the past, is not going to be an algorithm that gives you either a green light or a red light and has a determinative answer. Instead, they tell industries that here are a series of reasons why the Agencies might take action, and if they see either a delta or a move to 30, the Agencies maintain it could satisfy the statutory standard of “substantially lessening competition.” I don’t think that means that the Agencies are obliged to take action in all of these sorts of cases, but it does meet the threshold for the presumption.
Keyte: Let’s move on to Guideline 2, which is essentially unilateral effects and upward price pressure (UPP), apparently without any reference to market definition. Herb, is that how you read that, and is that problematic in the context of a statute that references a “line of commerce,” and presumably does not condemn all mergers between substantial competitors?
Hovenkamp: There is a lot to unpack in that question, even though it did not have a lot of words.
Number one, I don’t think you need to have a market definition in a unilateral effects case. I think you can do it econometrically. You need numbers having to do with the elasticities of substitution between the merging firms and also between the merged firm and nonmerging firms, and you need to know something about margins. You can do unilateral effects without defining a market, and I would favor that.
The real hindrance to that has been the Brown Shoe statement that the words “line of commerce” and “section of the country” in Section 7 of the Clayton Act refer to a relevant product market and a relevant geographic market.
The H&R Block case was very clear about this. The judge condemned that merger but said that as a matter of economics we don’t need to define a market here; but as a matter of law—this is one of those things preserved in amber really—we need to define a market in order to do unilateral effects.
I would prefer to see that idea jettisoned, given the number of times the Supreme Court has said that the measurement of market power and market power effects is a factual question. I think it is time to treat those issues more as questions of fact subject to Daubert and evaluation of expert testimony, rather than as subject to something that the Supreme Court said. Beyond that, I think unilateral effects has a lot of capability and I think it will be used more.
But as I said a while ago, I think by trying to expand the evaluation of unilateral effects beyond price, the 2023 Guidelines throw in a real monkey wrench—all of these numbers that unilateral effects theory uses about elasticities and upward-pricing pressure are focused on price changes in response to changes in structure.
Keyte: Let me push you a little on the UPP, having written a little bit about this. Under UPP, there will always be a positive effect on price between a merger of competitors, and you are going to have, potentially, even significant UPP effects with mergers that could be six-to-five or seven-to-six. It is argued that without defining a market, assessing other rivals that could constrain the merging parties, and evaluating potential supply responses, it becomes a simple math exercise if you are not capturing at all how markets actually function.
Given what you just said, Herb, in a sense endorsing pure mathematical UPP on price effects, how do you respond to that?
Hovenkamp: More merger retrospectives, more studies to see if we have hit the target. Recent merger retrospectives have been much better at that. We have good merger retrospectives on concentration index performance. We do not have very good merger retrospectives on unilateral effects performances, although we have some.
One real gap right now is that if the courts are going to start adopting this 30% Philadelphia Bank presumption, we really do not have any merger retrospectives that address that particular presumption, so I don’t know how empirically robust that is.
But the retrospective testing on unilateral effects suggests that, if our target is to hit mergers that are going to produce minor price increases, we are probably doing it about right. The current Guidelines do not change that, except for this complicated introduction of nonprice issues. I think that is a very strong anti-enforcement development in the 2023 Merger Guidelines if it is taken seriously as a result.
Keyte: Let’s look at potential competition. Guideline 4 seems pretty straightforward in terms of following Marine Bancorp and some of the other cases, but perhaps not. How do you both see the potential competition doctrine where, in today’s world, it is closely connected with nascent acquisitions. One is arguably outside the market; one is maybe nascent in the sense that it may be a complement that could turn into a more direct competitor.
One, do you agree with the potential competition recitation—I think you have written on this, Herb, a little bit—in the Guidelines, and two, does it become a little bit superfluous given the emphasis on nascent competitors?
Wu: I feel that to stay true to the Guidelines’ emphasis on reflecting the case law and suggesting a return to one of normal law enforcement practice—law enforcement has a series of options or styles of cases to bring—it had to be included. But as an academic (and also when I was in enforcement), I felt that these doctrines are “creaky and moth-eaten,” to quote my old boss. They are kind of weird.
First off, the names are really bad and, if anything, I wish they could have renamed them. People in the world of antitrust all the time are used to the idea of “actual” potential competition and “perceived” potential competition—and to the fact that they are opposite of the way you think they are going to be. They are confusing labels. They are kind of a very stylized theory of how a market will work. Actually, it sounds more like a defense argument that the ongoing possibility of entry is having this pressure and we prefer them to develop their own entity inside.
