Welfare Economics and Antitrust Policy: A Selective Summary

Abstract

This Article summarizes the most salient components of my study WELFARE ECONOMICS AND ANTITRUST POLICY. It (1) distinguishes moral-rights (justice) analysis from moral-good analysis, (2) defines and examines the relevant extensions of the liberal conception of justice, (3) defines various egalitarian conceptions of the moral good and argues that libertarianism is not morally defensible and would not in any event have many of the extensions some claim it has, (4) defines various categories of antitrust-policy-coverable conduct (including natural and contrived oligopolistic conduct of various sorts and predatory conduct of various sorts), (5) defines the components of the gap between a product’s price and marginal cost that need to be distinguished respectively in individualized-pricing and across-the-board-pricing situations, (6) defines “investment competition” and identifies the determinants of its intensity, (7) explains why markets cannot be defined non-arbitrarily, (8) defines “the impact of a choice on economic efficiency” and lists various categories of economic inefficiency that can be generated, (9) delineates the protocol for predicting the economic efficiency of a choice that is ex ante economically efficient (which considers the impact of choices on many categories of economic inefficiency that conventional analyses ignore, takes economically efficient account of the fact that the individual exemplars of Pareto imperfections that would cause economic inefficiency in an otherwise-Pareto-perfect economy can counteract as well as compound each other’s misallocative tendencies, and considers how to take appropriate account of the fact that both analysis and its financing are allocative-costly), (10) examines the policy-relevance of a choice’s impact on economic efficiency, and (11) executes some preliminary antitrust-policy analyses.

The Articles in this issue execute antitrust-policy analyses or focus on one or more positions I take in my two-volume study Welfare Economics and Antitrust Policy. It therefore seems appropriate to begin this issue with accounts of (1) the most salient features of my approach to antitrust-policy analysis and (2) the most significant antitrust-policy conclusions my study reaches. This Article contains three sections. The first delineates definitions of (1) various moral-analysis-related concepts and the (morally defensible) liberal and egalitarian moral norms that I think the United States and many other countries and confederations that have antitrust policies are committed to instantiating and one (to my mind) morally indefensible purported moral norm (libertarianism), (2) various categories of antitrust-policy-coverable conduct, (3) concepts that refer to the components of the gap between a seller’s price and conventional marginal cost and to the determinants of the intensity of investment-competition in an arbitrarily defined portion of product-space (ARDEPPS)—see Section 3 for my reasons for substituting the acronym ARDEPPS for the term “market,” (4) two concepts that relate to the motivation for or a particular consequence of an exemplar of antitrust-policy-coverable conduct that are legally relevant and policy-relevant (the concepts of “specific anticompetitive intent” and the concept of “lessening competition” as it should be understood in the antitrust-law/policy-analysis context), and (5) various concepts and one theorem that relate to economic-efficiency analysis. The second section provides a brief overview of the protocol for economic-efficiency prediction or post-diction (hereinafter “prediction”) that the study employs and explains how this protocol differs from the economic-efficiency-prediction protocol used by virtually all economists (including all economists who analyze the economic efficiency of antitrust-policy-coverable conduct and antitrust policies). The third section explains why the antitrust-policy-analysis protocol my study uses is not “market-oriented”—does not require any allegedly relevant market to be defined and, relatedly, does not make any use of estimates of the magnitudes of market-aggregated parameters (such as the market share of one or more alleged or actual perpetrator[s] of antitrust-policy-coverable conduct or the concentration of the seller-side of any [allegedly relevant] market).

I. Definitions

A. Definitions of Concepts That Play a Role in Moral Analysis, of Various Relevant Defensible Moral Norms, and of One Morally Indefensible Decision-Criterion

My study of antitrust-policy analysis proceeds on the assumption that in the polities with which I am concerned antitrust-policy-coverable conduct and antitrust policies are to be morally evaluated by their impact on the instantiation of the liberal conception of justice and, if the antitrust-policy-coverable conduct or the antitrust-policy response at issue is not mandated or prohibited by the liberal conception of justice, by the conduct’s or policy’s impact on a pure or mixed egalitarian conception of the moral good. In my terminology, the preceding conclusions reflect my assumption that the polities in question are liberal-moral-rights-based polities that are committed not only to giving lexical priority to the instantiation of the liberal conception of justice but also to the instantiation of an appropriate (see below) pure or mixed egalitarian conception of the moral good (to the extent that this second objective can be secured without violating any liberal moral right).

My study proceeds on this normative assumption not because I think one can demonstrate through Foundationalist argument ¹ (1) the universal moral correctness of engaging in a bifurcated prescriptive-moral discourse that distinguishes discourse about “the just” (about moral rights and obligations) from moral discourse about “the moral good,” (2) the universal moral correctness of giving lexical priority to justice-conclusions over moral-good conclusions, (3) the moral optimality of the liberal conception of justice, or (4) the moral optimality of egalitarian conceptions of the moral good but because when I apply the philosophically informed empirical protocols I argue should be used to derive conclusions about a polity’s moral type and commitments to the U.S. Declaration of Independence, the U.S. Constitution and its various Amendments, the Charter of Fundamental Rights of the European Union, the contemporary German constitution, and the contemporary French constitution and its historical antecedents, the conclusion that emerges is that (these polities)/(that confederation) are/is committed to giving lexical priority to the instantiation of the liberal conception of justice and to making decisions that instantiate some egalitarian conception of the moral good when the choice that would do so is not prohibited by the polity’s commitment to instantiating the liberal conception of justice. ²

I hasten to add that, even if you are not convinced that the above polities/confederation have the moral commitments I claim they have, my study’s and this section’s moral-norm discussions should be of interest because (1) antitrust policies must be morally evaluated from the perspective of some moral norm(s), (2) the liberal conception of justice and the various egalitarian conceptions of the moral good I will distinguish are morally defensible, and (3) this section will also explain why the various criteria that many antitrust-policy analysts use to evaluate antitrust policies (libertarian decision-criteria and criteria that assume that the only morally relevant desideratum/desiderata is/are economic efficiency, consumer welfare, or seller-competition) cannot bear moral scrutiny.

I recognize that the expressions italicized in the preceding paragraphs require elucidation. I will start by illustrating the distinction I draw between prescriptive-moral discourse about “the just” (about moral rights and obligations) and prescriptive-moral discourse about “the moral good.” To illustrate the difference between these two categories of prescriptive-moral discourse, I will examine a situation in which one moral-rights bearer is in a position to render assistance to another moral-rights bearer who is a victim of an automobile accident that has taken place. In any moral-rights-based polity, two sorts of moral questions can be asked about the moral desirability of the potential assistance-provider’s providing assistance: (1) does the potential assistance-provider have a moral obligation to render assistance, and (2) would it be morally good for the potential assistance-provider to render assistance. In a liberal-moral-rights-based society, the answer to the moral-obligation question will depend on whether the potential assistance-provider has promised to provide assistance, is in an intimate relationship with the injured party (or perhaps has a status-relationship to the injured party that is normally associated with intimacy), was a culpable cause-in-fact of the potential assistance-recipient’s injury or possible injury, or was uniquely well-placed to provide assistance to the injured or at-risk party that would significantly reduce the probability that that party would suffer an opportunity-to-lead-a-life-of-moral-integrity-imperiling loss without subjecting himself or herself to a significant risk of sustaining such a loss or devoting so much time and effort to the assistance-effort for the imposition of a duty to provide the assistance to reduce the probability that the assistance-provider and others would take their lives morally seriously by undermining their perceptions that they are the authors of their own lives. Regardless of whether the potential assistance-provider has a moral duty to provide the assistance (but admittedly less-morally-saliently if the potential assistance-provider does have such a duty), one can ask whether, from the perspective of various conceptions of the moral good, it would be morally good for the potential assistance-provider to provide assistance in the situation in question. An evaluator’s relevant moral-good assessment will depend on both the conception of the moral good to which the evaluator subscribes or has been instructed to use (see the following text) and the facts of the matter that that conception implies are relevant to the moral goodness of the relevant potential assistance-provider’s providing assistance.

The next “concepts” I wish to elucidate are the philosophically informed empirical protocols that I think should be used, respectively, to determine (1) whether a polity is moral-rights-based or is moral-goal-based; (2) if the society is moral-rights-based, the moral norm that grounds the conception of justice it is committed to instantiating; (3) if it is moral-goal-based, the moral norm that grounds the conception of the moral good it is committed to instantiating, and (4) if the society is moral-rights-based, the conceptions of the moral good that are and are not morally permissible for its government’s total decisions to instantiate (where the decisions’ moral-good-related defensibility depends on their relationship to the distribution of the polity’s members’ moral-good positions). (To determine whether a society/polity that does/[does not] distinguish justice from the moral good has moral integrity, one must determine whether the society’s/polity’s members’ claims/arguments/conclusions/perceptions/conduct conform sufficiently to the entailments of the conception of justice [the moral good] to which they best conform [see below] for the society to deserve to be said to have moral integrity.)

The first of the above protocols should be obvious. To determine whether a society/polity is a moral-rights-based society/polity, examine (its members’ claims/arguments/conclusions/perceptions/conduct)/(its constitution’s text) to see whether they manifest (1) a belief in the importance of distinguishing justice from the moral good and (2) a commitment to giving lexical priority to justice-conclusions over moral-good conclusions. The protocol for identifying the moral norm that grounds a moral-rights-based society’s justice-commitment has both a “goodness of fit” component and a “partially-exonerating explicability of non-fits” component. This protocol varies with whether it is focusing on the conduct and perceptions of a society’s members or on the content of a polity’s constitution and other facts that affect the correct interpretation of its constitution’s text. ³

When the focus is on the conduct and perceptions of a society’s members, the “goodness of fit” component of the protocol focuses on how well the various “candidates” for the title of being the conception of justice the society is committed to instantiating fit (1) the moral-rights claims its members make and do not make; (2) the arguments that observers, officials, claimants make and do not make for justice-claims; (3) the conclusions that officials and observers reach about the correctness of the moral-rights claims that were made or could have been made; (4) how close disputants, officials, disputants, and observers thought the “cases” in question were (enquoted because most of the “cases” will not be legal cases); (5) how certain disputants, officials, and observers were about the resolutions of the claims in question that were correct; and (6) the extent to which the relevant society’s members’ conduct fulfilled the moral obligations they would have if the “candidate” in question really was the justice-conception to which the relevant society was committed. The “partially-exonerating explicability of non-fits” component focuses on explanations of non-fits that reduce the damage the non-fits do to the candidacy of the relevant imperfectly fitting candidate. The list of partially exonerating explanations of any non-fit includes (1) the fact that the non-fit related to a claim/argument/conclusion/perception/conduct-choice of someone who was not morally perceptive; (2) the fact that the non-fit related to a claim/argument/conclusion/perception/conduct-choice that was made or a perception that was formed in haste (say, in an emergency situation) rather than as a result of or in response to a careful, thorough deliberation; (3) the fact that the non-fit reflected an empirical error (e.g., about the capacities of individuals of African descent or of women); (4) the fact that the non-fit reflected a conceptual error such as the conclusion that one cannot have a right to privacy in a public toilet (an error some U.S. courts have made when assessing the admissibility of evidence of homosexual sexual conduct in public toilets the police collected by drilling a hole in the roof of the facility and observing conduct through that peep-hole) or the conflation of the psychological-disposition, material-welfare, and moral-entitlement senses of “interest” and “concern” (an error that several U.S. courts have made when assessing whether newspaper stories that sacrifice the privacy interests of their subjects are morally and constitutionally legitimated by “the public’s right to know”); (5) the fact that the victim of what the candidate implies would be a moral-rights violation did not make a moral or legal claim because of the transaction-costliness of doing so (which would be double-edged if the candidate implied that the government has a moral duty to prevent such victims from being deterred from pursuing their claims by transaction costs); (6) the fact that the perpetrator of conduct that the candidate implies would be moral-rights-violative opposed the victim’s claim because doing so was rendered ex ante profitable by the possibilities that the victim would abandon its claim, that the claim would be incorrectly resolved, and/or that opposing the relevant victim’s claim would deter future victims from making or pursuing claims (an explanation that is also double-edged in that it disfavors the conclusion that the relevant society has moral integrity); (7) the fact that the maker of a choice the candidate would deem moral-rights-violative made the moral-rights-violative choice because it thought that the victim would not perceive that its moral rights had been violated, would not be able to identify the violator, and/or would be deterred from making and pursuing a claim by the transaction cost of doing so and/or by a belief that the legal system would not do it justice (explanations that are also double-edged for the reason that the sixth explanation in this list is double-edged), and so on. In any event, this overall protocol will yield the conclusion that a particular “candidate” is the moral norm that grounds the relevant society’s conception of justice if the society is moral-rights-based (or the society’s conception of the moral good if it is goal-based) and that the relevant society has moral integrity if that candidate (1) “discounted-fits” (i.e., fits appropriately discounting the associated non-fits) the relevant facts better than any other morally defensible candidate does and (2) discounted-fits the relevant facts sufficiently well for the society to deserve to be characterized as a society of moral integrity.

When the focus is on the constitution of the relevant polity, the protocol also has a goodness-of-fit and a partially exonerating explicability of non-fit component. The constitution-related “data” the protocol takes into consideration includes the text of the constitution and various preceding documents that inform the meaning of that document’s text (in the United States, for example, the text of the Declaration of Independence), other information that is relevant to the correct interpretation of such texts including the ordinary meaning of words in the text in context at the time of its promulgation and ratification, contemporaneous discussions of the meaning of the text by its drafters and ratifiers, information about antecedent events that led to the inclusion of particular texts in the constitution, information about other decisions the text’s drafters or ratifiers made just prior to, contemporaneous with, or shortly after they drafted or voted for the constitution, possibly exonerating explanations for non-fitting constitutional texts or omissions from the constitution that do not fit the conclusion that the constitution commits the polity to instantiating a particular moral norm (such as the argument that the U.S. Constitution did not prohibit slavery—indeed, implicitly recognized that slavery would continue—because the northern states needed to make this compromise to induce the southern states to join the Union or the argument that neither the original U.S. Constitution nor any specific Amendment in the Bill of Rights [as opposed to Amendment IX] guaranteed women the right to vote because at that time the society’s members believed that females lacked the capacity to vote competently).

The next moral-norm-related concept I need to define or explicate is “the liberal conception of justice.” On my understanding, the defining characteristic of the liberal conception of justice is that it assigns lexical priority to all moral-rights bearers’ (all creatures’ that possess the non-experience-generated neurological prerequisites for leading a life of moral integrity) having and seizing a meaningful opportunity to lead such a life by fulfilling their liberal moral obligations, developing self-consciously or implicitly a possibly-dialectically-evolving conception of the moral good, and making choices that conform to a hard-to-specify requisite extent to the conception of the moral good to which they, respectively, subscribe at each point in time.

I will now be more concrete. The basic liberal moral obligation of individual moral-rights bearers is to treat other moral-rights bearers with appropriate, equal respect and to show appropriate concern for their “welfare” in part in the enquoted term’s utilitarian conception but preeminently for such other creatures being able to lead a life of moral integrity. In my judgment, liberalism does not impose a duty on individual moral-rights bearers to show appropriate, equal concern for all other moral-rights bearers because it permits such individuals to favor their intimates over moral strangers. Liberalism permits such favoritism because participation in intimate relationships disposes the participants to take their lives morally seriously in general in that it involves the participants considering on their own, with their intimates, and with third parties how participants in intimate relationship should treat each other, how they and their co-participants have treated each other, and how they and their intimates should treat each other in the future.

