A PYMNTS Company

Potential Competition Mergers: Lessons from Outside the Box

BY | February 10, 2022

The puzzle of potential competition mergers is not theoretical but empirical — verifying that the firm not in the market is a likely significant competitor and that very few others…

The puzzle of potential competition mergers is not theoretical but empirical — verifying that the firm not in the market is a likely significant competitor and that very few others are.  It may help to look at potential competition in other antitrust contexts.  From collusion, “reverse payment” pharmaceutical cases work only because regulation and legislation identify a unique potential generic competitor.  From abuse of dominance, the U.S. Microsoft case suggests low requirements for establishing that Netscape was a potential competitor to Windows, but in practice the case became equivalent in evidence and outcome to one about excluding actual competitors in browsers.  We conclude by looking at the relevance of empirical methods for assessing competitive risk in mergers, shifting the burden of proof, adding objectives beside “consumer welfare,” and, perhaps most important, focusing on how U.S. merger law specifies illegality when competition “may,” not “will,” be substantially reduced.

By Tim Brennan[1]


In principle, potential competition mergers are not puzzling. Understanding how a merger between an ongoing firm and a supplier not yet in the market could reduce competition in the future is neither difficult nor requiring the latest advances in industrial organization theory or even Chicago School economics. Courts have recognized a potential competition doctrine going back to at least 1964.[2]

The hard part about potential competition cases is not understanding them, but proving them. The two empirical challenges to any competition case might be labeled “not enough” and “too much.” Suppose that one wants to challenge Firm A’s acquisition of Firm B on potential competition grounds, that is, that Firm B is not yet a competitor to Firm A but will present significant competition to Firm A in the future. The “not enough” label applies to the question of whether Firm B will in fact be a significant competitor to Firm A in the future, even though it is not yet in Firm A’s market. The “too much” label applies to the question of whether if Firm B is a potential competitor, how do we not know that Firm’s C, D, E, etc. are not also potential competitors, rendering a merger between A and B inconsequential.

To get a handle on these two challenges, it may be useful to think beyond potential competition mergers as conventionally understood. A first step is to look at lessons from non-merger antitrust cases — collusion and single-firm conduct — where the issues in question primarily involved potential and not actual competitors. From collusion, we can look at arguments against agreements between incumbent pharmaceutical companies and potential suppliers of generic competitors, so-called “reverse payments” cases.[3] From single-firm conduct, we can look at the case brought by the U.S. government against Microsoft, where the motivating concern was stifling potential competition to the Windows operating system.[4]

In light of the current prominence of challenges to the consumer welfare and other conventions of merger enforcement, we look at outside the box observations in four merger related areas. The first is the relevance or lack thereof to current methods for evaluating unilateral effects mergers and coordinated effects mergers. A second involves whether plaintiffs should bear the burden of proof of showing harm, or that burden should be reassigned to the merging parties. A third issue is whether rejecting the consumer welfare standards is likely to usefully contribute to the adjudication of potential competition cases. Last is what should need to be proven in a merger case, based on the wording of the Clayton Act.



The canonical “reverse payments” pharmaceuticals case goes as follows: Firm I, the incumbent, holds a patent for a drug, D. Firm G develops a generic version of D, before Firm I’s patent on D expires. I sues G for patent infringement. G countersues I, claiming that the patent is invalid. I and G settle these lawsuits, where in exchange for G delaying its entry, I makes a payment to G. The term “reverse payments” comes about because I initially sued G for infringing its patent, but nevertheless I pays G. The competitive problem is that the result of this reverse payment patent litigation settlement is an agreement between I and G not to compete, at least for the duration of I’s original patent.[5]

I do not claim how well this portrayal fits any particular putative reverse payments case. Nor do I attempt to resolve two main critiques of these cases. The first is that courts understandably prefer parties in a lawsuit to settle rather than litigate; contesting a reverse payments case runs counter to that disposition. The second critique is that a patent entails a presumption that the holder has the exclusive right to use its intellectual property over the lifetime of the patent.[6] My only point here is to note that this is a potential competition case in the guise of a collusion case. The potential entrant is the producer of the generic.

However, the challenge of defining the set of potential entrants with whom the incumbent might collude, and that this may be a problem, is notably easy in these cases. Although the question of competing drugs to address a particular condition remains, we have an apparent incumbent monopoly, defined by patent protection. Second, the set of potential competitors is not open-ended, but defined by legislation and regulation.[7] Generics have to register with the Food and Drug Administration to be able to undertake activities that might otherwise violate the incumbent’s patent rights.

