LEXIS®-NEXIS® Academic Universe

LEXIS(R)-NEXIS(R)

Document 5 of 47.

APRIL, 1985

85 Colum. L. Rev. 515

LENGTH: 27746 words

ARTICLE: EXTENSION OF MONOPOLY POWER THROUGH LEVERAGE.

Louis Kaplow *

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

* Assistant Professor of Law, Harvard University. Northwestern University, A.B., 1977; Harvard University, J.D., 1981; A.M., 1981. The helpful comments and assistance of Lucian Bebchuk, Erwin Chemerinsky, Brad Karp, and Steven Shavell are greatly appreciated.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

SUMMARY:
... The debate over the ability of firms to use restrictive practices to leverage their monopoly power from one market to another has continued throughout the history of the antitrust laws. ... For example, one common argument is that many such practices do not extend monopoly, but merely serve to permit profit maximization through price discrimination. ... Most recently, Easterbrook has advocated a new, more limited role for antitrust regulation that is based heavily on a static approach. ... For example, with regard to the well-known I.B.M. punchcard tying case, Posner argued that "[i]f the purpose of imposing the tie-in really was to protect good will rather than to discriminate, the specifications alternative was presumably less efficient than the tie-in; otherwise the manufacturer would have promulgated specifications voluntarily. ... Second, to the extent one has trouble determining which of many possible motivations can explain and which effects will result from the observed behavior, it would seem incongruous to permit a restrictive practice under the antitrust laws because, rather than inevitably resulting in anticompetitive effects, it sometimes results in other undesirable effects instead. ...

TEXT:
[*515] INTRODUCTION

The debate over the ability of firms to use restrictive practices to leverage their monopoly power from one market to another has continued throughout the history of the antitrust laws. The most common application of the leverage hypothesis involves tying arrangements. For example, a firm with monopoly power over one product is observed selling it to customers only on the condition that they purchase another of the firm's products as well. Courts and many commentators have long feared that such a tying arrangement will facilitate the firm's monopolization of the market for the second product. The leverage hypothesis underlies a substantial portion of the antitrust attack on many other restrictive practices, ranging from vertical mergers and reciprocal dealing arrangements to many tactics scrutinized under Section 2 of the Sherman Act.n1

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n1 15 U.S.C. § 2 (1982). Vertical mergers are those involving firms at different levels in the production chain -- e.g., between manufacturers and suppliers, or manufacturers and retailers. Reciprocal dealing involves an agreement to purchase from another firm on the condition that the other firm will make purchases from the initial firm -- e.g, a steel company agrees to buy its trucks from a truck manufacturer only if the truck manufacturer buys its steel from the steel company.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

In the past three decades, a growing body of commentators -- originating in the "Chicago School"n2 -- has sharply criticized this traditional position by arguing that such leverage is impossible. Restrictive practices seen by courts as devices to extend monopoly are instead argued to be benign or even efficient techniques by which firms maximize their profits given the market conditions they face. This particular critique constitutes the foundation for one of the primary fronts of attack on the antitrust laws today. The development of the traditional leverage approach and the details of the antileverage position are briefly described in Part I.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n2 See, e.g., Posner, The Chicago School of Antitrust Analysis, 127 U. Pa. L. Rev. 925 (1979) [hereinafter Posner, Chicago School]. But see Nelson, Comments on a Paper by Posner, 127 U. Pa. L. Rev. 949 (1979) (criticizing Chicago School positions); Scherer, Book Review, The Posnerian Harvest: Separating Wheat from Chaff, 86 Yale L.J. 974 (1977) (same).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

There are a number of deficiencies in the analysis of recent commentators who have attempted to proclaim the death of leverage theory. The basic mistake in their central thesis is that antitrust law should [*516] be indifferent to the exploitation of monopoly power because extant power is a fixed sum and thus will result in the same damage regardless of how it is deployed. Although of some superficial appeal, it can readily be demonstrated that their analysis is strongly counterintuitive. Consider the case of a terrorist on the loose with one stick of dynamite. The fixed sum thesis posits that since the power is fixed -- that is, the terrorist has one and only one dynamite stick -- we should be indifferent to where the dynamite is placed. It is all too obvious, however, that the potential damage resulting from power in this context, as well as in virtually any other we can imagine, is overwhelmingly dependent upon how it may be used. Part II of this paper develops the general theoretical invalidity of the fixed sum thesis in the leverage context and presents factors indicating that leverage, even when understood as extension of monopoly power, is possible in practice.

Part II establishes only that leverage must be taken seriously; other explanations offered by the critics of leverage theory to explain restrictive practices are not ruled out.n3 Part III criticizes the most frequent of these explanations. In addition, it develops alternative explanations for the use of these practices. These explanations, which have received too little attention from courts and commentators, derive from a firm's misperception of the effects of its actions or a firm's decision to pursue objectives other than profit maximization. Part III concludes with some warnings concerning how readily one can be misled by the methods used to compare competing hypotheses in this area.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n3 Section II-A, however, does indicate that some of the other explanations, even if valid, would not necessarily exonerate the defendant's behavior, as many of the critics claim.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

I. THE LEVERAGE DEBATE

The debate over the leverage issue has followed a rather simple pattern. Almost since the passage of the Sherman Act,n4 courts have advanced the often superficial assertion that many trade practices are attempts to extend monopoly power. This position generally continues to prevail. This Part briefly sets forth this position, and then explores what has become the dominant response of a growing number of commentators during the past three decades. This response asserts that it is simply impossible to extend monopoly power using such practices, and that these practices can better be understood as attempts to maximize profits than as attempts to extend existing monopoly.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n4 26 Stat. 209 (1890) (codified as amended at 15 U.S.C. §§ 1-7 (1982)).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Traditional leverage theory claims that a monopolist's use of its power in its own market to control activities in another market typically represents an attempt to spread its power to the other market. This argument was advanced in the dissent in Henry v. A.B. Dick Co.,n5 and was [*517] adopted by the Supreme Court in the Motion Picture Patents decision.n6 It has been applied frequently in the tying context,n7 and in other vertical arrangements cases,n8 as well as in the monopolization context under Section 2 of the Sherman Act.n9 The Court and various commentators have advanced other reasons to support their hostility to such practices,n10 a detailed discussion of which is beyond the scope of this Article.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n5 224 U.S. 1, 53 (1912) (White, C.J., dissenting) (patent misuse case, stating that tying allows patentees "to multiply monopolies").

n6 Motion Picture Patents Co. v. Universal Film Co., 243 U.S. 502, 518 (1917). Motion Picture Patents was not technically an antitrust case. The Court, in affirming a decision that held the patentee's tying clause to be invalid, found support for its views in Congress' recent enactment of Clayton Act § 3 (codified at 15 U.S.C. § 14 (1982)), but did not decide whether there was an antitrust violation. Id. at 517-18.

n7 See, e.g., Fortner Enterprises, Inc. v. United States Steel Corp., 394 U.S. 495, 509 (1969); Times-Picayune Publishing Co. v. United States, 345 U.S. 594, 611 (1953) ("[T]he essence of illegality in tying arrangements is the wielding of monopolistic leverage; a seller exploits his dominant position in one market to expand his empire into the next."); Carbice Corp. of Am. v. American Patents Dev. Corp., 283 U.S. 27, 32 (1931); United Shoe Mach. Corp. v. United States, 258 U.S. 451, 457-58 (1922); R. Posner, Antitrust Law: An Economic Perspective 171-72 (1976) (noting the traditional argument); L. Sullivan, Antitrust 431 (1977) ("The consistent judicial instinct has been that these arrangements have but a single purpose and effect, to extend the seller's power in the market for the tying product into that for the tied product."); Ferguson, Tying Arrangements and Reciprocity: An Economic Analysis, 30 Law & Contemp. Probs. 552, 561-62 (1965); see also R. Posner, supra, at 183 (indicating that the framers of § 3 of the Clayton Act believed in the leverage theory).

n8 See, e.g., Brown Shoe Co. v. United States, 370 U.S. 294, 332 (1962) (vertical merger); Standard Oil Co. v. United States, 337 U.S. 293, 305-06, 314 (1949) (exclusive dealing).

n9 See, e.g., United States v. Griffith, 334 U.S. 100 (1948).

n10 See, e.g., Blake & Jones, Toward a Three-Dimensional Antitrust Policy, 65 Colum. L. Rev. 422, 433-35 (1965) (protection of individual freedom and opportunity). Commentators who reject the leverage hypothesis and argue that various practices should be legal are necessarily rejecting such alternative rationales for existing prohibitions.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Although the critique of leverage theory has been advanced for some time, more recently is has assumed far greater prominence and attracted a wider range of advocates and adherents. Virtually all such criticism has focused on two basic points, which will be described here and explored in detail in Part II.

A. The Fixed Sum Argument

The critique first focuses on the fixed sum argument, which begins with the view that, if the existing degree of monopoly power was legally obtained,n11 the monopoly profit deriving therefrom is also legal. Restrictive practices are then defended by the fixed sum claim, which is simply that a firm with market power may be able to gain its profit all from its own market, all from another, or from any combination [*518] thereof, but the total amount of restriction that the monopolist will profitably be able to impose is fixed regardless of the practice that is used. An alternative statement of the same argument is that for a given amount of power, indirect expolitation through restrictive practices can cause no more damage than direct exploitation. Proponents of this position usually trace their arguments to the teachings of Aaron Director.n12 The analysis was well explained by Posner:

A [fatal] weakness of the leverage theory is its inability to explain why a firm with a monopoly of one product would want to monopolize complementary products as well. It may seem obvious that two monopolies are better than one, but since the products are by hypothesis used in conjunction with one another . . ., it is not obvious at all. If the price of the tied product is higher than the purchaser would have had to pay in the open market, the difference will represent an increase in the price of the final product or service to him, and he will demand less of it, and will therefore buy less of the tying product. To illustrate, let a purchaser of data processing be willing to pay up to $1 per unit of computation, requiring the use of one second of machine time and 10 punch cards, each of which costs 1¢ to produce. The computer monopolist can rent the computer for 90¢ a second and allow the user to buy cards on the open market for 1¢, or, if tying is permitted, he can require the user to buy cards from him at 10¢ a card -- but in that case he must reduce his machine rental charge to nothing, so what has he gained?

As this example suggests, in the absence of price discrimination a monopolist will obtain no additional profits from monopolizing a complementary product.n13

Bork refers to the argument thus criticized as the "fallacy of double counting" and applies this analysis in a number of contexts.n14 In particular, [*519] he argues that "[t]he law's theory of tying arrangements is merely another example of the discredited transfer-of-power theory, and perhaps no other variety of that theory has been so thoroughly and repeatedly demolished in the legal and economic literature."n15 It is this criticism that I refer to as the "fixed sum" view of monopolistic practices.n16

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n11 If the monopoly power itself is illegal, it is argued, the monopoly power should be the direct subject of any antitrust attack. This argument is discussed further in Section I-B, and criticized in Section II-A-1.

n12 Many of the proponents were Director's students at the University of Chicago. The only time Director published this line of argument was in Director & Levi, Law and the Future: Trade Regulation, 51 Nw. U.L. Rev. 281, 290, 292 (1956). Bowman's application of this argument in the tying context, Bowman, Tying Arrangements and the Leverage Problem, 67 Yale L.J. 19, 21 (1957), has much to do with its popularization among commentators. Much of the argument, however, seems to have been understood decades before, as indicated by the Supreme Court's opinion in the first United Shoe case, United States v. United Shoe Mach. Co., 247 U.S. 32, 65 (1918).

n13 R. Posner, supra note 7, at 173 (emphasis in original); cf. id. at 197 (applying the argument to vertical integration); 5 P. Areeda & D. Turner, Antitrust Law 165-66 (1980) (reciprocity); Posner, Chicago School, supra note 2, at 926 (tie-ins).

n14 See R. Bork, The Antitrust Paradox 140 (1978) (licensing practices in United States v. United Shoe Mach. Corp., 110 F. Supp. 295 (D. Mass. 1953), aff'd per curiam, 347 U.S. 521 (1954)); id. at 141 ("all the vertical merger cases under amended Section 7 of the Clayton Act," and in United States v. Griffith, 334 U.S. 100 (1948)); id. at 213 (vertical merger in Brown Shoe Co. v. United States, 370 U.S. 294 (1962)); id. at 258 (reciprocity in Federal Trade Comm'n v. Consolidated Foods Corp., 380 U.S. 592 (1965)); id. at 290 (vertical restrictions in United States v. Arnold, Schwinn & Co., 388 U.S. 365 (1967)); id. at 306 (exclusive dealing in Standard Fashion Co. v. Magrane-Houston Co., 258 U.S. 346 (1922)); id. at 367 (monopoly extension deemed "impossible" in tying context); id. at 373-74 (tying cases).

n15 Id. at 372.

n16 Others have followed this view. See, e.g., Note, An Economic Analysis of Royalty Terms in Patent Licenses, 67 Minn. L. Rev. 1198, 1218-19 (1983) (citing W. Bowman, Patent and Antitrust Law (1973); Note, An Analysis of Tying Arrangements: Invalidating the Leveraging Hypothesis, 61 Tex. L. Rev. 898, 909-11 (1983) [hereinafter cited as Texas Note].

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

B. The Profit Maximization/Monopoly Extension Distinction

This summary of the critics' argument is not sufficiently precise. It is well known that a monopolist will not garner the same reward regardless of its actions. For example, if it were to set its price at the competitive level (at marginal cost), it would receive no excess profits and cause no resource misallocation at all. Thus, the critics from the beginning have distinguished between practices that are merely profit maximizing for a given quantum of monopoly power and those that extend monopoly power into other markets.n17 The former are deemed legitimate on the ground that any challenge to such practices implicitly challenges the legitimacy of extant monopoly power, and thus such an attack should be directed at the monopoly itself.n18 if the monopoly is legal, the profit-maximizing practices should be deemed permissible; if illegal, remedial powers should be brought to bear directly on the monopoly. [*520] The latter -- monoploy extension -- which the critics associate with the traditional leverage argument, is conceded to be illegitimate but thought to be impossible because of the fixed sum argument.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n17 See, e.g., W. Bowman, supra note 16, at 54 ("the most pertinent question involves whether patentee-licensee agreements are profit-maximizing or are monopoly-extending"); id. at 240 ("No careful distinction between monopoly maximization and monopoly extension has been articulated and consistently applied [by the Supreme Court]."); R. Posner, supra note 7, at 29 (specially defining the phrase "unilateral noncoercive monopolization" to refer to monopolistic practices that extend monopoly power); id. at 171 (applying the distinction); Baldwin & McFarland, Tying Arrangements in Law and Economics, 8 Antitrust Bull. 743, 768 (1963); id. at 771 n. 54b (applying the distinction); Bauer, A. Simplified Approach to Tying Arrangements: A Legal and Economic Analysis, 33 Vand. L. Rev. 283, 292, 298, 299 n.52 (1980); Bowman, supra note 12, at 19 ("A distinction can usefully be made between leverage as a revenue-maximizing device and leverage as a monopoly-creating device. The first involves the use of existing power. The second requires the addition of new power." (footnote omitted)); id. at 23 (applying the distinction); Burstein, A Theory of Full-Line Forcing, 55 Nw. U.L. Rev. 62, 62 (1960); Comanor, Vertical Mergers, Marker Power, and the Antitrust Laws, 57 Am. Econ. Rev. (Papers & Proceedings) 254, 254 (May 1967) (distinguishing between "market power" and "market position"); Director & Levi, supra note 12, at 290 (price discrimination "might be considered more an enjoyment of the original power than an extension of it"); id. at 295; Markovits, Tie-Ins and Reciprocity: A Functional, Legal, and Policy Analysis, 58 Tex. L. Rev. 1363, 1369 (1980).

n18 See, e.g., Bowman, supra note 12, at 32.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

The analysis underlying this position has not been refuted directly. Opponents generally repeat the more superficial argument that was generated by the courts or develop exceptions that implicitly concede much of the ground to those advancing benign characterizations of restrictive practices. The next Part of this Article seeks to broaden substantially the scope of the controversy as well as to provide a general framework from which to understand and evaluate those refutations offered in the past.