In my judgment, liberalism also has two other implications that bear on the moral desirability of some antitrust policies. First, although liberalism does sometimes impose positive moral obligations on moral-rights bearers (to benefit others), liberalism is disposed not to impose positive obligations on non-government moral-rights/duty bearers. I base this admittedly contestable conclusion on the following “facts” and contestable empirical assumptions: (1) by definition, liberalism gives lexical priority to relevant creatures’ having and seizing a meaningful opportunity to lead a life of moral integrity; (2) moral-rights bearers are more likely to lead a life of moral integrity if they perceive themselves to be the authors of their own lives; and (3) the imposition of positive moral duties to confer equivalent-dollar gains on other moral-rights bearers is more likely than the imposition of otherwise-equally-onerous negative moral duties not to impose (equal) equivalent-dollar losses on others to undermine the prospective obligor’s and others’ perceptions that they are the authors of their own lives. Second, the test of moral culpability that liberalism claims is applicable when a moral-rights bearer makes a choice that imposes an equivalent-dollar loss on one or more other moral-rights bearers depends on whether the loss is opportunity-to-lead-a-life-of-moral-integrity-imperiling (might or would kill a victim, inflict neurological damage on a victim that would render the victim neurologically incapable of taking its life morally seriously, or perhaps might or would inflict pain and/or disability on a victim that strongly militates against the victim’s leading a life of moral integrity) or is not opportunity-to-lead-a-life-of-moral-integrity imperiling (in my terminology, is “a mere-utility loss” [regardless of the amount of utils that would or might be lost]). When a moral-rights bearer’s choice would or might impose an opportunity-to-lead-a-life-of-moral-integrity-imperiling loss on one or more other moral-rights bearers, liberalism implies that the choice is liberal-moral-rights-violative unless the injurer did appropriate research into its options and their relevant consequences and concluded that its choice would not imperil opportunities to lead lives of moral integrity all tolled (e.g., because, in the case of Pharmaceutical company, the amount of opportunity-to-lead-a-life-of-moral-integrity-imperiling losses it caused through the pollution generated by its production of successive units of a drug it produced was lower than the opportunity to lead a life of moral integrity preserved by the consumption of each unit of the drug it produced and no change in production process liberalism would obligate it to make would preserve opportunities to lead lives of moral integrity all tolled [perhaps because the relevant drug company would find it more profitable to shut down rather than to shift to the safer production process]). When a moral-rights bearer’s choice imposes mere-utility losses on its traditional victims, its choice will be liberal-moral-rights-violative only if it did appropriate research into its options and their relevant consequences and that research implied that the equivalent-dollar gain the choice should be predicted to yield the injurer is at least as high as the equivalent-dollar loss the choice should be predicted to impose on the chooser’s traditional victims. ⁴ I argue that in the mere-utility-loss cases the relevant metric for the gains and losses the relevant choice generates is equivalent-dollars rather than units of utility (utils) because (1) the choice between these two metrics will be critical only if differences in the average-utility values of the equivalent-dollars gained and lost cause choices that yield a net equivalent-dollar gain/loss to the chooser and its traditional moral-rights-bearing victims to yield a net utility loss/gain to these creatures and (2) the injurer will rarely be a cause-in-fact and never be a culpable cause-in-fact of the differences between the injurer’s and its traditional victims’ income/wealth positions or dispositions to obtain utility by spending or saving money that can cause the metric to affect the relevant differences and on that account the net benefit/cost comparison.

Liberalism also has many implications for the moral obligations of the government of a liberal-moral-rights-based polity. Most abstractly, (1) liberalism obligates such governments not only to show appropriate, equal respect for all moral-rights bearers for whom it is responsible but also to show appropriate, equal concern for their “welfare” (because government cannot enter into intimate relationships with any of the moral-rights bearers for whom it is responsible), and (2) liberalism is not indisposed to impose positive duties on government (since government does not have a perception that it is or is not the author of its own life). More concretely, liberalism imposes a moral duty on the government of a liberal-moral-rights-based polity (1) not to make choices that would violate the negative liberal moral duties of non-government actors and (2) to fulfill a wide range of positive moral duties—(A) to prevent the society’s members and participants from violating their liberal moral duties (by educating them, making liberal-moral-duty-violating choices illegal, legislating appropriate sanctions for liberal-moral-rights-violating conduct [which are not more severe ex post than the wrongfulness of the wrongful conduct merits], implementing the laws in question appropriately [without violating anyone’s legal-process-related liberal-moral rights]); (B) to give victims of liberal-moral-rights violations appropriate opportunities to secure legal redress; (C) to assure that the moral-rights bearers for whom it is responsible have the nutrition, clothing, living accommodations, medical care, psychological nurturing, formal education, knowledge of alternative “life-styles”/(conceptions of the moral good)/(religions), experiential opportunities (including opportunities to participate in intimate relationships), and privacy that significantly contribute to their having a meaningful opportunity to lead a life of moral integrity; (D) to give each moral-rights bearer for whom it is responsible an appropriate, equal opportunity to participate in political elections and government-decision-making processes and to influence government decisions; and (E) to make decisions that collectively give equal weight to its various members’ various morally defensible moral-good positions. (I am fully aware of the difficulty of operationalizing most of these moral duties.)

I turn now to the concept of “egalitarian conceptions of the moral good.” My study delineates five conceptions of the moral good that are egalitarian in that the weight they assign to any relevant creature’s obtaining a unit of any desideratum does not depend on any non-relevance-determining characteristic ⁵ of the relevant creature or any choice the relevant creature made. The five egalitarian conceptions of the moral good my study considers are the total-utility and average-utility variants of utilitarianism, the equal-utility egalitarian norm, the equal-resource egalitarian norm, and the various other equal-opportunity egalitarian norms.

The effect of a choice on the utilitarian norms’ instantiation depends on its impact on economic efficiency and material inequality. The effect of a choice on the non-utilitarian egalitarian norms depends solely on its impact on material inequality (given the metric for that inequality that is deemed appropriate).

My study also considers a number of positions that at least some evaluators who claim to subscribe to libertarianism take. I question the moral defensibility of libertarianism on the ground that this position denies that (with one telling exception) creatures whose choices’ moral status can sensibly be assessed never have a moral obligation to benefit any other creature and claims that (with that exception) it can never be demonstrated that it would be morally good for such a creature to benefit another such creature. (The exception is telling because it relates to the justness or “moral-goodness” of a promisor’s fulfilling its promise and liberalism acknowledges the moral desirability of fulfilling promises because it is in the self-interest of prospective promisors to be able to make binding promises.)

My study also argues that a number of purportedly moral conclusions that some self-styled libertarians claim libertarianism warrants are not morally defensible. Thus, I argue that (1) one cannot establish the moral desirability of allowing lucky finders to keep what they found by arguing that that conclusion imposes only a trivial ex ante loss on each of the potential future finders on whom it imposes a loss because the morally relevant loss is the ex post loss of the actual future finder, because the alleged triviality of the ex ante loss is morally irrelevant, and because in some instances the alternative finder is sufficiently identifiable ex ante for its ex ante loss not to be trivial; (2) the libertarian claim that each individual is morally entitled to receive resources whose allocative value equals the individual’s contribution to the total allocative product of the relevant economy is morally undercut by the fact that the vast majority of interpersonal allocative-product differences are attributable to factors over which the individuals in question had no control for whose magnitudes individuals are not morally responsible—for which they cannot claim moral credit; and (3) the related libertarian argument that government efforts to redistribute income are morally undesirable because markets allocate to each individual resources whose allocative value equals the individual’s allocative product would be morally indefensible even if individuals deserved moral credit for their allocative products because in actual Pareto-imperfect economies individuals’ gross wages do not equal their respective allocative products and the gross returns to saving and investment do not equal the contribution that the investment in question made to economic efficiency.

Finally, my study points out that (1) no matter how the following concepts are defined, increases in economic efficiency, consumer welfare, and seller-competition are not intrinsically morally valuable; (2) different analysts and the same analyst in successive sentences or successive paragraphs define these concepts differently; (3) although there is a “correct” definition of “the impact of a choice on economic efficiency,” economists implicitly define this concept in other ways that are “incorrect” in that their implicit definitions do not fit the intuitive understanding of those who refer to the concept and create a concept that cannot play a useful role in the analyses in which it is referenced; (4) there is no correct way to define the impact of a choice on consumer welfare or seller-competition; and (5) contrary to the implicit assumption of those economists who claim that antitrust-policy-coverable conduct and antitrust policies should be evaluated by their impacts on economic efficiency, consumer welfare, and seller-competition, these impacts are not monotonically related.

B. Definitions of Various Categories of Antitrust-Policy-Coverable Conduct

My study distinguishes four categories of investments that business entities can make. The first is quality-or-variety-increasing (QV) investments. Such investments create a new product, an additional distributive outlet, or additional capacity and inventory (which increase the average speed with which their maker can supply its product or service during a fluctuating-demand cycle). Most QV investments do not generate any scientific, technological, or commercial discoveries, but a few do.

The second category of investment is investment in production-process research (PPR). PPR is designed to discover a privately cheaper way to produce a relevant quantity of an existing product. I should acknowledge that some investments might be said to be investments in both product-research and PPR (seek to discover a new product that is cheaper to produce).

A third category of investment is investment that renews production facilities that will use existing production processes to produce existing products.

A fourth type of investment is investment in (1) research into the performance of an existing product, lifetime cost of maintaining and repairing an existing durable product, the external cost generated by the production and consumption of an existing product, or buyer misperceptions of these facts about an existing product and (2) communications of information about those issues to potential and actual buyers and/or the government.

My study distinguishes four “pricing-techniques” a seller can employ either when selling its product to an individual buyer or when selling its product to multiple buyers. This list includes (1) “single pricing”—setting a single per-unit price and allowing the buyer to purchase as many units of the seller’s product as the buyer wishes to purchase at that price; (2) perfect price discrimination, which can be practiced (A) by pricing the successive units of the seller’s product the buyer might want to purchase at usually different per-unit prices equal to the successive units’ respective dollar-values to the buyer, (B) by charging the buyer a lump-sum fee for the output of the seller’s product at which the buyer’s demand curve for the seller’s product cuts the seller’s marginal-cost curve from above—a lump-sum fee equal to the dollar-value of these units to the buyer minus any risk costs this arrangement imposes on the buyer, or (C) by charging the buyer the highest lump-sum fee it could pay without making itself worse-off ex ante for the right to purchase the just-referenced unit output of the seller’s product at a per-unit price equal to the height of the buyer’s demand curve for that product at the just-referenced output; (3) a mixed pricing-technique in which the seller charges the buyer a lower lump-sum fee than just described for the right to purchase units of the seller’s product for a higher per-unit price than the per-unit price just referenced; and (4) tie-ins and reciprocity arguments in which the seller reduces the transaction surplus it destroys in the process of removing a given amount of buyer surplus by charging the buyer a supra-marginal-cost per-unit price by conditioning its sale of one product for a per-unit price that is lower than the per-unit price it would charge the buyer in an independent (one-product) transaction on the buyer’s agreeing to purchase its full requirements of a second product from the seller for a per-unit price that exceeds the second product’s standard market per-unit price or by conditioning its sale of one product to a buyer for a lower per-unit price than the per-unit price it would charge that buyer for the product in question in an independent transaction on that buyer’s agreeing to supply the first firm with the “buyer’s” total output of a second product for a lower per-unit price than the buyer of the first product would otherwise have charged for the second product. A seller can also employ any of these four pricing-techniques discriminatorily when pricing its product or products to different customers.

My study also delineates clear definitions of “oligopolistic conduct,” “predatory conduct,” and “competition-distorting conduct.” These definitions are important because neither economists nor lawyers have ever defined these concepts explicitly, and, although the definitions I delineate and use are consistent with most of the definitions that economists’ and lawyers’ usages of these concepts implicitly adopt, my definitions are inconsistent with some of these usages (e.g., with the claim that choices that manifest any kind of recognized interdependence are “oligopolistic” or with the claim that any investment that the investor would not have made had it not believed that it would deter a rival investment is “predatory” [and therefore violates the Sherman Act] ⁶ ). (Economists and lawyers also fail to recognize the distinction I draw between contrived-oligopolistic choices and natural-oligopolistic choices—a distinction that is legally critical in the United States and does have some policy relevance.)

In my terminology, a business entity is said to have initiated or to have attempted to initiate an oligopolistic-conduct sequence if it made a choice it would not have found ex ante to be at least normally profitable had it not believed ex ante that one or more rivals’ responses (hereinafter “relevant-rival responses”) to that choice would or might be affected by the rivals’ perception that the initiator would or might react to uncooperative and/or cooperative rival-responses in ways that would render unprofitable for the rivals uncooperative responses they would otherwise have found profitable. I distinguish two subcategories of oligopolistic conduct by the reason the relevant rivals anticipate the initiator will react in the above way(s) to their responses. In my terminology, oligopolistic conduct is said to be “contrived” when the anticipated initiator-reactions are inherently unprofitable for the initiator and the initiator has communicated to its relevant rivals its intention to react in the way(s) in question despite their inherent unprofitability to induce the relevant rival(s) and perhaps other rivals as well to respond cooperatively to its future oligopolistic (and other strategic) conduct. In my terminology, oligopolistic conduct is said to be “natural” when the relevant rivals anticipate the initiator’s relevant reactions because the initiator will have an opportunity to react (perhaps because relevant buyers will give the initiator a chance to rebid) and the initiator will find the anticipated reaction inherently profitable.

A wide variety of types of choices can be made oligopolistically. An initiating seller’s price-choice will be oligopolistic if ex ante it would have found the price it charged to be unprofitably high had it not believed that relevant rivals’ responses would be influenced by their perception that it would or might react to their cooperative and/or uncooperative responses in one or more ways that would render unprofitable for them uncooperative responses they would otherwise have found profitable. In my terminology, the initiator’s oligopolistic pricing will be deemed contrived if it elicited its relevant rivals’ relevant expectations by communicating to them its intention to reciprocate to their cooperative responses by foregoing the opportunity to beat offers they make to their customers that contain prices that would make it inherently profitable for the initiator to beat their offers and/or to retaliate against their uncooperative responses (say) by offering inherently-unprofitably-low prices to their customers. (I should add that in my terminology an initiator’s rival’s cooperative responses, an initiator’s rival’s decision to take advantage of the initiator’s offer of reciprocation, the initiator’s fulfilling its promise to reciprocate, and an initiator’s retaliation against rivals’ non-cooperation are also deemed “oligopolistic.”) When the relevant rivals anticipate the initiator’s relevant reaction because they believe that the buyers that the initiator is charging an oligopolistic price will give the initiator a chance to rebid if its initial offer is beaten by a rival and the initiator will find it inherently profitable to beat any initially superior rival offer a rival would find profitable to make if that offer would be accepted by the buyer in question, the oligopolistic pricing is said to be “natural.” A seller’s decision not to make a QV investment, not to execute a PPR project, not to do research into (the quality of its product’s performance)/(the durability of its product)/(the lifetime cost of maintaining and repairing its product), and/or not to communicate what it knows about these matters to prospective buyers will be oligopolistic if its ex ante perception that these negative decisions are at least normally profitable is critically affected by its belief that at least one relevant rival’s response to such negative choices will or may be critically affected by that rival’s perception that the firm in question will react to its decisions (not to make)/(to make) expenditures of those kinds in ways that would render unprofitable for the rival in question decisions to make such expenditures that the rival would otherwise find profitable. (Such oligopolistic decisions can be either contrived or natural. For example, decisions by two rivals not to make an inherently profitable investment when both of their making an investment would be against each’s interest will be contrived-oligopolistic if the decisions not to invest was secured by an agreement in which each promised not to invest if the other did not invest or by threats of retaliation [by one or both of the firms in question indicating that it would react to the other’s investment by making inherently unprofitable decisions that would impose losses on the investor (e.g., decisions to charge the investor’s customers inherently-unprofitably-low prices or decisions to place inherently unprofitable advertisements that target the investor’s customers)].) By way of contrast, decisions by two rivals not to make an investment each would find profitable if it would not induce the other to invest when both of their making an investment would be against each’s interest will be natural-oligopolistic if each realizes that its making an investment will render it profitable for the other to make an investment that the other would not otherwise have found profitable because the other would have realized that its making an investment (that could not be profitably withdrawn) would make it profitable for the second firm to invest (since the second firm’s decision not to invest would no longer deter the investor from making the investment it had already made and would not induce the first investor to withdraw its investment)—that is, will be natural-oligopolistic when the decisions reflect the fact that the two firms confronted each other with what I denominate critical “natural-oligopolistic investment-disincentives” (disincentives equal to the loss each potential investor would anticipate its investment would impose on it not only directly but also by inducing its rival to make an investment the rival would not otherwise have made because the induced investment would not otherwise have been profitable for the rival).

