Antitrust Law: Economic Theory and Common Law Evolution

This page intentionally left blank

Antitrust Law Economic Theory and Common Law Evolution

This book is an effort to consolidate several different perspectives on antitrust law. First, Professor Hylton presents a detailed description of the law as it has developed through numerous judicial opinions. Second, the author presents detailed economic critiques of the judicial opinions, drawing heavily on the literature in law and economics journals. Third, Professor Hylton integrates a jurisprudential perspective into the analysis that looks at antitrust as a vibrant field of common law. This last perspective leads the author to address issues of stability and predictability in antitrust law and to examine the pressures shaping its evolution. The combination of these three perspectives offers something new to every student of antitrust law. Specific topics covered include perfect competition versus monopoly, enforcement, cartels, Section 1 doctrine, rule of reason analysis, boycotts, market power, vertical restraints, tying, exclusive dealing, and horizontal mergers. Keith N. Hylton has taught at the School of Law of Boston University since 1995. He previously served on the faculty at Northwestern University School of Law. Professor Hylton currently teaches courses in antitrust, torts, and labor law, and he writes widely in the field of law and economics, with more than forty publications in American law journals and peer-reviewed law and economics journals. Professor Hylton has also served as a director of the American Law and Economics Association.

Antitrust Law Economic Theory and Common Law Evolution

KEITH N. HYLTON Boston University

   Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo Cambridge University Press The Edinburgh Building, Cambridge  , United Kingdom Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521790314 © Keith N. Hylton 2003 This book is in copyright. Subject to statutory exception and to the provision of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published in print format 2003 - isbn-13 978-0-511-06922-2 eBook (EBL) - isbn-10 0-511-06922-7 eBook (EBL) - isbn-13 978-0-521-79031-4 hardback -  hardback isbn-10 0-521-79031-X - isbn-13 978-0-521-79378-0 paperback -  paperback isbn-10 0-521-79378-5 Cambridge University Press has no responsibility for the persistence or accuracy of s for external or third-party internet websites referred to in this book, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

To my parents and the memory of my brother Kenneth

Contents

Preface

page xi

1 Economics I. Definitions II. Perfect Competition Versus Monopoly III. Further Topics

1 1 9 21

2 Law and Policy I. Some Interpretation Issues II. Enacting the Antitrust Law III. What Should Antitrust Law Aim to Do?

27 28 30 40

3 Enforcement I. Optimal Enforcement Theory II. Enforcement Provision of the Antitrust Laws Appendix

43 43 47 64

4 Cartels I. Cartels II. Conscious Parallelism III. Conclusion

68 68 73 89

5 Development of Section 1 Doctrine I. The Sherman Act Versus the Common Law II. Rule of Reason and Per-Se Rule III. Conclusion

90 90 104 112

6 Rule of Reason and Per-Se Rule I. The Case for Price Fixing II. Per-Se and Rule of Reason Analysis: Further Developments

113 113

vii

116

viii

Contents III. Per-Se Versus Rule of Reason Tests: Understanding the Supreme Court’s Justification for the Per-Se Rule

129

7 Agreement I. The Development of Inference Doctrine II. Rejection of Unilateral Contract Theory

132 133 140

8 Facilitating Mechanisms I. Data Dissemination Cases II. Basing Point Pricing and Related Practices III. Basing Point Pricing: Economics

144 145 154 160

9 Boycotts I. Pre-Socony II. Post-Socony III. Post-BMI/Sylvania IV. Conclusion

166 166 170 181 184

10 Monopolization I. Development of Section 2 Doctrine II. Leveraging and Essential Facility Cases III. Predatory Pricing IV. Conclusion

186 186 202 212 228

11 Power I. Measuring Market Power II. Determinants of Market Power III. Substitutability and the Relevant Market: Cellophane IV. Multimarket Monopoly and the Relevant Market: Alcoa V. Measuring Power: Guidelines

230 230 235 237

12 Attempts I. The Swift Formula and Modern Doctrine II. Dangerous Probability Requirement

244 244 248

13 Vertical Restraints I. Resale Price Maintenance II. Vertical Nonprice Restraints III. Manufacturer Retains Title IV. Agreement

252 252 262 267 270

14 Tying and Exclusive Dealing I. Introduction II. Early Cases III. Development of Per-Se Rule IV. Tension Between Rule of Reason Arguments and Per-Se Rule

279 279 284 286

239 243

295

Contents V. Technological Tying VI. Exclusive Dealing Appendix

ix 301 303 307

15 Horizontal Mergers I. Reasons for Merging and Implications for Law II. Horizontal Merger Law III. Conclusion Appendix

311 311 317 330 330

16 Mergers, Vertical and Conglomerate I. Vertical Mergers II. Conglomerate Mergers III. Concluding Remarks

333 333 344 351

17 Antitrust and the State I. Noerr-Pennington Doctrine II. Parker Doctrine III. Some Final Comments: Error Costs and Immunity Doctrines

352 354 371

Index

379

375

Preface

This book is, at least in part, the result of frustration. As a teacher of antitrust law, I found that the available textbooks did not offer the range of perspectives I try to give my students. I hope this frustration is reflected here. It has led me to push hard to create something that does not fit into one of the molds offered by the existing textbooks. The available textbooks on antitrust approach the subject from one of two angles, law or economics. From the law angle, one finds either law casebooks or hornbooks.The law casebook,for those who have never seen one, is a compilation of excerpts from important court opinions, interspersed with discussions and questions contributed by the casebook authors. As a way of learning the important results of a court decision, the casebook is inefficient. One is forced to read through lengthy court opinions in the end to note one or perhaps as many as three important legal propositions. In addition, I have often found myself thinking, as I looked at a casebook, that the author is trying to push me in a certain direction without stating his opinions openly. I prefer to see opinions set out openly. The format of this book avoids what I see as the drawbacks of the casebook. The case summaries state concisely the important legal propositions, and clearly separate them from the less important statements. Following each case, I discuss whether the decision can be defended on economic or legal grounds. My aim is not to brainwash students; it is to push them to question all aspects of the decision, from legal validity – that is, consistency with prior decisions and relevant statutory law – to policy considerations. The other style of law textbook is the hornbook, in which the author summarizes the cases and provides some synthesis of the law. The benefit xi

xii

Preface

of this approach is that it avoids the inefficiency of the casebook – forcing the student to read four hundred pages in order to extract forty legal propositions. The key drawback of the hornbook is that it can die quickly. The law and our understanding of it change over time. The legal propositions associated with a decision that we think are important today may not be tomorrow, while the less important statements may gain in prominence. The casebook, by setting out the decision in full, is capable of changing as views of the law change. The standard hornbook does not have this adaptability. I hope that the format of this book avoids the early death problem. Bork’s Antitrust Paradox did not die early because it presents a sustained theoretically consistent critique of antitrust doctrine. Theory plays an important role in making any discussion of the law worthy of reading after the law has changed. Theory provides a method of understanding or explaining the law, and an objective standard for criticizing it. These standards exist independently of the statement of the law. The reader of a theoretical hornbook such as Bork’s picks up more than just the case law, and this is what makes the theoretical hornbook adaptable and capable of withstanding shifts in legal doctrine. I hope I have immunized this book in the same way. What does this book contribute to the many textbooks already out there? First, I have tried to integrate law and economics at a reasonably high level. Some of the discussions of economic issues are pitched at the advanced-undergraduate economics level. The issues are simply too deep to try to cover as many and remain at a level that assumes absolutely no training whatsoever in economics. Thus, one important difference between this book and most antitrust textbooks, especially law texts, is that I devote a great deal of energy to discussing the relevant economics. Of course, if one had the time, one could get all of this material out of some of the better antitrust casebooks. The footnotes of the recent Areeda editions (coauthored with Louis Kaplow) contain references to much of the relevant theoretical work in economics. But no one has the time to consult all of the references. I have tried to incorporate some of the recent teachings in a simple way, sparing the reader the task of hunting down journal articles. The reader who seeks a deeper treatment of the issues can read the articles cited in the footnotes of this text, or consult an advanced textbook such as Jean Tirole’s The Theory of Industrial Organization, and from there embark on a broader search of the literature.

Preface

xiii

The other side of the coin in this integration process is the law. Unlike most textbooks that stress the economics of antitrust, I have presented the law in a detailed way. I summarize and provide economic analyses of the major and minor legal propositions of each decision. In some instances, I draw attention to minor parts of opinions that have interesting economic implications. Second, this book attempts to present the material as simply and straightforwardly as possible. I have tried to avoid highfalutin economic and legal terminology. I have tried to explain economic concepts in simple terms with simple examples. As a result, even though the book deals with some advanced issues, the presentation should be easy to follow even if the reader has not had much training in either law or economics. Third, unlike either the law textbook or the economics textbook, this book attempts to integrate discussion of some issues in common law theory. By common law theory, I mean the exploration of such issues as certainty of law, the relative merits of rules versus legal standards, the process of legal evolution, and the capacity of courts to apply reasonableness standards to business practices. These issues are general, appearing in many fields of law outside of antitrust. I have taken a position on many of these issues, and, of course, the reader is free to disagree. For the reader who disagrees, the useful part of the discussion will be the examples in which case law and economics are integrated in a discussion of legal theory. The case law provides a set of concrete examples that illustrate some long-standing controversies in common law theory. The discussion of economics provides some content to the notion of reasonableness when used as an alternative to that of positive legal validity. For the student of legal evolution, the most general and frequently recurring problem in antitrust law is the tension between the economic conception of a reasonableness inquiry and the administrative concerns of courts and enforcement agencies. An economically defensible reasonableness inquiry would seek to determine whether a practice, challenged as a restraint of trade, generates economic benefits in excess of its social costs. A practice that enhances society’s wealth would be permitted under this standard. However, such a standard, in the antitrust realm, would tend to put courts and enforcement agencies in a disadvantaged position relative to defendants. Courts and enforcement agencies often do not have enough information to rigorously assess the economic reasonableness of a challenged practice. Such an assessment

xiv

Preface

often requires information privately held by the defendant and his business associates. Courts and enforcement agencies have responded to this imbalance by interpreting antitrust statutes in a manner that excuses the plaintiff in certain instances from having to demonstrate that the challenged practice is economically unreasonable. The Supreme Court started along this path in its first opinion interpreting Section 1 of the Sherman Act, U.S. v. Trans-Missouri Freight Assn;1 and post–Sherman Act antitrust statutes have avoided any reference to reasonableness. The problem this has generated is that courts find it very hard to move away from reasonableness standards. Indeed, to do so is to abandon the most basic feature of common law adjudication: the equation of legal validity with some notion of reasonableness. The common law process has traditionally developed, generated new law, by extending settled propositions. The extension is justified by appealing to reasonableness, whether based on cost-benefit balancing or a survey of the parties’ expectations. In other words, it has not been the practice of courts in the western world to say, “this is the law, and to hell with its justification.” Because the common law process requires justification of legal standards, antitrust courts are continually pushed in the direction of providing an economic reasonableness justification for every decision. The tension between economic reasonableness and administrative concerns seems to be the principal, or at least one of the principal, driving forces in the evolution of antitrust law as a field of federal common law. In several chapters of this book, I note that the law has moved from a reasonableness inquiry to a per-se standard and then back toward a reasonableness inquiry.2 The forces pushing in either direction are continually acting on antitrust doctrine. Indeed, it is possible to state a very simple model of antitrust evolution. Public enforcement agencies and federal statutes put pressure on courts to adopt per-se standards. This pressure pushes the law in the direction of per-se rules until, in some cases, a “validity crisis” is reached – a point at which the court can no longer defend its decisions by appealing to economic reasonableness arguments. At that point, the court either retreats from the per-se standard (Sylvania, Chapter 13) or reinterprets the reasonableness argu1 2

166 U.S. 290 (1897). See Chapters 5 (Section 1), 9 (boycotts), 13 (vertical restraints), and 15 (horizontal mergers).

Preface

xv

ments (Addyston Pipe, Chapter 5). As I point out in the text, this process is observed in several areas of antitrust law. The tension between economic reasonableness and administrative concerns also has implications for the role of economic theory in antitrust analysis. Economic theory is obviously helpful, and probably necessary, in getting a sense of the meaning of economic reasonableness; and I do not believe there is any area of economic analysis that is too abstract to be useful in this enterprise. However, there may be limits on the extent to which antitrust doctrine should reflect or take into account all of the concerns of the economic theorist. As Judge (now Justice) Breyer put it in his Barry Wright opinion, while technical economic discussion helps to inform the antitrust laws, those laws cannot precisely replicate the economists’ . . . views. For, unlike economics, law is an administrative system, the effects of which depend upon the content of the rules and precedents only as they are applied by judges and juries in courts and by lawyers advising their clients. Rules that seek to embody every economic complexity and qualification may well, through the vagaries of administration, prove counter-productive, undercutting the very economic ends they seek to serve.3

In view of this, the critic of antitrust doctrine must give some thought to the likelihood and cost of judicial error. Where the expected cost of error (the likelihood multiplied by the cost) is high, antitrust doctrine should stop short of requiring a full blown inquiry into all of the matters of concern to the economist. The controversial part, of course, is identifying those areas in which the expected cost of error is high enough to warrant this approach. If the costs of error were the same, whether an erroneous finding of innocence (false acquittal) or an erroneous finding of guilt (false conviction), then we might be led by administrative concerns to adopt rules of per-se legality and per-se illegality with equal frequency. However, economic analysis permits us to state presumptions or hypotheses regarding the relative costs of false acquittals and convictions in antitrust litigation. If, with respect to a certain type of claim (e.g., predatory pricing), economic reasoning leads us to hypothesize that the expected costs of false convictions are greater than those of false acquittals, we should prefer to adopt per-se legality rules to govern these claims, and conversely. This is the primary sense in which I think economic theory 3

Barry Wright v. ITT Grinnell Corp., 724 F.2d 227, 234 (1st Cir. 1983).

xvi

Preface

should inform antitrust doctrine, and I have tried in several parts of this book to use economic reasoning in this manner. The mixture of perspectives may make this book unsuitable for any particular class. The criticism of court decisions may seem too harsh to make appropriate material for law students. I leave these questions to the reader. At the least, I hope that this will provide answers to, or a better method of answering, many questions not covered in the standard presentations of antitrust law. As with any project that takes a lot of time, this one benefited from the input of many people. Mark Grady, whose office was across the hall from mine in my early teaching years at Northwestern, influenced my thinking on antitrust as well as other areas of legal scholarship. Readers who are familiar with Mark’s work will notice that influence in some parts of this book. Indeed, this book follows Mark’s course outline, which he gave me in my first year of teaching to help me get started. I also must thank Frank Fisher, who took the time to give me detailed, page-by-page comments on a very early draft. And there are friends and colleagues, like Roger Blair, who gave me useful advice on a chapter here or there, and I am sorry that I cannot list all of them here. Finally, I should thank my former students and research assistants, especially Deborah Loesel, Brian Kaiser, and Jessica Selb, who helped with the research and inspired me to see this project through.

1 Economics

In this chapter, I will introduce economic concepts that I plan to use throughout the text, set out the basic economics of monopoly, and compare monopoly with its polar opposite, perfect competition. I also will discuss some relatively new topics, such as transaction-cost and information economics, and their relevance to antitrust policy.

i. definitions A. Monopoly A monopolist is a single supplier of a good. However, this definition is too simple, because it includes firms that become dominant by being the lowest-cost competitor and those that obtain an exclusive franchise from the state. As far as antitrust policy is concerned, there is a big difference between these two cases. Because the simple, “single-supplier” definition is potentially misleading, one should focus on market conditions. The crucial feature of monopoly status is the absence of competition from other firms. The common example of monopoly in our lives is local telephone service, provided in most places in the United States by a regulated, privately owned monopoly. However, even here competition from wireless and optical fiber companies has eroded the monopoly status of the local telephone companies. One of the purest monopolies in recent memory was Aeroflot, the airline of the former Soviet Union. Before the breakup of the Soviet Union, there were no competing airlines. 1

2

Economics

Figure 1.1

B. Market Price A market equilibrium, where the quantity demanded by consumers equals the quantity supplied by producers, generates a market price, as shown in Figure 1.1. The downward sloping line is the demand curve and the upward sloping line is the supply curve. Think of the demand curve as a schedule of bids offered by consumers. Each point along the curve is a maximum price that at least one consumer is willing to pay. The horizontal axis measures the total quantity demanded at a given price, and since each consumer would accept the item at a lower price, quantity demanded increases as price falls. Similarly, one can think of the supply curve as a schedule of minimum asking prices stated by producers. Since each producer is willing to sell the good at a price at or above his asking price, the total quantity offered for sale at a given market price (measured by the horizontal axis) increases as price rises.1 1

More technically oriented treatments typically explain that each consumer has a schedule of bids for each quantity desired. Consumers offer less per unit for higher quantities because the utility gained per unit of consumption falls as consumption expands. The market demand curve is the “horizontal sum” of the individual demand schedules. The student trained in economics may prefer to think in these terms. I have attempted to simplify the presentation in the text.

I. Definitions

3

In a market with many producers and consumers, none of them actively sets the equilibrium price. It is, in a passive sense, determined by the actions of the marginal consumer and marginal producer. The marginal consumer (point C in Figure 1.1) is just indifferent between buying the good and going without it, given the market price. Similarly, the marginal producer (also at C) is indifferent between selling at the market price and keeping his output. The inframarginal consumer (to the left of C along the demand curve in Figure 1.1) is willing to pay more for the good than is the marginal consumer, and the inframarginal producer (to the left of C along the supply curve) is willing to part with the good for a lower price than the marginal producer would accept. The price in an exchange between inframarginal actors is indeterminate – it is any level between the maximum the consumer is willing to pay and the minimum the producer is willing to accept. To see the role played by marginal actors in the determination of equilibrium, suppose the price is initially set above the level that equalizes the amounts demanded and supplied. Suppliers would offer a quantity larger than consumers were willing to purchase, and as a result some sellers would be unable to find buyers. Among them would be inframarginal sellers, who would cut their asking prices in order to make a sale. This process would continue until the equilibrium price is reached.

C. Market’s Contribution to Wealth Because the marginal consumer determines price, all other consumers (inframarginal) gain by making trades in the market. Consumers’ surplus measures the gain to consumers from taking advantage of the market: some consumers would still buy the good at a higher price, but they can purchase it at the cheaper market price. Similarly, because the marginal producer determines price, producers’ surplus measures the gain to producers generated by market transactions. The diagram in Figure 1.1 also illustrates the incremental wealth generated by the market, which is equal to the sum of consumers’ surplus and producers’ surplus. Consumers’ surplus is the area ABC, and producers’ surplus is the area DBC. The total surplus, or the market’s contribution to wealth, is maximized when price is equal to the market equilibrium level p1 and quantity is equal to the market equilibrium level q1. Because total surplus hits its maximum at the market equilibrium, I will refer to this as the social optimum.

4

Economics

Although this may seem an unusual way to measure incremental wealth,2 this is the approach Adam Smith emphasized in arguing against the mercantilist policies followed by England and other European countries over the eighteenth century.3 The doctrine of mercantilism, still with us today in many quarters, held that a government should manage foreign trade in order to maximize gold reserves. To the mercantilists, this was how a country became wealthy. In practice, the doctrine necessitated a strategy of maximizing exports and minimizing imports. Adam Smith’s argument, startlingly counterintuitive at the time and still misunderstood by the majority of governments today, was that the mercantilists’ conception of wealth was invalid and that their policies were likely to reduce rather than increase wealth. A market’s real contribution to wealth is the difference between the value of the benefits a good provides and the resource cost of its production. Smith argued that a policy of free trade in competitive markets maximizes this measure of incremental wealth. Of course, understanding Smith’s argument requires some familiarity with the properties of competitive markets. I take up that topic next.

D. Defining Perfect Competition A competitive equilibrium satisfies the assumptions of the model of perfect competition, which are as follows. 1. Atomism. The output of each seller and the consumption amount of each buyer is a small fraction of the total output of the market, so no buyer or seller can have more than a very small influence on market price or quantity. Alternatively, each buyer and seller takes market price as given. We could speak generally of a spectrum with atomism on one end and monopoly on the other. Of course, the theoretical endpoints are hardly ever observed. Atomism, in its extreme version, requires an infinite number of infinitesimally small producers and consumers. Monopoly requires a single seller, but even where we do find a single seller of an item, often suppliers of close substitutes constrain the monopolist’s pricesetting decisions. 2

3

It is important to note the difference between stocks and flows. Consumers’ surplus is a flow while wealth is a stock. In view of this, I have referred to consumers’ surplus as a component of incremental wealth. An Inquiry into the Nature and Causes of the Wealth of Nations (Edwin Cannan, ed., New York: Modern Library 1994).

I. Definitions

5

2. Perfect Information. Consumers can distinguish between different goods. They also know if one seller is offering a particular good at a lower price than another seller. Really, all we need is that information must get around reasonably fast. The assumption of perfect information simplifies the matter. Obviously, the assumption is not an accurate description of the world.4 In the real world, we see firms advertising. We could make the model resemble the real world more closely by assuming that information is a commodity that must be supplied. However, once we assume information must be supplied, the reason for making the simplifying assumption of perfect information starts to become clear. The market for information is peculiar. Information is a public good, in the sense that a decision to supply it to one person generally means that the good is also available to others. For example, the purchaser of a newspaper may read it and then give it to a friend. Because the information can be shared, the producer may not receive compensation for the benefits conferred upon a large number of consumers, and in this case the market may provide insufficient incentives to produce news. This is illustrated in Figure 1.2. The forward-shifted demand curve includes the benefits of newspapers to nonpaying readers. The social optimum is at (p2, q2) rather than the market equilibrium (p1, q1). This example suggests that relaxing the assumption of perfect information immediately introduces some element of market failure into the model. Consider the case of advertising to inform consumers of the existence of a better mousetrap. Suppose there are competing sellers of this new mousetrap. A seller who advertises the mousetrap cannot be sure that the benefit will accrue to himself alone, because he cannot limit the message only to consumers who will purchase from him. Since some of the benefits may go to other sellers, his incentive to pay for informative advertising is attenuated. In the extreme case, failure of the assumption of perfect information can make a market virtually infeasible. The best example is the problem of adverse selection in the insurance market. Suppose there is a 4

A related and more fundamental criticism is that the assumption of perfect information ignores the central problem that needs to be explained: how privately held information is revealed and communicated among market participants, see F. A. Hayek, Individualism and Economic Order 77–106 (London: Routledge & Kegan Paul, 1949). Although Hayek’s critique has important implications for antitrust policy, space will not permit me to cover it here.

6

Economics

Figure 1.2

continuum of risk levels among potential insurance purchasers, and the insurer cannot determine the risk level of each applicant. The insurer’s price will be a weighted average of the prices that should be charged to each type, the weights reflecting the anticipated shares of each risk type in the insured population. If some relatively low-risk customers exit the relationship and insure themselves or do without insurance altogether, then the price must be increased for the relatively high-risk customers who remain. But this may lead others to drop their policies, and so on. In the end, only the most risky customers seek insurance, and with little to be gained from pooling their risks, the market vanishes. 3. Mobility. Resources flow easily from one market or sector of the economy to another: no barriers to entry exist. Without mobility, monopoly power becomes possible. Simply meeting the assumption of atomism does not eliminate the possibility of market power. To take a concrete example, consider the market for attorneys. In the United States, there are too many of them to count. It would seem, therefore, that the atomism requirement is satisfied. However, the market is not perfectly competitive because not everyone who could perform as an attorney is permitted to enter the market. Every attorney must pass

I. Definitions

7

a bar exam and be sworn in to the state in which he or she wishes to practice. The bar passage requirement reduces the total number of attorneys and allows them to earn a return in excess of the opportunity costs of the skills and resources employed in that profession. 4. No Third-Party Effects. The model of perfect competition assumes there are no externalities, that is, third-party effects. The parties who contract over the supply of a good or service bear all of the costs and benefits associated with the production of that good or service. Externalities lead to production levels that deviate from the social optimum. For example, consider the case of a company that produces chemicals and also pollutes the water as a byproduct. The company produces too much from society’s point of view. The total cost of the company’s output is more than the production cost borne by the company, it also includes the costs generated by the pollution. If the company were forced to bear the pollution costs, it would demand a higher price in order to supply the market. Put another way, the supply curve for the chemical producer would shift back, as shown in Figure 1.3, reflecting the higher price demanded for each level of output. The

Figure 1.3

8

Economics

upward-shifted supply curve in Figure 1.3 reflects the real costs of producing chemicals. As the diagram also shows, given any market price, the firm overproduces, relative to the social optimum (which is at q1). One way to correct the overproduction demonstrated in this example is to tax the chemical producer. The company’s supply should be reduced by taxing it at a rate that reflects the costs generated by the pollution it imposes on society. A more general approach to solving this problem was suggested by Ronald Coase.5 Coase demonstrated that in a regime in which transaction costs were zero, and property rights well-defined, resource allocation would be efficient. To see why this holds, consider again the example of the chemical producer who pollutes the water. Suppose a downstream firm finds that it must clean the water in order to use it in production. If it is less costly for the upstream chemical producer to reduce its production than for the downstream firm to clean the water, then the downstream firm will have an incentive to offer a payment to the upstream chemical producer in exchange for a reduction in the upstream producer’s level of output. The incentive for such a side payment remains as long as the gain from cleaner water to the downstream firm (area abcd in Figure 1.3) exceeds the loss from cutting back production to the downstream firm (area abc). As Figure 1.3 suggests, the side payments will continue until the upstream producer cuts back to the optimal level q1. 5. Homogeneous Product. Products are not differentiated. For example, a seller of wheat really sells standard wheat – nothing fancier or different from what every other wheat seller offers. This assumption implies that markets cannot be divided up into small enough portions to violate the atomism assumption. If, for example, the market for wheat could be divided into one million markets for different types of wheat, one of those one million markets could likely contain only one firm. Thus, the homogeneity assumption provides another way of avoiding monopoly. Homogeneity also helps avoid the informational problems suggested above. Suppose there were several brands of wheat and consumers could not distinguish one from another. Then an inferior brand might sell for the same price as a superior brand, because consumers were unable to make fully informed choices. 5

R. H. Coase, The Problem of Social Cost, 3 J. Law & Econ. 1 (1960).

II. Perfect Competition Versus Monopoly

9

E. Economic Profit Economic profit is the excess of revenue over costs, where costs include compensation for risk-taking and the opportunity cost of capital. This is not the same as accounting profit, which makes no attempt to include risk-taking and lost opportunities as elements of total cost. A firm may be earning positive accounting profits and negative economic profits. This is why one cannot infer monopoly power simply from observing the profit reports of a company. A simple story illustrates the concept of economic profit. Suppose a wealthy ice cream lover donates two plots of land to a company that runs a chain of ice cream parlors. One plot is in Quiet Square, a sleepy, smalltown intersection that rarely sees crowds. The other plot is on Busy Street, smack in the middle of downtown Busy City, an area full of pedestrians from sunrise to sunset. One would not be surprised to find that the Busy Street parlor makes a substantially greater accounting profit than the Quiet Square parlor. However, the relation between their economic profits may be the opposite. To measure the economic profit of the Busy Street parlor, one must subtract from accounting profit an estimate of the rental value, or opportunity cost, of the plot of land on Busy Street. Economic profit at either ice cream store is measured by the extent to which accounting profit exceeds the rental price for the location.

ii. perfect competition versus monopoly A. Perfect Competition The fundamental result of the model of perfect competition is the following: In long run competitive equilibrium, firms earn zero economic profits. This happens because of entry and exit. If firms earn positive economic profits, then rivals will enter the market. Entry continues until the increase in supply pushes price down to a level that just compensates for the cost of producing and the opportunity cost of capital and managerial skill. If firms earn negative economic profits, exit will occur until economic profits return to zero. It is important to keep in mind that entry and exit occur in response to economic profits, not accounting profits. Second, economic profits go to zero in the long run, not the short run. Nothing in the model of perfect

10

Economics

competition suggests that firms cannot earn economic profits in the short run. Indeed, the appearance of economic profits (economic losses) in the short run causes entry (exit). Although the five assumptions of perfect competition described in the previous section should be sufficient to generate the zero economic profits proposition, an intermediate set of assumptions (almost all of them derivable from the initial five) are useful in analyzing the long run equilibrium of a perfectly competitive economy. The first intermediate assumption is that each individual firm faces an infinitely elastic demand curve. The elasticity of demand measures the responsiveness of the quantity demanded to the changes in the price of the good. A zero demand elasticity means that a price change has no effect on the quantity demanded. Infinite elasticity means that a firm can produce as much as it wants to sell at the equilibrium price without that increase in quantity supplied having any effect on the market price. Because of this assumption, the firm in a competitive economy becomes a “price taker,” that is, it takes the market price as given – fixed, not subject to its influence. A firm can certainly charge a price different from the market price; however, the assumption implies that the firm has no incentive to do so. Suppose the firm sets its price above the competitive level. It will sell nothing, because consumers can buy at the market price from another firm. Suppose the firm sets its price below the competitive level. Then it sells the same amount as it would at the competitive price, but it will make less revenue because it sold at a lower price. The second intermediate assumption is profit maximization. In long run competitive equilibrium, economic profits are zero, which implies that price is equal to average cost. Let C = production cost, AC = average cost (C/q), MC = marginal (or incremental) cost. Then profit = pq - C = q(p - AC), so positive profit implies p > AC, and zero profit implies p = AC. Since the firm is also maximizing profits, price must equal marginal cost ( p = MC). Why? The firm is maximizing profits, which means it increases output until marginal revenue equals marginal cost (MR = MC), or that it will produce each unit that brings in as much or more revenue than it costs. When the demand curve is infinitely elastic, MR = p. Thus, in long run competitive equilibrium MR = MC = AC = p. Profit-maximization is not a strong behavioral assumption because a competitive environment more or less forces firms to maximize profits. Suppose a firm chose not to maximize profits. Since economic profits, among profit-maximizing firms, are zero in the long run, a firm that did not maximize profits would earn a negative economic profit. The owners

II. Perfect Competition Versus Monopoly

11

of the firm would then come under pressure to sell the assets or transfer them to some other use. The third intermediate assumption is that the individual firm faces a U-shaped long run average cost curve. It has this shape because of increased opportunities for specialization (as scale increases) and substitution of more productive inputs. Consequently, we can view the firm as initially drawing on factors of production uniquely suited to the firm’s activity. Because these specialized factors are unusually productive, they drive down long run average cost. To see this, let L represent the only (variable) input and let its price be w. Also, let APL represent the average product of L. In the long run (where all factors are variable), AC = C/Q = wL/Q = w/APL. Thus as average product increases, average cost falls. Later, as the gains from specialization are exhausted and the firm begins to draw on factors that are not so well suited, the long run average cost curve begins to rise. I will not present a detailed discussion of the relationship between short- and long-run cost curves, but I will note here that short-run cost curves typically are U-shaped for different reasons. Short-run cost curves fall initially because of fixed costs, and, in some cases, the rising average product of the variable factor; and begin to rise at some point because of diminishing returns. With one factor of production fixed, the variable factor becomes less productive as output expands. On the basis of these assumptions, a simple diagram illustrating the process of competition is shown in Figure 1.4. Suppose price rises above

Figure 1.4 Psr = short-run price P1r = long-run equlibrium price Area Psr abc = short-run profit

12

Economics

the competitive level (because demand shifts outward from D to D’). Firms expand output, and see profits in the short run. The profits provide a signal that leads to entry. Entry causes the industry supply curve (see S) to shift outward (S’), until price returns to the long-run equilibrium. At that level, price equals the minimum of the long-run average cost curve, which means that firms produce goods in a method that economizes on production resources. The long-run competitive equilibrium is efficient in the following sense. Total welfare is maximized because price, which measures the marginal benefit to consumers, equals marginal cost. Thus, no alternative price-output combination could increase society’s welfare relative to the long-run competitive equilibrium. Furthermore, because price is equal to the minimum of the long run average cost curve, the long run competitive equilibrium achieves this welfare-maximizing priceoutput combination in a manner that economizes on productive resources.

B. Monopoly 1. Basics. The monopolist, unlike the perfectly competitive firm, does not face an infinitely elastic demand curve. This is because the monopolist’s demand curve is the market demand curve. Like the competitive firm, the monopolist tries to maximize profit. To do this, the monopolist expands output until the increase in revenue attributable to the last unit just equals the incremental increase in cost (MR = MC). In order to achieve this, the monopolist restricts output and charges a higher price than would a firm in a competitive industry operating under the same cost conditions. This occurs because marginal revenue for the monopolist is always less than what it would be for a competitive firm facing the same price. Why? Because an increase in output raises total revenue by price times the increase in quantity, but also reduces per unit revenue by causing a decrease in the price. Thus, under monopoly MR < p. The results under monopoly are the following. (1) Economic profits are positive (p > AC). (2) Part of the wealth of consumers is transferred to the monopolist (see the rectangle PmABC in Figure 1.5). (3) Part of society’s wealth is wasted. This is illustrated in Figure 1.5, which assumes a horizontal average cost curve for simplicity. The area labeled “deadweight loss” (triangle ABD) measures waste. This waste of society’s wealth occurs because consumers are willing to pay a price that exceeds

II. Perfect Competition Versus Monopoly

13

Figure 1.5 Pm = monopoly price Qm = monopoly output Area PmABC = monopoly profit = monopoly transfer

the marginal cost of producing additional output, but the monopolist does not supply the additional output. Thus, the deadweight loss triangle contains the set of potential welfare-enhancing exchanges that are forgone by the monopolist. The transfer of wealth shown in Figure 1.5 does not always happen as described. Sometimes increasing costs consume all of the wealth: it is still a transfer, but not to the monopolist’s profits. For example, the transfer may go to the owner of the assets needed to acquire the monopoly. Consider the case of taxicab medallions in New York City. The last report of a sale of a taxi medallion in New York City listed the price as $182,000,6 which suggests buyers expect to earn monopoly rents. Medallions are an example of a rent-seeking expenditure: an investment that does not generate wealth for society. In the case of monopoly, people bid for rights to a portion of wealth transferred from consumers. After acquiring the right, they need to earn the rent to pay off the debt the acquisition generates. In some cases, the rent may just meet the debt payments. Indeed, it follows from the model of perfect competition that if the market for acquisition of monopoly status is competitive, rents earned by winners will be merely sufficient to cover the acquisition costs. Put another way, if the market for acquisition of monopoly status is 6

See Sheryl Fragin, New York’s Terror Taxis, Explained, The New York Times, Sunday, August 21, 1994, page 9, Sunday Business Section.

