The 2020 Vertical Merger Guidelines, now withdrawn by the FTC, did not represent sound merger policy, argues Steven Salop; rather, they were overly defendant-friendly and based on a procompetitive presumption not supported by empirical studies or economic theory. They should be rapidly replaced with a more enforcement-oriented alternative.
Editor’s note: catch up on previous entries in our debate over the FTC’s withdrawal of the Vertical Merger Guidelines here.
I support the FTC decision to withdraw the 2020 Vertical Merger Guidelines (VMGs), which in my view—and as I discussed here, here, and here—do not represent sound merger policy. Like Singer and Steinbaum, I think that the VMGs are seriously flawed and should be rapidly replaced with a more enforcement-oriented alternative, one that includes anticompetitive presumptions. I look forward to new VMGs, hopefully along the line of my own suggested VMGs.
The 2020 VMGs represent an implicit and strongly procompetitive bias that is not justified by either economic theory or valid empirical studies. They require too little evidence from the parties to justify claims that “elimination of double marginalization” (EDM)—which explains how a vertical merger can lead to lower output market price because the merged firm transfers its inputs to its downstream affiliate at cost, rather than at a higher price—will lead to price reductions, and their discussion of EDM ignores some key issues. The defendant-friendly language of the VMGs generally would also make it more difficult for the antitrust agencies to win meritorious cases at trial. They also overlook some anticompetitive theories of (i) coordination and (ii) regulatory and contractual evasion.
The VMGs Ignored the Fact That Empirical Studies and Economic Theory Do Not Support a Procompetitive Presumption
The empirical evidence does not justify presuming that vertical mergers in oligopoly markets benefit competition. Earlier surveys of econometric studies, relied upon by commenters who preferred a procompetitive presumption, involved studies that often were methodologically flawed or did not really distinguish between competitive harms and benefits. The Beck and Scott Morton article critiques these surveys and reviews more recent empirical studies that use a more advanced empirical toolkit. Beck and Scott Morton conclude that “taken as a whole, these studies do not provide evidence for the proposition that all or most vertical mergers are good for consumers.”
Nor does economic theory support a clear procompetitive presumption. While vertical mergers can lead to efficiencies, they also can lead to anticompetitive effects. When a vertical merger raises input foreclosure concerns, there are inherent adverse horizontal effects.
The VMG Mistakenly Downplayed Analysis of Accommodating Price Increases
When there is imperfect competition in the input market, a merger foreclosure strategy of raising rivals’ input prices may lead other input sellers to increase their own prices. This accommodating response will worsen the impact on the downstream firms’ customers. The VMGs fail to spell out this analysis. As a result, such accommodation is more likely to be assumed away by agency staff and the courts.
This is not an idle concern. As discussed here, the FTC apparently ignored such accommodating responses in the Staples/Essendant office supplies merger, despite the fact that Essendant only faced one main wholesale competitor. By contrast, accommodation analysis was central to Carl Shapiro’s testimony in the winning private antitrust case by Steves & Sons attacking the Jeld-Wen/CMI merger, a case that the DOJ failed to pursue.
Even if staff take this accommodating effect into account and allege such responses in court, they will be disadvantaged at trial, since the VMGs do not clearly embrace the concern. For example, accommodating price effects are only mentioned implicitly and in passing in the context of simulation modeling, which also might suggest their consideration only in such models. Yet, qualitative evidence of input market concentration, product differentiation, customer specialization to certain input suppliers, and capacity limits might be highly probative by themselves. Thus, the VMGs would make it more difficult for agency trial staff to convince a court that accommodating price responses are important.
The VMGs Overemphasized and Implicitly Required Quantitative Evidence
Section 7 of the Clayton Act does not require that the agency establish its prima facie case of competitive harm with quantitative evidence. The DC Circuit made this point in its 2019 AT&T/Time Warner opinion. While the VMGs suggest that quantitative evidence is used “where sufficient relevant data are available,” their repetitive discussion of comparing EDM downward pricing pressure to upward pricing pressure from foreclosure suggests a requirement of quantitative evidence.