The second category of antitrust-policy-policy-coverable conduct I want to define is “predatory conduct.” On my definition (which is also consistent with the overwhelming majority of economist-usages, though they have never formally defined the concept), a business entity’s choice is said to be “predatory” if its ex ante perception that the choice would be at least normally profitable was critically affected by its belief that the choice would or might increase its profits by driving an established rival out, (with an exception that will be delineated below) by inducing an established rival to substitute one or more investments that were less competitive with the perpetrator’s investments than the rival’s current investment(s) were, by deterring an established rival from making an additional investment that would be competitive with the perpetrator’s investment(s), (with the same exception) by inducing an established rival to substitute a new investment that would be less competitive with the perpetrator’s investment(s) for a new investment that would be more competitive with the perpetrator’s investment(s), and/or (with the same exception) by inducing a potential competitor to make an investment that was less competitive with the perpetrator’s investment(s) rather than an investment that would be more competitive with the perpetrator’s investment(s). Here is the exception: for reasons that relate to the definition of “conduct motivated by specific anticompetitive intent” that will be delineated below, conduct that induces rivals to substitute investments that are less competitive with the perpetrator’s investment(s) for investments that would be or were more competitive with the perpetrator’s investments are not deemed “predatory” if the investment-shift is economically efficient in that it increases economic efficiency by reducing the economically inefficient duplicativeness of the investments in the relevant portion of product-space.

A wide variety of business choices can be predatory. Business entities can practice predation by charging inherently-unprofitably-low prices to a target’s customers, by placing inherently unprofitable advertizements that are directed inter alia at a target’s customers, by making inherently unprofitable communications to its target’s customers that convey false or misleading positive information about the perpetrator’s product(s) or false or misleading negative information about the target’s product(s), by predatorily refusing to deal with a target or by entering into predatory full-requirements-purchasing or total-output-supply contracts (though I suspect that only a trivial percentage of choices of these latter kinds are predatory), by making a predatory QV investment or a predatory investment in PPR. (A QV investment or investment in PPR is predatory if the investor’s ex ante perception that it would be at least normally profitable was critically affected by its perception that it would or might increase its overall profits because the investment in question would be less competitive with its other investments than the investment its investment would deter would have been [and this investment-location shift is not economically efficient] and/or because the non-retaliatory responses the investor’s established rivals would make to its investment would reduce its other investments’ profit-yields by less than their non-retaliatory responses to the rival investment its investment would deter would have reduced its other investments’ profit-yields—in my terminology, because ex ante the predator perceived itself to have a critical “monopolistic investment-incentive” to make the investment in question.)

In my terminology (which I think is standard), conduct is said to “distort competition” if the conduct causes its perpetrator(s) to have competitive advantages over one or more rivals without causing the perpetrator(s) to be at least as allocatively better-placed than the disadvantaged rivals to supply relevant buyers. Conduct that distorts competition need not be predatory—need not be expected by its perpetrator(s) to drive the disadvantaged rival(s) out, to deter them from investing, or to induce them to decrease economic efficiency by substituting investments that are less competitive with the investment(s) of the perpetrator(s) for investments that are or would be more competitive with the investments of the perpetrator(s). Many types of conduct can distort competition. The relevant list includes blocking access to a rival’s place of business, communicating to potential customers inaccurate or misleading positive information about the distorter’s product(s)/credit-worthiness/(likely survival) or inaccurate or misleading negative information about the disadvantaged rival’s product(s)/credit-worthiness/(likely survival), establishing inaccurate quality standards that disadvantage inappropriately one or more rivals, bribing one or more of a target’s employees to perform poorly, providing inaccurate or misleading information about the target to potential employees or potential financers of the target, and so on.

C. Definitions of the Components of the Gap Between a Seller’s Price and Its Conventional Marginal Cost and Definitions of the Factors that Determine the Imperfectness of Investment-Competition in Any ARDEPPS and of Other Concepts That Play a Role in Investment-Competition Analysis

I start with the seller-price-competition-related concepts. The applicable conceptual systems vary to some extent with whether the sellers in question are setting “individualized prices” to their individual potential and actual customers or are setting “across-the-board prices,” which apply to all their potential customers. For simplicity, the account that follows ignores the possibilities that the price or prices a seller can obtain may be affected by errors it, its rivals, or the relevant buyers make.

In individualized-pricing contexts, the gap between the price a seller obtains from a buyer it is (privately) best-placed to supply (a buyer it can profit from supplying on terms contained in an offer that no rival will find inherently profitable to beat if the rival’s doing so would yield it a sale) equals the “basic competitive advantage” (BCA) it enjoys in its relations with the buyer plus “the contextual marginal costs the closest rival it faces for the relevant buyer’s patronage must incur to match the best-placed supplier’s highest non-oligopolistic-price-(HNOP)-containing offer to the buyer” plus the natural-oligopolistic margin (NOM) the seller can obtain from the buyer plus the contrived-oligopolistic margin (COM) the seller has obtained from the buyer. An individualized pricer’s BCA in its relations with a buyer it is best-placed to supply equals its buyer preference advantage (BPA) over its closest rival for this buyer’s patronage plus its marginal (or incremental) cost advantage (MCA) over this rival (where the sum of the BPA and MCA is positive though one of these figures may be negative—may be a disadvantage). The contextual marginal cost that the closest rival of a best-placed supplier must incur to match an HNOP-containing offer from the best-placed supplier (CMC_#2) equals the costs the second-placed supplier must incur to charge a price that will result in its offer’s matching the HNOP-containing offer of the best-placed supplier because #2’s matching-offer price is discriminatory and/or violates a minimum-price regulation. The gap between the across-the-board price that an across-the-board pricer will find most profitable to charge and its marginal costs is determined by the gap between its across-the-board HNOP and its marginal costs (which depends on the distribution of its BCAs and basic competitive disadvantages [BCDs] and on the amount by which its rivals’ prices exceed the respective rivals’ marginal costs) and on the NOMs and COMs the across-the-board pricer finds, respectively, possible and profitable to obtain (and does obtain). The increase in the price that an across-the-board pricer finds profitable to charge because its rivals are charging supra-marginal-cost prices is the across-the-board-pricing counterpart to CMC_#2 in the individualized-pricing context. The magnitude of this across-the-board-pricing counterpart to CMC_#2 depends in part on the order in which the relevant rivals lock themselves into announced prices (by advertizing them, posting them on shelves, entering them into computers, and making sales at the prices in question): this component of the across-the-board pricers’ (P−MC) gap will be larger if (roughly speaking) rivals that are second-placed or close-to-second-placed far more often than they are best-placed lock themselves into prices early in the producer-pricing sequence (since such firms will have more incentives to beat the offers of their rivals from a position in which they have BCDs).

My study uses three sets of concepts to explain the imperfectness of investment-competition in any ARDEPPS and to analyze the impact of exemplars of various categories of antitrust-policy-coverable conduct on the intensity of investment-competition in any ARDEPPS. The first set contains the profit-differential, risk, retaliation, and scale barriers to entry faced by the best-placed and various worse-than-best-placed potential entrants into an ARDEPPS and their counterparts for the established firms that are better-placed than any other established firm to make an investment in the ARDEPPS in question that would raise equilibrium investment in that ARDEPPS (on different assumptions—see below). The barriers in question are relevant because they reduce the post-investment lifetime supernormal rate-of-return the relevant investor will anticipate realizing on the most-supernormally-profitable investment it could make in the relevant ARDEPPS at the relevant point in time below the pre-investment lifetime supernormal rate-of-return that was generated by the most-supernormally-profitable investments in the ARDEPPS in question. The second set of investment-competition-analysis-related concepts contains the monopolistic investment-incentive, the monopolistic investment-disincentive, and natural-oligopolistic investment-disincentive an established firm may have to make an additional investment in its ARDEPPS, which refer, respectively, to (1) the amount by which the rate-of-return the prospective investor would realize on the investment in question would be increased because the investment would increase the profits-yields of the investor’s other investments in the ARDEPPS because it would deter a rival investment that would reduce those investments’ profit-yields more in that the deterred investment would be more competitive with the investor’s other projects than its own investment would be and/or would induce the investor’s rivals to respond in ways that reduced the investor’s other projects’ profit-yields by less than the responses they would have made to the deterred investment would have done, (2) the amount by which the rate-of-return the prospective investors would realize on the investment in question would be reduced because the investment would reduce the investor’s other investments’ profit-yields usually because it would not deter a rival investment and would on its own compete with the investor’s other investments and cause the investor’s rivals to respond in a way that reduces the investor’s other projects’ profit-yields but conceivably because the investor’s investment would reduce its other investments’ profit-yields more in these ways than they would have been reduced in these ways by the rival investment the investor’s investment would deter, and (3) the amount by which the rate-of-return the prospective investor would realize on the investment in question would be reduced because the investor’s investment would reduce its other investments’ profit-yields not only directly in the above two ways but also by inducing a rival to make an investment in the ARDEPPS in question the rival would not otherwise have made by changing the situation so that the rival’s decision not to invest would not deter the investor from making the investment it had already made. The third set of concepts I use to explain the imperfectness of investment-competition in an ARDEPPS and to analyze the possible effects of any exemplar of antitrust-policy-coverable conduct on the intensity of investment-competition in an ARDEPPS references two ARDEPPS-investment quantities: (1) the entry-preventing investment-quantity in the ARDEPPS and (2) the entry-barred established-firm-investment-deterring investment-quantity in the ARDEPPS. I will not go into more detail here, but the analysis of the impact of potential competition or any relevant exemplar of antitrust-policy-coverable conduct on the intensity of investment-competition in an ARDEPPS will depend inter alia on whether the entry-preventing investment-quantity in the ARDEPPS equals, exceeds, or is lower than the entry-barred established-firm-investment-deterring investment-quantity in that ARDEPPS.

D. The Definitions of “Specific Anticompetitive Intent” and “Lessening Competition” That My Study Employs (and That I Think Are Legally Correct in the United States and the EU)

I believe that the “specific anticompetitive intent” test of antitrust illegality is promulgated by Sections 1 and 2 of the Sherman Act and Articles 101 and 102 of the 2009 Treaty of Lisbon. ⁷ On my understanding, a business actor’s choice manifests its “specific anticompetitive intent” if ex ante the actor would not have found the choice to be at least normally profitable had it not believed the choice would or might yield it profits by reducing the absolute attractiveness of the best offers against which it would have to compete in one or more ways that would render the conduct profitable even though it would be economically inefficient in an otherwise-Pareto-perfect economy. ⁸

In my judgment, with one possible exception that I believe is empirically unimportant, choices that are motivated by specific anticompetitive intent are liberal-moral-rights-violative usually because they impose a net equivalent-dollar mere-utility-related loss on the customers of the chooser and the customers of the chooser’s product-rivals “combined” that exceed the equivalent-dollar gains the choices confer on their maker. The possible exception is related to the possibilities that (1) the relevant economy’s other Pareto imperfections would cause an antitrust-policy-coverable choice to confer a net equivalent-dollar gain on relevant creatures other than the chooser and its traditional victims and in so doing would render not economically inefficient a relevant choice that would otherwise have been economically inefficient or would cause an antitrust-policy-coverable choice to impose a net equivalent-dollar loss on relevant creatures other than the chooser and its traditional victims and in so doing would render the choice economically inefficient even though it would not otherwise be economically inefficient and (2) the relevant chooser was aware of these realities ex ante or would have been aware of these realities had it done the research into these possibilities that liberalism obligates it to do. Although I realize that the following conclusion is in one sense self-serving, I doubt that these two conditions will be fulfilled in more than a trivial number of instances. That is why my study does not address these possibilities when considering the liberal-moral-rights-violativeness of exemplars of various categories of antitrust-policy-coverable conduct and the antitrust-policy-related moral obligations of a government committed to instantiating the liberal conception of justice.

I believe that the “lessening competition” test of illegality is promulgated by the U.S. Clayton Act, ⁹ Article 101 of the 2009 Treaty of Lisbon, ¹⁰ and the European Merger Control Regulation. ¹¹ My study argues that the legally correct way to define the concept of “lessening competition” when interpreting and applying the U.S. Clayton Act, the concepts of “preventing or restricting” competition” when interpreting and applying Article 101 of the 2009 Treaty of Lisbon, and the concept of “impeding effective competition” when interpreting and applying the European Merger Control Regulation is to conclude that covered conduct has lessened, prevented or restricted, or impeded effective competition if the conduct imposed a net equivalent-monetary loss on the actual and potential customers of the perpetrator(s) and the actual and potential customers of the perpetrator’s or perpetrators’ product-rivals by worsening the absolute attractiveness of the best offers they, respectively, received from worse-than-best-placed suppliers. Although the following position is more contestable because it cannot be justified by any text in the relevant statute, treaty, or regulation, I think it is correct as a matter of law and morally desirable as a matter of policy to allow the defendant(s) to exonerate itself (themselves) by demonstrating that the conduct at issue would not have imposed an equivalent-monetary loss on the relevant buyers had it not increased the perpetrator’s or perpetrators’ organizational economic efficiency and thereby improved the perpetrator’s/perpetrators’ array(s) of competitive positions and concomitantly worsened one or more rivals’ array(s) of competitive positions sufficiently to cause it or them to exit or worsened one or more potential investors’ arrays of prospective competitive positions on their investment(s) sufficiently to render ineffective one or more potential rival investors that would otherwise have been effective potential entrants/expanders.

E. The Definition of Economic-Efficiency-Analysis-Related Concepts

The first economic-efficiency-analysis-relevant concept I want to define is the concept of “the impact of a choice on economic efficiency.” Two criteria should be used to evaluate such concepts as “the impact of a choice on economic efficiency”: the conformity of the definition to professional and popular intuitive understanding and the ability of the concept as defined to contribute to the analyses in which it plays a role. According to the definition that best satisfies these criteria, the impact of a choice on economic efficiency equals the difference between (1) the number of dollars (monetary units) that would have to be transferred to its winners to leave them as well-off as the choice would leave them if (A) they did not agree to the transfer, (B) they were either indifferent to the substitution of the transfer for the choice or were unaware of the linkage between the transfer and the relevant choice’s rejection, (C) their distributive attitudes toward such transfers and normative distributive commitments gave them no reason to prefer the transfer to the choice or vice versa, and (D) the financing of the transfer would not benefit or harm them indirectly by changing the conduct of others by altering their respective income/wealth positions or by satisfying or dissatisfying the winners’ “external preferences” (for or against specified others’ obtaining resources or having opportunities) and (2) the number of dollars that would have to be withdrawn from the choice’s losers (on analogous assumptions) to leave them as well-off as the choice would leave them. I should indicate that no other economist supports this definition, which uses a correctly elaborated version of what economists denominate “the monetized equivalent-variation” measures of a choice’s winners’ equivalent-dollar gains and choice’s losers’ equivalent-dollar losses. All other economists use the definition of this concept that is implicit in their use of the Kaldor–Hicks test for the economic efficiency of a choice, the so-called Scitovsky test for the economic efficiency of a choice, or the potentially Pareto-superior definition of an economically efficient choice, all of which are incorrect. ¹²

The second economic-efficiency-analysis-relevant concept I want to define is the concept of a category of resource-use. Resources can be used to implement a particular production process that is being employed to produce one or more units of an existing product, to create a QV investment, to execute a PPR project, to discover facts about the quality of the performance of a particular product, the durability of a particular product, the lifetime cost of maintaining and repairing a particular product, the externalities generated by the production and/or consumption of successive units of a particular product, the errors relevant potential and actual buyers make when estimating the dollar-value to them of a particular product or the lifetime cost to them of a particular product, to implement a particular pricing-technique, to practice various forms of oligopolistic/predatory/competition-distorting conduct, to consummate horizontal/conglomerate/vertical mergers or acquisitions, to execute horizontal/conglomerate/vertical joint ventures, to engage in horizontal/conglomerate/vertical internal growth, to influence the decisions of independent distributors and input-suppliers (by making pricing-decisions, by making decisions to include various kind of clauses in distributorship or input-supplier contracts, and/or by adopting particular sales policies), to conceal conduct that is or may be falsely deemed to be illegal, to defend oneself in trials or related legal negotiations, to develop and implement antitrust policies or other types of policies, to finance government policies in various ways, and so on.

