The U.S. Department of Justice and Federal Trade Commission 2023 Merger Guidelines: An Account and a Few Positive and Many Negative Assessments

Abstract

This Article summarizes and criticizes the DOJ/FTC’s 2023 Merger (M&A) Guidelines. Part I argues that the Agencies’ claim that the Guidelines are not binding, violates the antitrust laws’ addressees’ constitutional right to fair notice . Part II discusses the Agencies’ failure to articulate their understanding of the (M&A)-related tests of illegality the Clayton and Sherman Acts respectively, promulgate. Part III argues that the Agencies’ account of the U.S. antitrust law’s goals are ill-formulated and includes some goals of questionable desirability. Part IV explains why “market definitions” are inherently comprehensively arbitrary and why market-oriented approaches to analyzing the legality of (M&A)s are therefore inaccurate and their use by the Agencies is unconstitutional and avoidable. Part V delineates the various ways in which (M&A)s can affect the intensity of price-competition, analyzes the determinants of these possible impacts, and points out that the Guidelines mis-state the relevance of many such determinants and totally ignore many other such determinants. Part VI analyzes the various ways in which (M&A)s can affect the intensity of investment-competition, analyzes the determinants of the magnitudes of each of these possible impacts, points out that the Guidelines provide little information about the approaches the Agencies will take to these issues, and argues that the Agencies do not understand the determinants of the effectiveness of potential competition and may subscribe to the erroneous limit-pricing theory. Part VII delineates the correct way to analyze whether an (M or A) violates the Sherman Act and points out that the Guidelines provide almost no information about the way in which the DOJ will approach this issue. Part VIII criticizes various positions that the Guidelines take on the antitrust illegality of vertical (M&A)s.

As the Abstract indicates, the Article delineates and comments on eight components of the DOJ/FTC 2023 Merger Guidelines. ¹ Before proceeding, I will make three introductory points. First, this Article pays little attention to federal U.S. antitrust case-law and does not explicitly address antitrust-law-related Industrial Organization scholarship. The Article’s topic is the DOJ/FTC 2023 Merger Guidelines. It addresses the case-law only to assess the Agencies’ claims that some of the positions their Guidelines take are mandated by that case-law. It addresses the academic literature only implicitly in those instances in which the Guidelines’ positions it criticizes or supports are consensus positions in the IO literature. I have examined the case-law and academic literature in question exhaustively and exhaustingly elsewhere. ² Second, this Article approaches the 2023 Merger Guidelines as if they were a piece of scholarship—that is, does not view them as a political document or as a compromise reached by disagreeing Agency personnel. I adopt this approach inter alia because I believe that in the U.S. federal Agencies have a Constitutional duty not only to interpret and apply the laws they administer correctly as a matter of law but also to provide those that are affected by the Agencies’ choices with accurate information about how the Agencies will proceed. Third, this Article typically addresses the way in which the Guidelines treat a particular issue by explaining initially the way in which I believe it is correct as a matter of law (and economics) to address the issue in question and then comparing my conclusions with the statements that the Guidelines do and do not make about the Agencies’ approach or approaches to the issue in question. I am concerned that some readers may be led by this expositional strategy to conclude that the Article pays insufficient attention to the Guidelines: that perception is belied not only by the Article’s text but also by its multitudinous footnote-citations to Guidelines provisions.

I. The Binding Force of the 2023 Merger Guidelines

The 2023 Merger Guidelines state: “the Merger Guidelines create no independent rights or obligations, do not affect the rights or obligations of private parties, and do not limit the discretion of the Agencies including their staff, in any way.” ³ Although the following claim is certainly contestable, the Guidelines appear to me to reassert this claim in many later passages by stating that the DOJ and FTC will vary the methodologies they use across cases not only because the facts of those cases, the availability of those facts, and relatedly the techniques that are best adapted to discovering the relevant facts differ across cases (a practice that is completely unobjectionable) but because for some unexplained reason the Agencies will choose to use different methodologies in different cases (to exercise “discretion” to use whatever methodology whey want to use). I do not think it is morally defensible or Constitutional for the Agencies to regulate (M&A)s without issuing binding detailed Guidelines: operating without such Guidelines deprives the laws’ addressees of the fair notice to which they are morally and Constitutionally entitled. ⁴ In my judgment, the Agencies have a Constitutional obligation not only to follow the Guidelines they have promulgated but to issue appropriately-informative Guidelines.

II. The Tests of Illegality Promulgated by the Applicable Sections of the Sherman Act ⁵ and Clayton Act ⁶ and the Force of Judicial Precedents Interpreting and Applying Those Provisions

Section 1 of the Sherman Act prohibits any merger or acquisition (M or A) that constitutes a “contract [or] combination … in restraint of trade or commerce among the several states or with foreign nations.” Section 2 of the Sherman Act declares it to be a felony for any person to “monopolize or attempt to monopolize” any part of the trade or commerce among the several States, or with foreign nations …”in any way (including by consummating an [M or A] that violates this prohibition). Section 7 of the Clayton Act declares illegal any (M or A) whose “effect may be substantially to lessen competition or tend to create a monopoly in any line of commerce in any section of the country.” I will make six points or sets of points about the 2023 Merger Guidelines’ interpretations of these statutory provisions or the binding force of the federal judiciary’s interpretations and applications of these provisions.

The first is that although the 2023 Merger Guidelines recognize that the Sherman Act and Clayton Act promulgate different tests for the illegality of (M&A)s, ⁷ neither the Guidelines nor the DOJ or FTC elsewhere (nor for that matter, to my knowledge, any judge, any Industrial Organization economist, or any Antitrust Law law professor ⁸ ) has ever articulated the tests of illegality the two statutes make applicable to (M&A)s, much less offered a legal-argument justification for their test-of-illegality conclusions. I believe that the Sherman Act promulgates a “specific anticompetitive intent” test of illegality according to which an (M or A) is Sherman-Act-violative if the ex ante perception of one or more participants that the (M or A) was at least normally profitable was critically affected by a belief that it would or might reduce the absolute attractiveness of one or more of the best offers against which the participant(s) would have to compete in one or more ways that would render the (M or A) profitable even though it would be economically inefficient in an otherwise-Pareto-perfect economy. The part of this operationalization that begins with “in one or more ways” is included to render lawful investments that create new product variants, new distributive outlets, or capacity or inventory that increase the average speed with which the investor can supply a material good or service throughout a fluctuating demand cycle (what I denominate “quality-or-variety-increasing [QV] investments”) and investments in production-process research (PPR) despite the fact that ex ante the investor would not have found it at least normally profitable had the investor not believed that the investment would or might deter a rival investment if the investor’s ex ante perception that the investment was at least normally profitable was not critically affected by a belief that its investment would reduce the profits yielded by its other investments by competing against them and/or by inducing its rivals to offer better terms to potential buyers by less than those profits would otherwise have been reduced in these ways by the rival investment the investment in question would deter (in my terminology, unless the investor perceived itself to have a critical “monopolistic investment-incentive” to make the investment in question). I derive this test-of-illegality conclusion from the following legal arguments: the textual legal argument that focuses on the fact that as the nineteenth century progressed State courts came to define “restraint of trade” to be a ”restraint” imposed with specific anticompetitive intent, the historical legal argument that the legislators who passed the Sherman Act did so because they wanted to prohibit antecedent conduct they perceived to be motivated by specific anticompetitive intent, the historical argument that the legislators who passed the Sherman Act did not think that charging supra-competitive prices or realizing a supranormal rate-of-return constituted or inevitably manifested “monopolizing” conduct, and the Constitution-related legal argument that on my interpretation the Sherman Act would instantiate the liberal conception of justice the U.S. government is Constitutionally committed to instantiating (whose components are explanations of why the original U.S. Constitution and certainly the Constitution of 1890 commit the U.S. government to instantiating the liberal conception of justice, why that abstract commitment entails a concrete commitment to deter liberal-moral-rights-violative conduct and to give victims of such conduct appropriate opportunities to secure legal redress, and why business conduct that is motivated by specific anticompetitive intent violates the liberal moral rights of the perpetrator’s disadvantaged rivals and actual and potential customers). ⁹ If I ignore for the moment complications created by Clayton Act Section 7’s “may be” and “substantially” language and one other possibility I will discuss below, on my understanding, it would be correct as a matter of law to interpret Section 7’s “lessen competition” (and “tend to create a monopoly”) language to render any (M or A) illegal if it would impose a net dollar-loss on the potential and actual customers of the participants and their product rivals by worsening the absolute attractiveness of the best offers these buyers, respectively, received from any potential supplier that was not privately-best-placed to supply them—that is, that was at a competitive disadvantage when competing for the relevant buyer’s patronage (attributable to differences in the dollar-values the buyer placed on the privately-worse-than-best-placed supplier’s and the buyer’s best-placed supplier’s products and differences in the variable costs that the buyer’s relevant worse-than-best-placed supplier and its best-placed supplier would have to incur to supply the buyer in question when the relevant variable costs include what I call contextual costs [see below] as well as standard costs). The language that begins with “by worsening” is included because, in my judgment, it would be incorrect as a matter of law to find that an (M or A) violated Section 7 of the Clayton Act if it would not have imposed a net dollar-loss on Clayton-Act-relevant buyers had the (M or A) not created a resulting firm that was more able than the participants would have been to reduce the buyer surplus its customers obtained (1) by avoiding misestimates of its customers’ demand curves or its own marginal-cost curves and/or (2) by using “complicated” pricing-techniques (inter-buyer price discrimination, perfect price discrimination, a combination of lump-sum fees and supra-marginal-cost per-unit prices that does not constitute perfect price discrimination because the per-unit price exceeds the seller’s marginal cost at the output at which the relevant demand and marginal-cost curves intersect, and tie-ins and reciprocity agreements of various relevant kinds). The other possibility previously referenced is that it may be correct as a matter of law to interpret Section 7 to permit defendants whose (M or A) would otherwise violate it to exonerate themselves by proving that their (M or A) would not have imposed a net dollar-loss on Clayton-Act-relevant buyers had it not generated organizational economic efficiencies (for example, created a resulting firm whose marginal-cost curve[s] was [were] lower than the participants’ curve[s] would have been because the [M or A] created a resulting firm that could take advantage of real economies of scale or combined assets that are relevantly complementary for non-scale reasons) that worsened one or more rivals’ array(s) of competitive positions sufficiently to induce the rival(s) in question to exit: although no textual argument can be made for this organizational-economic-efficiency defense, it is favored by the fact that natural monopolies do not violate U.S. antitrust law as well as by the fact that U.S. IP law (and tax law) is designed to encourage firms to make investments that the government values because they are assumed to be economically efficient. In any event, I base my conclusion about the legally-correct interpretation of the Clayton Act’s “lessen competition” language on the fact that Section 7’s promulgators were responding to their realization that conduct that did not violate the Sherman Act could still impose dollar-losses on Clayton-Act-relevant buyers by freeing the participants from each other’s competition (and, possibly [see below], by creating a situation in which more Sherman-Act violations would be committed). To repeat: the first point I want to make in Part II is that the 2023 Merger Guidelines do not remedy the Agencies’ historic failure to define the concepts of “restraint of trade” and “monopoliz[ing]” conduct, which play critical roles in the Sherman Act’s tests of illegality, or the concepts of “lessen[ing] competition” or “tending to create a monopoly” (hereinafter “lessening competition”) that play a critical role in the Clayton Act’s Section 7 test of illegality.

Part II’s second set of points relates to the facts that the 2023 Merger Guidelines do not remedy the Agencies’ historic failure to define the expression “may be” and the word “substantially” that are part of its Section 7’s test of illegality. I do not think that much can be learned on this front from the fact that the Clayton Act was designed inter alia to prevent movements toward monopoly at their incipiency. ¹⁰ I admit that I do not have much useful to say on these “interpretive” issues. My disposition would be to interpret “may be” and “substantially” to authorize the Agencies not to challenge proposed (M&A)s or attack consummated (M&A)s that should be predicted to “lessen competition” in the Clayton-Act sense if they reasonably conclude that efforts to prevent their consummation or secure their dissolution would not further the goals of the statute, all things considered (considering the allocative transaction cost, public-financing-generated economic inefficiency, and distributive impact of any related government efforts and, possibly, those efforts’ deterrent effects).

Third, the 2023 Merger Guidelines indicate ¹¹ that the Agencies believe that it is correct as a matter of law to read the “any line of commerce” and “any section of the country” (emphases added) language of Section 7 to declare illegal every (M or A) that would (or did) impose (in the relevant way?) a net dollar-loss on Clayton-Act-relevant buyers operating in one or more than one line of commerce or one or more than one section of the country even if the (M or A) would not or did not impose a net dollar-loss or would or did confer a net dollar-gain on all affected Clayton-Act-relevant buyers. Admittedly, this legal conclusion is favored by a literal reading of Section 7’s text and might be rendered morally desirable if it implied that the Agencies were authorized to condition their approval of an (M or A) that would not impose a net dollar-loss on Clayton-Act-relevant buyers on the participants’ divesting parts of their operations when the divestiture would yield Clayton-Act-relevant buyers a net dollar-gain. However, I am not persuaded. Perhaps because of the difficulty of passing corrective legislation in our non-parliamentary system of government, U.S. courts often ignore statutory text that disserves what they perceive to be the goals of the relevant legislation—that is, often correct what the courts perceive to be legislative-drafting errors. And it is far from clear that it would be morally desirable to require a divestiture that would confer a net dollar-gain on Clayton-Act-relevant buyers in one or more lines of commerce or sections of the country in cases in which the proposed merger would not impose a net dollar-loss on all Clayton-Act-relevant buyers in the relevant way: requiring such divestitures in such cases might deter the proposal or consummation of (M&A)s that would be desirable and, even if it does not have this effect, ordering many of the relevant divestitures might be undesirable all relevant things considered even if doing so was in the interest of Clayton-Act-relevant buyers (e.g., because such Agency actions would generate allocative transaction costs and sacrifice efficiencies whose generation would not yield the relevant buyers dollar-gains that exceed the dollar-gain the divestiture would yield them by preventing the participants from freeing themselves from each other’s competition).

Fourth, the Agencies appear to believe that Section 7’s reference to competition in “any line of commerce” or in “any section of the country” legally obligates them to analyze the predicted or actual impact of a proposed or consummated (M or A) on competition in a defined product market or a defined geographic market and to use market-oriented protocols (in which predictions are substantially based on the magnitudes of market-aggregated parameters) to predict the competitive impact of any (M or A) in an allegedly-relevant defined market. My assessment of this Agency position is largely influenced by the conclusion of Part IV that market definitions (that definitions of “classical economic markets” and definitions of allegedly-functional “antitrust markets”) are inevitably arbitrary not just at their peripheries but comprehensively. If as I believe that conclusion is correct, the courts and the Agencies have two “interpretive options”: (1) declare the relevant sections of the Clayton Act unconstitutional on the ground that they do not provide the fair notice to the law’s addressees that is Constitutionally required or (2) execute a “saving construction” of the “line of commerce” and “section of country” language under which that language reinforces the implication of the word “substantially” that the Agencies are legally obligated not to challenge (M or A)s they predict will lessen competition in the Clayton Act sense when the lessening of competition they predict is too small for it to be desirable for the (M or A) to be challenged. Awkward as it may be, I think the second response is both morally desirable and consistent with our valid (because morally legitimate) judicial practice.

The fifth set of Part I points relates to the 2023 Merger Guidelines’ “Guideline 3: Mergers Can Violate the Law When They Increase the Risk of Coordination” and their reference to “anticompetitive coordination.” ¹² I will make two subsets of points about this Guideline.

The first subset is linguistic—that is, responds to fact that neither in the 2023 Merger Guidelines nor anywhere else have the Agencies either explicitly or implicitly defined what they mean by (non-tacit) “coordination” or “tacit coordination.” I start with “non-tacit coordination.” I believe that the Agencies are defining “non-tacit coordination” to refer to the type of interaction they also denominate “anticompetitive coordination” and I denominate “contrived-oligopolistic conduct.” On my definition (which fits the overwhelming majority of usages by economists [who have also not explicitly defined “oligopolistic conduct” or distinguished what I call “contrived-oligopolistic conduct” and “natural-oligopolistic conduct”]), “oligopolistic conduct” is conduct initiated by an actor’s making a choice it would not have found ex ante profitable had it not believed that its rival’s or rivals’ (hereinafter, “rivals’”) responses to that choice would be influenced by a belief that the initiator would or might react to their responses in one or more ways that would render unprofitable for them “uncooperative” (uncoordinated?) responses they would otherwise have found ex ante profitable. In my terminology, oligopolistic conduct is defined to be “contrived” if the reason that the initiator’s rivals anticipate that the initiator would react to such uncooperative responses in one or more ways that would critically reduce their profitability is that the initiator has communicated to the rival(s) (verbally or non-verbally) that it (1) will react to rivals’ unfavorable (uncooperative or “uncoordinated”?) responses by retaliating (by making one or more otherwise-unprofitable moves [for example, by charging uncooperative rivals’ customers inherently-unprofitably-low prices or by placing inherently-unprofitable advertisements that target the non-cooperator’s potential customers] to punish the non-cooperators for making their uncooperative responses to deter them and others from making such responses in the future) and/or (2) for analogous reasons, will react to rivals’ favorable “cooperative” (“coordinated”?) responses (decisions not to beat the initiator’s offers to its customers despite the inherently profitability to them of beating those offers) by reciprocating to the rivals’ cooperation (say, by enabling the cooperators to make more profits by supplying their customers by not beating offers the cooperating rivals make to their own customers despite the fact that it would be inherently profitable for the initiator to beat the cooperating rivals’ offers to their customers—[say] in cases in which the initiator was the cooperating rivals’ closest competitor for the relevant buyers’ patronage). I should admit that my claim that the Agencies are defining “non-tacit coordination” in the same way that I define “contrived-oligopolistic conduct” is disfavored by the facts that (1) in stating that a defining characteristic of tacit coordination is that it may not involve “an agreement,” ¹³ they are implying that a defining characteristic of non-tacit coordination is that it does involve the creation of an agreement and (2) on my definition of contrived-oligopolistic conduct, contrived-oligopolistic conduct that involves exclusively threats of retaliation (no promises of reciprocation) does not involve the creation of an agreement. I hasten to add that the use of the word “coordination” to distinguish “contrived-oligopolistic conduct” from other kinds of conduct is inapposite. Non-oligopolistic conduct in which each member of a set of competitors bases its decisions in part on the decisions made by its rivals involves “coordination” in the normal sense of that noun even if it involves neither “cooperation” nor the kind of conduct I define as oligopolistic or contrived-oligopolistic.

I turn next to the Agencies’ understanding of the concept of “tacit coordination.” I am disposed to conclude (perhaps generously ¹⁴ ) that the Agencies use the expression “tacit coordination” to refer to conduct that I denominate “natural-oligopolistic conduct.” In my terminology, oligopolistic conduct is “natural” when its initiator’s rivals anticipate that the initiator will or might react to any unfavorable (uncooperative) response they make in ways that would render ex ante unprofitable for them responses they would otherwise find ex ante profitable because they believe that the relevant buyer(s) will give the initiator a chance to rebid (to beat the rival’s [rivals’] superior offer[s]) and the initiator will find it both possible and inherently profitable to beat any rival offer whose acceptance would yield the rival profits because time is not of the essence and the reactive rebid of the initiator would have yielded it profits had it been made initially (relative to not bidding for the relevant buyer’s or buyers’ patronage at all) that exceed the extra cost generated by the making of the rebid. This interpretation of the Agencies’ understanding of “tacit coordination” is consistent with the Agencies’ claim that tacit coordination “would not itself violate the law.” ¹⁵ If “tacit coordination” is to be understood to be the type of “coordination” that is involved in natural-oligopolistic interactions, it would not violate the Sherman Act because it would not involve the creation of any contract, combination, or conspiracy in restraint of trade and (in my judgment) could not correctly be said to involve “monopolization” or “attempts to monopolize” as those words/expressions should be understood as matters of law in the Sherman-Act context. It should be clear that the Agencies’ use of the word “coordination” in the expression “tacit coordination” is inappropriate for the same reason that its use of the word “coordination” in isolation is inapposite: it ignores the fact that in normal usage conduct that is not “oligopolistic” in my sense certainly can accurately be described as “coordinated” even if it is not “cooperative.”

