How many cheers for antitrust's 100 years?

Demsetz, Harold

This article describes the ambiguity inherent is U.S. antitrust policy, arguing that it is a necessary consequence of the true, but not commonly understood, task of antitrust policy. Competition is multidimensional in form, and its different dimensions cannot be maximized together. Therefore, antitrust policy cannot maximize competition per se, but aims to achieve an efficient mix of competitive forms. Inadequate knowledge of the technical and preference tradeoffs involved guarantees that questions about the appropriate competitive mix will remain open to debate. The resulting policies, in the author's opinion, merit one cheer out of a possible three.

I. INTRODUCTION

The objective of our antitrust experiment is frequently taken to be the promotion of competition or the reduction of market power. This cannot literally be antitrust's goal. It presumes that competition is single dimensional or, if multidimensional, that it can be increased on all fronts simultaneously. The presumption is wrong. The intensity or effectiveness of one form of competition often correlates negatively with the strength of other forms. This means that maximization of competition is a meaningless goal. The goal is more correctly described as choosing a preferred mixture of competitive forms.

Thus, price competition between existing goods can be intensified by eliminating patent and copyright protection, but this reduces the effectiveness of competition to produce new goods. Similarly, competition between firms can be made more effective if competition between persons within firms, as between partners, is suppressed. The trade-off between different forms of competition is in fact recognized in antitrust by the distinction Judge Taft sought to draw in Addyston between ancillary restraints of competition and restraints whose main objective is to eliminate competition. The problem is recognized latter by Judge White when he attempted to forge a rule of reason in Standard Oil. The generality of the problem and its importance to a much broader range of antitrust issues is, however, not made explicit in the context of these cases. One form of competition is necessarily substituted for another, and prioritizing forms of competition by nomenclature such as "ancillary" simply diverts attention from the ever-present and quite general trade-offs that are involved.

Recognizing these often negative correlations between different dimensions of competition, I take as the goal of antitrust the achievement of an efficient mixture of competitive forms or its mirror image an efficient mixture of monopolistic forms. This is quite different from the goal of increasing competition or reducing monopoly, which, to be sensible, requires the absence of a general trade-off problem. It is in reference to the efficient mixture criterion that I judge the past 100 years of antitrust.

The efficiency of a mixture of competitive forms reflects the technically possible trade-offs between types of competition and the preference weights given to these types. There is no explicit market in which the preference weights for forms of competition are revealed, and the technical trade-offs between the various forms of competition, in principle determinable objectively, are so unknown that gross guess work must be used to judge them. The fact is, we do not possess enough general knowledge and understanding of the competitive trade-off problem to resolve differences in our evaluations of antitrust policy. Perhaps this is why there has been no systematic empirical study and appraisal of the efficacy of our antitrust policy. Unfortunately, evaluation remains a highly subjective task. That is why the debate over antitrust is never ending.

The problem is revealed in the Sherman Act itself. We are not dealing with the equivalent of a speed limit law. We have here something more like RICO. The Act did not legislate prohibitions that can be known with fair certainty to have been breached. The Sherman Act offers no clarifying discussion, even though monopoly and monopolization are not concepts with obviously clear meanings. Subsequent legislation failed to improve matters much. President Wilson, in preparation for passage of the Clayton Act, appointed a committee precisely for the purpose of recommending item by item devices by which competition might be eliminated and monopoly obtained. The committee found the task impossible, and cloaked its failure in simple, but nonetheless ambiguous, opposition to price discrimination.

This vagueness raises serious questions about the constitutionality of our antitrust laws. The Fifth Amendment's due process clause forbids Congress from enacting criminal statutes so vague as to disarm people of the ability to know when the law is being violated. Yet, virtually all of our century-long experience with antitrust can be characterized as judicial testing of alternative interpretations of monopoly and monopolizing. The constitutionality issue has come before our courts in a few instances. There, antitrust legislation has been judged to be sufficiently clear as to the actions and situations it prohibits as to satisfy constitutional requirements.