I agree in some cases, but the bottom line is they don’t strike me as doctrines. Whatever value or accuracy they had in the 1960s, they don’t feel like the kind of thinking that is going on in this century.
The nascent competition doctrine enforcement, again, would be my choice. I did some scholarship in this area as well, and I did some of the legal work. At first the idea was we will rehabilitate potential competition doctrine, but I think, academically at least, I have had an interest in trying to plot out a doctrine that is a little more sensitive to the ways businesses are thinking in this area.
Hovenkamp: I was not at all happy with the Guidelines’ treatment of potential competition doctrine. I thought it was practically a photograph of where we left the law in the mid-1970s when the Supreme Court abandoned the field with Marine Bancorp. The Guidelines offered nothing new. The Supreme Court abandoned it because they thought it was too unmanageable—too many predictions, too much speculation.
I think the most fundamental problem, however, is the link between potential competition and market definition, because the one major development in the last fifty years has been the rise of the hypothetical monopolist test, which looks not at who is currently in the market but who would be in the market in response to a small but significant non-transitory increase in price.
So the question I have is: is there a group of firms out there that would not be considered in the market under the hypothetical monopolist test, but that we would nevertheless want to worry about as potential competitors, and can we do that without excessive speculation?
I think there is a pro-enforcement outcome to that story, which is that many, many of those potential competition cases from the 1960s and 1970s might actually be horizontal merger cases today if we define the market under the hypothetical monopolist test rather than relying on a theory of potential competition. I am not saying it would pick up all of them, but it would pick up the beer case, where in response to a small price increase Falstaff would go into the market. Well, today we would say that is a horizontal merger because under the hypothetical monopolist test, Falstaff would be defined as already in the market, even though it is not making sales there at this minute.
I will conclude with one thing. Justice Douglas did that in El Paso Natural Gas. El Paso acquired Pacific Northwest. Pacific Northwest had been wanting to get into the California natural gas market; it had been attempting to do so but had never done so. El Paso continuously cut back its price to keep Pacific Northwest from entering and then finally acquired it. Douglas looked at all of that evidence and said: “This is a horizontal merger. Even though Pacific Northwest is not making any sales in California, this is a horizontal merger because it had shown that it wants to get into this market.”
Here is the rub. Once you can flip these things from potential competition mergers to horizontal mergers, your enforcement stance becomes much, much more aggressive. You start winning cases if you treat them as horizontal mergers; you lose them when you treat them as potential competition mergers.
Keyte: Very interesting; Tim?
Wu: I can add a little more, and where this ends up who knows? Like I said, I imagine it like an old musket on the wall; it needs to be there because it is still Supreme Court doctrine, but you wouldn’t want to take it out too often because when you fire it, it could explode on you. I feel that way. I also have a good law student who, like Herb, makes the point: “Why don’t you just prosecute this as a horizontal merger if it fits the market definition?”
I guess the final critique I have of the doctrine is that it has this kind of stylized idea that eliminates the fact that the acquirer could have gone into that market itself. I think it is not in some cases a very realistic undertaking. It ignores the whole buy-or-build business strategy theory, where sometimes companies just lack the expertise or specialty to be in an adjacent market—if it is truly an adjacent market, if we are not talking about the same market—and might want to be in that entirely separate market.
There is another set of questions that I think are better reflected by a nascent competition inquiry, which is: Are they buying this nearby company because they are afraid they are going to attack their real monopoly or be a threat to them? I think that is the question you should be asking, as opposed to imagining that in fact the monopolist was going to build its own competitor in that market. I just find it too stylized, and that is not how businesses think.
Keyte: Let me move to Guideline 5, which I find to be the most provocative Guideline, as it deals with foreclosing rivals and raising rivals’ cost, but without reference to whether the relevant transaction is vertical, conglomerate, or otherwise. It strikes me as rather open-ended. And it is not necessarily dealing with dominant firms or so-called “entrenched” firms. It appears more to be that if we think that an acquisition of any kind, of any asset, tilts the playing field somewhat in terms of what one may call the competitive process or even just bargaining-power, that is going to be good enough to invoke Section 7. Tim, do I have that right, and if so, do you think that is a good idea?
Wu: I agree with you that it is one of the most provocative and innovative parts of these new Guidelines. I think they are trying to overcome what I might call the almost segregation of vertical mergers into an entirely different category of merger that never gets successfully challenged. There was sort of a stigma, particularly in the Agencies, that, “We just don’t do that kind of thing.”