14

Economics

competitive, we should observe a phenomenon that can be described as ex ante rent dissipation. Sometimes, managers and other agents of the firm transfer the monopoly rents into production costs – a process that can be labeled ex post rent dissipation. We just considered the example of the taxicab medallion owner who buys into a monopoly and needs to earn the stream of monopoly rents in order to pay off the debt (an example of ex ante rent exhaustion). But suppose the owner already has the medallion, and a new law limits the supply of medallions, and suppose further that price regulation discourages price competition among the medallion owners. The owners may find other ways to compete. For example, they may purchase fancy cars to use as cabs, or offer drinks or food to customers. This type of competition occurred in the airline industry under regulation by the CAB. Airlines competed in nonprice categories, and the competition drove service costs up.7 Unions demanded a share of the rents and managers paid them off to avoid labor problems. This process converted monopoly rents over time into costs. Recall that the waste of society’s wealth shown in Figure 1.5 results because the monopolist forgoes several wealth-enhancing trades. Both sides could gain if a transaction took place within the deadweight loss triangle. The portion of the demand schedule between points A and D in Figure 1.5 shows the maximum prices consumers are willing to pay, which reflect their valuations of the benefits they derive from the monopolist’s product. The marginal cost the seller incurs by supplying additional output is given by the portion of the marginal cost curve between points B and D. Since the consumer’s willingness to pay equals or exceeds marginal cost at all output levels along these segments, both sides would gain at any transaction price less than the consumer’s maximum offer price and greater than marginal cost. Why doesn’t the monopolist supply the additional output? To do so while still earning at least the monopoly profit would require the firm to price discriminate by setting a lower price for consumers whose maximum offer prices are on the portion of the demand curve between points A and D in Figure 1.5. If the monopolist could perfectly pricediscriminate, by charging each consumer a price equal to the consumer’s maximum offer price, then the monopolist would expand output up to the competitive level (point D). Note that this implies that there would 7

See, for example, Richard A. Posner, The Social Cost of Monopoly and Regulation, 83 J. Pol. Econ. 807–27 (1975).

II. Perfect Competition Versus Monopoly

15

be no deadweight loss: the monopolist supplies the competitive level of output, and takes all of the additional surplus in the form of profit. In the more realistic scenario in which the monopolist cannot charge each new consumer a price equal to that consumer’s maximum offer price, the monopolist still has an incentive to expand output if he can sell additional output to the consumers between points A and D on the demand curve at some price between those two points (and note that unlike the perfect price-discrimination case, the new consumers will gain also). However, the monopolist will not go all the way to the competitive level of output in this case. The problem with price discrimination is this: how would the monopolist prevent the low-offer-price consumers from reselling to other (high-offer-price) consumers? The administrative costs of setting up and enforcing a price discrimination scheme (in which a monopolist charges different prices, unrelated to the costs of supplying those units, to different consumers) could outweigh the additional benefits to the monopolist. The standard analysis of monopoly implicitly assumes the administrative costs of price discrimination exceed the benefits to the monopolist. 2. Stability of Monopoly. Because monopolists earn profits in excess of opportunity costs, they attract entry. It follows that for monopolies to continue, barriers must prevent entry by competitors. But what is a barrier to entry? Is having to build a plant a barrier to entry? Generally the literature identifies two types of entry barriers: natural and artificial. Let us start with artificial barriers to entry. There are two kinds of artificial barriers: government created and privately created. Government created artificial barriers can include: (1) patents, (2) taxicab medallions, (3) government franchises (e.g., electricity supply) or exclusive contracts, and (4) licensing. Examples of privately created artificial barriers are: product differentiation, advertising, exclusivity contracts, and product tying. The government created barriers are fairly easy to understand, but the theory of private barriers poses some difficulty. Product differentiation sometimes acts as a barrier to entry because it creates brand loyalty and therefore makes it more difficult for a rival to enter and compete for consumers. Exclusive dealing contracts create barriers by foreclosing the market to rivals. For example, if firm A has an exclusivity contract with the only supplier of a vital input, it would be difficult for a rival to enter and compete against firm A. Product tying also tends to exclude

16

Economics

rivals by forcing them to enter at two levels (the tying and tied product) in order to compete against the seller of the bundled product. The theory of privately created barriers is a field of controversy in antitrust policy.8 On one extreme is the expansive view suggested in the work of Joe Bain,9 and on the other the view, suggested in the work of Harold Demsetz,10 that the government creates the only real entry barriers – and even then the concept is troubling in Demsetz’s view because government necessarily plays an important role in defining property rights. George Stigler11 took an intermediate position, labeling such things as product differentiation a barrier to entry only if the cost of differentiating a product is higher for an entrant than an incumbent firm. Admittedly, the private barrier theory can go to an absurd extreme. Consider the necessity of building a plant in order to produce the good. Is this also a privately created entry barrier? If building a plant is not a privately created entry barrier, then why is it a privately created barrier to form an exclusive dealing arrangement with the only supplier of a vital input? One could say that in the former case, any potential rival could build his own plant, while the latter example involves an exclusive arrangement that cannot be duplicated. But suppose the incumbent firm’s plant uses up the best location available, and suppose there are substitutes to the vital input? Natural barriers make up the second general class of entry barriers. The classic example is that of an “increasing returns” or “high fixed costs” monopoly, such as railroads, electricity suppliers, and water suppliers. The phenomenon of a falling average cost through the scale of production appears in each of these examples. Figure 1.6 illustrates the cost curves for a high-fixed-costs monopoly. The marginal cost curve lies below the average cost curve at all output levels. Because of this, the firm has an incentive to expand output even when it would not be able to cover all of its costs at the higher output level. If the average cost curve declines throughout, as shown in Figure 1.6, then competition will result in leaving one firm in the industry. That one firm will have a natural monopoly. It would be inefficient to have two 8

9 10 11

For an illuminating discussion of the problems of definition in this area, see Harold Demsetz, Barriers to Entry, 72 American Economic Review 47–57 (March 1982). Joe S. Bain, Barriers to Entry (Cambridge, MA: Harvard University Press, 1956). Demsetz, supra note 8. George J. Stigler, The Organization of Industry 67–70 (Chicago and London: University of Chicago Press, 1968).

II. Perfect Competition Versus Monopoly

17

Figure 1.6

firms supplying the good. Why, for example, should two firms run sewer pipes, or telephone wires through the same portion of a city? In many of the early Sherman Act cases, railroads argued that they suffered from the declining average costs phenomenon, and should therefore be allowed to enter into cartels.12 Unrestrained competition, they argued, would lead to ruinous competition: in the end, only one railroad would survive, which would then charge a monopoly price. Economists today generally do not accept this argument, and antitrust courts at the turn of the century rejected it. As stated, it is a weak argument. The fundamental problem is that average costs are unlikely to fall through the entire range of production. At some point, it would become difficult to gain efficiencies by simply expanding. No individual railroad ever grew so large as to test this proposition, but it is implausible anyway. In spite of the implausibility of the ancient ruinous competition argument, new research has suggested that the argument for price-fixing 12

See, for example, U.S. v. Trans-Missouri Freight Assn., 166 U.S. 290 (1897).

18

Economics

cartels may have had some validity. Indeed, the improbability of a firm experiencing declining average costs through the entire scale of production seems now, in light of recent research, to provide additional support to the policy advanced by the railroads. Chapters 4 and 5 will discuss this subject. 3. Deadweight Loss: Large or Small? There is a strong argument for the position that the deadweight loss due to monopoly is small, too small, in fact, to concern enforcement authorities. In 1954, Arnold Harberger estimated that deadweight loss triangles distributed across the economy add up to at most 1/10 of 1 percent of national income.13 Harberger’s estimate touched off several attempts to independently estimate the size of the welfare loss, most of which reached conclusions similar to his and a few that generated numbers as high as 4 to 7 percent.14 Although Harberger’s analysis remains a matter of contention, the clear implication is that the benefits of anti-monopolization efforts are small, and probably less than the costs of legislative and enforcement activity directed at the monopolization problem. The view that deadweight loss is large has received its best theoretical support from the literature on rent-seeking, which began in 1967 with an article by Gordon Tullock.15 Tullock made the point that the expectation of earning profits in excess of opportunity costs would generate efforts to gain entry or ownership of the right to monopolize. In a fully competitive ex ante market in monopolization rights, the winning bid would equal the expected flow of monopoly rents. The resulting winner

13

14

15

Arnold C. Harberger, Monopoly and Resource Allocation, 44 American Economic Review 77–87 (May 1954). Harberger’s approach tried to err on the side of overestimating rather than underestimating the resource misallocation due to monopoly power. One might argue that the effort was flawed from the start because the size of the misallocation was itself a reflection of the threat of antitrust enforcement in the period of Harberger’s sample, 1924–8. But it is unlikely that antitrust enforcement had a large impact on the size distribution of firms over the 1920s, see George J. Stigler, The Economic Effects of the Antitrust Laws, 9 J. Law & Econ. 225–58 (October 1966). Furthermore, there is evidence that early antitrust enforcement, by outlawing cartels and leaving mergers unregulated, may have caused an increase in market concentration levels, see George Bittlingmayer, Did Antitrust Policy Cause the Great Merger Wave?, 28 J. Law & Econ. 77–118 (April 1985). For an excellent summary, see F. M. Scherer, Industrial Market Structure and Economic Performance 459–62 (Houghton Mifflin, 2d ed. 1980). Gordon Tullock, The Welfare Costs of Tariffs, Monopolies and Theft, 5 Western Economic Journal 224–32 (1967). For a survey of the literature on rent seeking, see Dennis C. Mueller, Public Choice II 229–44 (1989).

II. Perfect Competition Versus Monopoly

19

would function as a zero-profit monopolist. Such a contest, however, leads to an unproductive use of resources. Theorists have long argued that efforts to acquire a monopoly waste resources. In fact, what we typically observe in the acquisition process seems to be a transfer of resources to others. Given that most if not all of the ultimate recipients of this transfer are themselves involved in legitimate activities, one might appropriately ask whether a real waste exists. There is a straightforward answer. Even if we assume that all of the resources devoted to the acquisition of monopoly status are transferred to people involved in legitimate activities (e.g., public-interested regulators, lawyers, economists), it remains true that the transaction ultimately ends in a transfer of wealth. Because that transfer adds nothing to the stock of goods and services, it seems appropriate to refer to the resources employed in effecting this transfer as wasted. In particular, activities that serve no purpose other than to create and maintain the monopoly position are thorough sources of waste.16 4. Inadequacies of the Criticism of Monopoly. Here I want to point out some rather sturdy criticisms of the antimonopoly position. Start with the short run–long run distinction. Joseph Schumpeter noted that short run profits must appear, for otherwise the incentive to enter and the incentive to innovate would never exist.17 But the attainment of a monopoly position often provides short run profits. Take, for example, the process of innovation. A firm develops a new production process that results in a dramatic reduction in costs, leading to an expansion of output and a short run monopoly. Over time, others learn how to mimic the process and entry reduces profit to zero. Schumpeter’s important insight was that the innovation-profit-entry process is a chain that is linked at several places, and that the causation is not necessarily unidirectional. 16

17

There is a second and less clear answer: to the extent that beneficial externalities result from the wealth-transfer process, perhaps not all of the resources employed are wasted. Suppose, for example, that lawyers employed in the transfer process gain skills and develop law that may be usefully applied in other, productive sectors of the economy. In this case it is less clear that the entire process is wasteful. Still, it is probably a waste. To make the case that the transfer process is productive under these conditions, one must show that the resources used in the transfer process could not be used more productively in some other sector of the economy in which the same or similar beneficial externalities are possible. This is a high burden of proof. And if it is not, this merely serves as an example of how difficult it is, when one introduces externality considerations, to reach any firm theoretical conclusions on the desirability of monopoly. Joseph A. Schumpeter, The Theory of Economic Development 128–56 (1934).

20

Economics

The prospect of earning large, temporary profits generates efforts to innovate. Successful innovation leads to large profits. Take away or reduce the size of profits, and you will see less innovation, and less entry. Schumpeter’s point can be made somewhat clearer by considering the expectations of actors. Schumpeter’s claim is that equilibrium requires an expectation of profit resulting from innovation. If innovation does not lead to profit, firms would not devote effort to innovation, because of its cost. However, if firms devote no effort toward innovation, the likely profits from innovating would be large. The prospect of short run profits must therefore always be present in a competitive economy. A second implication is that policies that tend to reduce short run profits also reduce innovation incentives. The Schumpeterian argument remains a very strong one that receives too little attention from antitrust policy makers. Operationally, it suggests that we should be careful about enforcing the anti-monopoly provisions of the Sherman Act. Aggressive efforts to dissolve businesses with large market shares and high profits as soon as they appear may lead in the long run to a reduction in society’s wealth. A second argument against zealous enforcement is a variation of the preceding one. In certain areas, innovation provides spillover benefits to other firms and other industries. The innovator cannot collect compensation for the spillover benefits. Hence, the attainment of a short run monopoly is the best an innovator can do. Our patent laws already embody this theory. Not every process innovation or neat idea can gain patent or copyright protection. Copyright protects expression, not ideas. Patent protection requires a certain degree of nonobviousness, and there are vast areas that cannot receive such protection, such as graphic designs with functional features, and mathematical formulae. In light of these large gaps in government protection, some incentive must be provided to innovators who fall within them. The prospect of short run monopoly provides this incentive. The third argument is that the prospect of attaining a short run monopoly may spur entrepreneurs to seek out and identify consumer tastes not sufficiently satisfied by the range of products already on the market. The result is the introduction of new, differentiated products. The differentiated-product monopolist enjoys a short term monopoly. As time passes, others will enter the field until economic profits fall to zero. However, because these firms do not face an infinitely elastic demand

III. Further Topics

21

curve, the differentiated-product monopolist will produce at a level below the long run cost-minimizing point. One might think that society loses to the extent that the differentiated monopolist’s production costs exceed the long run minimum. However, the relevant choice is not between the monopolistic competition equilibrium and the perfectly competitive equilibrium, it is between the monopolistic competition outcome and some other outcome that does not satisfy consumer preferences for product diversity.

iii. further topics To this point, I have covered the introductory topics discussed in standard antitrust textbooks. Here I will briefly introduce some relatively new areas of research that usually are not covered in much detail in introductory antitrust courses.

A. Conscious Parallelism Conscious parallelism refers to parallel pricing decisions by firms in oligopolistic industries, carried out in the absence of explicit price-fixing agreements. There are many examples, perhaps the most common provided by the airline industry. Two or three times every year, one of the major airlines announces a price increase or decrease, and parallel actions by competitors immediately follow the announcement. It is hard to escape the conclusion that the major airlines take each other’s actions into account in forming their own price and output decisions. Consciously parallel behavioral patterns have troubled antitrust commentators for many years. The standard model of competition describes every firm as a price-taker, which means that each firm takes the market price as given and then expands production until marginal cost is just equal to price. The only information the competitive firm needs is the location of consumers, the firm’s own costs, and the market price. A firm does not need information on the actions of competing firms. The troubling matter to antitrust commentators is that they do not observe this behavioral independence in markets that exhibit consciously parallel behavior. Firms in oligopolistic industries have a keen interest in the actions of their rivals, and take actions in response to those of rivals. Conscious parallelism presents the issue of whether the government should treat this behavior as independent, or resulting from tacit

22

Economics

agreement. If the former, then Section 1 of the Sherman Act, which outlaws price-fixing, does not prohibit the activity. If the latter, then Section 1 applies, and a violation may be found if the evidence is sufficient to establish agreement. In the 1950 article that initiated the debate among legal academics, James A. Rahl argued that recent research in economics indicated that consciously parallel behavior resulted from independent decision making.18 Rahl referred to research influenced in large part by the work of Edward Chamberlin in the early 1930s.19 However, the notion of interdependence can be traced back to the work of Augustin Cournot in 1838.20 In the Cournot duopoly model, total output equals the sum of the outputs of each duopolist. Thus, if q represents total output, q1 the output for firm 1, q2 the output for firm 2, then q = q1 + q2. The demand schedule specifies the relationship between market price and total quantity. Let p = D(q) be the demand schedule. The revenue of each firm is therefore R1 = D(q)q1 = D(q1 + q2)q1 for firm 1, and R2 = D(q)q2 = D(q1 + q2)q2 for firm 2. Each firm maximizes profit, which means that each firm expands output until marginal revenue equals marginal cost. However, because marginal revenue is a function of q1 + q2, each firm must anticipate the other’s output in the course of making a profit-maximizing output decision. The Cournot duopoly model reveals the manner in which an independent profit maximization decision process leads to interdependent behavior. The behavioral independence implied by the model of perfect competition appears as the number of firms expands toward infinity. As the number of firms grows large, the impact of each firm’s output decision on the market price becomes negligible. In the limit, each firm takes the market price as fixed, and profit maximization requires the firm to expand output until marginal cost equals price.

B. Transaction Cost Economics There are instances in which markets are competitive, in the sense that entry is easy and the market has a large number of competitors, and yet 18

19

20

James A. Rahl, Conspiracy and the Anti-Trust Laws, 44 Illinois Law Review (Northwestern University) 743 (1950). E. H. Chamberlin, Duopoly: Value Where Sellers Are Few, 43 Quarterly Journal of Economics 63–100 (November 1929); The Theory of Monopolistic Competition, Ch. III (Cambridge, MA: Harvard University Press, 1933). For a translation, see Augustin Cournot, Researches into the Mathematical Principles of Wealth (New York: A. M. Kelly, 1960).

III. Further Topics

23

bilateral monopolies arise between contracting parties. The bilateral monopoly may arise as a result of transaction-specific investments. Jobspecific or firm-specific training offers perhaps the best example. Initially, there may be thousands of workers who have the capacity to learn skills to perform a certain job. However, if the training involves firm-specific skills, the relationship deviates from the competitive model after the worker forms an employment relationship with the firm. The employee, armed with specific skills, can threaten to leave, imposing costs on the firm of retraining a new worker and possibly losing business secrets to a competitor. The employer, on the other hand, can threaten to reduce the employee’s wage well below the value of his work. Because the employee has specific skills, he cannot shop them around to other employers in order to maintain leverage at the bargaining table. In a competitive market, prices will adjust to reflect the risk of opportunistic action occurring after contract formation. Thus, if employees know that the employer may opportunistically reduce pay after the training period, they will rationally respond by asking for a higher wage initially. Similarly, if employers think that employees will act opportunistically ex post, they will offer a lower wage. Because the market is competitive ex ante, all expected rents accruing to opportunistic action ex post will be offset by ex ante concessions. Long-term contracts that include provisions allowing the concerned party to monitor and respond to opportunistic actions offer an alternative method of controlling opportunism. These contracts provide some measure of security to the parties’ investments by reducing the risk that one of the parties will exit or threaten to exit the relationship before the other has received an adequate return on its investments. Long-term contracts also insulate the parties from temporary changes in outside opportunities that influence bargaining positions. This model has been developed by the work of several economists,most notably Oliver Williamson.21 The central message for antitrust law is an alternative view of the foreclosure problem. The traditional foreclosure theory holds that long term contracts and vertical integration create barriers to entry by foreclosing access by competitors to important suppliers or important customers. The transaction cost model offers an alternative theory of the incentives for vertical integration, long term contracts, and other devices that have the effect of foreclosing competition. 21

See Oliver E. Williamson, Markets and Hierarchies: Analysis and Antitrust Implications (1975).

24

Economics

The courts often respond in foreclosure cases by requiring the defendant to adopt some less restrictive alternative. Transaction cost economics implies that long-term contracts, or more generally restrictions on competitive bidding during the term of a contract, may sometimes be the least restrictive method of carrying out a production or distribution plan.

C. Information Economics The model of perfect competition assumes that actors have perfect information about the price and quality characteristics of items on the market. This is obviously not true. However, the traditional approach in antitrust enforcement is to question practices that the model of perfect competition cannot explain. A better approach would start from the opposite pole and ask how a workably competitive market might respond to the failure of one of the assumptions of the model of perfect competition. Approaching antitrust policy from this perspective is more likely to generate useful tests for the presence of anticompetitive behavior. Suppose manufacturers offer two types of widgets, high and low quality. One half of the manufacturers produce each quality type. If consumers are perfectly informed, they will pay a price pL for the low quality widget and pH for the high quality widget, where pH > pL. If the consumers cannot distinguish between high and low quality, but do know the relevant probabilities they will offer the average valuation, (pL + pH)/2. This leads to a reduction in demand for high quality widgets and an increase in demand for low quality widgets, resulting in overconsumption of low quality and underconsumption of high quality widgets. The inefficiency could be considerably worse. Suppose some high quality manufacturers, unable to meet their revenue targets, withdraw from the market, leaving a higher proportion of low quality producers. Then the market price falls even more. But this would lead to a further exodus of high quality manufacturers. In the end, all high quality producers would exit, leaving only low quality producers. The scenario I have just described is a version of the adverse selection problem discussed earlier in this chapter. The specific version just described is called the “lemons problem,” based on an insightful article by George Akerlof.22 The name communicates the idea fairly well: if 22

This is a simple description of the “lemons” problem in economics, see George A. Akerlof, The Market for “Lemons”: Quality Uncertainty and the Market Mechanism, 84 Quarterly Journal of Economics 488–500 (1970).

III. Further Topics

25

prospective buyers expect that a substantial percentage of the used cars for sale are defective (i.e.,“lemons”), they will discount their bids accordingly. An owner of a used car, anticipating the discounting of bids, will not sell unless he is trying to get rid of a lemon. Hence, we observe an equilibrium in which low quality goods drive high quality goods out of the market. Return to the widget market. The lemons tragedy need not result if widget manufacturers could credibly communicate information about quality to consumers. There is no need for low quality manufacturers to communicate their quality. If consumers could easily verify quality reports before purchase, low quality manufacturers would have no incentive to report falsely. Under these conditions, high quality manufacturers would assert their quality, low quality manufacturers would either disclose theirs or reveal it by their silence, creating a market of perfectly informed consumers. However, a simple case with easily verifiable quality reports does not describe reality. Economists have distinguished “search” and “experience” goods.23 Consumers can verify the quality of a search good before purchase. Determination of the quality of an experience good requires some use of the product. With respect to experience goods, manufacturers may not report quality truthfully. The short-run gain from providing a false report may exceed the long-term damage to reputation, especially if the manufacturer has no plans to stay in the market long term. In light of these problems, what methods of credibly communicating quality do manufacturers employ? The method that seems most prevalent is the creation of a brand.24 The more valuable the brand, the more the producer loses from misreporting some aspect of product quality. Thus, advertising and the creation and maintenance of brands may be understood as one response to the fact that consumers receive imperfect information about quality.25 For the antitrust policy maker, the next step is to specify the practices that manufacturers might use in order to enhance or to protect brand names. It happens that several practices that have faced challenges for 23

24

25

Phillip Nelson, Information and Consumer Behavior, 78 Journal of Political Economy 311–29 (1970). Benjamin Klein and Keith B. Leffler, The Role of Market Forces in Assuring Contractual Performance, 89 J. Pol. Econ. 615 (1981); I. P. L. Png and David Reitman, Why Are Some Products Branded and Others Not?, 38 Journal of Law and Economics 207–24 (April 1995). See Demsetz, supra note 8, at 49–50.

26

Economics

being inconsistent with the model of perfect competition may serve this purpose.26 One example is product tying. In several antitrust cases, the manufacturer has tied service to the sale of an item. From the entry barrierforeclosure perspective, the practice of tying offers a method of foreclosing access by certain competitors to customers; or of forcing competitors to enter at two levels, that of the tying product and that of the tied product. Alternatively, tying is sometimes described as a mechanism that enables the manufacturer to price-discriminate.27 The information market perspective suggests that the tie-in of service to goods may be one method of protecting reputation, and thereby credibly communicating product quality. The alternative is to tell consumers how to service the item correctly, but there is no guarantee that consumers will follow the manufacturer’s directions, and problems that arise may be attributed to the producer rather than to the consumer. There are several other practices challenged in antitrust courts that may have served as methods of protecting reputation. Antitrust courts have shown reluctance to give serious consideration to arguments that a certain practice was employed to protect brand capital.28 The reluctance seems understandable if one views the model of perfect competition as the null hypothesis in examining any challenged practice. However, the model of perfect competition is probably an inappropriate benchmark for assessing a great deal of market behavior. 26

27 28

For a critique of antitrust law that stresses this point, see Harold Demsetz, How Many Cheers for Antitrust’s 100 Years?, 30 (2) Economic Inquiry 209–17 (April 1992). See Chapter 14 for a discussion of the price-discrimination theory of tying. International Business Machines Corp. v. United States, 298 U.S. 131 (1936) (rejecting goodwill defense for tying purchase of punch cards to computer); International Salt Co.v. United States, 332 U.S. 392 (1947) (rejecting goodwill defense for tying purchase of salt to processing machine).

2 Law and Policy

This chapter introduces some fundamental legal, historical, and policy issues in antitrust law. My aim is not so much to report the details as to give the reader a few simple explanations of the relationship between the Sherman Act and its historical roots, reasons for the enactment of the antitrust laws, and the policy issues recurring in antitrust cases. The focus of this chapter is the Sherman Act. The important parts of the statute are as follows: Section 1: Every contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States, or with foreign nations, is declared to be illegal. Every person who shall make any contract or engage in any combination or conspiracy hereby declared to be illegal shall be deemed guilty of a felony, and, on conviction thereof, shall be punished by fine not exceeding [$10 million] if a corporation, or if any other person, [$350,000], or by imprisonment not exceeding [3 years],1 or by both said punishments, in the discretion of the court. Section 2: Every person who shall monopolize, or attempt to monopolize, or combine or conspire with any person or persons,2 to monopolize any part of the trade or commerce among the several States, or with foreign nations, shall be deemed guilty of a felony [and is similarly punishable].

1

2

In 1990, the maximum fines for violations of the Sherman Act were increased to $10 million for corporations and $350,000 for individuals. 104 Stat. 2880 (1990). Section 8 of Sherman Act tells us that “person” includes corporations and associations under the laws of the United States.

27

28

Law and Policy

i. some interpretation issues A. Conspiracy, Section 1, and Section 2 We can contrast the two important sections of the Sherman Act by comparing the elements of a conspiracy charge. Section 1 is a conspiracy statute. Conspiracy traditionally consists of three proof elements. The first is proof of an agreement and intent to violate the law. The second is the so-called duality requirement, that there were at least two parties to the agreement. The third is a famous nonrequirement: the government does not have to prove that the parties either carried out or had the power to carry out the criminal act. Since the word conspiracy appears in Section 1 of the Sherman Act, the foregoing elements are part of antitrust conspiracy doctrine. Thus, in a price-fixing case, the plaintiff must prove intent and agreement to fix prices. Second, the plaintiff must show duality. But what does duality mean in this setting? If it means simply that there must be at least two individuals involved, then a plaintiff could bring a price-fixing suit against two members of a partnership. However, the courts have decided that the members of a partnership are not involved in a conspiracy to restrain trade; the partnership is treated as a single entity.3 Surprisingly, the Supreme Court has never articulated the degree of separation required to satisfy the duality element.4 It appears that all one can say is that the parties must be sufficiently separate that it would be inappropriate to view them as members of the same business. Third, and most interesting, Section 1 does not require the plaintiff to prove that the parties had sufficient control over the market to actually have a significant impact on the market price. Thus, a group controlling only 5 percent of the relevant market could face a price-fixing charge, and would not be able to rely on the defense that they are too small to affect the market price. Compare Section 2 along the same lines. First, Section 2 requires that the plaintiff prove intent, but the plaintiff does not have to demonstrate that an agreement existed. Second, Section 2 does not require duality, it 3

4

United States v. Addyston Pipe & Steel Co., 85 F. 271, 281–282 (6th Cir. 1898) (setting forth five types of contract, one of them the partnership agreement, that were upheld under the common law of trade restraints, and asserting that such contracts were outside of the reach of Section 1). For the most recent effort, see Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752 (1984).

I. Some Interpretation Issues

29

applies to unilateral conduct. Third, the plaintiff must show that the defendant has market power.5

B. Jurisdiction The statute refers to interstate commerce. If interpreted literally, this would imply that the statute applies only when an anticompetitive act affects the market in goods that are traded across state lines. The Supreme Court initially adopted this interpretation: in United States v. E. C. Knight Co.,6 the Court held that manufacturing was not interstate commerce and thus the statute did not cover a monopoly of sugar manufacture. But as the Court gave broader scope to Congress’s power to regulate trade, it also extended the reach of the Sherman Act. Now, it is virtually settled that the Sherman Act applies even to anticompetitive acts that affect only the market for a good within a single state. Ignoring the issue of state boundaries, there is a deeper jurisdictional problem. Suppose a group arranges a boycott that has the effect of excluding one small firm from the relevant market. For example, suppose a boycott prevents one doctor or one lawyer from practicing in his or her area of specialty, and that the relevant geographic market has many other local specialists. This boycott probably would not have a perceptible impact on the market. If there are one hundred local practitioners and one has been excluded, the market pretty much remains competitive.7 Can a court claim jurisdiction over this dispute under the Sherman Act? The Supreme Court answered this affirmatively in Summit Health, Ltd. v. Pinhas,8 but it was a 5-4 decision.9

5 6 7

8

9

U.S. v. Grinnell Corp., 384 U.S. 563 (1966). 156 U.S. 1 (1895). Assuming the market was competitive initially. However, there are reasons to doubt the competitiveness of markets in professional services. Though the number of competitors may be large in these markets, informational disparities make it difficult for consumers to shop intelligently among different providers. See Richard Craswell, Tying Arrangements in Competitive Markets: The Consumer Protection Issues, 62 Boston Univ. Law Review 661 (1982). 500 U.S. 322 (1991). The issue was whether the interstate commerce requirement of antitrust jurisdiction is satisfied by allegations that the defendants conspired to exclude plaintiff, an eye surgeon, from the market for eye surgery services in Los Angeles because he refused to follow an unnecessarily costly surgical procedure. The dissenters, all relatively young justices at the time, were O’Connor, Scalia, Kennedy, and Souter. Justice Thomas, who joined the Court after the Summit Health decision, would probably cast his vote with this group.

30

Law and Policy

The jurisdictional issue resurfaces occasionally in antitrust disputes, sometimes dressed in different legal terms. In Chapter 9, I describe it as the “Klor’s paradox,” after the case that frames the issue at its most general level.10 In Klor’s the defendants, instead of referring to the interstate commerce requirement, argued that the Sherman Act required some showing of public harm for the plaintiff to maintain his action. Whether one talks about public harm or interstate commerce the underlying issue is the same: if an exclusionary act has no perceptible impact on the market, should the plaintiff still have a valid Sherman Act claim? We will consider this question in more detail in Chapter 9. For now, ask yourself the following: What would happen if a defendant in a tort action said that tort law aims to deter reckless behavior, not to compensate specific victims, and thus the victim should not collect damages until he can show that failure to do so would lead to a significant reduction in precaution by potential injurers? How is this question linked to the jurisdiction issue in antitrust?

C. The Clayton Act The Clayton Act declared the following acts illegal (but not criminal): (1) price discrimination (Section 2, later the Robinson-Patman Act of 1936); (2) tying and exclusive dealing contracts (Section 3); (3) corporate mergers that tend to result in monopoly (Section 7); and (4) interlocking directorates, that is, common board members among competing companies (Section 8). The Act qualified each prohibition with the condition that the practice becomes illegal only where the effect may substantially lessen competition or tend to create a monopoly in any line of commerce. Another important provision is Section 4, which tells us who can sue: “[A]ny person who shall be injured in his business or property by reason of anything forbidden in the antitrust laws may sue in any district court of the U.S. . . . and recover treble damages and the cost of suit, including a reasonable attorney’s fee.”

ii. enacting the antitrust laws There are two reasons to look into the common law background of the Sherman Act. First, most of the legislators supporting the Act referred to the common law. Supporters claimed that they were seeking federal 10

Klor’s v. Broadway-Hale Stores, 359 U.S. 207 (1959).

II. Enacting the Antitrust Laws

31

enforcement of the common law’s prohibition of contracts in restraint of trade.11 Referring to language of a draft similar to the Sherman Act, Senator Sherman himself said that the act “does not announce a new principle of law, but applies old and well-recognized principles of the common law.”12 Is this a valid description? The second reason is found in Supreme Court opinions that refer to the common law as a basis for distinguishing the Sherman Act from other statutes. For example, in both Connally v. General Construction Co.13 and Cline v. Frink Dairy Co.,14 the Court held a state statute void for uncertainty, even though supporters of the statute argued that it was no more uncertain than the Sherman Act. The Court distinguished the Sherman Act on two grounds. First, the Sherman Act had been deemed not too uncertain to satisfy due process requirements in Nash v. United States.15 Second, since the Sherman Act’s “rule of reason” (see Chapter 5) embodied rules declared by common law precedents, it afforded a definite and certain objective standard. Of course, many other federal court antitrust opinions have referred to common law precedents, probably the most influential is United States v. Addyston Pipe & Steel Co.16 Given that modern courts perceive some connection between the Sherman Act and the common law, the interesting question is the extent to which the Act embodies the common law.

A. Common Law Background Four lines of earlier law have played some part in shaping modern antitrust law: (1) prohibitions against “interference with markets;”17 (2) case law on the enforcement of contracts in restraint of trade; (3) the application of the law of criminal conspiracy to combinations in restraint of trade; and (4) the common law of monopoly. 11

12 13 14 15 16 17

John C. Peppin, Price-Fixing Agreements under the Sherman Anti-trust Law, 28 California Law Review 297, 306 (1940). See, for example, id. at 306 n.29. 269 U.S. 385 (1926). 274 U.S. 445 (1927). 229 U.S. 373 (1913). For discussion, see Chapter 4. 85 F. 271 (6th Cir. 1898), aff’d, 175 U.S. 211 (1899). For discussion, see Chapter 5. Donald Dewey, Monopoly in Economics and Law 112 (1959). The discussion in this section borrows heavily from Dewey’s book, and from William Letwin, Law and Economic Policy in America (1965). For another very thorough source on the common law background, see Hans B. Thorelli, The Federal Antitrust Policy: Origination of an American Tradition (Baltimore, MD: Johns Hopkins University Press, 1955).

32

Law and Policy

1. Market Interference. Let us start with prohibitions against market interference. This category refers to statutes covering the related acts of forestalling, regrating, and engrossing.18 In today’s language, we would refer to much of this as arbitrage: buying low in one market and selling high in another (in a different location or later time). The statutes, which were generally of a protectionist nature, are really the English ancestors of modern statutes prohibiting below-cost pricing (and perhaps rules against predatory pricing). The main difference between these early statutes and the Sherman Act is that courts interpret federal antitrust law today as aiming to enhance competition, while the early market interference statutes served largely to suppress competition. A middleman who purchases the supply of corn in town A, and sells it in town B will obviously attempt to sell it at a higher price than he paid for it. The early statutes aimed to prohibit these transactions on the ground that they caused an artificial increase in prices.19 But if town B has a higher demand for corn than town A, the middleman has merely helped the market transfer corn to people who place the highest value on it, which does not harm competition. Indeed, it promotes competition by encouraging the entry of corn sellers. The market interference statutes were repealed in 1772.20 The United States did not exist as a separate country before then, so these statutes probably did not have a direct impact on American federal antitrust law, which would not begin to develop until 1890. And even if they did, the purpose of the early market interference statutes is apparently at odds

18

19

20

For a detailed discussion, see William Letwin, supra note 17, at 32–39. Forestalling is “[a]n unlawful attempt to raise prices; buying anything ‘before the due hour’ or ‘pass[ing] out of the town to meet such things as come to market.’ ” Id. at 33–34. Regrating: “Retailing, buying in bulk and selling in small lots.” Id. at 34. Engrossing: “to buy crops in the field before they were harvested or at least before they were ready to come to the market.” Id. See Thorelli, supra note 17, at 16; Roman Piotrowski, Cartels and Trusts: Their Origin and Historical Development From the Economic and Legal Aspects 145–52 (1933). 12 Geo. 3, c.71 (1772). In a passage discussing the common law treatment of monopolies, the Supreme Court described the reasons for repeal as follows: “From the development of more accurate economic conceptions and the changes in conditions of society, it came to be recognized that the acts prohibited by the engrossing, forestalling, etc., statutes did not have the harmful tendency which they were presumed to have . . ., but, on the contrary, such acts tended to fructify and develop trade. See the statutes 12th George III., chap. 71, enacted in 1772, and statute of 7 and 8 Victoria, chap. 24, enacted in 1844, repealing the prohibitions against engrossing, forestalling, etc., upon the express ground that the prohibited acts had come to be considered as favorable to the development of . . . trade. Standard Oil v. U.S., 221 U.S. 1, 54 (1911).