Usually, however, the issue before the court is not the vagueness of the notions of monopoly and monopolizing, as these are stated and barred by the Sherman Act, but a more specific practice. In Nash v. United States [1913], the Court upheld the constitutionality of the Sherman Act on the grounds that it criminalized only conspiracies to restrain trade. A conspiracy to reduce output and/or rig price is deemed by the court to be a sufficiently knowable violation of what the Sherman Act bars as to undermine the claim that the Act is unconstitutional for reasons of vagueness. However, monopoly and monopolizing have been read into business practices that have little connection to conspiracy. Even the existence of a high degree of market concentration has been deemed to violate the Act. Moreover, there is more reason than might be supposed to deem that price agreements serve socially useful, as well as socially harmful, purposes, but more of this later. Here I simply wish to point out that conspiracy, presumably a knowable criminal act, is not the entire content of monopoly and monopolization as the courts have interpreted these.

A similar point is raised by United States v. National Dairy Products Corp. [1963]. National Dairy was charged with violating paragraph 3 of the Robinson-Patman Act. This makes it a crime to sell goods at "unreasonably low prices for the purpose of destroying competition or eliminating a competitor." The Court deemed this sufficiently precise so that a defendant selling below cost, not to meet the price of a competitor, to dispose of distressed goods, or to engage in other varieties of socially useful competition, could anticipate easily enough that a jury might well find him guilty should he sell below cost. I shall argue later that predation reveals no way to use a cost standard to separate legitimate or socially useful competition from predation. Here I merely wish to point out that this section of Robinson-Patman has little to do with the Sherman Act's prohibitions and criminalization of monopoly and monopolization, and that the degree of vagueness it offers is considerably greater than for price-setting conspiracies.

The constitutionality of the Sherman Act, however, is not a central topic of this paper, but the problem of vagueness, it seems to me, cannot be solved by appeal to economics. If there is no state of real affairs describable as perfectly competitive, there is also no state of affairs free of monopoly elements. Indeed, the private property right system that underlies our economy does its work precisely by raising barriers to forms and intensities of competition thought to be undesirable. The goal of the property right system, it seems to me, is precisely that set out above for antitrust. Specifying barriers to various forms of competition is the same problem as specifying a mixture of allowed forms of competition, and its solution, in both versions, may be guided by the standard of efficiency.[1]

The perfect competition model used by economists to set forth an idealized concept of competition is not of much use in guiding us to the preferred mixture of competitive forms that is a meaningful goal for antitrust. Instead, the very notion of perfect competition suggests that maximizing the degree of competition is a meaningful goal. Moreover, the purpose of the model is quite different from that of breathing useful life into the concept of competition. Its true function is to understand decentralization, not competition. It is more aptly named the perfect &centralization model because it deprives authority of any role in the allocative process.

In a debate like that which engaged Adam Smith and mercantilists, and which now concerns the disintegrating communist world, the perfect competition model is a powerful tool for demonstrating the rationality of allocative outcomes in the absence of central planning. It is not very useful in a debate about the efficacy of antitrust precedent. It ignores technological competition by taking technology as given. It neglects competition by size of firm by assuming that the atomistically sized firm is the efficiently sized firm. It offers no productive role for reputational competition because it assumes full knowledge of prices and goods, and it ignores competition to change demands by taking tastes as given and fully known. Its informational and homogeneity assumptions leave no room for firms to compete by being different from other firms. Within its narrow confines, the model examines the consequences of only one type of competition, price competition between known, identical goods produced with full awareness of all technologies. This is an important conceptual form of competition, and when focusing on it alone we may speak sensibly about maximizing the intensity of competition. Yet, this narrowness makes the model a poor source of standards for antitrust policy.