II. THE POSSIBILITY OF LEVERAGE

The response that has been offered to rebut the simplistic leverage theory put forth by the courts has two basic deficiencies. Both derive from the principle illustrated in the introduction by the discussion of a terrorist's placement of a stick of dynamite. First, as Section A explores, the distinction between extension and maximal exploitation of monopoly is problematic because it is somewhat arbitrary and does not have the obvious implications for antitrust policy that are often advanced.n19 Second, the fixed sum argument, although an improvement in some respects upon the courts' analysis, is itself overly simplistic. Section B outlines a common set of problems tht arise in the application of the fixed sum hypothesis. The implication is that leverage, even as defined by proponents of the fixed sum argument, is more plausible in some setting than they are willing to admit.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n19 Given the former point, the latter should not be surprising.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

A. Failure of the Profit Maximization/Monopoly Extension Distinction

This Section advances two criticisms of the distinction between profit maximization and monopoly extension upon which the leverage critique rests. First, it reveals the fallacies underlying the assertion that only monopoly extension offers grounds for policy concern. Second, it argues that the distinction is meaningful only as a heuristic device to clarify the time frame of analysis -- short run versus long run -- and is thus without a priori normative significance. This interpretation also highlights the potential for error in using inappropriate models for analyzing restrictive practices.

1. The Distinction Has Limited Relevance. -- The critics concede that a monopolist's profits, and thus the welfare cost to consumers, will depend upon the tactics it employs. In the simple example of price setting, the monopolist cares greatly whether it must price at cost or is permitted to elevate its price to the profit-maximizing level. It is entirely conceivable that an appropriate competition policy would constrain [*521] the freedom to set prices. In fact, a wide variety of direct regulations at both the state and federal level do just that.

Those regulations, however, are not generally part of the existing antitrust regime,n20 but the point is that this result is not inevitable. When deciding whether a wide variety of restrictive practices that are either explicitly addressed by the antitrust statutesn21 or arguably within their general scopen22 should be deemed illegal, it is essential to consider the overall costs and benefits of permitting such practices.n23 While it may often be true that practices enabling a firm to maximize its return in a given market will be desirable from this perspective, there is no reason to believe that this would generally be the case.n24

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n20 The reasons typically offered to explain why this is the case include the courts' incapacity to regulate prices directly, see, e.g., United States v. Trenton Potteries Co., 273 U.S. 392, 397-98 (1927), and the view that legitimate monopolies deserve the reward to encourage efficient behavior, see, e.g., Kaplow, The Patent-Antitrust Intersection: A Reappraisal, 97 Harv. L. Rev. 1813 (1984).

It should also be noted that numerous antitrust decrees, such as that involving ASCAP and BMI, and the motion picture industry, do involve regulation of this sort, either directly or indirectly. In addition, various tests for liability, including those for predatory pricing, price discrimination, and price squeezes, involve scrutiny of the prices charged. See also infra note 149 (further exploring this inconsistency).

n21 See e.g., 15 U.S.C. § 14 (1982) (Clayton Act § 3, concerning tying and related practices).

n22 For example, virtually any practice used by a firm to recover monopoly profits can be argued to fall within the prohibition of § 2 of the Sherman Act, 15 U.S.C. § 2 (1982).

n23 Of course, issues concerning legislative intent as well as those concerning less narrowly defined economic ramifications, see, e.g., supra note 10, might also be relevant, or even decisive. Only the direct, narrowly defined, economic effects are discussed here.

n24 Slawson has properly criticized Bowman, and implicitly a number of others as well, for begging the question on this issue. See Slawson, A Stronger, Simpler Tie-In Doctrine, 25 Antitrust Bull. 671, 688-89 (1980). He indicates that all uses of monopoly are within the general concern of the antitrust laws, although he does not get the mileage out of his point that he could. For example, when discussing tying arrangements shortly after making his criticism, he responds to the price discrimination motivation by stating that "of course this cannot possibly be a use's only effect, even if it is the seller's only purpose . . . . The other effects should at least be examined in order to determine whether they are anticompetitive or harmful in other respects. . . ." Id. at 690 (all but first emphasis added). His earlier point, as explained in the text to follow, proves that one must do more than examine the other effects. The price discrimination effect itself must be scrutinized. Turner's analysis of tying arrangements also takes issue with the position of Bowman and others of similar views, but does not attack directly the claim that profit maximization in itself is unobjectionable. See Turner, The validity of Tying Arrangements Under the Antitrust Laws, 72 Harv. L. Rev. 50, 63 n.42 (1958) (noting that one must account for the effects on competing sellers).

Burstein, who is responsible for much of the more rigorous development of the analysis concerning the use of restrictive practices as profit-maximizing devices, see Burstein, The Economics of Tie-In Sales, 42 Rev. Econ. & Stat. 68 (1960); Burstein concludes that existing prohibitions are desirable despite the fact that in large part the prohibited practices enhance the gains from monopoly power rather than extending monopoly power, see Burstein, supra note 17, at 93, although he does not develop why he thinks it appropriate to take that position.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

The common reply is that this sort of criticism is better directed at [*522] the existence of the monopoly than at the restrictive practice. This superficially appealing argument is remarkably deficient. First, in many instances no simple remedy would restructure the market effectively, so prohibition of the profit-maximizing practice may be a good second best. Second, the cost of remedies where they are available may in fact be greater than the cost of permitting the monopoly to exist while controlling its behavior. Finally, it may well be desirable to permit some monopolies to come into and remain in existence, yet provide them with smaller rewards than those they might realize if permitted unlimited exploitation.n25

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n25 One area that obviously presents numerous instances subject to precisely this sort of analysis is the exploitation of a patent monopoly, explored in Kaplow, supra not 20.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

This perspective casts in a somewhat different light numerous defenses of various restrictive practices.n26 For example, one common argument is that many such practices do not extend monopoly, but merely serve to permit profit maximization through price discrimination.n27 This would be a benign interpretation only if it could be assumed that price discrimination were not undesirable. Of course, direct attempts at price discrimination are illegal under the Robinson-Patman Act.n28 Since this prohibition is the most sharply criticized of all the antitrust provisions,n29 it may not be surprising that the commentary [*523] concerning the use of various practices to facilitate price discrimination has proceeded along these lines. Yet the resulting position of such commentators is somewhat ironic. On the one hand, they criticize those who challenge restrictive practices on the ground that the criticism ultimately questions the legitimacy of the underlying monopoly when that monopoly itself has not been deemed illegal. On the other hand, they defend those same practices because they serve a supposedly benign purpose which has been deemed illegal. More careful commentary has on occasion both recognized this connection and noted that price discrimination can be more costly than is often assumed to be the case.n30

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n26 The text focuses on price discrimination, the most commonly noted of the supposedly benign explanations for restrictive practices of this kind. Other explanations are often offered in the same casual manner. For example, Bowman argues that tying may be used as a technique of evading price regulation, see Bowman, supra note 12, at 21-23 which in fact may have been the most plausible explanation for the tying arrangement in Northern Pacific Ry. v. United States, 356 U.S. 1 (1958). His observation that "no revenue can be derived from setting a higher price for the tied product which could not have been made by setting the optimum [i.e., profit-maximizing] price for the tying product," id. at 23, is inapposite since the purpose of the price regulation was to prevent just that. The best that can be said in defense of his position is that since evasion of the price regulation is the result, the regulatory commission, rather than the courts hearing antitrust controversies, should have jurisdiction. That, of course, is an alternative, but it is not obviously the best one and it would require substantial further analysis to establish that it was. Moreover, if a court examining a tying practice must decide which of many purposes motivates the restriction, see Part III, infra, it may be sensible to reject Bowman's conclusion if the most benign explanation the defendant can offer is that it is attempting to evade price regulation. No doubt, if asked by the regulatory agency, the defendant would deny this objective and claim to be extending its monopoly power in the unregulated market, which is beyond the agency's jurisdiction.

n27 See, e.g., R. Posner, supra note 7, at 173, 174-75; Bowman, supra note 12, at 19 n.4. Very similar arguments are made concerning the use of such practices to earn greater profits by exploiting the complementarity in demand between two or more products. See, e.g., id. at 25-27 & n.21.

n28 15 U.S.C. § 13 (1982) (§ 2 of the Clayton Act, amended by the Robinson-Patman Act of 1936, ch. 592, 49 Stat. 1526).

n29 See, e.g., R. Bork, supra note 14, at 382-401 ("antitrust's least glorious hour"); F. Schere, Industrial Market Structure and Economic Performance 580-82 (2d ed. 1980) ("an extremely imperfect instrument, [i]t is questionable whether the circle of beneficiaries extends much wider than the attorneys who earn sizable fees interpreting its complex provisions").

n30 See R. Posner, supra note 7, at 11-13, 177-79; Markovits, supra note 17, at 1445. It is not my purpose here to evaluate Posner's ultimate conclusion that tentatively would permit most restrictive practices, despite the difficulties related to price discrimination, R. Posner, supra note 7, at 178-80, but merely to highlight my argument that labelling a practice as profit maximizing rather than exclusionary leaves one far short of understanding what is necessary to decide whether the practice should be prohibited.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

2. The Distinction is Arbitrary. -- The position that various restrictive practices should be permissible because they are merely devices that maximize profit also suffers because a relevant conceptual distinction between profit maximization and monopoly extension cannot be maintained. Initially, practices in both categories are motivated by the firm's desire to increase its profits.n31 It has already been noted that practices merely increasing profits to an existing monopoly, without "extending" it, can increase the welfare loss that results. From the perspective of conventional economic analysis, practices that have similar effects on social welfare are viewed similarly, so any attempt to draw a relevant economic distinction between profit-maximizing techniques and monopoly-extending techniques must rely on grounds other than an evaluation of the net welfare effects in each instance. But why would one want to make such a distinction?n32

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n31 But see Section III-B, infra.

n32 In fact, precisely those commentators who emphasize the profit maximization/monopoly extension distinction advocate limiting antitrust concerns to this narrowly defined economic conception of social welfare. See, e.g., R. Bork, supra note 14, at 50-89, 110-15; R. Posner, supra note 7, at 18-22. Commentators who are more associated with the leverage hypothesis tend to argue that the goals of antitrust should not be so limited. See, e.g., Sullivan, antitrust, Microeconomics, and Politics: Reflections on some Recent Relationships, 68 Calif. L. Rev. 1 (1980).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

The position of these commentators can be given more meaning only by developing it more extensively. The most reasonable interpretation, I believe, is that their two categories represent an implicit attempt to distinguish between short-run and long-run phenomena -- or, to use more technical language, between static and dynamic models. [*524] Profit-maximizing practices are meant to refer roughly to those actions that can have fairly direct and immediate effects, while monopoly extension refers to behavior designed to have implications on the magnitude of profits and welfare loss in the future.n33 The prototypical example of a profit maximization device is a pricing decision by a firm with market power, a decision which can be implemented rather quickly.n34 By contrast, practices designed to affect the market share and elasticity of market demand might be labelled monopoly extension devices. These practices do not increase short-run profits, and might even decrease them. The firm's motivation is to chage the structural conditions it faces in the future in order that it may receive greater profits in the future. This perspective is not static, in the sense that it does not take the existing parameters as given; rather it is dynamic, in that it focuses upon how the parameters change over time.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n33 Typical measures of market power ask how far prices can be raised above cost under current market conditions. See, e.g., Landes & Posner, Market Power in Antitrust Cases, 94 Harv. L. Rev. 937, 939, 944-47, 959 (1981). Relevant factors include information such as the short-run elasticity of market demand and the existing market share of the firm. See id. at 944-47.

n34. Under this framework, schemes designed to exploit interdependency in demand among goods are placed in the profit maximization category. Bowman originally considered this as the one true possibility for leverage. See Bowman, supra note 12, at 25-27. More recently, Bowman has been sharply criticized for this categorization on the ground that practices with such effects are analytically similar to price discrimination, which is not monopoly-extending. See Texas Note, supra note 16, at 922-27. My attempt to understand the implicit rationale behind the profit maximization/monopoly extension disinction is consistent with this latter position. The very existence of this dispute over categorization is in the end largely semantic, reinforcing my overall conclusion that this long-used distinction is both indeterminate and irrelevant.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Earlier it was shown that this distinction has no a priori implications for the desirability of practices that affect profits in the short run, as opposed to those having implications in the long run. Additionally, it should now be apparent that this static versus dynamic distinction, although an aid to understanding,n35 is not one that even in theory could be helpful in the manner intended by those who rely on it so heavily. Static analysis has never been understood by economic theory as a privileged arena for behavior by firms. Such analysis, rather, is merely a simplification to help analyze complex behavior.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n35 I must concede at least this much as I use the distinction to help explain the issue I discuss in subsection II-B-2, infra.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Reconsidering the price setting issue helps illustrate this point. The prices charged by a dominant firm may influence entry into the industry, and thus long-rue market share and market power.n36 Even short run pricing behavior might have long-run effect. A rule that permitted [*525] only practices that had effects limited to the short run probably would permit nothing. A modification that prohibited only practices with adverse long-run effects, regardless of the short-run implications, would be rather incongruous. First, since the aggregate welfare effect of any practice is the discountedn37 sum of the effects in the present and future, one again wonders why the present should be ignored. Second, since no effects, even those from changes in pricing policies, are instantaneous, and since most long-run effects begin to be felt -- however slightly -- almost immediately, the idea of separating the two is most unpromising. Finally, such a rule may well be backwards in some cases. For example, a practice might produce tremendously beneficial shortrun effects and moderately adverse long-run effects. Presumably one would not ignore the former, but instead balance the two. The same reasoning is appropriate when the short- and long-run effects are reversed.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n36 See, e.g., Gaskins, Dynamic Limit Pricing: Optimal Pricing Under Threat of Entry, 3 J. Econ. Theory 306 (1971); Milgrom & Roberts, Limit Pricing and Entry Under Incomplete Information: An equilibrium Analysis, 50 Economertrica 443 (1982); M. Fortunato, The Welfare Effects of Entry Barriers and Antitrust Policy in a Dynamic Limit Pricing Model (Unpublished doctoral dissertation, Harvard University, 1983).

n37 Future effects are given less weight, that weight being a function of what is deemed to be the appropriate social discount rate.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Just as it matters where the terrorist leaves the stick of dynamite, it also matters how a firm chooses to exercise the power that it has. The fixed sum hypothesis is thus not an inherent physical law, on a par with the conservation of mass plus energy, but a position that when properly examined is at base counterintuitive.n38 It does not follow, of course, that the earliest statements of the leverage hypothesis -- which never was supported by in-depth analysis -- are correct, just because central portions of the criticism are off the mark. The leverage issue therefore presents a much more open question than is generally believed. A complete resolution is vastly beyond the scope of this Article. The remarks that follow are confined to analyzing the plausibility that leverage can be used in a harmful manner similar to that originally postulated by the courts and challenged by their critics.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n38 Even Bork does not appear to be wholly consistent in advancing his position that the fixed sum hypothesis is unalterably valid. For example, it is hard to understand how he can agree that the Court was correct in forbidding predation in Lorain Journal Co. v. United States, 342 U.S. 143 (1951), see R. Bork, supra note 14, at 344-45, without conceding that the fixed sum hypothesis is flawed in principle. He distinguishes United States v. Griffith, 334 U.S. 100 (1948) -- where he feels "the fallacy of double counting" led the Court astray -- primarily on the grounds that Lorain Journal had a high market share, predatory intent, and no apparent efficiency justification. R. Bork, supra note 14, at 345. Yet these arguments would have been irrelevant if the structure of the argument itself were logically fallacious, as most of Bork's discussion suggests.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

B. How Past Analysis Often Obscures the Potential for Leverage

My purpose in arguing that leverage is possible in some instances in spite of the criticism offered by those advancing the fixed sum hypothesis is two-fold. First, I will systematize the arguments that may be combined in a variety of contexts to explore the potential for leverage. [*526] Second, in the process, I will consider some factors that typically are overlooked by both critics and supporters of the courts' edicts.

