Conversation about 2023 DOJ/FTC Merger Guidelines

July 23, 2024

Moderator

• Ian Simmons - Partner & Co-Chair, O’Melveny & Myers LLP, Editor, Antitrust Magazine

Panelists

• Susan Athey - Economics of Technology Professor, Stanford Graduate School of Business; former Chief Economist, U.S. Department of Justice Antitrust Division
• Bruce Hoffman - Partner, Cleary Gottlieb; Former Director of the Bureau of Competition, Federal Trade Commission
• David Lawrence - Policy Director, Antitrust Division, U.S. Department of Justice
• Martha Samuelson - Chief Executive Officer & Chairman, Analysis Group
• Nathan Soderstrom - Associate Director for Merger Analysis, Federal Trade Commission

IAN SIMMONS: Good morning, everybody. My name is Ian Simmons, and it has been a privilege of mine over many years to be an editor on the ABA Antitrust Magazine, an important pillar of a great Section of the ABA.

Today we have convened a panel to talk about the 2023 Department of Justice and Federal Trade Commission Merger Guidelines, a very powerful and informative document that I commend to the audience. We have five terrific panelists today, and I would like to briefly introduce them before we launch into our discussion.

We have Susan Athey. Susan is The Economics of Technology Professor at the Stanford Graduate School of Business. She is a John Bates Clark Medal winner, and Susan did her PhD work at Stanford. As many listeners may know, Susan recently stepped down as Chief Economist of the Antitrust Division of the U.S. Department of Justice.

We have Bruce Hoffman, who is a Partner at Cleary Gottlieb and a former Director of the Bureau of Competition at the Federal Trade Commission. Bruce is a well-known merger and antitrust lawyer and got his law degree at the University of Florida.

We are privileged to have David Lawrence, who is the Policy Director at the Antitrust Division at the U.S. Department of Justice. David oversees policy development and supervises the Appellate, Competition Policy and Advocacy, and International sections. David clerked on the U.S. Court of Appeals for the Second Circuit.

We have Martha Samuelson, who is the CEO and Chairman of Analysis Group. Martha is a terrific economist. She did her JD at Harvard Law School, her MS in Management at the Massachusetts Institute of Technology Sloan School of Management, and her BA at Yale.

Finally, we have Nathan Soderstrom, who is the Associate Director for Merger Analysis at the Federal Trade Commission. Nathan did his JD at the University of Minnesota Law School.

Panelists, welcome. Why don’t we start in with our discussion?

If my recollection is correct, the first Merger Guidelines were written by Don Turner and issued in 1968; we then had the 1982 revision under Assistant Attorney General William Baxter; the renowned 1992 Merger Guidelines that were moderately revised in 1997; and then the 2010 Guidelines.

Maybe David can start with this question—and Susan and Nathan, please chime in—what animated the need for a new draft of Merger Guidelines?

DAVID LAWRENCE: Thanks for the question, Ian, and thanks for hosting this discussion with so many thoughtful people.

To me two things really animated the project. First, as you just went over, our Guidelines have been updated periodically every ten or fifteen years, going back to 1968, and we were thirteen years following the 2010 Guidelines, so a revision was in order to account for a lot of different developments in the intervening years.

Second, we had a real desire to be more comprehensive so that the guidance fully reflected how the agencies review mergers. That is a good government concept. I came to this realization leading the 2020 Vertical Guidelines process back when Bruce was at FTC. I think that was a useful project because it put out guidance on a topic where the Guidelines since 1992 had a pretty big hole. Both agencies had pursued quite a few vertical mergers in that time, but we had no guidance on the topic. Even after we got the 2020 Guidelines out, the combination of them with 2010 still left some theories unstated. For example, in Visa/Plaid in 2020 the Department alleged that the merger would entrench monopoly power by increasing barriers to entry, but that theory was not stated in either guidance document.

In my view, it is just good government to have a single guidance document that actually tells the reading public how you are likely to proceed. The Guidelines are not much use to a businessperson if they can read them, try to think about and apply them, and then find out later that their merger triggered some common but entirely unwritten concern. So combining in one document all the theories from 2010, from the 2020 document, and our cases over that period was a real priority. I would say similarly working in some of the key legal principles and how they interact with the analysis helped us make a more comprehensive guidance document that I think better informs the public, and that is very much good government.

IAN SIMMONS: Thank you, David.

Susan, you participated in the drafting. Would you or Nathan like to chime in or add anything to David’s comments?

SUSAN ATHEY: Yes. To pick up where David left off, there were a variety of topics that were not covered or not covered in sufficient detail in prior editions of the Guidelines.

To think about the process of updating, in general, when we have economic trends in the economy and trends in firm strategy, their practices, and their business models, those trends drive changes in M&A strategy and patterns in M&A strategy. Alongside that, economic research and antitrust practice also follow those trends and have to adapt and update their toolkits to meet that challenge. Those things in turn merit more development and detail in the Guidelines.

Some obvious trends in mergers over the last couple of decades include mergers involving multiproduct firms and mergers involving platforms. The 2023 Merger Guidelines provided clarification on topics for multiproduct firms, market definition for cluster markets, bundled products, or one-stop shopping. These were not controversial updates, but they can be really helpful for avoiding costly confusion and possibly obfuscation in advocacy and litigation, just make the whole process easier and more efficient.

Platforms are a meatier topic, where modern scholarship has highlighted a number of scenarios where, just as an example, a dominant platform’s acquisition of a platform participant like a buyer or seller on their platform could be anticompetitive, but analyzing the potential harms from that type of platform merger requires several steps in the chain of logic and there may be several reinforcing effects. The Guidelines lay out the steps in that chain of logic, and in my view that is very helpful because it enables merger advocacy and litigation to focus more on the facts that support the theory of harm rather than needing to start from scratch and lay out these complicated chains of logic just in the context of the litigation, so again it makes for more efficient communication.

A final example is a merger of buyers. That has been a popular topic in the labor economics research literature in the last fifteen years and there is a large body of evidence showing that labor market monopsony is an issue and also studying the impact of mergers on labor markets. The 2010 Horizontal Merger Guidelines clearly stated that everything applied to buyer markets, but the 2023 Merger Guidelines provide more details about how that applies to labor markets. That is something that again is reflecting agency practice where agencies now routinely look at the effects of mergers on labor markets.

To an economist it is very straightforward to write down equations where you switch the role of buyer and seller and see that the math and arguments about inefficiencies for market power carry over directly to mergers of buyers, and this is a classic textbook exercise in introductory economics classes. But what I found is that for a non-economist it can be quite confusing to change all the variables and translate everything from sellers to buyers, and it is not obvious that all of the ideas carry over, so fleshing things out and making that explicit saves a step in the analysis.

IAN SIMMONS: Before I invite Nathan to come in, Susan, correct me if I’m wrong, but I think these Guidelines should be commended because they are really the first Guidelines that formally and holistically discuss the concept of platforms and the economics of platforms. Am I right about that? I don’t think the 2010 or 1992 Guidelines did.

SUSAN ATHEY: No. This is the first one.

DAVID LAWRENCE: There was one of the examples in 2010 I believe related to a platform circumstance, so some of the economics were there—I want to give credit there—but I completely agree with Susan that there is a great deal more content here, in large part thanks to her expertise to help us work that through.

IAN SIMMONS: Nathan, would you like to add anything about what considerations animated the need for new draft Guidelines?