II. Enacting the Antitrust Laws

33

with that of the Sherman Act. The Sherman Act has never been interpreted as aiming to suppress arbitrage by middlemen. 2. Restraint of Trade. The restraint of trade cases were contract actions in which the defendant, who had breached the contract, asserted the defense that the contract violated the public policy against restraining trade. Example: a contract between a master and an apprentice stipulates that the apprentice will not set up in competition with the master within a year after completing his training. The apprentice sets up shop in competition with the master within a year of completion of training. Some textbooks state that in the earliest cases, courts held these agreements unenforceable. One of the cases cited in support of this claim is The Dyer’s Case,21 where the court denied an attempt to collect on a bond for John Dyer’s breach of his agreement not to use his art (dying) within the town for half a year. However, the early cases that condemned these agreements involve conditions that master craftsmen imposed on adolescent apprentices,22 which is true, for example, of The Dyer’s Case. The courts may have denied enforcement in these cases largely because the apprentice was not competent to enter into an enforceable contract when the parties made their agreement. In short, some of the early cases holding noncompetition agreements unenforceable may have had less to do with restraint of trade doctrine than with the general question of when a party is competent to enter a contract.23 A later case, Mitchel v. Reynolds,24 summarized restraint of trade doctrine as follows: covenants not to compete may be justified if reasonable and ancillary to some principal (legitimate) transaction and if limited in time and space. Mitchel also required a showing that the plaintiff had provided “good and adequate consideration.” The rule of Mitchel v. Reynolds is sometimes referred to as the Rule of Reason, and it remains the operative doctrine in restraint of trade actions today. Reynolds had a bakery. Mitchel leased it from him for seven years on condition that Reynolds would not reenter the local bakery market during that period. Reynolds reentered the market, in competition with 21 22 23

24

Year-Book, 2 Hen. V, 5f. (1414). Dewey, supra note 17, at 124 n.1. This is the interpretation offered in Dewey, supra note 17. An alternative interpretation is provided by Thorelli, supra note 17. Thorelli suggests that the common law of trade restraints became more flexible over time as courts began to recognize the economic benefits provided by the restraints, id., at 17–18. 1 P. Wms. 181, 24 Eng. Rep. 347 (K.B. 1711).

34

Law and Policy

Mitchel. Mitchel sued to collect his bond, and Reynolds said that the contract was unenforceable because it restrained trade. The court upheld the contract because the parties were competent, had provided consideration, and the restraint was reasonable. Although the court never articulated precisely why it was reasonable, it is not hard to see why. Few bakers would have paid much to lease the shop from Reynolds if they would have to put up with competition from him immediately after moving into the shop. If Reynolds had established a good reputation, they would likely lose in such a contest. But if almost no one would pay a good price for the shop when Reynolds decided to exit, what incentive would Reynolds have to maintain the quality of his shop near the end of his career?25 And if there is no incentive to maintain quality near the end, what point is there in attempting to establish a reputation for quality? Mitchel v. Reynolds is important because it was the first of the trade restraint cases to make a serious effort to state the doctrine in this area, and the formulation established has remained intact for more than two hundred years. Of course, there have been disputes over the meaning of the doctrine. One of the most important is the limiting case of an unreasonable contract. Mitchel v. Reynolds states one example of an unreasonable contract: one requiring Reynolds not to practice anywhere in England for the rest of his life. Should modern courts interpret that as a fixed boundary? Should a modern court invalidate a contract that prevents a firm from operating anywhere in England even though it operates, with the aid of modern technology, in all other areas of the globe? We can state the question more generally: does the common law rule of reason embody fixed principles, making certain actions unreasonable per se? Over time, courts have generally come to the conclusion that it does not. The rule of reason is a flexible, cost-benefit test that takes modern conditions into account in assessing the reasonableness of a trade restraint. To settle on fixed categories of reasonableness or unreasonableness would be inconsistent with the structure of the test. 25

On the general problem of incentives and reputation over time, see David M. Kreps, Corporate Culture and Economic Theory, in J. Alt and K. Shepsle (eds.), 90–143, Perspectives on Positive Political Economy (Cambridge: Cambridge University Press, 1990). Kreps shows that if reputation is a tradeable asset, then even an individual with a finite career may have an incentive to maintain his reputation for quality at the end period of his career. One can view the issue in Mitchel v. Reynolds as whether the court would allow parties to create a market in reputation.

II. Enacting the Antitrust Laws

35

3. Criminal Conspiracy. Another potential source for modern antitrust law is criminal conspiracy doctrine. Over the 1700s, “Parliament repeatedly outlawed attempts to exact unjust or unreasonable wages and prices”26 – referred to as Combination Laws. Also, under conspiracy doctrine, concerted action to bring about the commission of unlawful acts would support a conspiracy charge, even in the absence of a crime.27 Putting these together suggests that during the period the Combination Laws were in effect, business and labor cartels operating in England or in the American colonies risked possible conviction of conspiracy to fix prices. This is a reasonable inference, but the cases provide little support. The government mostly prosecuted working men who had attempted to raise their wages by obviously illegal means, such as threats and violent attacks against individuals and their property.28 Although the reports often fail to give a clear picture of what happened in these cases, it appears that none of them involved peaceful attempts to combine in order to raise prices on commodities, and only one English case may have involved a peaceful combination to raise wages.29 The doctrine of conspiracy applies to combinations that have an illegal purpose or use illegal means. Because the early cases typically involved clearly illegal methods, the case law has little in it suggesting that courts would find mere combination to raise prices unlawful. In addition, because they were internally inconsistent and seldom enforced,30 the Combination Laws did not clearly lead to the establishment of such a doctrine in the common law. In any event, the Combination Laws were repealed in 1824, more than sixty years before the passage of the Sherman Act. 4. Common Law of Monopoly. A fourth potential source for the law embodied in the Sherman Act is the common law of monopoly. The very 26 27 28 29

30

Dewey, supra note 17, at 116. Id. Id. But even this is unclear. Dewey reports that in The King v. Journey-men Taylors of Cambridge, the court upheld the defendant’s conviction on a conspiracy charge while conceding that their demands might not be unreasonable and the means employed to advance them were not unreasonable per se, id, at 116. In a footnote, Dewey notes that the accuracy of the report has been contested, id. A survey of the American labor cases reports that none of the early prosecutions involved peaceful combinations. Edwin E. Witte, Early American Labor Cases, 35 Yale L. J. 825, 825–28 (1926). Dewey, supra note 17, 115–18.

36

Law and Policy

existence of this common law is questioned by William Letwin, who states that it was “invented largely by Sir Edward Coke.”31 However, a little thought about the relationship between law and markets will reveal why it may seem that Coke invented the law and why that perception may be wrong. Monopoly is a short run phenomenon in markets with easy entry. A monopolist who charges the profit-maximizing price will encourage entry. Entry will continue until the growing number of firms drives profits to a level that just compensates the owners of capital for their forgone alternatives. Generally, stable or long-run monopoly status occurs when governments exercise control over entry into markets.32 It is therefore not surprising that the common law of monopoly arose largely in opposition to attempts by the state to restrict entry. Coke’s first mention of that common law regards patents; specifically, the invalidity of certain patents granted by the king. That the common law would set itself against monopolies should not be so novel an idea as to need Coke’s invention. Not many arguments favor a state-granted monopoly, unless designed, like our patent laws today, to encourage innovation. Outside of modern patents, protected monopolies lead to higher prices, lower quality, and generate an unrelenting push by interested parties to gain control over them. The enormous sums spent to obtain control over monopolies, observed in every country today that restricts entry into markets, is probably the largest source of economic waste generated by the state’s creation and maintenance of monopolies. Darcy v. Allen (The Case of Monopolies)33 provides the best evidence of a common law of monopolies. Queen Elizabeth granted Darcy a patent covering the manufacture and importing of playing cards. Allen made and sold some playing cards and Darcy brought an infringement suit. The court held the monopoly void because it was “against the common law.”34 The patent deceived the Queen, the court said, because

31

32

33 34

William Letwin, Law and Economic Policy in America: The Evolution of the Sherman Antitrust Act 19 (1965). For the other view, see Thorelli, supra note 17, at 21–24, who argues that the common law of monopolies had a solid basis in common law rules protecting the freedom of individuals to set up businesses and enter trades. For a history of monopolization that is consistent with this claim, see Piotrowski, supra note 19. 11 Coke 84, 77 Eng. Rep. 1260 (K.B. 1603). 11 Coke at 86a, 77 Eng. Rep. at 1266.

II. Enacting the Antitrust Laws

37

it was intended to enhance social welfare, but the grant only enriched Darcy while reducing social welfare. The relationship between the common law of monopolies and the Sherman Act is weak. The common law of monopolies attempts to constrain the power of the state to arbitrarily restrict entry into markets. The Sherman Act aims largely to restrict efforts by private firms to monopolize markets.

B. Summary We return to the question: Did the Sherman Act attempt to seek federal enforcement of common law principles? It seems that no common law action for conspiracy to restrain trade existed. There was a doctrine governing trade restraints, but it began as and remains part of contract law. There was a criminal conspiracy doctrine, but it had not been applied to mere business agreements that did not either aim to violate a criminal law or did not require such a violation for execution. It is hard to escape the conclusion that the Sherman Act attempted to take an area of private law, and to expand public enforcement into that realm. To do so required a new form of conspiracy. There were pieces of this new creature existing in ancient common law, but the whole creation was never observed before Congress brought it to life in 1890.

C. What Led to the Sherman Act 1890? One can take either a “public interest” or “public choice” view of the enactment of the Sherman Act. The public interest view holds that Congress was sincerely trying to do something to bring monopolies under control, to inject more competition into the economy, and to reassert the value of the individual consumer over that of the large conglomerate firm. The public choice view suggests that legislators used the Sherman Act in an effort to keep themselves in office. Implicit in the public interest view is the theory that the market itself cannot maintain a competitive system. According to the model of perfect competition presented in Chapter 1, when a monopolist is making profits, entry occurs, reducing the monopolist’s profits and restoring a competitive equality or balance among suppliers to the market. The model of perfect competition implies that there is no need for the government to ensure and maintain an atomistic market structure by constraining large

38

Law and Policy

businesses. The public interest view implies that the government may need to intervene in order to restore and maintain competition. The public interest view is set out in a number of antitrust textbooks. Here I will not attempt to explain the passage of the Sherman Act; I will only state a few reasons to question the public interest view. Several things came together in the late 1800s. Many individuals, including legislators, considered the railroads and the trusts (Standard Oil, American Tobacco) too powerful. The transportation and communication revolutions of the middle 1800s led to large businesses, which created jobs but also destroyed many through business expansion. Congress received numerous petitions to put the trusts and the railroads under the regulation of the federal government.35 At the same time, the trusts were influential in Congress, having pushed through a tariff bill very much in their favor in the same session that passed the Sherman Act.36 Should one believe that the same Congress that protected the trusts at the expense of the consumer, by enacting tariff legislation, would also seek to protect the consumer from the pricing tactics of the trusts? An unwavering adherent to the public interest view could offer a theory that reconciles the conflicting actions of the Congress that passed the Sherman Act. The probable explanation is that Congress sought to protect American workers from unfair competition from abroad, and also sought to protect the American consumer from unfair business practices at home. If the legislative process could not be influenced by special interests, this explanation would be acceptable. However, because the legislative process is influenced by special interests, it is worthwhile to at least consider alternative assessments of the motivations of Congress in 1890. The cynical view is that the Sherman Act was a harmless way of appeasing public demands for regulation, and at the same time allowing members of Congress to garner support for and deflect criticism regarding the protective tariff legislation.37 It was 35 36

37

Dewey, supra note 17, at 142. The McKinley Tariff Act of 1890 increased tariffs on wool, silk, cotton, agricultural products, and tin goods. The bill repealed the raw sugar duty but kept prohibitively high duties on steel rails, structural steel and iron, and copper. Sidney Ratner, The Tariff in American History 36–37 (1972). Passed by Republicans, the bill sought primarily to “counteract the dissatisfaction of western farmers who might be turning toward the rival Democratic and Populist parties.” Id. at 37. Thomas J. DiLorenzo, The Origins of Antitrust: An Interest-Group Perspective, 5 International Rev. of Law and Economics 73–90 (1985). This thesis is also suggested in Dewey, supra note 17, at 143. It appears that there were identifiable interest groups that promoted antitrust legislation. Perhaps the most prominent were rural cattle growers and

II. Enacting the Antitrust Laws

39

harmless because under the then-prevailing view of Congress’s power under the Commerce Clause,38 it was understood that enforcement of the statute required the aid of the states, primarily through state-level antitrust legislation. Any state would be free after the passage of the Sherman Act to devote a substantial part of its resources to antitrust enforcement, but at a cost of a reduction in business activity, something states have never been keen to do on their own.

D. What Led to the 1914 Legislation? The Clayton and FTC Acts of 1914 are easier to explain than the Sherman Act. They were both a response to the decision of the Supreme Court in Standard Oil Co. v. United States. Then-Chief Justice White had finally gathered a majority in support of his view that the Sherman Act condemned only unreasonable restraints of trade. The alternative view was that it condemned all restraints save a fixed set of contracts, such as partnership agreements, deemed harmless long before the passage of the Act. Under the rule of reason announced in Standard Oil, the Sherman Act condemns only unreasonable restraints of trade, as determined by purpose, power, and effect. Standard Oil left all sides unhappy. Pro-regulation interest groups thought it would emasculate enforcement. Pro-business interests thought it would give the government a free hand, since the rule of reason provided an ambiguous legal standard.39 As a result, the Clayton Act set forth a set of specific violations. If the specified acts tended to reduce competition, they counted as violations. Congress did not want courts to apply a general rule of reason test to these enumerated violations, and established the Federal Trade Commission with the authority to aggressively enforce the antitrust laws. Section 5 of the FTC Act, as now understood, permits the FTC to

38 39

butchers whose livelihoods were threatened by the growth of large meat-processing facilities. See Gary D. Libecap, The Rise of the Chicago Packers and the Origins of Meat Inspection and Antitrust, 30 Economic Inquiry 242–62 (1992); Donald J. Boudreaux, Thomas J. DiLorenzo, and Steven Parker, Antitrust before the Sherman Act, in The Causes and Consequences of Antitrust (Fred S. McChesney and William F. Shugart eds.), pp. 255–70 (Chicago and London: University of Chicago Press, 1995). See the discussion titled “Jurisdiction” in this chapter. I have left out special interests, but there is evidence that they played a part. See Robert B. Ekelund, Jr., Michael J. McDonald, and Robert D. Tollison, Business Restraints and the Clayton Act of 1914: Public- or Private-Interest Legislation?, in McChesney and Shugart, supra note 37, at 271–86.

40

Law and Policy

prosecute and find violations (of Section 5) in the event of discovery of anticompetitive practices that the Sherman or Clayton Acts may be unable to punish, due to the difficulty of meeting proof requirements under those statutes.

iii. what should antitrust law aim to do? Should the antitrust laws seek to enhance competition by maintaining an atomistic structure, in which numerous small businesses compete, or should it aim to maximize consumer welfare? A crude but fair summary of the development of antitrust law is that courts have shifted, though gradually, from an adherence to the former to an acceptance of the latter. The first Supreme Court antitrust opinion, U.S. v. Trans-Missouri,40 contains references to the desirability of maintaining a market in which numerous small businesses compete, and many later opinions have relied on such policy statements as justifications. Important modern cases, however, have clearly placed the consumer welfare goal ahead of the atomism goal.41 One can make plausible arguments for protecting small businesses from being pushed out by larger ones,42 especially if one considers the contribution of such businesses to employment growth,43 innovation, and other beneficial externalities. The notion is that if small businesses contribute disproportionally to innovation, or some other desirable end, then it may be a good policy to subsidize them or to protect them from competition. Such protection would merely compensate these businesses for the benefits they externalize to others. The problem with this argument is simple: where should the government draw the line on protection? Admit one case, and it becomes dif-

40 41 42

43

166 U.S. 290 (1897). For discussion, see Chapter 5. The best example is Continental T.V. Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977). For a critical review of the arguments for protecting small businesses (in the international trade context), see Paul R. Krugman, Is Free Trade Passe? 1 Journal of Economic Perspectives 131–44 (Fall 1987). Krugman’s discussion is equally applicable to the antitrust debate. The employment growth externality has been more or less assumed to be valid. However, a recent study demonstrates that a company’s contribution to employment growth is closely related to its share of total employment – in other words, large employers tend to make a greater contribution to employment growth than do small employers. See John Haltiwanger (University of Maryland), Scott Schuh (Federal Reserve Board), and Steven J. Davis (University of Chicago), Gross Job Flows in U.S. Manufacturing, published by the Census Bureau.

III. What Should Antitrust Law Aim to Do?

41

Figure 2.1 P0 = market price with competition restricted P1 = price under competition

ficult to deny others; and it is certain that one can reach a stage where protection from competition reduces society’s welfare. To see this point, consider the classic argument for free trade. In Figure 2.1, consumers gain area A + B as a result of competition from a more efficient supplier. The less efficient suppliers lose area B. The net gain from allowing efficient competitors to displace less efficient is positive. The classic case for free trade has been criticized on several grounds. Externalities (innovation or employment benefits of small companies), wealth effects on tastes,44 and increasing returns may imply that it is better to protect a small business in order to let it reach efficient scale. In theory, these are valid criticisms. However, the relevance of these criticisms and the drawbacks attending some of their implications should be kept in view. First, there is neither a sufficiently well-developed theory nor a sufficiently welldeveloped body of empirical research that would allow us to say whether the value of the external benefits enhanced by protection outweigh the 44

The argument associated with Figure 2.1 is that the gain to winners, A + B, exceeds the loss to losers, B. This is known in the economics literature as Kaldor-Hicks efficiency. However, the price reduction shown in the figure generates an increase in the real wealth of consumers. That increase in wealth may in turn affect preferences (wealth effects), altering the demand curve. If wealth effects are significant, the Kaldor-Hicks efficiency test becomes uncertain.

42

Law and Policy

consumer welfare benefits of unrestrained competition. The current state of the literature does not allow us to answer this question. This is an important weakness in the argument for protection; for if there is no way to distinguish important from trivial externalities generally, then there is no good reason for thinking that each supposedly beneficial externality enhanced by restricting competition is not offset by a negative externality created or another positive externality forgone by not permitting unrestrained competition. Knowing nothing of the relative importance of various externality arguments, the sensible inferential rule is to assign equal weights to each claim. But this leads to the answer that it is best to rely on the classical model of free trade. There are so many plausible externality arguments that in the absence of solid empirical evidence, the appropriate assumption is that they cancel each other out. For each innovation incentive that protection provides, there is an innovation disincentive provided, and so on. Second, there is a general procedural problem that externality claims introduce, especially when presented as arguments for protecting small businesses from competition: who decides when an externality is important enough to justify protection from competition? If it is a human making the decision, then he or she is subject to error, and will be inundated with requests for protection from businesses in every sector of the economy. How can these decisions be made in a disinterested, objective manner? One might well say that who decides is not as important as making sure that such decisions are made by someone and are subject to review, that is, that due process is provided. But this misses the point that the market allows us to avoid this issue altogether. The market is an objective, disinterested mechanism that answers the question of who should be allowed to produce what, at which price, which quantity, and so on. To put such decisions in the hands of individuals, which the externality defense of protection would require, substitutes individual decisions for market outcomes. Efforts to experiment with this approach on a large scale have failed spectacularly.

3 Enforcement

Enforcement of the federal antitrust laws occurs through the actions of government agencies and the lawsuits of private parties. In this chapter, I set out the mechanics of these enforcement processes and consider whether the level of enforcement is optimal from an economic perspective. I also examine some predictability and fairness issues generated by the current enforcement framework. If the antitrust laws were designed to provide the socially optimal level of deterrence, then they would discourage only those acts that reduce society’s wealth. I will say that the antitrust laws overdeter if they discourage conduct that on balance increases society’s wealth. Similarly, I will say that the laws underdeter if they fail to discourage conduct that reduces society’s wealth. I conclude, in this chapter, that the penalty provisions and the rules governing damages, when viewed in isolation, are likely to underdeter covert anticompetitive activity, such as price-fixing. However, it is difficult to say whether the antitrust laws, viewed in their entirety, underdeter or overdeter. The overall picture is complicated, because while some features of the enforcement structure suggest the laws underdeter, other features suggest the laws may overdeter.

i. optimal enforcement theory Assume that the objective of the antitrust laws is to maximize consumer welfare. What is the appropriate fine for an antitrust violation?1 That 1

I follow the standard approach in the literature by examining fines. Since the Sherman Act already imposes fines for antitrust violations, it makes sense to consider the

43

44

Enforcement

question received its clearest answer in an article by William Landes,2 who relied on the enforcement theory of Gary Becker.3 Under the assumption that enforcement is costless and that the government will identify and punish all antitrust violators with probability one, the consumer welfare-maximizing fine is easy to state: the optimal fine equals the sum of the portion of deadweight loss borne by consumers and the monopoly transfer. This is the optimal fine level because it “internalizes” to the monopolist all of the external costs of monopoly, given the assumptions of costless and certain punishment.4 A fine set at a level higher than optimal deters some monopolizing activity that actually increases society’s wealth, and one set lower than the optimal level fails to deter some monopolizing activity that decreases society’s wealth. Figure 3.1 illustrates this argument, in the simple case with constant marginal cost. Suppose the process that enables a firm to monopolize a market (e.g., a merger) also reduces its production costs, so average cost falls from AC0 to AC1 as shown in Figure 3.1. The deadweight loss after the merger is represented by the area ECG in Figure 3.1. The monopoly transfer is represented by the area GDAE. If the fine is set at the optimal level, which is equal to the sum of areas ECG and GDAE, then as long as the rectangular area GBHD exceeds the deadweight loss triangle ECG, the monopolizing entity will have an incentive to continue in its activity. Society gains because the additional wealth going to owners of the monopoly exceeds the additional welfare loss to consumers. This argument becomes more complicated if we consider the possibility that consumers might collect damage awards against the monopolist. Suppose, in particular, that the fine is set at the optimal level (i.e., the fine is equal to the monopoly transfer plus the deadweight loss) and that it is transferred from the monopolist to consumers. Suppose, further,

2

3

4

structure of optimal fines. The more general question is whether fines are preferable to subsidies or some other mechanism for influencing the monopolist’s incentives. I do not consider that question here. William M. Landes, Optimal Sanctions for Antitrust Violations, 50 University of Chicago Law Review 652 (1983). See also Gregory Sidak, Note, Rethinking Antitrust Damages, 33 Stanford L. Rev. 329 (1981). Gary S. Becker, Crime and Punishment: An Economic Approach, 76 J. Pol. Econ. 169–217 (1968). A proof of this claim is provided in the appendix of this chapter. The “costless” punishment assumption implies that the state incurs no cost (on the margin) in imposing the penalty, regardless of its size.

I. Optimal Enforcement Theory

45

Figure 3.1

that consumers know beforehand that they will receive the monopoly transfer plus the deadweight loss. Under these assumptions, consumers will discount the costs of dealing with the monopolist. Their consumption plans will not be affected as in the standard monopoly discussion (where they reduce consumption in response to a monopolistic price increase). Furthermore, if the monopolist knows that he will have to compensate consumers, then the possibility of fines will dampen his production incentives further. Consider the special case in which both the monopolist and the consumer use an accurate assessment of damages when deciding how much to produce and how much to consume, respectively. To the monopolist, the real price charged to consumers is the price of the good less the expected per-unit damage payment. To the consumers, who are by assumption aware of the potential damage award, the perceived price is the same. The result is that damages have no effect

46

Enforcement

on behavior: the monopolist continues to produce at the monopoly output level.5 The ultimate effect of private damages on the monopolist’s decision is difficult to state, because it depends on assumptions as to whether the parties accurately predict the cost of damage awards to the monopolist, the likelihood that suit will be brought, and the cost of litigation. It is sufficient here to note that Figure 3.1 reflects the assumption that the consumer and the monopolist do not take expected damage awards into account ex ante. Now let us return to the optimal fine question. To provide a benchmark for comparison with actual fines, let us consider the simple case with linear demand schedules, and the marginal cost and average cost curves flat and equal to each other. It is straightforward to show that under these assumptions transfer + consumer deadweight loss =

Ê 3ˆ( transfer ). Ë 2¯

If the marginal cost curve is upward-sloping rather than flat, this expression overstates the total external loss (assuming monopoly output is the same in both cases). As we consider steeper marginal cost curves, the consumer deadweight loss portion shrinks to zero. Thus, (–32 )(transfer) is an upper bound on the optimal fine and the transfer itself provides a lower bound for the optimal fine. If the marginal cost curve is flat over the relevant output range, the upper bound is an accurate measure of external loss. I will use the upper bound through much of the discussion below as a benchmark. Now suppose enforcement is costly and apprehension by the state less than certain. Specifically, let the state’s cost of punishment (given apprehension) be C and the probability of apprehension (and conviction) be p.6 Then the optimal fine is 5

6

See Jonathan Baker, Private Enforcement and the Deterrent Effect of Antitrust Damage Remedies, 4 Journal of Law, Economics, and Organization 385 (Fall 1988). To see this in Figure 3.1, suppose the demand curve shifts up by the expected per-unit damage award, reflecting the subsidy consumers receive in the form of compensation. At the same time, the marginal cost curve for the monopolist shifts up by the same amount. It is straightforward to show that the monopoly pricing equilibrium is the same. I am assuming that the enforcement process works without error. Thus, every apprehension is assumed to lead to a conviction. For simplicity, I will refer to p as the probability of apprehension, though what I really have in mind is the probability of apprehension and conviction.

II. Enforcement Provisions of the Antitrust Laws

47

Ê 1ˆ( transfer + consumer deadweight loss) + C . Ë p¯ We must multiply the transfer and deadweight loss components by –1p in order to “correct” for the likelihood that the typical monopolist will get away with his conduct. An upper bound on the optimal fine is therefore Ê 3 ˆ( transfer ) + C Ë 2p ¯ The state’s cost of punishment must be added to the fine in order to force the monopolist to bear all of the costs generated by his activity. For those monopolists who are apprehended, the additional charge C is appropriate because their apprehension forces the state to spend C in order to punish them. We can propose several variations on this formula, depending on the assumptions one chooses to work with; however, the basic implication remains the same: internalize the transfer and deadweight loss, plus incremental enforcement costs. Can this rule be implemented? Surprisingly, the answer may be yes. Many large antitrust cases now involve experts trained in econometrics who could provide reliable estimates of the transfer and forgone consumer surplus components. The enforcement costs will introduce an additional measurement difficulty. In the section below, I will use the theory presented here to see whether the actual features of antitrust enforcement are consistent with those of an optimal enforcement framework.

ii. enforcement provisions of the antitrust laws A. Who Enforces? Who enforces the antitrust laws? The leading figure is the Antitrust Division of the Justice Department, which is responsible for enforcing the Sherman, Clayton, and Robinson-Patman Acts. The Justice Department, together with the offices of the United States Attorneys, can enforce these laws through either civil or criminal prosecution. Sections 1, 2, and 3 of the Sherman Act, Section 3 of the Robinson-Patman Act, and Section 14 of the Clayton Act provide for criminal penalties. The Justice Department may bring civil suits under Sections 1, 2 and 3 of the Sherman Act. Sections 2, 3, 7, and 8 of the Clayton Act (as amended by the Robinson-Patman Act) provide only for civil suits with enforcement either by the Antitrust Division or the Federal Trade Commission (FTC).

48

Enforcement

Because of special features of our federal antitrust laws, the Antitrust Division plays a far more important role in the enforcement process than that of a mere policeman over antitrust activity. Our notion of a policeman is someone who is told what the law is and goes out on the street and tries to apprehend those who violate the law. Of course, every police officer has some discretion; he must decide whether to haul in a jaywalker for ticketing or to chase down a purse snatcher. The Antitrust Division, by comparison, is part police department and legislature. Its discretion is much broader than that of the typical enforcement agency. For example, the Antitrust Division issues guidelines that detail its enforcement plans in certain areas of antitrust law, and where there are no guidelines, there are practices on which the business community has come to rely. But federal law does not fix the guidelines and the customs of the department. A new Division Head can revoke a set of guidelines at any moment. Activity once considered safe may become subject to criminal prosecution. The FTC shares enforcement of the Clayton Act, largely a civil statute, with the Justice Department. It is also primarily responsible for the enforcement of the FTC Act, which outlaws unfair competition and unfair or deceptive advertising (Section 5 of the FTC Act). The agency’s mission has two components: antitrust enforcement (Bureau of Competition) and consumer protection (Bureau of Consumer Protection). Like most other federal agencies, it enforces rules and promulgates new rules (which operate industry-wide) through the rule-making procedure detailed in the Administrative Procedure Act. Also, like most such agencies, the FTC carries out its “in-house” enforcement activity through administrative proceedings which culminate in either a finding of a violation or no violation. If the defendant violated the FTC Act, the typical remedy is a cease and desist order, which defendants can appeal within the agency to the Commissioners. If the Commission upholds the decision of the administrative law judge, the defendant may appeal to the United States Court of Appeals. If the appeals court upholds the Commission, the Commission may seek enforcement and civil penalties from a federal district court. As for non-“in-house” activity, such as civil prosecutions under the Clayton Act, the FTC must proceed, as would an ordinary private plaintiff, by going to the federal district court to seek an injunction. The FTC may also seek an injunction from a federal court pending the outcome of an administrative hearing. As Section 4 of the Clayton Act makes clear, any person, whether an individual or a business entity, injured by conduct in violation of the

II. Enforcement Provisions of the Antitrust Laws

49

antitrust laws, may sue to recover treble damages, costs of suit, and a reasonable attorney’s fee. This is an unusually generous statute to plaintiffs. It is no wonder that the federal courts have been on guard for plaintiffs who have tried to convert ordinary business torts (and even mundane gripes) into antitrust claims. Finally, state attorneys general have authority under Section 4C of the Clayton Act to bring federal antitrust suits as parens patriae on behalf of state residents.

B. Are Penalties Optimal? Violations of the Sherman Act are criminal7 and are punishable by up to three years’ imprisonment and fines up to $350,000 for an individual and up to $10 million for a corporation. The Criminal Fines Improvement Act of 1987 provides that fines may exceed these levels to the extent of twice the gross gain derived by the defendant from the offense or twice the gross loss resulting to others, unless imposition of such a fine would unduly complicate or prolong the sentencing process.8 Consider these amounts in light of the optimal fine discussed in the previous section. If the “gross loss” incurred by a consumer of the monopolist can be measured by the monopoly transfer, then the maximum level is twice the transfer. Indeed, in the simple linear case, both the gross gain and the gross loss will equal the monopoly transfer. Thus, if we take two times the monopoly transfer as the statutory maximum fine, we see that it is greater than –32 times the monopoly transfer, which is the simplest approximation of the optimal penalty. If the probability of detection and apprehension is less than one, this comparison becomes more difficult; for in this case the optimal penalty is –32 times the transfer divided by the probability of apprehension. If the probability of apprehension is –12 , then the optimal fine is three times the transfer, which is greater than the statutory maximum fine. It is easy to see that if the probability of apprehension is greater than –43 , then the statutory maximum fine will exceed the optimal fine, and conversely. In addition, 7

8

The Department of Justice can also prosecute civilly under the Sherman Act. For a discussion of the frequency of criminal and civil prosecutions by the Justice Department, see Richard A. Posner, A Statistical Study of Antitrust Enforcement, 13 J. Law & Econ. 365–419, 385–86 (1970). 18 U.S.C. §3571. For discussion of the statute, see Jeffrey S. Parker, Criminal Sentencing Policy for Organizations: The Unifying Approach of Optimal Penalties, 26 Am. Crim. L. Rev. 513 (1989).

50

Enforcement

the greater the cost on the margin of enforcing the law, the more likely that the optimal fine exceeds the statutory maximum. We have demonstrated that under the assumption of costless enforce3 ment – 2p times the monopoly transfer provides an upper bound on the optimal fine, where p is the probability of apprehension. If the marginal cost schedule is flat in the relevant range of output, this measure is precise. This suggests that if p is greater than –34, the statutory maximum exceeds the optimal fine. The question then becomes: how likely is it that the probability of apprehension exceeds 75 percent? The likelihood of apprehension and punishment depends on the defendant’s activity.9 Price-fixing agreements, for example, are kept secret and therefore are unlikely to be discovered by enforcement authorities. The probability of apprehending and punishing price-fixers is probably far below 75 percent. On the other hand, mergers typically take place in public view and are discussed in the business sections of newspapers. The probability that a merger creating monopoly power will lead to antitrust enforcement may easily be as high as 75 percent.10 There is a further difficulty having to do with the type of legal standard being enforced. The analysis so far has assumed that the punishment authority fines the defendant without inquiring into the reasonableness of his conduct. Put another way, I have assumed the enforcement authority applies a rule of per-se or strict liability. However, under a rule of reason standard, a court or punishment authority will examine the reasonableness of the defendant’s conduct (i.e., conduct a cost-benefit analysis of sorts), and hold him liable only if his conduct is deemed unreasonable.11 Under this type of test, the change in the defen9 10

11

Herbert Hovenkamp, Federal Antitrust Policy 600 (1994). For an interesting study that estimates enforcement probabilities in the merger context, see Malcolm B. Coate, Richard S. Higgins, and Fred S. McChesney, Bureaucracy and Politics in FTC Merger Challenges, 33 J. Law & Econ. 463–82 (October 1990). Coate, Higgins, and McChesney compute the probability of a merger challenge given certain findings with respect to factors in the merger guidelines. In general, if both the Bureau of Competition and the Bureau of Economics agree that the merger guideline factors justify a challenge, the likelihood of a challenge is high (97 percent). Perhaps their most interesting finding concerns the case in which the two bureaus disagree. If the lawyers of the FTC’s Bureau of Competition think none of the factors justify a challenge, while the economists of the Bureau of Economics think that all of the factors justify a challenge, the likelihood of a challenge is only .003, which indicates that the lawyers have greater influence than the economists. Since the economists are more likely to reach the right conclusion regarding the importance of the merger guideline factors, this is a troubling result. For an introduction to the rule of reason and the per-se liability standards, see Chapter 5.