At least four components should be part of the typical leverage inquiry.n39 First, it is important to understand the potential cost to the firm itself of employing the restrictive practice. Practices that come cheap are more likely to be used in attempts to secure effects in the long run even when those effects may be somewhat speculative. Second, as explored previously, it is necessary to focus explicitly on the relationship between the short and long run, noting that the firm will be willing to incur positive net costs from a static perspective in order to achieve greater overall profits after taking into account long-run effects. The other two components concern the strength of self-corrective market forces. The third point is that free rider problems may often inhibit the ability of competitors, purchasers, or suppliers to counter the designs of a firm employing restrictive practices. Fourth, market imperfections may create additional opportunities for restrictive practices to modify market structure. All four points are commonly overlooked or underestimated by the leading critics of the courts' proscriptions.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n39 This list does not attempt to be exhaustive; rather, it emphasizes what is most often overlooked. I would hope that the potential for further development in this area will be enhanced by clarifying the elements as much as possible.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

1. Failure to Note Minimal Cost of Monopolistic Practices. -- The point here is extremely simple: to the extent restrictive practices cost the firm using them very little, they are much more likely to be used. The question of how much restrictions might cost the firm has frequently been overlooked. Director and Levi note when discussing United States v. Griffithn40 that "it would seem that in order to impose additional coercive restrictions on the suppliers, as, for example, on the suppliers for competitive markets, the monopoly owner would have to pay the suppliers for these additional restrictions."n41 But how much would the monopoly owner have to pay? Bowman notes that "the demand for the tying product at any price will be less than the demand before the imposition of the tie-in."n42 How much less? If different brands or different outlets are fungible, or at least rather close substitutes, it would cost next to nothing to induce a customer or supplier to deal with one firm rather than another. Although this point is frequently ignored, it was noted long ago by Turner in the tying contextn43 and by Blake and Jones as lapplied to foreclosure through vertical mergers.n44 In fact, it is even possible in some contexts that the restrictive practice would yield [*527] modest cost savings, so it might break even or offer a modest gain in the short run.n45 This point alone does not make the leverage argument. However, in light of such low (or even negative)n46 costs, the expected payoff need not be overwhelmingly high to induce firms to make attempts at leveraging their way to greater profits in the future.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n40 334 U.S. 100 (1948).

n41 Director & Levi, supra note 12, at 292.

n42 Bowman, supra note 12, at 21 n.8.

n43 See Turner, supra note 24, at 61 (discussing Northern Pac. Ry. v. United States, 356 U.S. 1 (1958)).

n44 Blake & Jones, supra note 10, at 454-55 (discussing Brown Shoe Co. v. United States, 370 U.S. 294 (1962)).

n45 See Bauer, supra note 17, at 299 (discussing decreased marketing costs in tying context). The same argument can be made in the other contexts as well. For example, see the discussion of Brown Shoe in Blake and Jones, supra note 10, at 454-55.

n46 If the costs truly were negative, i.e., the firms saved more in sales expenditures and the like than the inducements cost them, one would be inclined to approve of the restrictive practices simply because they are an efficient sales practice. Of course, this conclusion depends upon the leverage effects being remote and also assumes that other considerations, see supra note 10 and accompanying text, would not lead to an opposite conclusion.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

2. Use of Static Models. -- The distinction between profit maximization and monopoly extension can be understood, if at all, as a distinction between effects that can be observed in a static model and effects that only emerge from a dynamic perspective. Thus, if the courts' critice are attempting to demonstrate that monopoly extension cannot be accomplished using the restrictive practices they examine, one would expect their analysis to focus exclusively upon dynamic models. Quite surprisingly, the opposite has often been the case. To the extent this is true, their criticism is wholly beside the point, because the argument attempts to disprove the existence of long-term effects (that by definition only appear in a dynamic framework) by using static analysis (that by definition ignores such long-term effects).

The practice of predatory pricing illustrates this point.n47 Assume that a dominant firm lowers its prices substantially for a period of time in order to drive out competitors or scare off new or potential entrants. It is conceded that the firm will suffer losses in the short run.n48 The strategy is only pursued because the firm expects to make a greater profit in the long run than it would have in the absence of its predatory strategy.n49 Static analysis -- determining whether the firm profited immediately by its price cut -- would produce the misleading conclusion that predation was never profitable. One could only discover the potential profitability of this strategy by taking into account the dynamic [*528] effect on the firm's market position. In the predation context, commentators have long recognized this point,n50 but in examining other exclusionary practices, particularly with regard to leverage analysis, many have failed to appreciate the need to consider how the effects of practice over time can be dramatically different from its immediate impact on the firm employing it.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n47 No attempt is made here to offer any arguments concerning either the plausibility that predatory pricing occurs with any substantial frequency or the appropriateness of the many legal rules that have been offered to address the issue. The purpose here is solely to describe the static/dynamic distinction in a well-understood context.

n48 The nature of the losses that are necessarily borne depends upon how predation is defined. Under any definition, there are losses relative to the price the firm would otherwise have charged. For the purposes of this illustration, however, it would be simplest to assume that there are net operating losses (i.e., price does not cover variable cost) during this period.

n49 See, e.g., 3 P. Areeda & D. Turner, supra note 13, at 151 (1978) ("[T]he classically feared case of predation has been the deliberate sacrifice of present revenues for the purpose of driving rivals out of the market and then recouping losses through higher profits earned in the absence of competition.").

n50 It has still been the case that, in this context, criticism of static analysis has been in order. See, e.g., Comanor & Frech, Strategic Behavior and Antitrust Analysis, 74 Am. Econ. Rev. (Papers & Proceedings) 372 (May 1984); Williamson, Antitrust Enforcement: Where It's Been, Where It's Going, 27 St. Louis U.L.J. 289, 301 (1983) (quoting Baumol Quasi-Permanence of Price Reductions: A Policy for Prevention of Predatory Pricing, 89 Yale L.J. 1, 2 (1979)).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

A number of the leading attacks on the view that leverage can extend monopoly implicitly or explicitly take a static perspective. For example, Director and Levi's argument that it is costly "to impose additional coercive restrictions on . . . suppliers"n51 implicitly takes a static outlook, for their point is merely that the imposition will have a net cost in the present. Bowman indicates that if "the same output of the tied product can still be produced under circumstances consistent with competitive production of the tied product, no additional or new monopoly effect should be assumed."n52 That is true for the present, but Bowman does not assess the long-run impact of shifting sales of the tied product toward the frim employing the tying arrangement. Bork's arguments often implicitly assume that the firm employing the practice is not motivated by potential long-run effects on market structure.n53 Posner sometimes makes similar mistakes,n54 although on at least one [*529] occasion he explicitly notes such errors when made by others.n55 Other commentators in this tradition often take a static perspecitive.n56 Most recently, Easterbrook has advocated a new, more limited role for antitrust regulation that is based heavily on a static approach.n57

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n51 Director & Levi, supra note 12, at 292 (discussing United States v. Griffith, 334 U.S. 100 (1948)).

n52 Bowman, supra note 12, at 20. Bowman continues his argument by stating that "[i]f the tying seller gave a compensating advantage to the buyer, he might not be displaced. But in that event the tie-in would no longer be useful." Id. at n.6. The criticism in text applies equally to this point.

n53 See R. Bork, supra note 14, at 306-07 (noting lack of impact of exclusive dealing in one-store towns, not considering whether there will be any long-run impact on the market structure of the upstream industry); id. at 372-81 (noting potential short-run profit-maximizing effects of tying; dismissing entry barrier argument because no current monopoly profit results); see also Williamson, Book Review, 46 U. Chi. L. Rev. 526, 528 (1979) (criticizing Bork for failing to deal with strategic issues absent from static models).

n54 For example, he argues that if there are two manufacturers, each with half of the market, neither will be able to put the other out of business by purchasing all the distributors because "[t]he minimum price for the distributors' assets will be the cost to the second producer of being forced out of business, since the second producer will pay any price up to that cost in order not to be forced out." R. Posner, supra note 7, at 199. It is unclear why Posner believes this indicates that such a purchase of all distributors could not occur. That the winning bid would be high does not imply it would not be made if the winner can expect monopoly profits in the future as a result. Cf. Kreps & Wilson, On the Chain-Store Paradox and Predation: Reputation for Toughness 41-57 (Stanford University Graduate School of Business Research Raper No. 551, June 1980) (examining the role reputation plays in the context of a long-run analysis). Posner also argues that IBM's price reduction in response to entry by Telex, see Telex Corp. v. International Business Mach. Corp., 367 F. Supp. 258 (N.D. Okla. 1973), rev'd, 510 F.2d 894 (10th Cir. 1975), was desirable since it was above cost and brought lower prices to consumers. See R. Posner, supra note 7, at 194-95; id. at 193 ("can exclude only a less efficient competitor"). In making a similar analysis of price discrimination used as an entry deterrent, he claims that "[i]t is indefensible, however, to try to justify such a prohibition on the ground that it makes entry less attractive [since] [t]he effect of the . . . lower prices . . . is to increase output in [such] markets -- a good effect of discrimination." Id. at 206. Both of these arguments are true if one makes a static comparison to an identical state of affairs where the price is not reduced, with the result that output is less. But this ignores the effect of permitting such behavior on a firm with a long-run outlook, whose ability to lower prices in response to entry enables it to maintain elevated prices in the pre-entry state of affairs. See, e.g., Hay, A Confused Lawyer's Guide to the Predatory Pricing Literature in Bureau of Economics, Bureau of Competition Federal Trade Com'n, Strategy, Predation, and Antitrust Analysis 155, 161-66 (S. Salop ed. 1981) (presence in market of less efficient rival may substantially reduce price from monopoly level) [hereinafter cited as S. Salop]; Ordover & Willing, An Economic Definition of Predatory Product Innovation in S. Salop, supra, at 301, 365-70 (vertical leverage with similar effects). Whether this result is desirable is a very complicated question, see, e.g., sources cited supra supra note 36, but one that must be addressed before a conclusion such as Posner's can be reached.

n55 Posner correctly discusses how it may be possible for a firm to establish an effective threat of localized predatory behavior, enabling it to deter entry without often actually having to spend resources engaging in predation. See R. Posner, supra note 7, at 185-86; Posner, Chicago School, supra note 2, at 939-40. This observation concerns only one dynamic aspect of a firm's pricing strategy, as noted in the preceding footnote.

n56 See, e.g., Texas Note, supra note 16, at 911 n.70 (using static analysis to criticize Slawson's "fanciful[ ] suppos[ition]," which was in fact grounded in part on entry barriers -- a dynamic effect, see Slawson, supra note 24); id. at 917 (same). In fact, this Note, carrying the subtitle "Invalidating the Leveraging Hypothesis," employs only static analysis. See, e.g., id. at 925 ("tying arrangements can increase a monopolist's profits only if its customers have differences in demand elasticities that are indicated by their combining the commodities in different proportions" (emphasis added)).

n57 Although Easterbrook's approach is based upon a strong belief in the dynamic forces of competition, see, e.g., Easterbrook, The Limits of Antitrust, 63 Tex. L. Rev. 1, 5-6 (1984), much of this argument as to specific practices, as well as his proposed framework for assessing antitrust liability, is strongly dependent upon a static perspective. The problem is most generally illustrated by his framework, which requires that antitrust plaintiffs meet five proof requirements as a prerequisite to further analysis of the defendant's practices. The third is that there is widespread adoption of identical practices in the industry. See id. at 29-31. This filter limits antitrust to horizontal action, implicitly ruling out all single-firm activity that has been the focus of § 2 of the Sherman Act and § 3 of the Clayton Act. Such an approach only makes sense in a static framework that takes market structure as given and sees collaborative actions among existing competitors as the only possible dangers.

His fourth requirement, which limits antitrust challenges to situations where output and market share fall, see id. at 31-33, even more obviously limits analysis to a static perspective. See Markovits, The Limits to Simplifying Antitrust: A Reply to Professor Easterbrook, 63 Tex. L. Rev. 41, 83 (1984) (offering a similar criticism of Easterbrook's fourth requirement). As the text describes, practices are alleged to have the opposite effect in the short run (when actions are likely to be challenged), which would, independently of the third requirement, exonerate all single-firm monopolization (and some group monopolization). Long after the fact, a plaintiff might be able to overcome this hurdle, but this assumes that the plaintiff is still in existence, that causation (which much of Easterbrook's article suggests would be well beyond the competence of courts to determine) can be demonstrated, and that long after-the-fact enforcement is the preferred alternative. In addition to using the static frame of analysis, Easterbrook gives it an explicit slant in favor of antitrust defendants. "If manufacturer's sales rise, the practice confers benefits exceeding its costs. If they fall, that suggests (but does not prove) that there are no benefits. 'Does not prove' because other things in the market may have happened at the same time." See Easterbrook, supra, at 31 & n.65 (emphasis added). It is of course curious how "other things" can explain away changes in one direction (presumptively adverse to defendants) but cannot explain away changes in the other (presumptively adverse to plaintiffs).