This emphasis on quantitative evidence will lead to certain concerns being downplayed. For instance, it is more difficult to quantify the impact of coordinated effects in a way that can be balanced against EDM. The same point applies to concerns that the merger will lead to rivals’ delayed access to new, higher quality products, a competitive concern suggested by the DOJ’s discussion of the abandoned LAM/KLA merger.
Requiring quantitative simulation studies raises the agency’s evidentiary burden in court. Even the most rigorous empirical studies can be criticized. The data is never perfect. There are always assumptions that can be nitpicked. There is an unlimited number of possible alternative assumptions and functional forms, and the VMGs flag the idea that specific simulation models can be attacked by proffering other models that do not predict “substantial price increases.” (And why are “substantial” price increases required?) A simulation model can always be made more complex, and then a court might ridicule the resulting complex model as a Rube Goldberg machine, as Judge Richard Leon did to Shapiro’s model in the AT&T/Time Warner case. (Shapiro’s recent article points out that AT&T’s economics witness, Dennis Carlton, then surprisingly agreed with the Judge’s colorful but misconceived characterization.)
While agency staff may be well qualified to separate the serious criticisms from the mere debating points, many judges are not. The complexity of the quantitative evidence in the AT&T/Time Warner merger trial illustrates this issue; and Judge Leon’s deficiencies, unfortunately, are not unique. Thus, a mandate to rely on quantitative evidence will significantly and inappropriately advantage the merging parties at trial.
Requiring quantitative evidence of competitive harms also inappropriately raises the agency’s evidentiary hurdle in another way. There is an intrinsic anti-enforcement bias in standard statistical tests. An American Statistical Association statement in 2016 made it clear that the usual p-values test for statistical significance do not distinguish true from false claims because they focus only on avoiding false positives, not false negatives. As a result, these tests provide the parties an opportunity to get their deal cleared, even if the expected (estimated) price increase is positive. This anti-enforcement bias is a problem, particularly in light of the Section 7 legal standard, which places greater weight on false negatives. Indeed, this same point applies to horizontal merger analysis.
“THE 2020 VMGS REPRESENT AN IMPLICIT AND STRONGLY PROCOMPETITIVE BIAS THAT IS NOT JUSTIFIED BY EITHER ECONOMIC THEORY OR VALID EMPIRICAL STUDIES.”
The VMGs Uncritically Embraced Elimination of Double Marginalization
The VMGs uncritically embrace efficiency claims involving elimination of double marginalization (EDM). EDM is repeated nineteen times in the VMGs, which are only fourteen pages long. Merger-specific EDM consumer benefits sure may occur, but they are not inevitable and they may not eliminate consumer harm.
The VMGs’ analysis suffers from several problems. The VMGs do not also require the parties to provide sufficient evidence that downstream prices will be reduced. This is important because if prices are not reduced, there is no consumer benefit. The VMGs also fail to explain the key technical point that the incentive to lower prices is mitigated or even eliminated when the merged firm takes account of its lost profits from the foreclosure reducing its input sales to foreclosed downstream rivals. They do not mention that downstream prices are less likely to be reduced if the merged firm fears a destructive price war. Downstream prices also are less likely to be reduced if there are MFNs, which are contractual provisions that require the seller to offer the buyer its best price. If a seller offers MFNs to all of its customers, it cannot offer any particular buyer a lower price than the others, a constraint that can deter price reductions. If the merging firms never even attempted to eliminate EDM with a contract in the period before the merger, it would be important to know why. Perhaps the evidence would indicate that the merger would not lead to lower prices charged to consumers.
The VMGs also erroneously presume that EDM is merger-specific—that is, could not be attained absent the merger. A general claim that there were “bargaining frictions” is not evidence; nor is it an adequate answer. For example, in a horizontal merger, the agencies would reject a justification that the merger is procompetitive because it would allow the parties to end their costly patent infringement litigation. The agencies would say that the parties could just settle the case instead. Reliable evidence justifying the “bargaining frictions” claim should be required. After all, the parties faced bargaining frictions in negotiating their merger agreement, but were able to reach a mutually acceptable deal. Indeed, if the parties’ only explanation for failing to achieve EDM is contractual bargaining frictions, the agencies would do better by introducing the parties to a top-notch mediator, rather than permitting an otherwise potentially anticompetitive merger.