The third economic-efficiency-analysis-relevant “concept” I want to define is a trilogy of related concepts—the concept of “a resource allocation,” the concept of a category of resource allocations, and the concept of a category of resource misallocation (economic inefficiency). Resource allocations involve the withdrawal of resources from one or more specific uses and their devotion to a particular use. Different categories of resource allocations can be distinguished by the functions from which each withdraws resources and the function to which each devotes resources. For each category of resource allocations, there is an associated category of resource misallocation (economic inefficiency). In part, because I think it is ex ante economically efficient to study separately the economic efficiency of the various separate resource allocations that are components of resource allocations from more than one category of use to one category of use, I believe it is ex ante economically efficient to distinguish and analyze the economic efficiency of the following kinds of resource allocations (and related categories of resource misallocation [economic inefficiency]): inter alia, inter-ARDEPPS and intra-ARDEPPS UO-to-UO resource allocations and misallocations, inter-ARDEPPS and intra-ARDEPPS QV-to-QV resource allocations and misallocations, inter-ARDEPPS and intra-ARDEPPS PPR-to-PPR resource allocations and misallocations, UO-to-QV and QV-to-UO resource allocations and misallocations, UO-to-PPR and PPR-to-UO resource allocations and misallocations, PPR-to-QV and QV-to-PPR resource allocations and misallocations, allocations and misallocations of resources from the use of one known production process to the use of another known production process, allocations and misallocations of a final good to one consumer rather than to another consumer, allocations and misallocations of time among various uses (leisure-production/consumption, the performance of UO-producing market labor, the performance of QV-investment-creating market labor, the performance of PPR-executing market labor, the performance of non-market legitimate labor, the performance of criminal acts), consumption-allocations/misallocations between future and current consumption, allocations or misallocations of resources among the various pricing-techniques that could be used or among the various techniques an imperfectly-vertically-integrated firm could use to control independent contractors, allocations and misallocations of resources among the various techniques vertically integrated firms can use to control their employees, allocations and misallocations of resources to the consummation/execution of various kinds of (M&A)s/joint ventures/internal growth, allocations or misallocations of resources to the design, implementing, and financing of various government policies.

The fourth economic-efficiency-analysis-related concept I want to define is the concept of a Pareto imperfection. Seven categories of Pareto imperfections are identified: imperfections in seller-competition, imperfections in buyer-competition, (real) externalities, taxes on the margin of income, resource-allocator “non-sovereignty” (ignorance of facts that affect the advantageousness to the resource allocator of a resource-use it could execute), resource-allocator non-maximization (resource-allocator failures to make the choice that the information at its disposal implies would best serve the resource allocator’s interest on the resource-allocator understanding of its interests), and buyer surplus (which I substitute for the standard “public good” imperfection). The defining characteristic of Pareto imperfections is that in an otherwise-Pareto-perfect economy an individual exemplar of any one of them will either cause or tend to cause economic inefficiency.

The fifth economic-efficiency-analysis-related “thing” I want to define is a theorem, not a concept. On my construction, The General Theory of Second Best ¹³ holds that, because the various Pareto imperfections an economy may contain (each of which would cause or tend to cause economic inefficiency in an otherwise-Pareto-perfect economy) will in general be as likely to counteract as to compound each others’ misallocative tendencies, one cannot accurately predict whether a choice will increase or decrease the amount of economic efficiency generated in a particular economy by determining whether it will decrease or increase the Pareto-imperfectness of that economy (put equally crudely, whether it will increase or decrease the number of and magnitudes of the Pareto imperfections the economy contains).

The sixth economic-efficiency-analysis-related concept I want to define is the concept of “a distortion” in a private-benefit, private-cost, or profit figure. In my terminology, any such private figure is “distorted” if it differs from its allocative counterpart. More specifically, I write that a private figure is “inflated” if it is higher than its allocative counterpart and is “deflated” if it is lower than its allocative counterpart. The “distortion in the profits a resource allocation yields the actor to which the resource allocation allocates resources” equals the difference between the profits the resource allocation yields the actor to which the resource allocation allocates resources and the economic efficiency of the resource allocation in question (which equals the distortion in the private benefits the actor obtains by using the allocated resources in the way the actor uses them minus the distortion in the private cost to the actor to which the resource allocation allocates resources of the resources the resource allocation allocated to that actor). To see why economic-efficiency analysis is concerned with profit-distortions, focus on resource allocations to UO-producing, QV-creating, and PPR-executing uses and note that (1) if the last exemplar of a particular category of resource allocations to a particular ARDEPPS yields zero profits (say, when the executed resource-use is the production of a marginal unit of an existing product) or zero supernormal profits (say, when the executed resource-use is the creation and use of a QV investment or the execution of a PPR project and use of the discovery it yields), the conclusion that the profits were inflated implies that the resource allocation in question was economically inefficient (reduced economic efficiency by the absolute value of the prevailing distortion) and suggests that various intra-marginal resource allocations in the relevant category to the relevant ARDEPPS may also have been economically inefficient and (2) if the last exemplar of a particular category of resource allocations to a particular ARDEPPS yields zero profits, the conclusion that those profits were deflated implies that the resource allocation in question was economically efficient (increased economic efficiency by the absolute value of the deflation) and suggests that additional resource allocations in the relevant category to the relevant ARDEPPS may have been economically efficient (that from the perspective of economic efficiency the relevant economy devoted too few resources to the relevant category of resource allocations to the relevant ARDEPPS). Section 2 outlines the (in my view) ex ante economically efficient (distortion-analysis) protocol for economic-efficiency prediction that builds on the preceding conclusions—that is, that is not “first-best” because first-best economic-efficiency analyses ignore the central insight of The General Theory of Second Best, is not second-best in that second-best economic-efficiency analyses ignore the allocative cost of doing and financing the theoretical and empirical work that would improve the accuracy of the analyst’s economic-efficiency prediction, but is third-best in that it takes into account The General Theory of Second Best, the contribution that additional research and known-information consideration can make to economic efficiency by increasing the economic efficiency of government choices, and the allocative cost both of doing and of financing the relevant work. ¹⁴

II. The Protocol My Study Uses to Predict the Economic Efficiency of Any Exemplar of Antitrust-Policy-Coverable Conduct or of Any Antitrust Policy

Section 1 explained that the impact that an exemplar of antitrust-policy-coverable conduct or an exemplar of antitrust policy has on economic efficiency is relevant to its serving or disserving the instantiation of any conception of the moral good that values the total or average utility of creatures whose utility it values (though the fact that a choice increased/decreased economic efficiency does not guarantee that it increased/decreased the total or average utility of such creatures). Section 1 also raised the possibility that the economic efficiency of a choice that imposes only mere-utility losses on its traditional victims might be relevant to its liberal-moral-rights-violativeness if (1) the interaction of the relevant economy’s Pareto imperfections (A) caused the choice to confer net equivalent-dollar gains on relevant creatures other than the chooser and its traditional victims that rendered economically efficient a choice that would otherwise have been economically inefficient or (B) caused the choice to impose net equivalent-dollar losses on relevant creatures other than the chooser and its traditional victims that rendered economically inefficient a choice that would otherwise have been economically efficient and (2) (A) the chooser was aware of one of these realities before making the choice or (B) would have recognized one of these realities before making the choice had it done the research into these possibilities that liberalism obligates it to do. Certainly on the former of these accounts and possibly on the latter, antitrust-policy analysts should investigate the economic efficiency of the conduct and policies they are called upon to evaluate.

Unfortunately, the economic-efficiency-prediction protocol that antitrust-policy analysts use is deficient in at least five respects. First, the conventional protocol ignores many of the categories of economic inefficiency whose magnitudes antitrust-policy-coverable conduct and antitrust policies can affect—for example, the economic inefficiency that results because (1) given the amount of resources the relevant economy devotes to QV-investment creation, a net equivalent-dollar gain could have been generated had more resources been devoted to QV-investment creation in some ARDEPPSes and fewer resources been devoted to QV-investment creation in other ARDEPPSes; (2) given the amount of resources the relevant economy devotes to PPR, a net equivalent-dollar gain could have been generated had more resources been devoted to PPR in some ARDEPPSes and fewer resources been devoted to PPR in other ARDEPPSes; (3) given the total amount of resources the relevant economy devotes to the production of units of existing products, to QV-investment creation, and to PPR, a net equivalent-dollar gain could have been generated had more resources been allocated to one or two of these categories of use and fewer resources been allocated to one or two of these categories of use.

Second, conventional antitrust-policy-related economic-efficiency analyses (1) pay insufficient attention to the various ways in which antitrust-policy-coverable conduct and antitrust policies can increase or decrease private transaction costs (the private-transaction-costliness of the pricing-techniques that are used, the private transaction costs generated by oligopolistic/predatory/competition-distorting conduct of all kinds, the private transaction costs of negotiating [M&A]s and joint ventures and of coordinating their participants’ conduct, the private transaction costs that non-vertically integrated producers incur to use contractual clauses and sales policies to induce their independent input-suppliers and independent distributors to make the choices that are most profitable for the producers and the private transaction costs that vertically integrated producers incur to use hierarchical controls to influence the decisions of their input-producing and distributive employees, the private transaction-cost effect that antitrust policies have by deterring private-transaction-costly conduct, by inducing the law’s addressees to conceal their illegal conduct in private-transaction-costly ways, by inducing the law’s addressees to “defend” themselves in related law-enforcement negotiations and litigation, by involving the government’s making private-transaction-costly efforts to implement its policies [to determine whether particular parties have broken the law, to negotiate plea-deals with suspects, to litigate cases, to operate the courts, to run the prison and parole systems, and to finance the policy in question, etc.]) and (2) ignore the difference between the private transaction-cost effects in question and their allocative counterparts (a failing that is associated with the third deficiency of the conventional economic-efficiency-analysis protocol—its ignoring The General Theory of Second Best [its first-best character]).

The third deficiency of conventional (antitrust-policy) economic-efficiency analyses is their first-best character—that is, is (1) their assumption that any choice that increases/decreases the magnitude and/or incidence of any type of Pareto imperfection (which would generate or would tend to generate economic inefficiency in an otherwise-Pareto-perfect economy) will increase/decrease the amount of economic inefficiency that is generated in the relevant economy and (2) their related failure to use a protocol for analyzing the impact of any antitrust-policy-coverable choice or any antitrust policy on the amount of economic efficiency in any category that is generated that takes appropriate account inter alia of the fact that the individual Pareto imperfections that cause or tend to cause any category of economic inefficiency in an otherwise-Pareto-perfect economy are (roughly speaking) as likely to counteract as to compound each other’s misallocative tendencies.

The fourth deficiency of conventional antitrust-policy economy-efficiency analysis is that it ignores the economic-efficiency cost of financing any antitrust policy (not only the allocative transaction cost of the policy’s financing if the policy is financed by raising taxes or selling government-supplied goods or services that would otherwise have been made freely available but also any economic inefficiency any related tax increases or government-supplied-product/service pricing would generate even if the policy’s design and implementation did not generate any allocative transaction costs).

The fifth deficiency of the conventional antitrust-policy economic-efficiency-prediction protocol is that it does not instruct the analyst to do relevant work if and only if it is ex ante economically efficient to do the work and, relatedly, does not provide the analyst with any guidance about how to determine the ex ante economic efficiency of executing particular relevant theoretical or empirical research projects or of devoting additional time to considering the economic-efficiency implications of what is already known. In my terminology, the fifth deficiency of the conventional antitrust-policy economic-efficiency-prediction protocol is that it is not “third-best-economically-efficient.”

The antitrust-conduct/policy economic-efficiency-prediction protocol my antitrust-policy study uses is a protocol on which I have been working for many years that I fully developed in my 2020 study Welfare Economics and Second-Best Theory: A Distortion-Analysis Protocol for Economic-Efficiency Prediction. ¹⁵ I do not have the space to describe this protocol in great detail here, but I will outline its twelve components.

The first component is an instruction to the economic-efficiency analyst to do all work but only that work that is ex ante economically efficient. The second component lists the various categories of resource allocations any economy can execute, the related categories of economic inefficiency any economy can generate, and the various categories of choices that can affect the amount of private and allocative transaction costs that can be generated in a given economy. The third component derives (1) mathematical formulas that indicate separately for each category of resource allocations the way in which the relevant Pareto imperfections and sometimes other parameters interact to distort the profits that exemplars of the category of resource allocations in question will yield the actor to which the resource allocation allocates resources and (2) a mathematical formula that indicates the way in which the relevant Pareto imperfections and sometimes other parameters interact to distort the private-transaction-cost effects of various choices. The fourth component instructs the analyst to select a third-best-economically-efficiently-large, random sample of ARDEPPSes on which the economic-efficiency analysis in question will initially focus. The fifth component instructs the analyst to estimate the relevant policy’s effects on the amount of resources devoted to each category of resource allocations in the ARDEPPSes to be studied—the resource allocations in each category the policy will deter in each studied ARDEPPS and the resource allocations in each category the policy will elicit in the studied ARDEPPSes. The sixth component instructs the analyst to estimate the profits that the resource allocations the policy will elicit in the ARDEPPSes to be studied will yield the actors to which those policy-elicited resource allocations allocate resources and the losses that the resource allocations the policy will deter in the ARDEPPSes to be studied would have imposed on the actors the policy caused not to be allocated resources. The seventh component instructs the analyst to estimate the magnitudes of the parameters in the formulas for the distortions in the profits that were yielded or would be yielded by exemplars of each category of resource allocations in the ARDEPPSes to be studied. The eighth component instructs the analyst to combine the conclusions of the fifth, sixth, and seventh components to estimate the impact that the policy would have on the amount of economic inefficiency in each category generated in the studied ARDEPPSes—to estimate the economic efficiency or economic inefficiency of the elicited resource allocations in each category in each ARDEPPS and the potential economic efficiency or economic inefficiency of the resource allocations in each category the policy would deter in each studied ARDEPPS. The ninth component instructs the analyst (1) to estimate the possibly different ratios of the allocative value of the resources devoted to each category of resource allocations in the whole economy to the allocative value of the resources devoted to that category of resource allocations in the studied ARDEPPSes, (2) to multiply the estimate of the economic-efficiency impacts that the policy had by deterring and eliciting resource allocations in the relevant category in the studied ARDEPPSes by the relevant ratio just delineated to generate an estimate of the policy’s impact on the amount of economic inefficiency in the relevant category the economy generated by executing economically inefficient resource allocations in that category and by failing to execute economically efficient resource allocations in that category, and (3) to add together the conclusions reached for each category of resource allocations about the policy’s economy-wide impact on the amount of misallocation in the associated category the economy generated by executing economically inefficient resource allocations in that category and by failing to execute economically efficient resource allocations in that category to yield an estimate of the policy’s impact on the total amount of economic inefficiency the relevant economy generated by executing economically inefficient resource allocations in all categories and failing to execute economically efficient resource allocations in all categories. The tenth component instructs the analyst to estimate the allocative transaction costs the design and implementation of the policy would generate if its implementation’s financing would not generate any allocative transaction costs—that is, to estimate (1) the policy-design allocative transaction costs by estimating the private policy-design costs and the parameters in the formula for the ratio of the associated allocative to private transaction cost and multiplying the estimated policy-design private transaction costs by the above ratio (calculated by combining the formula with the estimated value of the parameters it contains), (2) the allocative transaction costs the policy’s implementation will involve the government’s generating “directly” (by estimating the private implementation costs the government will have to incur to implement the policy throughout the whole economy and the magnitude of the associated allocative transaction costs to their private counterparts [by combining the formula for that ratio with estimates of the magnitudes of its parameters]), and (3) the impact that the policy’s implementation will have on the allocative transaction costs the policy’s addressees and potential violators of other antitrust laws will generate (a net effect that could be positive or negative depending inter alia on the relevant extent to which the policy’s implementation deters allocative-transaction-costly violations, induces its addressees to conceal their illegal behavior in allocative-transaction-costly ways, causes accused and guilty violators to generate allocative transaction costs to defend themselves or try to obtain lesser sanctions)—that is, by estimating the policy’s impact on the private transaction costs potential and actual violators generate and multiplying that estimate by an estimate of the ratio of the allocative-transaction-cost effect to the private-transaction-cost effect (a ratio derived by combining the relevant formula with estimates of the parameters it contains). The eleventh component instructs the analyst to estimate the economic-efficiency cost of the financing of the policy in question, which depends both on the allocative transaction costs the financing generates and the extent to which the tax-increases, increases in the prices the government charges for goods and services it supplies, canceled expenditures, and national-debt increases used to finance the policy’s design and implementation increase or decrease economic efficiency, allocative-transaction-cost consequences aside (a task that requires the analyst to estimate the proportions of the relevant financing that will be secured by these various methods, the private and allocative transaction-cost consequences of using each of these methods, the economic inefficiency or economic efficiency of the tax increases or price increases that are used to finance the design and implement the relevant policy, the economic inefficiency or economic efficiency of the canceled expenditures, and the economic inefficiency of any associated increase in the national debt). I could say more but will not. The twelfth component of the antitrust-policy economic-efficiency-prediction protocol my study uses instructs the analyst to combine the economic-efficiency conclusions of the ninth through eleventh components to yield an overall estimate of the impact of the policy in question on economic efficiency.