Easterbrook's fifth requirement is basically that suits brought by rivals should not be entertained. See id. at 33-39. Again, from a static perspective, anticompetitive action is associated with higher prices, which would benefit rivals. Hence, rivals' complaints are seen as evidence against this possibility. From a dynamic perspective, the static picture could reveal lower prices and damage to rivals, that damage directly leading to the long-run effects on market structure. Easterbrook suggests, however, that the only two possibilities are static anticompetitive effects (reduced output and higher prices in the short run) and efficiency, see, e.g., id. at 36 (discussing GM-Toyota agreement), implicitly ruling out dynamic effects. See also Markovits, supra, at 84 (noting that Easterbrook also omits the possibility that both of his hypothetical effects might exist simultaneously).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

[*530] It is hard to understand why so much of the criticism of leverage theory operates primarily in a static framework when even some of the earliest and most unsophisticated statements of the leverage theory were explicitly grounded in a dynamic model.n58 For example, statements concerning foreclosure typically look to the long-run effect on the market position of competitors.n59 Arguments concerning the erection or maintenance of entry barriersn60 also have been grounded explicitly in a dynamic context, as have more sophisticated arguments concerning reputation effects of practices by established firms,n61 strategic positioning,n62 and other effects on firms' costs.n63 There have been [*531] ad hoc discussions of, and replies to, such arguments, but they would have assumed a more central role in the debate if the focal point for the primary analysis were not so fixed on static models.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n58 One possibility is that dynamic models are much more imposing and less likely to lead to determinate conclusions than static analysis. I find Spence's reply compelling: "[T]he state of our understanding of both dynamic strategy and intertemporal market performance is currently insufficient to justify confident conclusions with respect to rules and standards. But I do think that it is better to admit ignorance than to defend rules based on incomplete static models of industries." Spence, Competition, Entry, and Antitrust Policy in S. Salop, supra note 54, at 45, 87.

n59 See, e.g., Standard Oil Co. v. United States, 337 U.S. 293, 308-09 (1949).

n60 See infra subsection II-B-4.

n61 See, e.g., Kreps & Wilson, Reputation and Imperfect Information, 27 J. Econ. Theory 253 (1982); Milgrom & Roberts, Predation, Reputation, and Entry Deterrence, 27 J. Econ. Theory 280 (1982); Williamson, supra note 50, at 304-06; Kreps & Spence, Modeling the Role of History in Industrial Organization and Competition, at 24-29 (Harvard Institute of Economic Research Discussion Paper No. 992, July 1983); Kreps & Wilson, supra note 54. An example of an analysis in which these long-term effects are ignored is R. Bork, supra note 14, at 153 (discussing long-term predation).

n62 See, e.g., Dasgupta & Stiglitz, Uncertainty, Industrial Structure, and the Speed of R & D, 11 Bell J. Econ. 1, 11-15 (1980); Dixit, The Role of Investment in Entry-Deterrence, 90 Econ. J. 95 (1980); Eaton & Lipsey, Exit Barriers Are Entry Barriers: The Durability of Capital as a Barrier to Entry, 11 Bell J. Econ. 721 (1980); Gilbert & Newbery, Preemptive Patenting and the Persistence of Monopoly, 72 Am. Econ. Rev. 514 (1982); Salop, Strategic Entry Deterrence, 69 Am. Econ. Rev. 335 (1979); Spence, Entry, Capacity, Investment and Oligopolistic Pricing, 8 Bell J. Econ. 534 (1977); Williamson, supra note 50, at 311 ("Strategic behavior is an interesting economic issue only in an intertemporal context in which uncertainty is featured." (emphasis in original)).

Easterbrook has argued that although such models of strategic behavior "are cause for serious reflection[,] . . . [they] are not a sufficient basis for antitrust condemnation of ambiguous business practices, a view shared even by proponents of some of the strategic models." Easterbrook, Is There a Rachet in Antitrust Law?, 60 Tex. L. Rev. 705, 710 (1982) (emphasis added). In the footnote that follows, he explicitly refers to disclaimers in the Eaton and Lipsey article I have just cited and in Spence, supra note 58. Although caution is surely in order, as strongly contended in Section III-C, infra, Easterbrook's claim is somewhat overstated. Eaton and Lipsey's statement that "[a]pplication to policy is clearly premature," Eaton & Lipsey, supra, at 728, quoted in Easterbrook, supra, at 710 n.21, hardly supports Easterbrook's advocacy of rolling back antitrust proscriptions on the ground that preliminary findings such as Eaton and Lipsey's will ultimately prove unimportant or wrong. Also misleading is Easterbrook's comment that Spence "caution[s] against applying strategic models to hunt for predation in antitrust cases." Easterbrook, supra at 710 n.21 (quoting Spence, supra note 58, at 81). Spence also states that regulation of predation "should not serve as the main weapon in the arsenal for dealing with the market power and related performance problems." Spence, supra note 58, at 81 (emphasis in original). He cautions against the "widely accepted myth that the pursuit of profit by 'legitimate' means will always result in the right results. Not only is it not true . . .; acceptance of it biases the process of changing the market outcome. . . ." Id. Easterbrook's comment that "Spence and others who would formulate models of strategic interaction are also careful to observe that the conclusions of their work depend on adherence to the conditions" specified therein, Easterbrook, supra, at 710 n.21, similarly fails to support his conclusion that exclusionary conduct is not a significant cause for concern -- a conclusion that assumes that such strategic interactions are implausible.

n63 See, e.g., Williamson, supra note 50, at 309 ("The upshot is that the [predatory pricing literature's] equally efficient rival criterion is primarily suited to static circumstances where historical differences and continued cost asymmetries may be presumed to be absent." (emphasis in original)).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

3. Overlooking the Free Rider Problem. -- When analysis of anticompetitive practices moves beyond the static context, it becomes necessary to evaluate more carefully the impact of the self-correcting processes of the marketplace. In a competitive market, if a single firm raises its price, others will maintain their original, lower prices, and the firm cannot profit because it is unable to sell its product at the higher price. Firms that can raise prices above marginal cost and still retain a substantial portion of their customers are said to have market power, which necessarily implies that other firms will be unable or unwilling to produce at lower prices the output necessary to deny the powerful firm its profit.n64 In the long run, however, other firms in the industry will be [*532] able to expand their output, and new firms will be encouraged to enter the industry that offers prospects for high profits.n65 In many settings, market power tends to erode over time. It may still be desirable to regulate conduct that may only be viable in the short run. After all, costs incurred for merely a few months, years, or decades are costs nonetheless. There is, for example, little disagreement that price-fixing should be broadly proscribed even in those industries where the prospects for substantial long-run excess profits are not great.n66

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n64 See, e.g., Landes & Posner, supra note 33, at 937, 944-51.

n65 See sources cited in supra note 36.

n66 See, e.g., L. Sullivan, supra note 7, at 186, 192.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

This analysis must be adjusted slightly to deal with the leverage context. Here the issue often concerns the ability of a firm over a period of time to enhance its power in some market.n67 For the firm to succeed, and thus for society to lose out, existing competitive forces must be insufficient to stifle the firm's progress.n68

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n67 Similar analysis is applicable to some attempts to maintain entry barriers.

n68 Another issue, which is precisely the question addressed in the preceding paragraph, is whether significant power is possible in the new market. Since that issue is often not decisive and raises a number of questions unrelated to the focus of this Article, it will not be explored here, except for the discussion below in subsection II-B-4.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

One potential force that might limit the build-up of market power is action by firms that will be the losers in the future -- e.g., the customers and suppliers of the industry undergoing change -- which have an interest in stopping the firm attempting to leverage its way to a second monopoly. The problem, however, is that their interest will often be insufficient to motivate them to act in a manner that will prevent the leverage from being effective. Resistance by such firms is subject to the free rider problem. If, for example, many customers are buying the product of an industry that is becoming more concentrated, each buyer will be unwilling to incur a significant expense in preventing the concentration because it bears the total cost of any of its efforts but only receives a benefit in proportion to its share of the market. Each buyer reasons that it can take a "free ride" on the efforts of the other buyers who will bear the expense of preventing the rise of concentration. If the others are inclined to reason in much the same way, it would not have made much difference if the buyer had tried to prevent the inevitable, except that the valiant buyer would be out whatever the hopeless defenses had cost.n69

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n69 This refusal to go along with the collective best interest corresponds precisely with the incentive for firms in a cartel to cheat. There may be circumstances in which a group of buyers could overcome the free rider problem in fighting the firm attempting to monopolize their source of supply, but it can hardly be assumed that this will generally be the case, as is discussed infra pp. 534-36.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Commentators arguing that leverage will generally be impossible often rely upon the ability of potentially affected firms to thwart efforts to extend monopoly, failing in any way to take account of the free rider problem. Director and Levi, when criticizing the Court's decision in [*533] United States v. Griffith,n70 argue that "it [would not] seem to be in the interest of the suppliers to encourage the growth of monopoly among the exhibitors."n71 Concededly the suppliers would be displeased with the prospect, but no reason is offered indicating why it would have been in the individual interest of any supplier to undergo the cost of carrying on the fight. In discussing United Shoe,n72 Bork argues that "exclusion would of course be detrimental to the shoe manufacturers [who bought from United]. They would prefer not to have to deal forever with a monopolist."n73 Their preferences generally stated, however, do not determine whether it would have been in their financial interest to act on such preferences. Bork repeats the mistake when discussing exclusive dealingn74 and price discrimination.n75 When analyzing reciprocity, Areeda and Turner claim that "the defendant's customers will not lightly join in the destruction of his rivals, for they would not want to be totally dependent upon him."n76 The free rider argument indicates that although they will not "lightly join in," they are likely to join nonetheless. Perhaps this is why Areeda and Turner continue by noting that "[a]t least when other rivals."n77 That reservation is precisely correct, but it moots their argument in this context. Presumably other things will not be equal -- there will be some cost so long as the defendant is offering some inducement.n78

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n70 334 U.S. 100 (1948).

n71 Director & Levi, supra note 12, at 292.

n72 United States v. United Shoe Mach. Corp., 110 F. Supp. 295 (D. Mass. 1953), aff'd per curiam, 347 U.S. 521 (1954).

n73 R. Bork, supra note 14, at 140.

n74 "The large firms, here the theaters, would not stand for such nonsense for a moment. They would support new entrants in the production and distribution of advertising films or enter that activity themselves. They would certainly not use their own market strength to give a monopoly to their suppliers." Id. at 308 (discussing FTC v. Motion Picture Advertising Serv. Co., 344 U.S. 392 (1953)). Since the market consisting of the buyers was quite concentrated, the magnitude of the free rider problem would be less, but it is still not obvious how the theaters would have behaved.

n75 "[S]ellers who saw a monopoly developing at the customer level would offer the lower prices to other customers to prevent that outcome." Id. at 390. Whichever seller takes this altruistic action must, however, bear the full cost, while the benefits accrue to all.

n76 5 P. Areeda & D. Turner, supra note 13, at 175.

n77 Id. It is uncertain from this discussion whether Areeda and Turner are aware of the implications of the free rider argument in this context. In his earlier analysis of the tying issue, Turner noted that purchasers going along with the sellers engaging in the restrictive practice may suffer in the long run, but comments that "the risk was deemed negligible by the buyers and lessees concerned." Turner, supra note 24, at 61. Turner may well have been right on the facts of the case he was discussing, Northern Pac. Ry. v. United States, 356 U.S. 1 (1958), but the free rider problem would have provided an independent explanation for the buyers' and lessees' willingness to go along.

n78 As noted in subsection II-B-1, such inducements may cost the defendant very little.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Posner is much more aware of the issue than the other commentators, [*534] although his treatment is also unsatisfactory. For example, he argues that purchasers will be in a position to thwart predatory pricing attempts by continuing to buy from higher-priced sellers and thus can preserve the competitive structure for the source of supply.n79 Unlike the others, however, Posner admits that "[t]his analysis does not, however, completely negate the possibility of effective predatory pricing" because, among other things, the purchaser "may decide to act as a 'free rider.'"n80 Posner also argues that purchasers would not go along with exclusive-dealing contracts that were exclusionary.n81 Here he concedes that "each shoe manufacturer might reason [along the lines suggested by the free rider argument; however], there is no certainty that he would. If people always thought in such ways, no cartel, other than one that was legally enforceable, would be even partially effective."n82 This brief concession in the first instance is excessively mild under the circumstances, and his refusal to concede any ground in the second instance is indefensible. Of course it is not absolutely certain that every shoe machinery manufacturer would succumb to self-interest, or that the self-interest of each would necessarily prevent action. The point of the free rider argument is not that "people always [think] in such ways," but that they orten do. Presumably, effective cartels are not observed in every industry in every region at all points in time precisely because firms often think in such ways, but cartels are in fact sometimes observed, suggesting that sometimes they do not. The free rider problem should be identified and taken seriously where conditions are conducive to its existence.n83

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n79 See R. Posner, supra note 7, at 184.

n80 Id. at 184-85 (emphasis added).

n81 See id. at 202 (generally); id. at 203 (applying argument to United Shoe). Easterbrook makes a similar argument when discussing full requirements contracts, see Easterbrook, supra note 57, at 25 & n.52, which Markovits criticizes for reasons parallel to those offered here, see markovits, supra note 57, at 78 & n.88.

n82 R. posner, supra note 7, at 204 (emphasis in original).

n83 Posner argues that a "second producer can overcome the free-rider problem by offering to lease the machine . . . on the condition that a specified minimum number of shoe manufacturers lease machines from him. Each lessee would find it advantageous to enter into such a contract . . . ." Id. at 204. Posner surely is right since, by definition, if individual actors subject to the free rider problem can act in concert there is no more free rider problem. The question is whether concerted action is plausible under the circumstances. Such action may be hard to coordinate, subject to cheaters, and illegal as well. See infra note 87. One could also argue that because such coordinated responses have not been observed in reported cases, there must have been no threat to begin with. My point here is that the free rider problem could plausibly have inhibited such response. On the choice between such competing explanations, see Section III-C, infra.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Finally, Easterbrook, writing in the context of predatory pricing, has recently made the first significant attempt to deal with the free rider problem. His primary argument expands upon the idea that buyers can act collusively to protect the dominant firm's rivals and thus preserve [*535] competition in the long run.n84 His conclusions, however, remain unpersuasive. The complexity of the analysis makes it impossible to examine fully each aspect here. Three defects are most notable. First, he implicitly makes highly unrealistic assumptions concerning the lack of transaction costs.n85 Second, he extrapolates from indeterminacy in formal models and uncertainty in the real world to confident results that by no means represent the only possibilities and in some circumstances seem quite unlikely.n86 Third, he does not fully escape even rather direct [*536] applications of the free rider problem to his hypothesized solution.n87 Despite these criticisms, it should be noted that Easterbrook has advanced the analysis of an issue that most have long ignored in numerous antitrust contexts.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n84 See Easterbrook, Predatory Strategies and Counterstrategies, 48 U. Chi. L. Rev. 263, 271-72, 287, 293 (1981).

n85 The hypothetical possibility that buyers may get together with victims of predation or new entrants is itself of no significance, for such arrangements are always possible in a world with perfect costless bargaining. The same analysis demonstrates the lack of any need for antitrust laws since consumers could always erode market power in collaboration with entrants and fringe firms. Moreover, this would hardly be necessary because if consumers could overcome free rider problems and combine against a monopolist or cartel they could directly bargain for an efficient result, just as the public would not need protection from pollution laws if such collective action were possible. Easterbrook's flaw is that he does not show that the strategy he contemplates is realistic. For example, he assumes that large groups of consumers will be able to strike long-term contracts setting both prices and quantities with firms about which little may be known and that are under attack by the predator. See Zerbe & Cooper, An Empirical and Theoretical Comparison of Alternative Predation Rules, 61 Tex. L. Rev. 655, 697 (1982) (footnotes omitted):

[Easterbrook] develops an elaborate scenario in which purchasers band together in order to defeat attempted predation at the producer level. This line of reasoning is ultimately indecisive. Offering possible solutions does not demonstrate that the problem will actually be solved. Indeed, Professor Easterbrook's solution of having customers execute long-term supply contracts may involve considerable private costs and ultimately is improbable. For each scenario where predation fails, it is possible to construct another one where it succeeds.