People talk about the outside effects of these Guidelines, but the effects inside of the Agencies also are important because in a lot of areas you need to buck up the staff attorneys to the effect, “Listen, you have the authority here.”
Note also that the Clayton Act does not actually use the word “vertical” or anything like that. But most have grown up in a culture where we would look at a transaction and say: “Well, it’s vertical. Okay, what’s next?” That had become the level of analysis for many vertical mergers. I think Guideline 5 is designed to be corrective in that respect.
I think the Fifth Circuit’s Illumina/GRAIL challenge is to some degree a vindication of this broader approach, though it may have just been very strong facts. It is in the framework of trying to not create a safe harbor for vertical mergers and say, “Listen, if you have a strong foreclosure case, bring it.”
Keyte: Herb, I want you to address this because it has almost to me a European feel of protecting rivals irrespective of market-wide effects.
Hovenkamp: I, too, think Guideline 5 is one of the sharper, more creative things in the 2023 Guidelines. In part this is because it is moving merger analysis in this area in a manner that is favored by economists. It calls for more economics rather than less.
Economists have been telling us for years that there is nothing particularly distinctive about vertical integration. It is simply a different kind of merger of complements; vertically related firms are complements, just like other types of complements, and we should treat them all the same way. That is what Guideline 5 does. It diminishes the role of vertical mergers as a distinctive category. It is really all about mergers of complementary products and tries to treat them under a common rubric.
Does it do it correctly? Well, in the notes, it changes the foreclosure threshold to 50%, which is a whole lot less than it used to be. But let’s observe that the standards we are currently using for tying and exclusive dealing are more in the 30%-to-40% range, so it is still not as aggressive as exclusive-dealing law and tying law.
I think overall Guideline 5 is a right effort, and I think it serves to reduce the distinctiveness of vertical relationships as a classification.
Keyte: Herb, let me press you on that a little bit. How do you deal with—to take what you described before—as a purely complementary acquisition that may affect a rival’s access to an optimal asset, but the Agencies can rely just on harm to rivals for Section 7?
Hovenkamp: I would add to the Guidelines that one needs also to show a market-wide output reduction or a price increase. That is, not every exclusion of a rival is competitively harmful, but ones that facilitate the exercise of market power are.
My main point was that this distinction between complementary and vertical is wholly artificial. If you buy a camera where you buy the body from one company and the lens from another company, you are looking at an acquisition of complements. If you are looking at a system in which the camera maker buys the lens and installs it and then you buy the complete package, you are looking at a vertical acquisition. There is absolutely no difference other than the way the product passes through the distribution chain, and there is really no reason to treat them differently for most purposes.
The real question we are asking is: What are we trying to get at here? I think a shortcoming in many of the Guidelines is that there is a punchline: that is, if they don’t lead to lower output or higher prices, then you have not crossed the threshold of showing that they are anticompetitive.
Keyte: Tim, would you agree with Herb that the principles in Guideline 5 still need to be connected to market-wide harm?
Wu: A lot depends on what you mean by the word “harm” there. As I have said, if we mean harm in the sense of price increase or a demonstrated obvious price increase projected right now, I think that ignores a lot of exclusionary harm or innovation-related harm, which over the long run are often more important.
I don’t think the government can ever win a case in a federal court without a theory of harm. I think Illumina/GRAIL is a good example of how things should function, which is the agency did not look at that and say, “Okay, it’s a vertical.” Ten years ago, they might have said: “Forget it, it’s vertical; we don’t do vertical cases.” The government did not go in there saying, “We don’t think there is any harm here, but we have this new Guideline we wrote.” I think they came with a strong theory of harm.
What the exact definition of harm is may vary in cases, and we may debate what that means, but I don’t think the government is ever going to go to court without a theory of harm.
Hovenkamp: Illumina/GRAIL was too easy a case. The total was 100 percent.
Wu: I am saying that ten years ago it might have been like, “Well, it’s vertical.” The cultural shift is as important as anything to me, which just says that vertical cases can be normal cases.
Maybe Live Nation/Ticketmaster would have been blocked if we had this Guideline in place, but I think instead they concluded, “Okay, so it’s vertical, let it go.” They did not have 100, but they had extremely high market shares.
Keyte: I think the scope of Guideline 5 is interesting and provocative, but the question of whether it must fundamentally be connected to market-wide harm remains, no?
Wu: Well, there is a broader point here, which is that it is a mistake to read these Guidelines as a statute as opposed to what they are, which are enforcement Guidelines; these are starting points. These are parameters of when they may bring a case, but not in every situation.