II. Enforcement Provisions of the Antitrust Laws

51

dant’s expected liability as he crosses the reasonable conduct threshold may be so large that it exceeds the change in marginal social harm, resulting in overdeterrence.12 To see this, suppose the social harm when the defendant engages in some potentially anticompetitive conduct (such as acquiring monopoly power) is $20, and the harm when he forbears from such conduct is $10 (because he exploits his potential power in a less harmful manner). Thus, the marginal social harm is only $10. Under a reasonable conduct standard, he will be held liable only when he does not forbear, so the change in liability may be as large as $20. Since marginal liability exceeds marginal social harm, the result may be overdeterrence if there is some uncertainty in determining the reasonable conduct threshold. Summing up, this analysis suggests that with respect to covert activities, such as price fixing, the statutory maximum penalty is less than the optimal fine. This is especially likely where the activity is covert and the legal standard is one of per-se liability. Since both of these statements are true in the case of price fixing, this suggests that the likelihood of underdeterrence is especially high in the case of price fixing. Recall that the lower bound on the optimal fine is simply the transfer divided by the probability of apprehension. If the probability of apprehension is less than 50 percent, then the statutory maximum fine falls below the lower bound on the optimal fine. Again, the preceding analysis suggests this is likely to be true for covert activities such as price fixing. I have argued that the fines levied against some antitrust violators are probably well below the optimal level. However, to demonstrate this carefully requires a great deal of empirical evidence that is not at hand. We need data on the likelihood of apprehension and punishment as a function of the defendant’s activity (price fixing, product tying, mergers). With such data we could compare optimal fine estimates with the actual statutory maximum fines. Until such estimates are available, we will have to rely to some extent on conjecture. In view of the probable inadequacy of fines, the threat of imprisonment is probably necessary if the threat of criminal prosecution alone is going to generate the optimal deterrence of covert antitrust violations. Of

12

For a more detailed discussion, see the Appendix. For a general discussion of the deterrence issue in the context of reasonableness rules, see Richard Craswell, Deterrence and Damages: The Multiplier Principle and Its Alternatives, 97 Michigan Law Review 2185 (1999); for a model of torts yielding a similar result, see Keith N. Hylton, Costly Litigation and Legal Error under Negligence, 6 J. Law, Econ. & Org. 433 (1990).

52

Enforcement

course, public enforcement is not limited to criminal prosecution, the government can also prosecute civilly.13 The typical civil remedies are generally equitable remedies such as injunctions, dissolution, or divestiture orders.14 However, under Section 5(a) of the Clayton Act, a judgment in an antitrust case against a defendant creates a presumption of guilt (or prima facie evidence of guilt) in a subsequent private action based on the same violation. Thus, a defendant who litigates against the government to the point of judgment in a civil antitrust prosecution will find himself vulnerable to damage claims from private parties injured by the defendant’s antitrust violation. The question is whether this additional threat of private liability makes up for the inadequate deterrent threat of public enforcement. The discussion of private actions later in this chapter will shed light on this question.

C. Fairness Recall that in Standard Oil, Justice White attempted to import the reasonableness standard employed in the common law restraint of trade cases into Sherman Act jurisprudence (see Chapter 2). That immediately raised a thorny issue. The common law decisions were, and remain to this day, based on contract actions. The Sherman Act, however, is a criminal statute. Legal standards appropriate for civil litigation are not necessarily appropriate for criminal proceedings. Indeed, this was the point of the defendant’s argument in Nash v. United States.15 In Nash, the defendant argued that the Sherman Act failed to meet constitutional due process requirements because of uncertainty. Specifically, he claimed that conviction on a finding of unreasonable conduct was unfair because it required an estimation of reasonableness, which may differ among fair-minded people. The defendant was not saying that the estimate might differ among reasonable people; that would be false. A finding that a defendant’s action is unreasonable has long been considered equivalent to a finding that reasonable people would not knowingly take the defendant’s action. However, fair-minded refers in this context to people who do not have a specific intent to harm. Fair-minded

13

14 15

On the frequency of civil and criminal prosecutions, see Richard A. Posner, A Statistical Study of Antitrust Enforcement, 13 J. Law & Econ. 365–419, 385 (1970). For a discussion of civil remedies, see id. at 385–6. 229 U.S. 373 (1913).

II. Enforcement Provisions of the Antitrust Laws

53

people might differ about the reasonableness of action, even though reasonable people would not. Reasonable people are necessarily fairminded, in the law, but the converse is not true. Outside of the set of fairminded people, we have criminals, on one end of the spectrum, and incompetents on the other. The Court rejected the defendant’s claim in Nash on the ground that the law is “full of instances where a man’s fate depends on his estimating rightly . . . some matter of degree,”16 and gave examples of instances in which the criminal law required estimation of reasonable behavior.17 Consider an antitrust example (not from Nash): Suppose that a court had to adjudicate a Sherman Act Section 2 criminal prosecution involving a merger that would, in a civil case, be declared an unlawful combination because of an unreasonable economic effect, although it recognized that fair-minded people could reasonably conclude otherwise.18 I find it hard to avoid the conclusion that this is unfair. The rule of reason standard is analogous to the negligence test in tort law. Both require a balancing of the social costs and benefits of the defendant’s behavior. A court can find a defendant negligent in the absence of criminal intent. In theory, it is possible for a new head of the Antitrust Division of the Justice Department to decide to bring criminal prosecutions against routine business conduct never before prosecuted under the Sherman Act. In practice, it does not happen because the department has customarily reserved criminal prosecution for flagrant cases. In addition, the Department has taken to issuing guidelines in recent years to inform the business community of the actions that are likely to give rise to criminal or civil proceedings. Furthermore, the practice of the Justice Department tends to limit criminal proceedings to situations where: (1) the offense is clearly unlawful (price fixing), (2) the defendant has acted with “specific intent to restrain trade or monopolize,” or has engaged in “predatory practices,” (3) the defendant knows that courts have held similar

16 17

18

229 U.S. at 377. “If a man should kill another by driving an automobile furiously into a crowd he might be convicted of murder however little he expected the result. . . . If he did no more than drive negligently through a street he might get off with manslaughter or less. . . . And in the last case he might be held although he himself thought that he was acting as a prudent man should.” 229 U.S. at 377. Phillip Areeda and Louis Kaplow, Antitrust Analysis: Problems, Text, Cases 56 (5th ed., Aspen Publishing Co., 1997).

54

Enforcement

conduct unlawful in other cases, or (4) the defendant has previously committed an antitrust offense.19 All of this is reassuring, but one should not lose sight of the issues raised by the considerable discretion enforcement authorities have under the Sherman Act. This is an enforcement framework that contradicts some of the sturdiest notions of procedural fairness: an unspecific, short, vague criminal statute that requires the application of a balancing test, coupled with an enforcement agency that binds itself by following custom and periodically issuing guidelines to the community of potential defendants. One is reminded of the story of Caligula’s edicts, that were hung so high people could not read them.20 It is a requirement of due process under our Constitution, and, going back further, a fundamental common law maxim, that criminal statutes should be specific, clear, and strictly construed. Specificity binds the hands of the enforcement authorities, and prevents them from choosing particular individuals for punishment, whether or not they had committed a wrong. Our federal antitrust enforcement framework represents a step backward when viewed from this perspective. The troubling nature of the enforcement structure is amplified by the many examples of actions that could give rise to either criminal or civil proceedings under the antitrust laws. For example, Section 3 of the Clayton Act outlaws tying and exclusive dealing. However, the Court has held that Sherman Act Section 1 can apply to the same activity; and as my earlier example suggested, Clayton Act Section 7 and Sherman Act Section 2 could both apply to mergers. Suppose that the Justice Department overturns enforcement custom by deciding to criminally prosecute some action that historically has been a subject of civil proceedings. Is the potential defendant entirely at the mercy of the Antitrust Division? As a matter of doctrine, the Court answered this in United States v. United States Gypsum Co.,21 holding that criminality required a showing of mens rea. The Court defined the requisite mental state as knowledge that one’s conduct would result in anticompetitive consequences, or a specific intent to violate the statute. Precisely what specific intent means in the area of antitrust law poses an

19

20 21

Phillip Areeda and Herbert Hovenkamp, Antitrust Law: An Analysis of Antitrust Principles and Their Application, vol. III, ¶ 311a2, p. 30 (Boston: Little, Brown, and Company, 1995). William Blackstone, Commentaries on the Laws of England, Volume I, p. 46. 438 U.S. 422 (1978).

II. Enforcement Provisions of the Antitrust Laws

55

interesting question. All businesses aim to maximize market share or profit, destroy competitors, corner the market, grow unboundedly, and so on. One can question whether internal memoranda expressing a desire or intent to “destroy the competition” should be taken as evidence of a specific intent to violate the antitrust laws. To be meaningful as a constraint on enforcement authorities, specific intent in antitrust should refer to an intent to take some action that is unambiguously harmful to consumer welfare.

D. Equity Proceedings Sherman Act Section 4 and Clayton Act Section 15 confer jurisdiction on the federal courts to “prevent and restrain violations of the antitrust laws” and direct the government to institute proceedings in equity to prevent and restrain antitrust violations.22 In some cases, the Justice Department or the FTC will seek a preliminary injunction. In most merger cases, this has the effect of a permanent injunction. One example is Royal Crown Co. v. The Coca-Cola Company.23 Coca-Cola and Dr. Pepper abandoned their proposed merger after the court issued a preliminary injunction in favor of the FTC. Clayton Act Section 16 permits private persons to obtain injunctions against actual or threatened antitrust injuries. The Supreme Court in the 1990–91 term24 decided that private parties can seek divestiture under Section 16. Clayton Act Section 16 also provides that “in any action under this section in which the plaintiff substantially prevails, the court shall award the cost of suit, including a reasonable attorney’s fee, to such plaintiff.”25 Preliminary injunctions are available to a private plaintiff where they would be available to the government: (1) the Court will consider the 22

23

24

25

For example, Philip Areeda and Louis Kaplow, Antitrust Analysis: Problems, Text, Cases 60 (5th ed. 1997). 887 F.2d 1480 (11th Cir. 1989). For discussion, see John E. Kwoka and Lawrence J. White, eds., The Antitrust Revolution 80–98 (Glenview, IL: Scott, Foresman, and Company, 1989). California v. American Stores Co., 495 U.S. 271 (1990), held that Section 16 of the Clayton Act entitles any person to divestiture as a form of injunctive relief against threatened loss or damages by violation of antitrust laws. For Section 16 divestiture suits, equitable defenses such as laches are available and may protect consummated transactions from belated attacks by private parties. 15 U.S.C. §15 (1988).

56

Enforcement

probability of plaintiff victory on the merits, (2) the hardship of a preliminary injunction on the defendant, and (3) the possibility that final relief could undo anticompetitive harm in the interim.26 Some scholars have suggested that courts tend toward leniency on the likelihood of success requirement.27 One can offer an explanation for this. Put yourself in the shoes of a district court judge looking at the complaint of a plaintiff who claims that a merger between two of his competitors will result in anticompetitive injuries. It would require an expert on the industry to assess the plausibility of the allegations, and judges do not have this expertise. The judge has a choice: deny the injunction, allowing the merger to go through or grant the injunction. If he chooses to deny the injunction, he faces the risk that the litigation will continue, the plaintiff will prove anticompetitive harms in court, and remedies will be difficult and costly to provide. In addition, he may be criticized at the appellate level. The appellate judges probably will scold him for not giving them time to look over the allegations while the deck was clear; at the least, they will say, he should have issued a stay pending an appeal of the decision to deny the injunction. If he grants the injunction, he avoids this potentially embarrassing criticism. The cost of the injunction, if it issued, is borne by the defendants, and by consumers if the merger would have enhanced consumer welfare. But the appellate court will not criticize the judge for this aspect of his decision. Obviously, the stressminimizing course of action is to issue the injunction. Putting all of this together, a problem emerges: an incentive exists for hold-up litigation in the merger area. Consider the incentives of a firm that sees its competitors about to merge into a powerful rival. What is the rational thing to do? Seek a preliminary injunction under Section 16. Is this a desirable state of affairs? Probably not. If the merger violates Section 7 of the Clayton Act, then how can it also hurt a competitor? If it violates the Clayton Act, then it tends to produce a monopoly or reduce competition substantially. But, in most cases, a reduction of competition between two (or more) firms opens up new competitive options for their rivals. That is, if two competing firms merge in order to raise the price charged for the good they sell, then a rival would have an oppor-

26

27

For example, DFW Metro Line Serv. v. Southwestern Bell Tel. Co., 901 F.2d 1267, 1269 (5th Cir. 1990). Philip Areeda and Louis Kaplow, Antitrust Analysis: Problems, Text, Cases 63 (5th ed. 1997) (“[A] number of courts do not require the private plaintiff to show a likelihood of success on the merits.”)

II. Enforcement Provisions of the Antitrust Laws

57

tunity to undercut the new merged entity and increase its market share. However, there is one important class of exceptions to this general rule. In the “essential facility” cases (see Chapter 10), we see transactions that allegedly result in monopoly power and a reduction of competition. For example, if a group of competing firms merge and acquire all of the ports of entry into a certain geographic market, the new merged entity could put its rivals at a disadvantage by charging discriminatory rates for access to the geographic market.28 The monopoly profits earned in this case are, technically speaking, the rents earned on the possession of exclusive access to the ports of entry. Outside of the class of essential facility cases, it seems that as a general rule the merger of two firms in order to reduce price competition between them cannot hurt a competitor. In view of this, some commentators say that courts should presumptively deny standing to competitors seeking to block mergers.29 What can the potential defendants who wish to merge do? One option is to seek a “business review letter,” in which the Justice Department states its intention not to bring suit. Alternatively, they can seek an advisory opinion from the FTC approving a proposed course of conduct. But these clearances are not binding. The agency reserves a free hand, should it later discover that the merger does violate the antitrust laws. Also, these letters do not constrain private plaintiffs and courts. The letters do not claim to grant immunity to the recipient, and even if they did, courts have made it clear that they will recognize immunity only if Congress, or some agency given immunization authority by Congress, has granted it.30

E. Private Antitrust Suits Private actions are by far the greatest part of antitrust litigation. Over the 1980s, the ratio of private to government antitrust suits was 28

29

30

This is an example of the general “raising rivals’ costs” strategy emphasized by Steven Salop. See Steven C. Salop and David T. Scheffman, Raising Rivals’ Costs, 73 Am. Econ. Rev. 267 (1983). See, for example, William Baumol and Janusz Ordover, Use of Antitrust to Subvert Competition, 28 J. L. Econ. 247 (1985). Of course, there is another side to this debate. One could argue that because competitors are virtually the only private parties with the information and resources necessary to bring a successful merger challenge, their role in merger enforcement must be accepted. See Joseph F. Brodley, Antitrust Standing in Private Merger Cases: Reconciling Private Incentives and Public Enforcement Costs, 94 Michigan Law Review 1, 1996. See Chapter 5.

58

Enforcement

10–1.31 A simple glance at Clayton Act Section 4 should reveal why private actions are such a large part of antitrust enforcement. Section 4 authorizes anyone injured in business or property to sue for treble damages plus costs, plus a reasonable attorney’s fee. There is probably no more plaintiff-friendly statute in all of the federal laws. 1. Treble Damages. Suppose a cartel fixes price at the monopoly level, as shown in Figure 3.1. The injury to consumers equals the sum of the transfer portion and the forgone consumer surplus. However, the only compensable portion is the transfer. Clayton Act Section 4 permits the consumers to sue for three times the transfer. Under what conditions is the treble damages rule optimal? The optimal damage formula, recall, has an upper bound equal to Ê 3 ˆ( transfer ) + C Ë 2p ¯ where C now represents the plaintiff’s cost of bringing suit and p is now the probability that the plaintiff brings and wins his lawsuit. Assume for simplicity that this formula is the optimal penalty.32 Because the Clayton Act permits the plaintiff to recover the cost of bringing suit, the crucial variable in any comparison of actual to optimal damages is the frequency of suit. Put another way, given that the successful plaintiff will be awarded 3(transfer) + C, in order to determine whether private damages 3 exceed the optimal level we need only compare (– 2p)(transfer) with 3(transfer). It follows from this that if the probability of suit is less than one half, treble damages fall below the optimal level. Is there any evidence on the frequency of suit in the antitrust area? I am aware of none. However, studies of tort litigation suggest that the 31

32

Steven C. Salop and Lawrence J White, Economic Analysis of Private Antitrust Litigation, 74 Georgetown Law Journal 1001, 1002 (1986). One can demonstrate that this is the optimal damage award under certain conditions. In particular, if the deterrence benefit (avoided losses net of forbearance costs) exceeds the total litigation costs, on the margin, for every lawsuit, then the optimal damage award, under strict liability, will internalize both the harm to victims and the victim’s cost of litigation. See Keith N. Hylton, Welfare Implications of Costly Litigation under Strict Liability, 4 American Law and Economics Review 18 (2002). Moreover, if for some plaintiff-defendant pair the deterrence benefit is less than the total litigation cost, the pair will have an incentive to enter into a waiver agreement in which the potential plaintiff agrees not to sue the defendant. Under the assumption that parties have exhausted potential gains from waiver contracts, the deterrence benefit from litigation will exceed total litigation costs. For an analysis of deterrence and waiver contracts, see Keith N.

II. Enforcement Provisions of the Antitrust Laws

59

frequency of suit conditional on negligent conduct is low, on the order 1 33 of – . If the frequency of suit is roughly the same in antitrust, then it 10 would follow that treble damages fall below the optimal level and antitrust damages underdeter. However, the findings of tort studies may not be transferable to the antitrust context. Because of treble damages and fee shifting, the frequency of suit conditional on a violation is probably greater in antitrust than in tort law. A few other qualifications are in order. First, whether antitrust damages over- or underdeter depends in part on the type of legal standard being applied. The foregoing argument assumes courts are applying a per-se standard. Under a reasonableness standard, underdeterrence is less likely, for reasons discussed earlier.34 Second, I have assumed so far that courts do not make mistakes and plaintiffs do not bring baseless claims. If we relax these assumptions, then the comparison between actual and optimal damages becomes 3 more difficult. The optimal level of damages may be less than (– 2p) (transfer) + C because it may be desirable to lower the level of liability in order to reduce the social costs of baseless legal claims. However, if courts have other tools at hand sufficient to control baseless claimants (e.g., sanctions for bad faith litigants), then the optimal scheme would 3 set damages (– 2p)(transfer) + C for antitrust violators and assess a separate penalty on baseless claimants. It happens that courts do have authority to sanction bad faith litigants. If the threat of sanctions works as a deterrent to bad faith litigants, then the existence of baseless claims would not require alteration of the optimal damages formula.35 What if the goal of antitrust damages is not to provide optimal deterrence but to provide full compensation to antitrust victims? If so, then one should note that the fully compensating damage level is approximated by Ê 3 ˆ( transfer ) + C . Ë 2p ¯ The Clayton Act allows the plaintiff to collect 3(transfer) + C, so therefore overcompensates the plaintiff.

33

34 35

Hylton, Agreements to Waive or to Arbitrate Legal Claims: An Economic Analysis, 8 S. Ct. Econ. Rev. 209 (2000). For discussion of empirical evidence on the likelihood of a tort suit, see Michael J. Saks, Do We Really Know Anything About the Behavior of the Tort Litigation System – And Why Not?, 140 U. Penn. L. Rev. 1147 (1992). See note 12 and accompanying text. For further discussion, see the Appendix.

60

Enforcement

The upshot of all this is that while the treble damages rule probably overcompensates victims, it is unclear whether it underdeters potential antitrust violators. Determining whether the trebling rule over- or underdeters requires empirical research on the frequency of private antitrust claims, the types of claims (per se, rule of reason), and the level of damage awards. The mere fact that damages are trebled does not imply overdeterrence. Some have suggested that the solution to the overcompensation problem (if it is a problem) is a decoupling of antitrust damages.36 Under this scheme, the successful plaintiff would receive compensatory damages plus an attorney’s fee. The remaining portion, two times damages, would go to the federal government. There are several problems with the decoupling proposal. First, if antitrust damages underdeter, then decoupling would exacerbate the problem, leading to a regime of full compensation of successful litigants and even more severe underdeterrence. The reason underdeterrence should worsen is that fewer plaintiffs would bring suit without the lure of receiving treble damages. Second, any substantial difference between the amount the defendant pays and the amount the plaintiff receives introduces an incentive for parties to settle out of court, with the defendant paying the plaintiff slightly more than the compensatory level. Increasing settlement may be a desirable goal, but the problem is that very few defendants would ever face up to paying the damage surcharge demanded by the government. This suggests a further reduction in the deterrent effect of antitrust suits. 2. Standing: Policy and Law. Standing doctrine aims to tell us who can sue under the antitrust statute. Clayton Act Section 4 seems to imply that anyone can bring suit, but the law on standing indicates otherwise. A theory of standing policy based on the optimal enforcement framework developed at the start of this chapter is easy to develop.37 Return

36

37

See A. Mitchell Polinsky, Detrebling Versus Decoupling Antitrust Damages: Lessons from the Theory of Enforcement, 74 Georgetown Law Journal 1231 (1986); Warren Schwartz, An Overview of Antitrust Enforcement, 68 Georgetown Law Journal 1075 (1980). For a more complete application of optimal enforcement theory to standing doctrine, see William H. Page, The Scope of Liability for Antitrust Violations, 37 Stan. L. Rev. 1445 (1985).

II. Enforcement Provisions of the Antitrust Laws

61

to Figure 3.1. Consider the parties who might claim injury from a monopolist’s actions. First, a consumer who has to pay the monopoly price clearly suffers from the monopoly overcharge and should therefore have standing to sue. Second, consider the union of employees of the monopolist. It is not clear that the monopolist’s activity harms the union. Certainly the union’s membership level falls to the extent the monopolist reduces output below the competitive level. However, the union probably bargained for higher than competitive level wages, as did airline employees before deregulation. Third, what about almost-employees, individuals who could have worked in production, had the monopolist not reduced output below the competitive level? Yes, the monopolist has harmed them, and in theory they should have standing. The same goes for almost-consumers, who would have consumed the product had the monopolist not set price above the competitive level. However, it should be clear that a rule granting standing to either of these parties would create huge administrative difficulties. A potentially very large set of consumers and workers could claim standing on this theory. Because of the prohibitive administrative costs of distinguishing true victims from phonies, the economically rational rule denies standing to these parties.38 This theory has clear implications for unmeasurable portions of damages. Each consumer who has purchased two widgets at the monopoly price could potentially claim that he would have purchased three widgets at the competitive price. But how can courts determine the validity of such claims? They cannot, so courts do not permit consumers to recover on such speculative components of injury. The foregoing suggest three simple principles: (1) consumers get automatic standing, (2) union members and other providers of inputs have an unclear case for standing, and (3) plaintiffs who cannot prove injury do not have standing. The law is generally consistent with these principles. The statute authorizes recovery only for actual injury to “business or property.” A loss of employment or reduction in wages is not generally

38

One could argue that this provides a key reason for public enforcement of the antitrust laws: it is impossible to identify all of the victims. However, I have assumed in the earlier section discussing optimal damages that damages for identifiably injured victims could be multiplied by a factor that effectively internalizes to the monopolist the costs that unidentifiable victims bear.

62

Enforcement

considered to be an injury that justifies standing, unless the plaintiff’s job is operating a business. However, courts have permitted employees to sue under the Sherman Act if their employers conspired to reduce wages. On the other hand, a claim brought by a union on the theory that a larger workforce would have brought it more dues probably would fail. The doctrine in this area has not been set out in bright lines; however, courts seem to be making an effort to distinguish cases of clear harm from those with mere speculative harm. As an analogous case, consider a lessor who collects a percentage of the lessee’s sales, and the lessee belongs to a price fixing cartel. In theory, this is similar to the case of the union that would like to sue the monopolist-employer for lost dues from decreased hiring. On one hand, the union did not receive the revenue from additional members that it would have received in a competitive output market; but on the other, the union probably shared in the monopoly rent. The same may be said of the lessor in this example. The Ninth Circuit looked at this case, treating it as one of “derived injuries,” and held that the lessor did not have standing.39 What about the case in which a second consumer also absorbs the injury? Should both the initial and second consumers have standing, or just one of them? Suppose the monopoly price is Pm and the competitive price is Pc. The amount sold the first purchaser is Q1, and that purchaser then sells Q2 to the second (at the same price), and so on. The monopoly overcharge is therefore

(Pm - Pc )Q1 = (Pm - Pc )(Q1 - Q2 ) + (Pm - Pc )(Q2 - Q3 ) + . . . + (Pm - Pc )(QN -1 - QN ) + (Pm - Pc )QN . A rule granting standing to every consumer for no more than the amount of his harm could result in N consumers having the right to sue for the overcharge on the units they purchased. The litigation could be expensive. More troublesome, however, is the possibility that with small purchase quantities and high litigation costs, perhaps not one consumer would have an incentive to bring suit. A rule that minimizes unnecessary litigation costs and maximizes deterrence would restrict the number of plaintiffs and increase their potential awards, or require the consumers to join in a class action.

39

See R. C. Dick Geothermal Corp. v. Thermogenics Inc., 890 F.2d 139 (9th Cir. 1989).

II. Enforcement Provisions of the Antitrust Laws

63

The answer the law provides comes from the Hanover Shoe/Illinois Brick rule. Under Hanover Shoe, the defendant normally cannot claim that the plaintiff passed on the price increases to its customers and therefore suffered no injury. Under Illinois Brick, the consumer purchasing from an innocent middleman may not recover from the manufacturer who unlawfully agreed with rival manufacturers to fix prices. The rules essentially permit the first consumer in the above example to bring suit.40 Although courts seem not to have noticed the connection, the passingon problem is essentially the economic source generating the doctrine governing derived injuries. In general, derived injuries refers to the set of cases in which some party linked to a second party who has dealings with a price-fixer claims to have suffered some derivative harm. The classic example is the shareholders or creditors of a firm that has purchased inputs from a monopolist. With respect to shareholders of the injured firm, the general rule denies standing. If such shareholders had standing, then it would be similar to the case in which all individual consumers received standing regardless of how far down they were in the chain of passed-along injuries. Litigation costs are minimized and deterrence enhanced by granting standing to a more immediate victim. Although the law on standing for derived injuries is vague,41 precise theoretical concerns drive the arguments for or against standing. They are two: (1) whether the plaintiff really suffered an injury as a result of the defendant’s practices, and (2) whether deterrence becomes greater by concentrating the power to bring suit in one individual, rather than spreading it over a large population of victims. Finally, there is a general requirement that the plaintiff prove antitrust injury. Under the rule announced in Brunswick Corp. v. Pueblo BowlO-Mat,42 the plaintiff must show (1) that he has suffered an injury (or the threat of an injury), (2) that the defendant’s illegal conduct caused the injury, and (3) that the defendant’s conduct was the sort that the antitrust laws were designed to prevent. This is essentially a threshold causation

40

41

42

California v. ARC, 490 U.S. 93 (1989), complicates matters considerably for the potential defendant. ARC held that the federal antitrust laws do not preempt a state antitrust law allowing indirect purchasers to recover overcharges. See Associated General Contractors v. California State Council of Carpenters, 459 U.S. 519 (1983), which announced several factors to consider: (1) the risk of duplicative recoveries, (2) danger of complex apportionment, (3) nature of the alleged injury, (4) relationship between the alleged violation and alleged injury, (5) the existence of more direct victims of the alleged conspiracy. 429 U.S. 477 (1977).

64

Enforcement

requirement. Thus, a plaintiff who claims that his business is being destroyed by the price cutting of a competitor is denied standing if the price cuts are not predatory.43

appendix A. Optimal Penalty, Basic Result Proposition: The optimal antitrust penalty is equal to the sum of the monopoly transfer and the foregone consumer surplus. Proof: Assume there are two states, perfect competition and monopoly. Let i index the state with i = 0 representing competition, i = 1 representing monopoly. Let CS(i) be consumer surplus in state i, R(i) be firm profit in state i, PS(i) be producer surplus in state i. Total Welfare in state i is W(i) = CS(i) + PS(i). First, it is straightforward to show that PS(i) = R(i) + F, where F is the firm’s fixed cost of operation. Thus, any policy that maximizes PS(i) necessarily maximizes R(i). Under competition, W(0) = CS(0) (because R(0) = 0). Monopoly is present in state 1, thus, welfare is W(1) = CS(1) + R(1). Welfare is enhanced by monopoly if and only if CS(0) - CS(1) < R(1). It follows that a fine equal to CS(0) - CS(1) will induce monopolization only when it is wealth-enhancing. Thus, the optimal penalty is equal to the monopoly transfer plus the forgone consumer surplus.

B. Optimality of Treble Damages Here I extend the discussion of treble damages by taking into account the possibility of judicial error, and that a suit may be brought against a firm in compliance with the antitrust law. Let v1 = the probability of injury, given the firm violated the law; v2 = the probability of injury, given the firm has complied with the law; p = the probability of a lawsuit, given the firm violated the law; q = the probability of a lawsuit given the firm has complied with the law; Cp = the plaintiff’s litigation costs; Cd = the defendant’s litigation cost; h1 = probability plaintiff loses, given defendant violated the law; h2 = probability plaintiff wins given defendant did not violate the law; T = monopoly transfer; L = liability (damages). 43

Atlantic Richfield Co. v. U.S.A. Petroleum Co., 495 U.S. 328 (1990).

Appendix

65

1. Plaintiff Uninformed. I assume first that plaintiffs do not know whether the defendant violated the antitrust law. The probabilities p and q differ only because the likelihood of bringing a suit is higher when the defendant violates the legal standard. In this case, there are no frivolous or baseless lawsuits. Consider the defendant’s incentive to comply with the law. If the cost of compliance is x, the firm will comply when x < [␯1 p(1 - h1 ) - ␯ 2 qh2 ]L + (␯1 p - ␯ 2 q)Cd , where the right hand side represents the increase in liability experienced by a firm that chooses not to comply with the law. Under assumptions made in the text, it is socially desirable for the firm to comply when Ê 3ˆ x < (␯1 - ␯ 2 ) T + (␯1 - ␯ 2 )(C p + Cd ) Ë 2¯ Equating (v1 - v2)(–32 )T + (v1 - v2)Cp with [pv1(1 - h1) - qv2h2]L + (pv1 qv2)Cd and solving for L yields the following formula for the optimal level of damages: (1)

ÈÊ 3 ˆ ˘ L* = {(␯1 - ␯ 2 ) [␯1 p(1 - h1 ) - ␯ 2 qh2 ]} Í T +Cp ˙ ÎË 2 ¯ ˚

This analysis suggests that the optimal antitrust damage formula has two components: a measurement multiplier of –32 , to incorporate forgone consumer surplus in the damage award; and a deterrence multiplier equal to (v1 - v2)/[v1p(1 - h1) - v2qh2]. Note that a priori, the deterrence multiplier may be greater than or less than one. For example, if p (the probability of bringing and winning a lawsuit given noncompliance) is much greater than q (the probability of bringing and winning a lawsuit given compliance), the deterrence multiplier may be less than one, which implies that the optimal damage award should provide for less than full compensation to antitrust plaintiffs. This scenario is particularly likely under a rule of reason standard. This is because the probability of winning given noncompliance is substantially larger than the probability of winning given compliance under the rule of reason standard.44 44

This reasoning is consistent with the argument in Craswell, supra note 12.

66

Enforcement

Now suppose courts never err, h1 = h2 = 0. In this case, the deterrence multiplier simplifies to [(v1 - v2)/v1p], which is less than one if (v1 - v2) < v1p. This is the simplest case in which the marginal social harm is less than the marginal liability – and again, this is more likely to be the case under a rule of reason standard. Note, however, that if the question of factual causation is rigorously considered by the Court, this should never occur. The reason is that a factual causation inquiry would cause the Court to hold the defendant liable only for those harms attributable to the failure to comply with the standard.45 Thus, the Court would reduce the plaintiff’s expected recovery, so that marginal liability risk becomes (v1 - v2)p. This implies the multiplier is simply 1/p as shown in the text. Finally, consider the special case in which v1 =1 and v2 = 0. This describes the case, fairly common in antitrust law, where harm is certain to occur if the defendant does not forbear, and certain not to occur if the defendant forbears. Consider, for example, price fixing. In this case, the deterrence multiplier simplifies to 1/p and the optimal damages formula becomes (2)

L* =

Ê 3 ˆ T + Cp Ë 2p¯

which is the formula stated in the text. 2. Plaintiff Informed. Now assume the plaintiff knows whether or not the defendant violated the legal standard. In this case, raising the fine deters anticompetitive acts (as in the previous example), but also increases the number of baseless legal claims. The level of damages alone is inadequate as a tool for bringing about optimal deterrence of antitrust violations and deterrence of baseless claims. Optimality could be achieved by a two-part scheme in which losing defendants are required to pay the damages specified in (1) and a penalty is imposed on baseless claimants (or a penalty is imposed on losing plaintiffs). If the level of damages is the only tool used to deter antitrust violations and control baseless claims, the optimal damages formula will be of the form (3)

45

ÈÊ 3 ˆ ˘ L* = {(␯1 - ␯ 2 ) [␯1 p(1 - h1 ) - ␯ 2 qh2 ]} Í T +Cp ˙ - S ÎË 2 ¯ ˚

This point is due to Mark F. Grady, A New Positive Economic Theory of Negligence, 92 Yale L. J. 799 (1983).

Appendix

67

where S is a subsidy to defendants.46 The subsidy ensures that the optimal level of damages equates the marginal social cost of inadequate deterrence with the marginal benefit of reducing the number of baseless claims. The new optimal level of damages will be unambiguously less than the level that optimally deters antitrust violations (in the absence of baseless claims). Whether the new damages formula is less than treble damages is unclear. 46

For a rigorous treatment of optimal damages in a setting in which control of the victim’s conduct is important, see Keith N. Hylton, Optimal Law Enforcement and Victim Precaution, 27 RAND J. Econ. 197 (1996).

4 Cartels

Collusion with respect to price violates Section 1 of the Sherman Act. However, because most collusive agreements are not explicit, Section 1 of the Sherman Act is difficult to enforce. The participants in a collusive agreement have strong incentives to hide it, given that it is unlawful and that consumers would object to it anyway. Thus, the interesting problems in applying Section 1 to collusive agreements, or cartels, are typically issues of inference and of determining the scope of the statute. In other words, the difficult issues are: when does the evidence demonstrate collusion, and what constitutes unlawful collusion? This chapter provides an overview of these problems.

i. cartels A. Basic Theory At the heart of Section 1 of the Sherman Act is a concern over the effects of cartels. A cartel is a group of firms that seeks to increase profits by restricting price and output competition among themselves. The fundamental result of the theory of cartels is that cartels are unstable, in the sense that they are not “self-enforcing.” The theory draws on the classic Prisoner’s Dilemma of game theory.1 Consider an 1

The problem is as follows: The police hold two criminals on charges of, say, robbery. They are put in separate cells. The police officers offer each the following deal. If both confess, both will be sentenced to jail for three years. If one confesses, and the other criminal does not confess, the one who confesses will be sentenced to jail for one year. The one who refuses to confess will be sentenced to six years. The criminals are each aware that

68

I. Cartels

69

Firm 1/Fim 2

Collude

Compete

Collude

($3B, $3B)

(–1,5)

Compete

(5, –1)

(0,0)

Figure 4.1

example of collusion between two competing firms, with the following payoffs shown in Figure 4.1. The first element is the payoff to firm 1, the second the payoff to firm 2. If both firms collude, they receive $3 billion each. If one colludes (i.e., sets a high price) while the other competes, the competitor receives $5 billion, and the “colluder” loses $1 billion. If both compete, they receive zero profits, as assumed in the model of perfect competition. Note that a policy of always competing yields two possible payoffs: $5 billion, if the other firm sets price high, and 0 if the other firm sets price low. A policy of colluding yields $3 billion, if the other firm colludes, and a $1 billion loss if the other firm competes. No matter what your competitor does, you are better off competing. Using game theory terminology, competing is a dominant strategy in this example.2 Does this imply that price fixing, the core evil the antitrust law aims to deter, never occurs? No. The point is that collusion is unstable, not that it never happens. But the question generated by the example is a bit more complicated. One can accept the point that collusion can occur, without being convinced that the expense of antitrust enforcement is justified. For enforcement to be a sensible policy, collusion must have sufficient stability to cause injuries that justify enforcement. If collusive agreements occur occasionally, lasting for one day and then falling apart,

2

if they refuse to confess, the likely sentence is two years. The “payoff” for confessing is either three years or one year in jail. The payoff for not confessing is either six years or two years in jail. They are better off, as a group, if both refuse to confess. However, each has a strong incentive to confess, not knowing what the other will do. I will use the term “dominant strategy” in later discussions of the cartel problem in Chapters 4 and 7. On the concept of a dominant strategy, see R. Duncan Luce and Howard Raiffa, Games and Decisions 79–80 (New York: Wiley, 1957). The connection with the Prisoner’s Dilemma, discussed in note 1, should be clear. In the Prisoner’s Dilemma, “confess” is a dominant strategy for each individual prisoner, even though they would be better off as a group if they refused to confess. In this example, “compete” is a dominant strategy for each firm, even though they would be better off colluding on price.