n86 See Easterbrook, supra note 84, at 269 ("predator's rival, after all, has the same incentive as the predator to ride out the price war and collect monopoly profits once one of them has collapsed"); id. at 285 ("[It is] unlikely that any one of these strategies dominates all others. The soultion is indeterminate. . . ."); id. at 285 n.47 ("[t]here is not often a solution to an n-player game of the sort described in the text"); id. at 286 ("If each firm is uncertain of the other's costs, neither has a clearly superior strategy."); id. at 293 ("there is no clearly dominant strategy here"). Easterbrook's ability to reach confident conclusions in spite of this problem has been aptly criticized. See Zerbe & Cooper, supra note 85, at 697-98 (footnote omitted):

The interaction between a predator and its target is similar to a game of poker. With each player able to bluff or play straight, the outcome depends on both the play and the other player's reaction to that play. This interdependence means that the ultimate outcome hinges on expectations. In a game of sufficient complexity, game theory is indeterminate, and arguments claiming that one player will always or usually win are invalid.
Where he does make determinate statements at the more detailed analytical level, his conclusions are simplistic. For example, one of his major points rests on the well-known argument that, in a sequence of markets, once all but the last has been entered, a monopolist A will have no incentive to engage in predation if the last is entered; given that result, the same argument applies to the next-to-last market. "The process can be extended to show that A would not predate at all," id. at 286. (Analogous reasoning would prove that strikes and wars would never occur.). If one assumes that the number of markets, or the number of potential entrants, is not fixed in advance but rather is subject to some uncertainty, his argument may fail. See Kreps & Wilson, supra note 61, at 275. In addition, if reputation effects are taken into account, see sources cited in supra note 61, the opposite is quite plausible. These particular criticisms illustrate the more general point that when a more complete account is taken of real differences between, for example, a large predator and a potential entrant, the worries Easterbrook criticizes using oversimplified formal analysis are seen to be more serious.

n87 Easterbrook's argument does not clearly indicate how a group of buyers will be able to accomplish the coordinated long-term contracting arrangement. If any buyer expects the contract to keep the market competitive, it will have an incentive to purchase in the interim from the predator at the lower predatory price and then pay the competitive price in the future after the predation has failed. If the contracting arrangement is expected to fail, it will not protect the buyer's future but will still cost the buyer something in the short run. To suggest that a victim or entrant could overcome this problem by a requirement that the contract not be effective unless most buyers agree -- in addition to raising possible legal problems of its own -- is analogous to suggesting that voluntary contributions could support national defense or bribe polluters to change their ways so long as contributions are contingent on sufficient participation. Overcoming the free rider problem is possible in some contexts, but such solutions cannot fairly be taken for granted, especially when far more plausible conjectures concerning the likelihood of predatory behavior are rejected as uncertain and indeterminate.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

It is particularly perplexing that the free rider problem has received so little attention by commentators critical of leverage theory. These commentators tend to be the same people who are critical of the Court's prohibition of resale price maintenancen88 and who were critical of territorial restrictions until the latter were sanctioned in Sylvania.n89 The leading efficiency justification offered for these vertical arrangements is that they allow manufacturers to circumvent the free rider problem at the retailer level.n90 Those commentators who advocate the curtailment of antitrust intervention seem to display this economic wisdom when it supports their position, but ignore or cast aspersions on such wisdom when it cuts against them.n91 Hopefully this subsection makes clear that free riders travel on two-way streets.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n88 See Dr. Miles Medical Co. v. John D. Park & Sons Co., 220 U.S. 373 (1911).

n89 Continental T.V. v. GTE Sylvania, 433 U.S. 36 (1977).

n90 See, e.g., R. Bork, supra note 14, at 290, 430-31, 435; R. Posner, supra note 7, at 149-50; Posner, The Rule of Reason and the Economic Approach: Reflections on the Sylvania Decision, 45 U. Chi. L. Rev. 1, 6-10 (1977).

n91 Interestingly, Posner's summary of the "Chicago School" positions notes the use of the free rider problem argument in the resale price maintenance context and omits any analysis of its applicability in the long-run contracting context, both within the span of a single page. See Posner, Chicago School, supra note 2, at 927.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

4. Tendency to Assume Perfect Markets. -- Markets do not always function in accordance with the textbook model of perfect competition, and the economic analysis of any situation must be adjusted accordingly. In [*537] fact, the whole of antitrust concerns the study of imperfect markets. Thus, it seems obviously counterproductive to carry on antitrust analysis without carefully addressing the possibility that defects in the competitive process are central to understanding the behavior under scrutiny. Although the cry to take note of market imperfections is hardly novel, the frequent failure to heed this warning in the context of analyzing the plausibility of leverage warrants its consideration here. But since this issue arises not merely as an incident of the leverage debate, but rather forms the core of much of antitrust analysis, I will allow one well-known example to serve as the necessary reminder.

The most common area of dispute in this respect concerns the capital market. The issue arises with vertical integration, as with mergers, where the argument offered by those suspicious of the practice is that potential entrants will be less able to regulate the behavior of existing firms when it becomes necessary to enter at two levels in order to compete effectively. It is argued that it will be more difficult to raise the capital for this endeavor than if, for example, it were only necessary to enter at the manufacturing level, relying upon existing retailers to sell one's output. Director and Levi's response was that [i]t is not evident whether the argument is based on an imperfection in the capital market, on the reluctance to assume the consequent risks, on the economies associated with raising large amounts of capital, or on the less efficient scale imposed on rival firms."n92 What is surprising is not that these questions are raised, as they should be, but that they are often deemed sufficient to end the discussion. Director and Levi only claim to be raising the issue,n93 but the general tone of their articlen94 as well as the views often attributed to Director on this subject suggest that their questions are intended to be leading. Bork argues that "[t]here [is] nothing to the notion that an established firm might integrate vertically in order deliberately to raise the capital requirements of entry" primarily because "[c]apital suppliers take risks when the stakes are high," and it is assumed that excessive profits can be reaped upon entering the integrated industry.n95

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n92 Director & Levi, supra note 12, at 291.

n93 See id. at 293.

n94 See, e.g., id. at 296 (concluding sentences: "We do suggest that in the future there may well be a recognition of the instability of the assumed foundation for some major antitrust doctrines. And this may lead to a re-evaluation of the scope and function of the antitrust laws.").

n95 R. Bork, supra note 14, at 241-42, 323 (emphasis added).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

This position ignores a number of obvious realities.n96 For example, [*538] the position assumes that no substantial risk arises from the fact that the new entrant is inexperienced, that there is no greater risk of failure when two new ventures must be launched rather than one, and that any such risks can be avoided by coordinating a simultaneous entry at both levels by two independent firms.n97 Posner, who is also critical of the entry barriers argument, takes a more moderate position. In the context of predatory pricing, for example, he concedes that "[t]his might lengthen the time it took for entry -- . . . more time may be required to launch a large enterprise than a small one -- but it should not preclude entry entirely."n98 Posner's analysis indeed discusses how the delay may be far longer in some industries than others, in particular noting the example of U.S. Steeln99 -- whose market share took decades to decline substantially -- which is particularly relevant to the vertical integration question.n100 Whether expressed as a diminished probability of entry or as delay before new entry, there is obviously a cause for concern. As noted previously,n101 for example, were it not for the fact that such delays and uncertainties in entry are widespread, there would be no need for the proscription against price-fixing. The discussion of entry barriers in this context serves as a reminder that market imperfections of this sort are not matters of occasional relevance [*539] to antitrust analysis, but lie at its very heart.n102

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n96 This point has been noted by, e.g., Porter, Strategic Interaction: Some Lessons from Industry Histories for Theory and Antitrust Policy in S. Salop, supra note 54, at 449, 466 ("There is widespread belief among managers that the diversified firm gains resulting advantages in access to capital compared to single-business firms, implying imperfections in the external capital market." (footnote omitted)); Williamson, supra note 50, at 301:

The possibilities that remediable impediments to entry might arise and that such circumstances are identifiable ought to be considered. Consistent with his neglect of strategic factors, Bork seems unwilling to entertain such possibilities. This unwillingness is due chiefly to his implicit assumption that labor and capital markets operate without friction, so that every market outcome is presumptively a merit outcome and further discussion is pointless. Once transaction costs are admitted, however, the assumption of frictionless markets no longer applies, the possibility of introducing strategic impediments to entry arises, and the main argument needs to be qualified.

n97 See generally J. Bain, Barriers to New Competition 155-56 (1956) (vertical integration requiring increased capital is necessary to avoid cost disadvantage in steel, copper, automobile, and farm machinery manufacturing, and in petroleum refining); L. Sullivan, supra note 7, at 448 (vertically integrated entry requires additional capital and a broader range of entrepeneurial skills and judgments); Blake & Jones, supra note 10, at 447-48 (where outlets or suppliers are tied up by a few vertically integrated companies, new entrant or unintegrated firm may be barred from market by preemption of most desirable outlets, substantial inelasticity of outlets or suppliers, or lack of capital to establish parallel vertically integrated operations); Blake & Jones, In Defense of Antitrust, 65 Colum. L. Rev. 377, 392-93 (1965) (integrated entry in fully integrated industry requires both more capital and more costly capital because risks are cumulative); Kaserman, Theories of Vertical Integration: Implications for Antitrust Policy, 23 Antitrust Bull. 483, 506-07 (1978) (vertically integrated entry is more difficult because a barrier to entry at any stage threatens the whole enterprise); Williamson, Book Review, 83 Yale L.J. 647, 656-68 (1974) (capital costs of vertically integrated entry can be a barrier because inexperienced firms are considered more risky by lenders).

n98 R. Posner, supra note 7, at 187 (footnote omitted).

n99 See id. at 58.

n100 See id. at 197-98. He emphasizes that the effect is one of delay rather than preclusion. See id. at 197 n.41.

n101 See supra p. 532.

n102 See supra pp. 536-37.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

III. ALTERNATIVE EXPLANATIONS OF ALLEGEDLY MONOPOLISTIC BEHAVIOR

This Article has established only that the traditional understanding of leverage is plausible in some circumstances, not that leverage is the inevitable result of various restrictive practices.n103 As Section II-A demonstrated, however, it does not follow that the absence of the traditional leverage effect eliminates all cause for antitrust concern. Depending upon how one resolves the controversy concerning other uses and effects of restrictive practices, it is quite plausible that they are sometimes desirable, sometimes undesirable, and sometimes irrelevant. It then becomes necessary to investigate which effects are most likely in various circumstances. Such an analysis could lead one to favor: (1) nearly absolute prohibitions if the effects were almost never desirable, and if it were difficult to isolate any exceptions; (2) qualified per se rules if fairly accurate and administrable rules of thumb could be developed to distinguish groups of cases; (3) a rule of reason if it were thought possible to implement through the litigation process, and deemed necessary to examine each case independently because of great case-by-case variations; or (4) no regulation if it were thought that detrimental effects were rare and difficult to segregate from beneficial effects.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n103 Markovits, who has written extensively on the leverage theory as applied to tie-ins and reciprocity, at times characterizes the position of those advocating restriction as believing that the profitability and leverage effect of such restrictions is inevitable. See Markovits, Tie-Ins, Reciprocity, and the Leverage Theory Part II: Tie-Ins, Leverage, and the American Antitrust Laws, 80 Yale L.J. 195, 204, 243 (1970). If that were the case, tie-ins and reciprocity would be seen between all conceivable goods and services everywhere one looked. Some statements by courts and commentators can be so interpreted, but it seems an unnecessary diversion to focus much attention on such an extreme position. A more limited claim would be that sometimes the restrictions are profitable, and when the restrictions are used the leverage effect is inevitable. Some commentators explicitly take this position, although the impact they claim inevitably accompanies the restrictive practice is more modest than the simple second monopoly position. See, e.g., Slawson, supra note 24, at 676 ("foreclosure which any tie-in effects" (emphasis in original)), 690 ("there is always at least some lessening of competition" (emphasis in original)).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

To reach any of the above conclusions, it would be necessary to explore not only whether and when the traditional leverage effect was possible, but also what competing explanations for an antitrust defendant's conduct might be more plausible. Section A demonstrates that some of the most commonly offered alternative explanations, whether or not they would exonerate a particular defendant, are not very plausible in many circumstances. It may often be the case that all the regularly debated theories have rather weak claims to explaining observed behavior. Section B examines some alternative explanations that have [*540] received too little emphasis in the debate over restrictive practices. Section C explores the process by which competing hypotheses are evaluated in this area.

A. The Plausibility of Alternative Explanations Derived from Profit Maximization

It is not my purpose here to add another round to the endless debate involving each of the many explanations offered for every variety of restrictive practice. Instead, I will briefly note three recurring weaknesses in many of the arguments which have been advanced to explain restrictive practices. I do not intend these criticisms to apply either to every argument, or to the application of any argument to every case. The weaknesses I note, however, arise and are overlooked with sufficient frequency to deserve systematic presentation. The following arguments generally indicate either that a hypothesized motivation is implausible or that, even if the imputed motivation is correct and deemed worthy, little would be lost if the restrictive practice were prohibited.