II. SUMMARY CONCLUSION

The consequence of all this is that my evaluation of the past century of antitrust is, and must be, highly subjective. In formulating my evaluation, I hew to the legal standard applied in criminal law, that the public interest is served by presuming innocence unless the court finds generally accepted and appropriate theory and/or substantial fact by which to find a defendant guilty. Highly speculative belief about behavior or its consequences does not satisfy such a standard, even when endorsed by expert economic witness. Although many antitrust cases involve civil and not criminal law, so that a preponderance of evidence rather than an innocent until proven guilty standard may be thought more appropriate, the latter seems close in spirit to that adopted in Matsushita v. Zenith [1986], in which the Supreme Court held that "conduct that is as consistent with permissible competition as with illegal conspiracy does not, without more, support even an inference of conspiracy." When judged by this standard and the goal of forging an efficient mixture of competitive forms, the record of the last 100 years has parts that meet this standard and parts that do not. The last fifteen years of antitrust has, in general been better than the thirty years from 1940 to 1970. On the entire record of antitrust over this period, I give one cheer out of three possible.

This is a more positive evaluation than one ! gave several years ago. In 1973, at a watershed conference on market concentration, Phil Neal, then Dean of the University of Chicago Law School, asked during a question-and-answer session whether I would repeal the Sherman Act. I replied "As it is presently being carried out, yes.''2 I have become somewhat more positive about antitrust because beneficial changes have taken place during the last fifteen years. A few of the more important changes can be described in summary fashion. The dangerous standard that seemed to be established in the 1945 Alcoa case, and which threatened to prevail for several years--that which made high levels of market concentration virtually a per se violation of the Sherman Act--has been abandoned. The ridiculous extreme reached in Brown [1962] in the antitrust treatment of mergers now appears to be just that. Territorial restraints after Sylvania [1977] no longer seem to be governed by the per se doctrine. And, in the just decided matter of Atlantic Richfield Co. v. USA Petroleum Co. [1990], the Supreme Court has refused to affix a predation label to the low prices set by ARCO in its aggressive marketing of gasoline in California.

These last fifteen years of antitrust have been guided to an uncommon extent by common sense. Were this to continue, my expectations about antitrust's future would be noticeably more positive than my evaluation of its past. However, given the ups and downs of antitrust history, it would be presumptuous to forecast a rosy future on the basis of the past fifteen years. The longer history that underlies my present evaluation of antitrust is impossible to discuss in detail. Yet, some discussion is required if the reasons for my one cheer out of a possible three judgment are to be understood. A few substantive areas of antitrust may be discussed in more detail.

III. PRICE AGREEMENTS

There is no doubt that the authors of the Sherman Act intended to attack the private cartelization of industry, and, indeed, private cartelization has been severely hampered by antitrust if it has not been completely eliminated. The consequences of this have been twofold. The direct impact has been to make the economy more price competitive. The indirect impact has been to encourage firms to substitute other methods of controlling price, and, quite possibly, to substitute increased price competition for other forms of competition. The indirect impact has raised production cost and reduced product-service competition relative to what these would otherwise be, but it seems to me that the negative consequences are overwhelmed by the beneficial increase in price competition.

Cartel agreements allow cooperation between would-be rivals without forcing them to undertake serious and lasting changes in the internal organizations of their firms. Separate identities are preserved as is control over production methods (even though some minor compromises in these may be required to obtain cooperation). Neither the adoption nor the rescinding of a price agreement seriously undermines organizational integrity or impairs the internal efficiency of the cooperating firms. (The agreement, of course, may misallocate output between firms.) Cartel agreements therefore offer an appealing way for firms to restrict output and raise price, and they are likely to be used just for this purpose, rather than for other efficiency-serving purposes also. Therefore there is a theoretical base for expecting that many price agreements will make the mix of competitive forms less efficient.

Moreover, the court is in a good position to have fairly objective evidence available to it if there is an operative price agreement. The placing of an antitrust burden on explicit price agreements seems appropriate to me. The factual problem becomes more difficult if firms rely on conscious parallelism. The court is then forced to take a more speculative approach when reaching its decision, and this, it seems to me, reduces the probability that guilty findings will be beneficial.

To find merit in antitrust attacks on price agreements is not to defend the type of attack that has emerged. Price agreements are held to be per se violations of our antitrust laws. The per se policy has certain practical advantages, primarily in reducing legal uncertainty and enforcement cost. However, I do not believe these advantages outweigh the disadvantages of inflexibility of response. Efficiency justifications for price agreements, while less likely than for other business practices, are possible. These remain beyond the pale of legitimacy under the per se doctrine.