70

Cartels

enforcement would not be a wise policy. But how frequently must collusion occur before enforcement becomes justified? This question has not been answered. Theory tells us two things: collusion can occur, and it is unlikely to last forever. Further, because antitrust enforcement is costly, to justify the expense of enforcement collusion must occur with substantial frequency and duration. What conditions will bring about reasonably stable collusion? For stability to occur, the parties must be reasonably sure that the other members of the cartel will comply with the collusive agreement. It is not enough that A believes that B generally keeps his word. Because if A believes that B believes A will defect from the tacit collusive agreement, A will have an incentive to defect. Let us distinguish “one shot” and “repeat-player” situations. In the former, the parties choose whether to comply with or defect from the tacit collusive agreement, and then (after the game ends) they never see each other again. In the repeat player case, they interact in later periods. In the one-shot case, cartel members cannot punish firms that deviate from the collusive agreement.3 That is because there is no later period in which punishment can be carried out. In view of the infeasibility of private enforcement, the parties are unlikely to develop the expectation of compliance necessary for successful collusion. In the repeat-player case, cartel members can punish firms that deviate from the collusive agreement. Monitoring compliance and punishing deviators increases the prospects for cartel stability. However, monitoring compliance is often difficult. The only way a cartel enforcer could check compliance with a price-fixing agreement is through monitoring transaction prices. But this information is typically shared only between the buyer and the seller, and business records can always be distorted. Punishment is also a troubling issue. The common law of contracts and modern antitrust law share the policy against enforcing collusive agreements.4 Thus, since courts will not enforce an agreement to fix prices, cartel enforcers have only extralegal means. Predatory (below-cost) pricing offers one possible enforcement response, but this is a costly policy for the enforcers. If none of the firms possesses a cost advantage, the firms that predate must suffer losses during the predatory campaign, as they are pricing below cost. These losses come after the 3

4

I am ruling out the alternative in which firms post a bond at the start of the game, which is forfeited by the firm that deviates from the collusive agreement. My one-shot assumption excludes earlier periods in which such an agreement might be set up. That is, collusive agreements that were not deemed reasonable by the court. See Chapter 5.

I. Cartels

71

firms have already suffered losses as a result of the deviator’s refusal to comply with the collusive price agreement. Organized crime networks offer another private enforcement mechanism. The cartel members might employ the local mafia to coerce a competitor into raising its price. But this is a potentially costly approach to enforcement, given the risk of extortion and of being double-crossed. The costs of employing the local mafia as cartel enforcers could easily outweigh the benefits. Informal exchange relationships provide yet another mechanism for enforcement. Referrals from other businesses are helpful in some industries. A firm that deviates from a price-fixing agreement may find that its rivals will no longer refer customers and transfer work to it. In certain fields such as the professions, where sellers have an enormous advantage over buyers in judging the quality of the services provided, these informal networks have a powerful influence. Given all of this, the problem should be clear: enforcement of collusive agreements is difficult generally. Because of this, the threat to enforce such an agreement by punishing defectors often lacks credibility. Even assuming the feasibility and credibility of policies to monitor and punish defectors, there remains the issue of entry. If entry is easy, a cartel that generates profits will invite new competitors to enter the field. Thus, difficulty of entry is a necessary condition for cartel stability. A closely related concern is the extent to which the parties deal with each other. Monitoring compliance and punishing deviations from policy require repeated contacts. In a market of one-shot exchanges, where new sellers constantly enter and old ones leave, the repeated dealings necessary to carry out an enforcement policy will not exist.5

B. Stigler on Monitoring George Stigler’s “A Theory of Oligopoly”6 sets forth one of the early and influential explorations of the monitoring issue. His conclusions are

5

6

There are, to be sure, important cases of relatively stable cartels, such as OPEC and the U.S. shipping cartels. Empirical studies of these cartels generally yield results that are consistent with cartel theory. On OPEC and cartel theory, see James M. Griffen and Weiwen Xiong, The Incentive to Cheat: An Empirical Analysis of OPEC, 40 J. Law & Econ. 289 (1997). On shipping and cartel theory, see Stephen Craig Pirrong, An Application of Core Theory to the Analysis of Ocean Shipping Markets, 35 J. Law & Econ. 89 (1992); William Sjostrom, Collusion in Ocean Shipping: A Test of Monopoly and Empty Core Models, 97 J. Pol. Econ. 1160 (1989). 72 Journal of Political Economy 4 (1964).

72

Cartels

worth repeating here since they have implications for the enforcement of the Sherman Act. Stigler’s first conclusion is that a cartel will prefer to make an agreement on output rather than on price, because participants can more easily monitor output than transaction prices. The simplest cartel agreement fixes quotas for total production, or divides markets into different segments. Whatever form the agreement takes, the monitoring problem remains. Suppose the firms attempt to supervise the agreement by following market share figures closely. A firm whose market share increases dramatically will face suspicion of shading under the cartel price. The difficulty in enforcing such a scheme is that each firm’s market share may fluctuate randomly. Because of these random shocks, the firms will have to devise some means of determining when market share deviates substantially enough from the norm to give rise to reasonable suspicions of cheating. The problem is further complicated by the fact that consumers may lie about the deals they received from other sellers. Indeed, if consumers know about the cartel, they will have strong incentives to lie in order to create suspicion among the members. In response to this, the firms will need to create some method for separating noise from informative signals in the sample of reports from consumers. There are two central features in the monitoring problem outlined by Stigler: (1) the problem of discovering expected market shares in the presence of random fluctuations, and (2) the problem of sorting valid from dishonest reports by buyers. Several conclusions follow from Stigler’s analysis. First, cheating becomes harder to conceal in a market with many buyers. This follows from the law of large numbers. As the number of buyers increases, the typical realization of market share within a given period will more closely approximate the underlying real market share. Second, collusion will be more effective, other things being equal, on honest buyers. It follows that collusion will be observed more frequently in settings where the buyer must report honestly on the offers of other sellers. The classic example is bidding for government projects. More generally, verifiability of reports is the central issue. Even if buyers report falsely, if the reports are easily verified, the sellers can separate valid from false reports. Knowing of the ease of verification, buyers would have little incentive to lie. It is the nonverifiability of consumer reports that makes the monitoring task difficult.

II. Conscious Parallelism

73

Third, collusion is more effective where brand loyalty exists and the identities of buyers do not change frequently. If consumers make repeat purchases, then a stable population of consumers should permit each firm to make share estimates of expected market share. But if the identities of buyers change frequently, the subpopulations of consumers linked to various brands will also change frequently, making it difficult for firms to distinguish random events from shifts in the underlying mix of consumers. The upshot is that Stigler’s factors, in combination with the factors favoring cartel stability discussed earlier (repeat players, buyer-seller informational disparity), suggest a list of conditions associated with collusive pricing. Given the difficulty of actually observing price-fixing agreements, these factors could be used as circumstantial evidence of collusion. Moreover, other things equal, we should observe higher prices where these factors are present.

ii. conscious parallelism A. Introduction to the Problem Conscious parallelism is parallel behavior that typically appears in markets with small numbers of sellers. It is not the result of an explicit agreement. It occurs because in markets with few sellers, firms take the reactions of competitors into account when deciding how much to produce or what price to set. Although it is hard to find a precise definition of it, the term conscious parallelism refers to a form of tacit collusion in which each firm in an oligopoly realizes that it is within the interests of the entire group of firms to maintain a high price or to avoid vigorous price competition, and the firms act in accordance with this realization. Air travel provides an example of an industry in which parallel behavior (not necessarily consciously parallel) is observed.7 Every few weeks the newspapers publish articles reporting a decision by a major airline to raise or lower its price, and the parallel actions of competitors. If, in 7

There are several other industries that have been described as exhibiting signs of tacit collusion. One prominent example is the “Ready to Eat” breakfast cereal industry, see F. M. Scherer, The Breakfast Cereal Industry, in W. Adams, ed., The Structure of American Industry (New York: Macmillan 1982); R. Schmalensee, Entry Deterrence in the Ready-to-Eat Breakfast Cereal Industry, 9 Bell J. Econ. 305 (1978).

74

Cartels

the case of a price increase, these actions result from a common perception that a move toward a higher price would lead to higher industry profits, then conscious parallelism accurately describes the firms’ conduct. The airline example also reveals a difficulty in defining conscious parallelism. Two different types of case fall within the general class of conscious parallelism antitrust cases. One type involves inference: we think there is an explicit agreement among the firms, but there is not enough evidence to meet the legal standard of proof. In the other, there is no explicit agreement, but we think there is tacit collusion. From an economic perspective, these are different cases. However, from the perspective of the antitrust enforcer they are for practical purposes the same. Declaring the second class of case immune from Sherman Act prosecution means that you effectively cannot prosecute cases within the first class. Among legal academics focusing on antitrust law, James Rahl probably first noted the problems generated by the theory of oligopoly and the traditional conspiracy requirements of antitrust law.8 Although the later work of Donald F. Turner9 and Richard A. Posner10 has overshadowed his discussion of conscious parallelism, I will use Rahl’s piece as a starting point because it does a good job of setting out the fundamental issues. Rahl focused on two problems: conscious parallelism and the intraenterprise conspiracy doctrine. Both are similar, he argued, in that they twist conspiracy doctrine beyond recognition. Though vague about the specifics, Rahl suggested that the conspiracy requirement serves useful purposes in antitrust and other areas of criminal law. Doctrinal innovations that threaten to water down or eliminate the conspiracy requirement should be disfavored in his view, in spite of the theoretical work supporting them. Rahl also expressed concern that the conscious parallelism and intraenterprise conspiracy doctrines might lead to the extension of conspiracy doctrine into new areas that were best left as matters of private law.

8

9

10

James A. Rahl, Conspiracy and the Anti-Trust Laws, 44 Illinois Law Review (Northwestern University) 743 (1950). I refer here to antitrust law commentators. The problems discussed by Rahl had been identified earlier, on a more general level, by Edward Chamberlin and other economists. Donald F. Turner, The Definition of Agreement under the Sherman Act: Conscious Parallelism and Refusals to Deal, 75 Harv. L. Rev. 655 (1962). Richard A. Posner, Oligopoly and the Antitrust Laws: A Suggested Approach, 21 Stanford L. Rev. 1562 (1969).

II. Conscious Parallelism

75

These issues are very much alive and perhaps more relevant today. The theory of oligopoly has continued to expand, and is now probably the largest area of research in applied game theory.11 Economists have focused on the analysis of strategic interactions. The theoretical literature supporting aggressive enforcement under Section 1 has expanded far beyond the stage it had reached when Rahl came to the problem. Increasingly, enforcement authorities face a decision whether to abandon the strictures of Section 1 conspiracy doctrine or to ignore some sophisticated research in economics.

B. Strain on the Conspiracy Doctrine The heart of the tension Rahl described is not hard to understand. First, start with the definition of conspiracy in the criminal law. It consists of three elements: (1) agreement, that is, the prosecution must prove the existence of some sort of agreement; (2) duality, that is, the agreement must involve at least two parties; and (3) irrelevance of the probability of harm, in the sense that there is no need to prove that the parties carried out or had the ability to carry out the illegal act, it is enough that they had a plan to carry it out. Sherman Act jurisprudence has adopted this definition of conspiracy in full. Now consider the economics of conscious parallelism. To the extent that it leads to tacitly collusive behavior, it has the potential to produce the same social losses as does monopoly or explicit price fixing. The interesting twist provided by the then-emerging wave of theoretical literature12 was the suggestion that collusive behavior could result in the absence of an explicit agreement. Firms recognizing their interdependence could tacitly agree to set price at the monopoly level. The conflict between theory and doctrine should now be evident. Theory suggests that consciously parallel behavior is not conspiracy in the traditional sense. However, it also suggests that parallel pricing potentially violates the antitrust laws because the end result may have an impact equivalent to price fixing. 11

12

To see the inroads applied game theory has made into oligopoly theory, see Jean Tirole, The Theory of Industrial Organization Ch. 6 (1988). The research was heavily influenced by Edward Chamberlin, Duopoly: Value Where Sellers Are Few, 43 Quarterly Journal of Economics 63–100 (1929); id., The Theory of Monopolistic Competition (Cambridge: Harvard University Press, 1933). See also Paul Sweezy, Demand under Conditions of Oligopoly, 47 Journal of Political Economy 568–73 (1939) (kinked-demand curve model of oligopoly).

76

Cartels

C. A Digression on the Theory of Interdependence Although Rahl cited the work of Chamberlin and other economists, he made no attempt to provide an explanation of the theory of interdependence, or how interdependence could lead to tacit collusion. This is a good place to provide the missing pieces of the argument, which more recent developments in industrial organization theory have updated.13 As noted in Chapter 1, the theory of interdependence probably began with Augustin Cournot’s mathematical model of competition in 1838. In the Cournot duopoly model, total output equals the sum of the outputs of each duopolist. Market price, which is a function of total output, is therefore determined by the output choices of the duopolists. Since the marginal revenue of each duopolist is a function of market price, and since profit-maximization occurs when marginal revenue equals marginal cost, the profit-maximizing output choice of each firm will be influenced by the output of the other firm. In particular, when one firm is deciding whether a certain output level will maximize its profits, it will have to anticipate the output level of the other firm. Hence, the duopolists’ output decisions are interdependent. While Cournot’s framework clearly illustrates the concept of interdependence, the model is less useful for analyzing pricing decisions. Joseph Bertrand explored pricing decisions in a duopoly in 1883. He concluded that in equilibrium both firms set price equal to marginal cost.14 Thus, in order to provide a rigorous economic account of tacit price collusion, one has to avoid Bertrand’s paradox. The Bertrand paradox – that duopoly results in competitive pricing – is the central thorn in the side of tacit collusion theories. The work on which Rahl and other antitrust commentators relied had not really resolved the problem.15 However, recent developments in game theory have provided a solution of sorts to the Bertrand puzzle, and thus an 13

14

15

For a clear and rigorous presentation, see Tirole, Theory of Industrial Organization, 240–62. See, for example, Tirole, The Theory of Industrial Organization, at 210. The result and the reasoning are similar to the Prisoner’s Dilemma and “cartel instability” problems discussed in notes 1 and 2 of this chapter. That is, not under the assumptions Bertrand imposes. There were solutions based on modifications of the assumptions of the Bertrand model. For example, Edgeworth introduced capacity constraints as a solution; and Hotelling introduced product differentiation. For discussions of these solutions, see Chamberlin, The Theory of Monopolistic Competition (7th ed., 1960), supra note 12, at 37–40, 226–8 (respectively).

II. Conscious Parallelism

77

account of the conditions under which interdependence gives rise to tacit price collusion. There are several solutions to the Bertrand puzzle suggested in the game theory/industrial organization literature. I will not attempt to survey them here.16 However, one approach assumes the game (i.e., the simultaneous choice of prices) repeats infinitely. Suppose the firms respond to deviations from the monopoly price by pricing at average cost (zero profit level) forever – a type of “trigger” strategy.17 Then unless the firms discount the future very heavily, the outcome in which they maintain monopoly price (i.e., collude) is an equilibrium. Alternatively, the monopoly price equilibrium may persist if firms play a “tit-for-tat” strategy of responding to every deviation from the monopoly price with a round of competitive pricing.18

D. Should a Conspiracy Statute be used to Regulate Consciously Parallel Behavior? How should a conspiracy statute deal with the problem of conscious parallelism? Rahl saw no reason in theory not to attack this conduct under Section 1; the serious problems, as he saw them, arose in practice. First, there is the evidence problem. The conspiracy requirement means there must be evidence of an agreement. But, by definition, consciously parallel behavior means that there is no such evidence. One solution is to relax the requirement of proof of an agreement in order to permit the Sherman Act to reach these cases. Rahl claimed that several cases of that time seemed to have taken this step. One example he cited is Interstate Circuit,19 in which the Court upheld a jury finding of conspiracy in violation of Section 1 on the basis of circumstantial evidence. He cited American Tobacco20 as another example, in which the Court found cigarette manufacturers guilty of a conspiracy to monopolize, largely on circumstantial evidence. 16 17

18

19 20

For a thorough survey, see Tirole, supra note 11. On trigger strategies and maintenance of collusion, see, for example, Robert Gibbons, Game Theory for Applied Economists 91 (Princeton, NJ: Princeton Univ. Press, 1992). For a general discussion of the stability of tit-for-tat strategies in this and other contexts, see Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, Inc., 1984). See also D. K. Osborne, Cartel Problems, 66 Am. Econ. Rev. 835 (1976); Griffin & Xiong, supra note 5. Interstate Circuit v. United States, 306 U.S. 208 (1939). American Tobacco v. United States, 328 U.S. 781 (1946).

78

Cartels

Rahl envisioned this relaxation continuing until the courts adopted something similar to a res ipsa loquitur approach under Section 1.21 Having reached this point, conspiracy would essentially be inferred from structure. But this is the point at which the weaknesses in an effort to use the Sherman Act to attack conscious parallelism become evident. First, the Sherman Act generally controls conduct, not structure. A firm with monopoly power can set its price at the monopoly level without violating the Sherman Act, as long as it has not engaged in exclusionary conduct that violates the Act.22 An enforcement strategy based entirely on circumstantial evidence, itself heavily determined by structure, arguably breaches this fundamental policy. Second, there is a core inconsistency that may frustrate enforcement efforts. Theory informs us that consciously parallel behavior is unilateral, in the sense that it results not from agreement but from each firm doing what is best for itself, taking into account the likely decisions of competitors. Enforcing the Sherman Act in an industry in which consciously parallel behavior is observed requires one to infer conspiracy from an observation of structure. But the inference of conspiracy is inconsistent with the theory at the foundation of the attack. Rahl noted some other problems. One is essentially that of Nash v. United States.23 The criminal laws specify acts that are punishable. But an approach that infers conspiracy from structure avoids the business of ever spelling out which acts are punishable. Another is that the res ipsa approach, after becoming an accepted practice under the Sherman Act, might spread to and distort other areas of criminal enforcement.

E. Remarks on Intraenterprise Conspiracy Rahl shifted focus to the intraenterprise conspiracy doctrine, which courts have relied on to uphold findings of conspiracy among parent and subsidiary corporations. Section 1 of the Sherman Act does not apply to actions by persons within a single business enterprise, or to acts involving a corporation and its officers, or parties to a partnership agreement.24 21

22 23 24

The doctrine permits a judge to submit a question of negligence to the jury even though the plaintiff is unable to present direct evidence on the precaution efforts of the defendant. See William L. Prosser, Handbook of the Law of Torts 211–28 (1971). See Chapter 10. See Chapter 3. See United States v. Joint-Traffic Assn., 171 U.S. 505, 567–8 (1898).

II. Conscious Parallelism

79

The intraenterprise conspiracy doctrine has operated outside of these clear boundaries. The doctrine dispenses with the duality requirement and permits courts to find conspiracy within a commonly owned enterprise. The intraenterprise conspiracy and conscious parallelism doctrines generate similar difficulties in distinguishing conduct from structure. The original justification for the doctrine was that it prevented organizations from hiding behind the shield of common ownership when they conspired in violation of Section 1. For example, in Yellow Cab v. U.S.,25 the Court said that the defendant could not escape liability by directing attention to its form (single business enterprise), because the intraenterprise conspiracy doctrine aims at substance rather than form. Rahl argued that the doctrine accomplished the opposite, because there is always agreement within a single enterprise. The conscious parallelism doctrine permits (or invites) courts to infer anticompetitive conduct from observations on the structure of the market and of parallel behavior among competing firms. The intraenterprise doctrine permits courts to infer anticompetitive conduct from observations on structure and evidence of agreement among defendants. While the former weakens the requirement of proof of agreement, the latter weakens the requirement of duality. Both doctrines, in effect, permit the Court to substitute or to put greater weight on observations of structure to make up for some missing piece of the conspiracy proof. To the extent that the intraenterprise conspiracy doctrine shifts the evidentiary burden toward the structure prong of conspiracy, it raises problems similar to those generated by the conscious parallelism theory. Again, because structure assumes a more important role in the proof of conspiracy, and the Sherman Act has been interpreted as punishing conduct rather than structure, the prosecution’s argument will suffer from an internal inconsistency. In addition, the failure to indicate precisely what sort of behavior violates the Sherman Act raises notice problems for potential defendants. The Supreme Court’s most recent pronouncement on intraenterprise conspiracy, Copperweld Corp. v. Independence Tube Corp.,26 scrapped a large part of the doctrine. The issue in Copperweld was whether a parent and its wholly owned subsidiary are capable of conspiring in violation of Section 1. The Court held that they are not. 25 26

332 U.S. 218 (1947). 467 U.S. 752 (1984).

80

Cartels

The ruling solves the problems Rahl identified but suffers from the artificiality that the Court was trying to avoid during the period when it accepted the intraenterprise conspiracy doctrine. In Copperweld, the Court merely redrew the boundary separating “persons who can make up a conspiracy” from those who cannot. Before Copperweld, “persons who could conspire” included a corporation and its wholly owned subsidiary. It was understood on the other hand, that this did not include a corporation and its officers or employees. To see the artificiality of the approach in Copperweld,27 ask yourself why the decision seems sensible. The reason is that even though the firm itself is a contract that restrains trade, the social benefits probably outweigh the social costs.28 Copperweld dealt with a case of a wholly owned subsidiary, but what about 80, 70, or 51 percent control? Copperweld did not address this squarely, and subsequent lower court decisions seem to be in disarray if one focuses only on the percentage of ownership by the parent corporation.29 However, the percentage of ownership is not the only question of concern. The rationale in Copperweld suggests that the holding applies to instances in which the “parent may assert full control at any moment if the subsidiary fails to act in the parent’s best interests.”30

F. Turner on Conscious Parallelism Following Rahl, the next substantial treatment of the conscious parallelism issue was that of Donald F. Turner.31 Turner agreed with Rahl’s 27

28

29

30 31

Why, then, would the Court take such an artificial approach? Probably because of the per-se illegality rule (Chapter 5), which applies to price-fixing conspiracies. The existence of such a rule introduces an incentive on the part of the Court to declare that this or that agreement is “not a conspiracy” rather than assert the more ambiguous argument that even if this is a conspiracy it is not an unreasonable one. This is the approach the Court took in Continental T.V. v. GTE Sylvania, 433 U.S. 36 (1977), when it analyzed the legality of nonprice vertical restraints. Some lower courts have extended the Copperweld rule to 82, 80, and 51 percent ownership cases; see Viacom Int’l, Inc. v. Time, Inc., 785 F. Supp. 371, 374, 384 (S.D.N.Y. 1992) (82 percent); Rosen v. Hyundai Group (Korea), 829 F. Supp. 41, 45 n.6 (E.D.N.Y. 1993), Novatel Communications Inc. v. Cellular Tel Supply, Inc., 1986–2 Trade Ca. (CCH) Parag. 67.412 (N.D. Ga. 1986). Other courts, however, have not applied Copperweld in cases involving 79, 75, and 54 percent; see Tunis Bros. Co. v. Ford Motor Co., 763 F.2d 1482, 1495 n.20 (3d Cir. 1985); Aspen Title & Escrow, Inc. v. Jeld-Wen, Inc., 677 F. Supp. 1477 (D. Or. 1987); American Vision Cntrs. Inc., v. Cohen, 711 F. Supp. 721 (E.D.N.Y. 1989). 467 U.S. at 771–2. Donald F. Turner, The Definition of Agreement under the Sherman Act: Conscious Parallelism and Refusals to Deal, 75 Harv. L. Rev. 655 (1962).

II. Conscious Parallelism

81

view that consciously parallel decisions should not constitute an agreement. However, he also argued that even if such decisions were held to constitute an agreement, they should not be deemed unlawful. Turner started by noting that the mere observation of consciously parallel behavior is not informative. The difficult part is drawing an inference on the nature of coordination in the market. Turner distinguished four patterns. a. parallel behavior accompanied by strong evidence of independence. This is a case of independence where the evidence suggests that each producer would have maintained his approach even if the others changed policy. For example, suppose a group of banks refuse to loan money to someone with poor credit history. An individual bank would still refuse to make the loan even if a rival bank suddenly began extending loans to bad credit risks. b. perfect competition. Assume an infinite number of sellers. Competition yields the same price, but this is no evidence of agreement. Evidence of parallel behavior is by itself uninformative in this setting. c. real conspiracy. Assume a large number of sellers who follow one leader; or who maintain price after a fall in costs, or after a decline in demand leaves each firm with excess capacity. In this case, it is hard to imagine that this could happen within a large group without an agreement among the players. What is important here is that the facts suggest an absence of rational behavior on an individual level. d. interdependence. Suppose three sellers maintain price after a decrease in the price of an important input, or after a fall in demand leaves them with excess capacity. This is a case of interdependence because this behavior may possibly result even without an explicit agreement. On the other hand, there may be tacit agreement since each player knows the impact of its decisions on the others. Turner pointed to the interdependence case as raising difficult issues under the Sherman Act. He referred to it as an “agreement” or an “agreement to agree,” but in either case he viewed it as a conspiracy. He also described this as independent behavior, in the sense that each firm is simply trying to maximize its own profits. Interdependence results because in an oligopolistic market, a firm that tries to maximize its own profits will have to anticipate the actions of its rivals. However, the fact that consciously parallel conduct can be characterized as independent under the most conservative view was not enough to discard the Sherman Act concerns according to Turner. There are instances in antitrust law where certain conduct presents an issue under

82

Cartels

the statute because a firm has market power, whereas the same conduct would not be a concern if the firm operated in a competitive industry.32 Turner distinguished these cases on the ground that they involve a policy of containment, of preventing firms from leveraging monopoly power into nonmonopolized markets. Indeed, all Section 2 monopolization cases involve some wilful maintenance, acquisition, or extension argument.33 Thus, according to Turner, if an attack on consciously parallel conduct uses the theory that the involved firms should be treated differently from firms in a competitive setting; that is, suit is brought under the Sherman Act for the offense of charging oligopoly prices, then it must remain consistent with the other cases based on this theory. This implies that oligopoly power obtained by means that would not violate Sections 1 or 2 should be immune from antitrust prosecution, which would immediately cut out a large set of cases that enforcers would have wanted to prosecute. What about those cases that are not immune from prosecution under Section 1? The difficulty, in Turner’s view, is fashioning an appropriate injunction. If consciously parallel behavior is unilateral or independent, then how can it be restrained under the statute? What should an antitrust court tell a firm to do: stop taking the actions of your competitors into account in pricing decisions? Because there is no easily enforceable injunction, the only solution Turner could see is a form of price regulation administered by the courts. But this would convert federal courts into public utility regulatory commissions, a level of judicial intervention that was not anticipated by the framers of the Sherman Act.34 The remainder of the Turner article discusses examples of conscious parallelism. Turner’s purpose probably was to show that his argument did not entail a total retreat from the challenge of controlling consciously parallel pricing under the Sherman Act, but his examples appear in retrospect to be poor. One was basing-point pricing. Turner asserted that these are easy cases, in the sense that it is easy to demonstrate that the practice violates the Sherman Act. However, in Chapter 8 we will see that modern research throws this conclusion into serious doubt. Another example of an obvious violation of the antitrust laws, in Turner’s view, is

32 33 34

See, for example, discussion of refusals to deal in Chapter 10. See Chapter 10. The earliest Supreme Court decisions under the Sherman Act have made this point clear. See Chapter 5.

II. Conscious Parallelism

83

retail price maintenance. Again, recent research casts doubt on this conclusion. Further, the law itself has virtually swung around to the position that resale price maintenance is not harmful. It remains per-se unlawful,35 but the Court has essentially accepted the reasonableness justifications for the practice.36

G. Reexamining Rahl and Turner under the Cournot and Bertrand Models Although the Rahl and Turner analyses of conscious parallelism differed in many respects, they agreed on the premise that the conduct at issue could be described as unilateral, and therefore determined largely by market structure. In this section, I will use the Cournot and Bertrand duopoly models to shed additional light on their positions. I have noted already that the simplest version of the Bertrand model yields the competitive price as the equilibrium; unless the firms are able to collude with respect to price, in which case they would choose the joint-profit maximizing (or monopoly) price. In the Cournot model, each firm chooses its output level in order to maximize its own profits. In the case of two firms, total output is simply the sum of the output of each firm. As one firm increases output it must consider the direct effect of its output increase on the market price, and the indirect effect its increase has on the market price by altering the incentives of its competitor. The simplest version of the Cournot model generates a rich set of potential equilibrium outcomes. One is the joint-profit maximizing equilibrium, where firms produce quantities that maximize joint profits and charge the monopoly price. Another is the Cournot equilibrium, which results when each firm takes the other’s output as fixed and maximizes its own profits with respect to the remaining market. Because each firm assumes that the output of the competitor is fixed, the indirect effect of expansion on the market price is zero. The direct effect causes the firm to reduce its output below the level that it would choose if the firm took the market price as given (i.e., acted as a price-taker). However, as the number of firms increases, the direct effect of each firm’s output increase declines; thus, the Cournot equilibrium price approaches the perfectly competitive price as the number of firms increases. 35 36

Dr. Miles Medical Co. v. John D. Park & Sons Co., 220 U.S. 373 (1911). See Continental T.V. v. GTE Sylvania, 433 U.S. 36 (1977).

84

Cartels

Figure 4.2

The third equilibrium of interest is the Stackelberg outcome (or “first mover” case), of which there are two. Suppose firm A knows how firm B will respond to its output decision, and firm A can commit itself to a certain output level. Suppose, on the other hand, firm B takes the output choice of A as given. Then A, moving first, will choose the “Stackelberg leader” quantity. In the equilibrium in which A is the Stackelberg leader, A chooses an output level that maximizes its profits after taking into account B’s reactions. B, maximizing its own profits while taking the output of A as given, chooses the “Stackelberg follower” quantity. In this equilibrium firm A produces more and, since both charge the same price, makes a larger profit than B. Figure 4.2 illustrates the several duopoly outcomes, under some simplifying assumptions.37 The segment DI is firm B’s schedule of

37

Specifically, Figure 4.2 assumes demand and cost curves are linear in output. For a clear presentation, see Michael D. Intriligator, Mathematical Optimization and Economic Theory 206–10 (1971). If both duopolists assume correctly that the other will move along his schedule of best responses (reaction curve), then the equilibrium is the Bertrand result with price equal to marginal cost, see Timothy F. Bresnahan, Duopoly Models with Consistent Conjectures, 71 Am. Econ. Rev. 934–45 (1981).

II. Conscious Parallelism

85

“best responses” to a given output choice by A. That is, the schedule represents the output levels chosen by B when it maximizes its own profits, taking A’s output as fixed. Yet another way of describing this is that the segment DI shows B’s output choices when it acts unilaterally, taking A’s output choice as given. The segment HC is A’s schedule of best responses. Where the two best response schedules intersect, at point E, one finds the Cournot equilibrium. The segment CD shows the set of output combinations which maximize joint profits for the duopoly. Note that it intersects the best response schedules only at the monopoly output levels for the individual firms. Thus, any jointprofit maximizing outcome in which both firms produce some output (e.g., an equal market division scheme) cannot occur through unilateral action. For our purposes it is sufficient to distinguish the outcomes that can be described as unilateral or structurally determined from those that require either explicit or tacit collusion. The competitive, Cournot, and Stackelberg leader outcomes all qualify as unilateral, or structurally determined. In each of these, firms act independently. The differences among these equilibria result entirely from market structure, the sequence of actions, and the information possessed by the firms. The monopoly or joint-profit maximizing outcome is not unilateral or structurally determined, since it requires either explicit or tacit collusion to reach and maintain. Return now to the conscious parallelism theories of Rahl and Turner. Since both argued that consciously parallel behavior could be described as unilateral, the Cournot and Stackelberg leader outcomes serve as illustrations of their positions. However, the monopoly or joint-profit maximizing outcome seems inconsistent with the Rahl and Turner theories, because joint-profit maximization cannot result, in the models considered here, from purely independent action. This insight seems to be at the heart of Richard Posner’s critique of Rahl, which I will take up in the next section. The only remaining question is whether it is possible to describe a tacit collusion equilibrium as structurally determined. If so, then it is possible to say that the Rahl and Turner theories of conscious parallelism embrace a broader set of equilibrium outcomes, including the monopoly price outcome which most concerns antitrust authorities. If not, then their theories encompass a rather limited set of equilibrium outcomes, most if not all of which are beyond the reach of Section 1 of the Sherman Act.

86

Cartels

H. Posner on Conscious Parallelism Richard A. Posner argued that Section 1 of the Sherman Act can and should regulate consciously parallel oligopoly behavior.38 He began his argument by setting forth several criticisms of the theory of interdependence on which Turner relied; for example, that with sufficiently long time lags a firm could get away with a price reduction without being hurt by the responses of competitors, or that if the price cutter’s new sales came largely from new customers, competitors may not be badly hurt. In both of these cases each firm may have strong incentives to cut price even though there is a small number of firms in the market. But Posner’s key criticism was of a price leadership model he found at the heart of Turner’s interdependence theory. Posner’s criticism runs as follows. If it is rational for others to follow the leader in maintaining a high price, then why doesn’t the leader have an incentive to cut his price and restore it, enough to make a quick profit and then return to the leadership position? In other words, if the theory is that people will follow a leader who restrains competition because they realize that competition hurts all, then the argument cuts the other way, providing an explanation for the leader’s price cutting. Posner distinguished his theory of conscious parallelism from those of Rahl and Turner by asserting that consciously parallel oligopolistic behavior, if it is really having the effect of a price-fixing cartel, needs enforcement and coordination, however small the group. And it almost always requires a small group. As I suggested in the previous section, Posner’s theory of conscious parallelism focuses on (and therefore clearly embraces) the joint-profit maximizing outcome, and rejects the theory that joint-profit maximization could result from independent conduct. Posner relied to some extent on Stigler’s model (Part I.B of this chapter) to spell out the requirements of coordination and enforcement. Central to Posner’s argument is the assumption that stable tacit collusion requires monitoring and enforcement. Once these elements are present, all that is necessary for the application of Section 1 is in place. In view of the need for explicit coordination and enforcement to maintain reasonably stable collusion, Section 1 should be applied in this case. 38

Richard A. Posner, Oligopoly and the Antitrust Laws: A Suggested Approach, 21 Stanford Law Review 1562 (1969).

II. Conscious Parallelism

87

As for the conspiracy proof problem – that is, that there is no explicit agreement – Posner suggested a unilateral contract interpretation. If someone runs an ad offering to pay $10 to the person who finds and returns his dog, then the person who meets this condition has an enforceable contract for the promised award, even though he never communicated his willingness to accept the promisor’s offer. In the same sense, a seller who communicates by leading to a higher price makes a unilateral contract which others accept through their adherence. Thus, if American Airlines announces a value-pricing plan, and all other airlines adopt the terms of the plan, then there is agreement.