Moreover, as the Congressional Research Service has said:

The Hatch-Waxman Act further provided prospective manufacturers of generic pharmaceuticals with a reward for challenging the patent associated with an approved pharmaceutical. The reward consists of a 180-day generic drug exclusivity period awarded to the first generic applicant to file a paragraph IV certification. Congress hoped that this entitlement would encourage generic applicants to challenge a listed patent for an approved drug product.[8]

In other words, no one else can be a potential entrant for six months. Potential competitors in merger cases, on the other hand, will lack this kind of statutory and regulatory identification. The relative ease in identifying potential competitors in reverse payments collusion cases illustrates, by contrast, the absence of a clear means to do so in merger cases.



The familiar justification for the Microsoft case brought in the late 1990s and settled in the early 2000s was that Microsoft undertook tactics to protect its Windows computer operating system monopoly from future competition posed by the ability to use web browsers, then primarily Netscape, to serve as a platform for applications, such as (Microsoft’s) Word, Excel, and PowerPoint. The Java programming language would run on the browser and enable applications such as these to run on it. To the user, Netscape and Java together would, in effect, be that future operating system. Users would not have to run Windows and, importantly, personal computer manufacturers would not have to purchase licenses for Windows to install on their machines to enable users to do anything with them.[9]

To many, the Microsoft case would seem a model for how to assess potential competition. The trial judge found that Microsoft had monopolized a market for operating systems designed to run on personal computers using the Intel chip,[10] and on with regard to legal liability, the D.C. Court of Appeals largely agreed.[11] Although the D.C. Court of Appeals rejected the remedy the trial court had proposed, and the case was ultimately resolved via settlement, the finding that Microsoft had violated the antitrust laws directed at a potential entrant stood and, one might think, serve as a guide for potential competition merger cases.

Some aspects of the Microsoft case may be encouraging to those who want to bring such cases. It turns out that the record in this case did very little to support a potential competition case.[12] In particular, it did not establish a relevant market in future operating systems and applications platforms, and then show that this combination was the most likely potential entrant into this market.

This may seem like good news for advocates of potential competition cases. Despite this lack of a record, the Court of Appeals did ratify the trial court’s finding that the Microsoft had monopolized a market in Intel-based PC operating systems. One might infer that the evidentiary burden needed to find that had Microsoft instead purchased Netscape, similar evidence would suffice to prove an illegal potential competition merger.

However, the news is not so good when one considers the ultimate resolution of the case. Had the plaintiffs believed that Microsoft was preempting the development of operating systems through practices that directed users to its Internet Explorer browser and away from Netscape, the direct remedy would have been to force Microsoft to divest Internet Explorer, similar to a horizontal divestiture to resolve a merger case. This bullet was not bitten. Instead, and making a long story quite short, the parties eventually settled with Microsoft agreeing to halt practices that prevented personal computer manufacturers from featuring Netscape or other browsers along with or instead of Internet Explorer.

The eventual outcome of this case was identical, then, to a case that had been not about excluding future competition with the Windows operating system, but about excluding present competition with the Internet Explorer browser. It ended up being a case about harming competition in a present market, not about potential competition. Claims that the case was about potential competition in personal computer operating systems turned out to be eye-catching window (so to speak) dressing that had little to do to explain or justify the eventual outcome of the case. U.S. v. Microsoft may be less helpful in designing potential competition merger cases than some might wish.



If the lessons for potential competition mergers from non-merger cases are limited, looking at developments in merger assessment may be instructive.

A. Are Standard Merger Assessment Tools Helpful?

One place to look is the expansion of methods for assessing mergers.[13] Neither the coordinated effects-based market definition method nor the more modern unilateral effects approaches are useful for potential competition mergers. Regarding coordinated effects, the concern rarely if ever with potential competition mergers is that it will facilitate collusion with a larger but still small set of other firms in a relevant antitrust product and geographic market. Moreover, the crux of a potential competition merger case is that there are not such firms out there; otherwise, the unique threat posed by the acquisition of an entity not yet competing with the acquiring form is undercut.

This invites turning to methods for assessing unilateral effects mergers, where the effect of the merger is a reduction in competition among the merging sellers. The recent emphasis on this merger concern has led to (and likely been facilitated by) advances in the empirical assessment of the direct effects of on the incentives for those sellers to raise price following the merger. These include using “diversion ratios” — the fraction of sales lost by one merging firm to the other if the former raises its price — to estimate “upward pricing pressure” — the incentive to raise price following the merger. With sufficient data and reasonable assumptions, one can use merger simulation to estimate the effect on price when the merging parties maximize profits together compared to when they did so separately, taking into account pricing responses of other competitors.

However, because the potential competitor by definition is not yet in the market in any significant way — otherwise the word “potential” would not apply — there are no data to support those methods. There have not been any sales to divert, so diversion ratios cannot be used. There has been no ongoing competition from which one can estimate firm-specific demands and costs, from which one can simulate the effects of the potential competition merger.