The economic perspective toward price agreements is dominated by inferences drawn from the perfect competition and monopoly models. The contrast between these models seems to justify the per se doctrine. These models, by focusing exclusively on output rate as the source of cost variations, largely ignore and divert attention from information, free-rider, and opportunism problems. These also affect cost. Recognition of them weakens the justification for the per se rule. When these sources of cost variation are important, price agreements can bring about an improvement in the mixture of competitive forms. A few examples reveal the need for flexibility.

The major exception to the per se doctrine is Appalachian Coals [1933]. This case is viewed by antitrust scholars as an exception, not to be treated as a serious challenge to the per se doctrine. Yet it offers an interesting, unexplored example of how a price agreement in the form of a joint sales agency might improve matters.

Coal is purchased by size. When servicing an order for a particular size, a coal producer automatically creates other sizes of coal. These by-product sizes are placed on the market for what price they can fetch because the cost of inventorying coal is asserted to be too high. If the price obtained is less than the cost of transporting and marketing the coal, a loss is incurred. This loss becomes simply a cost of producing the ordered coal size.

We have here a joint supply problem. It gives rise to costs that can be reduced through the use of a joint sales agency. If customer A orders large size coal from company X, then the smaller sizes automatically produced by X should be the source of coal for those who demand small nuggets. However, if a demander of small nuggets places his order with a second coal company instead of with X, an unordered quantity of large nuggets is automatically produced. Coordinating orders, say by channeling both to company X, can reduce the redundancy in the supply system. A sales agency accomplishes this, and it may accomplish it more cheaply than other institutional arrangements. One such alternative would be an intervening layer of coal brokers. These would receive customer orders and channel them to appropriate producers. However, if each coal producer has his own price schedule the broker's problem in serving customers expeditiously becomes quite complex. Lower prices might be offered by producers who plan to take longer to deliver. The broker's task becomes more simple if a single price schedule attaches to delivery commitment. But, then, we have the essentials of a sales agency.

The question is whether the productive efficiency offered by a sales agency is more than offset by its ability to raise price. In Appalachian, the court recognized that the joint sales agency provided only a small fraction of the coal supplied in its geographically relevant market, and it held that this deprived the agency of any control over market price. The court concluded that the agency did not violate our antitrust laws even though a price agreement clearly was involved. This seems to me to be a defensible decision and one that should not be treated as a depressionrounded exception to antitrust precedent.3

In Arizona v. Maricopa County [1982], the setting of maximum prices for medical procedures by physician groups helped insurers reduce uncertainty about potential claims, thereby creating a productive efficiency presumably of benefit to insurance purchasers. The agreement allowed for the specification of prices by which to determine maximum insurer liability but it did not bar doctors from charging either more or less. Under the per se doctrine the agreement was held illegal, and no serious thought was given to whether the price agreements might have served efficiency purposes without seriously compromising price competition.

That price agreements can yield efficiency is also shown in emerging conflicts over franchise-type operations. If we suppose that customers seek standardization, and that price is in some instances a relevant aspect of standardization as well as a device by which to mitigate opportunism problems, then a franchiser who requires franchisees to set a uniform price is serving his customers. Alternatively, the franchise operation can be viewed simply as a guise for rivals to jointly set price. The issue turns on how firm-like the franchise is and on what share of the market it possesses. The more firm-like it is and the smaller the share of the market it possesses, the more likely it is that uniform prices yield efficiency. In Sealy [1967], the court took the strange position that Sealy's non-excessive market share would have protected it from plaintiffs charges had it been a single firm. As a single firm there would have been no price agreement and no application of the per se doctrine. Given that it was a loosely knit franchise, the same non-excessive market share provided no protection from the per se rule.