NATHAN SODERSTROM: Happy to. Like Dave said, it had been thirteen years since the last update and we’d had significant developments in markets, certainly digital markets, the rise of tech platforms, et cetera. We’d had various studies showing increased concentration in various markets, as well as some empirical work, retrospectives that seemed to show anticompetitive effects from consummated mergers the agencies had reviewed. So, I think certainly the Commission was interested in thinking about ways to expand our toolkit and otherwise rethink some of our approaches to merger review.

One more point, going back to what Dave and Susan just said. We had the 2010 Horizontal Merger Guidelines and we had the 2020 Vertical Merger Guidelines, but our guidance documents did not reflect some of the cases we had brought, especially in the potential competition arena, or wanted to bring when it comes to entrenching dominance or labor markets, for example. I think that was important.

We do not want agency silence on a theory to be read as an agency-imposed limitation on our ability to enforce the statute. And having guidance documents that discuss the full menu of ways in which mergers can violate Section 7 we thought was important both for external purposes as well as internal guidance at the agencies.

IAN SIMMONS: Helpful, Nathan. I have been practicing for thirty-three years, and I am not sure I have ever met an antitrust lawyer who would take agency silence as acquiescence, so I think you are probably safe there.

Nathan, can you tell us about the process leading to the revisions in the 2023 version of the Guidelines?

NATHAN SODERSTROM: It was a really robust process. I won’t go into great detail, but I will mention a few points.

First, I think the AAG and others have been public about the fact that in drafting these Guidelines one work stream involved reading every appellate and Supreme Court decision touching Section 7 of the Clayton Act toward the goal of understanding what courts have said and furthering our understanding of what the law is. To some doing something like that may seem obvious, but in my mind it was an important innovation in getting this final product to one that was not only really comprehensive but also well-grounded in the case law.

Other than that, one aspect of the process that I thought was remarkable and frankly really well done was the level of staff engagement. The drafting team worked through each section of the draft Guidelines with staff who were specialists in the topics mentioned in the Guidelines. This was not just going through the motions or box checking. Certainly, staff in the Bureau of Competition was not shy about identifying places where earlier drafts of the Guidelines may have fallen short or needed clarification or amendment.

I think the drafting team, to their credit, was engaged and responsive to staff feedback. This led to a process that was admittedly lengthy, but also ultimately to a much stronger final work product that has real buy-in from staff who are doing the day-to-day work of merger review.

Engagement with external stakeholders was also really constructive. We had RFIs and comments on the draft, we hosted three public conferences after releasing the draft, agency representatives attended events organized by others, and certainly we read all the comments carefully. We were certainly gratified that many commenters recognized where revisions responded to their comments. I thought that was very constructive.

IAN SIMMONS: Nathan or others, can you think of one set of comments in particular that prompted what you consider to be the most notable revision? In other words, you put it out for comments, you received comments; did you ever say to yourself, “I am going to go and change this” because of this comment?

DAVID LAWRENCE: There was a joint comment cosigned by a number of prominent economists and academic experts focusing in particular—and we will talk a little bit more about this—on how the document used the term “competition,” the relationship of structure and market power. It was really, really useful in thinking about the issues and some of the ways the draft was written had unintended interpretations by some, so we worked closely with some of those folks on the revision and I think made some changes that helped address the issues that they raised. It was very constructive.

SUSAN ATHEY: Picking up on that, I think a lot of people were confused about the role of rebuttals in the previous draft, and I think that was a drafting issue of clarity, but it was such an important clarity issue that we tried to respond by making sure that people understood that all of the Guidelines were rebuttable.

Actually, that interacts with the competition definition because one way to evaluate each Guideline is to ask, “Is it specific enough?” It has to have a “Goldilocks” level of specificity—not too much, not too little—and it needed to be specific enough when in combination with the general description of what competition is and what harm to competition is so you would know what to rebut. I feel that the final draft was much clearer in that if you put together the pieces and the framing and the discussion of what competition means, it creates an understanding what a rebuttal would look like.

One last thing, which I think we can talk about a little bit more later as well, is that there was a lot of feedback on entrenchment. What we did was try to clarify our focus mergers that would lead to anticompetitive, in particular we tried to link the writing to concepts from Section 2 where there is a whole body of work supporting those words that made it more clear what we had in mind.

IAN SIMMONS: Terrific, Susan.

As an antitrust student, I do commend this set of Guidelines for its new topics—including a more-robust discussion of platforms—but I also commend it for its coherence in terms of the Division’s and the FTC’s other bodies of competition enforcement. It is not a unicorn. It does cohere. Someone might disagree with the philosophy that it is cohering, but it does cohere, and I commend that.

Martha, did you have anything you wanted to add about process before I move on?

MARTHA SAMUELSON: Yes, just a small set of observations as a participant. We felt very welcomed, I will say. We reached out to DAAG Doha Mekki and she set up a time for us to speak with David, Susan, and Ron Drennan (who was at that time the Acting Economics Director of Enforcement), and we talked actually at some length, Susan, as I remember, about the merging parties’ defenses. It just felt in the initial draft as if there was too much real estate or distance between setting out all of the potential competition concerns with mergers before one arrived at any discussion of merging parties’ potential rebuttals. Pro-competitive defenses were raised so much later in the back in an appendix and it felt as a result unlike a balancing of possible anticompetitive versus procompetitve outcomes.

Walking into the meeting, I guess I was not sure how receptive the DOJ was going to be about issues like this, but we experienced our discussion as productive and constructive participants. So, I thought it was a wonderful process as somebody not a drafter, outside, an external stakeholder.

IAN SIMMONS: Terrific.

I want to do a roll call on this one and bring Bruce and everybody into the discussion, but I want you to limit yourself to one or two sentence answers: What do you see as the most significant change in the 2023 Guidelines and why?

We will start with David on this one.

DAVID LAWRENCE: I think the most significant change is the use of the word “competition.” The new Guidelines use the word “competition” about 300 times more than the old ones, and there is a reason for that: we wanted to hew very closely to the statutory text and to the binding Supreme Court precedent, which of course uses the word “competition” consistently. We felt strongly that a guidance document can describe how we enforce the law but should not try to rewrite it, so it seemed imperative to center everything on the text and focus on that word “competition,” which just required a lot of rephrasing throughout.

This is also an area where we made some significant changes from the draft to the final. The comment I just mentioned and some other comments noted that when we removed the statement saying that the unifying theme of the Guidelines was market power we lost explanatory content that helped people understand what the Guidelines meant by competition. People asked, “Well, do we still think of competition in a way that relates to market power?”

The final Guidelines have in paragraph 2—and I urge readers to take a look at this—an explanation of what the Guidelines mean by “competition” and how it relates to market power. I will not quote it in full here, but I will just note that that was one of the most carefully crafted paragraphs in the document, one we actively shared during our public workshops and worked through down to the comma with the attorneys and economists who were engaged in providing us feedback. I have to say I am very happy with where it landed and very appreciative of the input we got from the community throughout the drafting process to get there.

IAN SIMMONS: Bruce, we haven’t brought you in yet. What do you see as the most significant change in these Guidelines and why do you say that?

BRUCE HOFFMAN: I have three. I will be quick.

I actually want to start by going back to something that Nate said earlier about the justification for the Guidelines. I agree generally that the timing was right for revised Guidelines and also that unifying the vertical and horizontal guidance made sense. I think the points Dave made about that are quite on point.

But I think the notion that there is empirical evidence of increased concentration or reduced competition in the U.S. economy and much less that any such trend, if it exists at all, has anything to do with antitrust policy, does not stand up to examination. There has been a fair amount of literature on this. If you are looking for an empirical foundation for changing Guidelines that is predicated on some notion that prior enforcement had failed in some way, that foundation does not exist. With that, let me give my three points on what I think are the biggest changes.