The inapplicability of either coordinated or unilateral effects methods does not mean there cannot be a potential competition case. However, such a case will not rest on economic techniques. It will depend on documents, particularly analyses within the acquiring firm looking at the potential competitive threat from the acquired firm. (Presumably, an acquiring firm knowing this do what it can to minimize or eliminate the set of such documents.) Testimony from industry experts as to what may happen in the future will also be important. The key lesson, however, is that one cannot expect empirical economic methods to be as informative as they may be in assessing mergers between ongoing enterprises.

B. Shifting the Burden of Proof?

A second possibility to increase the likelihood of the plaintiffs prevailing when bringing a potential competition case is if the burden of proof were shifted from the agencies to the parties in merger cases. Such a shift has been most prominently suggested in a recent report of the majority staff of the House Antitrust Subcommittee.[14] Such a shift is not without precedent. Contested mergers and other proceedings before the Federal Communications Commission are designated for hearings where the merging parties (“applicants” for the transfer of relevant licenses) bear the burden of proof.[15]

Obviously, shifting the burden would make it easier to bring potential competition cases. Whether this is a good idea is another question. With law enforcement in general, the view is that the harm of incorrect guilty verdicts is sufficiently great that the state should bear the burden of proof. Mergers are not matters of criminal law, however, so this general principle need not be valid. The argument for the state bearing the burden in mergers has been that blocking a beneficial merger creates irrevocable harm, while the market will correct for anticompetitive mergers that erroneously make it through the adjudicative process.[16] This presumption regarding relative errors has received extensive recent criticism.[17]

If the recent criticism prevails, potential competition cases may become easier to bring. My standpoint remains more conservative. Much of the concern regarding too little enforcement of antitrust, especially with regard to potential competition, has been in big tech.[18] However, a major incentive to innovate in that sector is the prospect that a major current company will pay handsomely to add the new product to its service offerings. It remains uncertain that making potential competition cases easier to bring by shifting the burden of proof will give the public better products at lower prices.

C. Going Beyond Consumer Welfare?

A third theme to consider are proposals to move antitrust beyond what some see as the narrow confines of “consumer welfare.” Whether this would be good for consumers or these other goals is a debate I will not enter into here.[19] The question is whether additional objectives will make potential competition cases easier to bring and win.

We have little if any experience in court with the sort of evidences judges will want to see to factor those additional objectives into the assessment of conventional mergers, so at this point I can only speculate. My present guess is that additional considerations are not likely to help win potential competition cases. It is already difficult to determine that an acquired firm will be a significant competitor to the incumbent acquirer, if it is not yet in the market at the time of the acquisition. It is hard to imagine that a court will find it helpful to determine that the acquisition will have a significant effect on the distribution of political power, economic inequality, the distribution of wealth between capital and labor, sustainability, or other goals some want pursue through antitrust prosecution. But time may tell a different story.

 D. “Will” or “May”?

A final aspect that could affect the prospects for potential competition cases comes from a perhaps odd source: textualism, that is, interpreting statutes on the basis of the meaning of the terms in them at the time they were written. The historical aspect of this is not necessary; we can go directly to the phrasing of the Clayton Act, which proscribes acquisitions where:

the effect of such acquisition may be substantially to lessen competition, or to tend to create a monopoly.[20]

The terms of interest are “may” and “substantially.”[21] In antitrust, especially unilateral effects cases where the concern is suppression of competition between the merging firms — and, as noted above, the likely concern in potential competition cases — in practice the plaintiffs have to show that the merger “will” lead to a substantial lessening of competition. However, that is not what the statute says. One could imagine that a court would require plaintiffs (assuming they retain the burden of proof) to show that if there is an effect on competition from the merger, that the effect is substantial. However, they would need to show only that there may, not will, be that substantial effect.

Having to show only “may” would presumably make potential competition merger cases more attractive to bring. To get a sense of how radical a departure this might be, though, consider what “may” might mean if it false short of “will.” Presumably it should mean more than “could” or “might.” Should it be “more likely than not”? The difficulty in coming up with an adjudicable standard can explain why, a century after passage of the Clayton Act, merger law would have evolved toward a “will” test. But taking textualism seriously suggests that a plaintiff need not have to show that elimination of the potential competitor will cause substantial harm, only that it may cause substantial harm.



The rationale for contesting potential competition mergers is clear; the problem is validating that the acquired firm is a significant potential competitor and that others are not. “Reverse payment” collusion cases involve potential competitors, but there the potential competitor is identified by regulators and protected by statute. The Microsoft case appears to support a loose requirement for identifying potential competitors. But that case turned out to be equivalent to excluding an actual competitor from the market for browsers rather than excluding a future competitor from the market for operating systems.