To end where I began. The protocol I use instructs the analyst to devote resources to executing the second-through-twelfth components described when but only when it is ex ante economically efficient to do it. Two additional related points should be made. First, to follow this instruction, the policy-analyst must continuously revise his or her estimate not only of the allocative cost of further work but also of the allocative value of the information the work might discover (e.g., how likely it is to change any conclusion one reached about the economic efficiency of the policy at issue or of any other policy in a way that increases economic efficiency by changing the identity of the policy-choice that is made). Second, as the parenthetical in the preceding sentence indicates, this instruction requires the analyst to take account not only of the economic-efficiency gains that work that is relevant to the prediction of the economic efficiency of the policy at issue will yield by improving decisions related to that policy but also of any economic-efficiency gains the work in question will generate by discovering information that will increase the economic efficiency of other decisions the government will make.

I will close this section with two economic-efficiency or sets of related economic-efficiency conclusions that affect my predictions of the economic efficiency of many exemplars of antitrust-policy-coverable conduct and many exemplars of antitrust policy. The first conclusion is that choices that reduce/increase poverty tend to increase/decrease economic efficiency on that account. This conclusion reflects the facts that poverty generates economic inefficiency (1) by causing too few resources to be devoted to investment to the human capital of the children of the poor and to the human capital of poor adults from the perspective of economic efficiency; (2) by leading the poor to make purchasing choices that are against their respective individual interests and on that account are economically inefficient by causing the poor to be misinformed about the performance-attributes/durability/lifetime-maintenance-and-repair cost of relevant products and to make purchases that the information they possess should lead them to conclude would not be in their parochial interest (say, to misestimate the price per ounce of products); (3) by causing poor individuals to make purchasing choices that are economically inefficient despite the fact that they are in the purchaser’s parochial interest because the consumption of the product whose purchase is most advantageous to the purchaser generates critically more external costs than the consumption of alternative products would have generated (by making it advantageous to the poor purchaser to buy and drive a more rather than less ugly, accident-prone, breakdown-prone car and to rent a more rather than less ugly, disease-and-fire-spreading housing-unit); (4) by causing poor individuals to cause economic inefficiency by committing crimes (by alienating the person who is poor and making him or her less concerned with the loss the crime imposes on its victim[s], by reducing the difference between the experience of being imprisoned in any particular facility and the experience of not being imprisoned but living with the financial constraints that would apply to the individual if he or she did not commit a crime, by increasing the probability that non-sovereignty and non-maximization would lead the individual to commit a crime when doing so was not in its parochial interest); (5) if medical care and food is subsidized for the impoverished members of the relevant polity, by making it advantageous for poor individuals to make economically inefficient decisions to do perilous work; (6) if as I assume the political influence of individuals increases with the expenditures they can make on political parties/campaigns/candidates and on participation in legislative, administrative, and adjudicative decision-making processes, by increasing the amount of economic inefficiency the government generates by making economically inefficient choices that confer equivalent-dollar gains on the better-off and impose equivalent-dollar losses on the poor, and so on. Although this conclusion is contestable, I believe that tax and spending policies that reduce poverty increase economic efficiency despite any tendency that taxes on the better-off have to cause economic inefficiency by deterring them from performing economically efficient market labor and from making economically efficient decisions to save and invest. (The probability that the last conclusion is correct will be higher if as I will assert next the wage of QV-investment-creating workers is inflated and the profit-yields of QV investments are inflated, at least if QV-investment-creating workers and the owners of QV investments are materially better-off than the average member of the relevant polity.)

The second set of antitrust-policy-relevant economic-efficiency conclusions for which I have argued (most recently, in detail, in my study Welfare Economics and Second-Best Theory) is that (1) the profits generated by the production of units of output with resources withdrawn exclusively from QV-investment creation-and-use and PPR execution-and-use are deflated, (2) the profits generated by the creation and use of QV investments with resources withdrawn exclusively from UO-production-and-consumption and PPR execution-and-use are inflated, and (3) the profits generated by the execution and use of PPR projects with resources withdrawn exclusively from UO production-and-consumption and QV-investment creation and use are deflated—that is, at least if the relevant resource-users are sovereign maximizers, given the total allocative value of the resources any relevant economy devotes to the production and consumption of units of existing products, to QV-investment creation and use, and to PPR execution and use, from the perspective of economic efficiency, too many resources are devoted to QV-investment creation and use and not enough to the production and consumption of units of existing products and to the execution and use of PPR. (I acknowledge that this conclusion is more contestable for PPR and for QV investments that are predicted to discover scientific, technological, or commercial knowledge, but I believe that the above conclusions are also correct for such QV investments and for PPR.)

III. The Non-Market-Oriented Character of the Antitrust-Policy-Analysis Protocols My Study Develops and Applies

The antitrust-policy-analysis protocols my study develops and applies are not market-oriented—that is, do not deem relevant the magnitude of any market-aggregated parameter such as (1) the share(s) that actual or alleged perpetrators have of the sales made in an allegedly relevant market or the share(s) they own of the investment in an allegedly relevant market or (2) the concentration of the seller-side of any allegedly relevant market (whether measured in the traditional way by the total share [of market-sales or market-investment] of the leading four or eight firms in the allegedly relevant market or in the more modern way by the sum of the squares of the market-sales shares of all firms placed in the allegedly relevant market—by that market’s Hirschman–Herfindahl Index [HHI]). Admittedly, my analysis of the impact of exemplars of antitrust-policy-coverable conduct or antitrust policies on investment-competition and, relatedly, on economic efficiency do refer to four categories of barriers to investment that specified potential investors may face at particular points in time and do define those individual barriers to refer, respectively, to particular (sets of) factors that reduce the post-investment supernormal profit rate that would be yielded by the most profitable investment the relevant potential investor could make if the investment in question would increase total investment in the ARDEPPS in which it was made to a quantity that would become the ARDEPPS’s equilibrium investment quantity below the lifetime supernormal profit rate that the ARDEPPS’s most-supernormally-profitable investments would yield if the ARDEPPS’s equilibrium investment quantity were the quantity of investment it contained before the investment in question was made. However, no economic-efficiency or moral-desirability conclusion (and no legal conclusion) I reach is critically affected by the way in which any ARDEPPS is defined. In short, my analytic protocols do not make any critical use of the magnitudes of any market-aggregated parameters or suppose that any allegedly relevant market can be defined correctly or non-arbitrarily.

This section justifies this feature of my antitrust-policy-analysis (and antitrust-law-analysis) protocols. The section has three subsections. Subsection 3A provides the most important justification—explains why the definitions of any traditional, “ideal economic market” and the definition of any “antitrust market” are inevitably arbitrary not just at their peripheries but comprehensively. Subsection 3B justifies my rejection of market-oriented protocols for determining whether a particular firm has engaged in contrived-oligopolistic individualized pricing, natural-oligopolistic pricing, predatory pricing, or competition-distorting conduct (all of which are liberal-moral-rights-violative) by explaining why—even if markets could be defined non-arbitrarily—it would be neither accurate nor morally defensible to use estimates of an alleged perpetrator’s share of the allegedly relevant market and/or estimates of the concentration of the seller-side of the allegedly relevant market to determine whether an alleged perpetrator did engage in any of these categories of conduct. Subsection 3C justifies my rejection of using market-oriented protocols to resolve antitrust-policy-related issues by explaining why it would be both inaccurate and morally undesirable to use estimates of (M or A)-participants’ shares of the markets in which they are deemed to be operating and estimates of the concentration of the seller-sides of those markets to decide (1) whether an actual/proposed (M or A) was/(would be) liberal-moral-rights-violative and did/would disserve the instantiation of any egalitarian conception of the moral good or (2) whether any actual/proposed (M or A)-related antitrust policy did/would fulfill the (M or A)-related moral obligations of the government of a polity committed to instantiating the liberal conception of justice or did/would promote the instantiation of the various egalitarian conceptions of the moral good.

A. The Inevitable, Comprehensive Arbitrariness of Definitions of Ideal Economic Markets and of Definitions of Antitrust Markets

Economists distinguish two categories of “market concepts”—(1) what they call “economic markets” and I call “ideal economic markets” and (2) what economists call “antitrust markets.” Traditionally, economists’ discussions of “market definitions” focused on “ideal economic markets”—that is, on economic markets whose definition purportedly satisfies three assumptions about the competitiveness of all products placed inside an economic market and the difference between the competitiveness of all product-pairs in an economic market and the competitiveness of any product in a particular economic market and any other product not placed in the market in question: (1) all pairs of products placed in a given economic market are highly competitive with each other; (2) all pairs of products placed in a given economic market are equally competitive with each other; and (3) any product placed in a given economic market is more competitive with any other product in that market than it is with any product not placed in that market. Unfortunately, for seven reasons, definitions of ideal economic markets are inevitably arbitrary not just at their peripheries but comprehensively. For simplicity, the following list will refer to products’ being equal-best-placed, best-placed, second-placed, and close-to-second-placed rather than to producers’ occupying these competitive positions:

1. There is no non-arbitrary way to determine the quantity of sales that must be made of a set of products for that set of products and its sellers and buyers to be said to constitute an economic market;

2. There is no non-arbitrary way to define “product-pair competitiveness”: the concept could be defined by (A) the sum of the dollar-sales that both members of the product-pair are privately-equal-best-placed to make and the dollar-sales that the members of a product-pair are, respectively, privately-best-placed and privately-second-placed to make, (B) (the dollar-sales that the two members of a product-pair are equal-best-placed to make times the average amount by which in these cases the two products are better-placed than the respective third-placed suppliers of the respective buyers in question) plus (the dollar-sales for which the two members of a product-pair are privately-best-placed and privately-second-placed times the average amount by which in those cases the privately-second-placed product is better-placed than the privately-third-placed product), (C) “(A)” or “(B)” above and the dollar-sales that one member of the product-pair is best-placed to make and the other is second-placed or close-to-second-placed (which will affect the extent to which the second product will reduce the profits the producer of the first product can realize by engaging in horizontal price-fixing);

3. Even if one could define the competitiveness of each pair of products, there would be no non-arbitrary way to define the competitiveness of all pairs of products: there would be no non-arbitrary way to decide whether the metric should focus exclusively on the mean of the distribution of all product-pair-competitiveness scores or should take into account the dispersion of that distribution and if it should take account of that dispersion how that dispersion should be measured and taken into account;

4. Even if there were a non-arbitrary way to measure the competitiveness of all product-pairs, there would be no non-arbitrary way to define the equality of all product-pair-competitiveness scores: there would be no non-arbitrary way to decide whether the equality should be determined solely by the mean deviation of the distribution of the product-pair-competitiveness scores or whether and if so in what way more weight should be given to scores that deviate from the mean by a larger-than-mean amount;

5. Even if there were a non-arbitrary metric for product-pair competitiveness, there would be no non-arbitrary way to measure the difference between the competitiveness of market-insider product-pairs and the competitiveness of various market-insider products with the market-outsider product with which they are, respectively, more competitive than they, respectively, are with one or more products placed in the market in which they are placed;

6. Even if one could give non-arbitrary scores to all possible market definitions’ satisfaction of the three criteria that market definitions should ideally satisfy, there would be no non-arbitrary way to pick the market definition that is most ideal when no market definition best satisfies each of the three criteria; and least importantly

7. Even if one could resolve the sixth difficulty, there would be no non-arbitrary way to resolve ties.

Starting in the 1950s, at least some economists became interested in another type of market and related market definition—what they denominated “antitrust markets.” Antitrust-market definitions were to be assessed not by whether the markets they defined were “ideal” in the sense in which I have just defined that adjective but by their functionality—by whether their definition and estimates of the magnitudes of the market-aggregated parameters with which they are associated would contribute to the execution of the most-legally-correct or most-morally-desirable resolution of antitrust-law or antitrust-policy issues. (The preceding sentence’s lack of specificity is attributable to the fact that antitrust-market advocates do not distinguish between antitrust-law analysis and antitrust-policy analysis and do not specify the criteria that should be used to evaluate the performance of an antitrust-law-analysis protocol or an antitrust-policy-analysis protocol.)

I will now explain why antitrust markets cannot be defined non-arbitrarily. The primary reason is that, regardless of the morally defensible criterion one uses to evaluate the performance of any existing-antitrust-law-application protocol or any antitrust-policy-evaluation protocol, no such protocol in which the magnitude of any market-aggregated parameter plays a role can be economically efficient or morally desirable because the magnitude of the non-market-aggregated parameters one uses to define any antitrust market are better predictors of the legally and morally relevant motivation and consequences of any antitrust-law/policy-covered conduct and the morally relevant consequences of any antitrust policy than are the magnitudes of market-aggregated parameters—because, given the cost of defining any antitrust market, decision-protocols that make use of antitrust-market-aggregated-parameter estimates achieve what the British call “the remarkable double” of increasing cost while decreasing accuracy.