Hovenkamp: James, let me add that this would have been a good opportunity to bring the law of mergers into a little closer alignment with the law of tying and exclusive dealing, because it still leaves us with an open question. You can get condemned for something as unlawful exclusive dealing with a 30 or 40% foreclosure rate, but a vertical merger of exactly those two same products would be let go. In most circumstances that is not rational, but nevertheless the Guidelines really don’t even attempt to bring the law of vertical or complementary acquisitions into alignment with the law of tying and exclusive dealing.
Keyte: Let’s look at Guideline 6, which talks about dominance, entrenchment, and things like leveraging.
First, the language of dominance has always had a European feel to it, and if I have it right—I will start with you, Herb—I think it was the old Heublein case where it was Pitofsky who wrote in effect: “Well, if you are going to engage in post-merger bundling, we can wait and see if you do that and bring a Section 2 case.” This entrenchment Guideline also cites A&P, which appeared to involve scale efficiencies applied to a different marketplace. How do you react to this Guideline from either of those perspectives?
Hovenkamp: I think the Guideline does a miserable job of distinguishing probably the clear majority of circumstances in which entrenchment results from lower costs or product superiority and those where there is serious, net competitive harm.
I don’t think there are very many of those. If you go back and look at the draft under Guideline 6, it twice cited the Allis-Chalmers case, in which a maker of steel mills acquired a maker of the electric hookups for those mills. The Third Circuit condemned the merger and said: “Okay, this gives Allis-Chalmers a competitive advantage because it is going to be the only company that is able to sell a ready-to-go steel mill; everybody else is going to have to negotiate with two different firms; and that will entrench the post-merger firm’s dominant position.”
Then you look at what happened to that theory under Brown Shoe. Foremost Dairies—another FTC case right after Brown Shoe—condemned the merger of a dairy firm and an ice cream manufacturer because the post-merger firm would be able to deliver milk, butter, and ice cream out of the same truck as it was doing its delivery services.
With these older cases, we have to start asking ourselves the question: Do we really want to use merger law to prohibit firms from making better products or products that consumers prefer or reducing costs simply on the theory that these product improvements or cost reductions are going to make it more difficult for other firms to compete?
What the Guideline lacks, even in the final version, is any attempt to distinguish—it actually has a qualifying line, “We don’t want to go after procompetitive developments.”
I think the big majority of these kinds of acquisitions are going to be ones where the harm is going to come from reduced costs, preferred products, or some other kind of feature that makes the product better but in the process tends to entrench its dominant position. So I am not a fan of Guideline 6.
Keyte: Tim, how do you react to that criticism, as well as the notion that I think extends from that, that there is a little bit of “big is bad” in Guideline 6?
Wu: Most of all I see Guideline 6 as being a major pushback against the great weakness of the consumer welfare standard or the emphasis on putting a burden on the government to prove a price harm from every merger.
But the consumer welfare standard has never been great at dealing with exclusionary conduct or entrenchment because many of the things that you would care most about—like barriers to entry, switching costs, and similar forms of conduct—are difficult to reduce to the question of, “will this increase the price of this particular product by this much money?” People are still trying to do it, but it is challenging and often fails. So I see this section in particular, a little bit like the last one, as a pushback against what I would say is an area about evident harm to competition.
I hear Herb’s point that this may be an area where the defending merging parties will want to suggest that they are creating efficiencies. I guess my response would be that there is nothing here that says you cannot try to bring in your procompetitive justifications. In fact, I think the Guideline mentions it; I don’t think it deprives defendants of their normal opportunity to do so.
But I think the Guideline is trying—and you can almost see the cases here denying scale, and you can think of some of the economists behind them—to give the lawyers permission where it seems very evident that the whole point of the transaction is to create a new moat or wall, or add a moat to the wall. So that even if the Agency cannot prove an immediate likely price increase, the antitrust law still stands for the proposition that “You don’t get to make acquisitions to build new moats or castle walls.” That is what I think Guideline 6 is doing here.
Keyte: One thing the Guideline says is that you can’t rely on efficiencies if the proposed transaction tends to “create a monopoly.” So, to some extent, the Guideline dealing with entrenchment is addressing so-called “preexisting dominant firms” creating or maintaining a monopoly position. Tim, is this how you see it?
Wu: There is this overlap between Section 2 monopoly maintenance and the Clayton Act style. I think this is in part an effort to harmonize that, to ensure that the lawyers in their brains do not think: “Well, this is a Section 2 case, so we can’t bring it. We can’t do anything here.”