I. Rahl/Turner v. Posner While I agree with Posner’s criticism of the interdependence theory, its force is limited to the price-leadership model at the heart of Turner’s theory. However, as the review of interdependence theory earlier in this chapter makes clear, the price leadership model is not the only potential account of tacit collusion. A collusive, monopoly price equilibrium may come about through the adoption of strategies (e.g., trigger strategies, or tit-for-tat) that effectively deter price-cutting. If a plausible explanation could account for the maintenance of such an equilibrium through unilateral conduct, then the Rahl/Turner theories of conscious parallelism would again become important. But I think this is unlikely. I also think Posner was correct in concluding that agreement, monitoring, and enforcement are the essential attributes of a price-fixing cartel, and that once these features appear, Section 1 enforcement becomes justified. However, there are soft spots in his argument. The biggest one is his light treatment of the agreement requirement. His example of the unilateral contract is essentially an argument for dispensing with the requirement of proving agreement. In place of such a proof, Posner would prefer antitrust authorities to offer proof of monitoring and enforcement efforts or capabilities. But this is where Posner’s argument appears weakest, because he provided little guidance to enforcement authorities on this element of proof. He did explain how some pieces of circumstantial evidence (e.g., price discrimination, price rigidity) might aid in inferring conspiracy. But using circumstantial evidence to aid the inference of conspiracy is not equivalent to using it as a substitute for proving conspiracy. Presumably more evidence is required for the latter task, but how much more?

88

Cartels

Even if it were possible to set forth a detailed description of the type of monitoring and enforcement evidence sufficient to prove a high probability of conspiracy, there remains a significant problem. If the firms have adopted strategies that effectively deter price-cutting, then enforcement efforts will, by hypothesis, never be observed; and if information gets around fast and the firms’ penalizing strategies are sufficiently harsh monitoring may not be necessary. It follows that those instances in which evidence on monitoring and enforcement is available will be those in which tacit collusion efforts were least effective. And in the absence of such evidence, we have only circumstantial evidence, which returns us to the heart of the problem, that circumstantial evidence often points to several different hypotheses. Evidence that seems consistent with tacit collusion could be observed in a setting in which firms are operating in a noncollusive manner.39 Finally, I think there is something inescapable to the point Rahl/Turner stress that an aggressive enforcement approach toward consciously parallel behavior would raise notice problems. Posner suggested that it is sufficient for the authorities to set a penalty that appropriately deters anticompetitive behavior. But antitrust doctrine creates a set of rules, within the framework of a criminal statute. If firms are penalized for taking actions that have been routine practices within their industries, how are they supposed to know when they are in compliance with the law? The problem was introduced with the passage of the Sherman Act and is the issue at bottom in Nash v. United States.40 Recall that the Sherman Act criminalized an area of activity that private law had traditionally governed.41 Although it was not his aim, Turner demonstrated the potential folly in inferring conspiracy from evidence that falls short of proving agreement. Recall his clear cases of tacit collusion: basing point pricing and 39

40 41

For example, one empirical study of the “Ready to Eat” breakfast cereal industry (often described as an example of a tacitly-collusive market) finds that price-cost margins are almost entirely explained by product differentiation and “multi-product firm pricing.” See Aviv Nevo, Measuring Market Power in the Ready-to-Eat Cereal Industry, 69 Econometrica 307 (2001). Product differentiation raises price-cost margins by giving the firm a small degree of monopoly power. “Multi-product firm pricing” refers to the incentive of a producer of two imperfect substitutes to charge a higher price than would be set for the two by two different producers. 229 U.S. 373 (1913). For discussion, see Chapter 3. See Chapter 2. One often finds uncertain standards in areas of private law, such as the negligence standard in torts. But the traditional view is that the standards of criminal law should be clearer.

III. Conclusion

89

retail price maintenance. Modern research suggests that these activities may be economically efficient. Thus, a policy of inferring conspiracy from the mere existence of a certain pricing convention in an industry could be mistaken; practices that are efficient but not well understood would be taken in some cases as evidence of price-fixing.

iii. conclusion This chapter has introduced the basic theory of cartels and discussed its implications for enforcement of Section 1 of the Sherman Act. Generally, there are two issues that make enforcement difficult. One is determining precisely what conduct constitutes a conspiracy for purposes of Section 1. Put another way, what degree of explicitness is necessary for a court to find an agreement? The other issue is determining the amount of evidence necessary to prove the existence of a conspiracy. This issue is closely intertwined with the first, because if courts demand evidence of an explicit agreement (e.g., a written contract), then Section 1 would be ineffective against consciously parallel conduct and practically ineffective against many explicit cartel agreements.

5 Development of Section 1 Doctrine

Section 1 litigation today applies two legal tests: the rule of reason or reasonableness test and the per-se illegality rule. Under the rule of reason standard, the court holds a practice unlawful only if it unreasonably restrains trade. Under the per-se rule, the court may find a practice unlawful without an inquiry into its reasonableness. What is the connection between the per-se rule and reasonableness tests? What is the connection between the reasonableness test under the Sherman Act and the reasonableness test employed in the closest common law analogue, the restraint of trade cases? This chapter aims to answer these questions. I also try to provide some sense here of the underlying tensions shaping antitrust doctrine. The fundamental tension is between the administrative and evidentiary burdens imposed on public enforcement agents by a reasonableness standard, and the traditional tendency of courts to apply reasonableness tests in the common law process. This tradition tends to direct courts toward an inquiry into the economic reasonableness of a defendant’s conduct.

i. the sherman act versus the common law The first Section 1 case to reach the Supreme Court was United States v. Trans-Missouri Freight Assn.1 Although it receives scant attention in

1

166 U.S. 290 (1897). This was not the first federal antitrust case. The first Section 1 prosecution was United States v. Jellico Mountain Coal & Coke Co., 46 Fed. 432 (C.C.M.D. Tenn. 1891), which involved a cartel of coal mining companies. The first antitrust case to reach the Supreme Court was United States v. E.C. Knight Co., 156 U.S. 1 (1895) (holding

90

I. The Sherman Act Versus the Common Law

91

antitrust casebooks, it is the most important federal antitrust opinion, for reasons that should become clear shortly. It involved an effort by eighteen railroads controlling traffic west of the Mississippi River to eliminate rate wars. The railroads created the Trans-Missouri Freight Association to establish freight rates for all participants. In January 1892, the United States brought suit under the Sherman Act to have the agreement declared unenforceable, the association dissolved, and its practices enjoined. The defendants admitted the formation of the association, and conceded that it constituted a restraint of trade, in the sense that their contract limited each member railroad’s freedom of commercial action. However, they urged, first, that they were exempt from the Sherman Act because they were regulated carriers under the Interstate Commerce Act; and, second, that the rates fixed were legal because they were reasonable and therefore valid under common law. The lower court accepted these arguments.2 The Supreme Court reversed. Responding to the claim that the Interstate Commerce Act made the railroads exempt, the Court held that [t]he Interstate Commerce Act does not authorize rate bureaus. In particular, the Act prohibits pooling. The general type of contract setting up a rate bureau is not mentioned in the Interstate Commerce Act.

In the absence of explicit authorization – and indeed in the presence of an apparent denial of authority to set up rate bureaus – the Court refused to find an exemption merely because another federal statute also regulated the railroads. The question whether regulation displaces antitrust arises in a number of areas. The Court’s decision in Trans-Missouri establishes the basic principle that federal regulation does not automatically imply antitrust exemption. The modern law of exemption, a difficult area, has three major categories. First, there is exemption implied by federal regulation, which the decision in Trans-Missouri seems to reject. However, in some cases, courts have found that federal regulation implies exemption. For example, McLean Trucking v. United States3 involved the ICC’s approval of a merger of two trucking companies. Because the ICC had to consider the effect of a merger on competitors and on competition, the Court held

2 3

that manufacturing was not “interstate commerce” and therefore that the Sherman Act did not cover a monopoly of sugar manufacturing). 58 F. 58 (8th Cir. 1893). 321 U.S. 67 (1944).

92

Development of Section 1 Doctrine

that ICC approval implies antitrust exemption. The second form of exemption is express, or explicitly authorized. For example, the McCarran-Ferguson Act exempts the insurance industry from federal antitrust law, as long as it is regulated by the state. Third, there is exemption implied by state regulation. The doctrine is complicated, but the basic rule is that exemption will be implied when the state has erected a scheme that displaces competition according to a clearly articulated state policy.4 Recall that the defendants in Trans-Missouri also claimed that their rate agreement did not violate the Sherman Act because the rates fixed were reasonable and therefore valid under common law. The defendants were in essence telling the Court that the decision standard under the Sherman Act should be the same as that employed by common law courts in restraint of trade cases. The common law of trade restraints equated legal validity with “reasonableness.” If federal courts had accepted this principle in reviewing Sherman Act decisions, then contract enforcement under the common law of trade restraints and legality under the Sherman Act would both turn on proof of reasonableness. The Trans-Missouri Court, through Justice Peckham, rejected the common law standard. First, the Court held that the statute condemned “every contract in restraint of trade,” which includes some of the types of contract that would satisfy the common law reasonableness test. The Court noted that a noncompetition provision in a contract for the sale of property, included for the purpose of enhancing the price, might not fall into the category of condemned contracts – an apparent reference to Mitchel v. Reynolds. Although the Court’s reading of the Sherman Act appears to be literal, it is not quite: it is a holding that the set of contracts illegal under Section 1 is larger than and contains the set of contracts deemed unreasonable under the common law. Second, the Court responded to the defendants’ attempt to prove reasonableness. The defendants had argued that the high fixed costs in their industry made their rates reasonable. If they competed vigorously, they could not cover fixed costs, which would lead to a failure to maintain equipment and structures, bankruptcy for some businesses, and inconvenience to customers. Competition would drive out all competitors, leaving one firm, which would then raise its price to the monopoly level.

4

The state action exemption was articulated in Parker v. Brown, 317 U.S. 341 (1943), and has been elaborated in subsequent decisions. The opinions in this area reveal that: there can be no immunity under Parker without (1) adequate state supervision and (2) a clear state policy or purpose to displace competition, see Chapter 17.

I. The Sherman Act Versus the Common Law

93

Thus, according to the defendants, their agreement maintained a competitive structure in the long run. The Court said that it could not practically determine the reasonableness of rates, because it would require too much information and tax the capacity of judges. Courts would have to review business records to issue a decision, and continually monitor them to ensure that practices remained reasonable. The framers of the Act, the Court suggested, did not envision this level of supervision and intervention by courts.5 While it is almost surely valid that the framers of the Sherman Act did not intend federal court judges to review business records in the manner of a public utility regulatory agency, the administrative burden argument has questionable relevance in this case. Administering the rule of reason would turn into an extremely expensive task if the Sherman Act required courts to determine whether the defendant charged reasonable prices. But this was not the issue before the Court, and it was not the issue in the trade restraint cases that gave rise to the common law rule of reason standard. The real issue was a version of the question that had come before courts in restraint of trade cases. Stated generally, the question is this: should the firms’ competition-restricting agreement be declared legally invalid because its net effect, taking into account possible consumer benefits, is probably harmful to the public? In the common law cases, a decision against the firms meant that the agreement could not be enforced. Under the Sherman Act, a decision against the defendant implies the agreement is illegal. Common law courts were able to answer the question without delving into the reasonableness of the cartel’s price. And, since common law judges had proven that they could analyze this question, there is no reason to believe the Supreme Court could not have tackled the issue in Trans-Missouri. Antitrust students should keep in mind that this was a 5–4 decision. The Court decided the central issue, whether the Sherman Act should employ the same decision standard as the common law courts, by one vote. This is troubling because the common law of contracts in restraint of trade had developed over a period of roughly five hundred years before the passage of the Sherman Act. With one bold stroke, the Supreme Court cut the cord connecting then-undeveloped Sherman Act 5

Note that this is essentially the “public utility” argument discussed in Chapter 4 (discussion of Donald F. Turner’s position on conscious parallelism).

94

Development of Section 1 Doctrine

case law to its most likely foundation in common law doctrine. Justice White noted this in a spirited dissent. One of the concerns Justice White raised was that the majority’s literal approach carried the implication that many ordinary business organizations such as partnerships violated the Sherman Act because they restrained trade. The Court addressed that issue directly in United States v. Joint Traffic Assn.6 The case involved another association formed to fix rates among competing railroads. The defendants tried to prove reasonableness by making the same ruinous competition argument employed by the defendants in Trans-Missouri. The Court rejected the defendants’ proof of reasonableness, and addressed the argument that its construction of the Sherman Act in Trans-Missouri rendered illegal all partnerships and other forms of business organization. These, said the Court – referring to partnerships, corporations, leases, and sales of goodwill accompanied by restrictive covenants – do not come within the legal definition of restraint of trade. In short, the Court repeated the key holding in Trans-Missouri, that the reasonableness test of the common law is not to be applied, with minor elaborations. In place of the reasonableness test, the Court recognized a fixed set of contracts outside of the legal definition of trade restraint (noncompetition covenants connected to a sale of property or goodwill, partnerships, corporations). Under the reasoning of Joint Traffic, contracts outside of this fixed set violate the Sherman Act if they restrain trade.

A. Aside on Ruinous Competition The defendants in Trans-Missouri and in Joint-Traffic attempted to justify their agreements as efforts to avoid ruinous competition. What is that? Ruinous competition appears in industries with high fixed costs. A high level of fixed costs implies that average cost (cost per unit of output) declines as output expands. Thus, a firm can expand output, continually reduce its price, and still make a profit on the margin, provided that the decreasing unit cost phenomenon holds over the entire range of output. In this cost structure, marginal cost is always less than average (or unit) cost (see Chapter 1). Thus, if the firm follows the ordinary profitmaximization rule of producing up to the point where price equals 6

171 U.S. 505 (1898).

I. The Sherman Act Versus the Common Law

95

marginal cost, it will have an incentive to expand output, even though it may not be able to cover its per-unit cost. Economists have typically responded to the ruinous competition argument by pointing out that it is inconsistent with the evidence. If costs fall through the entire range of output, price-cutting competition would continue until only one firm survived. Alternatively, mergers would result in the same end. If several firms remain in competition many years after the start of the industry, as happened with railroads, then unit costs probably do not fall through the entire scale of output. Recently theorists have argued that the fact unit costs do not fall through the entire range of output is not inconsistent with and may lend support to the ruinous competition thesis. Suppose costs rise sharply at some level of output. That level establishes a capacity constraint for the firm. Suppose all firms are capacity-constrained, and suppose the sum of their capacities exceeds the entire demand for the industry’s output. Figure 5.1 provides an illustration for the case of two firms. Under these conditions, competition may lead to cycles of price wars, and firms would suffer losses as a result. For example, if you choose any allocation of total output among the firms in Figure 5.1, where price is greater than marginal cost, at least one firm will have an incentive to expand output by cutting its price. This incentive remains in the short run even when price is below average cost. The likely result of competition is an outcome in which both firms fail to make enough revenue to cover their total costs. The short-run losses would cause some firms to fail to invest in fixed, specialized assets or to exit the market; and these

Figure 5.1

96

Development of Section 1 Doctrine

actions, in turn, would lead to interruptions in supply or service to some segments of the market. Consider the following example.7 Suppose widgets are produced by two firms under the conditions shown in Figure 5.1. The maximum consumers are willing to pay for a widget is $4. Widget makers incur only fixed costs; the total cost of producing widgets is $5 (regardless of the quantity produced) and the marginal cost is $0. The two widget makers, call them A and B, each have one plant with enough capacity to produce five widgets (per month). The total number of widget consumers demand (per month) is eight. Thus, as in Figure 5.1, total capacity (ten widgets) exceeds total demand (eight widgets). Suppose widget makers A and B start off with an agreement to divide the market in half, each charging $4 per widget. Since each sells four widgets, they each incur a unit cost equal to $5/4 = $1.25. Thus, each makes a profit equal to ($4 - $1.25)4 = $11. Suppose B defects from this agreement by cutting price to $3.5. If A keeps his price at $4, B will sell five widgets, use his full capacity, and lower his unit cost to $5/5 = $1. Thus, B’s profit, assuming A does not match his price cut, is ($3.5 - $1)5 = $12.5. A’s profit, assuming he does not match the price cut, would be ($4 - $5/3)3 = $7. On the other hand, if A matches B’s cut, and they continue each getting half the sales, they will each get a profit of ($3.5 $1.25)4 = $9. Note that this example looks like the cartel problem described at the start of Chapter 4. Indeed, the payoffs in this example follow the classic Prisoner’s Dilemma structure described in that chapter. If A sticks with the high price of $4, his payoffs are either $11 (if B prices at $4 too) or $7 (if B cuts price to $3.5). If A cuts his price his payoffs are either $12.5 or $9. Cutting price to $3.5 thus appears to be a “dominant strategy,”8 in the sense that A is better off cutting his price no matter what B does. The outcome of this “one shot game,” in which A and B face the option of sticking to a price of $4 or cutting price to $3.5, is one in which both firms charge $3.5 and make a profit of $9 each. But the game does not end here. We have merely returned to our starting point in terms of the incentives governing A and B. Taking the price of $3.5 as the starting point, with each firm selling four widgets and making a profit of $9, we can repeat the preceding argument to show that they will drive the price

7

8

This example is based on the excellent informal discussion in George Bittlingmayer, The Economic Problem of Fixed Costs and What Legal Research Can Contribute, 14 Law & Social Inquiry 739 (Fall 1989). See note 2 of Chapter 4.

I. The Sherman Act Versus the Common Law

97

down to $3, and then $2.5, and so on until they are both failing to cover their total costs. To see the last step, suppose they have both driven price down to $1.25, so that they are breaking even when selling four widgets. Even at this price, each firm has an incentive to cut price, say to $1.10. But this results in both firms failing to cover their costs when selling 4 widgets, since $1.10 - $1.25 = -$.15. The price-fixing agreement attempts to avoid the price war problem just described and to bind the firms to a long-term course of action that permits them to invest appropriately in equipment and structures. If new competitors can enter the industry easily, the parties to the price-fixing agreement will not earn more than the competitive rate of return. In other words, the threat of new entry will force the firms to set price close to average cost (or at average cost, net of the cost of entry). Although the standard position in the economics literature remains one of skepticism toward the destructive competition thesis, recent research has led some to rethink the matter.9 The topic remains controversial,10 but it is probably more open to debate now, in the sense that the available research expresses a variety of opinions, than at any other time. If the destructive competition thesis were valid, one might think that the early Supreme Court antitrust decisions would have led to the decline of the railroad industry. However, the industry’s long history of regulation makes it difficult to test this hypothesis. Return to TransMissouri. The Trans-Missouri Freight Association replaced an earlier traffic pool that ceased after the enactment of the Interstate Commerce Act of 1887. At the time of the government’s suit, the Interstate Commerce Commission favored and aided the Justice Department’s efforts to enforce the Sherman Act against the railroads.11 However, over the early 1900s, the ICC’s scope of regulation increased and the agency entered into a cooperative relationship with the railroads.12 The ICC 9

10

11 12

See Lester G. Telser, Economic Theory and the Core (1978); id., A Theory of Efficient Cooperation and Competition (1987); William W. Sharkey, The Theory of Natural Monopoly (1982); George Bittlingmayer, Decreasing Average Cost and Competition: A New Look at the Addyston Pipe Case, 25 Journal of Law & Economics 201–29 (1982); Stephen Craig Pirrong, An Application of Core Theory to the Analysis of Ocean Shipping Markets, 35 Journal of Law & Economics 89 (1992). For a recent skeptical view of core theory, see John Shepard Wiley, Jr., Antitrust and Core Theory, 54 Univ. of Chicago Law Review 556 (1989). Donald Dewey, Monopoly in Economics and Law 154 (1959). Specifically, the job of checking for rate discrimination was greatly simplified if ratemaking bureaus did some of the work (by not filing rates that appeared to be discriminatory). See Dewey, supra, at 154. This is consistent with notions of regulatory capture,

98

Development of Section 1 Doctrine

tacitly approved horizontal price fixing for a period of many years beginning some time after the Trans-Missouri decision.13 Later, when the Justice Department renewed its efforts against the railroads, the ICC opposed it. Congress resolved the conflict between the Justice Department and ICC by passing the Reed-Bulwinkle Act of 1948, which legalized rate agreements among railroads and motor carriers that received the approval of the ICC. Railroad pricing is largely unregulated today.

B. Validity Crisis and Resolution The two earliest Supreme Court decisions on the Sherman Act, TransMissouri and Joint-Traffic, established that the set of contracts made illegal by the Sherman Act is larger than and contains the set of contracts held invalid under the common law of trade restraints. This is an easy proposition to understand, but the Supreme Court could not maintain it for long. To see the problem confronting the Court, start with the basic proposition that the common law has equated legal validity with reasonableness. Although one can find many discussions of the role of reason in the common law process, William Blackstone’s Commentaries provided the following early and influential account. [I]t is an established rule to abide by former precedents, where the same points come again in litigation; . . . Yet this rule admits of exception, where the former determination is most evidently contrary to reason. . . . For if it be found that the former decision is manifestly absurd or unjust, it is declared, not that such a sentence was bad law, but that it was not law; . . . And hence it is that our lawyers . . . tell us, that the law is the perfection of reason, that it always intends to conform thereto, and that what is not reason is not law.14

In other words, judges must justify legal rules by an appeal to reason. In the commercial context, we typically see common law rules justified by economic reasonableness arguments, that is, arguments that refer to the relative costs and benefits of alternative arrangements induced by changes in the law. The rule of reason articulated in the common law restraint of trade cases is based on such an argument. For example, the trade restraint examined in Mitchel v. Reynolds was deemed reasonable

13 14

see George J. Stigler, The Economic Theory of Regulation, 2 Bell J. Econ. & Management Science 3 (1971). Dewey, supra note 11, at 154. William Blackstone, Vol. I, Commentaries on the Laws of England *69–70.

I. The Sherman Act Versus the Common Law

99

because the benefits it brought to the contracting parties and the public seemed to outweigh the potential public harm. Because the common law process has equated legal validity with reasonableness, for the Supreme Court to defend a decision standard that differs from the common law would eventually require either: (1) some proof that the common law decisions were actually unreasonable, so that the Sherman Act merely granted federal courts a chance to start over at crafting a truly reasonable law of competition, or (2) that the two legal regimes in fact had the same decision standard, so the Sherman Act merely afforded courts an opportunity to clarify and expand the law of competition. One of these arguments was necessary because if the common law was reasonable, and the Sherman Act adopted a standard foreign to the common law, then it follows that Sherman Act jurisprudence could not have a firm base in economic reasonableness. Thus, the principal holding of both Trans-Missouri and Joint-Traffic could only serve as a temporary answer to questions concerning the Sherman Act’s validity, because the holding invited an endless stream of legal challenges.15 The escape hatch was opened by Judge Taft in United States v. Addyston Pipe & Steel Co.16 The Government sought to enjoin the activities of six manufacturers of cast-iron pipe who had formed the Associated Pipe Works for the purposes of rigging bids on municipal government contracts, pooling the receipts, and fixing prices. The complicated agreement provided that when municipalities advertised for bids, the privilege of submitting the lowest bid would go to the member that offered to pay the largest bonus into a common fund.17 The defendants argued that their prices were reasonable, and that the arrangement was a necessary response to what would otherwise be destructive competition. The important part of the Court’s opinion comes in the treatment of the respective boundaries of the Sherman Act and the common law. Taft read the common law as voiding all price-fixing agreements not ancillary to some legitimate cause. Furthermore, the legitimate causes recognized by common law courts were limited in his view. They consisted of roughly 15

16 17

In using the label “validity crisis” to describe this tension, I have borrowed from the more general treatment of validity crises and the common law in Gerald J. Postema, Bentham and the Common Law Process (1986). 85 F. 271 (6th Cir. 1898), aff’d, 175 U.S. 211 (1899). Why do it this way? Answer: to whom is the contract worth the most? The most efficient member of the cartel will put in the largest bid for the contract. Assigning the contract to the most efficient member is the profit-maximizing strategy for the cartel.

100

Development of Section 1 Doctrine

the following five types of covenant: (1) by the seller of property not to compete with the buyer; (2) by a retiring partner not to compete with the firm; (3) by a partner pending the partnership not to interfere with the business of the firm; (4) by the buyer of property not to use it in competition with the business retained by the sellers; and (5) by an apprentice not to compete with the master after expiration of his time of service. While Taft did not say that this is an exhaustive list of the restraints that had been found reasonable, he suggested that this is all there is to the set of reasonable contracts. In addition, he warned that to attempt to determine reasonableness outside of this set is to “set sail on a sea of doubt.”18 Taft noted that some common law courts had sailed in this fashion, but only because they failed to understand “the proper limits of the relaxation of the rules for determining the unreasonableness of restraints of trade.”19 Taft’s response to the defendants’ attempt to prove reasonableness was consistent with his general argument. He rejected the notion that the Court has the responsibility to determine economic reasonableness, because it would lead in his view to a vague and indeterminate standard. He argued that common law courts did not usually operate in this fashion, that is, they did not attempt to determine reasonableness by applying general cost-benefit reasoning in every case. He refused to establish what he regarded as a novel and uncertain standard under the Sherman Act. Thus, Taft’s solution to the validity crisis generated by Trans-Missouri read the common law in a manner that indicated that the set of contracts invalid in common law courts is equivalent to the set of contracts the Sherman Act declared illegal. This solution has endured: the Supreme Court quickly climbed on board and has remained ever since. The validity of Taft’s argument is an altogether different matter. One could view this simply as an empirical matter: Count up the common law restraint of trade opinions and see whether few or many fall outside of the five categories Taft identified. The evidence does not support Taft’s position. A substantial portion of the trade restraint opinions apply the rule of reason to agreements that were not ancillary to some other business transaction.20 While a quantitative empirical study of the common law

18 19 20

85 F. at 284. Id. at 283. For a discussion and economic analysis of many of the cases, see Mark F. Grady, Toward A Positive Economic Theory of Antitrust, 30 Economic Inquiry 225 (April 1992). A few of

I. The Sherman Act Versus the Common Law

101

opinions would be worthwhile, the invalidity of Taft’s argument can be determined by examining Mitchel v. Reynolds. The theory of reasonableness articulated in Mitchel v. Reynolds is obviously more general than its caricature by Taft.21

C. The Case Against The Common Law Standard I noted earlier that Justice White wrote a spirited dissent in TransMissouri, and remained on the losing side in the early Sherman Act decisions. He was able to get enough votes to command a majority in Standard Oil Co. v. United States,22 but that case came too late to reverse the course the Court had set in Trans-Missouri. White had become Chief Justice in 1910 and a new Court majority allowed him to claim, in a remarkably hard-to-follow opinion, that the decision standard under the Sherman Act was the same as that employed by common law courts in restraint of trade cases. Thus, Section 1 of the Sherman Act condemned only undue or unreasonable restraints of trade. The reasonableness of a restraint depends on its purpose, the power of the parties, and the effect of their actions. White argued that the earlier Supreme Court decisions were consistent with this theory. It is highly doubtful that the earlier cases were consistent with White’s theory. Moreover, those earlier decisions provided the best source of

21

22

them are: Hearn v. Griffin, 2 Chitty 407 (K.B. 1815) (two coach masters on identical routes agreed not to set price higher than the other and set a common schedule that would assign each different days to travel on the common route); Herriman v. Menzies, 115 Cal. 16; 46 P. 730 (1896) (California Supreme Court upheld agreement between stevedores in San Francisco that arranged fixed prices); Ontario Salt Co. v. Merchants Salt Co., 18 Grant Ch. 540 (Ont. 1871) (court upheld agreement between two salt producers that organized “common selling arrangement”); Central Shade Roller Co. v. Cushman, 143 Mass. 353, 9 N.E. 629 (1887) (Three manufacturers of roller shades agreed to form a fourth corporation to sell their shades at a certain price. Court upheld agreement though partly on ground that firms possessed separate patents that could have been “mutually blocking”); Wickens v. Evans, 148 Eng. Rep. 1201 (1829) (Court upheld agreement among three trunk manufacturers who divided England into three exclusive marketing zones). The opinion in Mitchel v. Reynolds clearly reflects an awareness on the Court’s part of the social costs and benefits of a contract restraining trade. For example, the Court notes that firms often have incentives to enter these contracts in order to gain monopoly power, Mitchel v. Reynolds, at 189. The Court also notes that there are social benefits that come from agreements in which a firm sells its goodwill to another, agreeing to stay out of the market afterward, Mitchel v. Reynolds, at 190. The Court’s opinion suggests that courts should attempt to balance these considerations in deciding whether to enforce a trade restraining agreement. 221 U.S. 1 (1911).

102

Development of Section 1 Doctrine

guidelines for future decisions on price fixing – indeed, the facts of Standard Oil did not involve price fixing, they centered on the company’s attempt to monopolize the market. Thus, in spite of Justice White’s effort to import the common law standard into Sherman Act jurisprudence, the decision standard under the Sherman Act remained for practical purposes the standard announced in Trans-Missouri. Putting aside concerns over the reasonableness of the decision standard, what arguments might we make for or against the rule of reason? This chapter has already set forth the key argument in favor of the rule of reason: the necessity of an anchoring body of common law decisions to guide courts in applying the Sherman Act. Here it may help to consider the utilitarian arguments against adoption of the common law rule of reason as the standard of decision under the Sherman Act. The argument against the rule of reason is based on litigation experience. In both Addyston Pipe and Trans-Missouri, the government lost at the trial court level because it had failed to show the unreasonableness of the rates the associations fixed. As the Trans-Missouri opinion suggests, the Supreme Court believed the government’s argument that the reasonableness test would leave too much power in the hands of defendants. It would grant defendants an enormous advantage in antitrust prosecutions because only they would have access to much of the information necessary to prove or disprove reasonableness. Why did this not present an equally serious problem under the common law regime? Think about the typical case: the plaintiff attempting to enforce an anticompetition covenant against a defendant who had agreed to the provision. In such a case, neither party really has an informational advantage.23 Both plaintiff and defendant were parties to the contract, and presumably know all of the reasons for entering into the agreement. Certainly, both sides will possess private information bearing on the reasonableness of the agreement, but most likely neither will enjoy an informational advantage in litigation. The reasons for believing that plaintiff and defendant have some relative equality in contract litigation do not hold when the government prosecutes under the Sherman Act. The government, not having taken part in the agreement, and knowing little initially about the industry, is in a considerably worse position as a litigant than is the private defen23

On the importance of informational advantages in litigation, see Keith N. Hylton, Asymmetric Information and the Selection of Disputers for Litigation, 22 J. Legal Studies 187 (1993).

I. The Sherman Act Versus the Common Law

103

dant in a contract action. Replacing the rule of reason with a rule declaring all or certain trade restraints unlawful irrespective of reasonableness (per-se illegality) reverses the effects of the informational imbalance, and in this sense equalizes the positions of the parties. If the interest of the government in enforcing the Sherman Act were the overriding concern, then one can make a strong case for scrapping the rule of reason as a legal standard. However, the government’s interest is not necessarily the most important one. The interests at stake are many, and a rule of reason standard would at least enable or encourage courts to properly weigh them. As an illustration, consider the facts of the common law trade restraint case Mitchel v. Reynolds (Chapter 2). The case involved a noncompetition clause that permitted the seller to transfer the goodwill of his business to the buyer. The buyer brought suit against the seller after the seller breached the noncompetition agreement. Obviously, this case did not involve a government enforcement agency, but suppose it did. Imagine a similar case in which the parties maintained their agreement, and a government enforcement agent brought suit against them to have their agreement declared unenforceable because it restrained trade. A rule of per-se invalidity would minimize the administrative costs for courts and for an enforcement agent vested with the responsibility of prosecuting individuals who make such agreements. However, it would also severely limit if not eliminate the potential gains the parties might share as a result of a noncompetition agreement. Those gains result from the enhanced incentives for both parties to make brand capital investments, which also provide benefits to consumers. The rule of reason the court adopted in Mitchel resulted from an effort to balance the potential costs to consumers against the potential benefits to consumers and the contracting parties. If the market affected by the parties’ contract was one in which new rivals could enter easily, the potential costs to consumers from a noncompetition agreement between two existing firms would be minimized by the potential of new entry. It follows from this that a court could conclude under the economic reasonableness test that a noncompetition clause between two firms is on net beneficial to social welfare. Furthermore, note that such an analysis does not involve the Court taking into account its own administrative costs or the enforcement costs of the government. In the end, the Supreme Court’s decision to abandon the common law rule of reason made antitrust enforcement less difficult for the government. However, the Court had to justify the abandonment. The Court

104

Development of Section 1 Doctrine

eventually based its justification on Taft’s reasoning in Addyston Pipe, namely that the common law never really employed a rule of reason. Contemporary antitrust opinions continue to cite Addyston Pipe, perpetuating Taft’s misreading of the common law.

ii. rule of reason and per-se rule The next important step in the development of Section 1 came with the Supreme Court’s decision in Chicago Board of Trade v. United States.24 The Chicago Board of Trade required members to establish their offhour trading price for “to arrive” grain at a special session referred to as the “Call.” All trades that took place after the Call (which almost always ended by two p.m.) and before the opening of the Board on the next business day had to use the price set at the end of the Call. The government charged that the purpose and effect of the rule fixed the price for trading in “to arrive” grain. The Court rejected the government’s claim of per-se illegality and went on to give a detailed description of the factors that merit consideration in applying the rule of reason. The Court noted that every agreement restrains trade, so “the true test of legality is whether the restraint imposed is such as merely regulates and perhaps thereby promotes competition or whether it is such as may suppress or even destroy competition. To determine that question the Court must ordinarily consider the facts peculiar to the business to which the restraint is applied.”25 The Court then proceeded to analyze the purpose, power, and effect of the Call rule. First: The nature of the rule. . . . It required members who desired to buy grain to arrive to make up their minds before the close of the Call how much they were willing to pay during the interval before the next session of the Board. Second: The scope of the rule. . . . It is restricted to a small part of the total grain market, and a small part of the business day of grain traders. Third: The effects of the rule. The rule had no effect on market prices. But the rule improved market conditions.26

The Court held that the exchange’s call rule was “a reasonable regulation of business consistent with the provisions of the Anti-Trust Law.”27 The rule regulated competition instead of suppressing or preventing it.

24 25 26 27

246 U.S. 231 (1918). Id. at 238. Id. at 239–41. Id. at 239.

II. Rule of Reason and Per-Se Rule

105

The doctrine articulated in Chicago Board of Trade falls short of an uncompromising application of the common law rule of reason. Chicago Board of Trade deems an agreement reasonable if it regulates and thereby promotes competition. The focus remains on competition. A full blown common law rule of reason analysis would allow for an agreement to restrain competition if other benefits outweighed the harms caused by the competitive restrictions. If this had been the Court’s final statement on the law of Section 1, it would be fair to conclude that the Court had adopted a uniform standard close to the common law rule of reason. Of course, this would seem inconsistent with my earlier claim that the decision standard announced in Trans-Missouri prevailed, for the most part, even after Standard Oil. The answer to this puzzle is that a considerable gray area exists in the continuum between the common law rule of reason and the per-se rule. In the early years of the Sherman Act the Supreme Court more or less struggled between these two extreme positions, that is, either the Court would adopt the common law rule of reason approach or the per-se rule of Trans-Missouri. The contrasts between alternative positions became less harsh as time passed, and the Court rejected certain extreme views. The evolution of case law in this area illustrates Holmes’s theory of the process by which common law tends toward certainty.28 Holmes argued that cases concerning extreme propositions are first decided: for example, that crossing a railroad track without looking is negligence, and on the other extreme that crossing after looking both ways and listening is not negligence. Later cases deal with more difficult issues, for example, when the victim looked in the direction that the train ordinarily approaches, but was caught off guard because the train came from the other direction. The gray area between the extremes of clear negligence and clear nonnegligence are filled in by decisions specifying that Acoupled-with-B does not indicate negligence, but that A with C can support a negligence claim. Eventually the cases define a line separating descriptions of negligent and nonnegligent behavior, but the line is to some extent arbitrary.29 The Court rejected the extreme position of an uncompromising application of the common law rule of reason standard in Trans-Missouri. Chicago Board of Trade seemed momentarily to reintroduce the 28 29

Oliver Wendell Holmes, Jr., The Common Law 111–29 (1881). As Holmes noted, the line could be drawn a little further to the right or to the left without disturbing the basic message of the cases.