Number one, picking up on David’s last point about incorporating case law into the guidance and Guidelines, I think these Guidelines are more of an advocacy document than prior Guidelines in that they generally do not articulate merger benefits or positives. They do a lot of what I would describe as “selective citation” of case law, all of which is fine in a way—I don’t want to sound overly critical; I think there are a lot of useful things in these Guidelines—but I would say that in terms of providing guidance, compilations of case law in particular are less useful to practitioners and much less useful to courts, who can read cases as well as agency lawyers can, than guidance that incorporates economic and practical points about agency practices, things that courts and practitioners may be less likely to be familiar with.

The second big change is the downward revision of the concentration thresholds, which I think we will talk about more later.

The third is that these Guidelines introduced a number of theories such as entrenchment, or that talked about trends towards concentration or vertical integration that are new and that were not in prior Guidelines, so I think they are worthy of attention just for that reason.

I would say also that I think there are some serious questions about a number of these new concepts, what they actually mean, whether they actually make sense, and how they would actually be applied in practice in any way that would be useful in actual merger analysis.

Those are my somewhat critical-sounding points about the Guidelines. That said, though, I do think there is a lot of utility to the Guidelines in the way they came out, and we are using them in practice now, so we will see how that continues to develop.

IAN SIMMONS: Thank you, Bruce.

SUSAN ATHEY: Just to pick up on something that Bruce said, one of the other areas that we did receive a lot of comments on in the process were the citations to the case law. I think one of the small but I think helpful adaptations we did was to try to focus those citations at the start or setup of each section and then make more clear the structure of connecting the case law to the economics, which became a little bit more of a focus in the final revision.

What I wanted to talk about in terms of the most significant change was the change to operationalize the risk-assessment framework. The 2010 Guidelines also emphasized the words “may be” in the statutes, but they did not elaborate very much about how the risk-assessment framework affects the analysis.

Backing up, we have to observe that the Clayton Act is different than the Sherman Act. The Sherman Act in some cases balances harms and benefits to determine the overall net effect, while the Clayton Act requires an approach that is preventative and inherently involves probabilities.

The 2023 Merger Guidelines state that the agencies examine the totality of the evidence available to assess the risks the merger presents in order to determine if the merger has a reasonable probability of substantially lessening competition or tending to create a monopoly. I just want to explain why I think that risk-assessment framework is so important.

First of all, it helps us align economic evidence and expert testimony more closely with best research practice. In research, economic models and empirical evidence rely on assumptions and they necessarily focus on a smaller set of forces and outcomes, and researchers have to quantify uncertainty in their estimates and conduct sensitivity analysis; but they otherwise have to reason about how the findings of a more stylized and simplified model translate into real-world assessments of impacts, and that is a more qualitative assessment of interpretation of quantitative evidence. If a stylized model or an empirical exercise shows evidence of harm, the researcher might conclude that there is a reasonable probability of harm in the real world.

This conclusion, that there is a reasonable probability of harm, might be unchanged even if changing modeling assumptions changes the magnitudes of findings. So, that research practice is more in line with the way the Guidelines are set up now because the expert does not need to feel pressured to express greater precision or certainty than is consistent with a scientific research interpretation, and that makes it easier for non-specialists to step back and focus on the big picture of what the evidence points to and the credibility of the risk and focus on issues and modeling that have a large impact or change the conclusion about the risk of harm rather than the fine details of the magnitude.

The second reason this is such an impactful change is that many theories of harm—especially those involving related products, entrenchment of dominance or platforms—are inherently dynamic. A merger, for example, might increase the incentive and the ability of a merged firm to take actions that deter future entry or increase switching costs that affect future competition or otherwise protect its future profits at the expense of future consumers.

But, unlike short-term harm from raising price, it can be harder to quantify these long-term effects. We don’t know which firms would enter and what benefits those firms would bring if they did enter. The risk-assessment framework allows us to bring that kind of analysis in line with the Clayton Act so that we can show that there are many scenarios that together lead to a reasonable probability of harm to competition without needing to fully specify and quantify each one.

IAN SIMMONS: Helpful.

Would Martha or anyone who has not spoken on the change issue like to comment on this?

MARTHA SAMUELSON: I think this is a significant change. It is interesting actually. I have been—we all have—to some number of presentations in which DOJ discussed whether the Guidelines did introduce significant changes, and I thought that in many of these conversations DOJ’s perspective was that not much had changed.

DOJ’s position seems different here. I agree with all of the participants in this discussion think a lot has changed in these guidelines. I think that the guidelines are a very thoughtful document and that the document lays out frameworks and types of evidence that can be used in a very comprehensive way, but it does also feel more like an advocacy document to me.

In the 2010 Guidelines, merger analysis is described as assessing “what will likely happen if a merger proceeds as compared to what will likely happen if it does not.” That seems like simple and clear goalpost to me. I think of that description as a seesaw where you balance the evidence on one side and you balance the evidence on the other side and where the seesaw tips is where you want to be.

I think in the 2023 Guidelines, the focus is on the possibility of “harm to competition” and primarily on the possibility that a merger “may substantially lessen competition.” It seems to me that the implication is that the concern about possibly clearing an anticompetitive merger is larger than the concern about blocking a potentially procompetitive merger. It feels like the seesaw is weighted a little bit differently to me.

On the one hand, that troubles me. It provides, to me at least, less of a neutral goalpost and it opens the door to blocking mergers where the possibility of harm to competition is non-zero but also very speculative. Is that really a circumstance where we want to block a merger?

But I also think, as I think about it more—and this is the point you were raising, Bruce, about studies about the impact of concentration—as if some of the thought process is really more about belief systems than it is about empirical analysis of the impact of increasing concentration. If the concern is simply that there is more of a worry about the impact on competition of not blocking an anticompetitive merger than there is about the impact on competition of blocking a procompetitive merger, then this is where you would land.

I think the risk assessment is associated with that belief system, and I think even the concept of seeking to “arrest anticompetitive tendencies in their incipiency” is associated with that belief system. Incipiency means it is early in the process, you don’t know very much about what the outcome of a merger could be or the competitive implications of blocking a merger, but it feels as if there is a belief system in back of this that one type of concern is more important than the other type of concern.

DAVID LAWRENCE: If I could interject—I think the belief system that Martha is alluding to, the liability standard, is a decision that Congress made when it passed the Clayton Act. This area of risk assessment was another area where the imperative that we respect the statutory text played a big role. The statute uses the words “may be.”

The seesaw I think Martha was describing is very consistent with a lot of Sherman Act rule-of- reason balancing, which Congress rejected in drafting the Clayton Act and setting forth what the Supreme Court has recognized as a higher bar in this incipiency standard. Whatever you think this liability test requires, it has to be consistent with the words “may be” that Congress chose in the statute. I would agree that that leads you to an assessment of what could occur. It cannot be an illusory probability. It needs to be a reasonable probability.

MARTHA SAMUELSON: I completely understand what you are saying, David. I presume that the drafters of the 2010 Guidelines took the language of the Clayton Act seriously, and I think that the seesaw has shifted in these Guidelines and that there is a weight on one side of the seesaw in a way that was not present in the 2010 Guidelines.

IAN SIMMONS: Let me interject here. I want to stay on this issue, David’s point about the statutory language and Martha’s point about belief systems and the seesaw are critical.

Merger advocacy and merger litigation—and for the record, I am predominantly a litigator—is very interesting. In most litigation we are arguing about the past: Who did what to whom, why, and what was the effect? We have a record. We have documents. In merger advocacy and litigation we are arguing about the future, so there is a heightened degree of uncertainty—I won’t call it “speculative,” but we don’t know the future yet and we are trying to divine it through a process of point/counterpoint.