But a historical problem inside the agencies is thinking that: “Okay, we’ll let the crew over in the anticompetitive conduct shop bring a Section 2 case later.” So, Guideline 6 it is an effort to inject Section 2 thinking into merger analyses. You can see this with transactions that depriving rivals of scale, or generate network effects. You can also see Microsoft in here, as well as parts of the Google complaint. So, Guideline 6 is making sure that, in people’s brains, something that violates Section 2 should not be an exception to Section 7.
Hovenkamp: I think one place where I would want to push back a little is where the harm in Guideline 6 is simply the ability to deliver a better product. The real question is not whether there is an efficiency defense, but rather what is the source of the harm? Like I said, in Foremost Dairies the harm is that Foremost Dairies was able to deliver milk and ice cream at the same time to independent grocery stores. Is that really the kind of harm we want to be able to use in merger law?
In Allis-Chalmers the harm was, “We can deliver a complete, ready-to-go steel mill with its electric hookups so you don’t have to negotiate with two different contractors.”
Those are not Section 7 merger harms; those are basically consumer benefits. Yes, they do exclude rivals who are unable to meet them, but is that really the kind of harm that we want to use the merger laws to condemn?
Keyte: A good observation. Let’s skip ahead to rollups of private equity. I think serial acquisitions has always been out there, no? Anybody want to address this?
Wu: I do want to talk about the rollups of private equity, as this is one area where the Guidelines are trying to push the envelope a little bit.
In my time in government, I often thought that with private equity, when it came to mergers in the 1970s or so, government officials were saying: “There is something going on here. We don’t really understand it.”
Even the idea to try to identify certain forms of private equity conduct is an important step. That is all I want to say. It by no means ends of the conversation. I think it is very important for the government to have the conversation about what is good or bad and what the rules of the road should be for private equity.
Keyte: Herb, given Grinnell and Continental Ore, is this new?
Hovenkamp: I thought it was way too complicated. I think we should have a legal rule about serial acquisitions. I think what the agencies need to do is pick a presumptive time period that says: If you acquire, say, three firms in the same market within one year or maybe two years, you have to do a little empirical testing to see where you want to locate the standards, and then you can aggregate the acquisitions for the purposes of evaluation.
To the extent the Guidelines make this an examination of intent, I think it is inconsistent with the language of Section 7, which looks only at effects, and it also makes evaluating these mergers much more complicated.
All you really need to know is that if a large firm acquires A, B, and C spaced out over three to six months, one can aggregate their market shares in order to apply the Guideline standards. It would be a lot simpler, and though not perfect, it would be presumptive, and it would help provide an administrable standard.
Wu: The pushback is that the government does not feel like it really has a handle on what’s going on here, and if it sets a formal standard—private equity has shown a good ability to adapt very carefully to things like thresholds. I guess the concern is that they are kind of leaving themselves room.
I agree it is complicated. I think it reflects the government wanting to understand this area better and to be able to look at a broader range of evidence in this area. In a way, I think it does not do anything other than to say, “Well, I’m not going to ignore it.”
Keyte: Let’s talk about an area where I think they want to make it so broad that they are covering all potential bases, which is Guideline 9 on platform competition. My reaction to it is that there really is not guidance other than, “We’re going to use every other Guideline we can to stop pretty much any acquisitions, particularly of Big Tech firms.” Am I reading that too cynically?
Hovenkamp: You are reading it right. I do think it was very important for new merger Guidelines to add something about platforms. We are starting to see platform mergers. Now we have this announcement of Capital One and Discover in our future. There are going to be more mergers of platforms and it was time to take them seriously.
The way this Guideline treated them is an embarrassment. It is very one-sided.
First of all, the biggest thing that happens when digital platforms merge is that they become more valuable; that is to say they have more people on each side. If you merge two magazines, you get more advertisers, more subscribers, so you get huge efficiency gains.
To say that a little differently, the efficiency gains from platform mergers are way more than the efficiency gains from old-economy firms, provided—this is the ghost that is coming back to haunt Mark Zuckerberg—that you actually merge the platforms. The problem with Facebook is that Facebook purchased Instagram and then held it as a separate firm with its own separate network. But if networks are merged, you get enormous cost savings, enormous improvements in quality, and as a result I think they need to be treated more leniently overall than mergers in old-economy industries. That does not mean we should not be looking at them.
A critical question is: Is this merger going to be accomplished by actually consolidating the two platforms into one, or is it going to be more like the Facebook/Instagram story? None of that appears in this Guideline. I think it is completely useless as a guide to enforcement.