These examples reveal that the productive efficiency of price agreements is more than mere abstract speculation. Price agreements that allow production to take place at lower cost, thus enhancing cost competition, may yield sufficient benefits to offset the reduction in price competition that may also result.4 The preferred mixture of competitive forms is unlikely to be that which gives 100 percent weight to price competition and zero weight to cost competition. The trade-off easily could be encountered enough to warrant sacrificing the practical advantages of treating all price agreements as per se violations of our antitrust laws. If the case can be made that there is an efficiency yield from a price agreement and if the aggregated market share controlled by the firms (or franchisees) making this agreement is too small to allow it to influence price significantly, its legality ought to be at least a possibility.

This would seem to require the softening of the per se prohibition of price agreements. Although I do not believe that the per se treatment of price agreements is worse than leaving these agreements untouched, better results might be obtained by introducing a somewhat more flexible doctrine. This depends on the number of exceptions that would arise if the flexible doctrine were to be applied correctly. Very few exceptions might make flexibility not worth its price in terms of practicality. Moreover, a flexible doctrine might be subject to abuse in its application. These considerations might weigh against a flexibility doctrine even though we recognize that some price agreements improve efficiency.

IV. MARKET STRUCTURE

Market concentration began its antitrust life as an auxiliary aspect in price agreement cases (Addyston in 1898 and Trenton in 1927) and predatory behavior cases (Standard Oil and American Tobacco, both in 1911). It became a central consideration in U.S. Steel [1920], where it was held that mere size of firm was no offense against our antitrust laws. In Alcoa [1945], concentration took on a transcendent importance. Finally, in cases like Berkey Photo [1979], its diminished importance put it back to auxiliary status. The various roles played by market concentration reveal the difficulty it poses as a criterion for antitrust courts, for it indexes one of the most obvious trade-offs between forms of competition.

The more concentrated is a market, the higher is the probability of successful collusion. As a general theoretical proposition this is probably true. However, it is unlikely that markets have become concentrated in order to make collusion easier. Concentration sought purely to control price is a dangerous and probably seldom played game. Lasting changes in a firm are wrought through growth in its size. This means that disaster is courted by a firm if its growth cannot be defended on the basis of lower cost or superior products. In this respect concentration differs from price agreements, which, when entered into, do not require colluding firms to become internally inefficient. High levels of concentration are likely to result from successful, and legitimate, cost and product breakthroughs, made by some firms in an industry but not by others, or from scale economies strong enough to tolerate only a few large firms. A high level of market concentration is probably a temporary state of affairs if not supported by such underlying causes.

Hence, an attack on market concentration probably is also an attack on those who have taken the risks of investing to obtain scale economies and/or make product and cost breakthroughs. Statistical evidence supports this belief.5 In cases in which market share has provided the dominant reason for deciding against a defendant, as in Alcoa, antitrust arguably has reduced efficiency by giving too little weight to competitive product and cost accomplishments and too much to price competition.

Market concentration ought to influence but not dominate antitrust decisions. When it appears coupled to price agreements, it ought to influence decisions more heavily. Moreover, both concentration and price agreements are in principle determinable by factual evidence, although disputes arising over market definition and unwritten agreements surely must arise. Reliance on speculative imagination is therefore reduced.

Should there be a distinction between concentration arrived at through internal growth and concentration arrived at through merger? The question is difficult to answer. What seems relevant to me is whether merger is more like internal growth or more like a price agreement, and the criterion for determining this is whether the merger puts the merging firms at serious risk if it is not based on cost justifications. A merger of furniture makers into a loosely connected franchise, with each producer remaining in control of his own plant and permitted to produce his own private label furniture, and in which central control is exercised with respect to the price of franchise label furniture but not with respect to cost or product quality, exhibits the characteristic of a price agreement in that it can be reversed fairly easily if it fails to be justified by cost conditions. It is therefore more likely to be entered into solely for the purpose of controlling price if a larger market share is involved. A merger such as U.S. Steel's, that combines plant and management in an integrated fashion and permits no independent product lines, is another matter. It is unlikely to be entered into solely to control price. It will not be entered into if cost justifications are not expected. Although U.S. Steel's dominant position reflected the culmination of a merger of mergers, its nature was such that it wrought real and considerable organizational changes. It ought to have been treated as internal growth more than as a device for controlling price.

