I will restrict myself here to discussing the four most original and significant aspects of the law-study's analysis of conglomerate mergers. The first is the law-study's explanation of how conglomerate mergers that create a merged firm that faces one or more conglomerate rivals, each of which operates in at least one market in which each merger partner operates, will increase the profits that the merged firm can realize by practicing contrived oligopolistic pricing above the profits that the merger partners could have earned by doing so; will increase the profits that the merged firm's rivals can earn by practicing contrived oligopolistic pricing; will increase the profits that the merged firm can earn by raising the retaliation barriers to expansion faced by the rival conglomerates in question; will increase the profits that the merged firm's relevant rival conglomerates can earn by raising the retaliation barriers to expansion faced by the merged firm above the profits they could have earned by raising the retaliation barriers to expansion faced by the merger partners; and will increase the profits that the merged firm can earn by targeting the relevant conglomerate rivals for predation and vice versa. Thus, the law-study shows that a conglomerate merger can increase the ability of the merged firm to induce a rival conglomerate that operates in at least one market of both merger partners not to undercut its contrived oligopolistic prices by enabling it to engage in cross-market retaliation and/or to reciprocate to such a conglomerate rival's abstention from undercutting not only in the same market but cross-markets (67) (as well as by creating a merged firm that inherits the reputation for engaging in strategic conduct of the merger partner that has the stronger such reputation).
Relatedly, the law-study explains that a conglomerate merger will increase the profits that a conglomerate rival of the merged firm that operates in at least one market in which each merger partner operates can earn by practicing contrived oligopolistic pricing by enabling that rival to retaliate against the merged firm's undercutting not only within the market in which the merged firm had undercut its price(s) but also across markets against the merged firm if it undercuts such a rival's price(s) and to reciprocate to a merged firm that has abstained from undercutting not only within the market in which the merged firm has foregone that opportunity but also in a different market in which both it and the merged firm are operating. The law-study also explains that a conglomerate merger (1) can increase the profits that the merged firm can make by increasing the retaliation barriers to expansion faced by a rival that operates in at least one of each of the merger partners' markets by enabling the merged firm to retaliate across markets to that rival's QV investment and/or to reciprocate across markets to such a rival's decision to forego making an otherwise-profitable QV-investment expansion and (2) can increase the profits that a rival of the merged firm that operates in at least one market in which each merger partner operates can realize by erecting retaliation barriers to the merged firm's expanding in the same ways in which it can enable the merged firm to earn more profits by confronting the relevant rivals with (higher) retaliation barriers. Finally, a conglomerate merger that creates a conglomerate firm that operates in one or more ARDEPPSes in which each merger partner operates can also increase the profits the merged firm can realize by driving the kind of rival in question out of a market in which it was operating by enabling the merged firm to practice predatory pricing in two or more of the markets in which it and its target are operating and vice versa.
Second, the law-study provides an analysis of the conditions under which potential competition or a particular potential competitor will be effective, which is relevant to the determination of whether conglomerate mergers that eliminate a potential competitor will tend to yield Sherman-Act-illicit profits and reduce competition on that account. More specifically, the law-study demonstrates that a potential competitor will be effective if (1) the entry-barred expansion-preventing QV-investment quantity in the relevant ARDEPPS is lower than the entry-preventing QV-investment quantity (if the relevant [[[[GAMMA].sub.D] + R + S + L].sub.E] + [M or O] total where the subscript E stands for the established firm that is the relevant best-placed potential expander in the relevant ARDEPPS is higher than the [[[[GAMMA].sub.D] + R + S + L].sub.N] total for the best-placed potential entrant where the subscript N stands for the best-placed potential entrant into in that ARDEPPS) and (2) the potential competitor is either (A) uniquely-best-placed to enter the ARDEPPS in question or (B) sufficiently-well-placed to enter that ARDEPPS (i) to make it profitable for an established firm to make a limit QV investment (a QV investment it would not otherwise have found profitable) to keep the entrant out or (ii) to find entry profitable after all better-placed potential competitors have entered and the established firms have made all the limit QV investments they found profitable. (68) I should point out that this analysis implies the incorrectness of both the assumption that the three best-placed potential entrants will always be effective and the assumption that no worse-than-third-placed potential entrant will ever be effective (assumptions made in the U.S. DOJ's 1984 Conglomerate Merger Guidelines). In many situations, no potential competitor will be effective, and in many situations, more than three potential competitors will be effective.
Third, the law-study explains the effects that effective potential competition will have on QV-investment competition and, derivatively, on price competition in a relevant ARDEPPS and, concomitantly, the Sherman-Act-illicit profits and losses to relevant buyers that a conglomerate merger that eliminates an effective potential competitor will generate on that account (ignoring any effects it has by generating dynamic efficiencies). More specifically, the law-study argues that, except in two or possibly three types of situations, effective potential competitors will affect the competitive equilibrium by entering themselves or leading established firms to make limit QV investments--conclusions that imply that conglomerate mergers that eliminate an effective potential competitor will yield the merger partners Sherman-Act-illicit profits on that account by obviating their making a limit QV investment (which would have been unprofitable but for its deterring a rival entry) or by preventing an entry that would otherwise have been executed and would have reduced the supernormal profits their other QV investments in the relevant ARDEPPS yield while harming Clayton-Act-relevant-buyers by precluding them from securing the buyer surplus they would have obtained by purchasing the product or distributive service the eliminated QV investment would have created as well as the extra buyer surplus the eliminated QV investment would have enabled them to obtain when purchasing the preexisting products in the ARDEPPS by lowering the prices charged for them. (69)
The preceding conclusions reflect the law-study's demonstration that, if one excludes situations in which the relevant potential entrant is an established-firm buyer or seller that is considering vertically integrating backward or forward into the potential limit pricer's ARDEPPS, effective potential competitors will induce established firms to engage in limit pricing in only two types of situations, each of which rarely arises--viz., (1) when the potential limit pricer is a relatively-inefficient producer in a product niche that (A) fears that--if the actual profit-potential of its niche becomes known--it will be replaced by a more efficient newcomer and (B) cannot sell its knowledge of the potential of its business to a buyer that could do more with that knowledge or (2) when the potential limit pricer fears that--if the existence of one or more profitable QV-investment opportunities becomes known--potential investors will overinvest and raise total QV investment above its equilibrium quantity.
The fourth and final part of the law-study's analysis of conglomerate mergers that I will discuss at this juncture is its critique of the so-called "toe-hold merger" doctrine that several U.S. courts have used to analyze the legality of geographic-diversification conglomerate mergers. (70) Assume that a firm that wants to engage in geographic diversification (call it K for conglomerate or prospective conglomerate) merges with or proposes a merger with an established firm (E) that sells the same type of product K sells but operates in a geographic market in which K does not currently operate. According to the "toe-hold merger" doctrine, for such a merger to be deemed lawful under the Clayton Act, K must demonstrate both (1) that it would not enter E's market independently if it were prohibited from merging with or acquiring an established firm in that market and (2) either (A) that the E in question was a relatively small firm in its market (was an [E.sub.s]) or (B) if the E was relatively large firm ([E.sub.L]), the K had made reasonable, unsuccessful efforts to identify an [E.sub.s] firm that it could profit from acquiring or merging with if the alternative were engaging in no merger or acquisition in the [E.sub.s]'s market at all.
The law-study criticizes the "toe-hold merger" doctrine on four grounds. First, it questions the doctrine's assumption that, across all cases, the less-profitable but still-profitable K-[E.sub.s] merger the doctrine would require Ks to execute would yield Clayton-Act-relevant-buyers larger net dollar gains than would the more profitable K-[E.sub.L] merger the doctrine would require Ks to reject. It explains that, for this assumption to be true, two conditions must be fulfilled: (1) the positive difference between the profits that the K-[E.sub.L] and K-[E.sub.s] mergers would yield must be bigger than the positive difference between the equivalent-dollar benefits the K-[E.sub.L] and K-[E.sub.s] mergers would respectively confer on Clayton-Act-relevant-buyers (taking account of the fact that in some instances the K-[E.sub.s] merger may benefit relevant consumers more than the K-[E.sub.L] mergers would)--i.e., from the perspective of the goal of benefiting Clayton-Act-relevant-buyers, there must be a bias in favor of K-[E.sub.L] mergers relative to K-[E.sub.s] mergers--and (2) this bias must be sufficiently large relative to the differences in the profits yielded respectively by K-[E.sub.L] and K-[E.sub.s] mergers to make the K-[E.sub.L] mergers as a class less beneficial to relevant consumers than the more-profitable K-[E.sub.s] mergers would be. The law-study then explains why there is little reason to believe that, from the perspective of benefiting Clayton-Act-relevant-buyers, there is any bias in favor of K-[E.sub.L] mergers over K-[E.sub.s] mergers, much less a requisitely-large bias in favor of K-[E.sub.L] mergers. Thus, the law-study shows that--in relation to the profits that K-E mergers will generate for the merger partners because the K can take better advantage of the E's tax losses than the E firm could and because the owner of the E wants to retire and liquidate his or her assets, there is probably a relevant bias in favor of K-[E.sub.s] mergers over K-[E.sub.L] mergers. The law-study also explains that--in relation to the profits that a K-E merger will generate for the K firm when the K firm is able to purchase the E firm at a distress price as a result of its threatening to target it for predation--there is also likely to be a bias in favor of the K-[E.sub.s] merger over the K-[E.sub.L] merger. The law-study goes on the indicate that whether--in relation to the profits the K-E merger will generate for the merger partners by yielding static efficiencies--there is a bias in favor of K-[E.sub.L] mergers over K-Es mergers is unclear: the direction of any such bias depends on whether [E.sub.L] firms or Es firms have a higher ratio of best-placed to second-placed or close-to-second-placed positions. These arguments and others like them strongly favor the conclusion that it is extremely unlikely that, across all cases, there will be a bias much less a critical bias in favor of K-[E.sub.L] mergers over K-[E.sub.s] mergers.
Second, the law-study criticizes the "toe-hold merger" doctrine by questioning its implicit assumptions (1) that it will not be "cost-effective" to analyze on a case-by-case basis whether there is a critical bias in favor of K-[E.sub.L] mergers and (2) that it is legitimate for courts not to execute case-by-case analyses when in some sense it would not be cost-effective for them to do so.
Third, the law-study criticizes the "toe-hold merger" doctrine by pointing out that even if, as a class or in individual cases, the K-[E.sub.s] mergers would benefit consumers more than would the K-[E.sub.L] mergers for which they could be substituted, the doctrine might harm consumers by deterring Ks from investigating the possibility of entering the E market by merger or acquisition by reducing the profitability of the relevant research by increasing its cost and reducing the profitability of the merger that is executed (search-costs aside).
Fourth, the law-study criticizes the "toe-hold merger" doctrine because it implicitly rejects the do-nothing baseline for competitive-impact assessment that I think is part of U.S. antitrust law. Thus, the doctrine will lead to the prohibition of some geographic-diversification K-[E.sub.L] conglomerate mergers despite the fact that they would not lessen competition (indeed, despite the fact that they would increase competition) relative to the do-nothing alternative in the (unjustified and probably incorrect) belief that such mergers would benefit relevant consumers less than would a K-[E.sub.s] merger that would be more profitable for the K firm than doing nothing.
5. Pricing Techniques, Contract-Clause and Sales-Policy Surrogates for the Hierarchical Controls That Vertically-Integrated Firms Can Use to Induce Their Managers and Staff to Act in the Firm's Interest, and Vertical Mergers and Acquisitions
A firm can engage in a wide variety of vertical practices: it (1) can use various pricing techniques, (2) can subsidize its independent distributors' advertising expenditures, in-store trade-promotion displays, and shelf-placement (slotting) decisions, (3) can enter into tying or reciprocity agreements and maximum and minimum resale-price-maintenance agreements, (4) can contractually impose vertical territorial restraints and vertical-customer-allocation restrictions on its independent distributors, (5) can establish single-brand exclusive dealerships and enter into long-term full-requirements contracts and long-term total-output-supply contracts, (6) can inform its dependent distributors that it has adopted a sales policy of discontinuing its supply of any independent distributor that does not conform to one or more of its expressed wishes, and (7) can vertically integrate through merger, acquisition, joint venture, or internal growth. The Sherman Act covers all these species of vertical conduct; the Clayton Act covers (1) price discrimination as it idiosyncratically defines the practice (in Section 2), (2) full-requirements and total-output-supply contracts or contractual clauses and single-brand exclusive dealerships (in Section 3), and (3) (since 1950) vertical mergers and acquisitions (in Section 7). I believe that--with one possible exception--Article 101 covers all these variants of vertical conduct either as agreements among undertakings or (in the case of successful sales-policies) as concerted practices: the possible exception is unsuccessful sales-policies (sales policies that do not succeed in inducing any independent distributor to conform to the manufacturer's wishes). Article 102 covers all these categories of vertical practices when perpetrated by a dominant firm or the members of a set of collectively-dominant rivals. The EMCR covers vertical mergers and acquisitions and full-function vertical joint ventures.
The law-study's analyses of vertical conduct would constitute a substantial book by itself: they occupy 280 pages of Volume II. (71) Obviously, I cannot summarize most of these analyses here. Instead, I will focus on the elements of the law-study's relevant analyses that I think are original and/or important and underemphasized in the literature.