The preceding argument does not address an equally important reason why antitrust markets cannot be defined non-arbitrarily. If the assumption is that the antitrust-market-definition-generating protocol that is “correct” is the protocol that minimizes the erroneousness of antitrust-law decisions and makes what I take to be the realistic assumption that any protocol for defining antitrust markets will in combination with the rest of the market-oriented protocol for deciding antitrust cases result in some legally incorrect decisions being made, it will not be possible to choose an antitrust-market-definition-generating protocol non-arbitrarily because there is no non-arbitrary way to measure the seriousness of the errors with which any antitrust-market-definition-generating protocol will be associated and it is not the case that one such protocol will yield “smaller errors” than any other protocol regardless of the metric one uses to measure the errors associated with each such protocol’s use. The first reason why one cannot determine the correct antitrust-market-definition-generating protocol or the best antitrust-market definition non-arbitrarily is that there is no non-arbitrary way to decide whether the errors that the use of a particular protocol or market definition of these kinds will cause to be generated should be measured by (1) the number of cases the protocol’s or antitrust-market definition’s use will result in being decided legally incorrectly, (2) the average “seriousness” of the errors associated with the use of each possible protocol or antitrust-market definition, (3) some function that gives weight to both the number of the relevant legally erroneous decisions and the average seriousness of the relevant erroneous decisions, and (4) if both the number and average seriousness of the relevant errors are deemed to count, the function that takes both into account. The second reason why one cannot identify non-arbitrarily the “correct” antitrust-market-definition-generating protocol or the correct antitrust-market definition is that there is no non-arbitrary way to measure the seriousness of an incorrect resolution of a given antitrust case. Thus, there is no non-arbitrary way to decide whether the seriousness of an incorrect antitrust-law application should be measured by (1) the magnitude of the empirical error the erroneous legal conclusion manifested, ¹⁶ (2) the seriousness (however measured) of the intellectual error the judge made when interpreting the relevant statute (when determining the test of illegality it promulgated), (3) the disservice the erroneous decision would be calculated to have done to the instantiation of the liberal conception of justice if it had no precedent-related consequences (which might depend on whether the error was to hold illegal a lawful [M or A] or to hold lawful an illegal [M or A] as well as on the penalties imposed on or damages awarded against an innocent party found guilty or liable), (4) the disservice the erroneous decision did to the instantiation of the liberal conception of justice by eliciting future liberal-moral-rights-violative conduct and by establishing a precedent that would lead to judges’ making mistakes when applying the antitrust law in the future, (5) the disservice the erroneous decision did to the securing of the goals the applicable antitrust law was designed to secure both directly (see [3] above) and indirectly (see [4] above)—a disservice that is rendered more difficult or impossible to calculate by the difficulty or impossibility of determining the goals the statute was designed to secure and the weight the law’s promulgators thought should be given to each of these goals, (6) the disservice the erroneous decision did to the instantiation of the conception(s) of the moral good that the relevant society was either obligated to instantiate or could instantiate without violating any moral obligation. Most of the reasons I have just given for concluding that the selection of an antitrust-market-definition-generating protocol or an antitrust-market definition must be arbitrary when the goal is to elicit antitrust-law applications whose total “erroneousness” as a matter of law is minimized also apply when the goal is to identify the antitrust policies that are morally desirable. Once more, any antitrust policy that requires determinations of illegality to be based on the magnitudes of market-aggregated parameters, which relate to defined markets, will yield some morally undesirable legality-conclusions regardless of the antitrust-market-definition-generating protocol and related antitrust-market definitions that are used and the use of no such protocol or set of related market definitions will eliminate all morally undesirable decisions or be better than all their alternatives from all normative perspectives that may be used in a liberal-moral-rights-based polity whose members also subscribe to morally defensible egalitarian conceptions of the moral good. When the relevant analysis is a policy analysis as opposed to a law-application analysis, some sources of arbitrariness may not be operative (e.g., one will not have to identify the goals of actual legislation or the weight that its promulgators thought should be assigned to each valued goal), but it may still be necessary to make an arbitrary choice of the pure or mixed conception of the moral good that is supposed to be instantiated and one may still have to make arbitrary choices about whether the total erroneousness of the erroneous decisions that are made depends on their number or on their number and their average erroneousness (e.g., on the relative moral cost of finding morally desirable conduct illegal or morally undesirable conduct legal or the relative moral cost of penalizing actors who have made choices that are not morally undesirable and failing to permit victims of morally undesirable choices to obtain legal redress from their “injurers”).

I will stop here. The point is that, even if—contrary to fact—an antitrust-market-oriented antitrust-policy-analysis protocol could be morally superior to an antitrust-policy-analysis protocol that did not make relevant the magnitudes of antitrust-market-aggregated parameters, there would be no non-arbitrary way to identify the antitrust-market-definition-generating protocol or the antitrust-market definitions that would be morally optimal.

B. The Inaccuracy and Moral Undesirability of Using Estimates of Market-Aggregated Parameters to Determine Whether a Firm That Is Setting Individualized Prices Has Practiced Contrived-Oligopolistic Pricing, Whether a Firm Has Practiced Natural-Oligopolistic Pricing, Whether a Firm Has Set Individualized Predatory Prices, and Whether a Firm Has Distorted Competition

All the categories of conduct referenced in this subsection’s heading are liberal-moral-rights-violative. Accordingly, in a liberal-moral-rights-based polity, the most important policy issue that relates to these categories of conduct is whether one can prove (without violating any relevant liberal constraint) that a suspected perpetrator of any such conduct has actually engaged in the conduct. This subsection argues that no estimate of a market-aggregated parameter can play a morally desirable role in the analysis of whether a firm has engaged in any of the referenced categories of conduct. The same conclusion is justified for the other variants of contrived-oligopolistic, natural-oligopolistic, predatory, and competition-distorting conduct that can be practiced: I do not have the space to consider these other variants here.

1. Contrived-Oligopolistic Individualized Pricing

Five categories of evidence (sets of facts) can be used to prove that a firm that is practicing individualized pricing has engaged in contrived-oligopolistic pricing. Even if the relevant market could be defined non-arbitrarily, none of these facts can be accurately or morally desirably inferred from the suspect’s share of the market in which the pricing is alleged to have been practiced or the concentration of the seller-side of the market in question.

The first category of evidence includes documents/emails/audio recordings/employee or rival testimony that manifest or claim that the suspect had discussed its plan to practice contrived-oligopolistic pricing, had discussed its having practiced such pricing, had communicated relevant promises of reciprocation and/or threats of retaliation to relevant rivals, and/or had fulfilled those promises or made good on those threats. Obviously, the suspects’ share of the allegedly relevant market or the concentration of the seller-side of that market has no bearing on the existence of such evidence.

The second category of relevant evidence contains “straightforward” evidence that the prices the suspect charged particular buyers for the relevant product exceed the seller’s HNOPs in its relations with those buyers. (I assume that the defendant will be allowed to exonerate itself by proving that any gap between the price it charged a buyer and its HNOP is attributable to one or more errors it, its rivals, or the relevant buyers made.) It seems obvious to me that evidence establishing that the suspect has a high market share and/or that the seller-side of the relevant market is concentrated have no bearing on whether a seller’s individualized prices exceed its HNOPs. In my judgment, the difficulty of estimating an individualized pricer’s HNOPs will usually prevent antitrust-law-enforcement authorities from using such “straightforward” evidence to prove contrived-oligopolistic individualized pricing.

The third category of evidence that can be used for this purpose may be more available—“comparative evidence” that establishes that an individualized-pricing seller (1) has charged different prices at different times in the same location or different prices at the same time in different locations when the price differences cannot be attributed to differences in the seller’s HNOPs at the different times or in the different locations or (2) has charged the same price for a given good at different times or at the same time in different locations when the seller’s HNOPs can be proved to be different at the different times in the same location or in the different locations at the same time. Although the price relationships in question could reflect the seller’s practice of predatory pricing instead of its practicing contrived-oligopolistic pricing at one time but not at the other time or in one location but not in the other location, I do not find this possibility problematic. This kind of evidence is more likely to be practicable than the second category of (“straightforward”) evidence just described because it is cheaper to measure differences in HNOPs than to measure HNOPs. Once more, I see no reason to believe that the magnitude of any market-aggregated parameter has any bearing on whether this category of evidence will be available.

The fourth category of evidence that would be probative of (individualized) contrived-oligopolistic pricing contains a variety of other sorts of behavioral evidence—evidence that the suspect has lost sales to buyers it was best-placed to supply, evidence that the suspect has charged retaliatory prices to some buyers (ideally to customers of rivals that made sales to the alleged contriver’s customers), evidence that the alleged contriver has allowed rivals to make sales to their own customers at prices contained in offers that the alleged contriver would have found inherently profitable to beat, evidence that the alleged contriver had attempted to consummate an (M or A) that would not otherwise have been profitable with a rival that would not cooperate with its contrived-oligopolistic pricing or had considered engaging in or actually engaged in predation aimed at driving the relevant non-cooperator out, evidence that the alleged contriver had engaged in contrived-oligopolistic or predatory conduct in the past (though I recognize that in some jurisdictions such evidence is not admissible in criminal trials). A defendant’s share of the market in which the contrived-oligopolistic pricing allegedly took place and/or the concentration of the seller-side of any such market has no bearing on the availability of any behavioral evidence of any of these kinds.

The fifth and final category of evidence that is probative of contrived-oligopolistic individualized pricing has two components: (1) evidence that the suspect would engage in such pricing if it were profitable even though the pricing was liberal-moral-rights-violative and illegal and (2) evidence that the alleged contrived-oligopolistic pricing would be profitable for the suspect. It seems obvious to me that a seller’s market share and a fortiori the concentration of the seller-side of the market in which the alleged contrived-oligopolistic pricing purportedly was practiced has no bearing on the disposition of a suspect to engage in the practice when doing so would be profitable. This point is salient because, although I am aware of considerable anecdotal evidence of businesses’ engaging in immoral, illegal conduct, I still think that the vast majority of U.S. businesses will not engage in profitable conduct that is immoral and illegal. It is more complicated to explain (1) why the magnitudes of market-aggregated parameters do not have much if any connection to factors that determine the profitability of contrived-oligopolistic individualized pricing to a suspect and (2) why, even if I underestimate the relevant connections, it would be both more accurate and more morally desirable to estimate the relevant factors directly rather than to use market-aggregated parameters as surrogates for them. I will now be more specific about some of the non-market-aggregated parameters that affect the profitability of contrived-oligopolistic individualized pricing to a seller. Those parameters include (1) the seller’s (HNOP − MC) gap (the higher the gap, the less-profitable contrived-oligopolistic pricing because the larger the amount of safe profits the seller must put at risk to practice contrived-oligopolistic individualized pricing), (2) the number of rivals that could profit by beating the offers that the potential contriver could make that would contain prices that contain different COMs—the number of rivals of the possible contriver that are second-placed or worse-than-second-placed to supply buyers the possible contriver is best-placed to supply by less than the COM the best-placed supplier is trying to obtain (which equals the number of rivals to which the seller must make [costly] contrived-oligopolistic communications and the number of rivals that might have beaten the seller’s COM-containing offer [and hence the cost to the seller of identifying its non-cooperating rival if it determines that is has lost a sale to a rival that was worse-than-best-placed to supply the buyer in question]), (3) the profits that different individual rivals could realize by not cooperating with the seller’s contrivance if it would not react to their non-cooperation and/or cooperation (the frequency with which given rivals are second-placed to supply buyers from whom the seller is trying to obtain a COM and are worse-than-second-placed to supply the possible contriver’s customers by an amount that is lower than the COM the seller is trying to obtain, the size of the COM the possible contriver would try to obtain from customers the relevant rival is second-placed to supply, and the difference between that COM and the average amount by which the relevant rival is worse-than-second-placed when it is worse-than-second-placed), (4) the frequency with which the possible contriver is second-placed to supply buyers its possible non-cooperating rivals are best-placed to supply and the average amount by which in those cases in which the relevant rival is best-placed and the possible contriver is second-placed the possible contriver is better-placed than the third-placed supplier of the buyer in question (which factors determine the benefits the possible contriver can confer on a relevant rival virtually costlessly by reciprocating to that rival’s cooperation), the number of sales that the possible contriver’s potential non-cooperators are best-placed to make that the possible contriver is second-placed or close-to-second-placed to make and the relevant rival’s average (HNOP − MC) difference in its relations with the buyers just referenced (which [roughly speaking] determine the ratio of [A] the loss that the possible contriver would have to incur to inflict different amounts of harm on the relevant rival by reacting to the rival’s non-cooperation by making sales to the rival’s customers by charging them retaliatory prices to [B] the harm that such retaliation would inflict on the non-cooperating rival [for different amounts of harm inflicted] where the loss the contriver would have to incur to take a sale away from a non-cooperator by setting a retaliatory price to the relevant buyer is one cent plus the amount by which the contriver was worse-than-second-placed to supply the relevant buyer and the harm inflicted on the non-cooperator is the non-cooperator’s [HNOP − MC] gap in its relations with that buyer [assuming that the rival is charging that buyer its HNOP]), (5) the amount by which the possible contriver’s practice of the relevant contrivance will increase/decrease the profits it can make by engaging in oligopolistic and predatory conduct by strengthening/weakening its reputation for engaging successfully in strategic conduct (depending on whether the contrivance is successful, on whether the contriver fulfills its related promises of reciprocation and threats of retaliation, on the amount of investment the potential contriver owns and will own in other ARDEPPSes, and on the extent to which rivals in other ARDEPPSes learn of the contriver’s conduct and that conduct’s outcome in the ARDEPPS in question), (6) the law-related costs the possible contriver should anticipate incurring as a result of its practicing the contrived-oligopolistic pricing (which if anything will tend to increase with its market share and the concentration of the seller-side of the allegedly-relevant market to the extent that buyers and antitrust-enforcement officials pay more attention to the conduct of sellers with high market shares and sellers in concentrated markets and to the extent that triers of fact are more likely to find that such sellers in such markets have violated the law on any given body of other evidence). I could go on, but enough is more than enough—Dayeinu.

I want to close with three points. First, regardless of how markets are defined, across all defined markets, the suspect’s market share and the concentration of the seller-side of the defined market are not significantly or strongly positively correlated with the profits a possible contriver could make by practicing contrived-oligopolistic individualized pricing. Thus, regardless of the market-definition protocol that is used, across all defined markets, a seller’s market share and the concentration of the seller-side of its market will not be significantly or strongly correlated with (1) its (average [HNOP − MC] = average [BCA + CMC_#2]) in its relations with those buyers it is best-placed to supply, (2) the number of rivals that are second-placed, equal-second-placed, or worse-than-second-placed by a smaller amount than the various COMs the potential contriver might be tempted to contrive, (3) the ratio of the number of a potential contriver’s potential undercutter’s customers that the potential contriver is second-placed to supply to the number of the potential contriver’s customers that the potential undercutter is second-placed or close-to-second-placed to supply, (4) the average amount by which the potential contriver is better-placed than the third-placed suppliers of those buyers the potential contriver is second-placed to supply and a relevant rival of the potential contriver is best-placed to supply, (5) the ratio of the number of buyers the potential contriver is second-placed or close-to-second-placed to supply that a relevant rival is best-placed to supply to the number of buyers that the potential contriver is best-placed to supply and the relevant rival is second-placed or close-to-second-placed to supply, (6) the average (HNOP − MC) gap for relevant rivals in their relations with buyers they are best-placed to supply and the potential contriver is second-placed or close-to-second-placed to supply. I realize that I am leaving it to you to explain to yourselves why these claims and other claims I make that pertain to the relevance of the magnitudes of market-aggregated parameters to the profitability of contrived-oligopolistic individualized pricing to any potential perpetrator are justified, but it seems obvious to me that they are justified and that it is equally obvious that one could not accurately base any assessment of the profitability of such pricing for a potential perpetrator on the magnitudes of market-aggregated parameters even if, contrary to my conclusion, high market shares and high seller concentration have some bearing on the profitability of contrived-oligopolistic individualized pricing to any possible contriver. Second, high market shares and high seller concentration disfavor the profitability of contrived-oligopolistic individualized pricing by increasing the law-related cost of engaging in the practice. Third, even if I am underestimating the relevance of the magnitudes of these market-aggregated parameters for the profitability of contrived-oligopolistic pricing to a suspect, it would clearly be more accurate and more morally desirable to estimate the profitability of such pricing to a particular suspect by estimating the magnitudes of the non-market-aggregated parameters I have just listed (not primarily because but also because doing so would probably be cheaper than executing the market-oriented protocol, given the cost of defining the allegedly relevant market).

2. Natural-Oligopolistic Pricing

I will focus on three differences between the contrived-oligopolistic individualized-pricing analysis and its natural-oligopolistic-pricing counterpart. None of these differences disfavors my conclusion that it would be inaccurate and morally undesirable to base any assessment of whether a seller would find it feasible to practice and/or has practiced natural-oligopolistic pricing on the magnitudes of market-aggregated parameters.

The first difference relates to the substance of the documentary (firm-meeting minutes), email, recorded-message, testimonial evidence that bears on whether a suspect has practiced natural-oligopolistic pricing. When the suspected conduct is natural-oligopolistic pricing, the evidence in this category is evidence that the suspect realized that and planned to or did take advantage of the fact that its rivals would not beat offers it made that they would find profitable to beat if they believed that the buyers would not give the natural-oligopolistic pricer an opportunity to rebid, statements by the alleged natural-oligopolistic pricer that other choices (it was considering making)/(did make) (would be)/were rendered more profitable by the fact that the increase in (HNOP − MC) differences they yielded/would yield would enable the seller to obtain NOMs, evidence that the natural-oligopolistic pricer’s relevant rivals did not bid for the affected buyers’ patronage even though they could beat the natural-oligopolistic pricer’s offer with an offer whose acceptance would yield them profits, evidence that indicates that those rivals decided not to bid in the above circumstances because they realized that they would not in the end be able to make a sale to the relevant buyers, testimony from buyers that they did not receive bids from their worse-than-best-placed suppliers even though one or more of these potential suppliers could profit by beating their best-placed supplier’s initial offer if the rival’s offer would be accepted, evidence that shows that the buyers considered responding to a superior offer by a worse-than-best-placed supplier by not giving their best-placed supplier an opportunity to rebid or chose not to giver their best-placed supplier an opportunity to rebid to induce that best-placed supplier and their other best-placed suppliers not to charge them NOMs in the future. The fact that the substance of the relevant documents, emails, audio recordings, testimony is different in natural-oligopolistic-pricing cases than in contrived-oligopolistic pricing cases is irrelevant to the point I am now making: the alleged natural-oligopolistic pricer’s market share and the concentration of the seller-side of the market in which it was placed have no bearing on the probability that these type of evidence of natural-oligopolistic pricing will be available.