106

Development of Section 1 Doctrine

common law standard, though in a narrower form. However, the cases immediately following it would demonstrate that even the narrower rule of reason announced in Chicago Board of Trade did not furnish the general decision standard under Section 1. Whether the decisions from Trans-Missouri to Chicago Board of Trade pushed antitrust law toward certainty is a more difficult matter. Trans-Missouri was not a good start: it cut federal antitrust doctrine loose from a large body of common law decisions that the courts could have used to determine the legality of an anticompetitive arrangement. Subsequent decisions provided some guidance on the proper standard, and in this sense enhanced the predictability of antitrust law; but one must remember that this was starting from a base of maximum uncertainty. In any event, the notion that Chicago Board of Trade had resurrected the common law rule of reason as the general standard under Section 1 was disproved by the Court’s opinion in United States v. Trenton Potteries Co.30 The makers of 82 percent of toilets and other bathroom fixtures belonged to an association that had fixed the prices of toilets. The court of appeals had reversed a criminal conviction on the ground that the jury had not been allowed to consider the reasonableness of the prices fixed by the association. The Court held that the Sherman Act prohibits price fixing by those controlling in any substantial manner a trade or business in interstate commerce, despite the reasonableness of the particular prices agreed on. The Court defended this holding by suggesting that the administrative costs of enforcing the antitrust laws under a reasonableness standard would far exceed the benefits. [Price-fixing] [a]greements which create such potential power may well be held to be in themselves unreasonable or unlawful restraints, without the necessity of minute inquiry whether a particular price is reasonable or unreasonable as fixed and without placing on the government in enforcing the Sherman Law the burden of ascertaining from day to day whether it has become unreasonable through the mere variation of economic conditions.31

The holding applies to price fixing among competing firms. The Court made this clear in its effort to distinguish Chicago Board of Trade.32 The

30 31 32

273 U.S. 392 (1927). Id. at 397–8. Referring to Chicago Board of Trade, the Court said that “[t]hat decision, dealing as it did with a regulation of a board of trade, does not sanction a price agreement among competitors in an open market such as is presented here.” Trenton Potteries, 273 U.S. at 401.

II. Rule of Reason and Per-Se Rule

107

implication is that while rule of reason analysis may apply to other anticompetitive agreements, it is not the decision standard in the area of price fixing. Read broadly, the holding suggests that all price fixing among competing firms is per-se illegal. However, the decision can be read narrowly as applying only to those cases of price fixing involving firms that “control in any substantial manner” the business in a certain item or service. One can see in this line of cases how earlier decisions are dissected and then woven together into new doctrine. Trans-Missouri rejected the common law rule of reason as the general standard. Chicago Board of Trade suggested that the rule of reason test would be the standard, though in a somewhat narrower form. Trenton Potteries suggested that the per-se illegality doctrine of Trans-Missouri applies in the area of price agreements; the government will not have to prove the unreasonableness of a price-fixing agreement. The remaining uncertainty at this point of the development of Section 1 doctrine was the scope of the Chicago Board of Trade doctrine, that is, the extent to which it rather than the per-se rule applied to activities other than price fixing, and the boundary of the doctrine itself. More precisely, although the Chicago Board of Trade doctrine is a narrower form of the reasonableness test applied by common law courts, the case did not make clear just how much narrower. Indeed, the parameters of rule of reason analysis remain fuzzy today. Much of modern antitrust law is concerned with specifying these parameters. There was uncertainty over the rule announced in Trenton Potteries.33 The language in the opinion did not say clearly whether it held unlawful all price-fixing agreements, or only those agreements among parties controlling a substantial part of an industry, or those agreements among parties who control a certain share of business (82 percent in Trenton Potteries). The language that seemed to limit the holding to agreements entered into by parties controlling a substantial part of the market was used in Appalachian Coals v. United States34 to distinguish the facts from Trenton Potteries. In Appalachian Coals, 137 coal producers formed a new company, called Appalachian Coals, that would serve as the exclusive selling agent for member firms. The company worked like a sales cartel. The members 33

34

See John C. Peppin, Price-Fixing Agreements under the Sherman Anti-Trust Law, 28 California Law Review 297; 667 (1940) (two-part article). 288 US. 344 (1933).

108

Development of Section 1 Doctrine

of the cartel produced 74 percent of the bituminous coal mined in the Appalachian area, and 12 percent of that east of the Mississippi River. The lower court, following Trenton Potteries, ruled in favor of the government. The Supreme Court reversed, citing Chicago Board of Trade for the proposition that elimination of competition among the parties does not in itself warrant condemnation of the arrangement. The Court accepted the defendants’ reasonableness proof; that their collaboration formed for the purpose of distributing “distress” coal in a controlled fashion and preventing “pyramiding” sales.35 Distress coal is the surplus that results from the unwanted byproducts of making coal in particular sizes. Pyramiding occurs when a producer authorizes several agents to sell the same coal and thus leads buyers to suppose that the quantity available exceeds the true supply. In addition, the presence of organized and large buyers created monopsonistic conditions on the demand side, depressing market prices. The general picture, then, is that the defendants made this agreement to solve a certain coordination problem that reduced the producers’ revenue. In the absence of a centralized selling mechanism, each producer rushed to sell as much of the distress coal as possible, which put downward pressure on the price – a situation similar to that described in Figure 5.1 and its accompanying example. Large buyers, negotiating among several producers or sales agents, could get the sellers to compete in undercutting each other. This was a cartel, plain and simple. Although no evidence showed that the producers made monopoly profits as a result of the combination, its purpose was to hold prices above the level that otherwise would have been observed. The Court’s reason for accepting the defendant’s reasonableness proof is instructive. The Court cited the depressed conditions of the industry, with capacity exceeding seven hundred million tons to meet a demand of less than five hundred million tons. The high level of reserves suggested that the cartel could not have been making a supracompetitive profit, for otherwise coal producers would have had incentives to produce as much as possible. Second, the Court noted that there were other producers in the area, implying that if the cartel operated profitably it would have attracted entrants, presumably the other coal producers making up 88 percent of the market east of the Mississippi River. The Court concluded that the agreement could not have significantly affected the market price. The Court used these arguments to distinguish 35

Id. at 362–63.

II. Rule of Reason and Per-Se Rule

109

Trenton Potteries, on the ground that the parties in Appalachian Coals, unlike those in Trenton Potteries, had neither power nor intent to fix prices. There is perhaps no other way to describe Appalachian Coals other than as an application of rule of reason analysis very close to what one observes in the common law of trade restraints. The Court cites the rather narrow doctrinal formulation of Chicago Board of Trade, but then goes on to discuss in an approving manner activities of the defendants that clearly reduced competition. Thus, the Court’s reasoning implies that a reduction in competition is permissible if other benefits seem, on balance, to outweigh the harms. This balancing is inconsistent with the doctrinal formula of Chicago Board of Trade, which permits regulation of competition only if the effect enhances competition. The Appalachian Coals Court did not abandon the doctrinal formula of Chicago Board of Trade, but by approving in substance a broader economic reasonableness analysis, the opinion increased the uncertainty surrounding the rule of reason test. Later opinions have generally stayed within the parameters set by Chicago Board of Trade. For this reason, antitrust commentators tend to treat Appalachian Coals as an outlier. However, currently there are pressures to expand rule of reason analysis,36 and perhaps one day the Court will adopt the reasonableness test of the common law.37 After suggesting in Appalachian Coals that the per-se test of Trenton Potteries requires proof that the defendants’ price fixing had a substantial market effect, the Supreme Court abruptly rejected that proposition in United States v. Socony-Vacuum Oil Co.38 The problem that motivated the price agreement in Socony was in some respects similar to that in Appalachian Coals. The Court’s opinion tells us that independent oil refiners had no storage facilities and therefore had been dumping 36

37

38

Recent decisions such as Sylvania, BMI, and FTC v. Indiana Federation of Dentists reflect these pressures. For a decision that comes close to embracing the common law standard, see United States v. Brown University, 5 F.2d 658 (3d Cir. 1993). MIT and several Ivy League colleges had been agreeing on financial aid packages offered to admitted students. The Justice Department brought suit under Section 1. The Third Circuit reversed the district court’s finding that the plan violated Section 1 because it failed to give adequate weight to MIT’s “procompetitive and social welfare justifications,” 5 F.2d at 661. While the courts should obviously consider social welfare justifications under a full blown reasonableness inquiry, they are generally outside of the boundaries set by Chicago Board of Trade. 310 U.S. 150 (1940).

110

Development of Section 1 Doctrine

“distress” gasoline at discounted prices. But to understand the economics underlying Socony it is helpful to take a broader look at the facts. The important players in the market were the major, vertically integrated oil companies and the smaller independent refineries. State conservation laws restricted the amount of oil that could be pumped out of the ground and the petroleum code of the National Industrial Recovery Administration set price floors for oil and gasoline.39 The selling of distress gasoline at below-regulated (not necessarily below-market) prices is easy to understand in this setting. The supply restrictions of the conservation laws created an incentive for producers to pump oil in violation of the state conservation laws and to sell the additional supply, known as “hot oil,” at less than the regulated price.40 Sales of gasoline refined from hot oil, or “hot gasoline,” put downward pressure on the price for legal gasoline (i.e., gasoline refined from legal oil). In response, a group of major oil companies agreed to buy and store gasoline from independent refiners in order to prop up the spot market price of gasoline.41 Increases in spot market prices directly benefitted the majors because their distribution contracts were indexed to spot market publications.42 39 40

For a discussion, see Donald Dewey, supra note 11, at 162–3. The supply restrictions were the result of the state conservation laws. Putting the price regulation (NIRA) and the supply restrictions together, the market for oil could be described by Figure 5.2. Sales made in the area ABC would be labeled “hot oil” sales.

Area ABC = “distress” sales Q1 = supply restricted P1 = regulated price

Figure 5.2 41

42

This was carried out in conjunction with an output-limiting agreement among the major oil companies. For a thorough examination, see D. Bruce Johnsen, Property Rights to Cartel Rents: The Socony-Vacuum Story, 34 J. Law & Econ. 177–203 (April 1991). Id, at 178. Johnsen concludes that the majors’ buying program had no effect on gasoline prices. The reason is that a program of buying gasoline and storing it would increase inventories, putting downward pressure on forward prices for gasoline. Purchasers would shift from the spot market to the forward market to take advantage of the price differential. Johnsen argues that the majors’ overall effort was effective only because they restricted production in addition to purchasing gasoline from independent refineries.

II. Rule of Reason and Per-Se Rule

111

The Court addressed the defendant’s claim that Appalachian Coals implied that their agreement should be upheld. The Court noted that the presence of injurious practices provided the only common feature, in its view, between this case and Appalachian Coals. The key distinctions, to the Court, were that the plan in Appalachian Coals affected only a small part of the market and had never become operational. The Court distinguished Chicago Board of Trade on the grounds that the Board’s Call rule fixed not prices but the period of price making, affected only a small “proportion of the commerce in question,” and had the virtue of creating a public market for the grain contracts subject to the rule.43 Having distinguished Appalachian Coals and Chicago Board of Trade, the Court examined the defendant’s reasonableness arguments under the doctrine of Trenton Potteries. The Court first addressed the defendants’ argument that their plan did not eliminate competition. The Court held that it was sufficient that the plan would have the effect of stabilizing or increasing the price, and that the evidence indicated that it did. That the plan did not eliminate competition was of no consequence, for it was sufficient, in the Court’s view, that the plan curtailed competition.44 The second reasonableness argument put forth by the defendants was the claim that the plan eliminated competitive evils. The Court responded that this is an impermissible defense, because it would require the government to prove unreasonableness, thereby “emasculating” the Sherman Act.45 The defendants made a third claim that they did not have power to affect the market price. The Court held that this, too, is an impermissible defense under Trenton Potteries. Footnote 59 of the opinion makes clear that the market power defense is excluded on the basis of traditional conspiracy doctrine. The argument of footnote 59 is hard to contest. Criminal conspiracy doctrine does not require the state to prove that the conspirators had either begun to carry out or had the means to carry out their plan. It seems to follow, then, that if conspiracy to fix prices is a criminal offense, the ordinary doctrine applies, and thus power to actually set market price is not a necessary element of proof on the government’s part.

43 44 45

This suggests that the Court’s conclusions with respect to the majors’ buying program were wrong. However, my focus in the text is on the key legal propositions of the Socony opinion. 310 U.S. at 217. Id. at 220. Id. at 221.

112

Development of Section 1 Doctrine

Socony now stands for the proposition that price fixing is illegal per se. The proposition is broader than that established in Trenton Potteries, which suggested that price fixing is illegal only where the parties control a substantial share of the market. It is now settled that there is no market power requirement under Section 1.

iii. conclusion This chapter has presented the development of Section 1 doctrine, and argued that the doctrine’s evolution has been shaped largely by the tension between the economic conception of a reasonableness inquiry and the administrative concerns of enforcement agencies. Courts traditionally have adhered to reasonableness tests in developing new legal rules. Enforcement agents, however, are disadvantaged in antitrust by the application of a reasonableness inquiry because they do not, unlike litigants in restraint-of-trade contract disputes, have access to the information necessary to refute a defendant’s economic reasonableness arguments. The Supreme Court began in Trans-Missouri by siding with enforcement agents and rejecting the common law reasonableness standard, adopting instead a general per-se illegality rule. The tensions created by this standard led the Court to accept Judge Taft’s novel, though incorrect, interpretation of the common law in his Addyston Pipe opinion. Eventually the Court moved (or retreated) to the narrower reasonableness test articulated in Chicago Board of Trade, with the per-se standard reserved for price fixing.

6 Rule of Reason and Per-Se Rule

This chapter examines the boundaries of the rule of reason. On one end is the so-called per-se rule, and in theory there is another endpoint where all practices are per se lawful. Before examining the doctrine, I will take a detour to discuss some economic reasonableness arguments in favor of price fixing.

i. the case for price fixing Is price fixing always socially undesirable? The simple answer is no, as the following example illustrates. Suppose you develop a process innovation. It lowers the cost of producing widgets from $3 per unit to $2 per unit. Using the new process, you can produce widgets and charge $2.80, undercutting the competition and still make a profit of $0.80 per widget. Now suppose another producer can make widgets at $1 per unit if he could license the innovation from you. Of course you should license it. But suppose he undercuts your price, and puts you out of business. That may lead to the efficient result, but probably not what you intended when you licensed the competitor to use the innovation. One could argue that there is a potential arrangement that would leave both you and the licensee better off (i.e., after your exit), specifically, an agreement in which the licensee pays $0.90 per widget and pockets the extra $0.90. You earn more than if you had remained in the business alone, and the licensee earns a profit. This illustrates the basic proposition that efficient arrangements increase the wealth of all parties involved. 113

114

Rule of Reason and Per-Se Rule

Still, if you own the license, you may think you can do better, for several reasons. First, you might have information or experience that is difficult to transfer to the licensee, but would nevertheless prove helpful in any effort to improve the technology. What incentive do you have to expend effort toward improving the technology if you no longer use it yourself? Second, what if the licensee wants to modify the agreement in order to reduce the license fee? Outside of termination of the agreement, you have little control over the licensee. You could terminate the agreement and reenter the market on your own, but if you have been out of the business for a long time that could be a risky and expensive option. The relationship between you and the licensee turns, ex post, into one of bilateral monopoly. Note that this is an example of the transformation, stressed in the transaction cost literature (see Chapters 1 and 15), from ex ante competition to ex post monopoly. The best arrangement for maintaining incentives to improve the technology and to avoid the ex-post-opportunism problem may be one in which the original developer continues producing widgets for some time after the licensing agreement. But how can the original producer stay in the game and license the product? Answer: price fixing. If the inventor enters into a contract preventing the licensee from undercutting his price, then he avoids the prospect of having the market swept from under him. Consumers would no doubt do better in the short run if the licensee could price in an unrestrained manner and undercut the inventor. But in the long run, consumers may benefit more if price fixing is permitted, because the option enhances incentives to innovate and to license innovations. And of course, the licensee is better off, in the short and long run, than under the alternative in which the inventor refuses to license. This is a simple example and should suffice to establish the point that price fixing is not always, and under all circumstances, a socially harmful arrangement. Are the instances of socially beneficial price fixing sufficiently numerous to justify a rule of reason approach? No one has answered this question. However, a thorough reading of the common law restraint of trade cases, as Mark Grady has argued,1 would tend to support the view that instances of socially beneficial price fixing happen sufficiently frequently to justify a rule of reason standard.

1

Mark F. Grady, Toward a Positive Economic Theory of Antitrust, 30 Economic Inquiry 225 (1992).

I. The Case for Price Fixing

115

Core theorists have developed one theory that explains and justifies price fixing agreements.2 I will not attempt to summarize the literature in this area. However, the theoretical argument revolves around the notion of an “empty core.” The core of a game (think of a bargaining session) consists of the set of stable equilibrium allocations. Stability means “coalition-proof,” in the sense that no subgroup of bargaining parties has an incentive to form a coalition for the purpose of altering the equilibrium allocation. The core of a game is empty if there is no set of stable equilibrium allocations. How does this notion carry over into price fixing? Suppose firms are capacity constrained, meaning that no single firm can supply the whole market, the total capacity of the firms exceeds market demand, and fixed costs are large (see Figure 5.1). Then choose any allocation of sales among firms, and at least one firm has an incentive to cut price in order to redistribute business in its direction (see Figure 5.1 and the accompanying example). Competition, in such a setting, is likely to be unstable. The firms know that they are better off by accepting a price-output allocation rule that allows them to provide service while earning a competitive rate of return. However, the incentive to deviate from such an agreement is strong; indeed, it is the same as the incentive, discussed in Chapter 4, to deviate from a price-fixing agreement. The short run incentive to cut price toward marginal cost results in bouts of competitive price cutting, making it difficult in such an industry for firms to earn enough to invest properly in fixed assets. Core theorists have argued that price-fixing agreements are likely to show up in industries that exhibit the characteristics giving rise to the empty core problem. While it is not clear that such agreements provide socially desirable outcomes, neither is it clear that they do not. The case for prohibiting price fixing in these industries is not an easy one. It remains to see how well core theory stands up to theoretical and empirical examination.3 In addition, it takes a long time for a theory such 2

3

See, for example, Lester G. Telser, A Theory of Efficient Cooperation and Competition (New York: Cambridge Univ. Press, 1987), id., Economic Theory and the Core (1978); William W. Sharkey, The Theory of Natural Monopoly (1982). For recent theoretical and empirical work, see George Bittlingmayer, Decreasing Average Cost and Competition: A New Look at the Addyston Pipe Case, 25 J. Law & Econ. 201–29 (October 1982); Stephen Craig Pirrong, An Application of Core Theory to the Analysis of Ocean Shipping Markets, 35 J. Law & Econ. 89 (1992). Although not part of the core theory literature, Michael F. Sproul, Antitrust and Prices, 101 J. Pol. Econ. 741 (1993), is also of interest here. Sproul’s article demonstrates that antitrust prosecutions for price fixing have generally been followed by an increase in prices of targeted firms. Sproul notes

116

Rule of Reason and Per-Se Rule

as this to have any effect on the thinking of legislators. Perhaps thirty years from now the core theory literature will lead legislators or judges to reexamine the desirability of a per-se rule against price fixing.

ii. per-se and rule of reason analysis: further developments The theoretical justification for the per-se rule is utilitarian. The rule of reason involves a potentially costly inquiry, and courts are likely to reach the wrong answer every now and then because of the extraordinary expense involved in guaranteeing correct decisions. It makes sense, then, that in reviewing potentially anticompetitive practices, courts develop bright line rules that both enhance deterrence, by sending a clear message to potential defendants, and save administrative costs, by dispensing with many of the lines of inquiry that a full blown rule of reason analysis would require. Price fixing, the quintessential antitrust violation, is highly likely to be anticompetitive, and therefore is per-se illegal. The rule of reason as applied today differs from its counterpart under the common law. The common law rule of reason involved a balancing of broadly defined social costs and benefits. The modern rule of reason in antitrust is a narrower test, in the sense that it examines a narrower set of issues. Chicago Board of Trade tells us that a restraint of trade is reasonable only if it can be defended as being procompetitive. If the restraint is clearly anticompetitive, then even if it provides some social benefits unrelated to competition, it is still illegal. There is no clear theoretical justification provided in the case law for the Chicago Board of Trade doctrine. It reflects the Court’s view of the purpose of the Sherman Act at the time of the decision. The traditional view was that the Sherman Act’s purpose is to enhance competition. Thus, any restraint that satisfies the rule of reason inquiry must be one that enhances competition. If one pushes deeper for a theoretical justification, at bottom it would have to be similar to the utilitarian argument of Trenton Potteries. If the Court permitted almost any kind of costbenefit argument to justify a restraint under the rule of reason, then the inquiry could potentially be very expensive and deterrence weakened. The cases below push at the boundaries of the per-se and rule of reason tests. They force us to reexamine the Trenton Potteries justificathat one potential explanation (among several) for his result is that most of the targeted cartels were colluding in a manner that reduced costs.

II. Per-Se and Rule of Reason Analysis

117

tion for the per-se test and bring more clearly into question the early view of the purpose of the Sherman Act.

A. The Per-Se Rule and its Utilitarian Justification Modern price-fixing case law has been concerned with the scope of the per-se rule, and that has forced the Court to consider two issues: (1) the definition of “price fixing” and (2) the case for creating exceptions to the per-se rule. The statement “price fixing is per se illegal” is easy to say, but it immediately raises the problem of defining price fixing. The classic cases of a group of firms agreeing on the prices that will be offered to consumers or rigging bids for government contracts are uncontroversial. But what about maximum price fixing, dissemination of information on prices, and territorial allocation schemes? Recall that the Court referred to Chicago Board of Trade as a case in which there was an agreement covering the period of price fixing, not the prices. If the courts were to draw such detailed distinctions consistently, the per-se rule would be limited to a narrow set of cases. In addition, potential price fixers, aware of the distinctions made by the Court, would choose a form of collusion that avoids application of the per-se rule. Perhaps aware of these problems, the Court demonstrated a tendency early on to interpret the term price-fixing in a liberal fashion. In KieferStewart Co. v. Joseph E. Seagram & Sons,4 the Court held that a maximum price-fixing scheme is a per-se violation of Section 1, “[f]or such agreements, no less than those to fix minimum prices, cripple the freedom of traders and thereby restrain their ability to sell in accordance with their own judgment.”5 The decision is questionable on economic grounds because a straightforward economic defense exists for the scheme in Kiefer-Stewart, which involved the fixing of maximum resale prices. When, because of strong local demand, a retailer can set a monopoly markup on an item, a manufacturer has an incentive to set a ceiling on the resale price in order to maximize sales. In this case, both the consumer and the manufacturer have an interest in keeping the retailer’s markup low. Recognizing the force of this argument, the Supreme Court recently overruled Kiefer-Stewart in State Oil Company v. Khan,6 which 4 5 6

340 U.S. 211 (1951). Id. at 213. 118 S. Ct. 275 (1997).

118

Rule of Reason and Per-Se Rule

holds that rule of reason analysis applies to vertical maximum price fixing. The early liberal interpretation of price fixing is also reflected in United States v. Sealy7 and United States v. Topco Associates,8 two cases that involved territorial allocation schemes. Territorial allocation may facilitate a price-fixing agreement, but it is not the same thing as price fixing. However, the Court called the territorial allocation scheme in Sealy “an aggregation of trade restraints including unlawful price-fixing and policing.”9 In other words, territorial allocation is worse than price fixing. United States v. Topco Associates,10 repeated the holding of Sealy and suggested that the Court would take an expansive view of the term price fixing. The case involved a cooperative association (Topco) of about twenty-five small- to medium-sized regional supermarket chains. Member chains had market shares in their respective areas ranging from 1.5 to 16 percent, with an average of about 6 percent. The chains formed the association (or joint venture) so that the stores could “cooperate to obtain high quality merchandise under private labels in order to compete more effectively against larger national and regional [supermarket] chains.”11 Topco members were too small to market private label goods on their own. The agreement limited each member to selling Topco products in a designated geographic market. Pause for a moment to consider why the association would operate under this scheme. Firms typically advertise new products. Imagine the incentives of a supermarket chain within a territory in which it sells the same private label brand as another chain. The chains would draw from the same pool of consumers. One chain could cut its costs and increase its revenue by spending nothing on advertising and promotion, and selling the private label brand, or other brands, at a lower price. Each chain would have an incentive to free ride on the promotional expenditures of another. The Court treated this as a horizontal territorial restraint and ruled that such market allocations are per-se illegal. The rationale was that only Congress should make a decision to “sacrifice competition in one 7 8 9 10 11

388 U.S. 350 (1967). 405 U.S. 596 (1972). 388 U.S. at 357. 405 U.S. 596 (1972). Id. at 599.

II. Per-Se and Rule of Reason Analysis

119

portion of the economy for greater competition in another portion,” not the courts or private parties.12 Taken to its logical extreme, this doctrine has troubling implications. Why not say that any new combination, not in existence at the time of the Sherman Act’s passage, that restricts competition in one area in order to enhance it in another is illegal unless Congress authorizes it? Under this view, all joint ventures not in existence in 1890 would require the approval of Congress. And what about new forms of business organization? Consider the copyright royalty agencies, American Society of Composers, Authors, and Publishers (ASCAP) and Broadcast Music, Inc. (BMI). They would almost surely fall under the general prohibition announced in Topco. They restrict competition among musical composers to the extent that the composers cannot distribute their work through the agencies and at the same time charge their own prices for access to the portion of their work the agencies distribute. On the other hand, the copyright royalty agencies enhance competition by making a new product – a blanket license for copyrighted music – available to radio stations, television stations, and other consumers. Considered in light of the facts in Topco, the Court’s reasoning remains questionable. As the dissent noted, before the chains formed the association, there was no Topco product line. The association introduced a new product, which enhanced competition. One might try to cite Chicago Board of Trade to support the claim that the Sherman Act would permit this: an agreement regulating competition for the purpose of enhancing it. However, the problem with relying on Chicago Board of Trade, as a justification for the agreement in Topco, is that the Court has distinguished Chicago Board of Trade as a case not involving price fixing. The Court’s conclusion that this was a territorial allocation scheme similar in effect to price fixing made the Chicago Board of Trade doctrine inapposite. Justice Burger’s dissent reflected the deeper concern that the Court seemed to be extending the per-se rule beyond the boundaries established in earlier cases, notably Addyston Pipe. He argued that because the association’s agreement was ancillary “to the creation of new private label product,” it should not fall under the per-se rule.13 12 13

Id. at 611. Id at 613–14.

120

Rule of Reason and Per-Se Rule

Some commentators have suggested that the Supreme Court effectively overruled Topco in its Sylvania decision, even though the Sylvania opinion includes a footnote expressing the view that Topco is still sound precedent. The Sylvania Court held that vertical nonprice restrictions of intrabrand competition made in order to promote interbrand competition are not per se unlawful. I find it hard to believe that Topco remains valid after Sylvania, but we will consider this question in more detail later. The remaining decisions allow us to see how valid the Topco doctrine remains, as a general proposition concerning horizontally restrictive agreements ancillary to productive activities. Although disturbing in its potential scope, the Court did not clearly scale back the Topco doctrine until its decision in Broadcast Music Inc. v. Columbia Broadcasting System (BMI).14 In BMI, the Court was confronted with a price-setting cooperative association that had obvious efficiency justifications. In order to collect copyright royalties, owners must sell rights and ensure enforcement of promises to pay royalties, otherwise the right to collect royalties would have little use. But put yourself in the shoes of a music composer. Would you try to do this yourself? Obviously, you would not. ASCAP and BMI act as license clearing houses for music composers. Those dealing with either of the two receive a blanket license authorizing the user to perform the entire collection without regard to the number of times the user plays the works. In return, the users pay a royalty measured either by a specified fee or by a percentage of the user’s advertising revenues. CBS sought a license on a per use basis from both ASCAP and BMI. When they refused, CBS sued under the Sherman Act. The issue before the Court was whether the blanket license arrangement fell within the per-se category. The Court held that a blanket license arrangement is not a per-se violation of the Sherman Act. The Court analyzed the facts under the rule of reason and concluded that there was no violation because the marketing arrangements the defendants adopted appeared “reasonably necessary” for the development and marketing of rights granted to composers under copyright laws. As usual, the Court offered additional justifications. One was that the defendants already operated under restrictions imposed by consent decrees from earlier 14

441 U.S. 1 (1979).

II. Per-Se and Rule of Reason Analysis

121

suits brought by the government, and this suggested that the anticompetitive potential of their marketing arrangements was slight. The Court also noted that Congress had employed a blanket license approach in parts of a copyright statute covering cable transmissions, and thus had given its stamp of approval, though unwittingly, to the efficiency claims of the defendants. Finally, the Court pointed out that it would be hard to accept CBS’s argument, and at the same time modify the ASCAP practice to some form that seemed less restrictive. Even a per use, category-specific contract would involve fixing prices within categories of ASCAP members and for particular uses under the contract. The logical solution would seem to require ASCAP, in order to avoid antitrust liability, to negotiate with each user over each use of the work of every individual composer. But the Court had already accepted the premise that this solution would prove infeasible, or prohibitively costly, as a general approach to marketing copyrighted music. Still, one might argue that the Court should have required some less restrictive, yet feasible, alternative, such as the per-use license CBS proposed. Here I think the only answer is that as a general matter, the rule of reason does not require the least restrictive alternative. If, in the Court’s view, moving to the less restrictive alternative provides only small benefits, but significant costs, the Court will not interpret the rule of reason as requiring adoption of the alternative. It may be helpful to explain some of the underlying economic issues. First, recall that the Court had conceded that a general marketing plan requiring ASCAP to engage in individual level negotiation was infeasible, a premise which is important in the Court’s reasoning. In the district court, CBS had argued that it was feasible for ASCAP to serve as a gobetween in individual level negotiations. However, the district court found that CBS could make individual arrangements, bypassing ASCAP. So why would CBS bring this suit if it could bargain at an individual level, and its own argument implied this was true? The answer is that from CBS’s perspective, individual level negotiation may have been feasible, but no performer would negotiate with them if the ASCAP arrangements allowed performers to share in monopoly rents. So the fundamental economic issue in this case is whether ASCAP’s arrangement resulted in “monopoly rents” or whether the rents resulted entirely from the reduced enforcement and distribution costs due to the use of blanket licenses. Second, owners of copyrighted material can monitor its use without requiring a blanket license. But the monitoring problem becomes more difficult for per-use and category specific licenses. ASCAP

122

Rule of Reason and Per-Se Rule

would have to determine whether a user had taken more than what the contract permits, and given the low probabilities of detection, it is hard to see how ASCAP could provide a financial deterrent to cheating by users. And rampant cheating would effectively scale back ASCAP’s services, to the detriment of both users and composers. What does BMI imply for the general rule regarding price fixing? Recall that Socony held that all price fixing by firms competing in an open market is per-se unlawful. BMI introduces an exception for pricefixing plans that are necessary for the introduction of a new product. This rule applies to pure, horizontal price-fixing schemes, and is the only absolutely clear doctrinal exception to the per-se rule against price fixing.15 In addition, BMI has implications for the Topco dictum declaring impermissible price-fixing plans that reduce competition in one area in order to enhance it in another. This dictum should not be taken seriously as a rule of antitrust law since the plan in BMI clearly violated it. Once again, the rationale for a per-se rule is that it supposedly enhances deterrence and reduces administrative costs. The exception BMI introduced might in theory be used by a large number of antitrust defendants. Almost any cartel agreement on price or territory will also involve some cooperative elements that defendants could frame as an alteration of the goods the consortium produces, or an introduction of a new product. But if every cartel could rely on the BMI exception, there would be nothing left to the per-se rule. The courts would have to examine each procompetitive theory put forth in every Section 1 case. Because BMI is the only exception to the per-se rule against horizontal price fixing, drawing a boundary to the BMI exception is equivalent to defining the scope of the per-se rule itself, which is equivalent, in turn, to saying when the courts should or should not abandon the utilitarian justification of Trenton Potteries. The difficulty in providing a good answer to this question was illustrated in Arizona v. Maricopa County Medical Society.16 Seventy percent of the medical practitioners in Maricopa County belonged to the Maricopa Foundation. The Maricopa Foundation established a plan whereby member-practitioners agreed not to charge patients more than a specified fee for identified services. Several insurance companies agreed to pay the full cost of the services the participating doctors provided. The Pima Foundation created a 15

16

For reasons that will be made clear later, the other important exception, stated in Sylvania (Chapter 13), applies to price fixing only by implication. 457 U.S. 332 (1982).