The first law-study point about vertical conduct I want to recount is that the specific-anticompetitive-intent test of illegality and the lessening-competition test of illegality are properly applied to different "units of vertical conduct." Thus, although (not surprisingly) the specific-anticompetitive-intent test of illegality is properly applied (1) to an individual actor's pricing-technique or sales-policy choices, (2) to one or more vertical agreements into which an individual actor enters with independent distributors or suppliers, (3) in the rare instance in which two or more rivals agree to enter into parallel vertical agreements, to the agreement they made, and (4) to an individual vertical merger, acquisition, or joint venture, the lessening-competition test is properly applied to a rule allowing all members of a set of rivals to engage in a covered category of vertical conduct. The latter conclusion is a corollary of the level-playing-field norm of U.S. antitrust law and E.U. competition law. A decision-rule that focused on the competitive impact of an individual firm's vertical conduct would violate that norm because, unless the lessening-competition test were supplemented with a natural-monopoly-type organizational-allocative-efficiency defense that eligible defendants could establish at requisitely-low cost, an individual-firm's-conduct-focused decision-rule would result in the courts' adopting a pari-mutuel approach in which well-established firms would be precluded from engaging in vertical practices that increased their businesses' organizational allocative efficiency/proficiency when the conduct would on this account be requisitely likely to induce a marginal firm to exit by worsening its array of competitive positions or to deter a potential competitor from entering by worsening its array of prospective competitive positions while marginal and potential competitors would be allowed to engage in the exact same vertical practice because their doing so would respectively help them survive and encourage them to enter by improving respectively their actual and prospective competitive-position arrays by increasing their organizations' allocative and presumptively private proficiency. (72)
The second vertical-conduct-related law-study point I want to recount is the only original point the law-study makes about the single-product-pricing techniques a firm can use. The law-study explains the private benefits and costs to a seller of lowering the lump-sum fee and increasing the per-unit price it charges a given buyer above the marginal cost of producing the unit of output whose production would bring the seller's total unit-sales to that buyer to the transaction-surplus-maximizing quantity, the quantity that would maximize the joint gains of the relevant seller and buyer--the quantity at which the buyer's demand curve for the relevant product cuts the seller's marginal cost curve from above. As the law-study shows, such a pricing-move will benefit the seller by reducing the losses it sustains because of its pessimism about the relevant buyer's quantity demand for its product, because of the buyer's pessimism about the buyer's quantity demand for the seller's relevant product, and because of the risk of buyer arbitrage and may also benefit the seller by reducing the sum of the risk costs the relevant transaction imposes on it and the buyer by shifting to the seller some of the risk that the buyer's quantity demand will be lower than both the seller and the buyer expect when the seller is less risk-averse than the buyer and/or when the pricing-move increases the seller's risk by more that it decreases the buyer's risk--say, because the seller produces an input that the buyer uses to produce a final product or the seller produces a final product that the buyer distributes and the seller is concerned only about variations in the total sales of all its customers while the individual buyers are concerned not only about variations in such total sales but also about variations in their shares of those sales. The law-study also explains that such a pricing-move will impose losses on the seller by reducing its unit-sales and hence the joint gain its transactions with the relevant buyer generate and by deterring the buyer from making jointly-profitable demand-increasing expenditures and slotting decisions by creating a situation in which the seller obtains some or more of the joint profits such expenditures yield it and the buyer (though this effect can be reduced by appropriate buyer-expenditure-subsidization policies). (73)
The third point the law-study makes about vertical conduct that I want to recount here may not be original but is often overlooked or, at least, underemphasized: to the extent that a manufacturer charges its distributors a per-unit price for its product that exceeds the marginal cost the manufacturer must incur to supply the distributor with the good in question, it will tend to deter the distributor from making demand-increasing expenditures or slotting decisions that are in the distributor's and manufacturer's joint interest (in addition to causing the distributor to charge a resale price for the manufacturer's good that is higher than the resale price that would be in the distributor's and manufacturer's joint interest). Manufacturers can combat the former tendency of its supra-marginal-cost per-unit pricing by subsidizing its independent distributors' demand-increasing expenditures, by creating single-brand exclusive dealerships, by using so-called full-line-forcing tie-ins to shift the locus of its supra-marginal-cost pricing to a less-differentiated product for which the tendency of such pricing to deter demand-increasing expenditures is less consequential, and by obligating its independent distributors to give it jointly-optimal shelf-space. (74)
The fourth set of original points the law-study makes about vertical conduct contains five points that relate to tie-ins and reciprocity. The first of these is the law-study's recognition that meter pricing is a variant of the mixed pricing-technique in which a lump-sum fee is combined with a supra-marginal-cost per-unit price. (75) This recognition is important inter alia because it makes clear that the profitability of meter pricing will often not depend on its generating discriminatory prices.
The law-study's second original tie-in/reciprocity-related point relates to so-called full-line-forcing tie-ins, in which a seller lowers the per-unit price it charges a buyer Y for a product A the seller X produces below the per-unit price X would find most profitable to charge Y for A in a separate transaction on A (charges the buyer a lump-sum fee plus a per-unit price on A that would yield the buyer unnecessary buyer surplus if nothing more were required of the buyer) on condition that the buyer agree to purchase its full requirements of a second product B (that the seller may or may not produce itself) for a higher per-unit price than the buyer would otherwise have to pay for that second product B. The law-study explains that such tie-ins will tend to be profitable if (1) the ratio of the decrease in seller surplus generated by the price-reduction on A to the increase in buyer surplus generated by that price-reduction is lower than the ratio of the increase in seller surplus generated by the increase in the price of B under the full-requirements contract to the decrease in buyer surplus generated by the increase in the price of B (a condition that will be more likely to be satisfied if the transaction-surplus-maximizing output of B under the full-requirements contract exceeds the transaction-surplus-maximizing output of A, if the slope of the full-requirements demand curve for B over the relevant output-range exceeds the slope of the demand curve for A over the relevant output-range, and if the slope of the marginal cost curve for B over the relevant output-range is gentler than the slope of the marginal cost curve for A over the relevant output-range) and (2) the reduction in A's price will increase the joint profits that the seller and buyer obtain respectively by producing and distributing A more by encouraging the buyer involved in the sale to advertise and promote A than the increase in B's price will decrease the joint profits that the seller and buyer obtain respectively by supplying and distributing B by discouraging the buyer involved in the tie-in from advertising and promoting B (a relationship that is likely to obtain when product A is more differentiated than product B). (76)
The law-study's third tie-in/reciprocity-related point is that (1) tie-ins in which a producer of an input A against which substitution is possible ties its sale of that input to the buyer's agreeing to purchase its full requirements of the input B that could be substituted for the seller's input (and equates the ratio of the price of A to the marginal cost of A to the input seller with the ratio of the price of B under the tie-in to B's market price) or of another input against which substitution is not possible can increase the tying seller's profits by preventing its supra-marginal-cost per-unit pricing from inducing its customer to make jointly-unprofitable substitutions against the seller's input and that (2) end-product-royalty schemes and contractual clauses that obligate the input-buyer to sell the input supplier the input-buyer's total output of its final product for a lower price than the input-buyer would otherwise charge for that final product can perform the same function for producers of inputs against which substitution is possible. (77)
The law-study's fourth and partly-related tie-in/reciprocity-related point is that reciprocity agreements can perform all the Sherman-Act-licit functions that tie-ins can perform (though reciprocity agreements will be somewhat less likely to perform some of these functions than are tie-ins). (78)
The law-study's fifth tie-in/reciprocity-related point that I want to discuss here is really a pair of points that the law-study makes about the leverage theory of tie-ins and reciprocity to which U.S. courts historically subscribed and still appear to subscribe and to which the EC and E.U. courts have always subscribed--the theory that the only function or at least an inevitable function of tie-ins that are employed by sellers that have monopoly power in the so-called tying-product market and of reciprocity that is practiced by a firm that has monopsony power in the market in which it acts as a buyer is to use that power to leverage itself into a position of monopoly power in the so-called tied-product market (in the case of reciprocity, in the market in which the relevant actor functions as a seller). The law-study explains that there is a generation gap in this "theory": if the seller with monopoly power or the buyer with monopsony power has taken full advantage of that power when pricing the "monopolized" good or setting the price it pays for the monopsonized good, no power will be left for it to use when selling the tied-product. (79) The law-study also argues that the recent attempt of some economists to revivify the leverage theory by basing it on the fact that an infinitesimally-small reduction in the price that a seller charges can yield buyers some gains without costing the seller anything fails because sellers cannot in practice make infinitesimally-small changes in their prices. (80)
The sixth set of the law-study's vertical-conduct-focused points I want to recount contains three points that relate to resale-price-maintenance agreements, vertical territorial restraints, and vertical customer-allocation clauses. The first point in this set is that the U.S. courts and the EC and the E.U. courts are incorrect in claiming that the antitrust laws they are respectively charged with interpreting and applying disfavor reductions in intra-brand competition as much as they disfavor reductions in interbrand competition. The U.S. statutes contain no text that supports this conclusion. Admittedly, the texts of Article 101(1) and Article 102 do contain text that favors this claim--clause (e) in Article 101(1) lists as an example of covered conduct that has the object or effect of preventing, restricting, or distorting competition clauses in contracts that "make the conclusion of contracts subject to the acceptance by other parties of supplementary obligations, which by their nature or according to commercial usage have no connection with the subject of such contracts," and clause (d) in Article 102 lists as an abuse of a dominant position precisely the same clauses in contracts. However, these Treaty provisions clearly reflect the failure of the drafters and ratifiers of the original 1957 treaty (the Treaty of Rome) in which they first appear (81) to understand the Sherman-Act-licit functions that resale price maintenance, vertical territorial restraints, and vertical customer-allocation clauses (and, for that matter, tie-ins and reciprocity agreements) perform: these practices and the Article 101(1) clause-(e)referenced and Article 102 clause-(d)-referenced contract clauses that effectuate them are intimately connected to the proper functioning of the contracts that include them and would on that account be a standard part of commercial practice if not prohibited by law. Given these facts and the reality that the relevant treaty-provisions do not condemn these practices directly but condemn them respectively only when they prevent, restrict, or distort competition or constitute an abuse of a dominant position, no textual argument can be made to support the conclusion that the practices in question are illegal under E.U. competition law. Nor can their prohibition be justified in "policy terms." As the law-study shows, all these practices perform a wide variety of Sherman-Act-licit functions, (82) the argument made in both the U.S. and the E.U. that prohibiting these vertical practices can be justified by arguing that doing so protects the liberty interests of independent distributors cannot bear scrutiny, (83) and prohibitions of these practices seem unlikely to secure the Treaty goal of promoting trade among E.U. member-states by increasing "parallel trade" because manufacturers that are forbidden to prohibit independent distributors in one country in which they want to charge lower prices (say, because the country is poorer) from reselling the good to distributors in another country in which the manufacturer wants to charge higher prices (because it is richer) are likely to respond by vertically integrating forward into distribution, by ceasing to discriminate in the poorer country's favor, or by ceasing to sell their products in the poorer country altogether. (84)
The second point or set of points that the law-study makes about resale price maintenance, vertical territorial restraints, and vertical-customer-allocation clauses is that both the U.S. and the E.U. authorities' treatment of these practices has improved. In the U.S., resale price maintenance, which used to be per se illegal, is now to be assessed through a rule-of-reason analysis in which an exemplar is illegal only if it reduces intrabrand competition by more than it increases interbrand competition (85) (a better decision-rule that is still misguided, though it does not seem to have prevented the lower courts that are obligated to apply it from reaching in virtually all cases what is in fact the correct legal conclusion that the practices to which it applies are lawful). In the U.S., the legality of vertical territorial restraints (which were never per se illegal because they did not involve "price-fixing"--an ignorant distinction that reflected the courts' failure  to recognize that vertical territorial restraints and vertical price-fixes perform the same Sherman-Act-licit functions and  to distinguish between horizontal price-fixing and vertical price-fixing) depends on the same trade-off, but lower courts have overwhelming found for defendants in such cases. (86) The U.S. DOJ and FTC have stopped attacking these practices altogether. Starting in 1997, the EC responded to criticisms that both European academics and the E.U. courts made of its handling of these practices by changing its block-exemption regulations (BERs) on them in ways that take at least some account of the legitimate economic functions that these practices can perform. However, as the law-study shows, the EC's treatment of these practices (and, for that matter, the E.U. courts' treatment of them) continues to manifest at best a patchy understanding of their actual functions and a tendency to micromanage behavior that the relevant officials understand only imperfectly. (87)
The sixth set of the law-study's vertical-conduct-related points I want to discuss contains three points or sets of points about the Sherman-Act-illicit functions that various vertical practices can perform even if they do not "foreclose competition." The first such point is that the law-study reiterates the standard conclusions that the following categories of vertical conduct violates the specific-anticompetitive-intent test of illegality: (1) package-pricing tie-ins whose profitability is critically affected by their concealing predatory pricing and/or retaliatory pricing that violates antitrust law, (2) the tiny number of resale-price-maintenance agreements that are designed to facilitate the resellers' horizontal price-fix by enlisting their suppliers as enforcers of that price-fix, (3) the tiny number of resale-price-maintenance agreements that are designed to deter violations of manufacturer horizontal price-fixes by reducing the profits that individual price-fixers can earn by cutting their prices (by cheating their fellow cartel-members) by militating against the cheating's increasing the cheater's sales by deterring the cheater's independent distributors from passing on any of the price-cuts, and (4) the tiny number of vertical territorial restraints and customer-allocation clauses that are part of a broader horizontal market-division arrangement in which each manufacturer restricts its distributors (and hence itself) to competing for the patronage of a subset of the relevant potential buyers that (ideally) does not overlap with the subset of such buyers to which any other participant in the network restricts itself. The second point, which is not standard, is that tie-ins, reciprocity agreements, resale-price-maintenance agreements, vertical territorial restraints, and vertical-customer-allocation clauses will violate the specific-anticompetitive-intent test of illegality if (1) they increase the perpetrator's profits by driving out a rival or deterring a rival QV investment by increasing the demand curve for the perpetrator's product, thereby reducing the actual/prospective demand curves for the exiting/deterred rival/potential rival's products, thereby critically reducing the profits that the exiting rival can earn by staying in the business and/or the profits that the deterred potential investor could earn on any investment it introduced into the perpetrator's ARDEPPS and (2) the perpetrator would not have found the relevant exemplar of the practice in question profitable ex ante but for its belief that the practice would or might have this series of effects. The third point in this set is that only a tiny number of the exemplars of any of the practices in question is likely to violate the specific-anticompetitive-intent test of illegality for this reason and that it will be exceedingly difficult to establish the illegality of any exemplar that actually does violate that test for this reason: even if one can prove that the practice increased the practitioner's profits by causing one or more of its rivals to exit or deterring a rival QV investment, it will be extremely difficult to establish the requisite probability that the practitioner's ex ante belief that the practice would be profitable was critically affected by the prospect of its doing so (given the fact that the practice would also generate Sherman-Act-licit profits).
The sixth set of the law-study's vertical-conduct-focused points that I want to recount at this juncture is a set of points that relates to the concern that standard long-term full-requirements contracts, long-term single-brand exclusive-dealership contracts (which actually are a variant of long-term full-requirements contracts), long-term total-output-supply contracts, and vertical integration (whether by merger, acquisition, joint venture, or internal growth) may yield Sherman-Act-illicit profits and lessen competition by "foreclosing competition."
The first foreclosure-related law-study point is a corollary of the first set of law-study vertical-conduct-focused points this article discusses: the unit of conduct whose "foreclosing" effects is of legal concern varies with the test of illegality that is being applied. When the relevant analysis seeks to assess the legality of conduct under a specific-anticompetitive-intent test of illegality, the relevant unit for foreclosure analysis is an individual firm's engaging in the conduct in question or the conduct of any set of rivals that have agreed to engage in the conduct. When the relevant analysis seeks to assess the legality of conduct under a lessening-competition test of illegality, the relevant unit for foreclosure-analysis is a rule allowing all members of a set of rivals to engage in the conduct in question.
The second foreclosure-related set of points the law-study makes explains the legal relevance of the foreclosure possibility: the fact that conduct forecloses competitors is relevant in specific-anticompetitive-intent cases because the conduct that produces this effect will on that account tend to generate Sherman-Act-illicit profits by causing one or more rivals of the perpetrator to exit or by deterring one or more of the perpetrator's active or potential competitors from making a QV investment in the relevant area of product-space by preventing one or more actual or potential distributor-rivals of a vertically-integrated manufacturer-distributor perpetrator from buying a product its manufacturing division produces or one or more actual or potential final-good-manufacturer-rivals of a vertically-integrated input/final-good-manufacturing perpetrator from buying an input from the input-manufacturing division of the perpetrator; the fact that two or more rivals' conduct forecloses competition is relevant in lessening-competition cases because it raises the possibility that the conduct will inflict a net equivalent-monetary loss on Clayton-Act-relevant-buyers by causing one or more of the perpetrators' rivals to exit or deterring one or more of the perpetrators' actual or potential competitors from making a QV investment in the relevant area of product-space by preventing one or more actual or potential distributor-rivals of the perpetrators from buying the vertically-integrated manufacturer/distributor perpetrators' products or one or more actual or potential final-good manufacturer rivals of vertically-integrated input/final-good-manufacturing perpetrators from buying inputs from the perpetrators.