The second difference relates to the other kinds of behaviors whose proof supports the conclusion that an accused seller has practiced natural-oligopolistic pricing. I have already referenced two of these “other” behaviors: (1) decisions by worse-than-best-placed suppliers not to make offers that beat the best-placed supplier’s initial offer despite the fact that the worse-placed rivals would profit if their initially superior offer were accepted, and (2) one or more decisions by a firm accused of engaging in natural-oligopolistic pricing to make choices that would have been unprofitable but for the fact that, by raising the accused’s (HNOP − MC) differences, the decisions enabled the accused to obtain NOMs when it could not otherwise have done so. Again, the accused’s share of the allegedly relevant market in which the accused is suspected of having practiced natural-oligopolistic pricing and the concentration of the seller-side of the market have no bearing on the availability of these other kinds of behavioral evidence.

The third difference relates not to the determinants of the probability that a seller suspected of engaging in natural-oligopolistic pricing would engage in the practice if it were feasible for the seller to do so despite the conduct’s immorality and illegality but to the determinants of the feasibility of natural-oligopolistic pricing, which differ from the determinants of the profitability of contrived-oligopolistic-pricing. More specifically, the feasibility of a seller’s practicing oligopolistic pricing naturally is directly related to its (HNOP − MC) gap and is inversely related inter alia to (1) the benefits the relevant buyer can obtain by foregoing the opportunity to profit in the short run by allowing its best-placed supplier to beat any initially superior offer a worse-than-best-placed potential supplier makes (because doing so puts pressure on the buyer’s best-placed suppliers not to charge it natural-oligopolistic prices in the future), (2) the cost the possible natural-oligopolistic pricer will have to incur to verify with any degree of accuracy the truthfulness of its customer’s claim to have received a superior offer, (3) the standard transaction cost the possible natural-oligopolistic pricer would have to incur to change its initial offer, (4) the costs that the possible natural-oligopolistic pricer would have to incur to improve its initial offer because doing so would encourage other customers to intensify their bargaining, would lose the goodwill of buyers that have not received superior rebids, or would create opportunities for arbitrage. Once more, regardless of the protocol that is used to define relevant markets, across all defined markets, these determinants of the feasibility of natural-oligopolistic pricing are not significantly or strongly correlated with the suspected natural-oligopolistic pricer’s market share or a fortiori with the concentration of the seller-side of the allegedly relevant market in question.

3. Predatory Pricing (Which Is Almost Always Individualized)

I will focus on four differences between the predatory-pricing-related analysis and the contrived-oligopolistic individualized-pricing analysis, none of which suggests that the accuracy or moral desirability of using estimates of market-aggregated parameters to assess whether a particular suspect has practiced predatory pricing is any greater than the accuracy and moral desirability of using such data to assess whether a particular suspect has engaged in contrived-oligopolistic individualized pricing.

The first difference relates to the substance of the evidence contained in firm-meeting minutes, in firm emails, in audio recordings, and in testimony by firm employees/predation targets/buyers that have received predatory prices. The evidence provided would consist of (1) alleged-predator discussions of (A) the determinants of the benefits the suspected predator could obtain by driving the target out (see below) and the likely magnitude of those benefits, (B) the amount of harm the predator would have to inflict on the target to induce it to exit (or to sell out to the predator at a knock-down price), (C) the identity of the target’s customers to which it would be most cost-effective to offer predatory prices and the harm-inflicted to loss-incurred ratio for offering different target-customers’ predatory prices, (D) the possibility that it would be more profitable to buy the possible target out than to drive the target out, (E) the profitability of making an investment to deter someone else from replacing the investment of a predation target that has exited, (F) the ex post profitability of predation that was already practiced, and so on; (2) predator-employee testimony that the firm suspected of having practiced predation did so; (3) predation-target testimony that (A) the suspected predator has threatened it with predation, (B) the suspected predator has offered it a price for its business that was lower than the value it would have had had the target not been the target of predation, (C) the suspected predator had taken sales from the target despite the fact that the target had made offers to the relevant buyers that the alleged predator would not have found intrinsically profitable to beat, (D) the target had occupied competitive positions via-à-vis the suspected predator that would render its exit beneficial to the suspected predator (was often second-placed to supply buyers the suspected predator was best-placed to supply and in those instances was significantly better-placed to supply the relevant buyers than their third-placed suppliers were and was often best-placed to supply buyers that the suspected predator was not only second-placed to supply but was significantly better-placed to supply than were their third-placed suppliers), and so on, and possibly (4) testimony by buyers whose patronage the alleged predator had secured that it had done so despite the fact that they much preferred the product of their regular supplier (the target) by offering them prices that were far lower than their regular supplier’s price (that the alleged predator had charged them prices that seem suspicious because they seem likely to be lower than the suspected predator’s marginal or incremental-variable cost). The point here is that the alleged predator’s share of the market in which the allegedly predatory prices were charged and/or the concentration of the seller-side of that market have no bearing on the probability that any of these kinds of evidence will be available.

The second difference relates to the “straightforward” (i.e., non-comparative) and comparative evidence that might be used to establish that a suspected wrongdoer’s prices were (price was) not legitimate. Of course, when the conduct at issue is alleged predatory pricing rather than alleged contrived-oligopolistic pricing, the concern is that the price(s) charged is (are) below rather than above the legitimate price. But I do not think that that difference is consequential. What may be consequential is the possibility that, when the concern is that the price (prices) being charged is (are) too low, the charging of those prices might be legitimated by the profits the extra sales that the low prices enable the seller to obtain yield the seller (1) by leading the buyer that made the induced purchases to buy the product in the future, (2) by inducing other buyers who observed the performance of the extra units sold or were told of their good performance by their purchasers to purchase the good in question in the future, (3) by enabling the seller to make profitable extra sales of complements of the good priced low, (4) by enabling the seller to make profitable sales of other members of the product-line to which the product priced low belongs (because buyers prefer to buy multiple members of a single product-line for aesthetic reasons or because all members of any single product-line have similar performance-attributes and methods of proper use but these features vary from product-line to product-line), (5) by enabling the seller to make profitable sales of other unrelated products it sells because the product priced low performs well and the performance of the additional units sold enhances the seller’s reputation for producing good products, and conceivably (though I am skeptical) (6) by reducing the costs the seller must incur in the future to produce the product priced low or other products because of information the production of the additional units sold yields the seller. The point that is relevant in the current context is that the market share of the suspected predator and the concentration of the seller-side of the market in which the alleged predation allegedly took place have no bearing on whether these possible legitimating consequences of charging low prices are present and make it less likely that the suspected predator’s suspiciously low price(s) really was (were) predatory.

The third difference relates to the other kinds of behavioral evidence that bears, respectively, on whether a suspect has engaged in contrived-oligopolistic individualized pricing on the one hand or predatory pricing on the other. The only difference in such other kinds of relevant behaviors that I suspect may exist relates to the price or terms that a practitioner of contrived-oligopolistic individualized pricing may offer a non-cooperator that it is seeking to acquire or merge with and the price or terms that a prospective predator would offer a target of its actual or prospective predation. I suspect that a predator will tend to offer the target of its on-going or planned predation a price for its business that is lower than the value the target’s business would have if the target were not a target of predation, whereas a contrived-oligopolistic pricer will not tend to offer a non-cooperator a price for its business that is lower than that business’ value. Once more, the point that is relevant in the current context is that the size of a suspected predator’s market share or the concentration of the seller-side of the market in which the predatory pricing is alleged to have taken place have no bearing on the possibility that such merger-term or acquisition-price evidence will be available.

The fourth difference relates to the determinants of the profitability of predatory pricing, some of which I have already referenced when discussing the documentary, email, audio-recording, and testimonial evidence and some of the other behavior-evidence that bear on the probability that a suspect has practiced predatory pricing. The determinants of the profitability of a firm’s practicing individualized predatory pricing include (1) the size of the losses that predatory pricing must impose on its target to induce it to exit (the supernormal profits the target’s targeted investment[s] would have generated absent the predatory pricing and the stake the target has in avoiding a reputation of being easily driven out), (2) the cost to the predator of financing the short-run losses it will have to incur to inflict various amounts of losses on its target through predatory pricing (which depends on whether the predator can finance those losses from retained earnings generated in the market in question and/or in other markets in which it operates), (3) the harm-inflicted to loss-incurred ratio for the predator for different amounts of harm it could inflict on the target through predatory pricing (which depends on the frequency with which the predator is second-placed or close-to-second-placed to supply buyers the target is best-placed to supply and the target’s average [HNOP − MC] gap in its relations with those buyers the predator is second-placed or close-to-second-placed to supply and the target is best-placed to supply), (4) the dollar-gains the exit of the target would (directly) yield the predator if the target’s investment would not be replaced by a rival ([the number of sales the predator was best-placed to make and the target was second-placed to make times the average amount by which in cases in which the predator and target occupied these positions the target was better-placed than the relevant buyers’ third-placed suppliers] plus [the number of sales the target was best-placed to make and the predator was second-placed to make] times [the average amount by which in those cases in which the above conditions are fulfilled the predator was better-placed than the buyers’ third-placed suppliers]), (5) the amount by which the dollar-gains referenced in Item (4) in this list would be reduced because the target’s investment would be replaced by a rival investment if the predator did not make an investment that would deter the relevant-rival investment (which depends inter alia on whether the barriers to investment any rival faced were higher than the barriers the target faced on its withdrawn investment and on the extent to which the rival investor’s investment would be less competitive with the predator’s investment[s] than the target’s investment was [were]), (6) the extent to which the predator could reduce the amount by which the possibility of a “replacement”-investment by a rival decreased the profit-yield of the predation by investing itself, (7) the amount by which the predator would increase the predator’s overall profits by enhancing its reputation for engaging in strategic conduct, and (8) the law-related cost of the predatory pricing. It seems obvious to me that (1) the magnitudes of an alleged predator’s market share and the concentration of the seller-side of the relevant market have little or no bearing on these determinants of the profitability of predatory pricing (as well as having no bearing on whether a particular alleged predator would violate the law by practicing [liberal-moral-rights-violative] predatory individualized pricing if it were profitable for it to do so). It also seems obvious to me that, even if I underestimate the relevance of magnitudes of the allegedly relevant market-aggregated parameters for the profitability of predatory individualized pricing to a particular accused or for that actor’s willingness to violate the law and liberal moral rights when it is profitable for it to do so by engaging in predation, it would be more accurate and morally desirable to investigate these issues directly than to infer the correct resolution of these issues from the magnitudes of the standard market-aggregated parameters, particularly given the cost of defining the allegedly relevant market.

4. Competition-Distorting Conduct

I have already listed many of the types of conduct that are competition-distorting. I suspect that in virtually all cases in which conduct is alleged to be competition-distorting there is no doubt that the conduct in question was competition-distorting and that in the vast majority of cases in which a defendant is alleged to have engaged in conduct that would be competition-distorting if practiced there is no doubt that the defendant engaged in the conduct in question. However, I acknowledge that in some cases there may be some doubt about whether the defendant “did it”—specifically, (1) did deter potential suppliers of credit or other inputs to a targeted rival from supplying that target by bribing them to withhold their supplies or by telling them negative lies about or communicating misleading negative information about the target’s credit-worthiness or probable longevity, (2) did deter or attempt to deter potential customers of a targeted rival from patronizing the target by providing the potential buyers with inaccurate negative information about the target’s likely longevity, the quality of the performance of the target’s product, and the lifetime maintenance and repair costs of owning the target’s durable products or by providing such buyers with inaccurate positive information about its own survival, its product’s performance, and the lifetime maintenance and repair costs of owning its product (when that product is durable), or (3) did bribe target-employees to provide lower-quality labor or to leave the target’s employ. I also acknowledge that, when the liberal-moral-rights-violative character of a suspect’s communications are at issue and, relatedly, the moral obligation of the government of a polity committed to instantiating the liberal conception of justice to (prevent inaccurate or misleading communications)/(give rival and buyer victims of competition-distorting communications an appropriate opportunity to obtain legal redress)/(punish communicators of inaccurate or misleading competition-distorting information) are at issue, it will be relevant to determine whether the competition-distorter acted innocently, negligently, or intentionally. The questions with which I am currently concerned are: (1) do estimates of an alleged competition-distorter’s share of an allegedly relevant market or the concentration of the seller-side of that market have any bearing on whether the suspect “did it” or if it did do it whether it did it innocently, negligently, or intentionally, and (2) if evidence on the magnitudes of such parameters does have some bearing on the answers to these questions, would it nevertheless be both more accurate and more morally desirable to address these issues directly than to attempt to resolve them either exclusively by drawing inferences from the magnitudes of the standard market-aggregated parameters or indeed by using a protocol that gives some weight to the magnitudes of these parameters? It seems obvious to me that the answer to the first of these questions is “no” and that the answer to the second of these questions is “yes.” Once more, the answer to the first question is “no” because the magnitudes of the standard market-aggregated parameters have no bearing on either the profitability of competition-distorting conduct of any kind to a suspect or the probability that a suspect would engage in such conduct despite its liberal-moral-rights-violativeness and illegality if it were profitable for it to do so.

C. Mergers and Acquisitions—(M&A)s

This section analyzes the accuracy and moral desirability of any analytic protocol that derives conclusions about (1) an (M or A)’s liberal-moral-rights-violativeness, economic efficiency, and impact on material inequality from estimates of the participants’ shares of the allegedly relevant markets in which they are deemed to be operating and the seller-concentrations of the allegedly relevant markets in question (where the [M or A]’s impacts on economic efficiency and material equality are relevant to its impact on the total and average utility of all creatures whose utility is valued and the [M or A]’s impact on material equality is relevant to its impact on the instantiation of the various non-utilitarian-egalitarian conceptions of the moral good) or (2) the moral duty of the government of any polity that is committed to instantiating the liberal conception of justice to adopt and implement various possible (M or A)-focused antitrust policies and the impact of any such policies on the instantiation of any egalitarian conception of the moral good.

I will start by examining the accuracy and moral desirability of using estimates of the magnitudes of (M or A)-participants’ market shares and the concentration of the seller-sides of the markets in which they are deemed to be operating for their (M or A)’s liberal-moral-rights-violativeness. Some (M&A)s are liberal-moral-rights-violative, and some are not. An (M or A) is liberal-moral-rights-violative if one or both of two conditions is satisfied: (1) if at least one participant believed ex ante that the (M or A) would create a resulting firm that would have more opportunities to profit from engaging in liberal-moral-rights-violative conduct than the participants would have had as separate entities and intended ex ante the resulting firm to take advantage of one or more of these additional opportunities and/or (2) if ex ante the perception of at least one of the participants in the (M or A) that the (M or A) was ex ante at least normally profitable was critically affected by a belief that the (M or A) would increase the participants’ profits by reducing the absolute attractiveness of the best offers against which the resulting firm would have to compete relative to the absolute attractiveness of the best offers against which one or both participants would have had to compete as a separate entity 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. I will now address whether it would be accurate or morally desirable to assess whether either of these conditions is fulfilled by using a protocol that attempts to derive conclusions about these issues from estimates of the participants’ shares of allegedly relevant markets and estimates of the concentration of the seller-sides of the allegedly relevant markets.