The second subset of points I want to make about the 2023 Merger Guidelines’ treatment of “coordination” contains legal as opposed to linguistic points. The Guidelines contain the following claim: “Because tacit coordination often cannot be addressed under Section 1 of the Sherman Act, the Agencies rigorously enforce Section 7 of the Clayton Act to prevent market structures conducive to such coordination.” ¹⁶ The Guidelines also implicitly assume that the Agencies are authorized to count against an (M or A)’s legality any evidence that it may “increase the likelihood, stability, or effectiveness of coordination [presumably non-tacit as well as tacit].” ¹⁷ I will make four related points, some of which at least in combination may seem paradoxical. My discussion will assume that the expression “non-tacit coordination” refers to what I call “contrived-oligopolistic conduct” and that the expression “tacit coordination” refers to what I call “natural-oligopolistic conduct.” First, because the fact that tacit coordination does not violate the Sherman Act or any other law implies or strongly favors the conclusion that actors whose ability to profit by engaging in natural-oligopolistic conduct (for example, by engaging in natural-oligopolistic pricing) will engage in the conduct in question, I think it legally warranted for the Agencies to count against an (M or A)’s Clayton-Act legality any tendency it has to increase the ability of the participants and/or their rivals to profit by engaging in tacit coordination. Second, the fact that it is extremely difficult to prove that actors have engaged in contrived-oligopolistic pricing or have contrived investment-restrictions favors the policy-case and may favor the legal case for the Agencies’ counting against an (M or A)’s Clayton-Act legality any tendency it has to “increase the likelihood, stability, or effectiveness” of non-tacit coordination. Third, it is certainly legally appropriate for the Agencies to count against an (M or A)’s Sherman-Act legality or Clayton-Act legality testimonial or documentary evidence that the participants in an (M or A) realized that it would increase their ability to profit from engaging in non-tacit coordination and intended to take advantage of that opportunity. Fourth and possibly problematically: despite (1) the fact that contrived-oligopolistic pricing (non-tacit coordination?) is liberal-moral-rights-violative while natural-oligopolistic conduct (tacit coordination?) may not be and (2) the fact that contrived-oligopolistic pricing is likely to be more economically inefficient and more likely to reduce total and average utility than natural-oligopolistic pricing, I do not think that it is legally permissible for the Agencies to count against an (M or A)’s Clayton-Act legality the fact that it would increase the profitability of non-tacit coordination (contrived-oligopolistic conduct) to the (M or A)’s participants because such conduct is Sherman-Act-violative and I believe that the probability that businesses will violate the Sherman Act when it would be profitable for them to do so is too low for it to be morally and legally permissible for the Agencies to adopt a contrary assumption when evaluating the Clayton-Act illegality of a proposed (M or A).

This Part’s sixth point focuses on the Agencies’ positions on whether they are legally obligated to follow the federal courts’ conclusions about the test of illegality the Clayton Act prescribes for application to (M&A)s and the protocol one should use to predict whether any (M or A) will lessen competition in the Clayton Act sense. I should state at the outset that I do not understand the positions the Agencies are taking on these issues in the 2023 Merger Guidelines. I quote in full the relevant paragraph:

These Merger Guidelines include references to appliable legal precedent. References to court decisions do not necessarily suggest that the Agencies would analyze the facts in those cases identically today. While the Agencies adapt their analytic tools as they evolve and advance, legal holdings reflecting the Supreme Court’s interpretation of a statute apply unless subsequently modified. These Merger Guidelines therefore reference applicable propositions of law to explain core principles that the Agencies apply in a manner consistent with modern analytical tools and market realities. References herein do not constrain the Agencies’ interpretation of the law in particular cases, as the Agencies will apply their discretion with respect to the applicable law in each case in light of the full range of precedent pertinent to the issues raised by each enforcement action. ¹⁸

I do not understand this paragraph. Are the Agencies saying that they are bound by judicial precedent or that they are not bound by judicial precedent even when it has not been reversed if it is inconsistent with “modern analytic tools” and/or “market realities”?

I will now provide my own answers to the following two questions the preceding quotation might be addressing: (1) Does judicial precedent legally bind the Agencies to use a market-oriented approach to predict whether any (M or A) would lessen or has lessened competition in the Clayton-Act sense, and (2) does the Clayton Act make illegal any (M or A) that lessens competition in one or more markets even if it does not lessen competition overall? I will make two preliminary points. First, I am not sure that the Supreme Court believes that the Clayton Act legally obligates lower federal courts and the Agencies to use market-oriented approaches to analyze whether any (M or A) lessens competition, though I do know that the federal courts have always used such approaches ¹⁹ and that many Agency employees believe that the federal courts believe that the Clayton Act requires that a market-oriented approach be taken to (M or A) competitive-impact prediction and that the Agencies are bound by this (supposed) judicial conclusion. Second, I do know that in at least one case ²⁰ the Supreme Court has stated in dicta that a demonstration that an (M or A) has lessened or would lessen competition in a single relevant market would render the (M or A) Clayton-Act-violative even if the (M or A) would or did not lessen competition in the Clayton-Act sense overall—indeed, even if it did or would increase competition in that sense overall. The point now at issue is whether the Agencies would be legally bound by any such position the Supreme Court or the lower federal courts took on these issues if the positions in question were unconstitutional or, for some other reason, were wrong as a matter of law.

I will analyze this issue on the assumption that in the United States prosecutors/regulators are authorized (not to bring prosecutions that are legally warranted)/(not to enforce lawful regulations) when after appropriate consideration they conclude that such a decision would be in the public interest (acceptably defined). The first issue is whether, if (contrary to my belief) the Supreme Court or the lower federal courts have held or have stated in dicta that the Clayton Act requires that a market-oriented approach be used to predict the competitive impact of any (M or A) (or other Clayton-Act-covered business conduct), that fact would legally obligate the Agencies to base their (M or A) approval/disapproval decisions on market-oriented competitive-impact analyses and to explain those decisions outside legal fora in market-oriented terms, and to make market-oriented arguments for their decisions in any related court proceedings. ²¹ In my judgment, the answer to each of these questions is “No!.” I base this answer to the market-oriented analytic-approach issue not only on the premise that the Agencies have discretion to use the analytic approaches they responsibly believe are in the public interest for them to employ but also (perhaps primarily) on the combination of the facts that (1) given the inevitable comprehensive arbitrariness of market definitions (see Part IV), the use of market-oriented approaches to predicting the competitive impact of any (M or A) (or any other type of business conduct) is unconstitutional and (2) like the President of the United States, DOJ and FTC officials are obligated “to the best of … [their respective] ability[ies], to preserve, protect and defend the Constitution of the United States.” ²² I believe that this obligation extends not only to avoiding engaging in unconstitutional behavior themselves but to explaining to the courts why any use they make of market-oriented approaches to competitive-impact prediction and any effort they make to require litigants to make market-oriented arguments is unconstitutional (though I am uncertain whether—after having explained to any court why the use of any market-oriented protocol would be unconstitutional and having justified their legality/illegality conclusion through a non-market-oriented argument—it would be unconstitutional for the Agencies to offer a spurious market-oriented argument for their conclusion that they think might be more persuasive to the court).

My legal justification for my conclusion that the Agencies are not legally obligated to (challenge all proposed mergers)/(attack all consummated mergers) that will not or did not lessen competition overall but that a court or an Agency concludes will or did lessen competition in one or more (arbitrarily-) defined markets (that is, to follow the Supreme Court’s dicta in Philadelphia National Bank) is based on the combination of the fact that the definition of the market in question is arbitrary and the fact that the Agencies have prosecutorial discretion to make this choice.

III. The 2023 Merger Guidelines’ Account of the “Goals” that Do (and Should?) Ground (U.S.) Antitrust Law

The 2023 Merger Guidelines list the various goals that the Agencies believe U.S. antitrust law is designed to achieve, which they appear to believe are desirable. Specifically, the Agencies claim that U.S. antitrust law aims to “safeguard … the Nation’s free market structures” in order to “ensure” “the preservation of economic freedom and our free-enterprise system” and thereby to obtain “the best allocation of our economic resources, the lowest prices, the highest quality and the greatest material progress” and to “preserv[e] our democratic political and social institutions” ²³ and that Section 7 of the Clayton Act seeks, by increasing competition, to secure “lower prices,” “improve wages and working conditions,” “enhance quality and resiliency,” increase “innovat[ion],” and “expand choices.” ²⁴

This Part (1) delineates the moral norms that should be used in the U.S. to evaluate its governments,’ citizens,’ and participants’ choices and considers the moral desirability of the tests of illegality that the Sherman Act and Clayton Act make applicable to (M&A)s from the perspectives of these moral norms and (2) considers the abstract meaning and extensions of the goals the Agencies claim U.S. antitrust law is designed to secure as well as the desirability of achieving those goals from the perspectives of the moral norms I claim the U.S. government is Constitutionally obligated or permitted to instantiate.

I believe that the U.S. Constitution obligates its citizens, participants, and governments to give “lexical priority” to instantiating “the liberal conception of justice” and obligates the governments of the United Sates to make decisions that that conception of justice does not control that in toto fit the distribution of (morally-defensible) “moral-good positions” of its competent citizens. I will try to make the expressions enquoted in the preceding sentence acceptably intelligible.

First, I distinguish “conceptions of justice” and “conceptions of the moral good” because I believe that Americans and the U.S. Constitution draw this distinction. For example, I believe that in the U.S. one addresses two different questions related to the moral position of an individual who is in a position to render assistance to an automobile-accident victim. The first is whether the potential assistance-provider has a moral obligation to provide assistance to the automobile-accident victim (whether justice requires the potential assistance-provider to supply aid)—a moral obligation that is binding regardless of whether the conception of the moral good to which the potential assistance-provider subscribes favors his or her provision of assistance. In the U.S., the answer to that question depends on whether the potential assistance-provider has promised to provide assistance to the victim in question or perhaps in the relevant type of situation, whether the potential assistance-provider was a culpable cause-in-fact of the relevant accident, whether the potential assistance-provider was an intimate of the victim (or perhaps had a status-relationship to the victim that is normally associated with intimacy), or whether the situation in question satisfies the complicated conditions for there to be a duty to rescue that does not depend on any of the facts just listed. ²⁵ I think that the preceding account is consistent with my claim that the U.S. is committed to instantiating “the liberal conception of justice” (see below), but the point that is salient in the current context is that in the U.S. even if one concludes that in the situation in question the potential assistance-provider has no moral obligation to provide help, one still asks a second moral question—whether, from the perspectives of one or more morally-defensible conceptions of the moral good to which individuals may subscribe, it would be morally good for the potential assistance-provider to supply help.

Second, I need to delineate my understanding of “the liberal conception of justice.” Space-constraints limit me to making five points. First, liberalism gives lexical priority to all moral-rights bearers (all those human and [possibly] those other creatures that possess the non-experience-generated neurological prerequisites for leading a life of moral integrity by fulfilling their moral obligations and conforming their conduct to a personally [though perhaps not self-consciously] chosen, morally defensible conception of the moral good ²⁶ ) having and seizing a meaningful opportunity to do so. Second, liberalism implies that all governments in polities that are constitutionally committed to instantiating the liberal conception of justice and all businesses in such polities have a moral duty to treat all moral-rights bearers in the relevant polities with appropriate, equal respect and to show appropriate, equal concern for all such creatures’ welfare. Third, except in some situations in which a moral-rights bearer has a need to be rescued from a situation in which its opportunity to lead a life of moral integrity is imperiled, liberalism does not imply that any business has a moral duty to benefit its potential customers or anyone else. Fourth, liberalism implies that business conduct that is motivated by specific anticompetitive intent (and/or constitutes unfair competition ²⁷ ) is liberal-moral-rights-violative in that it fails to show the obligatory respect and concern for the business’ potential customers and/or its disadvantaged rivals. Fifth, liberalism does not imply that businesses have a moral right to engage in profitable conduct that does not manifest specific anticompetitive intent and is not unfairly competitive but that does reduce the buyer surplus its potential customers obtain.

I turn next to the moral norms that in the U.S. ground the conceptions of the moral good to which different individuals subscribe either self-consciously or implicitly as manifest in the choices they make. In my judgment, all these moral norms are egalitarian. The list of relevant egalitarian norms includes moral norms that value positively and exclusively a choice’s positive impact on (1) the total or average utility of all creatures whose utility counts, ²⁸ (2) the equality of the distribution of utility, of resources (where a resource is valued by the net dollar-benefits its use would have generated had it not been allocated to the creature that got it), or of some non-liberal opportunity that is not valued exclusively for the utility the receipt or seizure of the opportunity gives the creature who receives or seizes it among all creatures who count (where the relevant distributive equalities are appropriately defined), and (3) two or more of the previously identified egalitarian ²⁹ desiderata. ³⁰

I will now outline the basis of my conclusions that the original U.S. Constitution and a fortiori the 1890 U.S. Constitution and the contemporary U.S. Constitution commit the federal government to instantiating the liberal conception of justice. My argument that the original U.S. Constitution does so focuses inter alia on (1) the Declaration of Independence’s ³¹ references to “the Laws of Nature” and Unalienable Rights” as well as to its insistence that the People have a right to “happiness” that their Government has an obligation to enable them to secure, (2) the fact that “happiness” was understood at that time to be the state of mind that is secured by leading a life of liberal moral integrity, ³² (3) the fact that many leading Founding Fathers subscribed to Enlightenment philosophy (which supported the liberal conception of justice), ³³ (4) the fact that the “Privileges and Immunities” that the original U.S. Constitution obligated the States to secure ³⁴ were understood to be what I denominate “liberal moral rights,” ³⁵ (5) the fact that the “Republican Form of Government” that the original U.S. “guarantee[d] … to every State” ³⁶ was understood to be a form of government that protected liberal moral rights, ³⁷ (6) the fact that the Founding Fathers valued the separation of powers and federal structure the original U.S. Constitution creates primarily because they believed that these institutional arrangements would preclude the development of liberal-moral-rights-violating oligarchy, ³⁸ and (7) the fact that the original U.S. Constitution contains a number of more specific provisions that protect liberal moral rights, including the provisions prohibiting the federal legislature from granting any “Title of Nobility” ³⁹ and prohibiting any State from “grant[ing] any Title of Nobility or passing “any Bill of Attainder, ex post facto Law, or Law impairing the Obligation of Contracts.” ⁴⁰

I recognize that the conclusion that the original U.S. Constitution committed the polity to instantiating the liberal conception of justice is disfavored by the various provisions of the document that (at best) acquiesce in the continuation of slavery. ⁴¹ I also recognize that the conclusion that the original U.S. Constitution committed the polity to instantiating the liberal conception of justice is disfavored by various other social realities—the apparent belief of the majority of white males that women have fewer political and other rights than men and that property qualifications for voting are acceptable. However, I believe that the force of these counterarguments is critically reduced by the moral discomfort with slavery that slavery-supporters manifested, by the felt need to make even abhorrent concessions to secure the Union, and by the fact that many of these illiberal positions are explicable in terms that reduce their subscriptions’ relevant argumentative force. ⁴²

Moreover, even readers who are not convinced that the original U.S. Constitution committed the U.S. to instantiating the liberal conception of justice should find the claim that the U.S. Constitution of 1890 (which includes the Bill of Rights ⁴³ and the Reconstruction Amendments ⁴⁴ ) and the contemporary U.S. Constitution (which includes Amendments XIX ⁴⁵ and XXIV ⁴⁶ ) do so.

I will now address the implications of liberalism and the various egalitarian norms I have identified for the moral desirability of the tests of illegality that the Sherman Act and the Clayton Act, respectively, apply to (M&A)s. As I have already indicated, liberalism implies that the government of any polity that is Constitutionally obligated to instantiate the liberal conception of justice is Constitutionally obligated to prohibit all (M&A)s (indeed, all conduct) prohibited by the Sherman Act (i.e., all conduct motivated by specific competitive intent), to deter such conduct in all ways that protect liberal-moral-rights-related interests all tolled, and to give victims of such conduct an appropriate opportunity to obtain legal redress. However, liberalism neither obligates nor “forbids” the government of the U.S. to prohibit (M&A)s that do not manifest specific anticompetitive intent but do “lessen competition” in the Clayton-Act sense.

What are the implications of the various egalitarian norms previously identified for the moral desirability of prohibiting (M&A)s that lessen competition in the Clayton-Act sense. I will focus initially on egalitarian norms that value exclusively the equality of the distribution of utility, resources, or relevant opportunities. I suspect that (1) in the vast majority of cases, conduct that imposes losses on Clayton-Act-relevant buyers will reduce the equality of those distributions and be valued unfavorably from these perspectives and (2) although analyses of the moral desirability of antitrust policies that prohibit such conduct from these perspectives are more complicated (because they must consider the distributive impact of the financing of the policies, the costs the policies impose on the law’s addressees by causing them to defend themselves and by punishing them, and the distributive impact of the policies’ general-deterrent effects), the overwhelming majority of policies that effectively prohibit conduct that violates the Clayton Act’s “lessening competition” test of illegality will increase the equality of those distributions and be valued positively from these perspectives. There will be exceptions when the weighted-average disadvantaged buyer is better-off than the weighted-average advantaged manager/shareholder/employee of the perpetrator(s) (where the weight assigned to each individual’s income/wealth position is proportionate to his or her share of the relevant group’s dollar loss or gain) because the good involved is an expensive product or service whose units are sold to individual final consumers and/or when the formulation, promulgation, and enforcement of the policy at issue have public-finance-related and other distributive impacts that critically reduce the equality of the distribution of a relevant desideratum.

The analysis of the implications of the two utilitarian norms for the moral desirability of exemplars of antitrust-policy-coverable conduct that lessens competition or for any antitrust policy that effectively prohibits such conduct is far more complex (1) because it involves not only (A) the consideration of the relative income/wealth positions of the conduct’s or policy’s winners and losers—the relative magnitudes of the average marginal-utility value of the dollars gained and the average marginal-utility value of the dollars lost—but also (B) the relationship between the total dollar-gains and the total dollar-losses the choice will generate—that is, the economic efficiency of the business choice or antitrust policy in question—and (2) because the relevant economic-efficiency analysis is complicated and may yield conclusions that make it far more difficult to assess the desirability of any choice from either the total-utility or the average-utility utilitarian perspective or from any mixed egalitarian perspective in which increases in total or average utility are positively valued. The protocol that should be used to analyze the economic efficiency of choices is complicated: I will simply assert here that the protocol for economic-efficiency analysis I deem ex ante economically efficient strongly favors the conclusions that poverty causes economic inefficiency ⁴⁷ and that developed economies devote an economically-inefficiently-large amount of resources to the creation of new products, distributive outlets, and capacity and inventory (to the creation of quality-or-variety-increasing [QV] investments). ⁴⁸ Since in the vast majority of cases the number of buyers who are disadvantaged by competition-lessening conduct who are poor is higher than the number of perpetrator-managers/shareholders/employees who gain from such conduct who are poor, my poverty-focused economic-efficiency claim almost always strengthens the utilitarian case for the Clayton-Act test of illegality. However, the implications of my conclusion that developed economies create an economically-inefficiently-large amount of QV investment for the utilitarian desirability of the Clayton-Act test of illegality is far less certain because, at the same time that increases/decreases in QV-investment competition benefit/harm Clayton-Act-relevant buyers, they decrease/increase economic efficiency by increasing/decreasing equilibrium QV investment in the relevant economy. My concern that the enforcement of the Clayton Act’s “lessening competition” test of illegality will lessen total and average utility in many cases is enhanced by the difference between the impacts that the organizational economic efficiencies a horizontal (M or A) generates will have, respectively, on Clayton-Act-relevant buyers and on economic efficiency.