The third "foreclosure"-related point the law-study makes relates to the fact that the inherently-profitable decisions of a vertically-integrated individual firm or of two or more members of a set of vertically-integrated rivals not to deal with one or more competitors or to charge those competitors higher prices for the perpetrator's or perpetrators' product than the shadow prices that the relevant vertically-integrated perpetrators "charge" the divisions of their companies that compete with the disadvantaged rival(s) may well put the disadvantaged rival(s) at a competitive disadvantage that has no connection to how allocatively-well-placed the disadvantaged rival(s) is (are) to supply relevant buyers. The law-study points out that, if in these circumstances the relevant decisions are said to "foreclose" the disadvantaged rivals even if they do not induce a disadvantaged rival to exit or deter a disadvantaged rival from investing (if the concept of foreclosure is defined to cover cases in which disadvantaged rivals may not be induced to exit or deterred from investing), the fact that conduct "forecloses competitors" in this expanded sense will be relevant to its legality under the "distorting competition" test of illegality promulgated by Article 101(1); and if the relevant perpetrator is a dominant firm or the relevant perpetrators are members of a set of rivals that are collectively dominant, the fact that its or their conduct forecloses competition in this expanded sense will be relevant to its legality under Article 102's "exploitative abuse of a dominant position" test of illegality to the extent that it increases the probability that the perpetrator or perpetrators are obtaining an exploitatively-higher percentage of the transaction surplus their sales to their customers generate.
The fourth set of foreclosure-focused points the law-study makes relates to the way in which foreclosure disadvantages the foreclosed competitors. The standard account focuses on foreclosure's precluding a foreclosed manufacturer from obtaining the quantity of inputs it needs or the quantity of sales to distributors it requires to realize a normal rate-of-return on its investment or precluding a foreclosed distributor from obtaining the quantity of final goods it needs to buy (to "resell") to realize a normal rate-of-return on its investment. The law-study notes that this account ignores the facts that not all varieties of a required input will be equally satisfactory to a manufacturer, that not all potential customers will be equally valuable to a manufacturer, and that not all varieties of final goods will be equally valuable to a distributor. The law-study also offers an alternative (and, I believe, superior) account of the way in which foreclosure disadvantages the "foreclosed" competitor--viz., by depriving it of the profits it would have made on the purchases or sales the conduct in question precludes it from making.
The fifth set of foreclosure-focused points the law-study makes relates to the "character" of the foreclosure's or foreclosures' foreclosing decisions. The law-study indicates that some such decisions are predatory--are decisions to reject inherently-profitable deals that would have been economically efficient in an otherwise-Pareto-perfect economy in order to reduce the absolute attractiveness of the offers against which the forecloser will have to compete by driving the foreclosed rival out or by deterring a prospective "foreclosee" from making a rival QV investment. The law-study also points out that many foreclosing decisions are not predatory--are inherently profitable. Thus, as the law-study argues, (88) many long-term single-brand exclusive dealerships, (89) long-term full-requirements contracts, (90) and long-term total-output-supply contracts (91) that foreclose competitors perform Sherman-Act-licit functions whose prospective performance would suffice to lead their perpetrator to conclude ex ante that the relevant conduct would be profitable. The law-study also explains why it will usually be inherently profitable for a vertically-integrated firm to charge independent customers per-unit prices that exceed the marginal cost the integrated firm must incur to supply them while instructing its downmarket divisions to base their decisions on the assumption that the cost to them of the inputs/final products the firm's upstream division supplies them equals their marginal cost to the upstream division (92) and why most vertical mergers that create firms that disadvantage independents in this way do not violate the specific-anticompetitive-intent test of illegality. (93)
The sixth set of law-study foreclosure-related points I want to reference here relates to the moves that a foreclosed competitor can make to reduce or eliminate the loss that the "foreclosing" conduct would otherwise impose on it: (1) vertically integrating backward or forward itself through internal growth (an option that will be more cost-effective if ARDEPPS demand is increasing through time and the firm faces low [[[PI].sub.D] + R] barriers to entering the relevant ARDEPPS), (2) vertically integrating backward or forward through a joint venture (that faces lower [[[PI].sub.D] + R] banders than it would face on its own), (3) inducing an independent firm to enter the ARDEPPS in which the foreclosing conduct was taking place (by identifying that firm and explaining to it why its entry would be profitable and/or by increasing the profits the independent could earn by entering by offering to enter into a long-term supply-contract with it or by directly subsidizing its entry), or (4) participating in a vertical merger or acquisition. (94)
The seventh law-study foreclosure-focused point I want to repeat here relates to the following question: If buyers are collectively disadvantaged by one or more sellers' inducing enough of them to lock themselves into long-term full-requirements contracts to drive a rival out or to render unprofitable what would otherwise have been a profitable (and buyer-benefiting) entry (or established-rival QV-investment expansion), why do enough buyers agree to lock themselves in to generate this outcome? The answer is that the buyers are "tyrannized by their small decisions" (an expression I am borrowing from my first economics teacher, Fred Kahn): each individual buyer's refusal to lock itself in supplies a collective good to all buyers and, if no individual buyer's refusal to lock itself in will critically affect whether enough buyers are unbound to make entry profitable either directly or by leading enough other buyers to join it, it will not be in any individual buyer's interest to forgo even the tiniest compensation (say, the receipt of the proverbial peppercorn) to lock itself in, and the relevant seller or sellers will therefore be able to profit by locking enough buyers in to foreclose the continued operation of an established rival or a QV investment by a rival. (95)
The eighth and final set of law-study points on foreclosure I want to recount relates to the U.S. and E.U. case-law on foreclosure. I start with three of the points the law-study makes about the U.S. case-law on foreclosure. First, the law-study argues that both the quantitative-substantiality test that the Supreme Court originally promulgated for determining the legality of long-term single-brand exclusive dealerships under the (Clayton Act's) lessening-competition test (96) and the qualitative-substantiality test that the Supreme Court then substituted for the quantitative-substantiality test in a Clayton Act long-term full-requirements contract case (97) cannot bear scrutiny: neither focuses (1) on whether enough suppliers/distributors are left free to allow marginal established firms to survive or potential QV investors that would otherwise have found it profitable to add to the relevant ARDEPPS' QV-investment quantity to profit by doing so, (2) on whether the locked-in firms were particularly-well-placed to supply the firm(s) whose competition might have been foreclosed or vice versa, or (3) on whether the possibly-foreclosed rival could critically reduce or eliminate the loss the relevant conduct would otherwise have imposed on it by vertically integrating itself through internal growth, merger, acquisition, or joint venture or by inducing an independent firm to enter the ARDEPPS in which the conduct alleged to be foreclosing took place or was expected to take place.
Second, the law-study points out (98) that the "essential facilities" doctrine (created by the Supreme Court, though the Court denies having even endorsed it, and applied by several lower courts either explicitly or implicitly) is incorrect as a matter of U.S. law. The "essential facilities" doctrine (which, in fact, has never been well-defined) states that, in some circumstances, at least some types of firms, joint ventures, or participants in other sorts of collaborative arrangements that own a facility that on some relevant definition is "essential" have a legal obligation to make that facility available on "reasonable" terms to others for which usage of that facility is deemed to be essential in the relevant sense. The law-study argues that this doctrine is wrong as a matter of U.S. law if it requires the owners of the essential facility to allow others to use it on "reasonable" terms when their refusal to supply access to it on such terms was not predatory because U.S. antitrust law does not require actors that have obtained competitive advantages legitimately to share any associated transaction surplus with their customers.
Third, the law-study summarizes the position that the U.S. courts have taken in various time periods on the likelihood that vertical mergers would violate the Shennan Act or (post-1950, when the Celler-Kefauver Act amended the Clayton Act to make it cover vertical mergers and acquisitions) the Clayton Act. (99) More specifically, the law-study explains that (1) prior to 1950, U.S. courts assumed that many if not most vertical mergers were designed to and successfully did lessen competition by foreclosing one or more competitors of the merger partners, (2) between 1950 and 1970, U.S. courts--possibly under the influence of leading economists who believed that vertical integration reduces competition by foreclosing competitors (100)--were even more hostile to vertical mergers and acquisitions than their predecessors had been, (3) in the 1970s, the Supreme Court justified its condemnation of a vertical merger in the last such case it has heard (101) with an argument that is a "cousin" of the foreclosure argument--viz., the argument that the acquisition under scrutiny would lessen competition by raising barriers to entry--and the lower courts also manifested their hostility to vertical mergers in opinions that (collectively) held that, although foreclosures of 5-6% were acceptable, foreclosures of 15% or higher justified the conclusion that the vertical mergers that generated them violated the Clayton Act's lessening competition test of illegality, but (4) by the late 1980s (perhaps under the influence of the Chicago School), federal courts had come to the conclusion that vertical mergers and acquisitions reduce competition only rarely. (The law-study also points out that  although the DOJ and FTC were historically as hostile to vertical mergers as the U.S. courts were, starting in about 1970, they became much more friendly toward them,  although the DOJ's 1984 Vertical Merger Guidelines (102) contain a lot of statements that cannot bear scrutiny, they are not particularly hostile to vertical mergers, and  since 1984, the DOJ has devoted few resources to investigating or challenging vertical mergers [though the FTC has paid somewhat more attention to them]).
I will now recount three of the points the law-study makes about the EC's and E.C./E.U. courts' positions on foreclosure. First, the law-study states that, in the early 1990s, the EC seemed to be using something akin to the U.S. "quantitative sustainability" test to assess the legality of allegedly-foreclosing conduct (103) while the European Court of Justice (ECJ) seemed to be using something akin to the U.S. "qualitative sustainability" test of illegality to assess the legality of allegedly-foreclosing conduct. (104) Second, the law-study points out that, more recently, the EC has "evinced a greater awareness than their U.S. counterparts have done of the fact that the foreclosing effect of any arrangement that locks in a potential customer or supplier of an established rival of the perpetrators) or of a rival potential investor in the area of product-space in which the perpetrator(s) is (are) operating depends not on the relationship between the quantity of purchases or sales that the possibly foreclosed rival must make to be able to survive or break even and the quantity of sales or supplies in the relevant area of product-space that are not locked in but on the relationship between the fonner of the above two quantities and the quantity of not-locked-in sales that a particular foreclosee was or would be well-placed to make or the quantity of not-locked-in suppliers whose supply to the relevant foreclosee (given the products' attributes and the locations of different suppliers) would have been most advantageous to it." (105) Third, the law-study also points out that the EC's statements on the foreclosure issue are sometimes inconsistent, sometimes incomplete, and sometimes mysterious and probably wrong. Thus, the law-study points out that the EC's statement that vertical mergers that create an entity that has market shares below 30% in markets whose postmerger HHI is below 2000 are "unlikely" to be "of concern" ignores not only the inevitable arbitrariness of market definitions but the related fact that some pairs of firms in any defined market will be far more rivalrous than are other pairs of firms in that market; the EC's recognition that foreclosure might be prevented by targets' using "effective and timely counterstrategies" is combined with a discussion of the particular counterstrategies that might be effective that is both incomplete and references one counterstrategy--pricing more aggressively-- that is unlikely to be of much use; and the EC's suggestion that potential foreclosees will be able to protect themselves from harm by exercising their buying power is both mysterious and wrong. (106)
The ninth and final set of law-study vertical-conduct-focused points I will rehearse at this juncture contains six points or sets of points that relate to vertical mergers and acquisitions. I have already recounted the first three: (1) the point that, although analyses of the legality of vertical mergers and acquisitions under a specific-anticompetitive-intent test of illegality focus on the motive of an individual perpetrator of one or more vertical mergers or, in the rare instances in which two or more of rivals have agreed to execute parallel vertical mergers or acquisitions, on the motives of these multiple perpetrators, analyses of the legality of "vertical mergers" under a lessening-competition test of illegality should as a matter of law focus on the competitive impact of a rule allowing all members of a set of rivals to execute such mergers or acquisitions (under Article 101(1) if the mergers in question would generate no organizational economic efficiencies) as well as on these actors' ability to establish an organizational-allocative-efficiency (natural-monopoly-type) defense for their conduct in a Clayton Act case or an Article 101(3) efficiency defence in an Article 101 case, (2) all the law-study points I have already discussed about the economics of foreclosure and the foreclosure-related positions and general positions that U.S. and E.U. courts and antitrust-enforcement authorities have taken, and (3) the law-study's explanation of why vertically-integrated firms may make nonpredatory decisions not to supply independents or nonpredatory but competition-distorting decisions to charge independents prices for inputs or final products that exceed the upstream division's marginal costs while "charging" its own downstream division a marginal-cost-equaling shadow price for the product in question.
I want here to recount three additional law-study points or sets of points about vertical mergers and acquisitions. The first additional vertical-merger-focused law-study point is that vertical mergers and acquisitions can yield Sherman-Act-licit gains in at least three intrinsically-lawful ways that their horizontal and conglomerate counterparts cannot:
1. by creating a merged firm that can earn additional profits because it can use hierarchical controls to influence its own employees' decisions as opposed to using contract clauses and sales policies to influence the decisions of independents,
2. by creating a merged firm that can take better advantage of continuous-flow economies, and
3. by creating a merged firm that can coordinate its own downstream and upstream divisions' behavior without revealing relevant information about its competitive position and plans to an independent that might use such information to enter another level of the information-provider's business or might sell that information to an actual or potential rival of the information-provider. (107)
The law-study also points out that a vertical merger or acquisition can enable a regulated firm to conceal its price or rate-of-return violations more cost-effectively. However, although the profits a vertical merger or acquisition yields on this account are not Sherman-Act-licit, they are also not Sherman-Act-illicit. (108)
The second additional vertical-merger-focused law-study point I want to recount relates to the claim that vertically-integrated concerns often violate the law (the applicable test of illegality is usually not specified in this context) by engaging in price-squeezes--i.e., by charging a rival manufacturer or distributor a price for an input or final product that is so high relative to the price the vertically-integrated concern charges for its product as a manufacturer or distributor to preclude the independent from breaking even. The law-study points out that, at least if (as I believe) it is not more difficult to establish the predatory character of a predatory price-squeeze than it is to establish the predatory character of refusing to deal with the potential target of a predatory "squeezing price," vertically-integrated firms will find it more profitable to practice predation by refusing to deal with their targets than by going to the trouble of subjecting their targets to predatory price-squeezes. The law-study also points out that, unlike the U.S. courts, which have finally decided to reject predatory-price-squeeze claims, (109) E.U. courts continue to find pricing conduct they place in this category to constitute an Article-102-violating exclusionary abuse of a dominant position when the pricing is done by a dominant firm. (110)
The third and final set of additional vertical-merger-focused law-study points I want to make here relates to the "post-Chicago-school" analysis of vertical mergers. The law-study offers five criticisms or sets of criticisms of the post-Chicago-school scholarship on vertical mergers. (111) The first is that this literature does not distinguish between policy-analysis and legal analysis--assumes that any policy-argument that can be made for or against vertical mergers will count respectively for and against their legality under U.S. antitrust law (since the relevant scholars are Americans who are focused on U.S. law). The second set of criticisms relates to this scholarship's analysis of the impact that vertical mergers will have on foreclosing conduct and what I call contrived oligopolistic conduct: the literature on foreclosure does not pay any attention to whether the foreclosing conduct is predatory or inherently profitable, and the literature on contrivance ignores many of the ways in which vertical mergers will affect the profitability of contrivance (e.g., by yielding static efficiencies that increase the safe profits the merged firm must put at risk to practice contrivance) and sometimes employs Cournot models to analyze the likely impact of such mergers on contrivance, which are both unrealistic and inappropriate. The third set of criticisms relates to the argument that vertical mergers will lessen competition because early vertical integrators will obtain an advantage over later vertical integrators since early integrators will be able to strike a better deal with merger/acquisition partners than will later integrators. The law-study points out that there is no reason to believe that early integrators will have such an advantage and that, even if they do, there is no reason to believe that that reality will cause vertical mergers to lessen competition (or to violate a specific-anticompetitive-intent test of illegality). The law-study's fourth criticism of this post-Chicago-school literature focuses on its acceptance of a particular Chicago-school policy-argument for vertical mergers--viz., the argument that the vertical merger of successive "monopolists" (say, a monopolistic producer and a monopolistic distributor) will increase economic efficiency by reducing the prices that are charged to ultimate consumers by converting a situation in which an independent manufacturer charges an independent distributor a supra-marginal-cost per-unit price for its product that causes the distributor to charge higher per-unit prices than it would otherwise charge into one in which an integrated firm instructs its distributive division to price its products on the assumption that the distributive division's marginal cost of goods sold equal the marginal costs the manufacturing division incurred to supply the distributive division. The law-study criticizes this argument on two grounds: (1) it assumes that the independent manufacturer would charge the independent distributor exclusively single per-unit prices when in fact many and probably most manufacturers charge a mixture of lump-sum fees and much lower per-unit prices to their independent distributors (i.e., the argument exaggerates the difference between the per-unit price the independent distributor would pay and the per-unit price that the distributive division of a vertically-integrated firm would be instructed to assume it is paying for the goods it sells); and (2) the argument's assumption that any merger-generated decrease in the price charged ultimate consumers will increase economic efficiency (though possibly correct) is asserted and not justified by scholars who fail to understand the central point of The General Theory of Second Best--viz., that, in an economy that contains other Pareto imperfections, there is no general reason to believe that choices that reduce one imperfection (e.g., cause a particular price or set of prices to be more competitive) will even tend on that account to increase economic efficiency. (Of course, supporters of this argument could abandon their economic-efficiency claim and claim instead that their argument establishes that vertical mergers between "monopolists" are desirable because they benefit the monopolists' customers.) The fifth post-Chicago-school argument the law-study criticizes is another Chicago-school argument for the desirability of at least some vertical mergers that the post-Chicago-school revisionists accept: the argument that vertical mergers between the manufacturer of an input against which substitution is possible and a user of that input will increase economic efficiency by preventing jointly-unprofitable substitutions against that input. The law-study criticizes that argument by pointing out (1) that it ignores the fact that an independent input-manufacturer would be likely to reduce or eliminate such jointly-unprofitable substitutions against its input by charging its customers a lump-sum fee for the right to purchase its input at a per-unit price equal to or close to the input-manufacturer's transaction-surplus-maximizing marginal costs, by using tie-ins to shift the locus of some or all of its supra-TSM-marginal-cost per-unit pricing to the input that can be substituted against its input or to an input against which substitution is not possible, or by using end-product-royalty schemes or total-output-supply reciprocity agreements that enable it to capture the buyer surplus that it would otherwise generate by charging the manufacturer a per-unit price for its input equal to the input-producer's TSM marginal cost and (2) that, once more, it ignores The General Theory of Second Best in assuming that choices that increase the chooser's joint profits without violating the specific-anticompetitive-intent test of illegality will increase economic efficiency (although I would agree that in the vast majority of this set of cases choices that increase the perpetrators' joint profits will be economically efficient).