1. Less Restrictive Alternatives Ignored or Underestimated. -- Many commentators underplay the importance of alternative business methods by which a defendant might achieve its objectives without the need to engage in the challenged restrictive practices. The ability of the defendant to accomplish its stated objectives without relying on the restrictive practice under consideration has two implications. First, it casts doubt on the hypothesis that the defendant's alleged, permissible purpose is its actual purpose, especially if the less restrictive approach appears to be more effective at accomplishing the stated goal. Second, to the extent the alleged purpose is deemed affirmatively desirable, the availability of an alternative indicates that a general prohibition of the restrictive practice would result in little, if any, cost.n104

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n104 Easterbrook claims that "alternatives may be more costly, but the defendant will not be able to show the amount of the difference." Easterbrook, supra note 57, at 9. This does not lessen the importance of considering the issue. First, if the costs of alternatives are the true motivation fot employing the more restrictive practice, one would expect that cost analyses and decisionmaking procedures contemporaneous with adoption of the practice would support the inference. Second, the belief that costs cannot always be assessed accurately does not necessarily indicate that the less restrictive alternative analysis will be unilluminating in all cases. The analysis itself is avoided or underplayed far too often, and rather confident conclusions concerning the relative costs of alternatives are often possible.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

The most common instance of downplaying or ignoring less restrictive alternative concerns the price discrimination motivation for restrictions, particularly tying arrangements.n105 It is often alleged that the monopoly good (the tying product) -- typically a machine -- is sold or leased with a second good (the tied product) -- typically something [*541] used with the machine -- that is priced above its cost. Thus a profit is recouped in proportion to the customer's use of the machine.n106 This practice is often called methered pricing because sales of the tied good act as a meter that measures use of the machine.n107 But why not simply use a meter if this is the purpose?n108 Or examine the company's books? Either of these alternatives may be necessary even if a tying arrangement is relied upon since the seller must verify that the buyer is in fact using its tied goods with the machine rather than those available at a lower price from rivals.n109 Moreover, unless the seller was, by a happy [*542] coincidence, already producing a sufficient quantity of the tied good,n110 the expense of umplementing metered pricing through expansion of business in a product area that typically has great technological differences with the machine itself may well make tying more costly.n111 Of course, there will be some instances in which tying may be the only feasible way to effectuate price discrimination.n112 But price discrimination, and many other objectives, can often be pursued as well or better without tying arrangements.n113

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n105 The less restrictive alternative issue in this and other contexts is explored in Kaplow, supra note 20, at 1863 n.172.

n106 One instance of the over-application of this argument is the use of International Salt Co. v. United States, 332 U.S. 392 (1947), as an illustration. See, e.g., M. Handler, H. Blake, R. Pitofsky & H. Goldschmid, Cases and Materials on Trade Regulation 1144 (2d ed. 1983); Bauer, supra note 17, at 294-95. Omitted from these renditions of International Salt's facts is the inclusion in the typing agreement of a provision that permitted buyers to purchase their salt elsewhere if International's prices were higher than those of any competitor. See 322 U.S. at 394 n.5.

n107 The example of tying staples to a shoe button stapling machine, see Heaton-Peninsular Button-Fastener Co. v. Eureka Specialty Co., 65 F. 619 (C.C.W.D. Mich. 1895), rev'd 77 F. 288 (6th Cir. 1896), was used by Bowman in describing this argument, Bowman, supra note 12, at 23-24, and has been replayed ever since, see, e.g., P. Areeda, Antitrust Analysis 735-36 (3d ed. 1981).

n108 Bowman seems to concede as much, Bowman, supra note 12, at 24, but fails to connect the implications of his argument to his conclusions. Ferguson seems to rely upon the costs of installing, inspecting, and preventing tampering with meters as the basis for the manufacturer's preference for tying in some circumstances. See Ferguson, supra note 7, at 556. While he may sometimes be right, his analysis overlooks the similar, but often greater, costs of tying and the likely need to meter in any event, as noted in the text follows. Moreover, if the issue is whether there would be a significant cost in banning tying arrangements that sometimes facilitate desirable price discrimination, it is relevant that the costs of the alternative are at most an insignificant portion of what is at stake.

n109 Markovits' assertion to the contrary, see Markovits, supra note 103, at 237; Markovits, supra note 17, at 1380, may sometimes be true, but he ignores most of the costs of the tying arrangement and limits his argument to the case where the tied product would be easy to identify. If he means identification at plant inspections, the advantages over meter-reading begin to vanish. And if plant inspections or sampling of the end-product is necessary, detection may be most difficult. Computer cards can be labelled, but salt and staples are another matter. However, Markovits ultimately seems willing to admit that tie-ins are not necessary to accomplish meter pricing, which, he indicates, militates against a prohibition since the results can be accomplished in any event. See Markovits, supra note 17, at 1440-41. This point is quite right. If price discrimination were considered undesirable, and typing were simply the most efficient way of bringing it about, it would not follow that a prohibition was in order if the result would be to cause firms to continue their price discrimination in ways that simply cost more. However, the argument can also cut the other way: if price discrimination is viewed as one of the few beneficial uses of tying, his argument would indicate that there is little if any cost in a flat prohibition. See id. at 1429-31. Of course, the more costly the alternatives are, the smaller would be the proportion of firms that would continue to find price discrimination worthwhile.

It is also possible that the use of the tied good will not be a perfect proxy for use of the tying product, as when there are variable proportions in production. In addition, using the buyers' books may allow monitoring of additional factors that better facilitate tailoring the price charged to the buyer's valuation of the machine. For example, Burstein's discussion of the use of full-line forcing to extract profits, Burstein, supra note 17, at 89, overlooks the fact that either simple franchise fees or annual charges based on book measures may be a far more precise method of accomplishing the intended result. In fact, without initial reliance on information from buyers' books, it may be difficult for the seller to collect the information it needs to determine what the prices for the involved products should be in the first place under a tying or full-line forcing arrangement.

n110 Of course, it is precisely in such cases, which include a substantial majority of those that have come before the Supreme Court, that the leverage hypothesis is a possible alternative explanation. See United States v. Loew's, Inc., 371 U.S. 38 (1962) (distributor producing sufficient quantity of tied product -- undesirable films); Northern Pac. Ry. v. United States, 356 U.S. 1 (1958) (tied product -- transportation services -- was fully installed); International Salt Co. v. United States, 332 U.S. 392 (1947) (seller was nation's leading producer for industrial use of tied product -- salt); International Bus. Mach. Corp. v. United States, 298 U.S. 131 (1936) (seller was already producing 81% of the tied product -- tabulating cards); Motion Picture Patents Co. v. Universal Film Mfg. Co., 243 U.S. 502 (1917) (holder of patented motion picture projector already producing sufficient quantity of tied product -- motion picture film).

n111 Another alternative would be to purchase the tied good from another manufacturer and resell it at a higher price along with the tying good. This both adds to transaction costs and makes monitoring, see supra note 109, more difficult. In addition, such an arrangement has not typically appeared in the leading typing cases. See supra note 110.

The argument has been made that it may often be cheaper to use a tying arrangement because, for example, service representatives from the manufacturer must visit the buyer periodically in any event, thus yielding possible economies in delivery. This argument is terribly weak in many contexts. First, if this were true, the buyer would voluntarily accept the other good since the manufacturer could afford to sell it for less than its rivals. Second, the idea that a service technician can carry a tool box under one arm and a few dozen boxcars of salt, computer cards, or whatever else under the other is rather fanciful. There may be an exception in the case of photocopy machines, if they can be trained to break down just before running out of paper and other supplies, but again there would be no need for compulsion.

n112 See, e.g., Sidak, Debunking Predatory Innovation, 83 Colum. L. Rev. 1121, 1138 (1983) ("The additional cost of designing and manufacturing the camera and cartridge in a uniquely compatible configuration surely is less than the cost of monitoring millions of amateur photographers over a number of years to determine whether they are buying only Kodak film." (footnote ommitted)).

n113 See, e.g., L. Sullivan, supra note 7, at 454 (voluntary arrangements apart from tying are superior in facilitating secret price concessions by oligopolists attempting to cheat). Bork argues that, for example, "the tying of salt to a salt-dispensing machine may be the equivalent of a requirements contract, and so lower both selling and manufacturing costs." R. Bork, supra note 14, at 379. The obvious alternative of permitting the buyer to negotiate voluntarily a requirements contract with the seller or some other producer would presumably be as efficient, or more so, to the extent some other seller of the tied product was more conveniently located or otherwise better suited to the buyer's needs. Another example is Sidak's argument that technological tie-ins facilitate risk allocation, Sidak, supra note 112, at 1135, which is justly criticized in Ordover, Sykes & Willig, Predatory Systems Rivalry: A Reply, 83 Colum. L. Rev. 1150, 1162 n.26 (1983) ("[E]ven if risk sharing is desirable, tie-ins may not be. A warranty, a contingent sale, or a leasing arangement may do a better job of risk allocation. For all of these reasons, we suspect that the use of technological tie-ins for risk sharing is quite infrequent.").

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

[*543] Finally, even where the less restrictive alternative is noted, there is a tendency to give it too little weight. For example, with regard to the well-known I.B.M. punchcard tying case,n114 Posner argued that "[i]f the purpose of imposing the tie-in really was to protect good will rather than to discriminate, the specifications alternative was presumably less efficient than the tie-in; otherwise the manufacturer would have promulgated specifications voluntarily."n115 Aside from other deficiencies in this argument,n116 this deduction in no way indicates the significance of the resulting additional cost. The facts of I.B.M. strongly indicate that any such cost would have been miniscule at best,n117 so Posner's point lacks appreciable significance in determining an appropriate rule.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n114 See International Bus. Mach. Corp. v. United States, 298 U.S. 131 (1936).

n115 R. Posner, supra note 7, at 175.

n116 If the manufacturer expected to receive any other benefit from the tying arrangement, this inference would not follow. This reasoning is further examined in Section III-C, infra.

n117 See 298 U.S. at 139.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

2. Unsupported Post Hoc Conjecture. -- Most of the proffered objectives for tying can only be realized if the firm implementing the practice is consciously pursuing them. For example, price discrimination and other more complicated package pricing schemes can work only if the prices are set in a manner dictated by the underlying economic theories.n118 Thus, in analyzing past cases, one would suspect that, if many [*544] of the alternative explanations just discussed were the true explanations for the restrictive practice, the defendants would have argued these points strongly before the court.n119 The fact is that not only did these theories receive virtually no attention in the leading Supreme Court opinions on tying,n120 but they were not even raised in the parties' [*545] briefs.n121 One explanation for this might be that such alternative theories were beyond the understanding of management at the time, which seems entirely plausible since they were not well recognized by economists until the past few decades.n122

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n118 The same can be said of the theory that the tying arrangement is designed to recoup depreciation of the leased machine, see Hansen & Roberts, Metered Tying Arrangement, Allocative Efficiency, and Price Discrimination, 47 S. Econ. J. 73 (1980).

Easterbrook, supra note 57, at 5, offers the reply that firms using various restrictive practices may simply not know how it is that these practices work, but nonetheless we can infer from their success and imitation that they indeed are efficient. First, this argument simply begs the question concerning the desirability of the practices to which it is applied since the competing hypothesis is that they successfully assist the monopolization process, which also would imply that the practices would both survive and be imitated. Second, Easterbrook ignores the consideration raised here concerning the need to understand many practices in order to implement them successfully. Some of the practices he notes -- e.g., tie-ins and block booking, see id. at 6 -- he would have us believe are complicated price discrimination schemes and thus are directly subject to the argument noted in text. Others, such as resale price maintenance and vertical territorial and customer restrictions, see id., might be seen as devices designed to alleviate free rider problems at the retail level, but that theory also assumes some understanding of the problem because not just any resale price or territorial restriction will do, and such restrictions will be against the firm's interest if used improperly. Moreover, they will be inferior to many more straightforward alternatives, such as direct manufacturer advertising, in a wide range of circumstances. Thus, if their use is really based on mitigating the free rider problem, the manufacturer must have a fairly good idea of what it is trying to accomplish. For both sets of practices, the notion that trial and error, without even an approximate understanding of the intuition behind the arrangements, would have produced the efficient results hypothesized seems rather farfetched. Finally, for arguments often noted in the tying context, such as quality control for protection of goodwill, the idea that firms were unaware of such a theory is strongly belied by the fact that it was so often alleged in defense of the practices. See infra note 121.

n119 If the defendants perceived that the court would find their alternative explanations equally reprehensiblem however, such explanations would not as likely have been offered. The prohibition against direct price discrimination, see supra p. 522, offers one possible example, although this prohibition was not strengthened until after some of the relevant court battles.

Easterbrook argues that once a practice has been declared unlawful, various efficiency defenses will not be raised because the plaintiff would be required to admit that the law had been violated. See Easterbrook, supra note 57, at 6-7. This position, with rare exception, is hopelessly unconvincing, as it ignores the availability of argument in the alternative, is belied by the fact that all the defenses suggested by Easterbrook and others have in fact been raised (and many expressly considered) in antitrust cases, and may be often rendered moot because such defenses are often not inconsistent with other positions and are often the only plausible ones because the fact that the defendant engaged in the practice is not seriously subject to dispute. The most blatant deployment of the argument is in Easterbrook's discussion of Monsanto Co. v. Spray-Rite Serv. Corp., 104 S. Ct. 1464 (1984), where "the defendant had not asked the district court to overrule the earlier Supreme Court cases, and thus the issue was foreclosed." Easterbrook, supra, at 7. He continues:

One wonders just how bold the Court expects a defendant to be. Must it concede that it engaged in prohibited acts and ask the district court to . . . disregard an opinion of the Supreme Court? If the defendant both denies that it did the prohibited thing and seeks a change in the law, it is at a substantial disadvantage. How can it argue the competitive benefits of the thing it denies doing?
Id. at 7 n.15. In addition to the above arguments, one should note that all the Supreme Court said in this case was that it "was tried on per se instructions to the jury. Neither party argued in the District Court that the rule of reason should apply to a vertical pricefixing conspiracy, nor raised the point on appeal." 104 S. Ct. 1469 n.7. All the defendant apparently had to do was challenge the jury instruction, and perhaps make an offer of proof, to preserve the issue. This would not have been heard by the jury, so it could not possibly have prejudiced the fact-finding. It should also be noted that in this resale price maintenance case, like many others, although it was under dispute whether the defendant had an enforced resale price maintenance agreement within the meaning of the law, see, e.g., United States v. Parke, Davis & Co., 362 U.S. 29 (1960); United States v. Colgate & Co., 250 U.S. 300 (1919), there was no dispute as to whether the manufacturer announced and hoped for adherence to resale prices. Any attempt by the defendant to claim that resale price maintenance was desirable would not have been the least bit inconsistent with its claim that, although it hoped there would be adherence to its announced policy, it did not engage in illegal agreements to enforce that result.

n120 These alternative explanations for restrictive practices have never commanded a majority of the Court. In fact, the notion that a tying arrangement can operate as a vehicle for price discrimination has been examined only in a single, rather old, dissenting opinion. In Henry v. A.B. Dick Co., 224 U.S. 1, 65 (1912), Chief Justice White's dissent addressed the explanation, now endorsed eagerly by commentators, that a tying arrangement can and should be best understood as a means whereby the seller of a machine can extract a royalty based on the output of a machine. During the past 73 years, the Court has had ample opportunity to embellish or simply reiterate this position, yet it has failed to do so.

n121 For example, the hypothetical explanation offered in Peterman, The International Salt Case, 22 J. Law & Econ. 351, 362-63 & n.19 (1979), concerning distribution costs would surely have been raised in the defendant's brief if it had in fact been the defendant's motivation. An examination of the brief demonstrates, however, that such a motivation was not discussed, or even mentioned. Indeed, the argument that the tied product operates as a counting device that is capable of measuring a machine's output and thus generating a royalty according to intensity of use has been profferred only twice (and both times in summary fashion) in briefs submitted by parties attempting to justify their tying arrangements. This alternative explanation for the tie-in has not appeared in a Supreme Court brief in the past 63 years. In United Shoe Mach. Co. v. United States, 258 U.S. 451 (1922), counsel for the appellant-defendant allocated one clause of a single sentence in a 1,562 page brief to the proposition that the tied products might serve as a meter to measure the use of the tying product. See Brief for Appellants, Vol. 1, at 225, United Shoe. As was adverted to previously, see supra note 120, in Henry v. A.B. Dick Co., 224 U.S. 1 (1912), counsel for the appellee-defendant similarly argued that the tied product operated as a meter capable of measuring the use of the tying product so as to facilitate the determination of an appropriate royalty. See Brief for Appellee, supra at 15-16, A.B. Dick.