These differences will be far more problematic from either utilitarian perspective if the Clayton Act is not interpreted to permit defendants whose conduct would otherwise violate it to exonerate themselves by demonstrating that their conduct would not have imposed a net dollar-loss on Clayton-Act-relevant buyers by reducing the most attractive offers they, respectively, received from any worse-than-privately-best-placed potential supplier had the conduct in question not led one or more rivals of the perpetrator(s) to exit by yielding efficiencies that worsened the exiting rivals’ arrays of competitive positions sufficiently to make it profitable for them to exit.

I will now consider the goals that the 2023 Merger Guidelines state U.S. antitrust law is designed to secure. ⁴⁹ I begin with five points about “safeguard[ing]…the Nation’s free market structures.” First, “free markets” not only can coexist with but require substantial government interventions—(1) can be achieved only if property and contract rights are legally recognized and enforced and if various kinds of business organizations are legally recognized and their ownership-structures and internal workings are appropriately regulated and (2) can coexist with legislation, regulations, and judicial rulings that legally obligate participants in external-cost-generating contracts, perpetrators of external-cost-generating property-uses, and perpetrators of acts covered by tort and environmental law to compensate their victims or, indeed, that prohibit such conduct. Second, if “free market structures” are equated with “perfectly-competitive market-structures,” such market structures will almost never be securable without sacrificing substantial economies of scale and/or economically-efficient product differentiation. Third, even if “free markets” could be created without sacrificing economies of scale or economically-efficient product-diversity, in an economy that was otherwise realistically-Pareto-imperfect, free markets would be unlikely to maximize economic efficiency, and freer markets might very well decrease economic efficiency. Fourth, even if free markets would maximize economic efficiency, free markets might not be morally desirable from the perspective of liberalism or from any of the egalitarian perspectives I distinguished because they might result in more people’s not having the resources and opportunities that will give them an acceptable opportunity to take their lives morally seriously and/or might produce a highly unequal income/wealth distribution. Fifth, even if I grant ad arguendo the moral defensibility of the libertarian claim that individuals morally should receive or have a moral right to receive resources whose allocative value equals their respective allocative products and (2) the assumption of some libertarians that markets cannot be “free” if taxes are levied for distributional reasons, the libertarian support for free markets so understood would be undermined by the fact that in realistically-Pareto-imperfect economies gross wages would rarely equal allocative products and gross returns on investment would rarely equal the investment’s economic efficiency.

I turn next to the goals of “preserv[ing] economic freedom,” “our free enterprise system,” and “our democratic political and social institutions.” These concepts can be defined in a number of morally-defensible, critically-different ways that vary with the defensible moral norm that grounds the definition. Thus, whether one concludes that “economic freedom,” “our free enterprise system,” or “our democratic political and social institutions” would be enhanced by particular laws or tests of (M or A) illegality will often be critically affected by whether these desiderata are valued for utilitarian or liberal reasons.

The same point applies to the goals of securing “the best allocation of our economic resources” or “the greatest material progress”: because (1) economic efficiency is not a value, (2) the economic efficiency of a choice is completely irrelevant to its impact on the instantiation of those egalitarian norms that focus exclusively on the equality of the distribution of utility, resources, or various opportunities, (3) even though the impact of a choice on economic efficiency is relevant to its impact on the instantiation of moral norms that value exclusively the total or average utility of those creatures whose utility counts, the choice’s impact on the instantiation of those utilitarian moral norms is not monotonically related to its impact on economic efficiency, and, more generally, (4) no conception of the moral good can be shown to be morally optimal, “the best allocation of our economic resources” needs to be defined relative to a particular, morally-defensible conception of the moral good or a particular conception of justice (when the alternative allocations of resources that might be secured may be deemed just or unjust from the perspective[s] of one or more justice-conceptions).

The 2023 Merger Guidelines also state that one of the goals of U.S. antitrust policy is securing “the highest quality.” This goal may not be favored by norms that value exclusively the equality of utility, resources, and/or relevant opportunities because choices that secure “the highest quality” will often if not usually impose dollar-losses on poor buyers and confer dollar-gains on better-off buyers. Increasing quality may also be undesirable from either utilitarian perspective not only because of the distributive impact of doing so but also because, in my opinion, from the perspective of economic efficiency, developed economies currently devote too many resources to the creation of quality and variety or at least to product R&D that does not generate scientific, technological, or commercial discoveries.

Not surprisingly, the 2023 Merger Guidelines state, respectively, in their first and second relevant paragraphs, that U.S. antitrust policy is designed to secure “the lowest prices” (perfectly-competitive prices?) and to secure “lower prices.” Both these objectives will be desirable from the perspective of those egalitarian norms that value positively and exclusively the equality of the distribution of utility, resources, or relevant opportunities in the vast majority of cases, in which the weighted-average buyer of the relevant good or service is less-well-off than the weighted-average shareholder/manager/employee of the companies on which the price-lowering antitrust intervention imposes losses. In most cases, the distributive impact of lowering product/service prices will also favor the moral attractiveness of lowering prices from either utilitarian perspective not only for pure distributive reasons but also because any associated reduction in poverty or equilibrium QV investment in the relevant portion of product-space will tend to increase economic efficiency when it produces goods or services as opposed to production processes. However, the contrary conclusion may be justified when the directly-affected buyers are wealthier on the weighted average than the directly affected managers/workers/shareholders, when the business in question “produces” PPR, and/or when the economic-efficiency gains generated by the real economic efficiencies the relevant (M or A) yields substantially exceed the benefits those efficiencies confer on buyers by lowering prices.

The 2023 Merger Guidelines also state that one goal of Section 7 of the Clayton Act is to “improve wages and working conditions.” I assume that the Agencies believe that Section 7 will secure these desiderata by preventing (M&A)s that would create resulting companies that would have more power as buyers of labor than the participants would have had as separate entities. I will make three points. First, to the extent that the legally-correct application of Section 7 prevents (M&A)s that would increase the prices their participants and their participants’ rivals charge for the goods and services they produce and/or prevents (M&A)s that would create a resulting company that would be more organizationally-economically-efficient than its participants would have been, it will tend on those accounts to worsen the wages and working conditions of the relevant firms’ employees by lowering those firms’ marginal-revenue-product curves. Second, any improvements in wages and working conditions an (M or A) prevention generates will tend to be associated with an increase in the prices that (M or A) proponents charge because it will be associated with an increase in the marginal-cost curves they face: admittedly, as Part V will reveal, the relevant analysis is complicated, and the relevant effects are mixed. Third, the immediately-preceding conclusion implies that the desirability of achieving this goal from the perspective of the various egalitarian norms I have identified depends on (1) the percentage of any associated wage-cost increases that are passed on to consumers, (2) the ratio of any associated dollar-losses the affected companies sustain to the dollar-gains their employees obtain by virtue of the associated improvement in their terms of employment, (3) the way in which the dollar-losses the companies sustain are divided among the firms’ shareholders/managers/employees and the original weighted-average income/wealth positions of the affected consumers, shareholders, managers, and workers.

The 2023 Merger Guidelines also state that Section 7 is designed to “enhance quality and resiliency.” I do not understand what is meant by resiliency: Does it refer to the durability of products, the likelihood that firms will avoid bankruptcy, the health of workers or the likelihood that they will continue to be employed by the (M or A) participants or their product-rivals? As I have already indicated, increases in product-durability can be either economically efficient or economically inefficient, and I believe that, from the perspective of economic efficiency, too much product-durability (like all quality) may be provided by contemporary economies. I suspect that increases in product-durability may not increase total or average utility and am concerned that increases in product-durability may decrease the equality of the distribution of utility, resources, or relevant opportunities (because the dollar-value of increases in product-durability to poorer buyers is lower than it is to richer buyers).

The 2023 Merger Guidelines declare that one objective of Section 7 is to increase “innovation.” I have already indicated, that in my judgment, from the perspective of economic efficiency, (1) too many resources are devoted to product innovation that does not create pure scientific, technological, or commercial knowledge, (2) too few resources are devoted to PPR, and (3) it is unclear whether too many or too few resources are devoted to knowledge-generating product R&D. I suspect that additional innovations that are secured by decreasing price competition will be associated with decreases in the equality of the distributions of utility, resources, and relevant opportunities but that additional innovations secured by increasing investment-competition will be associated with increases in the equality of the distributions of utility, resources, and relevant opportunities.

The final goal the 2023 Merger Guidelines claim Section 7 seeks to secure is to “expand choices.” I do not understand what that goal entails. As I have already stated, (1) I doubt the economic efficiency or moral desirability from any egalitarian perspective of increasing the variety of products produced by the economy, and (2) I suspect that increases in the range of production-process options available to producers would be economically efficient, would probably increase total and average utility, and could either increase the equality of the distributions of utility, resources, and relevant opportunities or decrease the equality of those distributions, depending on whether the expanded choices were secured by increasing price-competition or by increasing investment-competition. To the extent that the goal of “expanding choices” is understood to be secured by increases in the equality of the income/wealth distribution, I think that this goal would tend to be deemed desirable by utilitarians, would be favored by non-utilitarian egalitarians, and might be favored by liberalism if its achievement reduces the extent to which resource-deprivation disserves the interest of moral-rights bearers in having a meaningful opportunity to take their lives morally seriously. However, as I previously indicated, in some cases, the prohibition of an (M or A) that violates Section 7 will not increase the equality of the income/wealth distribution.

IV. The 2023 Merger Guidelines’ Use of Market-Oriented Approaches to Predicting the Illegality of (M&A)s

On my definition, an approach to predicting the competitive impact of or specific-anticompetitive-intent motivation for an (M or A) is “market-oriented” to the extent that it bases its legally-relevant predictions on the magnitudes of one or more market-aggregated parameters such as the participants’ individual or total pre-(M or A) or predicted-post-(M or A) shares of one or more allegedly-relevant defined markets (whether the shares relate to market-sales or market-investment), to the total pre-(M or A) or predicted-post-(M or A) share(s) of the relevant market(s) possessed by the market’s or markets’ largest four or eight sellers, or (more recently) to the sum of the squares of the pre-(M or A) or predicted-post-(M or A) market shares of all sellers placed in the allegedly-relevant market(s)—to the allegedly-relevant markets’ HHI(s) (Hirschman-Herfindahl Index[es]). The 2023 Merger Guidelines state that the DOJ and FTC intend to base their illegality-assessments substantially on the magnitudes of various market-aggregated parameters. ⁵⁰

Unfortunately, the use of any market-oriented approach to (M or A) competitive-impact or motivation assessment is indefensible because the market definitions that any such approach involves are inevitably arbitrary, not just at the alleged market’s or markets’ peripheries but comprehensively. ⁵¹ I will now summarize the argument for this conclusion both for the ideal-type market definitions that economists exclusively referenced until the 1950s and for the allegedly-functional type of market definitions that industrial-organization (IO) economists increasingly discussed after that. Traditionally, economists assumed that economic markets should and would be defined so that ideally the following three conditions are satisfied: (1) all pairs of products placed inside an economic market are highly competitive with each other in the sense that each (or its producer) is well-placed (equal-best-placed, second-placed, or perhaps close-to-second-placed) to obtain a substantial percentage of the customers that every other product in the market is best-placed or equal-best-placed to supply, (2) each product placed in a given economic market is approximately equally competitive in the above sense with all other products in that market, and (3) each product in an economic market is more competitive with any other product in that market than it is with any product placed in a different economic market. There are seven reasons why one cannot define such classical, ideal-type economic markets non-arbitrarily: (1) one cannot determine non-arbitrarily the volume of sales that must be made of a set of products for that set of products (and its producers and buyers) to constitute a product market; (2) there is no non-arbitrary way to define the concept of “the competitiveness of two products”: does the competitiveness of two products depend on their impacts on each other’s competitive advantages over its respective closest rivals for the patronage of those buyers it is best-placed to supply, on their impacts on each other’s natural-oligopolistic margins, on their impacts on each other’s contrived-oligopolistic margins, on their impacts on each other’s supernormal profits, supernormal profit-rates, etc.; (3) even if the claim just made is false, there would be no non-arbitrary way to define the concept of the overall product-pair competitiveness of all pairs of products placed in an economic market: there is no non-arbitrary way to determine whether one counts as a component of product-pair competitiveness (for example) the fact that product A3 would reduce the competitive advantages, natural-oligopolistic margins, or contrived-oligopolistic margins of product A1 if product A2 did not exist even though product A3 has no such effect given product A2’s existence; (4) even if one could define non-arbitrarily “the competitiveness of any pair of products,” there is no non-arbitrary way to define the equality of the distribution of product-pair-competitiveness scores: does the equality of the distribution of product-competitiveness scores of a set of products depend on the mean derivation of the distribution, on its variance, on some other measure of the dispersion of the distribution of the relevant set of products’ product-pair-competitiveness scores; (5) there is no non-arbitrary way to define the differences between the competitiveness of various in-economic-market product-pairs and the competitiveness of a product placed in that economic market with a product placed outside that economic market; (6) when no economic-market definition is dominant in the sense of fulfilling best the three ideal attributes of a product market, there is no non-arbitrary way to decide on the relative importance of the three considerations; and (7) there is no non-arbitrary way to decide between economic-market definitions that satisfy equally well all tolled the three conditions that an ideal market definition would satisfy.

Starting in the 1950s, Industrial Organization economists abandoned classical economic-market definitions in favor of allegedly-“functional” economic-market definitions (definitions of “antitrust markets”). The economic-market definition that was deemed most functional and therefore “correct” was the definition that could play the most useful role in the most “cost-effective” protocol for resolving antitrust-law issues. The basic objection to such allegedly-functional economic-market definitions is that, by any morally-defensible standard of cost-effectiveness, no protocol for resolving antitrust-law (or antitrust-policy) issues that focuses at all on the magnitude of market-aggregated parameters can be cost-effective: the use of economic-market definitions is inevitably dysfunctional in that the definition of allegedly-relevant economic markets is extremely allocative-costly and the magnitudes of market-aggregated parameters have less predictive power than the magnitudes of the non-market-aggregated parameters that are used to generate the allegedly-functional economic-market definitions. One could list several other related reasons for concluding that allegedly-functional economic-market definitions are inevitably arbitrary: (1) there is no non-arbitrary way to determine whether the quality of the performance of a market-oriented legal-analysis protocol relates to its impact on the correctness of the legal conclusions its use generates (however measured) or to its impact on the extent to which the law as applied achieves its proximate or ultimate goals (whatever they are); (2) there is no non-arbitrary way to measure the incorrectness of an individual legal conclusion—does it depend (A) on the severity of the intellectual error it manifests (however measured), (B) in a Clayton Act case, on the amount by which conduct found illegal benefited Clayton-Act-relevant buyers or, in a Sherman-Act case, on the amount of supernormal profits a defendant found guilty believed ex ante the conduct would generate for legitimate reasons, (C) on whether the error resulted in a guilty defendant’s being found innocent (an illegal [M or A]’s being found lawful) or an innocent defendant’s being found guilty (a legal [M or A]’s being found illegal), (D) on the disservice the error did to the securing of the proximate or ultimate goals of U.S. antitrust law; and (3) there may be no non-arbitrary way to identify the proximate or ultimate goals of U.S. antitrust law.

Inter alia, Parts V, VI, and VII of this Article will justify my conclusion that market-oriented approaches to competitive-impact prediction and (M or A)-motivation assessment are not cost-effective (1) by delineating the non-market-oriented protocol for analyzing the impact that any horizontal (M or A) will have on Clayton-Act-relevant buyers by affecting the intensity of price-competition and explaining why one will not be able to predict this impact accurately or desirably from the magnitudes of market-aggregated parameters, (2) by delineating the non-market-oriented protocol for analyzing the impact that any horizontal (M or A) will have on Clayton-Act-relevant buyers by affecting the intensity of investment-competition and explaining why one will not be able to predict this impact accurately or desirably from the magnitudes of market-aggregated parameters, and (3) by delineating the factors that determine whether an (M or A) was motivated by specific anticompetitive intent or was proposed or executed by a participant that intended to take advantage of what it believed would be the resulting company’s enhanced ability to profit more by engaging in Sherman-Act-violative (contrived-oligopolistic or predatory) conduct than the participants would have done as separate entities.

V. The Impact of Any Horizontal (M or A) on Price-Competition—That Is, On the Dollar-Effect an (M or A) Will Have on Clayton-Act-Relevant Buyers by Affecting the Best Terms They Are Offered for Relevant Goods by Any Supplier that Is Worse-Than-Best-Placed to Supply Them, the Ability of Any Market-Oriented Approach to This Issue to Predict This Impact Accurately or Desirably, and the Approach or Approaches the 2023 Merger Guidelines Indicate the Agencies Will Take to This Issue

I will assume initially that the relevant sellers are setting prices separately to each relevant buyer (are practicing what I denominate “individualized pricing” as opposed to “across-the-board pricing”). Individualized pricing is often practiced by input producers when selling to final-good producers, by final-good producers when selling to wholesalers or retailers, and by final-good sellers when selling expensive products (such as motor vehicles) or unique products (such as antiques). After completing the individualized-pricing analysis, I will address the across-the-board-pricing situation.

My individualized-pricing analysis will assume initially that the relevant (M&A)s will not generate any relevant efficiencies and will not change equilibrium QV investment in the relevant arbitrarily-defined portions of product-space (henceforth, ARDEPPSes). I will start by examining the determinants of the dollar-impact that such a horizontal (M or A) will have on Clayton-Act-relevant buyers by creating a resulting firm whose HNOPs, natural-oligopolistic margins (NOMs), and contrived-oligopolistic margins (COMs) are different from those that the participants would have had as separate entities and then examine the determinants of the dollar-impact that such an (M or A) will have on such buyers by affecting the HNOPs, NOMs, and COMs of the resulting firm’s rivals.

On our current assumptions, with one qualification, such a horizontal (M or A) will create a resulting company whose HNOPs exceed those that the participants would have faced as separate entities by (1) the number of product-units the participants would have been uniquely-equal-best-placed to supply times the average size of the basic competitive advantage (marginal-cost advantage plus buyer preference advantage) the participants would have enjoyed over the respective third-placed suppliers of the buyers in question plus (2) the number of product-units that the participants would have been, respectively, uniquely-best-placed and uniquely-second-placed to supply relevant buyers times the average amount by which in instances in which the participants occupied these positions the second-placed participant would have been better-placed than the relevant buyers’ respective third-placed suppliers. I assume it is obvious that, regardless of how allegedly-relevant markets are defined, one would not be able to infer the magnitudes of these parameters accurately or morally desirably from the participants’ market shares, the total market shares of any defined market’s leading four or eight firms, or the pre-(M or A) or post-(M or A) HHI(s) for the defined market(s). Here is the “qualification” to which I previously referred. The preceding account ignores (1) the fact that an individualized pricer’s (HNOP–MC) difference in its relations with a particular buyer it is best-placed to supply equals not just its conventional-marginal-cost-difference-generated and buyer-preference-generated “basic” competitive advantage (BCA) over its closest rival for that buyer’s patronage but the sum of that BCA and the “contextual marginal costs” (CMCs) its closest rival for that buyer’s patronage would have to incur to match the best-placed firm’s HNOP-containing offer because the closest rival’s matching offer would (almost always) contain a price that was lower than the price(s) that that rival charged buyers the rival was best-placed to supply (was discriminatory) ⁵² and (2) the related fact that any (M or A)-generated increase in the resulting firm’s BCAs would also increase the resulting firm’s HNOPs by increasing its closest rivals’ CMCs by increasing the relevant rivals’ HNOPs by increasing the CMC the resulting firm would have to incur to match the rivals’ HNOP-containing offers to buyers a participant was second-placed to supply by making given prices the resulting firm charged a rival’s customers more discriminatory than those prices would have been for the relevant participant to charge. It should be clear that this complication does not favor the accuracy or moral desirability of market-oriented approaches to predicting the impact of a horizontal (M or A) on the resulting firm’s HNOPs—that the magnitude of this CMC-related effect cannot be predicted accurately or morally desirably from the magnitude of any market-aggregated parameter(s).