The final category of business conduct that the law-study analyzes contains joint ventures and other sorts of functionally-related collaborative arrangements. I will restrict myself here to rehearsing only four of the points or sets of points the law-study makes about these categories of conduct. The first set of such points articulates the five or perhaps six Sherman-Act-illicit functions that joint ventures can perform: increasing the parents' profits
1. by freeing them from the price competition they wage against each other when selling their preexisting products (joint-sales-agency joint ventures) and by increasing the profits the parents realize by inducing their other rivals to cooperate with the parents' contrived oligopolistic pricing (by facilitating the parents' [A] joint communication of their contrived oligopolistic intentions, [B] sharing of sales-record information, information about the parameters other than rival undercutting that determine the percentages of their old customers they retain, of their rivals' old customers they secure, and of new buyers they secure, information about the competitive positions of particular joint rivals in relation to particular buyers, etc., [C] collaborative reciprocation to collaborators, and [D] collaborative retaliation against undercutters),
2. by confronting each other and their joint rivals with retaliation barriers to entry or expansion,
3. by practicing other sorts of predation against joint rivals (primarily by increasing their ability to coordinate their predatory moves),
4. by enabling the parents to substitute one joint-venture QV investment for the two QV investments the parents would otherwise have made (sometimes by including in the joint-venture agreement a clause prohibiting the parents from investing in the joint venture's market but anyway by giving each parent an interest [often, a 50% interest] in the joint venture's QV investment),
5. by enabling one or both parents to induce the other to agree not to enter its fellow-parent's markets when each agreeing parent would otherwise have been an effective potential competitor of the other parent, and perhaps
6. by enabling the parents to eliminate the competition they wage against each other as buyers (most obviously if the joint venture is a buyer-coop but in other cases as well)--"perhaps" because of the uncertainty of whether the Sherman Act and Articles 101 and 102 prohibit reductions in buyer competition. (112)
The second set of law-study joint-venture-focused points I want to recount here relates to the conditions under which clauses in joint-venture agreements that prohibit the joint venture from entering the parents' markets and/or that prohibit the parents from entering the joint venture's market or each other's markets are correctly said (in antitrust jargon) to be "ancillary" to a joint venture--a characterization that is legally significant because only those restraints that are deemed to be "ancillary" are considered to be lawful. The law-study points out that the U.S. courts, the EC, and the E.U. courts have all failed to define the concept of ancillarity clearly: have never declared whether a restraint is ancillary if (1) it increases the Sherman-Act-licit profits the joint venture yields (e.g., by making it profitable for one or more parents to communicate to the joint venture information that will increase the joint venture's proficiency and profitability [information whose communication to the joint venture would not have been profitable to the communicator had the joint venture or the other parent[s] been able to use it to compete against the communicating parent), (2) it critically increased the ex ante profitability of the joint venture, or (3) (in my view, the legally correct position) it would be deemed legal under a specific-anticompetitive-intent test of illegality. (113)
The third law-study joint-venture-focused set of points I want to reference is actually a series of points made in response to proposals that it would be cost-effective and legally appropriate for courts to use simple filters to dismiss joint-venture cases before detailed evidence about the specific-anticompetitive-intent of the parents or the competitive impact of the joint venture is introduced. The law-study uses a hypothetical to illustrate the reasons why no such filters can be sufficiently accurate to be cost-effective or justifiable as a matter of law--in particular, to demonstrate that (1) one cannot tell whether a joint venture created by an agreement that imposes no restraints on the joint venture or the parents would violate the specific-anticompetitive-intent test of illegality or the lessening-competition test of illegality without obtaining a great deal of detailed evidence about the relevant business situation (inter alia, about the barriers to investment facing the parents and various other potential investors and about the dynamic efficiencies the joint venture would generate) and (2) one cannot tell whether any restraints a joint-venture agreement imposes on the joint venture and/or the parents are lawful without obtaining a great deal of similar information about the situation in the areas of product-space in which respectively the joint venture and its parents operate. (114)
The fourth and final law-study point about joint ventures I want to rehearse at this juncture is that both the U.S. Antitrust Division's approach to joint ventures (115) and the EC's approach to joint ventures (116) respectively in "markets" in which the parents already operate or in "markets" that are "adjacent" to markets in which the parents already operate place too much weight on the seller-concentration of these markets.
This article summarizes what I take to be the most important theoretical contributions of my two-volume study of Economics and the Interpretation and Application of U.S. and E. U. Antitrust Law. The summaries are designed to put into appropriate context the other articles in this issue, but I hope they are valuable in themselves. I have tried to make this article's account of the law-study's positions comprehensible in themselves, but readers who want to pursue these ideas would probably be well-advised to turn to their more complete exposition in the study. Although this article also gives relatively detailed accounts of my interpretations of the U.S. antitrust statutes, the antitrust provisions of E.C./E.U. Treaty, and the EMCR, its treatments of the U.S. DOJ and DOJ/FTC Guidelines, various EC Guidelines and Communications, the EC's BERs (block-exemption regulations), and the U.S. and E.C./E.U. antitrust case-law are very selective and partial. The law-study contains exhausting if not exhaustive discussions of all this material.
The Introduction indicated that the law-study has a not-yet-published policy-sequel--The Welfare Economics of Antitrust Policy and U.S. and E.U. Antitrust Law: A Second-Best-Theory-Based Economic-Efficiency Analysis. The policy-sequel recognizes that antitrust policy should not seek solely to maximize economic efficiency--that, inter alia, antitrust policy should also prevent immoral business conduct (conduct that manifests specific anticompetitive intent), punish the perpetrators of such conduct, enable victims of such wrongful behavior to secure redress, effectuate defensible distributive-justice goals, and help to instantiate the liberal democratic ideal of giving all members of the relevant political community equal influence on their government's decisions. The policy-sequel also recognizes that increases in economic efficiency are not valuable in themselves--that "increases in economic efficiency" (in the monetized sense in which that expression is defined in Part 2A of this article) are morally valuable only to the extent that they are associated with increases in total utility, increases in the satisfaction of preferences, or increases in the extent to which the society's members' have a meaningful opportunity to take their lives morally seriously by fulfilling their moral obligations, developing morally defensible personal conceptions of the good, and leading a life that is consistent with their respective conceptions of the good. (117)
Nevertheless, the policy-sequel focuses almost exclusively on the economic efficiency of (1) the business conduct that antitrust policies cover and (2)(A) the different antitrust policies that could be devised, (B) the antitrust policies that were created by the relevant U.S. statutes, enforcement-agency Guidelines, and judicial decisions, and (C) the antitrust policies that were created by the E.C./E.U. treaty, the EC's Guidelines, Communications, and Regulations and the EC's and the E.C./E.U. courts' case-resolutions. The policy-sequel uses a unique protocol to analyze the economic efficiency of the antitrust-policy-covered conduct and antitrust-policy options it considers. This protocol is distinguished by the fact that it responds appropriately to the basic negative and positive corollaries of The General Theory of Second Best (118)--the theory that demonstrates that, if one or more of a series of sufficient conditions for the achievement of any optimum cannot be or will not be fulfilled, one cannot conclude without further argument that a choice that fulfills more of the remaining conditions will even tend on that account to bring one closer to the optimum. Roughly speaking, this conclusion is correct because--in general--the "imperfections" any choice will eliminate will be as likely to have counteracted the net effect of the remaining imperfections as to have compounded their net effect. When the relevant optimum is "maximizing economic efficiency," the relevant set of sufficient conditions for its attainment are the Pareto-optimal conditions--perfect seller competition, perfect buyer competition, no real externalities, no taxes on the margin of income, resource-allocator sovereignty (each resource allocator knows everything he or she needs to know to identify the choice that maximizes the satisfaction of his or her preferences, given the budget constraint he or she faces), resource-allocator maximization (each resource-allocator makes the choice that the information he or she possesses implies would maximize the satisfaction of his or her preferences, given the operative budget-constraint), and no (critical) buyer surplus. When the policy-choices whose economic efficiency is at issue are choices about engaging in antitrust-policy-covered business conduct or antitrust/ competition policy-choices, the basic negative corollary of The General Theory of Second Best is that one cannot justify the conclusion that business conduct or an antitrust policy that increases/decreases seller or buyer competition will increase/decrease or even tend to increase/decrease economic efficiency on that account by citing the facts that perfect seller competition and perfect buyer competition are Pareto-optimal conditions (belong to a set of sufficient conditions for the securing of the economic-efficiency optimum). Regardless of whether the choice whose economic efficiency is to be investigated is an antitrust-policy-covered business-choice or an antitrust policy, the basic positive corollary of The General Theory of Second Best is that the economically-efficient protocol for analyzing the economic efficiency of any choice must at a minimum instruct the analyst to devote economically-efficient amounts of resources inter alia (1) to identifying the various categories of economic inefficiency (defined by the category of resource allocation with which they are associated) whose magnitudes the choice may affect, (2) to analyzing the different ways that the economy's Pareto imperfections interact to cause or not cause each category of economic inefficiency as well as the ways in which other parameters affect the amount of each category of economic inefficiency in the economy, given the Pareto imperfections it contains, and (3) investigating the prechoice magnitudes of the economy's Pareto imperfections and other economic-inefficiency-affecting parameters and the impacts of the choice whose economic efficiency is at issue on the magnitudes of the economy's Pareto imperfections and other economic-inefficiency-affecting parameters.
The protocol for economic-efficiency analysis that the antitrust-policy sequel uses meets these three requirements. I call this protocol a distortion-analysis protocol because it focuses to a substantial extent on the prechoice and postchoice magnitudes of the aggregate distortions that the economy's Pareto imperfections generate in the profits yielded by relevant exemplars of all categories of resource allocation--i.e., on the differences between the profits they respectively yield the resource-allocating principal and their respective economic efficiency. Such aggregate profit-distortions are salient because virtually all economic inefficiency is caused by "critical distortions" in the profits yielded by resource allocations that were or were not executed--i.e., by positive distortions in the profits yielded by resource allocations that took place that resulted in their being at least normally profitable for their respective resource-allocator principals despite the fact that they were economically inefficient and by negative distortions in the profits that would have been yielded by resource allocations that did not take place that resulted in their being less than normally profitable for their resource-allocator principals despite the fact that they would have been economically efficient.
I will now describe in some detail the distortion-analysis protocol that the policy-sequel uses to predict or post-diet the economic efficiency of antitrust-policy-covered business conduct and antitrust policies. I should state at the onset that this protocol is designed to be "third-best allocatively-efficient"--i.e., allocatively efficient given the fact that theoretical and empirical research is always allocatively-costly and is usually imperfect. The fact that the economic-efficiency-analysis protocol that the policy-sequel uses is third-best allocatively efficient is manifest in two of its features. First, at each of its stages, the protocol instructs the analyst to execute only those theoretical or empirical research-projects whose execution is third-best allocatively efficient--i.e., to execute only those research projects whose predicted allocative benefits (the allocative benefits they would generate by increasing the economic efficiency not only of the choice under consideration but of other choices as well by increasing the correctness of relevant theoretical results and the accuracy of relevant parameter-estimates) exceed their predicted allocative cost (the allocative value that the resources devoted to any relevant research-project would generate in their alternative employs plus the allocative cost of delaying any policy-decision to enable the decision maker to take advantage of the relevant research-results plus the economic inefficiency the government generates to finance the relevant research-project). This third-best-allocatively-efficient character of the protocol for economic-efficiency analysis that the policy-sequel uses is also manifest in the protocol's instructing the analyst to base his or her conclusions about the magnitude of various parameters on estimates/guesstimates. Second, the third-best-allocatively-efficient character of the protocol is also manifest in its instructing the economic-efficiency analyst to base his or her conclusions about the effect of the choices they are reviewing on the amount of some categories of resource misallocation in the economy as a whole on TBLE estimates/guesstimates of their impacts on the amount of such resource misallocation in a TBLE-large random sample of the economy's ARDEPPSes. I should acknowledge that this account of the third-best allocative efficiency of the economic-efficiency protocol that the policy-sequel uses ignores an infinite-regress problem for which I do not have a good solution--viz., the problem connected with the fact that, to proceed in a third-best-allocatively-efficient way, the economic-efficiency analyst will have to make a third-best-allocatively-efficient decision about whether to think about whether to think about whether to think about whether to think about ... the economic efficiency of executing particular research-projects.