Keyte: Tim?
Wu: I would disagree with that. I share the sense that it is a little too much like an essay, as opposed to an enforcement Guideline, and a little bit of an “explain it like I’m five” Reddit post on platforms, although obviously it is a little more complicated than that.
Returning to a theme, I think it reflects a sort of mea culpa again over the last ten years and maybe a corrective to the sense that, “This is a platform, a complicated new business model, so let’s stay away from this.” There is a little bit of that flavor to it.
I have said many times that I think one of the goals of the Guidelines is to try to give the staff attorneys a sense that they should not be blinded by various labels when you are dealing with platform competition. I think it broadly assists in that goal.
Slightly in conflict with Herb, I do not think it is the Guidelines’ goal to give a pro and con. I think if anything, it is too essay-like and too unformed. I think they should identify situations where things can go wrong, as they do in other situations like foreclosure, where the writing and thinking is more mature. I don’t think it should name the worst-case scenarios, but it has this feeling that it was written like a bit of an essay.
Anyway, I think the major function of Guideline 9 is to try to do some battle with American Express and to disable the sense that you can simply ignore this particular merger because it is a platform. I don’t think it gives a very strong sense of when you do bring the cases or what the theory is.
Keyte: It is certainly a battle cry. Let’s talk a little bit about Guideline 10 on buyer acquisitions, workers, et cetera. What we have seen in the past on mergers here is the protection of very high-level writers, and now perhaps even rock stars, but maybe we are getting closer to traditional workers in the Kroger/Albertsons matter.
The question I have here is twofold: One, we have always had this as a marketplace, just flipped upstream; but there is this odd statement in the Guideline that these will “likely involve lower HHIs,”” and I don’t quite understand that. Why the asymmetry?
Hovenkamp: I think the asymmetry results from the fact—and this is a difficult fact for enforcers to address—that as you get lower and lower on the employee chain, markets tend to get bigger. This is something that is going to become an issue in the Kroger/Albertsons case.
The question is: For an employee who is working in a grocery store like Kroger, what is the relevant employment market for that person? Does that mean there is a market for grocery store employment? I don’t think so. I think most of the people who work in grocery stores can move over to shoe stores or Target or something like that. That is an empirical question. There is going to be a fairly high amount of mobility at the bottom. So it may be that the agencies are trying to compensate for that by saying, “We may want to use different concentration metrics.”
I am a little skeptical of the Kroger/Albertsons case. I will concede for a moment that there are some employees who work for large grocery chains who are fairly specialized—there are butchers, for example, managers perhaps—but for most of them the employment market is probably a whole lot broader than grocery stores, and grocery store concentration in most parts of the country is already pretty low. So I think it is going to be an uphill battle.
On the other hand, I applaud the Guideline for giving more attention to buy-side mergers, including mergers that have an adverse impact on labor, but we still have to go through a lot of hurdles with respect to proving that those kinds of harms will actually materialize.
For example, in Kroger/Albertsons, what is going to happen to the employees if that merger occurs? Even if there is an attempt to suppress wages, those workers are going to go to other forms of employment. There is a lot of stuff that has to be proved here. A lot of it lies in the ugly facts rather than the allegations.
Keyte: Tim, anything to add to that?
Wu: Yes, big picture. I think when we look back at this period, and even maybe the Biden Administration’s antitrust policies, I have often thought that the sleeper or the thing that may be regarded later as the most consequential change was taking labor markets seriously. I don’t doubt that it presents empirical challenges, but there is also broad agreement in this area. A lot of labor economists are saying, “for some of these mergers you are talking about maybe their effects on consumers we are not so sure about, but we are seeing effects on labor.”
Also, the Guideline is little closer to the original intent—this may be slightly disputed—I think of the Sherman Act or the Progressive Era, at least around the Clayton Act and a focus on labor, not just consumers.
I think the Guidelines had to have Guideline 10.
Hovenkamp: I agree.
Wu: It reflects what people are thinking, and it corrects a blind side. These cases are going to be hard to bring anyhow, and as Herb points out, so far the victories have been on behalf of bestselling authors—
Hovenkamp: Stephen King.
Wu:—like Herb Hovenkamp.
Hovenkamp: I’m afraid not.
Keyte: Maybe we’ll have some golfers in there next.
Wu: Golfers might be there and, outside of the merger context obviously, college athletes, so it may go in some slightly unpredictable directions. The bottom line is that I think this sort of had to be here, and this is reflective of what enforcement officials, lawyers, and economists are all thinking about.