Guideline 6 of the current Guidelines (“Mergers Can Violate the Law When They Entrench or Extend a Dominant Position”) says: “In some cases, the nascent threat may be a firm that provides a product or service similar to the acquiring firm that does not substantially constrain the acquiring firm at the time of the merger but has the potential to grow into a more significant rival in the future.”

And then Guideline 6 also uses language such as: “Firms with niche or only partially overlapping products or customers can grow into longer-term threats to a dominant firm. Once established in its niche, a nascent threat may be able to add features or serve additional customer segments growing into greater overlap of customer segments or features over time, thereby intensifying competition with the dominant firm.”

Let me be a cynic for a minute. A cynic might say that litigation in federal courts is about evidence, credibility, and measurability. That is a little bit in tension with merger litigation. The person wearing black robes on the bench needs to be able to write an opinion based on things that are somewhat measurable. A cynic might say that the language I just read involves contingencies built on contingencies. Is our cynic right?

BRUCE HOFFMAN: I would say the cynic is wrong in this sense: I don’t think there is any doubt that the idea of a nascent competitor and the acquisition of a nascent competitor could be a potential competition problem is correct. It is undoubtedly true that acquiring nascent competitors before they grow into serious competitive threats could harm competition, and there are plenty of empirical examples of that.

It is also true that when you are looking at evaluating a transaction that involves the acquisition of a nascent competitor you are making predictions about the future just as you do in any other merger analysis. You are always going to be bounded by burdens of proof and standards of evidence to try to determine if you can show that that nascent competitor is likely enough to grow into a competitive threat that you meet the standards articulated in the statute and the case law.

To the extent the Guideline is attempting to say that there is some lower standard here below the levels of probability that would be attached to anything else, I think that would probably not be a viable reading of the Guideline and probably would not get traction. But that is not really how I read this; I read this as articulating a fairly straightforward proposition.

I think in practice when you look at these issues the real trick is: What do the facts show? This is an area of merger enforcement where you do not typically have things like a concentration threshold—you don’t have a lot of numbers to play with—and you are looking at predicting the future.

I can tell that when we have had cases like this—and we brought a couple when I was Bureau Director—what really became the critical issue for us was documents: What did the people think and what did they say inside the companies both in terms of the acquiring firm and its evaluation of the potential competitive threat of the firm it was acquiring and the firm being acquired in terms of what its officers, directors, leaders, innovators, whoever they might be, thought about their future trajectory—what they intended to do, what they planned to do, and what they thought was realistic.

I don’t know in this area that you are going to be able to get any more precise than that because again this is the area where in particular you are going to have less in the form of empirical work or other kinds of evidence you could evaluate. You are really going to be looking at what do people think they could do.

DAVID LAWRENCE: I largely agree with Bruce on this.

Going back, Ian, to your concept of courts measuring, the question Congress asked the courts to answer is “What may be?” The Supreme Court said that is “reasonable probability.”

It is not the only area of law where judges are asked that kind of question. In the “ineffective assistance of counsel” area the judges need to determine—and it is the exact same words from the Supreme Court—whether there is a “reasonable probability” that the outcome of the trial would have been different, a similar kind of question.

The Supreme Court has told us quite clearly—going back to that seesaw—that it is not just the standard of proof in a civil jury trial. It is whether there is a reasonable probability—which could be something less than a 51% probability but still be a reasonable probability. I agree with Bruce that it is hard to get much more quantitative than that, but it is clear that it is not a completely symmetrical analysis on both sides.

MARTHA SAMUELSON: To my knowledge, earlier Guidelines did not discuss the concept of a nascent competitor or how to analyze the acquisition of a nascent competitor. I think the possibility that the acquiring firm may help make the nascent firm a more effective competitor is also something that needs to be considered in this analysis. The possibility of incubating and leveraging is another aspect of thinking about the future, and of course, as you say very wisely, Bruce, based on the business documents I think this is an important path to think through as well.

IAN SIMMONS: Let me ask you this, and, Bruce, maybe you want to go first because I was the cynic and you nicely answered my question. That same cynic might also say the government tends to credit company documents: “Why would they write this if they did not believe this?”

A cynic might say the government likes to credit documents impugning the rationale or the effects of the merger—“I want to take out my next closest competitor”—but the government does not credit documents saying, “No, I think this is going be good for customers”—is there any validity to the cynics who say that, that there is only one category of documents that are credited, those that may be used to impugn the rationale or the effects of the merger?

BRUCE HOFFMAN: I am going to give you two answers to that. One is extremely granular, and I will do that one second.

The first answer is I think sometimes that happens. It is sometimes warranted, though. When I am looking at internal business documents that are talking about what a merger is going to do, what the competitive effects of the merger are going to be, what the purpose of it is, or anything else like that, then I think it is not appropriate to discount documents that suggest that the merger would not harm competition or would improve customer service, et cetera, if those documents are created in the same way as documents that might say negative things about competition.

On the other hand, I do think there is some difference. As you go down the road and the merger is actually being reviewed, you want to think about: Why are people writing documents? That does not mean that documents that are created later should automatically be discounted, but I do think that the government can appropriately look at what were the conditions under which these documents were created and what was the audience. If you are saying things internally to other senior executives or to audiences like boards or that sort of thing, I think that has more credibility. People are not going to make misrepresentations to their board of directors. On the other hand, if you are writing documents that have no readily apparent purpose that say lots of great things about the merger, then I would be fairly skeptical. I think again this is something that in every particular investigation might get weighted differently, but in general I think what I have said is the correct way to approach this issue.

The granular point is I do think the agencies can suffer from a confirmation bias problem, and this expresses itself when you get the staff memo—this happened very often in my experience—where the staff has concluded that they want to bring a case, all the evidence supporting the case is in the text and often bold and set out in quotes in the middle and all the negative evidence is buried in footnotes with a sort of backhand treatment of it: “Oh, and there’s a document that said something different, but it really shouldn’t matter.” I don’t mean to suggest that this is the product of some kind of bias or malevolence—it’s a natural human tendency to start to conform your analysis to the viewpoint you’ve come to hold. But it can tilt the tone of the recommendation.

So it is pretty tough. When you are in the front office and you get a 100-page single-spaced memo and all the bad stuff is in footnotes, you sometimes do not get a realistic view of the effects. We used to do things such as for some cases have devil’s advocate-type meetings and things like that to try to make sure we were avoiding that kind of confirmation bias.

It is really a management issue. You just have to make sure that people are not incentivized to overstate their case or to fall into attempting to tell the same story as everybody else. It is hard.

SUSAN ATHEY: I want to jump in on that. Having had the privilege of working with these amazing career staff for two years, the economists are especially good at being not quiet.

IAN SIMMONS: I am not going to comment on that one.

SUSAN ATHEY: It is something that, at least during my time—I agree, Bruce, that it is kind of a management issue—one thing I was very proud of was celebrating and giving a voice to people who could coherently articulate dissenting views, making sure those get a full hearing, and having those fulsome debates looking at both sides, and also this devil’s advocate-type of approach.

The Agencies have been engaged in a lot more litigation in the last couple of years, and I think when you get to litigation your arguments have got to be airtight. If the people working on a case think litigation is a serious possibility and not just a settlement, I think it motivates everyone to surface both types of evidence.

While I have the floor, I want to go back and pick up on something I thought Bruce explained very well, which is that in the end it is an excellent point that we need to start with the motivation for the merger. If we have a dominant firm with entrenched market power or persistently high profits over a long period of time, we want to understand what they were afraid of: Is this merger dealing with something they were afraid of that was threatening their market power; or was it about something like Martha indicated, a way to make themselves a stronger competitor?