Notwithstanding the lore perpetuated by commentators, intricate pricing explanations for restrictive practices, to the extent two extremely brief passages constitute a genre of explanation, have been jettisoned entirely in the ensuing six decades in favor of various other justifications: (1) The protection of goodwill -- see Brief for Appellants at 13, 34, International Salt Co. v. United States, 332 U.S. 392 (1947); Brief for Appellants at 8-16, International Bus. Mach. Corp. v. United States, 298 U.S. 131 (1936); Brief for Appellants, Vol. 1, at 221-26, United Shoe Mach. Co. v. United States, 258 U.S. 451 (1922); Brief for Appellees at 13, Henry v. A.B. Dick Co., 224 U.S. 1 (1912); (2) Financing for research and development expenditures by generating rewards for innovations beyond that obtained via a single monopoly in the tying product market (an explanation directly at odds with the fixed sum thesis) -- see Brief for Respondents at 25, 27, Carbice Corp. of Am. v. American Patents Dev. Corp., 283 U.S. 27 (1931); (3) Creating or exploiting economies of scale -- see Brief for Appellants at 49-52, Times-Picayune Publishing Co. v. United States, 345 U.S. 594 (1953); Brief for Appellees at 116, United States v. Griffith, 334 U.S. 100 (1948). (Note that these summaries of necessity are limited to the leading cases and do not cover the immediate past.)

n122 The argument began to emerge, although still without direct application to antitrust policy, in Bailey, Price and Output Determination by a Firm Selling Related Products, 44 Am. Econ. Rev. 82 (1954). Compare the discussion in Section III-B, infra.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Both these points tend to eviscerate theories attempting to criticize past decisions on the ground that they overlook what were obviously the true motives of the defendants. Moreover, this argument suggests that the most plausible explanations for the defendants' behavior were [*546] those understood and argued about at the time. Of course, it now can be argued that in the future the situation will be different since defendants will have read the past three decades of literature and now know how to be efficient in ways they had never dreamed of before. Although this point is important, there is the danger that old practices will often continue to be done for old reasons; the only change may be in the sophistication of expert testimony and briefs offered in defense.n123

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n123 This possibility is enhanced if one adopts the perspective taken on a related issue advanced by Easterbrook. "Economists have developed procompetitive explanations for all these practices, sometimes several explanations for each practice. Then, too, practices that were deleterious yesterday may yield benefits today . . . ." Easterbrook, supra note 57, at 6. Easterbrook does not lend similar weight to the possibility that practices that were not deleterious yesterday may yield harms today. Elsewhere in the article, he criticizes claims of anticompetitive potential for restrictive practices as follows: "True, the world of economic theory is full of 'existence theorems' -- proofs that under certain conditions ordinarily-beneficial practices could have undesirable consequences. But we cannot live by existence theorems." Id. at 15. Thus, lists of possible anticompetive effects are to be disregarded because they only represent mere possibilities, while existence theorems concerning procompetitive explanations are to be accepted and seen as the basis for expecting even further, as yet unrecognized, benefits. This subsection argues that a more balanced approach is in order, and that one should by wary of after-the-fact rationalization of practices that have long been used because firms thought they would achieve anticompetitive effects but now are defended using expert testimony asserting various of the procompetive existence theorems. Many of these practices may in fact have been desirable all along, but establishing that conclusion requires argument beyond the level typically offered. See Section III-C, infra.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

3. Ambiguity of Questioning the Existence of Market Power. -- A common controversy involving the rules governing restrictive practices concerns the market power that ought to be required as a threshold to liability.n124 This issue has been central because of the link between traditional leverage theory and market power: to extend monopoly from one market to another, there must be some monopoly in the first market to begin with. What has generally been overlooked, however, is that demonstrating the lack of significant monopoly power in the first market casts doubt not only on the leverage explanation, but also on many of the others as well.n125 For example, price discrimination, as well as other more complicated package pricing schemes, requires that [*547] the firm employing the devices has monopoly power, for if it did not, it would lose to its competitors those consumers charged the higher net price.n126

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n124 In reference to tying and related practices, see, e.g., R. Bork, supra note 14, at 367-68; Slawson, supra note 24, at 684-91. On the more general issue of market power requirements as thresholds to liability, see Kaplow, The Accuracy of Traditional Market Power Analysis and a Direct Adjustment Alternative, 95 Harv. L. Rev. 1817 (1982); Landes & Posner, supra note 33; Schmalensee, Another Look at Market Power, 95 Harv. L. Rev. 1789 (1982).

n125 Bowman notes that "[a] tie-in is a useless device unless the supplier possesses substantial monoply over the tying product." Bowman, supra note 12, at 31. This observation is offered in connection with Bowman's criticism of the monopoly extension view; he fails to note its implication for the plausibility of the alternative explanations he presents. Similarly, Bork does not hesitate to offer price discrimination, see R. Bork, supra note 14, at 375-78, as one of his "useful [and 'most realistic'] explanations of observed tie-ins," id. at 375-76, without noting that his earlier criticism, cited supra note 124, of the courts' failure to consider the market power issue applies to his position as well.

Absence of market power does not cast doubt on all the other explanations for tying, such as the use of tying as a quality control, which is subject to many of the arguments noted supra in subsection III-A-1.

n126 See F. Scherer, supra note 29, at 315.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

The presence of market power is highly relevant to a complete theoretical analysis of tying. It is not, however, obvious that market power should be central to the courts' inquiries in this area because the degree of extent market power may not facilitate determining which among competing explanations for observed behavior is the most plausible.n127

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n127 This criticism is applicable to Easterbrook's recently-proposed framework that would require all antitrust plaintiffs to prove market power as the first of five thresholds before consideration of antitrust liability. See Easterbrook, supra note 57, at 19-23. (His framework is discussed further supra note 57.) His claim that market power is a necessary condition for adverse consequences, id. at 21, misses this argument. In addition, since he presumably would apply the requirement even to price-fixing cases, his approach would require plaintiffs to make what is often a most difficult showing even though there would be little harm in banning the practice in any event. Avoiding such results presumably is a substantial part of the motivation for Areeda's contrary position that Easterbrook criticizes, see id. at 29 n.62, although Easterbrook fails to respond to the arguments that seem to offer strong justification for Areeda's position. Easterbrook essentially attempts to use abstract logic to resolve an issue (whether there should be a market power requirement for a given practice) that depends upon highly pragmatic, factual considerations. Moreover, Easterbrook's assumption that the lack of market power will be "obvious," id., begs the real underlying question concerning whether the practice itself justifies an alternative inference and, more importantly, whether proof should always be required when it often (some would argue, almost always) will not be clear how much market power exists.

Market power is only the most prominent example of such misleading analysis. One commentator argues that leverage is an implausible "explanation of ties between goods unrelated in demand." Texas Note, supra note 16, at 917. Of course, such ties would be ineffective in accomplishing price discrimination, pricing of complementary goods, protection of goodwill, efficiencies in package sales, and most other purported objectives of tying as well. But see Section III-B, infra (non-profit-maximizing motivations). Thus, what may be a highly complex inquiry into whether there is any demand interrelationship between the tied and tying products or services may prove to be of little use in determining the underlying causes and effects of a defendant's behavior.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

B. Alternative Managerial Motivations

The existence of such a wide array of competing explanations for a defendant's behavior directs the analysis towards inferences about managerial motivations.n128 The standard assumption of neoclassical economic analysis is that firms seek to maximize profits. This perspective [*548] does not entail a shortsighted focus on current profits but considers long-run costs and returns. This assumption is invoked so oftenn129 as to make it almost beyond question in the minds of some. Typically, the assumption is implicit, as in the argument that an exclusionary attempt is unprofitable and thus it must be that the firm has some other purpose in mind. For many purposes, this assumption serves quite well, and it would obviously be difficult to maintain the extreme opposite assumption that profits are of little or no interest to most firms. Yet strict adherence to the profit-maximization assumption may obscure much of what explains the behavior in question. This Section explores some alternative managerial motivations that may help in understanding the behavior of firms that employ restrictive practices. Since profitmaximizing behavior is itself one of the assumptions of a perfect market, this whole discussion can be seen as an important and typically overlooked instance of the market imperfections point raised earlier.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n128 If a practice clearly cannot have detrimental effects, there may be no need to worry. However, if detrimental effects are possible, but not profit maximizing, then the analysis of this Section applies with full force.

n129 See, e.g., 5 P. Areeda & D. Turner, supra note 13, at 184 (analyzing reciprocity: "it is profit rather than volume that induces a supplier to make reciprocal purchases from the defendant who will then buy from him"); P. Areeda, supra note 107, at 649 n.13 ("the manufacturer's objective is not maximum sales but maximum profit"); R. Bork, supra note 14, at 145 (predation).

Easterbrook invokes this assumption in analyzing Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 104 S. Ct. 1551 (1984), where he argues that "[i]f the hospital had possessed power, though, it would have had no reason to use its power to increase the price (or reduce the attractiveness) of its anesthesiological service. The hospital already could have extracted monopoly rents for the use of the operating room." Easterbrook, supra note 57, at 26 n.53. It is illuminating to consider whether it is appropriate to analyze the behavior of a hospital -- even a private for-profit hospital -- using a profit-maximizing model rather than one that considers more directly the welfare of various groups of doctors that might be powerful in influencing the hospital's decisionmaking process.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

1. Misperception of Self-Interest. -- On occasion firms err in their estimation of the net profitability of various actions.n130 Of course, a priori, there may be no reason to expect any particular mistake, and it is clearly difficult to develop useful models that can predict random miscalculation or stupidity by firms.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n130 More extreme formulations are also possible. See, e.g., Priest, Cartels and Patent License Arrangements, 20 J.L. & Econ. 309, 312 (1977) ("There remains, however, a more serious objection to determining the competitive effect of a firm's actions on the basis of the intent of its managers. There is no reason to believe that company officials understand the mechanism by which any particular practice or policy affects profits.").

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

It is still possible, however, that certain systematic errors might be observed that would be relevant to antitrust policy.n131 To take the simplest examplen132 one might argue that even if a leveraging strategy is [*549] unprofitable or doomed to complete failure in the long run, many firms may cling to a misguided belief that they can succeed and thus will make repeated attempts in any event.n133 Some firms might believe that leverage will build entry barriers even though their construction will cost far more than their future worth, or that the short-run sacrifices in the profits of the tying product will pay off in the long run on profits of the tied product when there is in fact little prospect of such success.n134 Such conjectures, as noted previously, are quite plausible in light of the fact that courts and commentators fully believed such arguments for decadesn135 and in many instances continue to hold such beliefs today.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n131 See R. Posner, supra note 7, at 182, 184-87 (noting, but not analyzing, the possibility of "irrational behavior"); Williamson, supra note 50, at 312 (noting possibility of mistaken attempted predation).

n132 Precisely the same argument would be relevant to a number of similar antitrust issues, such as the prospect of long-run gains through predatory pricing, the likelihood that other firms will not cheat on the cartel price, or the prospect that a horizontal merger will create an atmosphere more conducive to collusive behavior.

n133 See Allen, Vertical Integration and Market Foreclosure: The Case of Cement and Concrete, 14 J.L. & Econ. 251, 270-72 (1971) ("Vertical mergers to protect one's historic market share may have simply seemed like the correct thing to do; whether such moves were really likely to enhance a firm's profits may have been a question getting close analysis only when it was too late.").

n134 Easterbrook argues that in the predatory pricing context, no antitrust challenge should be permitted if it appears that the period following predation is unlikely to give rise to monopoly prices that are sufficiently high. See Easterbrook, supra note 57, at 37. In addition to the difficulty of proving this effect and determining the desirability of limiting enforcement of after-the-fact relief, see supra note 57, Easterbrook does not address the issues raised here concerning whether antitrust laws should be applied -- or whether great efforts should be undertaken to insure that the antitrust laws are never, even accidently applied -- to miscalculations or behavior resulting from alternative motivations.

n135 Perhaps I reflect some personal bias in doubting that judges, lawyers, and commentators are so much less intelligent and informed than business managers that the continued belief in theories by the former has no bearing on the plausibility of the existence of similar beliefs by the latter. This position is reinforced because these groups of people often associate together and have similar educations and backgrounds, because information concerning managerial motivations has often been available to the former groups when analyzing the behavior of the latter, and because, as noted supra pp. [229-31], the managers, when their companies have been defendants, have often had every incentive to inform the former groups to the extent there was a divergence in belief. See also Blake & Jones, supra note 97, at 393-94.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

Although the apparent purpose of the antitrust laws is not to improve the effectiveness of management,n136 it does not follow that the possibility of misperceptions by management should be ignored. First, it is important to understand motivations to guide inferences concerning the likely effects of restrictive practices. Second, misperceptions [*550] might lead to the same, if not even greater, anticompetitive results as restraints of trade that are in fact profit maximizing. In the examples noted previously, I hypothesized only that the behavior would not ultimately prove profitable to the firm, without denying that it might have a serious exclusionary effect. Finally, even if the long-run exclusionary effect fails to materialize, the firm's behavior may still be costly. When firms make unprofitable decisions, it is often the case that society as a whole, rather than merely the shareholders and management, suffers.n137

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n136 See, e.g., R. Bork, supra note 14, at 249 ("But antitrust does not exist as a means for federal courts to review and revise management judgments about efficiency."); Allen, supra note 133, at 274 ("Perhaps a public policy against such vertical mergers protects both competitors and competition from firms' mistakes. But it does so by protecting firms from their own mistakes. Whether this is a worthwhile aim of public policy is at least debatable." (emphasis in original)); Baxter, Legal Restrictions on Exploitation of the Patent Monopoly: An Economic Analysis, 76 Yale L.J. 267, 318 (1966) ("no justification . . . for using the antitrust laws to assure that private economic interests are correctly perceived"); cf. R. Posner, supra note 7, at 205 ("implications for policy are unclear"); Easterbrook, supra note 84, at 279 ("It would . . . be legitimate to object that [social loss from unprofitable predation] has nothing to do with antitrust law, which cannot address all welfare reducing activities.").

n137 In a simple model of perfect competition, this conclusion would always be true. Since the relevant application to leverage assumes that there is some monopoly power to begin with, some actions that diminish profits, such as accidently charging a competitive price for a few decades, are obviously socially beneficial. But many, such as the use of resources to drive other firms out of the market or to erect entry barriers, typically will impose net social costs even if they do not accomplish their purpose. See, e.g., R. Posner, supra note 7, at 187-88 (predatory pricing).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

One might ask whether such problems should be addressed by the antitrust laws rather than some other, perhaps not-yet-existent body of law. Resolving this issue is beyond the scope of this Article, but a few observations are in order. First, if one is worried about various undesirable effects stemming from a single practice, it would often be sensible to deal with them by using the same regulatory apparatus. Second, to the extent one has trouble determining which of many possible motivations can explain and which effects will result from the observed behavior, it would seem incongruous to permit a restrictive practice under the antitrust laws because, rather than inevitably resulting in anticompetitive effects, it sometimes results in other undesirable effects instead.n138

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n138 See Kaplow, supra note 20, at 1857-58.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