Keyte: A quick question on efficiencies: Certainly, reading agency leadership, classic Neo-Brandeisian thinking—that efficiencies, especially in horizontal mergers shouldn’t really matter, and some of the leadership has been quite harsh about efficiencies—it seems to me the Guidelines gave back a lot of areas of efficiency that were a bit of a surprise.
We will have in this issue of Antitrust magazine an article by Mark Glick and others, and then there are obviously the writings of Eric Posner and others, that you should really not have efficiency arguments for most mergers.
I would ask Tim, do you think they caved on efficiencies, given where they might be more doctrinal in their views?
Wu: I have mixed feelings about it. It is hard to say you don’t even want to look at the efficiencies.
I agree with the critiques that say in some cases, maybe the Butterworth/Corewell Hospital merger, where you look at them and you are like: “Oh, my God, I can’t believe this. This gave the judge a way to give this a pass because he really trusted in his heart that the defendants were going to do the right thing.”
I just think it is part of our adversarial process, and maybe reflects again law enforcement, that you get to hear the other side of the argument, and it does demand that judges not just say, “Okay, they have a reason, you win.”
In other contexts, I do believe in per se rules. I do think the rule of reason has become very extensive outside of a merger context, but I think it is hard not to hear the other side.
Hovenkamp: I think the 2010 Guidelines were much better on efficiencies than the 2023 Guidelines.
Number one, by requiring that the consequence of provable efficiencies be such that prices are no higher or output no lower than they were prior to the merger, they made that argument about whether there is a distinct efficiency “defense” irrelevant. That is to say, one consequence of efficiencies is that there is not any harm to competition, and if you can show that, then we don’t need to worry about whether there is a separate efficiency defense.
The other point is that this is an area where we are just starting to see good evidence in merger retrospective studies. There have been two recent ones, very big, well-funded ones. The NBER has one from about November or so of 2023, and then there is a summary of another one in the Journal of Antitrust Enforcement from only three or four months ago.
They are very interesting because they show that even in mergers in highly concentrated markets—in fact, the NBER study was done under the 2010 Guidelines—pretty much every single merger in that study would have crossed the 2023 threshold and been challengeable. What that study found was that basically 25 percent of those mergers—this is over a period of close to twenty years—led to higher prices; on the other end, another 25 percent led to lower prices; and then in the middle was a fairly significant number of mergers that did not have measurable price impacts at all. I think those are the kinds of studies we need to be looking at more today.
The fact is I think we have seriously underestimated the extent of merger efficiencies. Secondly, we have not developed very good metrics for going after them.
By the way, the NBER study basically concluded that the metrics, pretty close to what the 2023 Guidelines adopt, are in about the right place, although they would still end up condemning a few efficient mergers and leaving some inefficient mergers alone.
This is one place where merger retrospective studies can do a lot of work.
Wu: One of my sidelines is that I teach a strategy class about mergers at the Business School. I think it is always eye-opening to hear from the other side some of the thinking behind merger decisions.
I will admit the possibility of efficiencies, but I think the trend you have to fight against is the discrepancy between doing a merger for one set of reasons and then coming up with efficiencies later.
It is not always completely easy to discern that, but the problem with the efficiency defense is that you can always pay to have lawyers and economists come up with convincing-sounding reasons that have nothing to do with the real reason that they did the merger in the first place.
I teach at our Business School that usually the reason we do a merger is to eliminate competition and increase profitability one way or another. I think occasionally it is about synergies, which is business school language for efficiencies, but more often it is either about building your wall higher or eliminating competition to raise prices. So maybe there is the possibility that you are efficient even though it was not really your intent.
If I were going in any direction, I would have said: “You know, if you find evidence that this was not the reason they were doing it, you should throw this efficiency evidence out the window. If you find significant evidence that they saw this as a way to get rid of this maverick or whatever it was that was irritating them with its price cutting, then don’t listen to this other stuff. Treat this a little more like a regular court case where you ignore a witness if they are not credible because they have said something different.”
That can still happen, but that is what I would have put in here, like trying to get the real reason for the merger, which sometimes is synergies, but I feel like in a small number of cases.
Keyte: Looking at the new Guidelines overall, some view the novel and provocative parts of the Merger Guidelines as being a little bit of a disincentive—I don’t know if it rises to a “scare tactic,” but it certainly adds costs on investigations and other things and maybe disincentivizes mergers.