One of the things that I really like about the Guidelines, putting together all of these different pieces, is that they make it a little bit easier to match up the case and the way the facts are organized with what the firms were worried about—with appropriate skepticism, as Bruce very nicely articulated.

One of the other areas we have in here is that the Guidelines discuss technology transitions (see, e.g., Section 2.6.A: “Nascent threats may be particularly likely to emerge during technological transitions . . . . A merger in this context may lessen competition by preventing or delaying any such beneficial shift or by shaping it so that the incumbent retains its dominant position”). That is an example where maybe it is something that is hard to describe and might seem hard to define, but when you break it down into the economics, the technology transition has this economic feature that it is a time when previously high entry barriers might be temporarily lowered and there may be a group of customers willing to switch products that weren’t before because they are adopting a new technology or they are anticipating bearing switching costs anyway, and that is a moment where a new firm can get a toehold.

If we look over history, there might be a long period of time where it is extremely hard to challenge an incumbent and then some periods of time where something changes and a competitor could actually get a toehold. Those moments are moments where we would want to be particularly careful about buying nascent competitors.

I think that connects back to what is going on with the merger. If they are saying, “We are terrible at innovation and people are doing this new thing,” where they are doing more cloud collaboration or they are moving to mobile, but something is changing and they are switching how they interact with products; and, “Gosh, our innovation team is pretty slow and our product team is behind, so we better buy somebody or else we might lose” versus “This is something that is going to make the market more competitive.” Buying a competitor is a lot worse for customers than being highly motivated to innovate, than having competition in innovation.

I think it is actually often not that ambiguous what the motivation is. Even if this new submarket that is taking place on the new technology is nascent and small, if you lose the unique competitive energy and innovation energy of the incumbent in the old technology, that is a real loss too. If they felt like their whole profit was at stake and they needed to innovate to create a new competitor, they might work pretty hard to do that. That is an important potential for loss of competition.

NATHAN SODERSTROM: If I could add to what Susan said but maybe take it in a slightly different direction, just on the question of pursuing our full statutory mandate, which is something the drafting team thought quite a bit about. As David and Susan both mentioned, the statute says “may” and not “will,” and that is an important word and it affects how we think about risk assessment.

Two other points there. The statute proscribes mergers whose effect may be substantially to lessen competition or tend to create a monopoly. And the back half of the statute—we saw references to it in the 1968 and 1982 Guidelines, but it disappeared from 1992 and 2010 as well as the 2020 Vertical Merger Guidelines.

I am not the first to observe this, but the disjunctive construction indicates that “tend to create a monopoly” must mean something. And you see the phrase “tend to create a monopoly” show up thirty-three times in the updated Guidelines. So just to say when we have mergers involving potential entrenchment, involving existing dominance, we want to signal that we are going to be vigilant about preventing mergers that may tend to create a monopoly, and we think we have the statutory authority to do so.

The other point I wanted to make is on mergers that may entrench a dominant position and whether those can violate Section 2 of the Sherman Act. Footnote 34 of the Guidelines discusses that possibility. And it is not just the current Commission or the current Antitrust Division—the Commission and Division under the prior administration took a similar position—the idea that we have a lower threshold for showing that if a firm has existing monopoly power and they acquire a potential competitor that kind of acquisition can violate Section 2 of the Sherman Act.

I think the underlying idea is that in those types of cases there is necessarily some speculation and some guesswork relative to, say, a merger of existing head-to-head competitors. But merger reviews are already about predicting the future. It is already tough to crystal ball these things, but we are not going to shy away from it. Like Bruce said, it still comes down to the facts, including the documents, what the parties and third parties say, and the economics. But again, just because there is necessarily more speculation or guesswork when thinking about potential competition, if anything, we think it may be more important to protect that type of competition.

IAN SIMMONS: Go ahead, Martha.

MARTHA SAMUELSON: On the technology transition that Susan raised—this, too, is a concept that has not been discussed previously in earlier Guidelines. And while a lot of what Susan says is correct, I would also note that many firms aspire to develop transformational technology. But it is a lot easier to see a technology transition when you look in the rearview mirror when there actually has been a significant change in technology that really ushers in a reconfiguration of an industry.

What concerns me is, because we don’t know and because everybody is aspiring all the time to be the firm that ushers in something like that, I worry that the thought process at the end of the technology transition analysis is simply that big and successful firms cannot make acquisitions, because that does strike me as not where we want to land.

IAN SIMMONS: Before we move on, I will take a brief editorial license. I am intrigued by page 20’s reference to technology transitions, but I do ask myself: How do we define it? And, Martha, to your point, don’t we know once we have passed through the tunnel but we did not necessarily know while in the tunnel?

Coming into this administration there was a view that enforcers had been overly concerned with false positives, and I think there is some validity to that. Doug Melamed wrote a wonderful piece about that.

I want to come back to this measurability point. It is all well-taken, David, Susan, and Bruce, everything that you said. After all, advocacy in litigation is about inferences, and it can be about inferences about the future. Inferences by definition require some degree of a leap of faith based on articulable facts.

But let me ask you this on measurability: The Guidelines (in Section 4.3) describe multiple paths to market definition including the Brown Shoe factors, which are Supreme Court precedent, and the hypothetical monopolist test. With respect to the hypothetical monopolist test, these Guidelines introduce the idea of a Small but Significant Nontransitory Worsening of Terms—SSNIPT in the nomenclature of these Guidelines. Can you give me concrete examples of how that test may work in practice? How do we gauge and evaluate a “worsening of terms” post-merger, something other than price, which we are used to thinking about?

DAVID LAWRENCE: This depends on the nature of the market of course, the dimensions over which firms compete. Imagine if my doctor’s office tomorrow were to change their terms so that they would sell my private medical information to the highest bidder, that would be a worsening of terms. It would not be a change in price, but you could use that kind of a change, depending on the context, to undertake the hypothetical monopolist exercise: Could they do that without me running to another doctor?

BRUCE HOFFMAN: Or if they just make you wait longer to get an appointment.

DAVID LAWRENCE: Yes, for sure. They already do that, but it could get worse.

MARTHA SAMUELSON: I think it is particularly useful in nonprice markets, zero-price markets, where what if with Meta you start to see a lot more ads when you log into Facebook? I think it is a really valuable addition and enrichment of the concept. It just fills in something that should have been there previously, the quality aspect.

BRUCE HOFFMAN: I take slight issue with that just in this sense. I think it is useful that it is articulated here. I don’t think it was ever missing. I think there was a little bit of a problem here with communication to the non-antitrust world. I think in the antitrust world we always understood the Small but Significant and Nontransitory Increase in Price (SSNIP) to include in the price term all other effects. It is just convenient to use price because I can write a number and an equation and assign in that equation some values to these other things, it just allows me to express it mathematically. I don’t think we ever meant to exclude effects on quality or things like that, but I think a lot of people did not understand that.

MARTHA SAMUELSON: Calling it out is valuable, I think.

IAN SIMMONS: That is well said, Bruce. I agree with that.

DAVID LAWRENCE: In fairness, I think that has always been the community’s understanding, but in litigation I have multiple times litigated against a defense side brief where the defense attorney who maybe in another setting would admit to that reading is writing to the court that it is actually only price and that there is no basis for defining the market because of the lack of price evidence. We all agree that that is economically and legally incoherent, but it still shows up in briefs, so clarifying it in the Guidelines seems useful.

IAN SIMMONS: Well said, David.