In any event, the economic-efficiency-analysis protocol that the antitrust-policy sequel employs has five components. The first focuses on the impacts that the various types of business conduct that antitrust policy could cover and various antitrust policies would have on the first six categories of resource misallocation listed in Part 2B if they would not affect what I call the organizational allocative efficiency of any firm by enabling it to combine or precluding it from combining assets that are complementary for scale or non-scale reasons. All of these categories of economic inefficiency are "relative total allocative cost" (RTLC) resource misallocations--i.e., all result because resources are allocated in economically-inefficient proportions among uses in a single category in different ARDEPPSes or between two categories of use. The second component of the protocol focuses on the impacts that the various types of business conduct that antitrust policies could cover and various possible antitrust policies would have on the seventh and eighth categories of resource misallocation listed in Part 2B if they would not affect the extent to which one or more firms combined assets that were complementary for scale or nonscale reasons. All of these categories of economic inefficiency are non-poverty/inequality-generated "discrete choice" (DC) resource misallocations--i.e., resource misallocations that result because a resource allocator makes one rather than another discrete resource-allocating choice (e.g., chooses to use one known production process rather than another or chooses to buy a homogeneous product from one supplier rather than another). The third focuses on the impacts that antitrust-policy-covered conduct and possible antitrust policies have on economic efficiency by enabling firms to increase or preventing them from increasing their organizational allocative efficiency (the allocative efficiency of the production process they use to produce their outputs of given goods, the allocative efficiency of the QV investments they create at a given allocative cost, or the allocative efficiency of the production-process-research projects they execute at a given allocative cost) by combining assets that are complementary for scale or nonscale reasons. The fourth component of the protocol investigates the impact that various types of antitrust-policy-covered conduct and various antitrust policies have on the amounts of various subcategories of economic inefficiency the economy generates because of the poverty and/or income/wealth inequality in the society in question. And the fifth focuses on the impacts that various types of antitrust-policy-covered business conduct and various possible antitrust policies would have on three categories of allocative costs--the allocative transaction costs generated in the economy in question, the risk and uncertainty costs that relevant individuals bear and the allocative costs that the prospective bearers of such costs generate to reduce them, and the economic inefficiency the government generates when financing its operations.
The component of the protocol for analyzing the impact that any antitrust-policy-covered business conduct or any antitrust policy will have on the amounts of the various RTLC categories of resource misallocation the economy contains probably responds to The General Theory of Second Best more complicatedly than does any other component of the economic-efficiency analysis protocol that the antitrust-policy sequel uses. Roughly speaking, the RTLC-resource-allocation-focused component of the economic-efficiency-analysis protocol instructs the analyst to predict or post-diet the impact of any antitrust policy on the six RTLC categories of resource misallocation in the following way:
1. derive (presumably-different) formulas for the aggregate distortion in the profits yielded by any economics-marginal resource allocation associated respectively with each RTLC category of resource misallocation--six formulas that collectively manifest the different ways in which all relevant Pareto imperfections in an economy interact to distort the profits yielded by any economics-marginal resource allocation (the least-profitable but not-unprofitable allocation in that category in a specified ARDEPPS) associated respectively with each RTLC category of resource misallocation;
2. identify a third-best-allocatively-efficiently-large random sample of the economy's ARDEPPSes--the sample of ARDEPPSes on which this component of the distortion-analysis protocol will focus;
3. generate third-best-allocatively-efficient estimates or guesstimates of the pre-choice magnitudes of the parameters in the formulas derived in Step (1) for the ARDEPPSes to be studied and calculate the pre-choice and post-choice magnitudes of the aggregate distortions in the profits yielded by the economics-marginal resource allocations in the categories associated with the RTLC categories of economic inefficiency in all ARDEPPSes that are to be studied;
4. in each studied ARDEPPS, generate a third-best-allocatively-efficient estimate or guesstimate of the attributes of the "resource-allocation marginal allocative product curve" (MLP .../...) for each category of resource allocation associated with a RTLC category of resource misallocation in any ARDEPPS between the resource-allocation marginal-allocative-product figures associated with the estimates/guesstimates of the prechoice and postchoice aggregate distortions in the profits yielded by the usually-changing economics-marginal resource allocation in the specified category in the specified ARDEPPS where the curve in question is a curve in a diagram whose vertical axis measures dollars and whose horizontal axis measures the total allocative cost of the resources devoted to a curve-specified category of use in a specified ARDEPPS (specified by an entry that replaces the first ellipsis in the subscript of the symbol for the curve) after having been withdrawn from a curve-specified category of use in one or more specified ARDEPPSes (specified by an entry that replaces the second ellipsis in the subscript of the symbol for the curve) and the height of the MLP .../... curve at any TLC quantity indicates the net allocative benefits that would be generated by the use of the last (say) SI in resources (measured by their allocative cost) that would bring TLC for the specified category of resource allocation from and to specified ARDEPPSes to the quantity in question and where the height of any MLP .../... curve at any TLC quantity at which the last specified resource allocation yields zero profits (as it would if the allocation were both economics-marginal and mathematics-marginal and were executed by a sovereign maximizer) equals SI minus the aggregate distortion in the profits the relevant resource allocation yields;
5. in each ARDEPPS, for each category of resource allocation associated with a RTLC category of resource misallocation, derive from the formulas generated in Step (1) and the parameter estimates/guesstimates made in Step (2) and Step (4) estimates/guesstimates of the impact of the policy on the quantity of that category of RTLC resource misallocation the ARDEPPS contains; and
6. sum the estimates/guesstimates of the policy's impacts on each RTLC category of resource misallocation in all studied ARDEPPS to calculate the policy's impact on the total amount of each such category of resource misallocation the studied ARDEPPSes contain and derive an estimate/guesstimate of the policy's impact on the total amount of each such category of resource misallocation the economy contains by multiplying each such studied-ARDEPPSes choice-generated change-in-resource-misallocation total by the ratio of the estimated allocative cost of the resources devoted to the category of resource allocation in question in the economy as a whole to the estimated allocative cost of the resources devoted to the category of resource allocation in question in the studied ARDEPPSes.
Although the following conclusion certainly is contestable and may be wrong, I believe that it will prove to be third-best allocatively efficient to use a protocol to analyze the impact of antitrust-policy-covered business conduct or any antitrust policy on the DC categories of resource misallocation that is less complicated than the protocol that I think will be third-best allocatively efficient to use when impacts on RTLC categories of resource misallocation are at issue. More specifically, I think it will prove to be third-best allocatively efficient to use the following protocol to analyze the impact of antitrust policies on the categories of resource misallocation referenced in the seventh and eighth items in the Part 2B list:
1. develop formulas for the aggregate distortion in the profits yielded by any DC (discrete-choice) category of resource misallocation (the distortion in those profits generated by all exemplars of all types of Pareto imperfections) that equate those aggregate profit-distortions with the sum of what I call the seven step-wise distortions that all exemplars of each individual type of Pareto imperfection would generate in the profits of each type of DC resource allocation--viz., the distortions that respectively all exemplars of each type of Pareto imperfection would generate in such profit-figures given the exemplars of the other types of Pareto imperfections the relevant economy contains;
2. devote a third-best-allocatively-efficient amount of resources to determining (A) the frequency with which the sum of the six step-wise distortions in the profits yielded by each category of DC resource allocation that are generated by all exemplars of each individual type of Pareto imperfection other than imperfections in seller competition are non-zero when the imperfection-in-seller-competition- oriented step-wise profit-distortion in that profit-figure is non-zero, (B) the relative absolute sizes of the sum of the six relevant non-imperfection-in-seller-competition-oriented step-wise distortions and the imperfection-in-seller-competition-oriented step-wise distortions in those cases in which both this summed-step-wise-distortion figure and the imperfection-in-seller-competition step-wise distortion are non-zero, and (C) the correlation between the summed-step-wise-distortion figures and the imperfection-in-seller-competition-oriented step-wise distortions when both deviate significantly from zero; and
3. analyze the impact of the business conduct or antitrust policy under review on each DC category of resource misallocation (A) by devoting a TBLE amount of resource to estimating/guesstimating the amount of that category of DC resource misallocations the relevant economy would contain if the business conduct under review were not engaged in or the antitrust policy under review were not adopted--i.e., (i) to estimating/guesstimating (a) the incidence of discrete choices in the relevant category whose profit-yield would be inflated in those circumstances (for which the aggregate profit-distortion would be positive), (b) the relevant DC-by-DC magnitudes of the positive distortions the profits yielded by those DC choices whose profits would be inflated in question, and (c) the profits yielded by the relevant DC choices and (ii) to generating an estimate/guesstimate of the amount of that category of DC resource misallocation the relevant economy would contain in the above circumstances by ascertaining the number of DCs in the relevant category for which the positive profit-distortion exceeded the profits yielded and the average amount by which in those cases the positive profit-distortion exceeded the profits yielded and multiplying the two amounts, (B) by devoting a TBLE amount of resources to making the same estimates/guesstimates and calculations for the situation that would prevail if the relevant business conduct were engaged in or the relevant antitrust policy were adopted (inter alia, by examining how the conduct or policy would affect the aggregate distortion in the profits yielded by DC choices in the relevant category by altering the incidence and magnitudes of the economy's various Pareto imperfections), and (C) comparing the estimates/guesstimates of the total amount of the relevant category of DC misallocation in the economy as a whole pre-conduct/policy and post-conduct/policy: at a more-refined level, if (as 1 suspect is true in the vast majority of relevant cases) (i) the sum of the six non-imperfection-in- seller-competition-oriented step-wise distortions in the profits yielded by DC resource allocations for which the absolute imperfection-in-seller-competition-oriented stepwise profit-distortion is significantly different from zero either is near zero or has an absolute value that is much smaller than the absolute value of the imperfection-in-seller-competition-oriented step-wise distortion in the relevant profit-figure and is uncorrelated with the imperfection-in-seller-competition- oriented step-wise distortion in the relevant profit-figure and (ii) it would be not only allocatively expensive but given (i) prohibitively allocatively expensive to identify the cases in which the above relationships do not obtain, analyze both the pre-choice magnitudes of the relevant aggregate profit-distortions and the impact of the conduct or policy under review on the aggregate distortion in the profits that the relevant discrete choices generate on the assumption that the former figure equals the imperfection-in-seller-competition-oriented step-wise distortion in those profits and the latter figure equals the conduct-generated or policy-induced change in the imperfection-in-seller-competition-oriented step-wise distortion in that profit-figure, but if there is reason to believe that the assumptions delineated after (i) and (ii) above are not accurate either in general or in relation to some identifiable DC resource allocations, analyze both the pre-choice aggregate profit-distortions and the conduct/policyinduced changes in the aggregate profit-distortions in a way that takes appropriate account respectively of the sums of the six relevant pre-choice non-imperfection-in-seller-competition individual-Pareto-imperfection-oriented step-wise profit-distortions and the predicted impact of the conduct or policy on those sums as well as of the relevant pre-choice imperfection-inseller-competition-oriented step-wise profit-distortions and the predicted impacts of the policy on the latter step-wise distortions. (N.B.: this account reflects my assumption that it will not be TBLE to base estimates/guesstimates of economy-wide DC resource misallocation on estimates/guesstimates of their magnitudes in a sample of the economy's ARDEPPSes because I suspect that DC resource misallocations of any sort are most likely to occur in particular ARDEPPSes that will be TBLE to identify.)
The third component of the distortion-analysis protocol for predicting or post-dieting the economic efficiency of antitrust-policy-covered business conduct or antitrust policies focuses on the possibility that the business conduct or antitrust policy in question will increase economic efficiency by enabling one or more firms to combine assets that are complementary for scale or nonscale reasons or will decrease economic efficiency by preventing one or more firms from combining assets that are complementary for scale or nonscale reasons--i.e., will on this account either increase or decrease the economic efficiency of the production processes the firm(s) use, one or more of the QV investments the firm(s) create(s) with resources of given allocative cost, and/or one or more of the PPR projects the firm(s) execute(s) with resources of given allocative cost (in my terminology, will on this account either increase or decrease one or more firms' organizational allocative efficiency). Business choices (to engage in or not to engage in mergers, acquisitions, joint ventures, or internal growth) and antitrust policies (to allow or prohibit mergers, acquisitions, joint ventures, and internal growth) can either increase businesses' organizational allocative efficiency by securing the combination of complementary assets or decrease businesses' organizational allocative efficiency by resulting in complementary assets not being combined.
The fourth component of the distortion-analysis protocol for predicting or post-dieting the economic efficiency of antitrust-policy-covered business conduct or various antitrust policies focuses on the ways in which such conduct or policies will increase or decrease economic efficiency by decreasing or increasing poverty and income/wealth inequality (say, by decreasing or increasing prices). This possibility is important because, especially in an economy that contains Pareto imperfections or would contain Pareto imperfections if some people were poor or income and wealth were distributed unequally, poverty and/or income/wealth inequality generate economic inefficiency (1) by increasing the amount of misallocation the economy generates because economically-efficient investments in the human capital of children and adults are not made, (2) by increasing the amount of misallocation that consumption-choices (say, to buy and operate ugly, breakdown-prone, accident-prone, noisy, air-polluting cars or to rent ugly and fire-and-disease-spreading housing units) generate because it is advantageous for individuals when they are poor to make external-cost-generating consumption-choices that are economically inefficient, (3) by increasing the amount of misallocation generated because the relevant economy's members make privately-disadvantageous consumption-choices that are economically inefficient--i.e., because (A) by reducing their preparedness for schooling and the quality of the education they receive both inside and outside schools, the poverty of the poor increases both the frequency with which the individuals who are poor fail to understand the attributes of products and their full cost to them and the frequency with which they do their maths wrong or make consumption choices unthinkingly and (B) by increasing their frustration and unhappiness, the poverty of individuals who are poor leads them to discount future benefits too highly from the perspective of their own lifetime welfare, (4) by increasing the amount of misallocation that poverty causes by inducing individuals who are poor to make economically-inefficient decisions to perform dangerous, lawful labor in all the ways that it causes poor individuals to make economically-inefficient consumption decisions that are not in their interest and, in a society in which poor individuals who have been injured at work or their families receive various types of government-transfers for which their non-poor counterparts would not be eligible, by rendering economically inefficient some decisions to perform dangerous lawful labor that are objectively beneficial for those who make them (perhaps given their concern for their families), (5) by increasing the amount of misallocation that people who are poor or people who have significantly less income and wealth than does the average participant in the relevant economy generate by engaging in economically-inefficient criminal activities by making them less concerned about the impact of their criminal choices on their victims (by alienating them), by reducing the difference between the attractiveness of life in prison and life without successful crime outside prison, by causing them to use too high a discount-rate to calculate the present value of their future welfare, and by causing them to be too optimistic about the profitability of crime for reasons other than the discount-rate they apply to their future welfare, (6) by reducing the political influence of the individuals who are poor and thereby increasing the amount of economic inefficiency the government of the relevant society generates because its choices are based on a calculation that places a lower weight on the average dollar gained or lost by the individuals who are poor or who have lower-than-average income and wealth than on the average dollar gained or lost by the individuals who are not poor or who have higher-than-average income and wealth, and hopefully (7) by increasing the economic inefficiency the economy generates because its distribution of income and wealth disserves (on balance) the "external preferences" of the members of and participants in the society in question (their nonparochial preferences for the resources and opportunities that others have). In any event, for the above reasons, the protocol that the antitrust-policy sequel uses to predict or post-diet the impact of various antitrust-policy-covered conduct and various antitrust policies on economic efficiency will take into account both the policy's impact on poverty and income/wealth inequality and the other parameters that the preceding list implies will affect the economic-efficiency consequences of the policy's distributive impact.