At the same time, if they are going to be useful Guidelines, they need to be tested in courts, and we have not really seen cases from the agencies testing the more novel or provocative theories. Is there a concern there? Do you think it is most useful if they actually take some of the more novel areas and litigate them and see if we can get some updated principles of law?
Hovenkamp: 96 percent—I am pulling the number off the top of my head, but a very high number—of the merger activity under these Guidelines is going to be in Guidelines 1 through 3.
Vertical mergers and potential competition mergers are going to be distant afterthoughts; they are not going to get much attention, maybe a little bit more than in the past. I think there is going to be more attention paid to buy-side harm, but there we have not yet done the empirical work, and that could prove problematic. That is one side thing we will see.
Like I said earlier, I think our future in terms of really novel events is going to be platform mergers, where I don’t think we have a good understanding at this point of either the competitive threats or the benefits. I think we will conclude that the benefits are much more powerful than the harms when they are compared with old-economy mergers, but we are not there yet.
Keyte: Thank you, Herb. Tim, thoughts on that question for you?
Wu: I don’t think that agencies should bring merger challenges to try to reinforce theories. I think they should bring them when they have a good case. I think you get in trouble when you are stretching to try to create a theory of a case. I hope they don’t become enamored with, “Hey, we really need to prove”—frankly, this happened already, but I don’t want to comment on it too much—“that doctrine X or Y really does still exist.” I think that is a terrible reason to bring a case. Let’s bring this out of the closet.
I am teaching these Guidelines for the first time this spring. The students suddenly key onto it. They are like, “Hey, the earlier ones are the important ones, and the later ones are kind of additive.”
I agree with Herb. With the exception of I think the platform mini-essay, the action is really in the early parts of these Guidelines. I do not think the outer reaches will be tested very often, but I think unavoidably some of what the agencies want to do will involve the Guidelines, so obviously I think they will end up as part of the court experience.
Keyte: I am going to make one observation before I ask for final observations. And that is I look forward to a court actually assessing the unilateral effects issue where we do not have any case law, there is no Supreme Court case, and there is no circuit court case. I think it is going to be a very interesting and complex area for the courts, even though it is early in the guidance.
Tim, your final thoughts, and then we will get Herb’s final thoughts.
Wu: Let me talk a little bit about the White House perspective on these. As I was saying, in the economic wing of the White House, we suggested to the President that he suggest that the FTC and Justice Department seek to revise the Merger Guidelines to better reflect current practice. Obviously, they did not need to do it, but it was our suggestion. I think it was out of a sense of macroeconomic consequence that while we were dealing with a series of difficult economic times we allowed too much consolidation of many industries to happen—I am talking about the Obama Administration—under our watch.
If I am going to pick apart several themes here, I think there are overlapping themes that this is first designed to be a corrective to a period where it is now felt that merger policy had become too lax, too green light, and too cold. Maybe ten years later we will feel that it has gone in the other direction. I think that is what economic policy should be frankly, sort of reflective.
I reject the idea that you can achieve absolute perfection. I think economies are dynamic, and you need to take a look at what has been going on and correct if you think things have gone wrong.
I think there is this corrective aspect, and I think there is a strong normalization aspect, again, abandoning the search for a perfect algorithm-like Guideline and making it more like regular law enforcement guidelines where you suggest, “Hey, here are the kinds of things that a prosecutor might bring a case about,” these are possible grounds for a case. Defendants might not always like the fact that it is slightly open-ended, but that has been our system.
I think it is important, finally, in terms of normalization or corrective that these Guidelines, if anything, are also important for the staffs at the agencies to give them the confidence in their lawyering and their economics that they can bring cases when they strongly believe they see harm.
Those are the things I see going through here which are more culturally and policy oriented than technical.
Keyte: Thank you, Tim. Herb?
Hovenkamp: I think the Guidelines do several good things. We have talked about them. I think they also do a lot of fairly bad things.
I think the problem is that they convey the impression that, overall, they are not really very much in the mainstream. That worries me a little bit, particularly if Mr. Trump should end up winning the election. That is, are these Guidelines going to go the way of the Bush Administration’s Statement on Section 2, which did not last half a year before there was a presidential election, and Obama’s Administration promptly yanked them.
The 2010 Guidelines were more moderate, they were more consensus-developed, and as a result they survived the first Trump Administration. I worry about that. What is going to happen if President Biden loses the next election?
There are some good things. I would stick to the first three Guidelines, with a little bit extra enforcement with respect to labor and networks. I think we fundamentally agree about most of that. The other stuff is all fluff as far as I am concerned.
Keyte: We will end on that. Thank you both for your time and your incredible, insightful thoughts.