The Court's position that mere size is no offense under the antitrust laws in U.S. Steel [1920] therefore seems to me unobjectionable, but its conclusion that U.S. Steel did not violate the antitrust laws is much more suspect. High concentration seems linked in the record to price agreements. The Gary price-setting dinners and Pittsburgh-plus pricing give more than a hint of price agreements formed and led by U.S. Steel. Combined with U.S. Steel's market share, which remained high even though it had fallen from its 1905 peak, there was enough in the situation to support a belief that the mix of competition could have been improved by finding U.S. Steel guilty.

U.S. Steel escaped a guilty finding partly because the Court's majority believed the Corporation did nothing to make life miserable for rival steel firms. That it kept rivals happy seems established in the record. That this should constitute a valid and important defense is puzzling if not incredulous.

The court used the comfort of rivals as a valid defense because of precedent set in earlier cases, particularly in the famous 1911 cases of Standard Oil and American Tobacco. In these, the court avoided basing its decisions on firm size or market concentration (unaccompanied by the charge of price agreements) by tying the guilty verdicts to predatory behavior. The combination of predatory behavior and high levels of concentration were sufficient in the court's opinion to find defendants guilty in these two cases. The convoluted consequence of relying on predation is fully revealed in U.S. Steel, where cozy relationships between would-be rivals were taken by the court as a signal that predation was absent. Obviously it was absent, but equally obviously predation is a dangerous notion to employ in antitrust proceedings.

V. PREDATION8

We have seen that to emphasize price competition exclusively is to den), the econom), other useful avenues of competition. Yet, to emphasize it not at a]] would be a strange antitrust policy indeed. Prohibitions on price cutting, to avoid predation, seem to attach little if any value to price competition. The most principled defense of such prohibitions is that they keep rivals alive and well so that they can be around to reduce the future monopoly power of today's predator. Competition tomorrow is substituted for competition today.

The policy is foolish and dangerous. First and foremost, it is so because guessing the future in those respects necessary to validate this policy requires heavy reliance on speculative flights of imagination. The known benefit of today's price competition is sacrificed for the highly uncertain benefit of increasing price competition tomorrow.

Second, it is so because it supposes we can distinguish between competitive pricing and predatory pricing. I know of no theory that makes this distinction plausibly operable. Indeed, an important reason for relying on competition is our inability to know what price should be in its absence. The principle of relating price to cost is of no help in making this distinction. A cut in price imposes no smaller loss on a rival if the new price remains above the price cutter's cost or falls below his cost. It makes no difference if the rival's cost is the same, lower, or higher; his and the price cutter's profits are reduced in all these cases as a result of the price cut. Moreover, correctly perceived, the price cutter never knowingly cuts a relevant measure of his price below his marginal cost. Correct measures of price are never willingly set below correct measures of marginal cost. Just as a new firm seeking to promote its product may give some away or sell some below its production cost because he expects to make up the loss on future sales, so a predator cuts price today because he expects to make up for it tomorrow. In both cases the relevant price is the expected revenue per unit wrought by today's price cut. The predator, like the promoter, cannot increase the value of his firm if this expected revenue is less than the expected marginal cost. We may discuss predation and contrast it to promotion, but I see no way to distinguish these cases in theory without relying on highly speculative and arbitrary modeling of the problem.

If the competitive benefit to be derived in the future from a policy of barring price cuts today is so speculative, and if there is no way in theory or practice to distinguish anticompetitive from competitive price cuts, what is there to say in defense of using the notion of predation to describe unreasonable behavior? Nothing. The sooner it is set aside by antitrust courts, the better will be the mix of competitive forms that guides economic behavior. In deciding the 1911 Standard Oil and American Tobacco cases, Justice White was right in seeking out a rule of reason, but he was wrong in his choice. The U.S. Steel case is only one example of the mischief created by this choice. Utah Pie is another.