The fifth and final component of the economic-efficiency-analysis protocol that the antitrust-policy sequel uses focuses on the impact of antitrust-policy-covered business conduct and antitrust policies on allocative transaction costs, risk and uncertainty costs (which are allocative as well as private) and the allocative costs that individuals generate to reduce the amount of such costs they bear, and the economic inefficiency that governments generate when financing their operations. I start with allocative transaction costs, which will usually not equal their private counterparts (because Pareto imperfections will usually distort the private value of the resources used up as transaction costs would have generated for their alterative employers and hence their private cost to the actor who used them up as transaction costs). The business conduct that antitrust policies cover (the execution of mergers, acquisitions, and joint ventures, the use of fancy pricing-techniques, the practice of contrived oligopolistic pricing, predatory pricing, or other types of predatory conduct, the use of various contractual surrogates for vertical integration, etc.) can generate allocative transaction costs directly and can also generate such costs indirectly (for example, by causing individuals who lose their jobs when a merger leads to the rationalization of production to apply for government-transfers). Antitrust policies can also affect the amount of allocative transaction costs generated in the relevant economy both directly (the allocative transaction cost the government generates directly when devising, adopting, and enforcing the policy) and indirectly (by inducing businesses to make allocative-transaction-costly moves to conceal their illegal behavior and causing them to generate allocative transaction costs to defend themselves in criminal and civil litigation, and [in the United States] by inducing private parties to generate allocative transaction costs to make legal claims against businesses that they claim have inflicted recoverable losses on them by violating the antitrust laws). Antitrust policies can also reduce the amount of allocative transaction costs that are generated by deterring business conduct (e.g., contrived oligopolistic pricing or the use of fancy pricing-techniques) that would have generated allocative transaction costs directly or indirectly (say, by causing employees it would render redundant to make claims for government transfers). The antitrust-policy sequel takes third-best-allocatively-efficient account (see below) of all such allocative-transaction-cost consequences.
Business conduct and antitrust policies can also affect the amount of risk and uncertainty costs individuals bear and the allocative costs that individuals generate to reduce the risk and uncertainty costs they bear. I suspect, however, that although antitrust policies will affect the costs that some face because of the risk of their being the plaintiff or defendant in an antitrust case and will induce some individuals to make allocative-costly moves to reduce such risk costs, it will not prove to be third-best allocatively efficient to consider these possibilities beyond the consideration given to them in the allocative-transaction-cost analysis.
The final "allocative cost" that the policy-sequel's protocol for analyzing the economic efficiency of any antitrust-policy-covered business conduct or any antitrust policy addresses is the impact that the conduct or policy will have on the fiscal position of the government and hence on the amount of economic inefficiency the government generates when financing its operations. Specifically, on this account, the protocol the antitrust-policy sequel uses instructs the analyst (1) to estimate/guesstimate the fiscal impact of the conduct or antitrust policy under review (its impact on the fines, tax revenues, and court fees the relevant government collects and the private transaction costs it incurs to devise, pass, and implement a relevant antitrust policy) and (2) to estimate/guesstimate the economic efficiency of the decisions that the choice under review causes the relevant government to make by altering its fiscal position--(A) if the analyst concludes that the choice under review will worsen the government's fiscal position, (i) estimate/guesstimate the extents to which the government will respond to this reality by raising tax-rates or imposing new taxes, by raising the prices it charges for goods and services it supplies, by printing money or selling bonds to "finance" the relevant deficit, or by eliminating other expenditures and (ii) analyze the economic-efficiency effects of these government responses and (B) if the analyst concludes that the choice under review will improve the government's fiscal position, (i) estimate/guesstimate the extents to which the government will respond to this reality by lowering tax-rates or eliminating some taxes, lowering the prices it charges for goods and services, destroying some of the money it possesses or retiring some government debt, or making other expenditures and (ii) analyzing the economic efficiency of each of these responses.
I will now attempt to make clear why it is important to use the distortion-analysis protocol just outlined to analyze the economic efficiency of a policy-choice--in particular, why this protocol will generate more accurate economic-efficiency predictions than will conventional first-best-allocative-efficiency analyses (which [usually implicitly] assume that the only Pareto imperfection in the relevant economy is the Pareto imperfection that the policy under review targets or directly affects). For this purpose, I will compare Oliver Williamson's canonical (first-best-allocative-efficiency) analysis (119) and a simplified version of the TBLE distortion-analysis-protocol analysis of the economic efficiency of horizontal mergers that convert a perfectly-competitive industry into a pure monopoly and, in the process of so doing, simultaneously (1) generate allocative production efficiencies that result in the merged firm's facing lower marginal costs than the merger partners would have faced after the date of the mergers and (2) despite this fact, create a situation in which the monopolistic merged firm's unit output is lower than the unit output that the merger partners would have produced after the date of the mergers (by creating a merged company whose profit-maximizing [supra-marginal-cost] price is sufficiently higher than its marginal cost to exceed the profit-maximizing price that each individual merger partner would have charged after the date of the mergers (the price that would have equaled the [higher] marginal cost each of the merger partners would have had to incur to produce their respective marginal units of output at the output at which their marginal cost curves cut their average total cost curves from below).
I will first compare Williamson's and my analyses of the two economic-efficiency effects of such mergers that Williamson does consider: (1) their reducing the cost the merged firm incurred to produce its post-merger output below the cost the merger partners would have incurred to produce that output after the date of the mergers and (2) its affecting economic efficiency by reducing the merged firm's unit output below the combined unit outputs that the merger partners would have produced after the date of the mergers had they not merged. For convenience, I will address the reduced-production-cost issue on the assumption that the merged firm faced and the mergers partners would have faced horizontal MC curves and that the mergers lowered costs exclusively by lowering the MC curve the merged firm faced below the MC curve that each merger partner would have faced after the date of the mergers. Williamson concludes that the economic-efficiency gain that the mergers generate by lowering the total private variable cost that the merged firm incurs to produce its unit output below the total private variable cost that all the merger partners acting separately would have incurred to produce that unit output after the date of the mergers equals the private-cost reduction just specified--i.e., equals the mergers-generated reduction in per-unit marginal cost times the merged firm's unit output. Williamson's conclusion derives from his implicit first-best assumptions that (1) the MC curve that would have been operative absent the mergers (the MC curves the merger partners would have faced after the date of the mergers) coincides with (is identical to) the marginal allocative cost (MLC) curve that would have been operative absent the mergers (the curve that indicates the allocative value that the resources devoted to producing successive units of the good in question absent the mergers would have generated in their alternative uses) and (2) the MC curve the merged firm faced after the date of the merger coincides with the MLC curve for the merged firm's production after the date of the mergers. Although these assumptions would always be correct if the economy contained no Pareto imperfection other than the imperfection in seller price-competition the mergers would generate, they will be correct only rarely and fortuitously in the real world, which contains a myriad of other Pareto imperfections that distort the private cost of the resources the merged firm did use and the merger partners would have used after the date of the mergers to produce the merged firm's unit output. For example, if the resources that the merged firm did use and the merger partners would have used to produce the merged firm's unit output were withdrawn/would have been withdrawn from the production of units of output of other goods by imperfect competitors that faced downward-sloping demand curves and did not engage in price discrimination or use any other fancy pricing-techniques, the imperfections in seller price-competition that these alternative resource-users faced would cause the MC curves that would have been faced by the merger partners after the date of the mergers to be lower than the MLC curve for the merger partners' production after the date of the mergers and would cause the MC curve the merged firm faced to be lower than the MLC curve for the merged firm--i.e., would deflate the private cost respectively to the merger partners after the date of the merger and the private cost to the merged firm of producing the merged firm's unit output by deflating those resources' private values to their alternative employers (which determines the prices that the merger partners and the merged firm, respectively would have to pay and did pay for them)--i.e., by causing these resources' private values to their alternative users to be lower than the allocative values the resources would have generated in their alternative users' employ. This conclusion reflects three facts: (1) the private value the resources would have generated for such alternative users is their marginal revenue products for these alternative users--(the marginal physical products the resources would have yielded their alternative employers) times (the [average per-unit] marginal revenues their alternative users would have respectively obtained by selling the output-units that the resources in question would have produced for them); (2) the allocative values that these resources would have generated in these alternative uses is their marginal allocative products in these alternative uses--if, for simplicity, one ignores the distorting effects of other types of Pareto imperfections, this marginal allocative product equals (the relevant resources' marginal physical products for their alternative employers) times (the prices for which the units of the other goods the resources would have enabled their alternative employers to produce could have been sold); and (3) marginal revenue is less than price for imperfectly-competitive sellers that face downward-sloping demand curves and do not engage in price discrimination. But if the two MC curves are lower than their MLC counterparts, Williamson's conclusion that that economic-efficiency gains the mergers generate by reducing the private cost of producing the merged firm's output equals the private-cost savings they enable the merged firm to realize--his conclusion that the per-unit reduction in marginal costs equals the per-unit reduction in marginal allocative costs--will be incorrect. If, for example, each MC curve understates MLC by 20%, the per-unit private-marginal-cost saving will be 20% lower than the per-unit marginal-allocative-cost saving: thus, if the merger reduces private marginal cost per unit from $10 to $5 and marginal allocative cost from $12 to $6, the private cost-saving in question will be $10-$5 = $5 per unit of output but the relevant allocative-cost saving will be $12-$6 = $6 per unit of output.
The actual relationship between marginal cost and marginal allocative cost will depend on the percentages of resources used to produce successive units of the product in question that are withdrawn from UO-increasing, QV-creating, and PPR-executing uses and on the signs and magnitudes of the percentage-distortions in the private benefits that the resources in question would have conferred on their alternative employers in each of these three categories of use. These percentage-distortion figures will depend not only on imperfections in seller price-competition in the rest of the economy but on the incidences and magnitudes of many other types of Pareto imperfections, which would probably be third-best allocatively efficient to take into account. I have neither the space nor the data to execute the relevant analysis here with anything like a TBLE degree of precision. However, I am confident that MC is lower than MLC and that Williamson's first-best-allocative-efficiency analysis underestimates the allocative-efficiency gain the mergers he is analyzing generate by reducing the private cost the merged firm incurs to produce its unit output below the private cost the merger partners would have had to incur to produce that unit output after the date of the mergers.
The first-best character of Williamson's analysis--i.e., the fact that it assumes that the relevant economy contains no Pareto imperfection other than the imperfection in seller price-competition the mergers at issue create--also undercuts his investigation of the second economic-efficiency issue he addresses--the economic-efficiency consequences of the reduction in the unit output of the good that the merger partners and the merged firm produce that he stipulates the mergers in question would cause. Williamson claims that this reduction in unit output will cause economic inefficiency equal to the area between the industry demand curve for the product in question and the MC curve for that good that would be operative after the date of the mergers if no mergers were executed between the good's higher no-merger output and its lower post-merger output. Once more, this conclusion would be correct if the relevant economy were otherwise-Pareto-perfect since in that case the industry demand curve for the good (DD) would coincide with the marginal allocative value curve for the good (MLV) (the curve that indicates the allocative value of successive units of the good) (120) and the no-merger MC curve for the good would coincide with the no-merger MLC curve for the good. However, in our actual, highly-Pareto-imperfect economy, the other Pareto imperfections the economy contains may well cause MLV to depart from DD and MLC to depart from MC. Thus, as we just saw, even if I ignore the possible divergence of MLV from DD, if the units of the good the merger partners produced that would not be produced by the merged firm would be produced with resources withdrawn from the production of other goods by non-discriminating imperfect competitors that face downward-sloping demand curves, MLC will be higher than MC (if no other Pareto imperfections critically affect this relationship).
Admittedly, the actual relationship between MLC and MC will depend on the percentages of the resources that would have been devoted to the production of the units of output that the merged firm would not produce that the merger partners would have produced after the date of the mergers that would have been withdrawn respectively from alternative unit-output-increasing, QV-creating, and PPR-executing uses and the percentage-distortions in the private benefits those resources would have generated in their alternative uses to their alternative users, which percentages will depend on the incidence and magnitudes of a large variety of Pareto imperfections. Once more, I have neither the space nor the data to execute the relevant analysis with anything like the TBLE degree of precision. Nevertheless, I will again state that I am confident that the relevant MLC curve is higher than the MC curve that Williamson assumes coincides with it. If MLC is higher than MC, Williamson will have overestimated the amount of economic inefficiency that the category of mergers he is examining will generate by reducing the unit output of the good the merger partners produce. Indeed, if MLC is sufficiently above MC, the reduction in unit output may actually increase economic efficiency. Put crudely, if the good in question (say, X) is produced with resources withdrawn exclusively from the production of units of other goods (say, Z1 ... n) and neither the good in question nor the other goods has any complements, the merger-induced increase in the price of X (in [P.sub.x]) will increase/decrease economic efficiency if it decreases/increases the difference between ([P.sup.*.sub.x]/[MC.sup.*.sub.x]) and weighted average ([P.sup.*.sub.z1 ... n]/[MC.sup.*.sub.z1 ... n]) where the asterisks over the P figures indicate that they have been adjusted to make each [P.sup.*] figure equal the associated marginal allocative value figure and the asterisks over the MC figures indicate that they have been adjusted to make [MC.sup.*.sub.x]/[MC.sup.*.sub.Z1 ... n] (where the weights assigned to the respective members of the set of products Z1 ... n are proportionate to the allocative value of the unit[s] of each product in that set that would be sacrificed to produce the marginal unit of X) equal the rate at which the relevant economy can transform the relevant bundle of Z1 ... n into X. (Not to worry!)
But I should not bog you down with too much detail. The important point is that, if I am right that [MLC.sub.x] exceeds [MC.sub.x] in our actual, highly-Pareto-imperfect economy, Williamson's canonical analysis will have overestimated the economy-efficiency loss that the mergers he is analyzing will cause by increasing the price and decreasing the unit output of the good the merger partners produce at the same time that it underestimates the economic-efficiency gain the mergers will yield by generating private production-efficiencies that have allocative counterparts. If the mergers in question and any antitrust policy one could adopt in relation to them had no other economic-efficiency consequences, an analyst that used my distortion-analysis protocol rather than Williamson's first-best approach to analyze the economic efficiency of the mergers in Williamson's category would conclude that such mergers were systematically more economically efficient than Williamson's analysis concluded. To the extent that the differences in question were critical--i.e., to the extent that analysts who used my approach would find economically efficient mergers that Williamson would find economically inefficient and/or (if my more specific conclusions are wrong) that analysts who used my approach would find economically inefficient mergers that Williamson would find economically efficient, the substitution of my approach for Williamson's would increase economic efficiency (if actual policy-choices were determined by their economic efficiency and analysts who used my approach took appropriate account of the allocative cost of the research they did).
So far, I have focused on the two categories of economic-efficiency impacts that Williamson's canonical analysis does consider. However, the distortion-analysis protocol would also address a number of other possibilities that Williamson ignores. I will now address four such additional possibilities.
Williamson's failure to consider the first two of these possible consequences probably reflects nothing more than the page-constraints that economics journals place on authors and Williamson's correct judgment that, across all cases, these issues are less important that the ones he analyzed. First, the distortion-analysis protocol would consider the transaction costs that the mergers Williamson scrutinized would generate and the transaction-cost consequences of any government responses to them. Perhaps more notably, the distortion-analysis protocol would take account not only of the referenced private transaction costs but also of the fact that various Pareto imperfections would distort those private transaction costs--would cause them to diverge from their allocative counterparts. Since all other economists fail to consider the possible difference between private and allocative transaction costs and no such differences would arise on Williamson's otherwise-Pareto-perfect assumptions, it seems fair to conclude that, had Williamson addressed transaction-cost issues, he would have equated private with allocative transaction costs.