I turn now to the dollar-impact that such a horizontal (M or A) will have on Clayton-Act-relevant buyers by affecting the NOMs the resulting firm can secure relative to those the participants would have obtained as separate entities. A seller will be able to obtain an NOM from a particular buyer it is (privately) best-placed to supply if those of the buyer’s inferior suppliers that could profit by beating the best-placed supplier’s NOM-containing offer if their offers would be accepted will be deterred from beating such offers by the combination of their beliefs that (1) making the initially-superior offer is costly, (2) the relevant buyer will give its best-placed supplier an opportunity to beat its inferior supplier’s initially-superior offer, and (3) the buyer’s best-placed supplier will find it possible and inherently profitable to beat its inferior-placed rivals’ initially-superior offer. Admittedly, the (M or A)s on which we are now focusing will not affect the fulfillment of some of these conditions—for example, will not affect (1) the conventional transaction cost to the rival of making the superior offer, (2) whether the buyer believes that its not giving its best-placed supplier an opportunity to rebid will deter its best-placed suppliers from charging it NOM-containing prices in the future, and (3) whether the need for speedy delivery precludes the best-placed supplier from making an effective rebid. However, to the extent that the (M&A)s in question raise the resulting firm’s HNOPs, they (1) will make it more likely that the resulting firm can obtain an NOM by increasing the profits it could have earned by supplying the relevant buyer at the rebid price that would make the best-placed firm’s second offer superior to its rival’s but (2) will make it less likely that the resulting firm can obtain an NOM by reducing the contextual marginal costs its rival would have to incur to beat its initial offer by making the rival’s superior-offer price less discriminatory. In my admittedly-contestable judgment, the first of these effects is bigger than the second: I suspect that the type of horizontal (M&A)s on which I am now focusing will also impose a net dollar-loss on Clayton-Act-relevant buyers by creating a resulting firm that obtains more NOMs than the participants would have obtained as separate entities. Once more, I see no basis for concluding that one will be able to predict this possible effect accurately or morally desirably by following any market-oriented protocol.

I turn now to the effect that a horizontal (M or A) that yields no relevant efficiencies and has no impact on any relevant ARDEPPS’ equilibrium QV investment will have on the COMs that Clayton-Act-relevant buyers will pay the resulting firm relative to those they would have paid the participants. I will assume ad arguendo that any such impact is legally relevant. A partial account of the factors that affect the profitability of a firm’s attempting to obtain COMs would include the following: (1) the cost it would have to incur to communicate its contrived-oligopolistic intentions to its possible undercutters—the number of rivals that could profit by beating offers that contained various COMs (the number of rivals that are second-placed or worse-than-second best-placed to supply its customers by less than the COM it is contemplating including in its price); (2) its ability to make the relevant communication non-verbally simply by charging a contrived-oligopolistic price (because it has a reputation for engaging in contrived-oligopolistic pricing and for not overestimating its HNOP or [HNOP + NOM] figure); (3) the cost the contriver must incur to determine whether it has lost sales to non-cooperators (given that some sales will be lost as a result of buyer-exits, buyer changes in taste, and the entry of new rival product-variants)—a function inter alia of the number of buyers involved, the consistency through time of repeat-sales-to-old-customers percentages, sales-to-former-customers-of-rivals percentages, and sales-to-new-buyers percentages, the availability of data on the parameters that have affected those percentages historically and the magnitudes of those parameters in the most-recent sales-period; (4) the cost to the contriver of identifying its cooperators and non-cooperators—a function of the number of its rivals, its information about their positions, etc.; (5) the ratio of (A) the amount of benefits the contriver can confer on a cooperator by not beating the potential cooperator’s offers to buyers the cooperator is best-placed to supply (roughly speaking, [i] the number of units the particular potential cooperator and the contriver are uniquely-equal-best-placed to supply times the average amount by which they are better-placed than those units’ respective third-placed suppliers plus [ii] the number of units the potential cooperator is uniquely-best-placed to supply and the contriver is uniquely-second-placed to supply times the average amount by which the contriver is better-placed than the respective third-placed suppliers of those units relative to [iii] the profits the contriver is asking the potential undercutter to forego [the number of units that the contriver is best-placed to supply and the relevant potential cooperator is second-placed to supply or is worse-than-best-placed to supply by less than the COM the contriver is attempting to obtain times the average difference in these cases between the COM the contriver is trying to obtain and the amount by which if any the relevant potential contriver is worse-than-second-placed to make the sales in question]); (6) the ratio of (A) the dollar-loss the contriver can inflict on a non-cooperator by making sales to non-cooperator customers that the contriver is not best-placed to make to the cost to the contriver of doing so: if one assumes that the non-cooperator is charging the relevant buyers its HNOPs and that those buyers accept the contriver’s initially-superior offers to them, for each buyer that is “stolen,” the ratio of the non-cooperator’s (HNOP–MC) difference to the amount (if any) by which the contriver is worse-than-second-placed to supply that buyer (a ratio that tends to decline from stolen buyer to stolen buyer if the contriver retaliation is most-cost-effective, a total ratio that declines as the contriver increases the amount of harm it inflicts on the non-cooperator through retaliation) to (B) the harm the contriver must inflict on the non-cooperator through retaliation (a function of the profits the relevant non-cooperator made by beating the contriver’s offers to the contriver’s customers—the number of unit-sales the non-cooperator “stole” from the contriver times the average difference in the cases of those “thefts” between the contriver’s attempted COM and the average amount by which if any the particular non-cooperator was worse-than-second-placed to supply the buyers in question, the number the potential non-cooperator assigns to the probability that if it steals a contriver’s customer the contriver will conclude that a theft has been committed and identify the non-cooperator as the thief, etc.; (7) the safe profits the potential contriver will have to put at risk to try to obtain a COM—its (HNOP–MC) plus NOM figures.

I will make two major points that relate to this list (which is a much-shortened version of the list I have delineated elsewhere ⁵³ ). First, it is far from clear that exemplars of the category of horizontal (M&A)s I am now considering will create a resulting firm that finds contrived-oligopolistic pricing more profitable than the participants would have done as separate entities. Admittedly, such horizontal (M&A)s may tend to make contrived oligopolistic pricing more profitable for the resulting firm in a number of ways: inter alia, by enabling it (1) to take advantage of enterprise-wide economies of scale in establishing a reputation for engaging in strategic conduct or of the stronger reputation of one of the participants for engaging in such conduct, (2) to combine contrived-oligopolistic communications related to both participants’ products, (3) to pool information about particular rivals’ positions and dispositions to not cooperate, (4) to pool information about past and current sales and their determinants that allows it to infer the existence of undercutting by an inferior from current and historical sales-records, (5) to take advantage of any excess reciprocatory power that one participant has in its relations with particular rivals, and (6) to retaliate more cost-effectively against a particular rival by increasing the extent to which one participant’s products are priced in a retaliatory way to a particular rival’s customers and decreasing the extent to which the other participant’s products are priced in a retaliatory way to that rival’s customers. However, the type of (M&A)s on which I am now focusing will also tend to make contrived-oligopolistic pricing less profitable for the resulting firm than it would have been for the participants as separate entities (1) by increasing the attention enforcement authorities pay to the resulting company’s conduct, (2) by creating a resulting company that is more vulnerable to defensive retaliation by non-cooperators who have been punished (by creating a resulting company whose defenses are more spread-out), and most importantly, (3) by increasing the resulting company’s (HNOP–MC + COM) figures (the amount of safe profits the resulting company must put at risk to try to obtain any COM). Moreover, although horizontal (M&A)s may increase the profitability of contrived-oligopolistic pricing by reducing the number of potential non-cooperators when the participants belong to each other’s sets of possible non-cooperators, (M&A)s can also increase the number of potential non-cooperators (when pre-[M or A] the participants were uniquely-equal-best-placed and a substantial number of rivals were equal-third-placed and/or close-to-third-placed or when pre-[M or A] the participants were, respectively, uniquely-best-placed and uniquely-second-placed and a substantial number of participant-rivals were third-placed or close-to-third-placed but far-worse-than-second-placed.

I admit that I do not have appropriate data on the magnitudes of the relevant parameters. However, if I had to guess, I would say that the kind of horizontal (M&A)s on which I am now focusing would tend to create a resulting company that obtained fewer and lower COMs than the participants would have obtained as separate entities.

The second major point I want to make at this juncture should by now be familiar. It should be clear that no market-oriented protocol can predict the impact of the category of horizontal (M&A)s now under consideration on the resulting firms’ COMs accurately or morally acceptably.

I turn now to the dollar-impact that the category of horizontal (M&A)s on which I am now focusing will have on Clayton-Act-relevant buyers by affecting the prices they pay for goods they purchase from the resulting firm’s rivals. As I have already indicated, horizontal (M&A)s in this category will tend to raise these prices by increasing the HNOPs of the resulting firm’s rivals. Thus, when a participant would have been a resulting firm’s rival’s closest competitor, such an (M or A) will increase the rival’s HNOP by increasing the contextual marginal costs the resulting firm will have to incur to match any offer its rival makes by increasing the resulting firm’s prices to its own customers (by increasing the resulting firm’s HNOPs and NOMs even if it does not increase its COMs [in my judgment, by more than it decreases the resulting firm’s COMs]). If the (M or A) would not affect the CMCs of any non-resulting-firm rival of a resulting-firm rival, the increase in the resulting-firm rival’s HNOP in those cases in which a participant would have been its closest competitor will equal the lower of the increase in the resulting firm’s CMCs and the amount by which the participant would have been better-placed than the relevant buyer’s third-placed supplier. Of course, because the (M&A)s in the category now under consideration would also tend to raise the prices that the non-resulting-firm rivals of each resulting-firm rival charge their own respective customers, such (M&A)s will also tend to increase each resulting-firm rival’s HNOPs to the extent that it increases the HNOPs (and NOMs and COMs—see below) of each such rival’s non-resulting-firm rival by increasing the resulting firm’s CMCs when the non-resulting-firm rival of a resulting-firm rival would otherwise have been the latter firm’s closest competitor by raising that rival’s HNOPs, NOMs, and COMs and thereby the CMCs it must incur to match given offers to buyers they are not best-placed to supply.

For reasons that have already been explained, any (M&A)-generated increase in the HNOPs (and [HNOP–MC] figures) of the resulting firm’s rivals will tend to increase the incidence and magnitudes of the rivals’ NOMs. The category of (M&A)s on which I am now focusing may also tend to increase the resulting firm’s rivals’ COMs, though such (M&A)s’ tendencies to increase these firms’ (HNOP–MC)s and NOMs will have the opposite effect. More particularly, the (M&A)s now under consideration will tend to increase the resulting firm’s rivals’ COMs (1) by increasing the dollar-benefits each can provide the resulting firm by reciprocating to its cooperation by increasing the frequency with which each is the resulting firm’s closest competitor and the average amount by which in those instances in which it is the resulting firm’s closest competitor it is better-placed than the relevant buyers’ third-placed suppliers, (2) by enabling each rival to take advantage of any excess reciprocity power it had in relation to one (M or A)-participant, and (3) by enabling each rival to lower the cost of inflicting through retaliation any given dollar-loss on the resulting firm relative to the cost of inflicting that loss on the participants combined (A) by increasing the frequency with which the rival was the resulting firm’s closest competitor or any given relevant amount worse-than-second-placed to supply the resulting firm’s customers, (B) by increasing the resulting firm’s ([HNOP + NOM]–MC) figures in relation to buyers the rival was second-placed or close-to-second-placed to supply, and (C) by enabling the rival to increase the loss-inflicted to cost-incurred ratio for any amount of loss inflicted by increasing the dollar-loss it imposes on the resulting firm by increasing the loss it imposes on the resulting firm by charging retaliatory prices to one participant’s customers and decreasing the loss it imposes on the resulting firm by charging retaliatory prices to the other participant’s customers.

I will now relax the assumption that the sellers affected by the category of (M&A)s whose effects are being analyzed are setting individualized prices. I do not have space to do more than point out five differences between the individualized-pricing analysis and the across-the-board-pricing analysis. ⁵⁴ The first difference relates to the definition of the relevant “HNOP.” When the prices being set are across-the-board prices, one must define not only the across-the-board HNOP of each relevant seller but the array of HNOPs for a relevant (arbitrarily-defined) group of sellers. That array contains the prices those sellers would charge if they announced their prices in a specified order (presumably in the order in which they did announce their prices) if each seller announced the price it would find most profitable given the prices other relevant sellers had already announced if those prices would not be changed and each seller assumed that its rivals’ responses to its price would not be affected by any belief that it could react to their responses. The second difference is that the (HNOP–MC) gap for any member of a relevant set of across-the-board pricing competitors will depend not only on its but on its product-rivals’ arrays not only of basic competitive advantages when competing for the patronage of the different potential buyers of their respective products (as well as on the order in which the sellers in question announce and lock themselves into prices). The third difference, which is connected to the second, is that the component of an across-the-board pricer’s (HNOP–MC) gap that is the counterpart to the contextual marginal costs an individualized pricer’s closest competitor for the patronage of an individual buyer the former firm was best-placed to supply would have to incur to charge that buyer a price that would result in the worse-placed rival’s matching the best-placed supplier’s HNOP-containing offer to that buyer is the extra margin the across-the-board pricer will find profitable to charge because the price that each of its rivals, respectively, charge its customers are also the prices each of them charges its own customers (because its rivals charge its customers supra-MC prices because they find it profitable to charge their own customers supra-MC prices not only because its rivals have BCAs in their relations with their own customers but also because each of its rivals knows that that rival’s rivals will be charging supra-MC prices to their own customers [for the same reasons]). The fourth difference is that, because (I believe) the component of any across-the-board pricer’s (HNOP–MC) gap the third difference references will usually be much higher than the closest-rivals’ contextual-marginal-cost-related component of an individualized pricer’s (HNOP–MC) gap, the (HNOP–MC) differences for across-the-board pricers will substantially exceed the average BCAs they, respectively, enjoy in their relations with those buyers they are, respectively, best-placed to supply. ⁵⁵ The fifth difference is that the average (HNOP–MC) gap for any set of across-the-board-pricing rivals will depend inter alia on the order in which the across-the-board pricers announce their prices (and make it costly for themselves to alter their originally-announced prices by advertising their original announced prices in media that reach final consumers, by printing their original announced prices on their products’ packaging, by posting their original announced prices on the shelves of the distributive outlets that sell them to final consumers, etc.): roughly speaking, the average (HNOP–MC) difference for a group of across-the-board pricers will be higher if those sellers that are second-placed or close-to-second-placed more often than they are best-placed and are best-placed by small amounts when they are best-placed announce their prices and lock themselves into their announced prices early in the price-announcement sequence.

The preceding discussion implies that the dollar-loss that horizontal (M&A)s that do not generate any relevant efficiencies or change equilibrium investment in the ARDEPPSes in which they are consummated will impose on Clayton-Act-relevant buyers by raising HNOPs when the participants and their mutual rivals charge across-the-board prices will increase with the extent to which those (M&A)s increase the BCAs of the resulting firm above those of the participants by freeing the participants from each other’s price-competition and the extent to which the resulting firm is better placed than the participants would have been to persuade relevant rivals to announce their prices in the order that raises the HNOP-array of the relevant sellers (especially, to prevent the firms that the FTC and DOJ characterize as “mavericks” ⁵⁶ from reducing the relevant array of HNOPs by delaying their price-announcements). Once more, there is no reason to believe that one will be able to predict accurately or morally desirably the extent to which any (M or A) that might be consummated by across-the-board pricers will increase their and their mutual rivals’ HNOPs from market-aggregated data.

I turn now to any differences the fact that a horizontal (M or A)’s participants and their product-rivals set across-the-board as opposed to individuated prices will have on the correct way to analyze the (M&A)s’ impacts on NOMs. I can think of three possible sets of differences. First, since the probability that a seller will be able to obtain an OM naturally increases with its (HNOP–MC) difference, any differences in the determinants of an (M or A)’s impact on individualized and across-the-board (HNOP–MC) gaps will cause differences in the appropriate impact-on-NOM analysis. Second, because a buyer’s rejection of an NOM-containing offer from its best-placed supplier that is better than any offer it received from any worse-than-best-placed potential supplier is more likely to be public when across-the-board prices are being charged, buyers may be more likely to reject their best-placed suppliers’ NOM-containing offers when across-the-board prices are being charged. Third, because any price-reduction an across-the-board pricer offers when some buyers reject its initial NOM-containing offer will presumably be across-the-board, the analysis of the cost and profitability of any firm’s improving its initial NOM-containing offer when some rivals respond to that offer non-cooperatively is different from and more complex than the counterpart analysis in individualized-pricing situations. To be honest, I do not know whether, all tolled, the horizontal (M&A)s on which I am now focusing will be more likely to increase or to decrease the NOMs obtained by the participants and their mutual rivals when all of them are charging across-the-board as opposed to individualized prices. However, I do know that one will not be able to predict this possible impact of the horizontal (M&A)s on which I am now focusing accurately or morally desirably from market-aggregated data.

The determinants of the profitability of contrived-oligopolistic pricing to a firm and hence of the impact that the category of (M&A)s on which I am now focusing will have on the profitability of such pricing to the participants and their mutual rivals will also be somewhat different when the relevant firms are charging across-the-board prices as opposed to individualized prices. I will limit myself here to pointing out four differences. First, because across-the-board prices will almost always be public whereas individualized prices will usually be secret, (M&A)s will not affect the profitability of contrived-oligopolistic pricing to the participants or their mutual rivals by changing the number of potential undercutters (by increasing/decreasing the cost of determining whether there has been non-cooperation by increasing/decreasing the number of rivals who could profit by undercutting the contriver’s contrived-oligopolistic offers [whose conduct must be investigated]). Second, if (as I assume) both undercutting and retaliation will be across-the-board when the prices that are set initially are across-the-board prices, the analysis of the cost to one or more contrivers of inflicting various amounts of harm on undercutters will be different when across-the-board pricing is being practiced (inter alia, will have to include analyses of the ability of contrivers that have been undercut to coordinate their retaliatory efforts as well as analyses of the likelihood that a cooperator will misperceive a rival’s retaliation against a non-cooperator to be non-cooperative conduct). I do not have space to analyze the determinants of the impact that the category of (M&A)s on which I am now focusing will have on the COMs that its participants and their mutual rivals would find profitable to attempt to contrive. However, it should be obvious that this impact cannot be predicted accurately or morally acceptably from data on the magnitudes of market-aggregated parameters.

I will now relax the assumption that the (M&A)s under consideration will not generate any efficiencies that would affect Clayton-Act-relevant buyers even if they did not affect equilibrium investment in the relevant ARDEPPS. In particular, I will focus on the impact on Clayton-Act-relevant buyers of any marginal-cost-curve reductions a horizontal (M or A) generates that relate to the resulting firm’s use of the participants’ investments. I suspect that by this point you will not be surprised to learn that, even if one controls for the size of the (M or A)-generated marginal-cost-curve reduction, the analysis of the relevant impact is complicated.