Section 3B’s analyses (1) of the determinants of the profitability of contrived-oligopolistic pricing, of the feasibility of natural-oligopolistic pricing, of the profitability of predatory pricing, and of the actual commission of the variants of competition-distorting conduct that can be practiced and (2) of the inaccuracy and moral undesirability of any protocol that purports to infer whether a particular suspect has engaged in any such liberal-moral-rights-violative conduct from the share the suspect has of the sales made in or investment located in the allegedly relevant market in which the alleged conduct allegedly took place or the concentration of the seller-side of any such allegedly relevant market implies the inaccuracy and moral undesirability of any protocol that purports to determine whether the participants in any (M or A) would expect the (M or A) to increase their opportunities to profit from engaging in one or more of these categories of conduct and intended ex ante to take advantage of one or more of any such opportunities they expected the (M or A) at issue to afford them. I hope that readers will find convincing or at least plausible my claim that analogous analyses would establish the same conclusion when the conduct whose engagement is at issue is contrived-oligopolistic across-the-board pricing, other kinds of contrived-oligopolistic and natural-oligopolistic conduct, predatory advertizing/refusals to deal/decisions to invest, and competition-distorting conduct of a kind I did not previously address.

This conclusion does not imply the inaccuracy and moral undesirability of any protocol that uses estimates of the magnitudes of market-aggregated parameters to assess whether an (M or A) is liberal-moral-rights-violative because at least one of its participants’ ex ante expectation that the (M or A) would yield it profits by reducing the absolute attractiveness of the best offers against which the resulting firm would have to compete below the absolute attractiveness of the best offers against which the participants would have had to compete as separate entities 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. To investigate this issue, I must (1) list the determinants of the amount of profits an (M or A) must yield (by raising the resulting firm’s profit-yield above the profits the participants would have realized as separate entities) for those profits to constitute at least a normal rate-of-return on the investment the acquirer has to make to execute and make use of the acquisition or the merging firms had to make to execute and make use of their merger, (2) list the different categories of liberal-licit profits an (M or A) can generate as well as the determinants of the magnitudes of each category of liberal-licit profits, (3) examine the relationship between, on the one hand, participants’ shares of allegedly relevant markets and the concentration of the seller-sides of those markets and the determinants referenced after “(1)” and “(2)” in this sentence, and (4) show either or both (A) that the magnitudes of the referenced market-aggregated parameters are sufficiently poor predictors of whether the liberal-licit profits that an (M or A)’s participants would anticipate ex ante that it would generate would constitute at least a normal rate-of-return on the relevant participant’s related investment for any protocol that bases its resolution of this issue on estimates of such parameters to be morally unacceptable even if no alternative to this protocol were available and (B) that alternative protocols that focus directly on the magnitudes of the relevant non-market-aggregated parameters would be both more accurate and more morally desirable.

To simplify, I will assume that the (M or A) is an acquisition. In this case, the profits the acquirer would have to expect ex ante the acquisition to generate for those profits to constitute at least a normal rate-of-return on the acquirer’s (M or A)-related investment will depend on the purchase price of the acquired firm, the costs the acquiring firm would expect to have to incur to integrate the operations of the acquired firm into the resulting firm, and the rate-of-return the acquiring firm considered to be normal on the acquisition in question. The amount of liberal-licit profits the acquiring firm would expect ex ante its acquisition to generate is more complicated to calculate. Those expected profits fall into at least seven categories. First, there are the liberal-licit profits the participants expect the (M or A) to generate by reducing the fixed costs the resulting firm will have to incur to renew one or more production plants or distributive outlets that both it and the participant that owned the old facilities would renew—profits that equal the static fixed-cost (private) efficiencies the acquisition would generate. Second, there are the liberal-licit profits the participants expect the (M or A) to yield them by generating static marginal efficiencies: if for simplicity I assume that those efficiencies reflect the (M or A)’s reducing by $α the height of the marginal-cost curve for producing a product that the resulting firm will produce and a participant would have produced, the associated profits depend on the magnitude of the α, the output of the relevant product that the relevant participant and the resulting firm would have produced had the (M or A) not reduced the marginal-cost curve for producing it, and all the factors that determine the extra profits the $α reduction in the MC curve enables the resulting firm to realize by increasing its output of the relevant product in response to the (M or A)-generated decrease in the relevant product’s MC curve. Third, there are the liberal-licit profit the (M or A) participants expect the (M or A) to yield them by generating dynamic fixed-cost and/or marginal efficiencies—that is, by enabling the resulting firm to make a more profitable investment than either participant could have made as a separate entity. Those profits themselves have many determinants, some of which depend on whether (1) the dynamic efficiencies cause the resulting firm to make an investment when no participant and no mutual rival of the participants would have made an investment, (2) the dynamic efficiencies cause the resulting firm to substitute a more-profitable investment for a less-profitable investment a participant would otherwise have made (when the profit-difference is not attributable to the resulting firm’s investment’s being less competitive with its other investments than the investing participant’s investment would have been with the investment[s] of the other participant), (3) the dynamic efficiencies cause the resulting firm to make an investment that deters a mutual rival of the participants from making an investment, or (4) the dynamic efficiencies convert a situation in which a mutual rival of the participants would have invested into a situation in which neither that firm nor the firm that results from the (M or A) invests by causing the resulting firm and the relevant mutual rival of the participants to confront each other with critical natural-oligopolistic investment-disincentives. To save space, I will not delineate here the different sets of determinants of the profits the dynamic efficiencies an (M or A) generates or could generate will yield the participants in each of the above four situations. Fourth, there are the liberal-licit profits that the (M or A)’s participants will expect it to yield by enabling the resulting firm and possibly the participants’ mutual rivals to substitute for the set of more duplicative/competitive investments they made or would have made a set of less duplicative/competitive investments that are not only less duplicative but also collectively more-economically-efficient because the investments that are substituted against were/(would have been) economic-inefficiently-duplicative. Fifth, there are the liberal-licit profits the (M or A)’a participants expected ex ante the (M or A) would yield them by creating a resulting firm that can take advantage of the tax losses that one or both participants would otherwise have sustained but not have been able to use (profits whose existence and magnitude depends on one or both participants’ sustaining tax losses, the number of years that the relevant polity’s tax law allows such tax losses to be carried forward [to be used to offset subsequent tax profits], the probability that the relevant participant[s] would generate enough taxable income within the allowed-carry-forward period to enable it/them to take advantage of their tax losses, the probability that the resulting firm would realize profits during the allowed-carry-forward period [ideally, profits at least as large as the participant’s or participants’ tax losses], the tax rate applied to the resulting firm’s taxable income). Sixth, there are the liberal-licit “profits” or equivalent-dollar gains the participants expect the (M or A) to yield them by enabling the owner or owner-manager of one of the participants to liquidate his or her assets and, when relevant, to escape the burdens of management (equivalent-dollar gains whose existence/magnitude depend on one of the participants being owned by such an individual and on the equivalent-dollar value that any such individual places on making his or her assets more liquid and escaping managerial burdens). Seventh, there are the possibly-liberal-licit profits the (M or A) participants expect the (M or A) to yield them by enabling them (and their mutual rivals) to overcome the public-good-type obstacles to their spending collectively as much money on (political parties)/(political campaign/political candidates) and participation in government-decision-making processes as is in the participants’ joint interest (profits whose determinants I will not discuss here).

I want to make three points about the significance of the preceding list in the current context. First, there is no or virtually no relationship between the magnitudes of the standard market-aggregated parameters and (the magnitudes of the listed determinants)/(the reality of the listed possibilities) that are relevant to the amount of liberal-licit profits an (M or A)’s participants will expect ex ante it to yield them. Second, even if there is more of a connection than I believe between the magnitudes of the referenced market-aggregated parameters and (the magnitudes of the relevant liberal-moral-rights-violativeness determinants)/(the probabilities that the listed liberal-moral-rights-violativeness-relevant possibilities are realistic), that connection would be far too weak to make any protocol that bases conclusions on whether an acquisition (or merger) is liberal-moral-rights-violative on the magnitudes of market-aggregated parameters acceptably accurate and morally desirable even if there were no other way to estimate (the magnitudes of the relevant non-market-aggregated determinants)/(the probabilities of the relevant possibilities). And third, even if the first two claims just made are wrong, it would clearly be both more accurate and more morally desirable to base conclusions about the liberal-moral-rights-violativeness of an (M or A) on direct investigations of the magnitudes of the listed non-market-aggregated determinants and the reality of the listed possibilities.

I turn now to the accuracy and moral desirability of any protocol for assessing whether an (M or A) or (M or A)-related policy will increase the total or average utility of all creatures whose utility is valued or will serve or disserve the instantiation of the equal-utility, equal-resource, or equal-other-opportunity egalitarian conception of the moral good that derives these conclusions from the magnitudes of the participants’ shares of all markets deemed relevant and the concentration of the seller-sides of those markets. The effects that are relevant to an (M or A)’s or an (M or A)-policy’s impact on total or average utility are the (M or A)’s or (M or A)-policy’s impacts on economic efficiency and material equality. The only effect that is relevant to an (M or A)’s or an (M or A)-policy’s impact on the instantiation of any non-utilitarian-egalitarian conception of the moral good is the (M or A)’s or (M or A)-policy’s impact on material inequality.

I begin with a partial list of the factors that determine an (M or A)’s impact on economic efficiency. These determinants include (1) the amount by which the (M or A) reduces the resulting firm’s fixed cost of renewing an existing investment; (2) the amount by which the (M or A) reduces the marginal-cost curves the resulting firm faces when producing the participants’ products below the marginal-cost curves the relevant participant would have faced as a separate entity when producing the relevant products; (3) the factors that determine the ratio of the allocative-cost saving associated with such static efficiencies to the private-cost saving that was generated; (4) the factors that influence the increase in unit output a marginal static efficiency causes and the economic efficiency that any such increase in unit output generates; (5) the factors that influence the magnitude of the dynamic efficiencies an (M or A) generates and whether those dynamic efficiencies (A) cause the resulting firm to make an investment when neither participant nor anyone else would have added an investment to the relevant portion of product-space, (B) cause the resulting firm to substitute a more-economically-efficient investment for the less-economically-efficient investment a participant or a mutual rival of the participants would otherwise have made, or (C) convert a situation in which a mutual rival of the participants would have invested into one in which neither that firm nor the resulting firm invests; (6) the QV-investment or PPR nature of the investments made in the relevant market(s) (ARDEPPSes), which is relevant because (in my judgment), from the perspective of economic efficiency, contemporary developed economies devote too many resources to QV-investment creation and not enough to PPR, and the factors that determine the extent to which the profits yielded by QV investments are inflated and the profits yielded by PPR are deflated; (7) the factors that determine whether the (M or A) will result in a more-economically-efficient because less-economically-inefficiently-duplicative set of investments being substituted for a less-economically-efficient because more-economic-inefficiently-duplicative set of investments; (8) the factors that determine whether the (M or A) increases poverty and thereby the economic inefficiency that poverty generates ((A) the factors that determine the extent to which the (M or A) raises prices by freeing the participants from each other’s price-competition [how often the participants were uniquely-equal-best-placed to supply particular buyers and in those instances their average competitive advantage over the buyers’ third-placed suppliers and how often the participants were, respectively, best-placed and second-placed to supply particular buyers and in those instances the average amount by which the second-placed participant was better-placed than the buyers’ third-placed suppliers], (B) the factors that determine the amount by which the static marginal efficiencies the [M or A] generates benefits relevant buyers [the magnitude of the static efficiencies, the frequency with which a participant was second-placed or worse-than-second-placed by less than the marginal efficiency generated]); (9) the magnitude of the dynamic efficiencies the (M or A) generates and the other factors that determine whether those dynamic efficiencies increased, decreased, or left unchanged equilibrium investment in the relevant market(s); (10) whether the relevant investments were QV investments or investments in PPR; (11) the percentage of the actual and potential customers of the participants and their mutual rivals that are poor or close-to-poor; (12) the factors that determine whether the participants share some of the profits the (M or A) would otherwise yield them with their managers and employees and the extent to which any additional compensation paid these individuals increases the tax rate applied to their marginal taxable income; (13) the factors that affect the extent to which the (M or A) increases the amount of money the participants and their rivals spend on elections and government-decision-making-process participation and the extent to which any such additional expenditures causes the government to make decisions that reduce economic efficiency; (14) the factors that affect the impact that the (M or A)’s tendency to increase P/MC ratios in the markets in which the (M or A) takes place affect the amount of inter-market UO-to-UO misallocation the relevant economy generates; (15) the factors that affect the impact that the (M or A) has on the amount of inter-market QV-to-QV misallocation or the amount of inter-market PPR-to-PPR misallocation the economy generates by increasing/decreasing equilibrium investment in the market of consummation; and (16) the factors that influence the amount of (A) UO-to-QV and QV-to-UO misallocation the economy generates, (B) PPR-to-QV and QV-to-PPR misallocation the economy generates, and (C) the amount of UO-to-PPR and PPR-to-UO misallocation the economy generates by increasing or decreasing the P/MC ratios of the products in the market of consummation and by increasing or decreasing the intensity of investment-competition in the market of consummation. I will stop here. The point is that the magnitudes of the participants’ market shares and the concentration of the seller-sides of the markets in which the participants are placed have little or no bearing on any of these determinants of the (M or A)’s impact on economic efficiency.

I turn now to the accuracy and moral desirability of any protocol for assessing the impact of an (M or A) or any (M or A)-related antitrust policy on material inequality. I begin with a partial list of the determinants of this impact. That list includes (1) the factors that determine the ratios of the weighted-average income/wealth position of the shareholders/creditors/managers/employees of the participants to the weighted-average income/wealth position of the shareholders/creditors/managers/employees of the participants’ rivals and to the weighted-average income/wealth position of the potential and actual consumers of the products of the participants and the products of the participants’ mutual product-rivals, (2) the factors that determine the profits the (M or A) yields the participants, (3) the factors that determine the equivalent-dollar effect of the (M or A) on the participants’ mutual rivals (the relative magnitudes of the negative impact the [M or A] has on them by yielding static and dynamic efficiencies to the positive impact the [M or A] has on them by freeing the participants from each other’s price-competition [by raising the prices the resulting firm charges the participants’ customers above the prices the participants as separate entities would have charged those buyers]), (4) the factors that determine the net equivalent-dollar impact that the (M or A) has on the actual and potential customers of the participants and their mutual rivals (the factors that determine the amount by which the [M or A] raises the prices the resulting firm charges the participants’ customers [and other buyers] by freeing the participants from each other’s price-competition, the factors that determine the extent to which the static marginal efficiencies the [M or A] generates benefits relevant buyers [the size of those static marginal efficiencies, the frequency with which the participants were second-placed or worse-than-second-placed by less than the static marginal efficiencies], the factors that determine the extent to which the [M or A] benefits/harms the relevant buyers by increasing/decreasing investment-competition in the allegedly relevant markets). Once more although I could be more specific about the relevant factors, I will stop here because the preceding list should suffice to establish the points I now want to make: (1) the magnitudes of market-aggregated parameters have no or virtually no bearing on whether an (M or A) would or did increase or decrease material inequality; (2) even if I underestimate the material-equality relevance of the magnitudes of the relevant market-aggregated parameters, any impact-on-material-equality prediction one could generate from data on market-aggregated parameters would be disturbingly inaccurate; and (3) it would clearly be both more accurate and more morally desirable to investigate the impact of antitrust-policy-coverable conduct and antitrust policies on material equality directly by estimating the magnitudes of the relevant non-market-aggregated parameters than to infer that impact from estimates of the magnitudes of the participants’ shares of allegedly relevant markets and of the concentrations of the seller-sides of allegedly relevant markets.

IV. Conclusion

I hope that this account of the most distinctive features my antitrust-policy study (and of the economic-efficiency-analysis study and the antitrust-law study on which the antitrust-policy study relies) accomplishes two goals: (1) encourages readers to devote their valuable time to my antitrust-policy study and (2) provides a basis for my comments on the contributions other scholars have made to this Antitrust Bulletin issue. My failure to make this summary shorter is not attributable to any time-constraints I faced. This summary’s length is primarily attributable to the fact that the moral, competitive-impact, economic-efficiency, and distributive-impact analyses my study executes all make use of conceptual systems I have developed that contain concepts that require definition—concepts that play an essential role in the policy analyses I execute, that (in my admittedly-self-serving view) critically affect my ability to generate morally correct and morally defensible antitrust-policy conclusions.