Second, the distortion-analysis protocol would consider the fiscal impacts of the mergers Williamson analyzed (e.g., their impacts on business profits and the taxes paid by businesses) and the fiscal impacts of the various responses government could make to them and, derivatively, the impacts that these mergers and any responses government made to them would have on the amount of resource misallocation government generates when financing its operations. Although Williamson's failure to consider these issues probably reflects the page-constraints he faced and a defensible set of priorities, it is worth noting that economists never address these issues--do not do so even when they are writing book-length economic-efficiency analyses whose lengths are not critically relevantly affected by page-constraints.
Williamson's failure to address the next two sets of possible economic-efficiency consequences of the mergers he analyzed that the distortion-analysis protocol would investigate is more troubling because economists in general fail to consider the impact of choices on these categories of economic inefficiency. The first set of such consequences is QV-investment related. Williamson is assuming that the product produced by the merger partners and merged firm is a homogeneous product whose attributes would not in any circumstances be altered. I think there are virtually no such products: even when no-one will ever find it profitable to alter the physical attributes of a product--i.e., to introduce a product variant with somewhat different physical attributes or to differentiate the image of a product with given physical attributes, the total attributes of a product include its average speed of delivery in all cases in which the demand for it fluctuates through time, and firms can increase the quality of their total product by making a QV investment in capacity or inventory that increases the average speed with which they can supply customers with units of their physical product throughout a fluctuating-demand cycle. Williamson ignores this reality--specifically, ignores the impact that the mergers he is considering will have on the equilibrium quantity of QV investment (on the equilibrium amounts of capacity and inventory) in the relevant area of product-space. In the one direction, Williamson's mergers will tend to increase equilibrium QV investment in the relevant ARDEPPS by raising the prices that will be charged for "the relevant product" at different ARDEPPS QV-investment quantities and hence the supernormal rate-of-return that all QV investments in that area of product-space will generate at any ARDEPPS QV-investment quantity and may increase equilibrium QV investment in that ARDEPPS by generating dynamic efficiencies. In the other direction, if the barriers to entry facing the best-placed potential entrant into the relevant ARDEPPS are critically higher than the barriers to QV-investment expansion that would have faced one or more of the merger partners and/or if Williamson's mergers would replace merger partners that would have faced no monopolistic QV-investment disincentives with a merged firm that faces substantial monopolistic QV-investment disincentives, Williamson's mergers would tend to reduce equilibrium QV investment in the relevant ARDEPPS on these accounts. The distortion-analysis protocol would instruct the analyst to estimate/guesstimate the impact of Williamson's mergers on equilibrium QV investment in the relevant area of product-space and then to investigate the impact of any change in equilibrium QV investment in that ARDEPPS on interARDEPPS QV-to-QV, UO-to-QV or QV-to-UO, and PPR-to-QV or QV-to-PPR resource misallocation.
The second member of this second set of possibilities that Williamson (like economists in general) ignores but that the distortion-analysis protocol would take into account are all PPR-related. Thus, the distortion-analysis protocol would consider the possibility that Williamson's mergers might deter economically-efficient PPR that aimed to discover a less-accident-and-pollution-loss-prone production process if, for example, the merger partners and the merged firm were liable in tort only for the consequences of their negligence, their failure to do PPR was never assessed for negligence, and no independent research firm could profitably execute the relevant PPR. The distortion-analysis protocol would also instruct the economic-efficiency analyst to investigate (1) the impact that the merger in question would have on the amount of PPR done to discover lower-marginal-cost production processes (by causing the pre-discovery unit output of the merged firm to be lower than the pre-discovery outputs that the merger partners would have "collectively" produced after the date of the merger and thereby [A] lowering the private cost-saving such a discovery would enable the merged company to obtain on its pre-discovery output below the benefits such a discovery would have yielded any merger partner that made it after the date of the discovery by lowering that merger partner's total variable costs on its pre-discovery output and putting it in a position to make money by licensing rivals to use its discovery to produce their pre-discovery outputs and [B] causing the profits such a discovery would enable the merged firm to realize by expanding its output to differ from the profits the discovery would have enabled any merger partner that made it to realize by expanding its output and licensing its rivals to use the discovery to expand their outputs). The distortion-analysis protocol would also instruct the economic-efficiency analyst to estimate/guesstimate the dynamic PPR-related efficiencies the mergers would generate and the impact that any such efficiencies would have on economic efficiency (1) by creating a merged firm that executes more-allocatively-valuable PPR projects than the merger partners would have executed with the resources (measured by their allocative cost) that they would have devoted to PPR after the date of the mergers and (2) by increasing or decreasing economic efficiency (inter-ARDEPPS PPR-to-PPR misallocation, QV-to-PPR or PPR-to-QV misallocation, and/or UO-to-PPR or PPR-to-UO misallocation) by creating a merged firm that devotes a different amount of resources (measured by their allocative cost) to PPR than the merger partners would have devoted to PPR after the date of the mergers. Finally, the distortion-analysis protocol would also instruct the economic-efficiency analyst to consider the possibilities that Williamson's mergers (1) might increase economic efficiency by creating a merged firm that executes less-economically-inefficiently-duplicative PPR projects than the merger partners would have executed with the resources that the merger partners would have devoted to PPR after the date of the mergers and (2) might either increase or decrease economic inefficiency by creating a merger firm whose ability to execute less-duplicative PPR projects than the merged firms would have executed after the date of the mergers makes it profitable for it to devote more resources to PPR than the merger partners would have found profitable to devote to PPR after the date of the mergers.
Because the distortion-analysis protocol would take into account many economic-efficiency-relevant impacts of the mergers Williamson studied that Williamson's analysis ignores, its assessment of the economic efficiency of these mergers will be more accurate than a Williamson analysis would be. Once more, to the extent that the distortion-analysis protocol correctly assesses mergers in Williamson's category to be economically efficient or economically inefficient when Williamson's analysis would have generated the opposite conclusion, the substitution of the distortion-analysis protocol for Williamson's analysis will increase economic efficiency if policy-decisions are determined by their economic efficiency and the distortion-analysis analyst takes economically-efficient account of the allocative cost of any analysis he or she executes.
This "coda" has delineated and made some effort to justify the distortion-analysis protocol that the antitrust-policy sequel to the antitrust-law books on which this special issue focuses uses to analyze the economic efficiency of antitrust-policy-covered business conduct and various antitrust policies. I have included this material for two reasons: (1) to make absolutely clear the difference between antitrust-policy analysis and the analysis of the legality of business conduct under U.S. or E.U. antitrust law and (2) to pique readers' interest in the antitrust-policy sequel and the Distortion-Analysis Protocol book I intend to publish before the policy-sequel to explain and justify the distortion-analysis protocol the antitrust-policy sequel will use to assess the economic efficiency of antitrust-policy-covered business conduct and various antitrust policies.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The author(s) received no financial support for the research, authorship, and/or publication of this article.
(1.) The two volumes of Economics and the Interpretation and Application of U.S. and E.U. Antitrust Law were published by Springer in 2014. The subtitle of Vol. I is Basic Concepts and Economics-Based Legal Analyses of Oligopolistic and Predatory Conduct. Vol. I will be cited hereinafter as Markovits, Economics of Antitrust Law 1. The subtitle of Vol. II is Economics-Based Legal Analyses of Mergers, Vertical Practices, and Joint Ventures. It will be cited hereinafter as Markovits, Economics of Antitrust Law II.
(2.) This policy-study will also be published by Springer, hopefully in 2017.
(3.) 26 Stat. 209 (1890) (codified as amended at 15 U.S.C. Sections 1-11) (hereinafter Sherman Act).
(4.) Natural oligopolistic conduct is conduct whose ex ante perpetrator-perceived profitability was critically affected by the perpetrator's belief that its rivals' responses would or might be influenced by their belief that it would have the opportunity to react to their responses and would find it inherently profitable to react to any undercutting or undermining response that they would otherwise find profitable in one or more ways that would render those responses unprofitable for them. The law-study uses the expression "undercutting response" to refer to a response to an individualized oligopolistic price that would render it unprofitable and the expression "undermining response" to refer to a response to an across-the-board oligopolistic price that would render it unprofitable. For the referents of "individualized price" and "across-the-board price," see Part 3C infra. The Sherman Act does not cover natural oligopolistic conduct because such conduct involves neither the making of an anticompetitive agreement nor the making and carrying out of an anticompetitive threat or promise. My contestable conclusion that the Sherman Act also does not cover any unsuccessful attempt to enter into an anticompetitive agreement (an agreement in restraint of trade) reflects three facts: (1) the section of the Sherman Act (Section 1) that prohibits agreements in restraint of trade does not explicitly prohibit unsuccessful attempts to form agreements in restraint of trade; (2) the laws of the U.S. central government do not contain a general attempt statute (i.e., a statute that declares illegal any unsuccessful attempt to commit an act that would be illegal if successfully completed); and (3) the U.S. courts have consistently refused to read attempt provisions into statutes that do not contain them. I think that these facts imply that it would be incorrect as a matter of law for U.S. courts either to read an attempt provision into Section 1 or to classify an unsuccessful attempt to enter into an anticompetitive agreement to be an "attempt to monopolize" (which is prohibited by Section 2). I should add that, for constitutional reasons, the Sherman Act should be given a so-called "saving construction" under which it would also be deemed not to cover any attempt (whether successful or not) by an individual actor to secure a legislative, administrative, or judicial decision that would reduce the absolute attractiveness of the best offer(s) against which it would have to compete in some way that would reduce economic efficiency in an otherwise-Pareto-perfect economy (see n.10 infra) or any agreement by two or more actors to cooperate in securing such a government decision so long as the conduct in question did not violate an appropriate criterion of fair play.
(5.) 38 Stat. 730 (1914) (codified as amended at 15 U.S.C. Sections 12-27) (hereinafter Clayton Act).
(6.) The Clayton Act defines "price discrimination" to be the requisitely-contemporaneous charging of different prices for the same commodity to different buyers. In the Clayton Act's terminology, a seller's charging different prices to different buyers is still called "price discrimination" when the differences in prices reflect differences in the marginal (or incremental) costs the relevant seller must incur to supply the buyers in question. According to the (superior) standard economics definition, pricing that the Clayton Act would denominate "cost-justified price discrimination" is not price discrimination at all. On the standard economics definition, a seller would also be said to have practiced price discrimination if it charged different buyers the same price despite the fact that it had to incur different marginal (or incremental) costs to supply them. On the Clayton Act's definition, the latter seller's pricing would not involve "price discrimination."
(7.) 38 Stat. 717 (1914) (codified as amended at 15 U.S.C. Sections 41-45) (hereinafter Federal Trade Commission Act).
(8.) 15 U.S.C. [section] 45(n) (1994).
(9.) Treaty of Lisbon Amending the Treaty on European Union and the Treaty Establishing the European Community (hereinafter 2009 Treaty of Lisbon) at Article 101, 2007 OJ C306 p.1 (December 13, 2007) (hereinafter Article 101).
(10.) An actor's conduct is predatory if its perpetrator's ex ante belief in the conduct's ex ante profitability was critically affected by the perpetrator's belief that it would or might increase its profits by reducing the absolute attractiveness of the best offers against which it would have to compete by driving an extant rival out, inducing an extant rival to relocate its existing investments further away in product-space from the perpetrator's investments, and/or deterring an entry or an established-firm investment-expansion or inducing the maker of such an additional investment to locate it further away in product-space from the predator's investments when these increases in profits would render the relevant conduct profitable though economically inefficient (see Part 2A of this Article) in an otherwise-Pareto-perfect economy. An economy is said to be Pareto-perfect if it contains no condition that would generate any economic inefficiency in an otherwise-Pareto-perfect economy--viz., no imperfection in seller competition, no imperfection in buyer competition, no externality, no tax on the margin of income, no critical imperfection in the information available to a resource allocator (no resource-allocator nonsovereignty), no failure of a resource allocator to maximize, and no critical buyer surplus.
(11.) A seller initiates a contrived-oligopolistic-pricing sequence if it charges a price that is higher than the price it would otherwise find ex ante profitable to charge because it believes that its rivals will or may be deterred from making otherwise-profitable undercutting (or undermining) responses to its offer by their belief that it may or will react to such responses in one or more ways that would render such responses unprofitable despite the inherent unprofitability of its doing so--a belief that the firm has created by threatening to retaliate against their undercutting and or to reciprocate to their abstention from undercutting.
(12.) See Markovits, Economics of Antitrust Law I at 117-121.
(13.) 2009 Treaty of Lisbon at Article 102 (hereinafter Article 102).
(14.) Council Regulation 139/2004 on the Control of Concentrations Between Undertakings, OJ L241 (May 1,2004) (hereinafter EMCR). The original EMCR was promulgated as Council Regulation 4064/89 on the New Control of Concentrations Between Undertakings, OJ L395/1 (1989).
(15.) I need to explain why I believe that the specific-anticompetitive-intent operationalization of (1) the "object" of "preventing or restricting competition," (2) an "exclusionary abuse" of a dominant position, (3) "a restraint of trade," and (4) "monopolizing" or "attempting to monopolize" is correct as a matter of law. I think that a relatively-straightforward textual argument justifies interpreting the Article 101(1) prohibition of covered categories of conduct that "have as their object.. .the prevention ... [or] restriction of competition" to promulgate a specific-anticompetitive-intent test of illegality. For a discussion of why in this context "object" should be interpreted to refer to "critical object" as opposed to "an object," see Markovits, Economics of Antitrust Law I at 117-120. 1 also think that my reading of the object-branch of the Article 101(1) test of illegality is favored by the fact that the drafters of this provision were aware of the Sherman Act's test of illegality (even if they could not have articulated it in the way in which I have) and the associated likelihood that they intended the object-branch of their test to replicate the Sherman Act's test. (For the same reason, I believe that the effect-branch of Article 101(l)'s test of [prima facie] illegality should be read to replicate the Clayton Act's lessening-competition test of prima facie illegality.) In my judgment, the conclusion that an "exclusionary abuse" of a dominant position refers to conduct by a dominant firm that violates the specific-anticompetitive-intent test of illegality is justified by the facts that (1) U.S. courts implicitly defined the concept of "exclusionary" conduct in this way before the passage of the Treaty of Rome, (2) the drafters and ratifiers of what is now Article 102 were well aware of this usage, and (3) this operationalization is consistent with the pejorative connotation of "exclusionary abuse." For further explanation, see id. at 131 and 134-139.1 wish 1 could make a textual or contemporaneous-usage argument for interpreting "agreements in restraint of trade," "monopolization," and "attempts to monopolize" to refer to covered conduct that violates the specific-anticompetitive-intent test of illegality. However, although some U.S. common-law cases in the second half of the nineteenth century seem to have used these terms and their cognates in ways that are consistent with my specific-anticompetitive-intent operationalization of them, the facts and opinions of these cases are too obscure for these opinions to justify my operationalization, particularly given the nonconforming and inconsistent ways in which this language was used in earlier U.S. and English cases. See Robert H. Bork, The Rule of Reason and the Per Se Concept: Price Fixing and Market Division, 74 Yale L. J. 775 at 783-785; Donald Dewey, The Common-Law Background of Antitrust Policy, 41 Va. L. Rev. 759 (1955); and William L. Letwin, The English Common Law Concerning Monopolies, 21 Univ. Chi. L. Rev. 355(1954). However, I do think that my conclusion that the Sherman Act promulgates a specific-anticompetitive-intent test of illegality is warranted by three facts: (1) the legislators who drafted and passed the Sherman Act did so in response to what they perceived to be horizontal price-fixing, predatory conduct, and horizontal mergers and acquisitions that manifested their perpetrator's or perpetrators' specific anticompetitive intent (though, of course, the relevant legislators did not use this language and probably could not have articulated this objection to the conduct in question); (2) contemporaneous discussions at the time of the passage of the Sherman Act manifest the fact that the Act was perceived to condemn conduct that manifested its perpetrator's or perpetrators' bad character; and (3) the fact that the Sherman Act made violations of its terms a criminal offense manifests the fact that it prohibited conduct that was deemed to be immoral (conduct that was deemed to be immoral for the reason my operationalization of the Act's tests of illegality articulates). It is not by chance that contemporary lawyers continue to claim that the Sherman Act is about "character," "malicious intent," and "bad guys." See Timothy J. Brennan, Ahead of His Time: The Singular Contributions of Richard Markovits, Antitrust Bull. (2016) at the paragraphs that respectively immediately precede the paragraph that contains n.45 and the paragraph that contains n.45.