I focus first on horizontal (M&A)s whose participants are charging individualized prices. I will focus separately on the impact that the static marginal-cost efficiencies a horizontal (M or A) will have on five different sets of buyers that are charged individualized prices. First, the efficiencies will have no effect on those buyers that both participants and the resulting firm are much-worse-than-second-placed to supply. The second set of buyers contains those buyers that the participants were considerably-worse-than-second-placed to supply but the resulting firm was either equal-second-placed to supply or close-to-second-placed to supply. In this situation, the efficiencies will not affect either the buyers’ best-placed suppliers’ HNOPs or the NOMs the buyers are charged but may reduce the COMs they are charged by increasing by one the number of sellers that could beat a relevant COM-containing offer the buyers’ best-placed supplier would otherwise make. The third set of buyers contains those buyers that the participants were either worse-than-second-placed to supply or uniquely or non-uniquely second-placed to supply but the resulting firm was uniquely-second-placed to supply. In such cases, the efficiencies the (M or A) generated will confer a dollar-gain on each buyer in this set equal to the amount by which they reduced the HNOP of their respective best-placed suppliers (the associated reduction in the resulting firm’s marginal costs minus any amount by which the relevant participant was worse-than-second-placed) plus any NOM the efficiencies deterred the buyers’ best-placed suppliers from obtaining by reducing their (HNOP–MC) differences. The fourth set of buyers consists of those buyers that neither participant was best-placed to supply but the resulting firm was either best-placed to supply or equal-best-placed to supply. The associated benefit to each of those buyers equals the competitive advantage that the firm that was uniquely-best-placed to supply it pre-(M or A) enjoyed pre-(M or A) plus any increase in the NOM the buyer had to pay because the resulting firm’s (HNOP–MC) difference was greater than the (HNOP–MC) difference of the resulting-firm rival that was the buyer’s uniquely-best-placed supplier pre-(M or A) or minus any decrease in the NOM the buyer had to pay because the resulting firm’s (HNOP–MC) difference was greater than its pre-(M or A) counterpart for the rival of the resulting firm that was the buyer’s best-placed supplier pre-(M or A) plus any increase or minus any decrease in the COM the buyer has to pay that the efficiency generated (for example, by changing the number of sellers that were second-placed or close-to-second-placed to supply the buyer and/or by increasing or decreasing the safe profits the buyer’s changing best-placed supplier [or uniquely-equal-best-placed suppliers] had to put at risk to try to obtain a COM). The fifth set of buyers contains those that one participant was best-placed to supply or that both participants were equal-best-placed to supply but the resulting firm was best-placed to supply by an amount that exceeded the competitive advantage of the relevant participant(s) by the marginal-cost efficiency that the (M or A) generated. In this case, at least if one assumes that the efficiencies do not change the number of units that the relevant buyers purchase, the efficiencies will tend to impose a dollar-loss on Clayton-Act-relevant buyers by making it more likely that they will be charged NOMs but will tend to confer a dollar-gain on those buyers by making it less likely that they will be charged COMs by increasing the safe profits the possibly-changing best-placed suppliers will have to put at risk to try to obtain a COM (though a full analysis of this COM-impact would have to consider the effect of the efficiencies on the ability of the resulting firm to reciprocate to cooperators and the retaliation-generated dollar-loss-imposed to dollar-loss-incurred ratio for the resulting firm for different relevant amounts of dollar-losses imposed).

The impact that the static marginal-cost efficiencies that a horizontal (M or A) generates will have on Clayton-Act-relevant buyers therefore depends not only on the magnitude of those efficiencies (which I assume no proponent of market-oriented competitive-impact-prediction protocols believes can be predicted from the magnitudes of market-aggregated parameters) but on a host of other factors that cannot be predicted from the magnitudes of market-aggregated parameters (such as the percentages of relevant buyers that fall into each of the five categories I have identified).

I will now relax the assumption that I have been making so far that the horizontal (M or A)s whose impacts on price-competition are at issue will not affect equilibrium investment in the relevant ARDEPPS. That assumption will almost never be accurate: (1) to the extent that a horizontal (M or A) increases/decreases the (P−MC) figures for the set of products in the relevant ARDEPPS at its pre-(M or A) equilibrium investment-quantity, it will tend to increase/decrease equilibrium investment in that ARDEPPS by increasing/decreasing the profitability of all existing and all possible additional investments in that ARDEPPS; (2) as Part V of this Article will explain, individual horizontal (M&A)s can also alter equilibrium investment in the ARDEPPS in which they are consummated by increasing or decreasing the intensity of investment-competition in the ARDEPPS in which they are consummated (which I will measure in one sense arbitrarily but usefully in this context by the difference between the quantity of investment in the relevant ARDEPPS at which its most-supernormally-profitable investments would yield just a normal rate-of-return and the equilibrium quantity of investment in the ARDEPPS in question); and (3) there is no reason to believe that the just-referenced change-in-price-competition effect on equilibrium investment and change-in-investment-competition effect on equilibrium investment will cancel each other out. To the extent that a horizontal (M or A) increases/decreases equilibrium investment in the ARDEPPS of consummation, it will tend on this account (1) to (confer dollar-gains)/(impose dollar-losses) on Clayton-Act-relevant buyers by lowering/increasing the HNOPs of the products that were supplied in the ARDEPPS pre-(M or A) and increasing/decreasing product-variety or average speed of supply through time in the relevant ARDEPPS, relatedly (2) to (confer dollar-gains)/(impose dollar-losses) on Clayton-Act-relevant buyers by lowering/increasing the NOMs those buyers pay, and (3) will have an uncertain impact on Clayton-Act-relevant buyers by virtue of its effect on the COMs they pay in that any decrease/increase in the relevant sellers’ ([HNOP−MC]+NOM} figures will increase/decrease their incentives to practice contrived-oligopolistic pricing by decreasing/increasing the safe profits they must put at risk to do so while any increase/decrease in equilibrium investment in the relevant ARDEPPS will tend to decrease/increase the relevant sellers’ contrived-oligopolistic pricing to the extent that (the decrease is generated by the exit of an established firm)/(the increase is generated by a new entry as opposed to by an investment by an established firm) since ceteris paribus the profitability of contrived-oligopolistic pricing in any ARDEPPS is inversely related to the number of sellers operating in it. Although for reasons of space I will not explain this assertion here, I believe that these conclusions apply regardless of whether individualized prices or across-the-board prices are charged in the ARDEPPS in question. In any event, for reasons that I have explained in this Part and for reasons that I will explain in Part V, the effect that a horizontal (M or A) would have on Clayton-Act-relevant buyers by virtue of its impact on equilibrium investment in the ARDEPPS of consummation cannot be predicted accurately or morally acceptably by any market-oriented analytic protocol.

So far, I have delineated the factors that determine the impact that horizontal (M&A)s will have on Clayton-Act-relevant buyers by affecting the prices they are charged (and any other terms in the relevant sales-contracts—a complication that I will henceforth ignore) for the products produced in the ARDEPPS of consummation and explained why this impact cannot be predicted accurately or morally acceptably by any market-oriented protocol. I will now delineate and assess the components or features of the approach or approaches the 2023 Merger Guidelines state the Agencies will take to predicting this impact of any (M or A).

First, the 2023 Merger Guidelines state that the Agencies will use a market-oriented protocol to predict the impact that any (M or A) will have on Clayton-Act-relevant buyers by altering the prices they are charged. ⁵⁷ I have already explained why no such approach can be acceptably accurate or morally or legally defensible. I therefore take some comfort in the fact that the 2023 Merger Guidelines state that “[m]arket concentration and change in concentration due to a merger are often [emphasis added] useful indicators of a merger’s risk of substantially lessening competition” ⁵⁸ —that is, in the fact that these Guidelines do not state that the Agencies believe that these parameters are always useful indicators of the Clayton-Act illegality of an (M or A). I also take some (but only limited) comfort from the following paragraph in the 2023 Merger Guidelines:

Relevant markets need not have precise metes and bounds. Some substitutes may be closer and others more distant and defining markets necessarily requires including some substitutes and excluding others. Defining a relevant market sometimes requires a line-drawing exercise around product features, such as size, quality, distances, customer segment or prices. There can be many places to draw that line and properly define a relevant market. The Agencies recognize that such scenarios are common, and indeed “fuzziness” would seem inherent in any attempt to delineate the relevant … market. Market participants may use the term “market” colloquially to refer to a broader or different set of products than those that would be needed to constitute a valid relevant antitrust market. ⁵⁹

Here! Here! My only disagreement is with the noun “fuzziness.” The correct modifier is “inevitable arbitrariness not just at the peripheries of the market definition but comprehensively.” I take only limited comfort from this paragraph because it manifests the Agencies’ stubborn refusal to bite the bullet—to admit that market definitions are comprehensively arbitrary and that any analytic protocol that ignores this reality cannot be defended.

Second, the 2023 Merger Guidelines also state that “the method by which the Agencies often define relevant antitrust markets” is the “Hypothetical Monopolist Test” (HMT). ⁶⁰ According to this test, the market in which any product produced by an (M or A) participant should be placed contains that product and the smallest set of other products whose placement under the control of one firm (“the hypothetical monopolist”) would render it profitable or most profitable (the Agencies have not been clear on this point) for that firm to “undertake at least a small but significant non-transitory increase in price (SSNIP) or other worsening of terms (SSNIPT) for at least one product in the group.” ⁶¹ I have elsewhere ⁶² explained in great detail why the SSNIP/”hypothetical monopolist” approach to market definition is not acceptable (why it is ill-specified and why no improvement in its specification would create an approach to market definition that would render acceptably accurate any market-oriented approach to competitive-impact prediction that incorporated an SSNIP protocol for defining allegedly-relevant markets).

Third, at one point, ⁶³ the 2023 Merger Guidelines do specify a wide range of parameters “observed market characteristics (practical indicia)”—whose identification the Agencies claim will contribute to the correct determination of the relevant market—in particular, “industry or public recognition of the submarket as a separate economic entity, the product’s peculiar characteristics and uses, unique production facilities, distinct customers, distinct prices, sensitivity to price changes, and specialized vendors.” The 2023 Merger Guidelines indicate that the market-definition protocol that focuses on these factors is different from the “hypothetical monopolist” approach—that in contrast to the approach that focuses on these factors “the hypothetical monopolist test” approach is “[a]nother common method employed by the courts and the agencies ….” ⁶⁴ In part, I object to this non-hypothetical-monopolist-oriented approach because I am not sure what is meant by “sensitivity to price changes” (is the reference to changes in the unit sales of one product that are generated by changes in the price of another product?); in part, I object to this approach because I suspect that products’ having different characteristics and being produced in distinctly-different production facilities have less bearing on their competitiveness than the Agencies seem to suppose; in part, I object to this approach because I doubt the relevance of “public recognition of … [a] submarket as a separate economic entity” (a doubt that the Agencies “share” in relation to market participants’ speaking colloquially about “the market” in which they are allegedly participating ⁶⁵ ); but, most importantly, I object to this approach to market definition because, even if the parameters on which it focuses were relevant to the prediction of “the impact of an (M or A) on price competition” in the Clayton-Act sense, it would be clear that those parameters should be taken into account directly rather than being used to define an allegedly-relevant market whose definition would allow market-aggregated parameters to be defined on whose magnitudes the relevant competitive-impact prediction would then be based: even if the parameters on which this alternative approach to market definition were relevant to the impact of a horizontal (M or A) on price competition, it would be both cheaper and more accurate to generate the relevant prediction from them directly rather than from the magnitudes of market-aggregated parameters associated with markets whose definition was based on estimates of these non-market aggregated parameters. This last objection is also my primary ground for opposing the Agencies’ use of evidence on the other parameters the 2023 Merger Guidelines state they will use to define allegedly-relevant markets—“[c]ustomers’ willingness to switch between different firms’ products,” ⁶⁶ the extent to which “customers consider the firm’s products and the rivals’ products to be closer substitutes,” ⁶⁷ “the fraction of the unit sales lost by … [an identified] first product due to changes in terms, such as an increase in its price, that would be diverted to “[an identified] second product”—the so called “diversion ratio.” ⁶⁸ I would be remiss not to acknowledge that the Guidelines’ Section 4.3.1. on Market Definition ⁶⁹ does contain a number of statements that warrant the conclusion that, as actually applied, the Agencies’ market definitions and market-oriented approach to competitive-impact prediction will be less inaccurate and morally/legally unacceptable than they would otherwise be: statements that the Agencies will take account of the facts that (1) some buyers may have to be taken into consideration even though they “travelled from outside the boundaries of the [defined] geographic market” “to mak[e] … purchases” ⁷⁰ and that (2) some buyers may have a preference for purchasing clusters of products (that are not competitive with each other) from the same supplier, ⁷¹ and (3) the statement that the Agencies will also “consider [ ] [to be] market participants” “[f]irms that are not currently active in a relevant market, but that would rapidly enter with direct competitive impact in the event of a small but significant change in competitive candidates, without incurring significant sunk costs.” ⁷²

Fourth, the Agencies use the Hirschman-Herfindahl Index to measure seller concentration and claim that rebuttable legal presumptions that (1) an (M or A) that increases by more than 100 points the HHI of a defined market whose pre-(M or A) HHI was greater than 1800 or (2) that increases the defined market’s HHI by more than 100 points and creates a resulting firm whose market share is over thirty percent are illegal are “highly administratable and useful tool[s] for identifying mergers that may substantially lessen competition.” ⁷³ For reasons I have already explained, I reject both the claim of administrability and the claim of usefulness. I should add, however, that the fact that, like the 2010 Merger Guidelines and unlike the 1992 Merger Guidelines, the 2023 Merger Guidelines make all HHI-oriented lessening-competition presumptions rebuttable constitutes a step in the right direction: the step would be larger if the relevant Guidelines were more forthcoming about the evidence that would rebut the HHI-oriented presumptions the Agencies claim are adminstratable and useful.

Fifth, my concerns about the Agencies’ understanding of the evidence that would rebut its presumptions is enhanced by the 2023 Merger Guidelines’ continuance of the Agencies’ historic mistake of claiming that “[t]he change in HHI from a merger of firms with shares of a and b is equal to 2 ab.” ⁷⁴ Regardless of whether market shares are calculated by reference to unit-sales, sales-revenues, productive-capacity, or QV investment more comprehensively, horizontal (M&A)s whose participants’ market shares are a and b (1) will virtually always for a host of reasons that the Agencies’ relevant claim ignores create a resulting firm whose market share is not (a + b) and affect the market shares of the resulting firm’s rivals and therefore (2) will be extremely unlikely to increase the allegedly-relevant market’s HHI by 2ab.

Sixth, I will comment on the positions that the 2023 Merger Guidelines take on the impact that an (M or A) that does not generate any relevant efficiency will have on Clayton-Act-relevant buyers by affecting their potential and actual suppliers’ HNOPs. I will make five points or sets of points. First, like their predecessors, the 2023 Guidelines analyze this impact (which they denominate “the unilateral effects” ⁷⁵ of an [(M or A)]) separately from the impact that (M&A)s have on Clayton-Act-relevant buyers by altering their suppliers’ OMs (by altering what they denominate seller “coordination”). Second, although the 2023 Merger Guidelines recognize that the magnitude of the unilateral effects an (M or A) will have on its participants’ prices will depend on how competitive their products are with each other (which they sometimes misdescribe as the extent to which the participants’ products are substitutes ⁷⁶ ), the 2023 Guidelines never provide the correct operationalization of the relevant concept (say that [1] “[f]irms are close competitors the more that customers are willing to switch between their products,” ⁷⁷ [2] the competitiveness of one firm with a rival increases with the extent to which the “firm’s competitive action results in greater lost sales for the rival … and [with] … the profitability of the rival’s lost sales ….” ⁷⁸ and [3] “[a] measure of customer substitution between firms [and of the strength of the competition between them] … is the diversion ratio”—“the fraction of unit sales lost by the first product due to a change in [its] terms, such as an increase in its price that would be diverted to the second product” ⁷⁹ ). Third, the 2023 Merger Guidelines ignore the fact that one component of an individualized pricer’s (HNOP−MC) gap in its relations with a particular buyer is the contextual marginal costs its closest rival for that buyer’s patronage would have to incur to charge that buyer a price that would result in the rival’s offer matching the best-placed supplier’s HNOP-containing offer (because the relevant rival-price is discriminatory or violates a minimum-price regulation) and that one component of an across-the-board pricer’s (HNOP−MC) gap is the extra margin it can charge because its rivals find it profitable for non-oligopolistic reasons to charge their customers and hence its customers supra-MC prices. Fourth and relatedly, the 2023 Merger Guidelines ignore the various ways in which an (M or A) can affect the contextual-marginal-cost components of any individualized pricer’s individualized HNOPs and the counterpart components of an across-the-board pricer’s across-the-board HNOP (both when the [M or A] in question generates no efficiencies and because of the efficiencies it generates). Fifth, the 2023 Merger Guidelines ignore the fact that even horizontal (M&A)s that generate no efficiencies can alter the individualized and across-the-board HNOPs of the participants’ rivals by creating a resulting firm whose individualized and across-the-board HNOPs and NOMs are higher than their participants’ HNOPs and NOMs would have been and thereby raising (1) the CMCs an individualized-pricing resulting firm would have to incur to match relevant offers its rivals might make when a participant would have been that rival’s closest competitor and by raising the CMCs the rival’s non-resulting-firm rivals are charging their customers by changing the resulting firm’s relevant CMCs and (2) the across-the-board HNOPs of its across-the-board-pricing rivals (A) by raising the across-the-board prices the resulting firm charges by changing its BCA/BCD array and/or (B) by creating a resulting firm that is better-placed to and does induce its rivals to announce their prices and lock themselves into their announced prices in an order that raises the HNOP array in the relevant ARDEPPS.

The seventh set of comments I will make on the 2023 Guidelines’ positions on the impacts that (M&A)s can have on Clayton-Act-relevant buyers by altering the prices they pay (and the other conditions of sale they accept) relate to any position the Agencies may have on the impact that (M&A)s that generate no relevant efficiencies will have on NOMs. I wrote “may have” because it is not clear that the Agencies recognize the phenomenon of “natural-oligopolistic conduct” or, relatedly, NOMs: it may be that their references to “tacit coordination” are references to “natural-oligopolistic conduct” as I have defined this concept, but it also may be that (like some scholars) they use the expression “tacit coordination” to refer to what I would call “contrived-oligopolistic conduct initiated by non-verbal communications.” Even if the Agencies recognize the phenomenon of natural-oligopolistic pricing, the Guidelines convey no information about the determinants of the feasibility of the practice or the ways in which (M&A)s that do not generate any relevant efficiencies can affect its feasibility and incidence (for example, can increase its incidence by increasing best-placed individualized pricers’ [HNOP−MC] gaps) both when the best-placed pricer is the firm that the (M or A) creates and when it is a rival of the resulting firm.

The eighth set of points I will make at this juncture relates to the Guidelines’ positions on (M&A)s’ possible impacts on what I call “contrived-oligopolistic pricing” and the Agencies denominate “coordinated (pricing) conduct.” I will make eight points or subsets of related points that relate to this issue.

First, although it is true that “[t]he fewer the numbers of competitively-meaningful rivals,” the more likely it will be that contrived-oligopolistic pricing will be profitable because the fewer the number of possible non-cooperators the fewer the communications a contriver will have to make and the less expensive it will be to identify the non-cooperator, it does not follow (as the Guidelines claim) that “[t]he fewer the number of competitively-meaningful rivals prior to the merger, the greater the likelihood that merging two competitors will facilitate coordination” ⁸⁰ because, as I explained, a horizontal (M or A)’s impact on the number of potential undercutters is not correlated with the number of firms that were second-placed or close-to-second-placed to supply a buyer a participant was best-placed to supply pre-(M or A): thus, if the participants were, respectively, uniquely-best-placed and uniquely-second-placed to supply a relevant buyer pre-merger and, pre-(M or A), a large number of resulting-firm rivals were third-placed or close-to-third-placed pre-(M or A), the (M or A) would increase the number of competitively-meaningful rivals the best-placed firm faced even if it faced only one such rival pre-(M or A). Second, for related reasons as well as several other reasons connected to the comprehensive arbitrariness of market definitions, the Guidelines claim that “[m]arkets that are highly concentrated after a merger that significantly increases concentration … are presumptively susceptible to coordination” ⁸¹ cannot bear scrutiny. Third, the Guidelines are correct in stating that what they term “coordination” is more likely “if a firm’s behavior can be promptly and easily observed by its rivals.” ⁸² Fourth, the Guidelines may or may not be correct in asserting that coordination is less likely if the relevant products are homogenous and “customers find it relatively easy to switch between suppliers”: ⁸³ this Guidelines assertion is supported by the fact that product homogeneity disfavors the profitability of contrived-oligopolistic pricing by increasing the number of possible non-cooperators (the number of sellers that are equal-best-placed or second-placed or close-to-second-placed when one seller is uniquely best-placed) given the number of sellers placed in the relevant ARDEPPS; however, this Guidelines position is undercut by the fact that product homogeneity favors the profitability of contrived-oligopolistic pricing by reducing the safe profits that a possible contriver must put at risk to practice contrived-oligopolistic pricing. Fifth, the Guidelines’ statement that “the likelihood of strong [uncooperative?] or rapid responses by rivals” to any attempt to secure a COM “increase[s]” when “the market has few significant competitors” ⁸⁴ is inaccurate in that it fails to take account of the fact that, regardless of the protocol used to define allegedly-relevant markets, not all product-pairs in any defined market will be highly competitive. Sixth, if the Guidelines are implicitly defining “a maverick” to be a firm that is “disruptive” inter alia in the sense that it does not cooperate with contrived-oligopolistic pricing, they are incorrect in asserting that an (M or A) that “eliminates a maverick” will “reduce the risk of [what the Guidelines call] coordination.” ⁸⁵ However, seventh, the Guidelines’ statement that “a firm with a small market share may have less incentive to coordinate because it has more to gain from winning new business than other firms” ⁸⁶ (may be more likely to be a maverick?) does not get it right: as I indicated earlier, the factors that affect the profitability to a firm of not cooperating are (1) the ratio of the number of units a firm is best-placed to supply to the number of units it is second-placed to supply or close-to-second-placed to supply and (2) (A) the average (HNOP−MC) figure for the firm if it sets individualized prices and (B) the firm’s (HNOP−MC) figure if the firm sets across-the-board prices. Eighth, the Guidelines do not recognize that horizontal (M&A)s can affect the COMs not only of the resulting firm relative to those that the participants would have secured as separate entities but also the COMs of the resulting firm’s rivals. I suspect that this omission is serious because my “best guess” is that horizontal (M&A)s will tend to increase the COMs of the resulting firm’s rivals both when the (M or A) generates no relevant efficiencies and especially when it does generate such efficiencies (though I suspect that horizontal [M or A]s will not tend to increase the resulting firm’s COMs above those the participants would have obtained because such [M or A]s will increase the resulting firm’s [HNOP−MC + NOM] figures and in so doing will deter its attempting to obtain COMs by increasing the safe profits it must put at risk to do so).