VI. VERTICAL RELATIONSHIPS

Not much needs to be said about vertical relationships, whether these involve mergers of upstream and downstream firms, tie-in sales, or territorial and price restrictions. There is little theory to support the notion that these reduce competition substantially, and there is much theory to support the notion that they either create productive efficiencies or facilitate price discrimination. Notions about market foreclosure are highly speculative. Usually they fly in the face of sound maximizing behavior or they reflect the prior existence of substantial market concentration. There are numerous reasons for suspecting that various vertical restrictions serve the purpose of mitigating externality, free-rider, and opportunism problems.

The opposition to these vertical relationships that worked its way through our antitrust courts during the decades of the 1950s and 1960s hardly gives grounds for celebration. The most that can be said of it is that it reduced price discrimination somewhat. Whether this is desirable or not is difficult to say. Price discrimination is likely to result in increased output, but not certainly so, and it may or may not reduce the aggregate of deadweight losses.

The record established by courts wrestling with vertical relationships is not to be admired, but it does offer a good example of how the substitutability between competitive forms is slighted. In a series of cases beginning with Dr. Miles [1911], a court-made distinction between vertical price and non-price restraints has emerged, with the former subject to a per se prohibition and the latter to a rule of reason. Yet, both types of restraints are substitutes in accomplishing a wide variety of objectives, although the relative costs of using them may differ enough in some situations to make them imperfect substitutes. To suppress one is to encourage use of the other, but sometimes at higher cost.

Price discrimination aside, the belief that these vertical relationships move us away from the efficient mix of competitive forms is highly speculative and not now rounded on robust accepted theory. These relationships did not deserve the harsh treatment they were given in the past. Nor do they now.

VII. A FINAL WORD

I could give two of three cheers for an antitrust policy that focused on price agreements but perhaps that allowed for more flexibility than is allowed under the per se doctrine. Mergers, when these yield very large market share and risk little loss of organizational integrity, may be treated as practical substitutes for price agreements. Otherwise, they should be treated as internal growth. A rule of reason approach to these applications of antitrust is needed, but the proper guide to flexibility is in the seeking of an efficient mixture of forms of competition and not in the attempt to reduce predation.

Each step beyond this narrow field of application, whether in the direction of predation, vertical relationships, or market concentration diminishes the value of antitrust. I can muster only one cheer because antitrust has strayed far afield and has relied too much on faulty notions such as predation. The more confined antitrust policy suggested above when joined to a policy of international free trade offers as much pressure for an efficient mixture of competitive forms as seems to me to be practically possible.

REFERENCES Case Citations Appalachian Coals, Inc., et al. v. United States, 288

U.S. 344 (1933). Arizona v. Mancopa County Medical Society et al.,

457 u.s. 332 (1982). Atlantic Richfield Co., v.U.S.A. Petroleum 495 U.S.

328 (1990). Berkey Photo v. Eastman Kodak, 603 E2d263 (1979). Brown Shoe Co., Inc. v.U.S., 370 U.S. 294 (1962). Continental T.V., Inc. et al. v. GTE Sylvania, Inc. 433

U.S. 36 (1977). Dr. Miles Medical Company v. John D. Park & Sous

Company, 220 U.S. 373 (1911). Matsushita Electric Industrial Co., Ltd. et al. v. Zenith

Radio Corporation et at., 475 U.S. 574 (1986). Nash v.U.S., 229 U.S. 373 (1913). Standard Oil Company of New Jersey et al. v.U.S.,

221 u.S. 1 0911). U.S.v. Addyston Pipe & Steel Company et al., 85 Fed.

271 (6th Cir. 1898). U.S.v. Aluminum Co. (Alcoa), 148 E2d 416 (2d Cir.

1945). U.S.v. American Tobacco Co., 221 U.S. 106 (1911). U.S.v. National Dairy Products Corp., 372 U.S. 29

(1963). U.S.v. Trenton Potteries Company et al., 273 U.S. 392

(1927). u.S.v. sealy, 388 u.s. 350 (1967). U.S.v. United States Steel Corporation et al., 251 U.S.

417 (1920). Utah Pie Company v. Continental Baking Company

et al., 386 U.S. 685 (1967).