2. Sales and Growth Maximization. -- In the past few decades, there has been increasing development, at both the theoretical and empirical levels, of the view that many large firms can be better understood as maximizing their total output, sales, or growth rates or, more generally, fulfilling the management's interests rather than those of shareholders.n139 These theories suggest that firms would be willing to spend [*551] more on various activities -- for example, price cutting, advertising, and plant expansion -- than a simple profit-maximization calculus would dictate. The reason most commonly offered is that management, not wholly controlled by shareholders, may benefit by trading off size for profits.n140 This theory has striking implications for much of antitrust.n141 For example, virtually all of the practices connected with the leverage issue operate to increase the sales or growth of the firm. Comments concerning tying often note the immediate effect in shifting sales of the tied product from other firms to the tying firm. Long-run growth in the market for the tied product, and entry barriers, as in the case of tying, vertical mergers, and a number of other restrictions, all concern future growth in the firm's sales. Thus, to the extent firms depart from profit maximization in the manner here described, one might observe widespread use of the restrictive practices discussed in this Article regardless of whether much of the leverage theory discussed in Section II-B or some of the competing explanations discussed in Section III-A are operative.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n139 See generally W. Baumol, Business Behavior, Value and Growth (rev. ed. 1967) (arguing that sales and growth maximization may better explain the motivations of many oligopolists than does profit maximization); R. Marris, The Economic Theory of "Managerial" Capitalism 46-61 (1964) (discussing noneconomic motives beyond profit maximization); O. Williamson, The Economics of Discretionary Behavior: Managerial Objectives in a Theory of the Firm (1964); Amihud & Kamin, Revenue vs. Profit-Maximization: Differences in Behavior by the Type of Control and by Market Power, 45 S. Econ. J. 838 (1979) (revenue maximization more prevalent in oligopolistic, management-controlled firms); Bailey & Boyle, Sales Revenue Maximization: An Empirical Vindication, 5 Indus. Org. Rev. 46 (1977) (sales maximization is a serious motivation for corporate managers in 70% of the firms analyzed); Hirschey & Werden, An Empirical Analysis of Managerial Incentives, 7 Indus. Org. Rev. 66 (1980) (empirical evidence of dual profit and sales incentives for managers of large industrial corporations). This theory has continued to be the subject of extended debate, see, e.g., F. Scherer, supra note 29, at 34-41, most of which will not be considered here. First, it is far beyond the scope of this Article to attempt a definitive resolution of that controversy. Suffice it to say that this view surely has enough plausibility to warrant far more serious consideration in antitrust scholarship. Second, one of the most serious criticisms offered to rebut this position concerns the ability of firms to survive if they behave in this manner, and the argument to follow in text provides some reasons why survival would be plausible in the present context.

n140 See Jensen & Meckling, Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, 3 J. Fin. Econ. 305 (1976). A similar example is management's potential gain from diversification -- e.g., through a conglomerate merger -- that produces no corresponding benefits to shareholders. See, e.g., Amihud & Lev, Risk Reduction as a Managerial Motive for Conglomerate Mergers, 12 Bell J. Econ. 605 (1981).

n141 This theory has been noted in this context before, although without analysis of its specific application or relevance to antitrust policy. See Blake & Jones, supra note 10, at 460-61. This branch of economic analysis was not nearly as developed when Blake and Jones wrote two decades ago. More recently, the connection has been noted by Porter, supra note 96, at 482-83 (footnote omitted):

Interacting with these considerations is the separation between ownership and control. The essence of the separation is that managers do not perceive their personal interests to be coincident with maximizing the long-run value of the firm. This can be because of bankruptcy fears, monetary incentives based on short-run profitability, criteria for promotion that often stress short-run performance, and other failures of reward systems that stem from imperfect information. Separation between ownership and control also allows other forms of managerial utility maximization, such as pursuit of status, exit barriers due to emotional attachments, and the like. Finally, separation of ownership and control, coupled with various transactions and information costs, also gives room for differences among companies in the decisionmaking power and authority of different functional departments or individual executives. The degree of separation between ownership and control and its internal consequences can and does vary among firms in a given industry, with the result that competitors can differ sharply in their motivations.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

One of the most common criticisms of these other theories is that [*552] they assume that the firm in question has some excess profits, or "slack;" otherwise, if the firm attempted to maximize sales at the expense of profits it would soon find itself out of business.n142 Even if one though that such slack was not present in most large firms, the issue here concerns the applicability of these theories to the situation in which slack is likely to be present since the defendant typically possesses at least some monopoly power.n143

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n142 See, e.g., F. Scherer, supra note 29, at 38.

n143 The firm need not even have excessive profits generated in connection with one of the products affected by the restrictive practice. It is enough that it have some excess from other areas of its activity, and that it believes the payoff in terms of sales expansion to be sufficiently great in the area to which it is applying the restriction. Such restrictions often will offer substantial sales at little cost, as noted in subsection II-B-1, supra.

This argument contrasts strongly with argument within the pure profit maximization paradigm where, as numerous commentators have asserted, see, e.g., 5 P. Areeda & D. Turner, supra note 13, at 182; R. Bork, supra note 14, at 144-45, the fact that a firm has excess profits in one market or product does not necessarily imply that they will be used to expand sales in another.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

As with the case of possible managerial miscalculation, alternative managerial motivations ought to be taken seriously. First, analyzing them is necessary if one hopes to understand and infer the likely effects of firm behavior.n144 Second, the implications of alternative motivations are quite consistent with, and perhaps even more suggestive than miscalculation, of the possibility that traditional anticompetitive effects result. Finally, even if antitrust were not designed to rectify such managerial behavior, it may nevertheless be sensible to prohibit practices adopted for the reasons discussed here, just as was the case with miscalculation.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n144 For example, two attempts at a comprehensive analysis of tying arrangements in well-known Supreme Court cases analyze the application of a wide array of possible motivations without even considering this rather straight forward explanation. See Cummings & Ruhter, The Northern Pacific Case, 22 J.L. & Econ. 329 (1979); Peterman, supra note 121 (International Salt).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

A more sophisticated analysis of what motivates firm behavior thus is of great importance to understanding much of antitrust. This Article only hints at some of the implications that would follow from a greater appreciation and application of some of these insights from economic analysis of the past few decades. Of all the issues examined in this Article, this one in particular assumes an importance quite large relative to the attention it has received. As a result, it offers one of the most fruitful paths for further inquiry.

C. Choosing Among Competing Hypotheses

Given the vast uncertainty concerning firm behavior, commentators on all sides of the issue tend subtly to manipulate available knowledge to yield spurious conclusions. The process often takes the form of argument by process of elimination in which a handful of competing [*553] motivations for a firm's behavior are presented and discussed. After all but the last possible motivation has been analyzed and shown to be uncertain, implausible, and highly contingent on unusual fact patterns, it is common to hear the proclamation that the final explanation (the author's favorite) must therefore constitute truth. The problem is that this ultimately victorious theory has not been subjected to quite the same scrutiny as were the theories previously eliminated.n145

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n145 This manipulation of the reader is possible for two reasons. First, each of the theories is so complicated and subject to so many reservations that it is possible both to refute (or confuse) any single theory ad nauseam and at the same time to make a powerful case for whichever theory the author favors, while holding back, in both tone and substance, a portion of the criticism of that position. Second, the author's choice of organization can exercise additional influence. As indicated, the process of elimination technique seems prominent. By the end of the discussion the reader hopes for something that works; the author then supplies the psychic gratification in the form of a plausible resolution to a seemingly insoluble problem. Moreover, by saving the affirmative argument for the end, the applicability to the author's position of criticism applied to earlier arguments may be well be overlooked. One such instance is the earlier discussion, supra pp. [232-33], of how the requirement of monopoly power for leverage to be successful tends to remain unmentioned when offering price discrimination as a competing hypothesis.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

This pattern of argument, which is quite plausibly unintentional, has been employed by advocates on all sides of the issue. For example, Turner argues as follows:

If in a described category of cases the tie-in serves no useful function, or if any useful function can be fulfilled in a large majority of instances by less restrictive devices, it is a reasonable assumption that the purpose of the seller in using a tie-in is to restrain competition in the tied product and that he has the power over the tying product which his purpose implies he has.n146

Bork takes the mirror image of this position:

[The firm] must offer something to the food canners to get them to sign requirements contracts, and it must offer that something for the life of the contract, which means that, in terms of cutting out rivals, the contract offers [the firm] no advantages it would not have had without the contract. The advantage of the contract must be the creation of efficiency. . . . In this situation, efficiencies are the reality, and the fear of foreclosure is chimerical.n147

[*554] Many others have made similar arguments.n148 To avoid this frequent problem, it is necessary to be more aware of the structure of argument leading to any conclusion, and to be wary of conclusions suggesting that one or a few simple explanations dominate all others.

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n146 Turner, supra note 24, at 62 (emphasis in original).

n147 R. Bork, supra note 14, at 304-05 (emphasis added). He argues somewhat similarly in an earlier article while discussing Brown Shoe: "Given a fragmented manufacturing industry such as that in Brown Shoe, the desire for monopoly profits could not inspire a merger trend. In such circumstances, therefore, the existence of the trend must be taken as prima facie evidence that greater concentration is socially desirable." Bork, Contrasts in Antitrust Theory: I, 65 Colum. L. REV. 401, 412 (1965). One could also develop the reasons why the integration would not be efficient -- for example, because if it were, internal expansion would be the preferred route -- and then argue that the actual motive must have been the desire for monopoly profits. See Blake & Jones, supra note 10, at 459 n.141 (criticizing along these lines Bork & Bowman, The Crisis in Antitrust, 65 Colum. L. REV. 363 (1965)); see also id. at 461 (same).

n148 Posner, for example, argues that the less restrictive alternative to price discrimination "was presumably less efficient than the tie-in; otherwise the manufacturer would have promulgated specifications voluntarily." R. Posner, supra note 7, at 174 (emphasis added). He then adds that the balance favors permitting the practice since the arrangement "is unlikely to have any competitive significance in the market for the tied product." Id. at 176. One could just as easily note the insignificance of any possible efficiency benefit, see supra p. 543, and argue that the balance must therefore come out the other way. Posner's treatment of the entry deterrence motive in this example is similar. See id. at 176; see also id. at 204 (after extensive criticism, rejecting the exclusionary motive for United Shoe's leasing practices in favor of a nonexclusionary motive that was mentioned more briefly and subject to no scrutiny).

Burstein deserves praise for the open-mindedness of his early treatment of the leverage issue. See Burstein, supra note 17, at 93-95 (concluding section). For example, he notes:

[I]t is important to stress the limitations of the method of science. To assert that a given model (say my own) is consistent with most of the data is not to assert that other models are consistent with none of the data. . . . [T]here remain cases that are consistent with previous theories, including that of extension-of-monopoly, and it is entirely possible that there are important and numerous classes of cases for which the latter model works better.

Id. at 94; see also Blake & Jones, supra note 10, at 459 (emphasizing mixed motivations rather than presuming efficiency or anticompetitive effects).

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -

This criticism might be applied to this Article as follows: The lengthy discussion of leverage in Part II, although hopefully succeeding in resuscitating the plausibility of the theory, may not have left one with the view that such leveraging as described could explain the wide use of some restrictive practices that is observed. The discussion in Section III-A leaves the reader highly dubious of all the other theories that have been put forth to explain why firms engage in restrictive practices. Finally, we are offered the author's explanation in Section III-B concerning alternative theories of managerial motivation. That theory, we observe, was not scrutinized nearly as strictly as were the others.

I hope to avoid this criticism by noting the wide range of qualifications throughout my argument as well as by framing the final argument explicitly as one that has yet to be examined seriously and is worth attention, rather than as the actual, final, true explanation of tying cases or antitrust generally. When the range of potential explanations and factors is so vast and complex, an overemphasis on the need to reach immediate, certain, and feasible solutions can seriously distort the analytical process. This is especially problematic when the analyst is an advocate for a particular outcome or a general viewpoint. A careful choice among the competing hypotheses thus requires both the synthesis [*555] of a substantial body of work and a willingness to recognize that even the best of existing studies leaves many questions unanswered.

CONCLUSION

This Article has developed a number of propositions central to the continuing debate over the leverage issue. As a theoretical matter, the profit maximization/monopoly extension distinction is misleading both as a basis for guiding policy decisions and as a method for analyzing the effects of restrictive practices. A more critical analysis of the potential for leverage shows that, when often implicit simplifications are relaxed, monopoly extension even as traditionally defined is possible. It is worth noting that the frequently omitted elements of the analysis have two common characteristics. First, the omissions correspond to the assumption that unregulated makets operate in conformity with simple textbook models, reflecting a faith consistent with a general hostility to antitrust. Second, the fact that all these points have long been understood intuitively by businesses, although perhaps only more recently bolstered by sophisticated economic analysis, suggests that conventional wisdom derived through experience, unrigorous as it might be, should not be rejected so quickly on the basis of elementary reasoning.

Those prepared to dismiss traditional explanations have frequently exhibited little reluctance in offering alternative explanations of firms' motivations, even when those theories were apparently unknown to the firms themselves. This cannot be a decisive ground for rejecting these alternatives altogether. However, the common pattern of offering hypothetical situations where such theories would apply does little to establish the validity of these claims when a variety of competing accounts are similarly plausible. The process of making definitive inferences is further complicated by the possibility that management decisionmaking is subject to systematic misperceptions of the effects of restrictive practices or is motivated by goals that diverge from the traditional assumption that firms seek only profits.

Taking all these issues together, it may come as a surprise that both courts and commentators express such confidence not only in explaining behavior in particular instances, but in making broad generalizations claiming to characterize behavior in large classes of cases. At least three explanations are possible. The first is carelessness, mistake, or insufficient sophistication in examining truly complex problems. Second, and more cynically, various courts and commentators on both sides of the question may be motivated more by advancing certain positions than by understanding the complexity before them. Finally, having accepted that market behavior is sufficiently problematic to reject faith in a model of perfect competition that would justify abolition of the antitrust laws, it may be that all are reluctant to admit that market behavior is also far too intricate to succumb to the model of simple [*556] limited intervention that antitrust has represented from the outset.n149

- - - - - - - - - - - - - - - - - -Footnotes- - - - - - - - - - - - - - - - - -

n149 This issue is well illustrated by Easterbrook's remark that "[t]he fundamental premise of antitrust is the ability of competitive markets to drive firms toward efficient operation. The entire corpus of antitrust doctrine is based on the belief that markets do better than judges or regulators in rewarding practices that create economic benefit and penalizing others." Easterbrook, supra note 57, at 24 (emphasis in original). As discussed in subsection II-B-4, it is well understood that quite the opposite is the case. The belief stated by Easterbrook justifies abolition of antitrust law; the "fundamental premise of antitrust" is the inability of competitive markets to operate in the manner described without assistance from courts. Yet there is an important grain of truth in Easterbrook's, statement, which may in fact be more what he had in mind. That is, antitrust does assume that competitive markets for the most part operate as Easterbrook believes they do. Antitrust seems to assume both that intervention is only occasionally necessary (relative to the number of firms, industries, and practices) and that necessary intervention generally can be limited to one-shot remedies and simple injunctions sufficient to establish a properly functioning market. If this were not believed, one might have expected to observe direct regulation or nationalization rather than antitrust.

It is necessary to distinguish what is believed and what is hoped, although the second can strongly influence the first. Certainly at the time the antitrust laws were enacted, operations of the economy were not sufficiently understood to form a wellgrounded view concerning what was either necessary or sufficient to achieve success. Moreover, this belief is cast into substantial doubt by the widespread existence of direct regulation, more extensive common law liability, and intrusive antitrust decrees (e.g., ASCAP, the motion picture industry, AT&T) combined with the court's frequently-expressed desire to do more -- cabined largely by recognized limitations on their ability to succeed. Finally, to the extent one is concerned about issues beyond economic efficiency, see, e.g., supra notes 10 and 32, one might also question whether very limited intervention is sufficient.

Thus, the concept of antitrust as limited intervention is both a hope and to an important extent a reality despite numerous exceptions, although not necessarily a belief -- well founded or otherwise -- concerning the actual operation of the economy. Against this background, much of the confusion concerning the leverage controversy can be understood as resulting from attempts to force the analysis of complex behavior into models consistent with aspirations concerning antitrust that may be inconsistent with reality.

- - - - - - - - - - - - - - - - -End Footnotes- - - - - - - - - - - - - - - - -