The ninth set of points I will make at this juncture relates to the 2023 Merger Guidelines’ positions on the possibility that an (M or A) might impose dollar-losses on Clayton-Act-relevant buyers in the forbidden way by raising the prices they are charged or worsening other terms they are offered by entrenching an (M or A) participant’s “dominant position.” ⁸⁷ I will make five associated points. First, this DOJ/FTC concern may be attributable to the Agencies’ awareness of the fact that E.U. competition-law prohibits “[a]ny abuse by one or more undertakings of a dominant position ….” ⁸⁸ Second, as I have explained elsewhere, ⁸⁹ there is no non-arbitrary way to define a firm’s dominance. One cannot do so non-arbitrarily by defining a firm to be “dominant if its share of the market in which it is said to be operating exceeds some number because there is no non-arbitrary way to define the market in which the relevant firm is allegedly operating.” And one cannot define a firm’s dominance non-arbitrarily in any way that does not focus on its share of an allegedly-relevant market because there is no non-arbitrary way to determine whether or the extent to which a firm’s dominance is a function of various non-market-share parameters—for example, is a function of its (1) total (HNOP−MC)s, NOMs, COMs, and/or (P−MC)s on the sales it made, (2) its average (HNOP−MC)s, NOMs, COMs, and/or (P−MC)s in its relations with those buyers it is best-placed to supply in individualized-pricing contexts or their across-the-board counterparts, (3) its supernormal profits, (4) its supernormal profit-rate, or (5) some combination of two or more of these parameters. Third, any market-share metric for a firm’s dominance is problematic not only because, as I have already indicated, the market definition on which any such metric must rely is inherently, comprehensively arbitrary but also because (as I have demonstrated in detail elsewhere ⁹⁰ ), regardless of the protocol one uses to define economic markets, there is no significant correlation between a firm’s market share and any of the firm-outcome measures previously listed. ⁹¹ Fourth, of course, if an (M or A) is deemed to have increased a dominant firm’s dominance (or to have made dominant a firm that would not have been dominant but for the [M or A]—a possibility that E.U. law has not considered) in one or more ways that impose dollar-losses on Clayton-Act-relevant buyers by reducing the absolute attractiveness of the best offers they, respectively, receive from any potential supplier that is not privately-best-placed to supply them, that fact will count against the relevant (M or A)’s Clayton-Act legality. However, fifth, the fact that an (M or A) imposed dollar-losses on Clayton-Act-relevant buyers in this way would count against the (M or A)’s Clayton-Act legality regardless of whether on this account it was deemed to entrench an already-dominant firm’s dominant position or to render “dominant” a firm that would not otherwise have been deemed “dominant.”

The tenth set of points I will make relates to the 2023 Merger Guidelines’ positions on the effects that (M&A)s can have on the prices that Clayton-Act-relevant buyers must pay by affecting potential competition and new entry. I will discuss these issues more thoroughly in Part VI. At this juncture, I will make only two points: (1) the Guidelines correctly state that new entry (including presumably any new entry an [M or A] causes to be executed) will “lower prices” and “increase [unit] output,” ⁹² and (2) the Guidelines suggest that the Agencies believe that perceived potential competition will tend to cause established firms to lower their prices ⁹³ so that a conglomerate (M or A) that eliminates an effective potential competition will tend on this account to lessen competition in the Clayton-Act sense.

The eleventh set of points I am making about the 2023 Merger Guidelines relates to their statements about the way in which the Agencies will assess the impact on price-competition of any efficiencies an (M or A) generates. ⁹⁴ (Part VI will analyze the relevance of some efficiencies an [M or A] can generate for its impact on investment-competition and the Guidelines’ treatment or non-treatment of this issue.) I will make six points. First, the Guidelines place the burden on the (M or A) participants of demonstrating that their (M or A) would generate efficiencies that would benefit Clayton-Act-relevant buyers. I think that this placement of the duty of coming forward with such evidence on the (M or A) participants is rendered legally correct by their better access to the relevant information. ⁹⁵ Second, the Guidelines do not distinguish between the static efficiencies an (M or A) can generate (which relate to the economic efficiency with which the resulting company uses or renews the participants’ investments) and the dynamic efficiencies an (M or A) can generate (which relate to the [M or A]’s creating a resulting firm that can make more-economically-efficient new QV investments or PPR investments than the participants could have made as separate entities). Third, the Guidelines do not distinguish between the static (marginal/variable)-cost efficiencies an (M or A) can generate (which can affect both Clayton-Act-relevant buyers and economic efficiency) and the static fixed-cost efficiencies an (M or A) can generate (which will affect Clayton-Act-relevant buyers only if they affect equilibrium QV investment in the relevant ARDEPPS by causing the resulting firm to renew a QV investment that its participant-owner would not have renewed in circumstances in which the non-renewal of that investment would not lead someone else to make a QV investment that would have the same effect on Clayton-Act-relevant buyers as the renewal of the non-renewed investment would have had). Fourth, the Guidelines contain no analysis of the factors that determine the impact that a given marginal-cost-curve reduction generated by the real efficiencies an (M or A) yields will have on Clayton-Act-relevant buyers. Fifth, the Guidelines state the Agencies’ correct position that even if the “real” efficiencies (see below) an (M or A) might be said (see below) to generate would benefit Clayton-Act-relevant buyers that fact will not favor the (M or A)’s Clayton-Act legality if the participants could have generated the same efficiencies without consummating the (M or A) under review. ⁹⁶ Sixth, the Guidelines indicate that the Agencies do distinguish between what I previously described as the “real efficiencies” an (M or A) can generate and the “cost savings” an (M or A) can generate by creating a resulting firm that has more “monopsony power” than the participants would have had as separate entities. ⁹⁷ However, the footnote that draws this distinction is puzzling in that it states that “[t]he Agencies will not credit efficiencies” that an (M or A) generates by creating a resulting firm with more monopsony power than the participants would have had ⁹⁸ –that is, will not count in favor of the (M or A)’s legality any related dollar-gain the (M or A) yields Clayton-Act-relevant buyers on this account—whereas elsewhere the Guidelines state the Agencies’ position that any “harm” to sellers an (M or A) generates by reducing “competition among buyers” will render the (M or A) illegal under the Clayton Act even if the (M or A) does not impose a net dollar-loss on Clayton-Act-relevant consumers in the forbidden way (indeed, even if it benefits such consumers) because Section 7 “prohibits mergers that may substantially lessen competition or tend to create a monopoly in any line of commerce and in any section of the country ….” ⁹⁹

The twelfth and thirteenth components of the 2023 Merger Guidelines I will discuss articulate positions of the Agencies that are clearly legally correct. The twelfth attributes to the Agencies a position that the Guidelines suggest but do not explicitly state the Agencies have taken: the position that, when analyzing the competitive impact of an (M or A), the Agencies will take account of future changes that may or will occur that are not caused by the (M or A) but will affect its competitive impact. ¹⁰⁰ The position just articulated is clearly correct as a matter of economics and law.

The thirteenth set of comments I will make focuses on “Guideline 8: When a Merger is Part of a Series of Multiple Acquisitions, the Agencies May Examine the Whole Series.” ¹⁰¹ This Guideline is economically justified and might be legally warranted even if its following would not increase the accuracy of the Agencies’ relevant competitive-impact predictions. To see why it will be more accurate to predict the competitive impact of a series of (M&A)s by analyzing the competitive impact of the series at one time rather than by analyzing the competitive impacts of each member of the series as if it were the only (M or A) that would be consummated and summing the competitive-impact conclusions for each (M or A) in the series on this assumption, note, for example, that (1) the series taken as a whole will be predicted to increase the resulting firm’s HNOPs when none of the separate (M or A)s in the series would be predicted to do so if (A) in relation to some buyers, three or more participants in the series were uniquely-equal-best-placed or (B) in relation to some buyers, one participant was uniquely-best-placed and two or more participants were equal-second-placed, (2) the series taken as a whole will increase the resulting firm’s HNOPs by more than they would be deemed to be increased if one summed the relevant effects of each member of the (M or A) series if, in relation to some buyers, (A) one participant was uniquely-best-placed and two or more participants were uniquely-equal-second-placed, (B) one participant was uniquely-best-placed, two or more participants were uniquely-equal-second-placed, and two or more participants were worse-than-second-placed by a smaller amount than was an alternative supplier, and (3) the series taken as whole will be predicted to increase the resulting firm’s COMs when an analysis that purports to predict this effect by predicting the impact that each member of the series would have on those COMs if it were the only (M or A) consummated and summing those predicted effects if none of the individual (M or A)s would affect the COMs taken in isolation (would critically affect the profitability of the resulting firm’s practicing contrived-oligopolistic pricing or the magnitudes of the COMs it would find ex ante profitable to try to obtain) but the cumulative effect of all members of the series would critically affect these “things.” To see why, even if, contrary to fact, the accuracy of the relevant predictions would not be improved by analyzing the competitive impact of the series taken as whole, focusing on the impact of the series taken as a whole might be legally warranted, note that, accuracy aside, there might be cost-related economies of scale in focusing on the series in one go and the Agencies have discretion to make those enforcement-decisions that are in the public interest—“prosecutorial” discretion that is ubiquitous in the United States whose existence in the instant context is supported (as I have already observed) by the facts that the text of Section 7 includes the words “may be” and “substantially.”

The fourteenth set of points that I want to make in this Part relates to the position that the Agencies take in the 2023 Merger Guidelines on what might be called “the compared to what issue”: When measuring the competitive impact of an (M or A), what outcome-baseline does one use—what alternative course of action do the Agencies assume the participants would take if they did not consummate the (M or A) at issue when determining the competitive impact (relative to something) of that (M or A). I have already indicated that the Guidelines state that the Agencies will credit an (M or A) with generating efficiencies only if the participants would not have generated the efficiencies the (M or A) might be said to generate in some other way had they not consummated the (M or A). This issue also arises in another specific context: when analyzing the competitive impact of an (M or A) executed by a “failing firm,” the Agencies state that they will conclude that any (M or A) a failing company proposes will be deemed not to lessen competition only if it “pose[s] a less severe danger to competition” than an alternative available (M or A) the failing company could have identified had it made “good-faith efforts” to identify “relevant” alternative (M or A)s (that presumably would be more profitable for the failing firm than consummating no [M or A] at all?). ¹⁰² I have two related concerns: (1) a concern that the legally-correct alternative choice whose consequences should be used as a benchmark for competitive-impact-prediction is “a choice to do nothing” and (2) an economic concern that the imposition of such a good-faith-effort duty may prove to be counter-productive—may cause the failing firm not to engage in any (M or A) by increasing the search costs it must incur to do so when the (M or A) it would otherwise have consummated (usually with a competitor) would have benefited Clayton-Act relevant buyers relative to the state of the world that its not consummating any (M or A) would yield. I hasten to add that this “compared to what issue” must be addressed generally—in relation to all choices the Clayton Act covers and all effects generated by the choices that the law’s addressees make. I am genuinely surprised that no academic, lawyer, or judge who has focused on U.S. antitrust law has ever formulated or addressed this general issue. My own view, which derives from the combination of my conclusions that (1) the U.S. is Constitutionally committed to instantiating the liberal conception of justice and (2) liberalism does not imply that relevant actors have a general positive duty to benefit (as opposed to a negative duty to avoid imposing losses on) those who are affected by the choices the actors make, is that the legally-correct baseline is the “do-nothing baseline.”

The fifteenth and final set of points I will make at this juncture relates to the 2023 Merger Guidelines’ statements about the “Analytic, Economic, and Evidentiary Tools” ¹⁰³ The Guidelines indicate that the Agencies will make use of. The relevant evidence is said to include information from the (M or A) participants, evidence about “how customers have shifted purchases in the past in response to relative changes in price or other terms,” information from the customers and workers of the participants, from the workers’ labor organizations, information from the participants’ rivals, information from the participants’ indirect customers, information from the participants’ distributors, information from consultants and industry analysts. The Guidelines also indicate that the Agencies will make use of evidence on post-(M or A) price-increases when consummated (M&A)s are at issue. Finally, the Guidelines state that the Agencies will make use of “economic models” and “merger simulations.” I have four comments. First, it is obviously good that the Agencies will make use of all these sources of information. Second, although the relevant sections of the Guidelines do not specify the information that these various sources will provide, I expect that much if not almost all of the information these sources will provide will be on the magnitudes of non-market-aggregated parameters. Third, this reality reinforces my puzzlement about the Agencies’ stubborn insistence that they are using a market-oriented approach to predicting the impact of (M&A)s on price-competition. Fourth, although I fully support the use of economic models and “merger simulations” (as well as efforts to derive predictions about the impact of a proposed [M or A] on price-competition from evidence about the impact that past [M&A]s in the relevant ARDEPPS have had on price-competition), (1) one cannot use these approaches cost-effectively or morally defensibly unless one has and employs a sound theoretical understanding of the determinants of the relevant effects and the ways in which those determinants interact to cause those effects and (2) the Guidelines do not make me optimistic about the Agencies’ relevant understandings.

This Part has ignored (1) the Guidelines’ discussions of the factors that influence the intensity of price-competition by influencing the intensity of investment-competition, ¹⁰⁴ (2) the Guidelines’ failure to address explicitly how the Agencies will analyze the Sherman-Act-violativeness of (M&A)s and their restricting their accounts of how the Agencies will address issues that are relevant to this question (A) to claims that vertical (M&A)s may sometimes lessen competition even if the resulting companies do not and did not intend to engage in Sherman-Act-violative refusals to deal and (B) to claims that the vertically-integrated firms may lessen competition by refusing to supply less-than-fully-vertically-integrated rivals, ¹⁰⁵ and (3) the Guidelines’ discussions of these two vertical-(M&A)-related issues. These issues will be addressed, respectively, in some detail in Parts VI, VII, and VIII of this Article.

VI. The Agencies’ Approach to Analyzing the Impact of (M&A)s on Investment-Competition (and on Product Variety and Quality)

Traditionally, IO economists, Antitrust Law law professors, the U.S. antitrust-law-enforcement Agencies, and U.S. courts focused exclusively on the impact of conduct on price-competition. It is not clear why they ignored the impact that business conduct or antitrust policy has on the intensity of investment-competition (however defined) and on product variety and quality. They may have done so (1) because, for some inexplicable reason, they were unaware of the fact or chose to ignore the fact that firms compete not only by lowering their prices toward their marginal costs but also by introducing new products, opening up additional distributive outlets, and increasing their average speed of supply throughout a fluctuating-demand cycle by increasing their capacity and inventory (or by doing additional production-process research if they are in the business of selling the right to use particular production processes) ¹⁰⁶ or (2) because they thought that any exemplar of business conduct or any antitrust policy would affect the intensity of price-competition and the intensity of investment-competition in the same direction and that any increase/decrease in price-competition would have the same effect on economic efficiency, total or average utility, or anything else that might be valued that the same increase/decrease in investment-competition (however the two are measured) would have on these desiderata (or, more modestly, that qualitative changes in the intensities of the two types of competition would have the same qualitative effects on each of these desiderata). In fact, none of these beliefs is true: (1) not only do the determinants of the intensity of price-competition and hence of the impact on price-competition of given exemplars of business conduct or antitrust policies differ from their investment-competition counterparts, but there is every reason to believe that given exemplars of business conduct or antitrust policies can increase price-competition while decreasing investment-competition or can decrease price-competition while increasing investment-competition (however the intensities of these two types of competition are measured), (2) not only will given changes in the intensity of price-competition have different effects on economic efficiency than “the same change” in the intensity of investment-competition will have, but in many circumstances given increases in price-competition will increase economic efficiency all tolled while the same increase in investment-competition will decrease economic efficiency all tolled, and (3) the distributive impacts of given increases in price-competition and investment-competition will not only be different but in some instances will differ in ways that critically affect their desirability from the perspective of various morally-defensible distributive norms.

In any event, credit where credit is due (even when it is partial). Not only the 2023 Merger Guidelines ¹⁰⁷ but also the 2010 ¹⁰⁸ and 1992 Guidelines ¹⁰⁹ recognize (1) that new entry can occur into a relevant portion of product-space and (2) that mergers can effect “Product Variety,” though they do not distinguish between the effects an (M or A) can have on product variety and innovation by changing the intensity of price-competition at the pre-(M or A) investment-quantity and the effects an (M or A) can have on product variety and innovation by changing “the intensity of investment-competition” (however measured).

I will now provide brief accounts of the determinants of the intensity of investment-competition in any ARDEPPS, on the various ways in which an (M or A) can affect that intensity, and on the 2023 Merger Guidelines’ discussions of these issues. I start by providing an arbitrary definition of the intensity of investment-competition in any ARDEPPS: on this definition, the intensity of investment-competition in any ARDEPPS is a decreasing function of (is inversely related to) the difference between (1) the quantity of investment that would have to be present in that ARDEPPS for the lifetime rate-of-return yielded by its most-supernormally-profitable investments to be just normal and (2) the equilibrium investment-quantity in that ARDEPPS. Inter alia, this definition manifests the fact that over all relevant total-QV-investment ranges the lifetime supernormal rate-of-return that will be generated by any QV investment in any ARDEPPS will be inversely related to equilibrium QV investment in that ARDEPPS: as total ARDEPPS QV investment rises, the supernormal rates-of-return generated by each QV investment in it declines because prices decline and the unit-sales of each (product produced in it)/(distributive outlet operated in it) declines. My excuse for this definition is that no legal (or policy) conclusion I reach will be affected by it.

I will next define two ARDEPPS-investment quantities: (1) “the entry-preventing investment-quantity for a given ARDEPPS at any point in time” is the smallest quantity of investment in that ARDEPPS at that time that would deter entry in that ARDEPPS (if all potential competitors were sovereign maximizers), and (2) “the entry-barred, investment-expansion-preventing investment-quantity in a given ARDEPPS at a given point in time” is the smallest quantity of investment in that ARDEPPS at that point in time that would deter all firms established in that ARDEPPS from adding to the ARDEPPS’ investment-total if all such established firms were sovereign maximizers and “entry were barred” (if no entry could take place, regardless of whether entry actually was barred and regardless of whether entry would occur if no established firm added to the ARDEPPS’ total investment).