Considerations specific to merger analysis do not seem more helpful for potential competition cases. Methods for identifying competitive harm through coordinated or unilateral effects lack the data to apply them. Shifting the burden of proof may be helpful as it would make any merger case to bring, and there is some precedent in FCC assessment of mergers, but whether that is a good idea will require a shift in judgment away from the view that harms from blocking good mergers exceed harms from allowing bad ones. Going beyond “consumer welfare” in antitrust prosecutions seems unlikely to give courts a firmer basis for identifying significant harms attributable to acquisition of a seller not yet in a market. The change most likely to affect potential competition cases would be to take that statutory “may” seriously — still prove that a reduction in competition would be substantial, but not require that the acquisition will bring about that harm. How to define “may” short of “will” remains to be seen.

1 Prof. (Emeritus), School of Public Policy, UMBC.

2 U.S. v. El Paso Natural Gas, 376 U.S. 651 (1964), cited in William Dorian, “The Potential Competition Doctrine: The Justice Department’s Antitrust Weapon under Section 7 of the Clayton Act,” 8 J. Marshall J. Prac. & Proc. 415 (1975).

3 FTC v. Actavis, Inc., 570 U.S. 136 (2013).

4 United States v. Microsoft Corporation, 253 F.3d 34 (D.C. Cir. 2001), settlement approved, United States v. Microsoft Corp., 231 F. Supp. 2d 144 (D.D.C. 2002)

5 Aaron Edlin, Scott Hemphill, Herbert Hovenkamp, and Carl Shapiro, “Activating Actavis,” 28 Antitrust 16 (Fall 2013).

6 See id. for responses to these critiques.

7 Congressional Research Service, The Hatch-Waxman Act: A Primer (Sept. 28, 2016) 5-7.

8 Id. At 9.

9 I don’t want to get beyond my technological expertise, but a Google Chromebook may be somewhat similar to the vision of a computer operating directly via a browser and not with a separate underlying operating system, at least not one visible to the user.

10 United States v. Microsoft Corp., 87 F. Supp. 2d 30 (D.D.C. 2000).

11 United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001).

12 The following draws from Timothy Brennan, “Do Easy Cases Make Bad Law? Antitrust Innovation or Missed Opportunities in U.S. v. Microsoft,” 69 George Washington L. R. 1042 (2001).

13 The following follows Timothy Brennan, “The Kinetic Rise and Potential Fall of Market Definition,” in Parcu, Pier Luigi, Giorgio Monti, and Marco Botta (eds.), Economic Analysis in EU Competition Policy: Recent Trends at the National and EU Level 68 (2021).

14 U.S. House of Representatives Subcommittee on Antitrust, Commercial and Administrative Law of the Committee of the Judiciary, Investigation of Competition in Digital Markets: Majority Staff Report and Recommendations 393 (2020).

15 47 U.S.C. 309(e). [“The burden of proceeding with the introduction of evidence and the burden of proof shall be upon the applicant ….”] I do not know the record, but legend has it that if the Commission is unsatisfied, the parties will generally if not always withdraw the merger, on the presumption that if they failed to persuade the Commission of that the merger is in the public interest, they would fail at a hearing on the merits. It would be interesting, if there is a way to do it, to determine if the different burden facing mergers in the telecommunications sector has led to better or worse outcomes than in the rest of the economy.

16 Frank H. Easterbrook, “Limits of Antitrust,” 63 Texas Law Review 1 (1984).

17 Jonathan Baker, “Taking the Error Out of ‘Error Cost’ Analysis: What’s Wrong with Antitrust’s Right,” 80 Antitrust Law Journal 1 (2015); Herbert Hovenkamp, “Antitrust Error Costs,” U. of Pennsylvania Institute for Law and Economics Research Paper 21-32 (Nov. 29, 2021), available at SSRN: https://ssrn.com/abstract=3853282.

18 See supra. n 13; Federal Trade Commission, Press Release: FTC Alleges Facebook Resorted to Illegal Buy-or-Bury Scheme to Crush Competition After String of Failed Attempts to Innovate (Aug. 19, 2021), available at https://www.ftc.gov/news-events/press-releases/2021/08/ftc-alleges-facebook-resorted-illegal-buy-or-bury-scheme-crush.

19 I’ve argued that this move will likely both hurt consumers and impede the achievement of other social goals apart from economic efficiency. Timothy Brennan, “Should Antitrust Go Beyond ‘Antitrust’?,” 63 Antitrust Bulletin 49 (2018).

20 15 U.S.C. 18.

21 Much of the following discussion is informed by discussions with Joe Farrell, but he is absolved of all errors and this view should not be attributed to him.