Bruce mentioned the Herfindahl-Hirschman Index (HHI) thresholds. I want to be constrained and short on this, but the Guidelines have altered what they consider to be a highly concentrated market from the 2010 Guidelines. I want you to be succinct and do a little bit of a roll call on this. Do you agree with this revision; if so, why; if not, why not? Do you think this was a revision that was warranted; if so, why; if not, why not?

BRUCE HOFFMAN: I will go first because I raised the issue before. I will say I don’t agree with this change because I don’t think it is empirically supported. The change in 2010 was a result of observed data on actual enforcement. I am not sure of any basis for the change back now and, as I said earlier, the actual data on concentration in the U.S. economy and the effects of merger enforcement do not support the notion that there has been under-enforcement or that there is an increase in concentration in any way that is relevant to antitrust policy.

DAVID LAWRENCE: I think the change is sensible. I will just disagree with Bruce on the empirical record first. I think in the last several years there has been an explosion in strong theoretical and empirical economic support for more aggressive thresholds than the 2010 Guidelines contain. You have Bhattacharya et al. (2023), Hoskin et al. (2018), Koch et al. (2021), Cooper et al. (2019), Dafny et al. (2012), Kwoka (2015), Ashenfelter & Hoskin (2010), to go along with folks like Carl Shapiro and Herbert Hovenkamp, who have been calling for this change on the basis of that empirical record. Frankly, I think the empirical record in support of these thresholds is stronger today than it has ever been, and I would just ask readers to go look at the actual record there.

The second reason I think it is a sensible change is that it is the law. These are the thresholds that were in place from 1982 to 2010, and there is an enormous universe of circuit court precedent applying those thresholds. That is where the law stands. You can count on one hand the number of circuit cases between 2010 and 2023 mentioning it and none of them interacted with a case in the intermediate HHI range such that they would form a holding of that circuit.

So, whether you are answering this from a policy perspective, empirically looking at the academic work, or you are answering it as a lawyer and trying to determine what the law really is, these thresholds seem well supported.

SUSAN ATHEY: Just to pile on a little bit, we got a lot of feedback. I spent eighteen months listening to a large number of people and reading very carefully what they had written. This particular change was one of the least controversial and it was something that many centrist economists argued for in their comments on the basis of this support.

One other comment on where it was before: It is sort of like having a speed limit: if the speed limit is sixty-five and you are saying, “Well, they are not giving tickets except to people going seventy-five, so let’s raise the speed limit to seventy-five,” but then they will start giving tickets above eighty.

There are a bunch of reasons why that can happen. It may continue to be the case that in the end the litigated cases—litigation is expensive and risky—probably will always remain above the threshold on average, but that does not mean that we don’t think there is a concern there.

One other thing that people sometimes forget about this is that when people argue about what is the theoretical support for the threshold and how does it link to the math of the economics, there are actually two considerations those thresholds are trying to support.

First of all, in a static Nash equilibrium, where everybody is best-responding to everybody else, what happens if you go from six to five or five to four and how does that affect prices? That is one way to think about what these thresholds should be.

But there is another thing, which is just interdependent behavior, not just explicit collusion but tacit collusion and cases where firms naturally are inhibited from taking aggressive actions because they can predict their rivals’ responses. That generally is related to the number of firms, and as things have gotten more and more observable it may become more of an issue going forward.

All of those reasons I think supported this threshold, but actually this was one where there seemed to be a broad centrist consensus that this change was reasonable in my reading of the comments.

NATHAN SODERSTROM: The 1800/100 HHI threshold was in place for roughly thirty years, from 1982 to 2010. My understanding is that in 1982 the reason those thresholds were chosen is that they roughly corresponded to four-firm concentration ratios courts had found problematic. So one can make the argument that they were based on what the courts had done; they were based on the case law.

In 2010 we had this observation that there was not a lot of enforcement below 2500/200. But, as Susan alluded to, there were reasons for that. I think the primary reason being that the agencies, as we all know, are resource constrained and can bring a finite number of merger challenges. By virtue of that, we are typically going to go after the most egregious examples of mergers we find illegal. And if we are choosing between a two-to-one and a six-to-five, we are usually going to go after the two-to-one.

So while it is true that you do not see a lot of merger challenges at these relatively lower thresholds, I think that is probably more a function of our resource constraints than it is speaking to the idea that mergers under that 2500/200 threshold but above 1800/100 necessarily are not a problem.

IAN SIMMONS: Excellent. Let’s move on a little bit, folks. Terrific conversation. Let’s talk for a moment about some of the newer aspects of the Guidelines. I am just going to lump them together given time constraints, and the panelists can answer as they see fit and put the emphasis where they deem appropriate.

Guideline 6 says that “mergers can violate the law when they entrench or extend a dominant position.” I thought it was an interesting and insightful complement to the other aspects of the enforcement agenda.

Guideline 7 says that “When an industry undergoes a trend toward consolidation, the agencies consider whether it increases the risk a merger may substantially lessen competition or tend to create a monopoly.” That sounds contextual and fact-based, so it seems like an interesting perspective.

Guideline 8 says that “When a merger is part of a series of multiple acquisitions, the agencies may examine the whole series.”

Guideline 9 says that “When a merger involves a multisided platform the agencies examine competition between platforms, on a platform, or to displace a platform.”

To this antitrust student, these are insightful and new perspectives. I don’t see why someone would say they are not warranted, but the audience is interested in the panelists’ views, not my views. What are your views on Guidelines 6 through 9?

MARTHA SAMUELSON: It is hard to lump them all together. They are quite different it seems to me.

Guideline 7—which discusses a merger in the context of industry trending toward consolidation—I think of as a plus factor. It seems quite straightforward to me. If there is a trend to consolidation in the industry, I understand why not do the analysis.

Guideline 8—which discusses a merger in the context of a series of deals—seems equally straightforward to me. If there is a cumulative set of deals and no single deal seems like it would be problematic—that is the Brown Shoe issue again obviously—that seems quite straightforward.

Guideline 9—which discusses a merger that involves in some fashion a multisided platform—I think is quite interesting. Platforms really are an unusual outcome in the economy and they are relatively recent. I think obviously it is important always with platforms to think about the incubation, what the procompetitive consequences of an acquisition might be, but I certainly understand what the concerns are as well.

BRUCE HOFFMAN: I will give a couple of reactions.

First of all, I agree with Martha, that these Guidelines cover such a wide range of things that it is hard to group them all together. There are certainly some things here that are helpful and positive and others that I think may be less so.

On the idea of entrenchment or extending a dominant position, I think that is a highly problematic concept. I think it is very difficult to articulate a situation in which that would be a predicted effect of the merger that would not otherwise be captured in the other forms of analysis you would do in the merger and where you have any confidence in an ex-ante prediction that a harmful result is even equally or less probable or more probable than a positive result because the conditions under which entrenchment or extension could occur are typically also the conditions under which the merger would likely have significant procompetitive effects.

In the past, we have typically said that these kinds of theories do not fit well with the idea of prospective enforcement because you are equally or more likely to harm competition or harm consumers by enforcement under these theories. I think there were some efforts to cabin this Guideline and to avoid mistakes of the past. I think it is instructive on that note to look at the U.S. commentary to the Organisation for Economic Co-operation and Development (OECD) in 2010 on conglomerate effects in mergers where it talks about entrenchment and points out serious errors in enforcement in the past by getting these ideas fundamentally wrong and condemning mergers that were very procompetitive because they hurt rivals, and that is a real danger here.