The next concept or set of related concepts I need to define and particularize is “the barriers to entry” faced by a particular potential competitor. The various “barriers to entry” faced by a particular potential competitor at a given point in time (at a particular ARDEPPS pre-existing QV-investment quantity) refer to various sources of the difference between the post-entry lifetime (hereinafter this modifier will be omitted) supernormal rate-of-return the potential entrant would anticipate realizing on the most-supernormally-profitable QV investment it could make at that point in time and the pre-entry supernormal profit-rate that was generated by the most-supernormally-profitable QV investment(s) in the defined ARDEPPS. Four categories of barriers to entry are distinguished ¹¹⁰ : (1) the profit-rate-differential barrier to entry (Π_D), which equals the difference between the weighted-average-expected (gross-of-risk-costs) rate-of-return that the ARDEPPS’ most-supernormally-profitable QV investments should be expected to generate at the ARDEPPS’ pre-entry QV-investment quantity and its counterpart for the relevant potential competitor’s most-supernormally-profitable possible QV-investment; (2) the risk barrier to entry (R), the difference between the normal rate-of-return for the investment that would be most-supernormally-profitable for the relevant potential entrant if it did not have to be concerned with retaliation to its investment (see below) and its counterpart for the owners of the most-supernormally-profitable QV investments in the relevant ARDEPPS ¹¹¹ ; (3) the retaliation barrier to entry (L), the difference between the retaliation-against-investment-related certainty-equivalent (weighted-average-expected plus risk) costs the relevant potential entrant faces and its counterpart for the owner(s) of the ARDEPPS’ most-supernormally-profitable QV investments on those investments; and (4) the scale barrier to entry (S), the amount by which the entry in question would reduce the supernormal profit-rate generated by all (I am simplifying) QV investments in the ARDEPPS in question. (For simplicity, the text has assumed and will continue to assume that at any ARDEPPS investment-quantity all members of the set of “most-supernormally-profitable investments in any ARDEPPS will have the same weighted-average expected rate-of-return and the same normal rate-of-return.”)

I turn now to concepts that relate to the possible difference between the supernormal rate-of-return that the owners of the most-supernormally-profitable QV investments in an ARDEPPS would realize on those investments if no additional QV investment were made in the ARDEPPS in question and the highest supernormal rate-of-return that a particular firm established in the ARDEPPS in question could realize on any QV investment it could add to the ARDEPPS’ QV-investment total. Seven such concepts are relevant. The first four are counterparts to the Π_D, R, L, and S barriers to entry that a potential competitor may face: established firms may face Π_D, R, L, and/or S barriers to making a (QV-investment) expansion in an ARDEPPS in which they are already operating. The fifth concept is the monopolistic investment-incentive (a negative figure [M] expressed as a percent of the private cost of the investment) an established firm may have to add an investment to its ARDEPPS’ total. An established firm will have a monopolistic investment-incentive to make an investment that would expand its total QV investment in an ARDEPPS in which it is already operating to the extent that the relevant investment will increase the profits its other investments in the relevant ARDEPPS yield not by increasing its reputation for quality or by expanding its product-line when some buyers value owning multiple products in a given line (for aesthetic reasons or because the method of use or performance-attributes of the products in different product-lines differ but are the same for products in the same product-line) but because the investment in question will be less competitive with the investor’s other investments in the ARDEPPS than the rival investment the investment in question would deter would have been and/or because the investment in question will elicit responses from established rivals in the ARDEPPS that are less damaging to the investor’s pre-existing-ARDEPPS-investments’ profit-yields than the responses that would have been elicited by a rival investment the investment in question would deter would have been (regardless of whether the deterred investment would have been made by an established rival or new entrant but particularly when it would have been made by a new entrant). The sixth concept is the monopolistic investment-disincentive (a positive figure [M], expressed in the same way as the negative M that symbolizes the investor’s monopolistic investment-incentive). An established firm will have a monopolistic investment-disincentive to make an investment that would expand its total QV investment in an ARDEPPS in which it already has investments to the extent that the relevant investment will reduce the profits yielded by its other investments in the relevant ARDEPPS by competing with them and inducing responses from one or more of its established rivals in that ARDEPPS by more than those profits would otherwise have been reduced directly and indirectly (in those ways) by the rival investment if any that the investment in question would deter. In almost all cases, an established firm will have a monopolistic investment-disincentive to make a particular QV investment in an ARDEPPS in which it already has such investments when the investment in question would not deter an established-rival QV-investment expansion or a new entry. The seventh and final concept is the natural-oligopolistic investment-disincentive (O—a positive figure) that two or more established firms face on the QV-investment expansions they are, respectively, contemplating. Established firms will face O disincentives in situations in which each member of a set of two or more established firms (say “two” for simplicity) is in a situation in which no established rival will make a QV investment if it does not though an established rival will invest if the potential initial investor does because (assuming that each such firm’s investment once made is not reversible) each will find that although it would find it profitable to invest if its investment would not induce an established rival to invest (perhaps despite the M disincentives it would have to do so) the supernormal profits its investment would yield under those circumstances are lower than the losses the rival investment its investment would elicit would impose on it. In any event, the sum of the barriers to expansion, monopolistic investment-incentives, or monopolistic or natural-oligopolistic investment-disincentives that would face the individual established firms that were successively privately-best-placed to add successive QV investments to a particular ARDEPPS entry aside (where one firm may be privately-best-placed to make more than one of the QV investments that are successively privately-best in the absence of entry) will presumably increase as one moves from first, second, … nth QV investment that could be made by established firms in the ARDEPPS in question. The entry-barred, expansion-preventing QV-investment quantity is the smallest quantity of ARDEPPS QV investment that would render it subnormally profitable for any established firm to add a QV investment to the ARDEPPS in question if entry were “artificially barred” (or unprofitable for any potential competitor) where the entry-barred clause is relevant because it affects the existence and/or magnitudes of relevant established firms’ monopolistic investment-incentives/disincentives or natural-oligopolistic investment-disincentives.

I will now discuss the implications of the preceding analysis for the effectiveness of potential competition or entry. First, although the preceding analysis did not address this issue, that analysis assumes or at least I believe that it can be shown that if potential competition is effective, its effectiveness will be manifest either in new entry’s resulting or in its causing one or more established firms to make “limit” investments they would not otherwise have made to prevent entry: I do not believe that effective potential competition affects outcomes by inducing established firms to charge lower, so-called “limit prices” to deter entry. ¹¹² Second, although the preceding analysis implies both (1) that there will be situations in which no potential competitor is effective (namely, when the relevant ARDEPPS’ entry-barred, expansion-preventing QV-investment quantity exceeds the entry-preventing QV-investment quantity [when the (Π_D + R + S + L + [M or O]) sum for the best-placed potential investment-expander in the ARDEPPS at its entry-barred expansion-preventing QV-investment quantity is lower than the (Π_D + R + S + L) sum for the best-placed potential competitor at the ARDEPPS’ entry-preventing QV-investment quantity]) and (2) that there will be situations in which a substantial number of potential competitors (perhaps more than 3 ¹¹³ ) will be effective (when the ARDEPPS’ entry-barred, expansion-preventing QV-investment quantity is substantially below the ARDEPPS’ entry-preventing QV-investment quantity (if I assume for simplicity that all the relevant investments are the same size) is more than three QV investments below the entry-preventing QV-investment quantity and there are many equal-best-placed potential competitors or the sum of the barriers facing the second-through-nth-placed potential competitors are only slightly higher than those facing the best-placed potential competitor where n is quite high. I should add that the number of limit QV investments that an ARDEPPS’ established firms will find profitable to make and the number of new entries that will be executed when the entry-barred, expansion-preventing QV investment number is a given number below the entry-preventing QV-investment number will depend on the rate at which the successively-worse-placed potential competitors’ (Π_D + R + L + S) sums increases and the rate of increase of the entry-unbarred (Π_D + R + L + S + [M or O]) figures for the established firms that would be successively best-placed to add QV investments to the ARDEPPS’ total. The presence of m potential competitors that would enter if no established firm made a limit investment might make one, two, … m limit investments profitable for the successive limit investors that made them.

The 2023 Merger Guidelines say virtually nothing about the conditions that determine the effectiveness of potential competition. They state only that “[t]he Agencies’ starting point for assessment of a [firm’s] reasonable probability of entry is objective evidence regarding the firm’s available feasible means of entry, including its capabilities and incentives.” ¹¹⁴ I do not understand the meaning of “means of entry.” At best, the word “capabilities” might be interpreted to refer to what I call the magnitude of the Π_D and R barriers that a potential entrant faces, and the word “incentives” might be interpreted to refer to those barriers and to the supernormal rate-of-return that would be yielded by the last (and least-supernormally-profitable?) QV investment that would have been made in the ARDEPPS if the ARDEPPS’ QV-investment quantities equaled the ARDEPPS’ entry-barred, expansion-preventing QV-investment quantity or the quantity of QV investment the ARDEPPS would have after its established firms made all the limit investments they would make: I am not the only position-taker who or that is flailing. Obviously, all these interpretations are generous stretches. Unlike its predecessors, the 2023 Merger Guidelines indicate that the Agencies believe that “[a] perceived potential entrant” (regardless of whether the perception of effectiveness is accurate?) “can prompt current market participants to … expand output [and] … lower product prices” ¹¹⁵ —a regrettable, new endorsement of limit-pricing “theory.” I hasten to add that the 2023 Merger Guidelines do correctly state that actual “[n]ew entry can yield a variety of procompetitive effects, including increased output or investment, higher wages or improved working conditions, greater innovation, higher quality, and lower prices.” ¹¹⁶ However, even if I could ignore the inevitable arbitrariness of “relevant market” definitions, the statement of the Guidelines that at least suggests that the competitive benefits that an entry will generate will be higher if the “relevant market” is “highly concentrated” ¹¹⁷ clearly would require justification and in my judgment would be doubtful.

I turn now to my analysis of the ways in which horizontal, conglomerate, and vertical (M&A)s can affect the intensity of investment-competition ¹¹⁸ and to the relevant positions that the 2023 Merger Guidelines take on these issues. I will start by making six sets of related points about the ways in which a horizontal (M or A) can affect the intensity of investment-competition. The first set relates to the dynamic efficiencies any (M or A) can generate—that is, to efficiencies that reduce the (Π_D + R) barriers to expansion the resulting firm faces below those its participants would have faced as independent entities. ¹¹⁹ The dynamic efficiencies an (M or A) generates will increase investment-competition when they cause the resulting firm to make a QV investment in the relevant ARDEPPS when neither participant nor any of their mutual rivals would have done so. The dynamic efficiencies an (M or A) generates will not cause it to increase investment-competition when they cause the resulting firm to make a more-profitable but no-larger QV investment than a participant would have made or when they cause the resulting firm to make a QV investment that deters a rival from making an equally-large QV investment. (In fact, even if the dynamic efficiencies result in the substitution of a more-profitable for a less-profitable QV investment, there would be no reason to believe that it would on this account confer a net dollar-gain on Clayton-Act-relevant buyers: I see no connection between the supernormal-profit-rate yielded by a QV investment and either the buyer surplus the sale of the product or service it creates generates or the investment’s impact on the buyer surplus that purchasers of rival products realize. Indeed, even if I ignore for the moment any connection between the dynamic efficiencies an (M or A) generates and the L barrier that various relevant actors face, the fact that an (M or A) would create a resulting firm that faced lower (Π_D + R) barriers to making a new QV investment might actually cause it to decrease investment-competition—could substitute for a situation in which a rival of the resulting firm would otherwise have invested inter alia because its investment would not induce a participant to invest a situation in which the resulting firm and its relevant rival confronted each other with critical natural-oligopolistic QV-investment disincentives. ¹²⁰

The second set of investment-competition-impact-related points I want to make contains two points or subsets of points that relate to monopolistic QV-investment disincentives. First, a horizontal (M or A) can reduce QV-investment competition by converting a situation in which one or possibly both participants would have made a QV investment when none of their rivals would have done so because the monopolistic QV-investment disincentives the participants faced were smaller into one in which the resulting firm does not invest because it faces larger monopolistic QV-investment disincentives than either participant would have faced in that the resulting firm’s profits are affected by the impact of any new investment it makes on the profit-yields of both participants’ pre-existing investments. Second, although a horizontal (M or A) may create a resulting firm that finds it profitable to make a QV investment that deters a mutual-rival QV investment when neither participant would have found it profitable to do so because the resulting firm has a larger monopolistic QV-investment incentive to do so in that the difference between the amount by which the resulting firm’s investment would reduce its pre-existing investments’ profit-yields and the larger amount by which the deterred investment would have done so is larger than the difference between the amount by which either participant’s investment would have reduced that participant’s pre-existing investments’ profit-yields and the amount by which the relevant participant’s rival’s deterred investment would have done so, the (M or A) in question will not increase investment-competition as I have defined it on this account and will have no tendency to yield Clayton-Act-relevant buyers a dollar-gain on this account when it has this effect.

The third set of investment-competition-impact-related points contains two points that relate to natural-oligopolistic QV-investment disincentives. First, a horizontal (M or A) can increase investment-competition by substituting for a situation in which (say) the two participants are the only firms that might find it profitable to add a QV investment to the relevant ARDEPPS but are deterred from doing so by the natural-oligopolistic QV-investment disincentives they cause each other to face into a situation in which the resulting firm does invest because the monopolistic QV-investment disincentives it faces are critically lower than the natural-oligopolistic QV-investment disincentives both participants faced (because, unlike the two participants taken together, the resulting firm can choose to make one and only one new investment). Second, it is conceivable that a horizontal (M or A) can lessen investment-competition by converting a situation (say) in which a participant and a resulting-firm rival confronted each other with natural-oligopolistic investment-disincentives that were not critical into a situation in which the natural-oligopolistic investment-disincentives the resulting firm faced were critical because they were larger (in that the rival’s investment would reduce the profit-yields of both participants’ investments).

The fourth set of points I will make at this juncture relates to the possible impacts that an (M or A) can have on investment-competition by affecting the L barriers faced by the resulting firm relative to the L barriers faced by the participants, the L barriers faced by those of the resulting firm’s rivals that are established in the relevant ARDEPPS, and the L barriers faced by potential entrants into that ARDEPPS. My previous discussion of the retaliation involved in contrived-oligopolistic pricing should have revealed how complicated a full analysis would be. I will restrict myself to the points I find most salient. First, an (M or A) could tend to increase the L barrier faced by the resulting firm relative to the L barriers faced by the participants (1) by increasing the resulting firm’s (HNOP + NOM + COM−MC) figures when it was best-placed above their counterparts for the separate participants when they were best-placed (that is, by increasing on this account the ratio of the price-retaliation-generated dollar-loss inflicted on the resulting firm to the dollar-cost-of-price-retaliation to the retaliating rival for given amounts of dollar-losses inflicted on the resulting firm for one or more resulting-firm’s individual rivals) and (2) by enabling the resulting firm’s individual rivals to reduce the cost it/they would have to incur to inflict a relevant dollar-loss on the resulting firm by engaging in price-retaliation below the cost it/they would have to incur to inflict the same total loss on the separate participants that would be equally effective at securing their separate cooperation by inflicting higher dollar-losses on one participant-division of the resulting company and lower dollar-losses on the other participant-division of the resulting company than they would have to inflict on the separate participants to achieve the same result (that is, by creating a resulting company whose defenses were more “spread-out”). Second, the (M or A) would tend to decrease the L barrier faced by the resulting firm relative to the L barriers faced by the participants as separate entities to the extent that rivals are deterred from retaliating against the resulting firm by their realization that the resulting firm would be disposed to take advantage of organization-wide economies of scale in building a reputation for not succumbing to threats or acts of retaliation (as well as for engaging in strategic conduct). Obviously, the first two L impacts will affect the (M or A)’s impact on investment-competition only if the increase in L deterred the resulting company from making a QV investment that would deter no rival QV investment that would have been as large or the decrease in L would cause the resulting company to make a QV investment that would deter no rival QV investment that would have been as large. Third, an (M or A) can also decrease investment-competition by increasing the L barrier facing a rival of the resulting company and thereby deterring it from making a QV investment whose execution would not have deterred the execution of an equally-large QV investment by another actor. An (M or A) can increase the L barrier faced by a rival of the resulting firm (1) by creating a resulting firm for which the dollar-loss-inflicted to dollar-cost-incurred ratio for price-retaliation against particular rivals that have invested or are considering investing is higher (by increasing the rivals’ [HNOP + NOM + COM−MC] figures, increasing the frequency with which the resulting firm is second-placed or close-to-second-placed to supply a buyer the relevant rival is best-placed to supply, and by enabling the resulting firm to do more retaliating via the pricing of one participant and less retaliating via the pricing of the other participant) and (2) by creating a resulting firm that can take better advantage of company-wide economies of scale in building and maintaining a reputation for engaging in strategic behavior. Admittedly, an (M or A) might also increase investment-competition by reducing the L barriers the resulting firm’s rivals face (1) by increasing the law-related cost the resulting firm must incur to engage in retaliation by causing the enforcement-agencies to pay more attention to the resulting company’s conduct and disposing triers-of-fact to find it guilty of illegal retaliation or making (illegal) retaliation-threats and (2) by making it more likely that a rival that was subjected to retaliation would engage in defensive retaliation (for the same reasons that [M&A]s can increase the L barrier that resulting firms face). I hasten to add that the rivals whose L barrier an (M or A) might affect could either be established rivals of the resulting firm or potential entrants to the ARDEPPS in question.

The fifth set of points I will make about the possible investment-competition impacts of (M&A)s could have been folded into the discussion of (M&A)s’ possible impacts on the (Π_D + R) or L barriers faced by the resulting firm’s rivals. It is sometimes argued (indeed the 2023 Merger Guidelines discuss the possibility ¹²¹ ) that “Mergers Can Violate the Law When They Create a Firm that May Limit Access to Products or Services That Its Rivals Use to Compete.” The (M&A)s in question are vertical (M&A)s and the concern appears to be that such (M&A)s may lessen not only price-competition but investment-competition by creating vertically-integrated firms that, in violation of the Sherman Act, will predatorily refuse to deal with their established rivals that are not fully-vertically-integrated or with potential competitors that are contemplating an incompletely-vertically-integrated entry. I will address this possibility in more detail in Part VIII. At the present juncture, I will restrict myself to indicating that (1) the conditions that must be fulfilled for such predatory refusals to deal to be effective are more demanding than the Agencies and others seem to appreciate, relatedly (2) the incidence of such predatory refusals to deal may be much lower than the Agencies appear to believe, and (3) since predatory refusals to deal violate the Sherman Act, since the predatory character of a refusal to deal may not be that difficult to prove, since the proliferation of vertically-integrated firms may not be economically inefficient or morally undesirable, and since (perhaps I should write: “I hope”) the vast majority of business-actors forego opportunities to profit by making Sherman-Act-violative business-choices, it may be incorrect as a matter of law (as well as morally undesirable from various normative perspectives) to count against the Clayton-Act legality of a vertical (M or A) the possibility that the resulting firm may engage in predatory refusals to deal.

The sixth and final point I will make at this juncture relates to the impact on investment-competition of conglomerate and vertical (M&A)s that eliminate an effective potential competitor. Obviously, all such (M&A)s will decrease investment-competition either by preventing a new entry or by deterring an established firm from making a limit investment. Four additional points are relevant. First, a potential competitor may be “effective” in the sense that it would enter if established firms did not make enough limit investments to deter it from entering without its participation in an (M or A)’s affecting equilibrium QV investment in the relevant ARDEPPS because enough other potential competitors are as well-placed to enter or are as close-to-as-well-placed to enter as it is for its elimination as an independent entity not to affect equilibrium QV investment in the ARDEPPS. Second and relatedly, a potential competitor need not be the best-placed potential entrant into a particular ARDEPPS for its participation in an (M or A) with an established firm in that ARDEPPS to lessen investment-competition in that ARDEPPS. Third, as I have already explained, potential competition (as opposed to an individual potential competitor) will be effective in a given ARDEPPS if and only if the ARDEPPS’ entry-preventing QV-investment quantity is lower than its entry-barred, QV-investment-expansion-preventing QV-investment quantity. Fourth, as I have also already explained, if an (M or A) that eliminates as a separate entity a better-placed potential entrant results in that potential entrant’s prospective QV investment’s being replaced by a less-profitable QV-investment made (say) by a worse-placed potential entrant, that fact will not favor the conclusion that the (M or A) imposed a net dollar-loss on Clayton-Act-relevant buyers. Fifth, in some situations, potential competition will be ineffective, and other situations each member of a large set of potential competitors (for example, more than three potential competitors) will be effective.