The VMGs Failed to Adopt Anticompetitive Presumptions
The VMGs should adopt rebuttable anticompetitive presumptions that a vertical merger harms competition when certain factual predicates are satisfied, as suggested in my article with Jonathan Baker, Nancy Rose, and Fiona Scott Morton. We proposed several anticompetitive presumptions, including vertical acquisitions by dominant platforms and acquisitions of “maverick” buyers or sellers who were disrupting pricing coordination before the merger. Anticompetitive presumptions are appropriate for vertical mergers that raise more likely concerns.
The FTC Majority Statement and Possible Next Steps
The FTC has been unfairly criticized for unilaterally withdrawing the VMGs. The DOJ unilaterally released the 1984 Merger Guidelines and the 2008 Section 2 Guidelines without the FTC. I certainly understand why the DOJ would not feel it was appropriate for the Acting AAG to withdraw the VMGs: The Obama DOJ waited until Christine Varney was confirmed to withdraw the Bush administration’s Section 2 Guidelines. But this does not mean that the FTC had to be paralyzed. The VMGs were never an FTC consensus document; two Commissioners dissented only a year earlier.
The FTC’s Majority Statement also made clear the concern that errors in the VMGs “could become difficult to correct if relied on by courts.” And in light of the pro-defense language of the VMGs, the fear that merging parties would use the VMGs offensively is real. The 1984 Waste Management and 1990 Baker Hughes decisions both cited the DOJ’s Merger Guidelines to rebut arguments made by the DOJ litigators.
The language of the Majority Statement is not perfect. The statement makes the point that the statute does not permit affirmative efficiency defense to an anticompetitive merger, and appears to criticize the VMGs as suggesting that they would entertain such a defense. This language may have confused some readers: I interpreted the VMGs instead as treating EDM and other vertical coordination benefits as leading to downward pricing pressure that could prevent any lessening of competition, rather than as suggesting an affirmative defense (like the “failing firm” defense) that would allow a merger that harms consumers with higher prices (e.g., a merger that raises prices but substantially decreases fixed costs). I interpret the efficiency section in the Horizontal Merger Guidelines similarly, as a rebuttal factor rather than as a defense that would permit a merger that would harm consumers.
In addition, as flagged by Carl Shapiro and Herbert Hovenkamp, the statement at one point mistakenly appears to suggest that EDM only can occur when there are successive monopolies and fixed proportions input usage. It is the case that EDM can occur in a wider variety of market conditions, subject to the usual “cognizability” requirements of verifiability, merger-specificity, and sufficiency to prevent consumer harms. The old “single monopoly profit” result only occurs when there is a protected monopoly at one level, perfect competition at the other level, and fixed proportions input usage. So perhaps the two ideas were inadvertently conflated. But I do not think that this demonstrates that economics no longer has any role at the FTC—the statement elsewhere correctly relies on various economics articles. I also expect that the Commission appreciates that its litigation efforts will fail without a strong economic foundation. To allay any fears, I hope that the Commission makes clear its support of modern economic analysis.
The Commission majority might be wishing to return back to the legal standard of Brown Shoe and the spirit of Justice Douglas’ dissenting opinion in Columbia Steel. I am not privy to Chair Khan’s legal agenda, but her writings (here and here) raise that possibility. If so, that journey will require walking down a long and bumpy road, and possibly coming to a dead-end at the Supreme Court. At least as an interim step, adopting a more moderate policy based on a solid economic framework, such as my suggested VMGs, would provide useful guidance.
Disclosure: Professor Salop is a Professor of Economics and Law at Georgetown University Law Center and a senior consultant at Charles River Associates. He regularly consults on vertical mergers for government agencies, merging parties, and merger opponents on matters that he is not authorized to disclose. The opinions in this article are his own and do not necessarily reflect the views of others. Professor Salop did not seek or obtain the views of any clients in preparing this article.
Learn more about our disclosure policy here.