I think when you get into the others, again you start talking about some very different areas. I do not think offhand there is anything wrong with at least looking at trends toward concentration or vertical integration, but I think it is very difficult to say what inference you should draw from that. For example, if you see a trend towards vertical integration, that probably means that there is something in the economics of the industry that makes vertical integration procompetitive, and therefore enforcement efforts that take a negative view of vertical integration because most firms are vertically integrating is likely wrongheaded and is likely going to produce a perverse result. On the other hand, if you see an industry that is consolidating rapidly, then it seems legitimate to look at the merger as part of that consolidation and figure out where you are going to end up.

On platforms, I think it is a good idea to talk about platforms and I think the Guidelines do a nice job of calling out different kinds of competition that occur, competition between platforms and so forth. I think there is a danger here in that the Guidelines use the terminology “conflict of interest” to talk about self-preferencing. Those are freighted terms; they sound negative.

We have had some prior discussions about this, but as I and others have written and as empirical work is increasingly supporting, platform self-preferencing tends to be procompetitive; it is more likely to be procompetitive than anticompetitive. It is not that it can’t be anticompetitive, but the basic reason why a platform would vertically integrate and self-preference is most probably that it is trying to provide better service to attract more users to its platform. The empirical results in this area tend to show that price falls to customers after platform self-preferencing, so using terminology like “conflict of interest,” which sounds bad to people, I think is unwarranted here. This is an area that I think should be revised frankly to be more reflective of the reality of what is going on with these kinds of behaviors.

DAVID LAWRENCE: I might just respond to what Bruce mentioned about Guideline 6—mergers that entrench or extend a dominant position—because I think that the history of entrenchment in merger review is really fascinating, and I think of these Guidelines as describing a sort of “third era” that we are now in.

In the first era Bruce alluded to, in the 1960s and 1970s, the agencies very actively pursued entrenchment of dominance theories without regard to whether that entrenchment was efficiency-enhancing or exclusionary, and the Supreme Court recognized these theories.

Beginning in the 1980s, which I think of as the second era, the agencies became, as Bruce mentioned, very concerned about discouraging efficiency-promoting mergers. Rather than begin to distinguish in entrenchment theories between exclusion and efficiency as occurs in Section 2 cases under Grinnell, the agencies entirely discarded entrenchment of power as a concern of merger review, including when it was exclusionary, and the Guidelines that followed purged any mention of entrenchment. And, as Bruce referred to, we even criticized other countries around the world in fora like OECD that were pursuing these theories.

This third era, the modern era that the agencies are now in, I think involves a more systematic and nuanced approach to entrenchment that is targeted at mergers that threaten exclusion harmful to competition. This was something where there were changes from the draft to the final and something where we looked very carefully at this issue.

I would just point readers to the first couple of paragraphs of Guideline 6 that pull in some of these concepts from Section 2 that enable drawing this distinction to make sure that we are well targeting these theories.

SUSAN ATHEY: To quickly respond to that, personally I did my very, very best to try to draw the distinction Dave said, and I think we did get commenters afterwards agreeing that we succeeded—for example there was a comment by Fiona Scott Morton, who had been critical of the draft and felt that the final 2023 release addressed this issue satisfactorily. I want to again highlight that there are established concepts in the Section 2 body of academic and case law that help draw that distinction.

If you think about a firm, it is not just that, “Oh, a big firm can’t buy anybody.” The question is: Does this acquisition entrench their market power; is the acquisition anticompetitive in that sense?

The “conflict of interest” term actually has some more precise definition in the Guideline, but also a high-level way to think about it is that in a simplistic textbook world if a firm is threatened, the only options they have is that they can lower price or improve quality. That is why textbooks say that competition aligns the firm with the consumers; there is an alignment of interests of the firm winning consumers’ business and the welfare of the consumers.

But in other settings the firm can take an action that is misaligned with their customers and misaligned with the consumer welfare, and that kind of misalignment can come up. When you buy a related product, and that in turn would allow you to, and incentivize you to raise switching costs, that is a misalignment with your customers’ interests in order to protect and entrench your profits. I think we would all agree that is something that is concerning.

It is really hard to draw these lines, but luckily we do have a history of trying to draw those lines and I think we have to engage with finding those lines.

BRUCE HOFFMAN: That’s right, but it is important to note that the term “conflict of interest” is not being specifically used in those broader contexts. It is specifically used with the idea that a platform will offer its own products on its platform. I think it is really misused there because that is not a situation where in the canonical case that is likely to be in tension with consumer welfare. It might be in tension with competitor welfare, but basically it says that it is going to be beneficial to consumers more likely than not.

SUSAN ATHEY: I completely disagree.

BRUCE HOFFMAN: You’re just wrong.

SUSAN ATHEY: Well, I’m an author of some of the theoretical work! I would just encourage readers to go back and read the chain of logic that is specified in the Guidelines and look at the examples that are given of how that could work. Those are very carefully written to be entirely consistent with modern economic theory.

The Guidelines do not say, and it is not true, that every time a platform offers its own product—for example, a grocery store offers a store brand—that that is a problem for consumers, but it absolutely hurts consumers if the way that this works out, if tracking through the chain of logic you show that a platform buying a platform participant, like a seller on the platform, if you show that hurts competition between the platforms and hurts competition on the platform, that it hurts the end consumers. Absolutely. That is 100 percent there in the logic and it is supported in the academic literature.

That type of harm to competition does not always happen. There are conditions where that happens, and that is where the fact-specific evidence should be: Do the facts support the logical scenarios under which that happens? But if you want to say that it is impossible that a platform offering its own product on the platform would lead to harm to consumers, that is just not correct.

BRUCE HOFFMAN: No, no, let me be clear. Last point on this. I am not saying that. What I am saying is that the conditions under which offering your own product on a platform are likely to harm consumers exist but they require specified conditions and they are not the base case, they are not the most probable scenario.

This is kind of a narrow point I am making, but “conflict of interest” is a freighted term that is viewed negatively, and the use of it here suggests that a behavior which is on balance more likely than not to be beneficial should generally be suspect, which I think is just flat wrong. Can it be suspect? Absolutely yes, and you lay out some conditions here in the Guidelines, I think rightly so, that express some of the circumstances under which that could be true. I think that is all right.

What I am talking about is that using that term here I think was unwise, and I think it creates an inference that less-sophisticated audiences will misunderstand.

DAVID LAWRENCE: Can I just clarify very briefly, Ian? I agree with Bruce that this is conditional and I want to be clear. If you look at page 24 in the platform Guidelines, the conflict of interest is presented as a conditional concept that can arise when this divergence in interest occurs, and I think we would all agree as the conditions there present it. So, it is not stated differently in the Guidelines.

IAN SIMMONS: Fascinating discussion. I could listen to you all day. I think a podcast is the next step.

Let me go briefly to efficiencies. The Guidelines retain the idea of merger-specific efficiencies. But how has the concept changed, if at all? Do you think the Guidelines’ treatment of efficiencies is at the right volume and pitch?

Bruce, why don’t we start with you on this one.

BRUCE HOFFMAN: I think a couple of things.

One is I think the overall treatment of efficiencies is generally consistent with what has happened before, where efficiencies are taken into account in determining what is the aggregate effect of the merger going to be. I don’t really think there is a big difference here. Those who wrote these could weigh in if they think there is a major difference; I don’t really see one.

There are two areas I would highlight. One is the elimination of what I think was Footnote 14—David always corrects me on this because I always get the footnote wrong—where there was an out-of-market efficiencies footnote in the 2010 Guidelines which I thought was appropriate, which basically said simply that out-of-market efficiencies can be taken into effect in circumstances where they would be significant and where they could not be maintained or obtained without the anticompetitive effects of the merger; in other words, something like a divestiture would not preserve the efficiencies without creating anticompetitive effects. I thought that was a good concept, I think it is a shame that it is not in here, but I do think it is something that the agencies should keep in mind.