I turn now to the 2023 Merger Guidelines’ positions on the possible impacts of (M&A)s on product variety and product R&D (inter alia, on the possible impacts of [M&A]s on investment-competition). There is good news for readers who would like to have to read less and bad news for those who are concerned with the Agencies’ applying the Clayton Act correctly as a matter of law and desirably as a matter of policy. Both the good news and the bad news reflect the same fact: I will write little on this issue because neither the 2023 Merger Guidelines nor any of their predecessors state much or a fortiori much correct about the various ways in which (M&A)s can affect the intensity of investment-competition or the approach the Agencies will take to analyzing this possible impact of (M&A)s.

VII. The Sherman-Act Illegality of (M&A)s

Part VII analyzes the determinants of the Sherman-Act-violativeness of any (M or A) and then addresses the positions if any the Agencies take on this issue in their 2023 Merger Guidelines. In my view, the Sherman Act prohibits those and only those (M&A)s that manifest a participant’s specific anticompetitive intent—in the language of its Section 1, prohibits those and only those (M or A)s that can correctly be said to involve “a contract or combination” “in restraint of trade.” An (M or A) will manifest a participant’s specific anticompetitive intent in two sets of circumstances: (1) when at least one participant’s ex ante perception that the (M or A) would be at least normally profitable was critically influenced positively by its belief that it would or might increase the resulting firm’s profits by reducing the absolute attractiveness of one or more best rival offers against which it would have to compete (relative to those against which a relevant participant would have had to compete) in one or more ways that would render the (M or A) at least normally profitable even though it would be economically inefficient in an otherwise-Pareto-perfect economy and/or (2)(A) at least one participant believed ex ante that the (M or A) would increase the profits the resulting firm could realize post-(M or A) by engaging in conduct motivated by specific anticompetitive intent relative to the profits the participants could have realized by engaging in such conduct post-(M or A) and (B) intended ex ante to take advantage of this opportunity. I will now address in more detail the categories of evidence that bear on whether either of these conditions is fulfilled and which party should bear the burden of supplying that evidence.

I will state at the outset that, because the relevant evidence (see below) is more securable by (M or A) participants than by the Agencies, it seems morally appropriate and legally correct to me to place the burden of producing that evidence on the participants. I should add that when the evidence of Sherman-Act-legitimizing motivation relates to static and dynamic efficiencies that the participants believe their (M or A) will generate for reasons that they have a rights-related interest in keeping secret the participants will have a rights-related interest in the evidence’s not being publicly released. I will assume in this Part that the Agencies do not have testimonial or documentary evidence of one or both participants’ specific anticompetitive intent—do not have statements by one or more participant-managers that they realized that the (M or A) would not be even normally profitable but for its freeing the participants from each other’s price-competition or investment-competition, statements by participant-managers that they believed that the (M or A) would render it profitable for the resulting firm to engage in more contrived-oligopolistic or predatory conduct and that they intended to take advantage of these opportunities, testimony of participant-employees that they heard participant-managers saying such things, testimony by mutual rivals of the participants that prior to the (M or A) the participants’ managers and/or employees told them that the resulting company would increase its retaliatory and predatory conduct post-(M or A), documentary evidence of such statements and/or messages, etc. On this assumption, two sets of evidence will be salient—(1) evidence that bears on the amount of profits that would constitute a normal rate-of-return on the (M or A) and (2) evidence on the magnitudes of the various Sherman-Act-legitimate profits the participants believed ex ante their (M or A) would generate. When the relevant conduct is a merger or acquisition, the information in the first category that is relevant is information on the private transaction cost the participants believed ex ante they would have to incur to negotiate the (M or A), the private cost the participants believed ex ante that the resulting firm would have to incur to integrate the two companies, and the participants’ ex ante perception of the normal rate-of-return for this type of expenditure. When the relevant transaction is an acquisition, the information in this category includes not only the purchase-price and the participant-integration costs but also any costs the acquirer must incur to finance the acquisition. Regardless of whether the relevant transaction is a merger or acquisition, the information in the second category focuses on the participants’ ex ante perceptions of the magnitudes of all the subcategories of Sherman-Act-legitimizing profits that their (M or A) could yield them: (1) the profits it would yield them by generating static variable-cost and fixed-cost efficiencies (which as we saw in the case of static variable-cost efficacies includes their ex ante perceptions of the [M or A]-generated MC-curve reductions, the unit outputs of the goods the resulting firm would have produced had its MC curve not fallen, the extra profits the MC-curve reductions would enable the resulting firm to realize by increasing the unit outputs of its products), (2) the profits it would yield them by reducing the fixed cost of renewing existing plant and equipment (by generating static fixed-cost efficiencies), (3) the profits it would yield them by generating dynamic efficiencies that rendered it profitable for the resulting firm to make one or more QV investment(s) that neither participant would have made, (4) the profits it would yield them by creating a resulting firm that could make one or more more-intrinsically-profitable QV investment(s) that would be substituted for the less-intrinsically-profitable QV investment(s) that would have been made by a participant, (5) the profits it would have yielded them by enabling the resulting firm to make a QV investment that would be profitable because it could stop investing after making one such investment whereas neither participant could profit by making a QV investment even though its investments were as intrinsically-profitable as the resulting firm’s because either participant’s investment would induce the other participant to make an investment, (6) the profits that it could yield by enabling the resulting firm to substitute for the participants’ existing, more-competitive QV investments “renewed” less-duplicative and on that account more-profitable QV investments, (7) the profits it could yield by enabling the owner/owner-manager of one participant to retire and liquidate his/her assets, (8) the profits it could yield the resulting firm by enabling it to take tax advantage of a tax-loss the other participant had sustained but could not utilize, (9) the profits it could enable the resulting firm to realize by increasing the NOMs it could obtain above those the participants could secure, (10) the profits it could enable the resulting firm to realize because it was better-placed than either participant or any mutual rival to cause it and its rivals to announce their prices and “lock themselves into” their prices in the order that would increase or maximize their across-the-board-HNOP array, (11) the profits it could enable the resulting firm to realize by helping it and its rivals to overcome the public-good-type problem that will tend to prevent them from spending as much money on (A) political campaigns, candidates, and parties, (B) legislative lobbying, (C) administration-process participation, and (D) adjudicating as is in their collective interest etc. Although some of these categories of profits may be morally dubious or worse, all of these categories are Sherman-Act-legitimating (with the possible exception of some exemplars of the last category of profits, which might be said to manifest a participant’s acting with anticompetitive intent). I hasten to state what should be obvious: if the participants in an (M or A) can establish the requisite probability that their ex ante perceptions that their (M or A) would be at least normally profitable did not depend on any belief that it would yield them profits by freeing them from each other’s price-competition or investment-competition or by enabling them to profit more from engaging in contrived-oligopolistic or predatory conduct, evidence that the (M or A) at issue would increase the participants’ profits in these Sherman-Act-non-legitimating ways would be irrelevant to the (M or A)’s Sherman Act illegality unless it was relevant to whether the participants believed ex ante that their (M or A) would increase the profits the resulting firm could realize by engaging in post-(M or A) Sherman-Act-violative conduct and intended ex ante to take advantage of this predicted effect of their (M or A).

I turn now to this latter possibility. I will make three relevant points. First, to prove that an (M or A) is Sherman-Act-violative on this account, it does not suffice to establish that the resulting firm did engage in Sherman-Act-violative conduct or even more such conduct than the participants would have engaged in as separate entities (even if one could in some unimaginable way establish the latter amount): for this purpose, the Agencies must show that (1) the participants realized ex ante that their (M or A) would increase the profits the resulting firm could realize by engaging in such conduct above the profits the participants could realize by doing so and (2) the participants intended ex ante to take advantage of this opportunity—both of which facts might be established even if the resulting firm did not actually engage in the conduct in question: I am not even confident about the probativeness on the above two issues of proof that the resulting firm did engage in the conduct in question. Second, it is extremely difficult to prove that contrived-oligopolistic conduct ¹²² or predatory conduct ¹²³ of any kind would be profitable or has been practiced. Third, the evidence that economists and judges claim is probative of contrived-oligopolistic conduct ¹²⁴ and predatory conduct ¹²⁵ is not in fact probative.

I will make three points about the 2023 Guidelines’ positions on the Sherman-Act illegality of (M&A)s. First, as I have already indicated, the 2023 Guidelines do suggest that the Agencies are aware that the applicable Sherman-Act test of illegality is different from the applicable Clayton-Act test of illegality. Second, even when the Guidelines consider possibilities that might be relevant to an (M or A)’s Sherman-Act illegality as well as to its Clayton-Act illegality ¹²⁶ or focus on a Sherman-Act-critical “nature and purpose of … [a] merger”—namely, “to foreclose rivals,” ¹²⁷ they do not consider the implications of their understanding of the conduct on which they are focusing for its Sherman-Act as opposed to its Clayton-Act illegality. Third, as I have already argued when discussing the Guidelines’ treatment of contrived-oligopolistic pricing and the creation of retaliation barriers to entry or established-firm QV-investment expansion and will argue in Part VIII when discussing the possibility of “vertical foreclosure,” the Guidelines’ treatment of these possible behaviors is inadequate.

VIII. The Relevance to a Vertical (M or A)’s Illegality of the Possibility that the Resulting Firm Would Refuse to Supply a Less-Vertically-Integrated Potential or Actual Rival

The 2023 Merger Guidelines address the possibility that a vertical (M or A) may create a vertically-integrated firm that will “foreclose rivals” ¹²⁸ that are not as vertically integrated by refusing to supply one or more such rivals with one or more goods or services the more-fully-vertically-integrated firm “produces” that the less-fully-vertically-integrated rival(s) need but do not produce—a possibility that the Agencies believe is legally relevant because of “the risk” that any such “limit[ation of] access” ¹²⁹ will lessen competition. Although the Guidelines claim that “the nature and purpose” of some vertical mergers is “to foreclose rivals, including by raising their costs,” the relevant Guidelines’ sentence ¹³⁰ says no more than that any conclusion that the “nature and purpose” of a vertical (M or A) is to foreclose one or more rivals is salient in that it “suggests the merged firm is likely to foreclose rivals”—does not state that any vertical (M or A) whose “nature and purpose” is to foreclose one or more rivals is Sherman-Act-violative on that account. ¹³¹ The Guidelines also do not provide much information about the factors that the Agencies believe determine whether or the extent to which foreclosing refusals to deal will lessen competition. Moreover, although the Guidelines do state that a vertical (M or A) may create barriers to entry “where it gives the merged firm increased visibility into the rivals’ competitively sensitive information” ¹³² and recognize that foreclosing conduct may raise the barriers to entry that a foreclosed potential competitor faces “by creating a need for the firm to enter at multiple levels,” ¹³³ the Guidelines do not refer to the reality that vertical (M&A)s (or vertical integration secured through internal growth or joint ventures) may increase the barriers to entry or established-rival expansion faced by rivals by creating a vertically-integrated firm that is more organizationally-economic-efficient than a less-vertically-integrated rival potential investor would be even if the vertically-integrated firm does not refuse to deal with any less-vertically-integrated actual or potential rival. Of course, I do not mean to imply that any tendency of a vertical (M or A) to lessen competition by raising rivals’ barriers to investment by making the resulting firm more economically efficient should as a matter of law count against its Clayton-Act legality. ¹³⁴

I will now make seven sets of points that relate to the likelihood that a vertical (M or A) will create a firm that will make more foreclosing decisions than the participants would have made as separate entities, the likelihoods that the foreclosing decisions that a vertically-integrated firm makes will be predatory or will lessen competition, and/or the difficulty of proving that a foreclosing decision was predatory or would lessen competition. First, it is important to recognize that no seller has a moral or legal duty to allow buyers to profit from patronizing it (to obtain buyer surplus by purchasing goods or services from it). Thus, decisions by a more-vertically-integrated firm to set prices to a less-vertically-integrated rival that deprive that rival of any buyer surplus could be problematic only if the seller’s efforts to remove all buyer surplus from the relevant customer were inherently unprofitable for it (say, because the research-cost the seller had to incur to prevent itself from charging the buyer a lower price than the buyer would have been willing to pay exceed the weighted-average or certainty-equivalent additional profits the research would otherwise enable it to achieve by obtaining higher prices from the buyer and the seller chose to incur those research-costs to induce the buyer’s exit or deter a potential competitor’s entry to reduce the absolute attractiveness of the best offers against which it would have had to compete).

Second and relatedly, the loss that an individual vertically-integrated firm can impose on a less-vertically-integrated actual or potential competitor by refusing to supply it depends not on the extent to which the relevant vertically-integrated firm is better-placed to supply that rival than is any other potential supplier of that rival but on the amount of buyer surplus the foreclosed firm would have obtained on its purchases from the forecloser if the forecloser did not “limit access” (predatorily or not).

Third, whether a vertically-integrated firm’s refusal to supply a potential investment-expander or potential entrant deters any such rival from investing depends not on the magnitude of the loss the foreclosure would impose on that potential investor (or established rival) but on whether that loss is larger or smaller than the supernormal profits that potential or actual rival would have realized on its investment absent the foreclosure.

Fourth, the benefits that a forecloser will obtain by deterring a rival investment by limiting the potential investor’s access will depend not only on how competitive the deterred investor’s investment would have been with the forecloser’s investments but also on (1) whether another rival would invest if the foreclosed and deterred investor does not (which depends both on the relative magnitudes of the barriers to investment faced by the deterred investor and the barriers faced by others potential investors and on the extent to which the forecloser’s foreclosing choices raises the L barriers faced by other potential investors by communicating to those rivals the forecloser’s intention to engage in relevant retaliatory conduct), (2) if another investor will invest, on any difference between the competitiveness of that investor’s investment with the forecloser’s projects and the competitiveness of the deterred investor’s investment with the forecloser’s projects (which may also be affected by the foreclosing conduct), and (3) on the ability of the forecloser to reduce its losses by making one or more limit investments.

Fifth, it is extremely complicated to determine whether a seller’s refusal to supply a particular buyer was inherently unprofitable (or whether the price that a seller charged a particular buyer was inherently-unprofitably-high). One cannot make these determinations simply by seeing whether the payments the buyer would have been willing to make to the relevant seller would have exceeded the conventional variable cost the seller would have had to incur to supply the relevant quantity of the seller’s good to the buyer in question. One must also consider inter alia (1) the pricing and other conventional transaction costs the seller would have to incur to negotiate a deal with the relevant buyer, (2) if the seller is setting across-the-board prices, the profits the seller would lose on its sales to other buyers if it reduced its price to the buyer in question below the price it would otherwise have found most profitable to charge, (3) if the seller is practicing individualized pricing, the contextual marginal costs it would have to incur to supply the buyer in question on various terms, (4) the possibility that the buyer in question had violated the contract, tort, or property rights of the relevant seller and the seller wanted to prevent the buyer from obtaining the buyer surplus the seller would not otherwise have found it profitable to remove to deter the buyer and others from violating its rights in the future or to induce the buyer to pay the seller the damages to which it was actually entitled, (5) the loss an input-supplier would sustain because the buyer would incorporate the seller’s input into an inferior final product whose poor performance would injure the reputation of the input supplier or the loss a final-good supplier would sustain because the buyer operated an inferior distributive outlet and the reputation of the final-good producer’s product(s) would be damaged by its (their) association with this inferior distributor, (6) the loss a supplier might sustain by supplying a particular buyer because that buyer had an unattractive moral or political reputation and some buyers have a dispreference for patronizing any firm that supplies such an unattractive customer, (7) the product whose supply is at issue is a desirable product whose performance depends on its being properly repaired and maintained or being used appropriately, and the supplier has reason to believe that the particular buyer will not treat, maintain, and/or repair the supplier’s product properly or use it appropriately and is concerned that it will be blamed for its product’s resulting poor performance because the buyer will bad-mouth it and/or others will observe the product’s poor performance and attribute that performance to the product, (8) the refusal was made by or an “access-limiting” high price was charged by a wholesaler to fulfill its lawful contractual obligations to a manufacturer that had morally and legally valid reasons for including vertical territorial restraints or resale-price-maintenance clauses in its contracts with the wholesaler, etc.

Sixth, although the Agencies may be aware of this reality, the Guidelines do not indicate that the amount by which a vertically-integrated established firm’s refusal to supply a less-vertically-integrated potential or actual entrant with a product the refused party needs will reduce the profits the refused party can realize may be reduced by the refused party’s ability to secure the refused good or service in question from another established supplier, to induce an independent firm to enter the business of producing the refused good or service and to supply it (by entering [say] into a long-term purchasing agreement with that firm), or to form a joint venture to produce the refused good or service.

Seventh, when the issue is the likelihood that a vertical (M or A) may create a resulting firm that engages in foreclosing conduct, it may be worth pointing out that, with one important qualification (see below), a more-vertically-integrated firm will not find it profitable to drive out or prevent the entry of a less-vertically-integrated firm that can perform the functions the latter firm does perform better than the relevant components of the more-vertically-integrated firm can do: the more-vertically-integrated firm will find it profitable to use the less-vertically-integrated firm to perform those functions at least if it can price the products and services the less-vertically-integrated firm buys from it so as to remove all buyer surplus from the less-vertically-integrated firm. Almost certainly, the more-vertically-integrated firm would find it more profitable to merge with or acquire the less-vertically-integrated firm than to drive it out or prevent its entry (at least if the [M or A] in question would be deemed legal). The qualification, which is important, is that the preceding argument will be less convincing if the less-vertically-integrated firm can obtain profits even in the short run on its purchases from one or more more-vertically-integrated firms because of the competition its potential suppliers wage against each other and may be in a position in the long run either if it does vertically integrate more or if it does not to obtain more profits on such purchases and to inflict more losses on the more-vertically-integrated firms.

I could go on. But the conclusion I want to establish is that the analysis of the foreclosure-related possibilities is complicated and that the 2023 Merger Guidelines do not recognize much less analyze satisfactorily the salient issues. One final point: even if relevant markets could be defined non-arbitrarily, the Guidelines’ claim that “[l]imiting rivals’ access to the related product will generally have a greater effect on competition in the relevant market if the merged firm and the dependent rivals face less competition from other firms” ¹³⁵ should be qualified by including the clause

controlling for inter-market differences in such parameters as the equality of the competitiveness scores (however defined) of all product-pairs placed in a market and inter-market differences in the ratios between the number of times that a firm in the market is best-placed and the number of times it is second-placed or close-to-second-placed.

IX. Conclusion

It would be not only inaccurate but remiss not to acknowledge that the 2023 DOJ/FTC Merger Guidelines are superior to their predecessors in some respects: they admit the “fuzziness” of at least some market definitions; they indicate that the Agencies will make use of more types of non-market-aggregated data than the preceding Guidelines referenced and will employ some analytic techniques that may not be market-oriented that the preceding Guidelines did not reference; and they devote more space to the possible impacts of (M&A)s on quality, variety, and innovation than their predecessors did. However, in my judgment these improvements constitute at most small steps in the right direction: the Guidelines continue to manifest the Agencies’ stubborn insistence on the value of market-oriented approaches; they mis-specify some of the non-market-aggregated parameters that are legally relevant that they do reference and continue to ignore a host of such parameters altogether; and they say virtually nothing useful or accurate about the ways in which (M&A)s can affect investment-competition and investment or those impacts’ determinants. Moreover, the 2023 Guidelines make some problematic and erroneous claims that their predecessors did not make—for example, refer to some antitrust goals that need additional specification or are dubious and seem to endorse “limit-pricing ‘theory.’”

This Article’s Abstract and the Article itself indicate that I have many objections to the 2023 Guidelines. I find these Guidelines deficient and frustrating even if their failings are not surprising.