(16.) 2009 Treaty of Lisbon at Article 101(3) (hereinafter Article 101(3)).
(17.) For a detailed discussion of these issues, see Richard S. Markovits, Truth or Economics: On the Definition, Prediction, and Relevance of Economic Efficiency 29-31 (2008).
(18.) Id. at 22.
(19.) See Markovits, Economics of Antitrust Law 1 at 165-181.
(20.) Id. at 249-299.
(21.) See id. at 20-26.
(22.) See id. at 26-39.
(23.) See id. at 43-49.
(24.) For a complete analysis, see id. at 49-68.
(25.) See id. at 345-348.
(26.) Id. at 349-351.
(27.) Id. at 352-367.
(28.) Id. at 367-387.
(29.) Id. at 387-414.
(30.) Id at 434-438.
(31.) See Markovits, Economics and Antitrust Law II at 118-120 for a discussion of the position that the Department of Justice and Federal Trade Commission 1992 Horizontal Merger Guidelines (hereinafter 1992 Horizontal Merger Guidelines) take on this issue at Section 2.12, last paragraph, 4 Trade Reg. Rep. (CCH) Section 13,104 (1992) and Markovits, Economics of Antitrust Law II at 157 for a discussion of the position that the Department of Justice and Federal Trade Commission 2010 Horizontal Merger Guidelines (hereinafter 2010 U.S. Horizontal Merger Guidelines) take on this issue at Section 7.2 (http:// www.justice.gov/atr/publis/guidelines/hmg-2010.pdf) (August 19, 2010).
(32.) See Markovits, Economics and Antitrust Law II at 121-122 for a discussion of the U.S. 1992 Horizontal Merger Guidelines' discussion of this issue at Section 2.12 last paragraph of those Guidelines, and Markovits, Economics of Antitrust Law II at 156-157 for a discussion of the U.S. 2010 Horizontal Merger Guidelines' treatment of this issue at Section 2.1.5.
(33.) Markovits, Economics of Antitrust Law I at 503-517.
(34.) Id. at 517-519.
(35.) Id. at 519-530.
(36.) Id at 531-543.
(37.) Id at 543-562.
(38.) Id. at 562-582.
(39.) For a detailed discussion of these objections to limit-pricing theory and the evidence that its advocates claim supports it, see Markovits, Economics of Antitrust Law 11 at 221-232.
(40.) For the law-study's analysis of predatory QV investments, see Markovits, Economics of Antitrust Law I at 582-609.
(41.) See Janusz Ordover & Robert Willig, An Economic Definition of Predation: Pricing and Product Innovation, 91 Yale L. J. 8 (1981).
(42.) For the law-study's analysis of predatory cost-reducing investments, see Markovits, Economics of Antitrust Law I at 609-611.
(43.) See Markovits, Economics of Antitrust Law II at 40-50.
(44.) Id. at 3-4.
(45.) Id. at 45 and 161-162.
(46.) Id. at 3-4.
(47.) The parenthetical primarily reflects the fact that, by increasing the merged firm's OCAs and NOMs above those the merger partners would have enjoyed, horizontal mergers will make contrived oligopolistic pricing less profitable for the merged firm than it would have been for the merger partners by increasing the safe profits the merged firm must put at risk to attempt to contrive oligopolistic margins above the safe profits the merger partners would have had to put at risk to do so.
(48.) This conclusion primarily reflects the fact that the merged firm can use the pricing and advertising of both merger partners' products to attack a predation target and also reflects the fact that the merged firm will be able to profit more than either merger partner could from the tendency of its practice of predation to maintain or enhance its reputation for engaging in strategic behavior.
(49.) Id. at 5.1 originally published this point in Richard S. Markovits, Predicting the Competitive Impact of Horizontal Mergers in a Monopolistically Competitive World: A Non-Market-Oriented Proposal and Critique of the Market Definition-Market Share-Market Concentration Approach, 56 Tex. L. Rev. 587, 611 (1978). I repeated it in a Distinguished Guest Lecture I gave to the U.S. Antitrust Division in the Spring of 1979. The U.S. antitrust-enforcement authorities (the Department of Justice [DOJ] and Federal Trade Commission [FTC]) reached something like the same conclusion in Section 2.21 of their 1992 Horizontal Merger Guidelines, though their articulation of it is imperfect and their discussions of the sources of both the merger partners' advantages over each other when they were respectively best-placed and second-placed and the merger partners' advantages over the third-placed supplier of a buyer when they were uniquely equal-best-placed to supply the buyer in question are incomplete and confusing. See Markovits, Economics and Antitrust Law II at 111.
(50.) Id. at 2.
(51.) The law-study makes these points at id. at 35-37.
(52.) Id. at 7-12.
(53.) See id. at 45 and 161-162.
(54.) See Mandeville Island Farms v. American Crystal Sugar, 334 U.S. 219,236 (1948). Three more recent U.S. cases follow the Mandeville Island Farms holding in declaring illegal the player-salary-reducing arrangements of various professional- sports leagues. See Smith v. Pro Football, Inc. 593 F. 2d 173 (D.C. Cir. 1978); Philadelphia World Hockey Club, Inc. v. Philadelphia Club, 351 F. Supp. 462 (E.D. Pa. 1972); and Mackey v. NFL, 543 F. 2d 606 (8th Cir. 1976).
(55.) See Khan v. State Oil Co., 93 F. 3d 1358,1361 (7th Cir. 1996), vacated on other grounds, 522 U.S. 3 (1997). See also United States v. Rice Growers Association, 1986-2 Trade Cas. (CCH) [paragraph] 67, 288 (E.D. Cal. 1986).
(56.) See 2010 U.S. Horizontal Merger Guidelines at [section] 1 [paragraph] 8: "Enhancement of market power by buyers, sometimes called 'monopsony power,' has adverse effects comparable to enhancement of market power by sellers." See also 1992 U.S. Horizontal Merger Guidelines at [section] 0.1: "The exercise of market power by buyers ('monopsony power') has adverse effects comparable to those associated with the exercise of market power by sellers."
(57.) See Robert Blair and Jeffrey Harrison, Monopsony (1993).
(58.) See EC Guidelines on the Assessment of Horizontal Mergers at [paragraph] 90, OJ C3225 (2004).
(59.) See Markovits, Economics of Antitrust Law II at 4-21.
(60.) See Markovits, Economics of Antitrust Law II at 22-35 for a much fuller account.
(61.) See id. at 50-97.
(62.) See Markovits, Economics of Antitrust Law I at 190-210 and Markovits, Economics of Antitrust Law II at 94-144.
(63.) Id. at 165-175.
(64.) Id. at 149-163.
(65.) See U.S. 1992 Horizontal Merger Guidelines at [section] 5.1.
(66.) See U.S. 2010 Horizontal Merger Guidelines at [section] 4.1.1 [paragraph] 4.
(67.) See Markovits, Economics of Antitrust Law II at 187-188.
(68.) Id. at 219-220.
(69.) Id. at 220-221.
(70.) Id. at 201-218.
(71.) See id. at 251-531.
(72.) See id. at 377-378.
(73.) See id. at 264-274 for a more detailed analysis that includes a discussion of the determinants of the ratio of the amount of transaction surplus a seller will have to destroy to remove a given amount of buyer surplus from a particular buyer of its product by raising its per-unit price above its transaction-surplus-maximizing marginal cost.
(74.) See id. respectively at 275-276, 287-300, 344-345, and 348-349.
(75.) See id. at 284-287.
(76.) See id at 267-272 and 287-300.
(77.) See id at 300-305 and 320-325.
(78.) See id at 308-325.
(79.) See id. at 389-390.
(80.) See id. at 488.
(81.) Treaty Establishing the European Economic Community, 298 U.N.T.S 11 (March 25, 1957).
(82.) See Markovits, Economics of Antitrust Law II at 276-325 (tie-ins and reciprocity), at 325-339 (resale price maintenance), and at 339-350 (vertical territorial restraints and vertical customer-allocation clauses).
(83.) Independent distributors do not have a "liberty interest" in making the choices these vertical practices prohibit them from making in the sense in which "liberty" is an important value in the U.S. and E.U. (we have liberty interests properly-so-called in making choices that contribute to the chooser's discovering the conception of the good to which he or she subscribes, that actualize his or her particular conception of the good, or that fulfill his or her moral obligations); the independent distributors have accepted the relevant restrictions on their opportunity-set under conditions that preclude the conclusion that their wills were overborne; and laws that prohibit sellers from engaging in these vertical practices are not likely in any case to expand the opportunity-set of the individuals who--absent those laws--would be independent distributors: manufacturers will be likely to respond to such prohibitions by vertically integrating forward into distribution, by changing their pricing technique (to one that involves higher per-unit prices and lower lump-sum [franchise] fees) to reduce the losses that these practices' prohibition would otherwise impose on them, etc. See id. at 449-450. For a thorough analysis of the concept of liberty, see Richard S. Markovits, Matters of Principle: Legitimate Legal Argument and Constitutional Interpretation 278-285 (1998).
(84.) See Markovits, Economics of Antitrust Law II at 448-449.
(85.) See Leegin Creative Leather Products v. PSKS, Inc., 127 U.S. 2705 (2007) and State Oil v. Khan, 522 U.S. 3 (1997).
(86.) See Markovits, Economics of Antitrust Law II at 407-409.
(87.) For a detailed discussion, see id. at 437-476.
(88.) For a summary of many of the individual sets of points about foreclosure that the following text will make, see id. at 642-645.
(89.) See id. at 344-346.
(90.) See id. at 350 and Markovits, Economics of Antitrust Law I at 642.
(91.) See id. and Markovits, Economics of Antirust Law II at 320-325.
(92.) See id. at 481-482. See also id. at 264-274, 485, and 501.
(93.) See id. at 476-483 for a list of the functions that vertical mergers can perform that contains a large number of Sherman-Act-licit functions.
(94.) See id. at 644-645.
(95.) See Markovits, Economics of Antitrust Law I at 643-644. I want to add one somewhat-related point that will certainly be contested and may well be contestable. In cases in which it is in the joint interest of the relevant buyers and the relevant established firms for enough buyers to lock themselves into the established sellers to deter new entry, I would conclude that the arrangements that lock the buyers in do not violate the specific-anticompetitive-intent test of illegality because they would be economically efficient in an otherwise-Pareto-perfect economy.
(96.) See Standard Oil Co. v. United States (Standard Stations), 337 U.S. 293 (1949) and Markovits, Economics of Antitrust Law I at 677-678.
(97.) See Tampa Electric Co. v. National Coal Co., 365 U.S. 320 (1961) and Markovits, Economics of Antitrust Law I at 676-677.
(98.) See id. at 673-676. The Supreme Court created the "essential facilities" doctrine in United States v. Terminal Railroad Association of St. Louis, 224 U.S. 383 (1912). The Supreme Court stated that the essential facilities doctrine was crafted by the lower courts and that it had "never recognized" that doctrine in Verizon Communications, Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 441 (2004).
(99.) See Markovits, Economics of Antitrust Law II at 502-513.
(100.) See, e.g., Corwin D. Edwards, Vertical Integration and the Monopoly Problem, 17 J. Marketing 404 (1953).
(101.) Ford Motor Co. v United States, 405 U.S. 562 (1972).
(102.) 49 Fed. Reg. 26, S23 (1984).
(103.) See Markovits, Economics of Antitrust Law I at 693, citing Hachette, 8th Annual Competition Report 114-115 and IRI/ Neilson, 1996 Annual Competition Report 63.
(104.) See Markovits, Economics of Antitrust Law I at 693, citing Delimitis, C-234/89, ECR 1-935 (1991).
(105.) See Markovits, Economics of Antitrust Law II at 448, citing Communication From the Commission--Guidance on the Commission's Enforcement Priorities in Applying Article 82 [now-Article 102] of the Treaty to Abusive Exclusionary Conduct by Dominant Undertakings at [paragraph] 20 inset four, 2009/C 45/02 (2009).
(106.) See Markovits, Economics of Antitrust Law II at 521 -524, citing EC Guidelines on the Assessment of Non- Horizontal Mergers Under the Council Regulation on the Control of Concentrations Between Undertakings at points 25, 39,49, 64, 67, 73, and 76, 2008C 265/07 (2008).
(107.) See Markovits, Economics of Antitrust Law II at 476-477.
(108.) Id. at 476.
(109.) See Markovits, Economics of Antitrust Law I at 574, citing Pacific Bell Telephone Company, dba AT&T California, et. al. v. Linkline Communications, Inc. et. al., 129 S. Ct. 1109, 1119 (2009).
(110.) See id. at 582, citing Deutsche Telekom, Case COMP/C-1/37.451, 37.578, 37.579 OJ L263 (2003) and Industrie des Poudres Spheriques SA v. Commission, Case T-5/97, p. 178 (2000).
(111.) See Markovits, Economics of Antitrust Law II at 486-493.
(112.) See id. at 535-538.
(113.) See Markovits, Economics of Antitrust Law II at 537,542, and 551. See also Markovits, Economics of Antitrust Law I at 647-649.
(114.) See Markovits, Economics of Antitrust Law II at 539-546.
(115.) Id. at 563.
(116.) Id. at 582.
(117.) For analyses of these normative issues, see Richard S. Markovits, Matters of Principle: Legitimate Legal Argument and Constitutional Interpretation (1998).
(118.) See R.G. Lipsey & Kelvin Lancaster, The General Theory of Second Best, 24 Rev. Econ. Stud. 11 (1956) for the first formal statement of the theory.
(119.) See Oliver Williamson, Economies us an Antitrust Defense Revisited, 125 U. PA. L. Rev. 699 (1997) and Economies as an Antitrust Defense: The Welfare Trade-Off 58 Am. Econ. Rev. 18 (1968). 1 should note that the trade-off Williamson is referencing is actually an economic-efficiency trade-off, not a "welfare" trade-off. Even if, ad arguendo, one equates welfare with utility, choices that increase economic efficiency will not increase welfare (i.e., utility) if the utility-value of the average dollar won by their beneficiaries is sufficiently below the utility-value of the average dollar lost by their victims. This conflation of "welfare" with "economic efficiency" is also manifest in the name of the field of economics that is concerned with economic efficiency--Welfare Economics.
(120.) For simplicity, the text ignores the fact that price-reductions can affect the marginal allocative value of intra-marginal units of a good by having distributive effects that alter the dollar value of intra-marginal units to their consumers.
Richard S. Markovits *
* The University of Texas at Austin, Austin, TX, USA
Corresponding Author: Richard S. Markovits, The University of Texas at Austin, Austin, TX 78705, USA. Email: email@example.com
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||D. Conglomerate Mergers through Conclusion, with footnotes, p. 52-83; Symposium: Economics and the Interpretation and Application of U.S. and E.U. Antitrust Law|
|Author:||Markovits, Richard S.|
|Date:||Mar 1, 2016|
|Previous Article:||A summary.|
|Next Article:||The tests of illegality under articles 101 and 102 TFEU.|