The only other comment I have on efficiencies is that I do think there is a mistake in the vertical arena. I don’t know if this is clearly articulated in the Guidelines, but I don’t think it is analytically correct to treat efficiencies in the vertical merger construct in quite the same way as we treat efficiencies in horizontal mergers because in vertical mergers the efficiencies are more inherent in the effects of the merger in the same way that any anticompetitive effect would be; and the same is true for pass-ons, so whether they are passed on or not is also symmetrical to whether anticompetitive effects are passed on.

I really think that in the vertical merger context you have to think a little bit differently about efficiencies and probably just start at a level playing field for both efficiencies and harms and evaluate them all together as opposed to putting a thumb on the scale in terms of allocating burdens of proof and requiring proof of pass-through and so forth. I just think that construct does not work well in vertical mergers. But that is not really a Guidelines point and that is something that predates these Guidelines.

IAN SIMMONS: Martha, any thoughts on efficiencies in our final moments here?

MARTHA SAMUELSON: Yes. I am going to back to the seesaw again a little bit. Merger specificity is and has been one of the factors that the Agencies consider when evaluating potential efficiencies. It seemed to me that in the 2010 Guidelines merger specificity was defined as benefits that likely would be achieved with the merger and likely would not be achieved without the merger.

It appears to me that the standard in the new guidelines is higher, that for efficiencies to be merger specific, it’s not just that they wouldn’t likely be achieved without the merger, but that it would be entirely impossible for them to be achieved without the merger. Put differently, efficiencies won’t be merger specific if they theoretically could be achieved without the merger even if they wouldn’t be achieved without the merger. For example, if additional scale enables the combined firm to invest in better production technology that it otherwise wouldn’t invest in, it is not clear to me that that would be credited here in the same way as it would have been credited in the 2010 Merger Guidelines.

DAVID LAWRENCE: I will agree with Martha inasmuch as I think the merger specificity prong here requires somewhat more than in 2010, and I think that is out of a recognition that contracts between firms are ubiquitous, lengthy, and sought for all sorts of problems in our economy these days.

Going to what Bruce mentioned about vertical mergers, they almost definitionally involve preexisting contracts, oftentimes very involved ones, between the parties. Taking a serious look at merger specificity seemed sensible, so we just revised that language to better fit the market realities we see out there.

IAN SIMMONS: In our final five minutes, what litigated cases do you believe are the most illustrative of the opportunities or challenges for enforcement under the Guidelines? These can be cases litigated to judgment or ultimately settled or deals abandoned, and they can be after the Guidelines or even illustrative of what the Guidelines say in the cases that took place before the Guidelines. If people could be succinct, I would like to do a roll call, and we’ll start with Bruce on this.

BRUCE HOFFMAN: The one I would highlight the most is Baker Hughes. I cannot recall if it is mentioned specifically in the Guidelines or not, but it lays out an analytical construct that I think courts tend to follow and I think courts are probably going to continue following that regardless of the Guidelines.

One issue with these Guidelines is that they do not really lay out a step-by-step process that a court can follow. Instead, they have these much more specific issues to consider—which, by the way, there is nothing wrong with that as a Guidelines matter; it just means that if you are a judge it may be difficult for you to follow. That is probably the one I would highlight the most.

I think beyond that there are a bunch of cases that are currently in various stages of litigation where these Guidelines may or may not start coming into play, and I think we will have to see how those evolve.

IAN SIMMONS: Thank you, Bruce.

Susan, do you want to go next?

SUSAN ATHEY: Sure. One case I will mention is the JetBlue/Spirit merger. This was an interesting case because there was going to be the removal of one competitor’s model that was proposed by the merging parties and they were going to turn all the Spirit planes into JetBlue planes and use JetBlue pricing.

One thing that I liked in that case was that the judge really took seriously the risk-assessment framework. Another is that he looked specifically at the harm to cost-conscious customers and concluded that there was a risk that there would not be competition for that segment of customers anymore.

It looks like a vanilla case to start with, but it actually raised some slightly nonstandard issues, and I think ultimately the risk and the “may be” was important in reaching the conclusion.

MARTHA SAMUELSON: I would mention Meta/Within. I think that in that case the government argued vigorously that Meta’s capacity to enter the virtual fitness market was sufficient and the court was not comfortable with that in the face of the business facts on the ground and the absence of a likelihood of Meta actually entering (without a merger). That strikes me as highlighting an aspect of the Guidelines where I do get concerned, focusing on capacity to enter instead of likelihood of entering.

IAN SIMMONS: Terrific.

David?

DAVID LAWRENCE: There are a number of merger and private cases where the Department participated as amicus while the Guidelines were under review that have vindicated, I think, some of the core legal principles here.

Obviously, Illumina v. FTC in the Fifth Circuit adopts a lot of what is reflected in Guideline 5 (“Mergers Can Violate the Law When They Create a Firm that May Limit Access to Products or Services That Its Rivals Use to Compete”).

In Deslandes v. McDonald’s in the Seventh Circuit, one of the debated issues about the labor markets Guideline was whether harm in a labor market alone really does suffice as the Guidelines say, and Judge Easterbrook answered that question strongly in the same direction as the Guidelines.

The Regeneron case in the Second Circuit—we talked about the market definition question earlier and the hypothetical monopolist versus Brown Shoe—earlier this year aligns completely with the view the Guidelines take that the agencies can have their choice of tools as plaintiffs for how to prove the relevant market.

Then there was the IQVIA decision. I was surprised at how quickly after the release of the Guidelines the 30 percent presumption that was drawn from Philadelphia National Bank was tested in a manner that is a holding. Because the defense expert had acknowledged I think a 31 percent market share, the judge found the Philadelphia National Bank presumption applied and rejected the same defense arguments we had been hearing in the comment process along the lines that old Supreme Court precedents don’t count.

IAN SIMMONS: Terrific. Thank you, David.

Finally, Nathan.

NATHAN SODERSTROM: As Bruce mentioned, we will learn a lot more in the coming months about how courts are going to read the new Guidelines. We have a historically busy merger litigation program at the Commission. We had Novant/CHS earlier this year, although of course the district court decision there was vacated. And then we have the Kroger/­Albertsons PI hearing in August, the Tapestry/Capri hearing in September, and Tempur Sealy/Mattress Firm looks like it is going to be in November. Those are all in active litigation so I won’t discuss them.

David mentioned IQVIA. Aside from the point he made, another thing I would mention about the IQVIA decision, which came out I think a week after the Guidelines were finalized, is that Judge Ramos spent quite a bit of time walking through the Brown Shoe factors as a pathway to market definition in finding that the Brown Shoe indicia supported our position that HCP programmatic advertising was a relevant product market.

That Brown Shoe pathway is certainly something the 2023 Guidelines emphasize, and it was gratifying to see the court implement that test and really dig into the Brown Shoe factors when thinking about our proposed market definition.

IAN SIMMONS: Thank you, Nathan.

That is our time, everybody. The red light is on here at the podium. I want to thank Susan Athey, Bruce Hoffman, Martha Samuelson, David Lawrence, and Nathan Soderstrom for a terrific discussion, and of course Michael Lindsay and Margaret Sharp of the ABA Antitrust Magazine. We only scratched the surface, but there is a lot in this transcript for people to read, learn, and think about.

To the panelists, thank you for your time and your hard work in appearing today. It has been terrific.