II. Per-Se and Rule of Reason Analysis

123

similar fee-scheduling plan, and its membership represented about 15 percent of the doctors in the county. Each foundation revised its maximum fee schedule periodically, subject to a vote by the membership. The question presented was whether Section 1 is violated by an agreement among competing physicians setting, by majority vote, the maximum fees charged to insurance plan policyholders. The Court concluded that such a plan does violate Section 1, even though the prices fixed were maximum fees.17 The defendants offered several reasons why the Court should analyze their plan under the rule of reason. First, they argued that because the medical profession already faces higher standards, it should not also be subject to the per-se rule. The Court noted that in previous decisions it had suggested that because of “public service aspects,” an exception to the per-se rule may be appropriate when the courts apply the Sherman Act to the professions.18 However, the Court distinguished those earlier decisions on the ground that they involved price fixing justified by ethical concerns, while the agreement in Maricopa had nothing to do with ethics. This distinction should trouble the student of antitrust law. Ethical justifications of price fixing should raise immediate concern because they tend to be the sort of slippery, difficult to disprove arguments that work best as concealing garments. Furthermore, in one of the prominent cases dealing with ethics and price fixing, Goldfarb v. Virginia State Bar, the price fixing occurred through the promulgation of a list of minimum fees by the state bar association. Goldfarb involved minimum fees set by a state-authorized entity, supported by the threat of punishment by the state. The fees in Maricopa, on the other hand, were maximum fees, set by a private association of doctors, who set up the plan in order to attract insurers. The insurers were a countervailing force,19 pressuring the 17

18

19

The Court cited two earlier decisions finding maximum price fixing unlawful, KieferStewart Co. v. Joseph E. Seagram & Sons, 340 U.S. 211 (1951), and Albrecht v. Herald Co., 390 U.S. 145 (1968). However, both of these cases involved a seller setting maximum prices that could be charged by retailers, and both have been overruled by State Oil Company v. Khan, 118 S. Ct. 275 (1997). In spite of this, the status of Maricopa (after Khan) remains unclear because the doctors’ arrangement in that case is probably better viewed as a horizontal agreement. The holding in Khan is expressly limited to vertical maximum price-fixing agreements. See, for example, Goldfarb v. Virginia State Bar, 421 U.S. 773 (1975), also Professional Engineers (this chapter). Why would insurers participate in a plan to gouge consumers? I find it hard to see why they would do this. The charges of doctors are a cost to the insurer. The policy premium

124

Rule of Reason and Per-Se Rule

doctors to keep their fees low.20 There is a big difference between the plans in Goldfarb and in Maricopa, and the Supreme Court drew precisely the wrong inference. The plan in Maricopa was far less worrisome. The defendants’ second argument was that the Court should analyze their plan under the rule of reason because of the judiciary’s limited experience with the health care industry. The Court rejected this argument on the ground that it would extend a rule of reason approach to every dispute involving an industry with which the Court had no previous experience, destroying the per-se rule. The defendants’ third argument was that the Court should analyze the agreement under the rule of reason because of the agreement’s procompetitive justifications. The Court never confronted the procompetitive justifications directly; rather, it rejected the defendants’ argument on the ground that per-se analysis does not permit an inquiry into procompetitive justifications. Invoking the utilitarian formula of Trenton Potteries, the Court said that the benefits of ferreting out the procompetitive justifications of such plans would be outweighed by the costs. These rejections are hard to square with BMI. The Court treated BMI as a case of price fixing, and yet listened carefully to the procompetitive justifications and applied rule of reason analysis. This is precisely what the defendants were asking for in Maricopa. If this were all to the opinion, the decision would have to be described as a failure to come to grips with the economics and facts of a dispute. However, the Court responded to the defendants’ efforts to rely on BMI in later passages. After repeating the Trenton Potteries formula, the Court noted that alternative methods of setting price would raise fewer anticompetitive concerns. For example, the insurers themselves could set maximum reimbursable rates, or a state agency could set the rates. One might assume these were less efficient methods of setting appropriate fees, since the options were available and the parties did not take them, but this “assumption” was “far from obvious” in the Court’s eyes.21 The

20

21

is on the revenue side. The insurer’s aim is to maximize the difference between the premium and the unit cost of the plan. That means the insurer should aim to minimize doctors’ fees, given a fixed premium. For an argument that countervailing forces should be considered carefully in antitrust analysis, see Barbara Ann White, Countervailing Power – Different Rules for Different Markets? Conduct and Context in Antitrust Law and Economics, 41 Duke Law Journal 1045 (1992). Alternatively, if the fees were high compared to the market, the insurers would have attempted to restrict the quantity of services provided under the plan. But the insurers had agreed to reimburse doctors for all of the services whose prices were specified under the plan, which is inconsistent with the theory that this was price fixing. 441 U.S. at 21.

II. Per-Se and Rule of Reason Analysis

125

Court held that on the basis of the evidence before it, the asserted efficiency gains from having the doctors, rather than insurers, set maximum fees could not justify the potentially anticompetitive effects. Later still the Court confronted the defendants’ BMI defense directly. It should be easy to see why the defendants would rely on BMI. They set their fees, and a large percentage of their patients came to them through insurers, and therefore the consumers never really examined the maximum fee schedule directly. They simply paid a flat fee for unlimited use of the good, like the customers in BMI. Of course, there are differences between this case and BMI. First, unlike the music composers in BMI, the doctors were already selling their services to customers and enforcing claims to payment. In other words, the significant transactioncosts barrier that prevented musicians from enforcing their claims to copyright fees did not stand in the way of doctors getting their pay. Also, the fees that the doctors set through the Foundation had the danger of becoming an agreement that operated throughout the entire market rather than on the subset comprised of insurance policyholders. The Court distinguished BMI on the ground that the Maricopa agreement did not allow the doctors to market a different or new product, since they already sold their services on the market. However, one could argue that the agreement did permit the introduction of a new product. It was a plan in which doctors made credible commitments to keep their expenses within limits, and to that extent allowed insurers to reduce their premiums, in the confidence that costs could be accurately predicted. BMI and Maricopa are not easily distinguishable if you focus on the new product criterion. However, the cases are easily distinguishable if you focus on the transaction costs that made the arrangement in BMI necessary, which clearly were not present in Maricopa.

B. Pressure on Rule of Reason Boundary I noted earlier that the cases of this chapter clarify the boundary between the rule of reason and the per-se rule. The cases discussed up to this point can be viewed as working within the domain of the per-se rule toward this boundary. The remaining cases of this chapter work within the rule of reason’s domain toward the same limit. National Society of Professional Engineers v. United States22 dealt with the canon of ethics of the Society, which prohibited competitive bidding. Members of the Society who complied with the code refused to 22

435 U.S. 679 (1978).

126

Rule of Reason and Per-Se Rule

negotiate or even to discuss the question of fees until after a prospective client had selected the engineer for a particular project. The question was whether the canon violated Section 1. The Court applied the rule of reason and held that the canon violated Section 1. The Court gave two reasons for its decision to apply the rule of reason. First, the case did not really involve price fixing, only a rule preventing price competition. Recall that the Court used this as one of its justifications for applying the reasonableness test in Chicago Board of Trade, where the Court noted that the “Call Rule” only limited the period of competitive price setting. Incidentally, this approach does not sit well with the Court’s tendency, in cases such as Topco, to treat horizontal market allocation schemes as per se unlawful. Market allocation schemes restrict competition, they do not require the parties to fix prices. The second justification for applying the rule of reason was that the defendants were members of a profession. While no doctrine requires application of the rule of reason to professions, in general, the professions have a greater claim to application of the rule of reason than do other industries. Indeed, this is the only exception the case law recognizes to the general, industry-wide application of the per-se rule.23 Although the Court decided to apply the rule of reason, it never got beyond a very general discussion of the defendants’ argument. The distinction between the treatment of the facts in this case under the rule of reason and application of the per-se rule is hard to see. The society claimed that the code was reasonable because competition among professional engineers violated the public interest; specifically, competitive pressure on engineering services would adversely affect the quality of engineering. The Court rejected this argument because “inquiry [under the reasonableness test] is confined to a consideration of impact on competitive conditions.”24 Recall that this is the doctrine of Chicago Board of Trade. Professional Engineers helps clarify it. The argument that defendants had to restrict price competition in order to enhance quality is impermissible under the

23

24

On the industry-wide application of the per-se rule, see United States v. Trenton Potteries, 273 U.S. 392, 397–8 (Agreements which create such potential power may well be held to be in themselves unreasonable or unlawful restraints, . . . without placing on the government in enforcing the Sherman Act the burden of ascertaining from day to day whether it has become unreasonable through the mere variation of economic conditions.) Recall that Maricopa firmly rejected the argument that the courts should deal with new or highly technical industries under the rule of reason. 435 U.S. at 690.

II. Per-Se and Rule of Reason Analysis

127

rule of reason, because it is not a claim that the challenged practice “regulates and thereby promotes competition.” Rather, it attempts to prove reasonableness by pointing to benefits that supposedly outweigh the harms associated with a reduction in competition. Professional Engineers reveals the narrowness of the Court’s interpretation of the Chicago Board of Trade doctrine. One can think of two types of tradeoff claims: those related the product and those unrelated to the product. An example of the former is the Professional Engineers case, in which the defendants claimed that a reduction in competition enhanced the quality of the product. An example of the latter would be competing firms who claim that they must fix prices in order to clean the environment or to enhance employment opportunities for minorities. That the Chicago Board of Trade doctrine would rule out the latter claim strikes me as uncontroversial. However, the claim of the defendants in Professional Engineers is not clearly unrelated to competitive conditions, because quality is an aspect of the product market. An attempt to fix prices in order to enhance quality is a claim that at the observed market price, the challenged practice enhanced quality. That is equivalent, in economic terms, to a claim that at the observed market quality (i.e., under the restraints imposed by the society), the challenged practice lowered the price. In other words, the defendants’ argument in Professional Engineers did address competitive conditions. Only an extremely narrow view of the term “competitive conditions” would lead one to think otherwise. The Court has had to struggle with this problem in more recent cases, and it is one that will not go away.25 The Chicago Board of Trade doctrine, as interpreted in Professional Engineers, takes an excessively narrow view of the realm of competitive conditions. In justifying this narrow view, the Court in Professional Engineers briefly reviewed the history of antitrust law. The Court cited Mitchel v. Reynolds26 as an early common law decision in which the court upheld a trade restraint because the benefits from the long run enhancement of competition exceeded the short run costs of the restriction. The 25

26

An attempt was made to modify the Professional Engineers doctrine in United States v. Brown Univ., 5 F.2d 658 (3d Cir. 1993). The Third Circuit, in a clear attempt to limit the scope of the Professional Engineers doctrine, identified certain quality-enhancement arguments made by the defendant (Massachusetts Institute of Technology) that it felt should have received greater weight by the district court in applying the rule of reason test. 1 P.Wms. 181, 24 Eng. Rep. 347 (K.B. 1711) (discussed in Chapter 2).

128

Rule of Reason and Per-Se Rule

discussion was designed to show that the Chicago Board of Trade rule was consistent with the common law of trade restraints.27 However, the cite to Mitchel v. Reynolds reveals a key weakness in the Court’s argument. Mitchel v. Reynolds was a case in which unfettered competition reduced quality (or more precisely, reduced incentives to invest in brand capital). The defendants in Professional Engineers could have justifiably cited Mitchel v. Reynolds in their favor. National Collegiate Athletic Association (NCAA) v. University of Oklahoma28 involved an agreement among NCAA colleges restricting the number of times each team’s football games could be televised. Under the plan, no member institution could appear on television more than a total of six times and more than four times nationally, with appearances divided equally between CBS and ABC. In 1979, CFA (College Football Association) began to advocate that its members have a greater voice in football television policy. CFA obtained a contract from NBC that would have allowed each member institution to appear on television more often. The NCAA declared that it would take disciplinary action. CFA brought suit in response. The issue was whether the NCAA plan violated Section 1 of the Sherman Act. The Court applied the rule of reason and held that it did. The Court cited both BMI and Sylvania as reasons for adopting a rule of reason test, noting that in the area of college football “horizontal restraints on competition are essential if the product is to be available at all.”29 The football league needs rules, even if they restrict output. Restraints that enhance competition on the playing field may have the effect of reducing the output of sellers. For example, the schools must agree on the length of the season, and this clearly has an output limiting effect. Second, Sylvania suggests rule of reason analysis should apply because the rules make a differentiated product which can compete more effectively with others. After citing BMI and Sylvania as reasons for applying the rule of reason test, the Court shifted direction and distinguished those cases. BMI was distinguished on the ground that while individual level negotiation was infeasible in the copyright market, it was quite feasible in this case. Consumers would still have access to the product even without the

27

28 29

Recall that this is the innovative, though doubtful, interpretation of the early case law provided by Taft in Addyston Pipe; see Chapter 5. 468 U.S. 85 (1984). Id. at 86.

III. Per-Se Versus Rule of Reason Tests

129

NCAA television plan. And the plan was not a crucial feature in the basic package of product-defining restraints, restrictions that had little to do with the method of marketing television rights. The BMI defense fails here, the Court argued, because the restraints had little to do with the definition, form, and existence of the product. The Court distinguished Sylvania on the ground that no good substitutes for the unique product of college football existed. Sylvania requires substitutes: Without reasonably good substitutes, the Sylvania doctrine cannot be applied. The Court also relied on the district court’s findings that the restraints had actually reduced competition by restricting output and raising price. The rest of the opinion applies the rules of Chicago Board of Trade and Professional Engineers.30 The defendants argued that the plaintiffs could not prove that they had market power. The Court responded that “[a]bsence of proof of market power does not justify a naked restriction on price or output, . . . some competitive justification”31 must be presented. The defendants also argued that the policy of restricting television rights was designed to protect attendance at games. The Court found this inconsistent with the evidence that television games were shown at all times, with no apparent effort to schedule them in a way that protected gate attendance. In addition, the Court rejected the argument because it was inconsistent with the rules of Chicago Board of Trade and Professional Engineers. One view of the defendants’ argument is that it was protectionist, that is, an effort to shield a weak product from competition, which violates Chicago Board of Trade. The other view is that game attendance serves other important social functions. The students get together, get drunk, and cherish the memories forever. But these “social benefit” arguments fall outside of the boundaries set by Professional Engineers.

iii. per-se versus rule of reason tests: understanding the supreme court’s justification for the per-se rule To this point, we have examined the Supreme Court’s justification for the per-se standard, and the general doctrinal boundaries separating 30

31

Indeed, the NCAA opinion is sometimes described by antitrust scholars as a “quick look” application of the rule of reason. The reason is that the opinion does not present a careful analysis of power, purpose, and effects issues. The Court largely applies rules developed in BMI, Sylvania, Chicago Board of Trade, and Professional Engineers to reach the conclusion that the NCAA’s plan violated Section 1. Id. at 109.

130

Rule of Reason and Per-Se Rule

per-se and rule of reason analyses. This is a good place to return to the Court’s justification in order to gain a deeper understanding of it. Because courts will sometimes make mistakes in deciding cases, standards of proof should be set with a view toward controlling or minimizing the costs of judicial error. The important types of mistakes are easy to state. An erroneous finding of nonliability (or, for simplicity “false acquittal”) happens when the Court holds that the defendant is not liable (innocent) when he in fact did violate the legal standard. Some commentators have referred to this type of mistake as Type-1 error,32 drawing on the terminology of statistics. An erroneous finding of liability (or “false conviction” or Type-2 error) happens when the Court finds the defendant liable (guilty) when the defendant in fact did not violate the legal standard. There are costs associated with these judicial errors. False acquittals have the effect of weakening or lowering the legal standard. As the frequency of false acquittals increases, potential defendants will have less incentive to comply with the legal standard, since the likelihood of having to pay damages for failing to comply falls. False convictions, on the other hand, have the effect of raising the legal standard. As the frequency of false convictions increases, even those potential defendants who are complying with the standard will be compelled to change their conduct in order to reduce the risk of being held liable. In other words, false acquittals lead to “underdeterrence” and its associated costs, while false convictions lead to “overdeterrence” and its associated costs. Choosing between per-se and rule of reason analyses is in part a choice between a greater risk of false acquittals and a greater risk of false convictions. It should be clear that potential false acquittal costs are greater under the rule of reason, and potential false conviction costs are greater under the per-se rule. Another important factor in choosing between rule of reason and per-se tests is the administrative cost of implementing the legal standard. Courts generally say that the rule of reason test is more costly, both for courts to administer and for enforcement agencies to meet in order to prosecute a case. It is easy to restate the traditional justification for the per-se rule in terms of these costs. Let T1C be the costs associated with false acquit-

32

See, for example, A. Mitchell Polinsky and Steven Shavell, Legal Error, Litigation, and the Incentive to Obey the Law, J. Law, Econ. and Org., vol. 5, pp. 99–108 (1989); Keith N. Hylton, Costly Litigation and Legal Error Under Negligence, J. Law, Econ., and Organ., vol. 6, pp. 433–52 (1990).

III. Per-Se Versus Rule of Reason Tests

131

tals, let T2C be the costs associated with false convictions, and let AC be the incremental administrative costs associated with the rule of reason test. The per-se rule is preferable to the rule of reason if AC > T2C T1C, that is, the incremental administrative cost associated with the rule of reason is larger than the difference between the error costs. It follows from this reasoning that if the potential false acquittal and false conviction costs connected with an activity are roughly the same, we should prefer the per-se rule. On the other hand, if false conviction costs are especially high, we should prefer rule of reason analysis. With this framework, we see that the Court’s justification for the perse rule is based on an assumption that potential false acquittal and false conviction costs are either roughly symmetrical, or the false acquittal costs are larger. The Court’s utilitarian justification for applying the perse rule to price fixing can be read as saying, that the false conviction costs are small, presumably because there are relatively few instances of socially beneficial price fixing, and the false acquittal costs are large, presumably because socially harmful price fixing would occur more frequently under the rule of reason.

7 Agreement

The cases in this chapter indicate the circumstances under which a court will infer an agreement. Because of the procedural posture of the cases, the legal holdings do not offer much insight, for in the typical case the Supreme Court upholds a lower court decision that in turn relies on facts and inferences that are difficult to assess on the basis of a cold record. Not coincidentally, dicta, specifically general comments about the evidentiary requirements of Section 1, make up the important parts of these opinions. The decisions seem to reveal a trend: From roughly 1940 to the early 1950s, the opinions suggest that not much more than purely circumstantial evidence would suffice to prove conspiracy in violation of Section 1. This coincided with a period some antitrust commentators referred to as the New Sherman Act. With the New Sherman Act, according to the commentators, courts would set aside burdensome proof requirements under Section 1 and replace them with a doctrine that brought antitrust law into agreement with modern industrial organization research.1 That research indicated that cartel pricing could occur in the absence of explicit agreement. Starting in 1954, the year of the Court’s decision in Theatre Enterprises v. Paramount Film Distributing Corp.,2 courts began issuing

1

2

The expression “New Sherman Act” was coined by Eugene V. Rostow, who set forth the thesis in several influential articles, see Eugene V. Rostow, The New Sherman Act: A Positive Instrument of Progress, 14 U. Chic. L. Rev. 567 (1947); id., Monopoly Under the Sherman Act: Power or Purpose?, 43 Ill. L. Rev. 745 (1949). 346 U.S. 537 (1954).

132

I. The Development of Inference Doctrine

133

decisions that disappointed New Sherman Act proponents. The postTheatre Enterprises opinions showed little tolerance for plaintiffs who brought claims of conspiracy supported by purely circumstantial evidence.

i. the development of inference doctrine The first Supreme Court case in this area is Eastern States Retail Lumber Dealers’ Assn. v. United States,3 which antitrust casebooks often cover in the section on boycotts. The Court upheld the trial court’s finding that retail lumber dealers had agreed not to buy from wholesalers who sold directly to consumers. This agreement violated Section 1. The conspiracy was inferred from the dealers’ circulation of a list of offending wholesalers, and the fact that many dealers stopped doing business with offending wholesalers. As Eastern States illustrates, we are dealing with a problem of inference. Agreement to fix prices is per-se unlawful. However, the plaintiff has the burden of proving the existence of such an agreement. The result in Eastern States implies that a finding of conspiracy may be based on circumstantial evidence. But the hard questions begin at this point. What are the minimal evidentiary standards? What behavioral assumptions are permissible? How far can courts carry the process of inference in order to find a violation? The case law attempts to answer these questions, and Eastern States provides some preliminary answers to each of them. As to minimal evidentiary requirements, note that the Eastern States decision relied on the fact that the defendants knew of the list of offending wholesalers, which implies the necessity of some common information. As to behavioral assumptions, Eastern States suggests that the assumption that the defendants will use the shared information in a rational, self-interested manner is permissible, for otherwise inference would be impossible. Finally, Eastern States held that the evidence supported a finding that the parties had agreed to restrain trade. If the Court had found that the evidence indicated merely an agreement to share information, leaving the participants free to decide what to do with it, the Court would not have upheld the Section 1 violation.4

3 4

234 U.S. 600 (1914). See Cement Mfrs Protective Ass’n v. United States, 268 U.S. 588 (1926).

134

Agreement

A. Interstate Circuit Modern inference doctrine more or less begins with Interstate Circuit v. United States.5 Eight motion picture film distributors with 75 percent of the feature film market had made identical modifications in their contracts with two affiliated Texas movie theater owners that exhibited firstrun films, Interstate Circuit and Texas Consolidated. Interstate had a monopoly of first-run theaters in six Texas cities, and Texas Consolidated dominated the movie market in the cities in which it operated. The manager of Interstate and Consolidated sent to each of the eight distributors a letter in which he asked compliance with two demands as a condition of Interstate’s continued exhibition of the distributors’ films in its first-run theaters at an admission price of 40¢ or more. The first condition prohibited subsequent runs of first-run films at prices less than 25¢, the second eliminated double features on first-run films. The manager addressed the letter to all of the distributors, so each knew that the others received the same offer. The district court held that there was an agreement in violation of Section 1. The issue before the Supreme Court was whether the evidence was sufficient. The Court held that the government had met the proof requirement, adding that “[a]cceptance by competitors, without previous agreement, of an invitation to participate in a plan, the necessary consequence of which, if carried out, is restraint of commerce, is sufficient to establish an unlawful conspiracy under the Sherman Act.”6 The Court gave several reasons, but three are important. First, each knew of the offer Interstate made to its competitors. Second, no evidence in the record suggested that the behavior of the film distributors was individually rational, that is, independent. The proposal involved a significant change from previous practice, and there was some evidence of disagreement from the distributors’ local managers. Third, the complexity of the proposed changes made it difficult to believe that they were coincidental. We can summarize the argument in three phrases: shared information, independence, complexity. These words provide some of the detail that the term agreement fails to convey. In the absence of hard evidence of conspiracy, the courts will have to use evidence on the amount of shared information, the extent to which the parties acted independently, and the 5 6

306 U.S. 208 (1939). Id. at 227.

I. The Development of Inference Doctrine

135

complexity of the alleged agreement in order to decide the conspiracy question. Shared information is important because a collusive scheme requires some minimal knowledge basis on which to carry out the scheme. The members of the cartel must know the price level at which they can remain without fear of losing customers to other firms. They also need to be sure that new entrants will not undercut the price level the cartel sets. Recall that in our discussion of cartels in Chapter 4, we found that maintenance of the joint-profit-maximizing output allocation requires cooperation and coordination among cartel members. The informational requirements of this level of cooperation are likely to be steep. If the two firms in a duopoly happen to have the same average and marginal cost schedules, they may be able to choose a joint-profit-maximizing, equal-shares market division scheme without having to supply additional information to each other. However, it is unlikely that two firms will be so much alike. In the more likely case that the firms have different production technologies, they will need to share information on costs in order to find a market division scheme that maximizes joint profits. Why is independence an important issue? The reason is suggested by the Prisoner’s Dilemma game introduced in Chapter 4.7 Collusion is a rational policy, under reasonable assumptions, only when each member of the cartel feels certain of the cooperation of the other members of the cartel. Without this certainty, the better strategy is to undercut the competition. Indeed, recall that under certain conditions price competition is a “dominant strategy” in the sense that each firm would do it regardless of what the other firms choose to do. Evidence of this attitude – that is, that each firm would have chosen its course of conduct regardless of the decisions taken by the others – suggests strongly that firms did not collude. While strong evidence of independence goes against an inference of conspiracy, the proper conclusion in the absence of such evidence is less clear. If successful collusion requires monitoring and enforcement, then it would seem that even in the absence of evidence of independence (or in the presence of evidence of interdependence), plaintiffs still should be required to bring forth evidence of a monitoring and enforcement mechanism.8 In any event, evidence of monitoring and enforcement certainly 7 8

For discussion, see Chapter 4. See Chapter 4 (discussing Posner’s views on conscious parallelism).

136

Agreement

bolsters the charge that the parties were not acting independently. The facts of Interstate Circuit indicate the presence of a monitor and enforcer: the manager of the affiliated Texas theaters. Lastly, complexity presents an obvious concern. If the parties are simply answering “no” to a request, then there would be little reason to believe that they had worked things out in advance. Take for example, the case of several banks turning down an individual for a loan. Although when examining the loan request, the banks process detailed and complex information, the final result is a simple yes or no. Since it is quite possible that they could all independently conclude that the applicant failed to qualify for a loan, the answer “no” could hardly serve as evidence of conspiracy. However, if each bank answered “no” and provided a detailed explanation, not the result of a standardized process, that matched that of the other banks, then the conspiracy inference would gain some plausibility. The absence of complexity does not imply that the parties could not have conspired, and the presence of complexity is not sufficient to prove the existence of a conspiracy. Two firms that have adopted a jointprofit-maximizing agreement may be able to implement their plan through a simple market division scheme, such as “you take East, I’ll take West.” Two firms that have matched each other’s products down to the level of the most minute specifications may have done so because of competition rather than collusion. Of the three elements considered – shared information, independence, and complexity – it should be clear that independence is the most important. Information sharing may or may not be observed in a cartel, depending on the technological differences between the firms. Complex agreements may or may not be observed. However, independence goes to the heart of the matter; proof of independence should put an end to the case.9 There is an important difference between the terms independent and unilateral. If a firm’s incentives are independent of other firms, then its conduct will clearly be unilateral. The converse is not necessarily true: a firm that acts unilaterally may not be acting independently. In other words, even if a firm’s incentives are not independent of others, it may still have acted unilaterally. For example, in the two firm oligopoly 9

This was the view advanced by Donald Turner in his article The Definition of Agreement under the Sherman Act: Conscious Parallelism and Refusals to Deal, 75 Harv. L. Rev. 655 (1962).

I. The Development of Inference Doctrine

137

studied in Chapter 4, each firm chooses its output conditional on the anticipated output choice of its rival, a case of interdependence coupled with unilateral conduct. This suggests the importance of having additional evidence (e.g., monitoring or enforcement efforts) before one draws an inference of conspiracy from the fact of interdependence.10 To illustrate, consider the following example.11 Suppose a large bookseller – call the firm Bookbuster – happens to have a standard form contract with three publishers. Bookbuster wants the publishers to impose a minimum resale price agreement on all other booksellers in their supply network. The publishers do not want the agreement because it would only serve to enhance Bookbuster’s profits while reducing theirs (in other words, this is a case of anticompetitive resale price maintenance). In particular, assume the agreement would reduce the profits of each publisher by $10,000. Finally, suppose Bookbuster has a clause in the contract that says that if a majority of the publishers under the agreement accept the renegotiated terms, they apply to all of the publishers under the agreement. Ordinarly, one would assume that Bookbuster could not get the publishers to accept this new deal unless it compensates each publisher for the $10,000 loss in profits. Suppose Bookbuster offers $5,000 in exchange for a promise to accept the new agreement. Even though this is only half the amount needed to compensate each publisher for its loss under the new agreement, there is a good chance the publishers will accept the new deal. First, note that if one publisher knows that the other two have rejected the deal, it may as well promise to accept. It would gain $5,000 and lose nothing, since a majority would have rejected the deal. Second, if one publisher knows that the other two have promised to accept, he may as well accept, thereby cutting his loss to $5,000. The only case in which a publisher has the appropriate incentive is when one publisher has promised to accept and one has rejected – in other words, when his vote is the tie-breaker. But as we expand the number of publishers in this example, it becomes unlikely that any publisher will view his vote as the tie-breaker, unless the publishers are all in communication with each 10

11

For an examination of this issue in the context of Interstate Circuit, see David A. Butz and Andrew N. Kleit, Are Vertical Restraints Pro- or Anticompetitive? Lessons from Interstate Circuit, 44 J. Law & Econ. 131 (April 2001). Butz and Kleit show, among other things, that the film distributors in Interstate Circuit may have operated under conditions in which their incentives were interdependent, but their conduct was unilateral (or individually rational). The example is based in part on the analysis in Butz and Kleit, supra note 10.

138

Agreement

other. Given this, there is a good chance the publishers will end up accepting Bookbuster’s deal.12 If this is the outcome, it will have some interesting properties. First, it may look like collusion, since it involves all of the publishers accepting a complicated contractual provision, just like the film distributors in Interstate Circuit. Second, it is a case of interdependence, since the majority rule provision in the Bookbuster’s contract makes the publishers’ incentives interdependent. Third, the end result seems to violate the individual rationality criterion, because each publisher is worse off under the new agreement. It should not be difficult to find evidence that a few of the publishers (if not all) did not want the new agreement. In short, this hypothetical could present a strong case for inferring conspiracy among the publishers under the Interstate Circuit doctrine, even though it involves unilateral conduct.

B. Unilateral Contract Theory Interstate Circuit introduced a “unilateral contract theory” of inference into antitrust law.13 According to contract law, a party may be bound by the promises issued in a unilateral contract. For example, consider a man who offers to pay $100 to the person who returns a dog named Spot, bearing certain features. The man is liable for the reward money to the person who returns Spot, even though the finder never formally accepted the offer. Several passages of Interstate Circuit suggest that acceptance of an offer to fix prices can in itself establish a violation of Section 1.14 Probably more than any other statement in an antitrust opinion, the unilateral contract theory provided support to the view, expressed by some commentators, that the courts were in the process of shredding the conspiracy requirement.15 However, the commentators showed little concern for the troubling scope of the unilateral contract theory. Consider the routine case of one airline announcing a price increase, and then competing airlines following with similar price increases. The unilateral contract theory implies that this pattern could be sufficient to establish a violation of Section 1. 12

13 14 15

For a rigorous examination, see Zvika Neeman, The Freedom to Contract and the FreeRider Problem, 15 J. Law, Econ. & Org. 685 (October 1999). See also Eric B. Rasmusen, J. Mark Ramseyer, and John S. Wiley, Naked Exclusion, 81 Am. Econ. Rev. 1137 (1991). See Chapter 4 (discussion of Posner). 306 U.S. at 226–7. Eugene V. Rostow, The New Sherman Act: A Positive Instrument of Progress, 14 U. Chic. L. Rev. 567, 575 (1947).

I. The Development of Inference Doctrine

139

We can describe the further development of Section 1 inference doctrine as a process of refining the inference rules developed in the earlier cases and at the same time confining, limiting, or narrowing the unilateral contract theory. However, before the process began, the Court blew another gust of wind in the sails of the unilateral contract theory with its decision in American Tobacco Co. v. United States.16 A jury convicted the defendants for violating Sections 1 and 2 of the Sherman Act. The jury had based its verdict to a large extent on circumstantial evidence.17 The Court faced a narrow issue: whether actual exclusion of competitors is necessary to the crime of monopolization under Section 2 of the Sherman Act. The Court held that neither proof of exertion of power to exclude nor proof of actual exclusion of existing or potential competitors is essential to sustain a charge of monopolization under the Sherman Act. The expansion of inference doctrine occurred in this case, as in the others, through comments tangential to the holding. The Court discussed the circumstantial evidence in a manner that suggested approval of the jury’s finding, then went on to make some general observations about what plaintiffs must show in order to prove conspiracy. Three of them attracted a great deal of attention. (1) It is not of importance whether the means used to accomplish the unlawful objective are in themselves lawful or unlawful. Acts done to give effect to the conspiracy may be in themselves wholly innocent acts.18 (2) No formal agreement is necessary to constitute an unlawful conspiracy. Often crimes are a matter of inference deduced from the acts of the person accused and done in pursuance of a criminal purpose.19 (3) The essential combination or conspiracy in violation of the Sherman Act may be found in a course of dealing or other circumstances as well as in an exchange of words.20

Let us return, again, to the discussion of cartel theory started in Chapter 4. I noted that two types of case fall in the “conscious parallelism” category. The first really presents a problem of inference. The

16 17

18 19 20

328 U.S. 781 (1946). The evidence was as follows: (1) The defendants had raised prices on cigarettes during the depression, when demand, the Court assumed, had fallen. (2) The defendants had raised prices simultaneously and by the same amount. (3) The defendants had high profit margins. 328 U.S. at 791–8. 328 U.S. at 809. Id. at 809. Id. at 809–10.

140

Agreement

prosecution’s theory is that an explicit agreement exists, but it is difficult to find evidence of it. The second involves instances where the firms agree tacitly, but not formally; they happen to see what is in each other’s interests and act accordingly. The second case accurately describes conscious parallelism, while the first is conspiracy. I noted in Chapter 4 that these cases appear different if viewed from an economic framework, but from the perspective of the antitrust enforcement agent they are indistinguishable. Thus, a decision that Section 1 does not apply to cases of the second type, because of the absence of conspiracy, would render the statute ineffective with respect to a substantial portion of cases of the first type. The Supreme Court’s words in American Tobacco apply to both types of case equally. The Court made no effort to suggest that the second case type (conscious parallelism) is one with respect to which it is doubtful whether the Sherman Act applies. The Court’s refusal to distinguish the two case types signaled an expansion of the scope of Section 1.

ii. rejection of unilateral contract theory The Court began to distance itself from the unilateral contract theory in Theatre Enterprises v. Paramount Film Distributing Corp.,21 and the distance continued to grow for some time after this decision. The plaintiff owned a suburban theater that had repeatedly sought to obtain first-run feature films from major motion picture distributors. The distributors turned him down, adhering to a policy of restricting first-runs in Baltimore to the eight downtown theaters. At bottom, the issue was whether the distributors’ conduct toward the plaintiff stemmed from independent decisions or from agreement, tacit or express.22 The Court upheld a verdict that the distributors had not conspired to deny the suburban theater owner access to first-run films, because the defendants had offered sufficient indications of the independence of their actions. In other words, each distributor would have continued to refuse to license first-run films to the suburban theater owner, even if one or more of the other distributors had granted the license. The most important justification offered by the defendants was 21 22

346 U.S. 537 (1954). There were a few procedural matters standing between the Court’s analysis and this underlying issue. Relying on Interstate Circuit, the plaintiff had sought a directed verdict, which the lower court denied. The jury found for the defendants. Thus, the precise issue was whether, taking the jury’s evidentiary findings as valid, a reasonable jury could conclude that the defendants had not violated Section 1.

II. Rejection of Unilateral Contract Theory

141

that the plaintiff’s suburban theater (1) would have to hold an exclusive license, and (2) could not, given its location, match the revenue brought in by the downtown theaters.23 A comparison to Interstate Circuit reveals the key piece of evidence missing in the plaintiff’s case. Recall that in Interstate Circuit there was evidence of disagreement among the film distributors’ managers with respect to the proposals the exhibitors made. Here, there was no evidence of disagreement among the film distributors about the policy of licensing new films only to downtown Baltimore theaters. There was no evidence to contradict the claim that the distributors had acted on the basis of independent incentives. The unilateral contract theory would seem to require a finding of a Section 1 violation in this case. There was a common policy, all the defendants adhered to it, and the policy restrained trade. However, the Court distanced itself from the unilateral contract theory with the following statement: “[B]usiness behavior is admissible circumstantial evidence from which the fact finder may infer agreement. But this Court has never held that proof of parallel business behavior conclusively establishes agreement or . . . that such behavior itself constitutes a Sherman Act offense. Circumstantial evidence of consciously parallel behavior may have made heavy inroads into the traditional judicial attitude toward conspiracy, but ‘conscious parallelism’ has not yet read conspiracy out of the Sherman Act entirely.”24 The unilateral contract theory has never regained the stature that it had in the period 1939 to 1954, that is, the period of the New Sherman Act. The modern decisions involve summary judgments granted to defendants. The doctrine that has emerged requires parallelism “plus” something else in order to survive a summary judgment motion.25 Generally, the plus factor is some set of facts or circumstances that suggests that the defendants were not acting independently.26 23 24 25

26

346 U.S. at 540. 346 U.S. at 540–1. Phillip Areeda, Antitrust Analysis 372 (3d ed. 1981). For a summary of cases, see Phillip Areeda and Louis Kaplow, Antitrust Analysis 301–14 (4th ed. 1988). For a recent case illustrating the hurdles placed before plaintiffs, see Reserve Supply Corp. v. OwensCorning Fiberglass Corp., 799 F. Supp. 840 (N.D. Ill. 1990) (rejecting circumstantial evidence, and concluding that “parallel behavior and the hope that something further can be developed at trial is not sufficient to warrant a trial on the merits”). For example, in Ambook Enterprises v. Time, 612 F.2d 604 (2d Cir. 1979), the Court reversed a summary judgment for the defendants because of evidence suggesting coercive pressure had been a feature of the defendants’ arrangement.

142

Agreement

Some of the most important statements from the Supreme Court appear in Matsushita Electric Industrial Co. v. Zenith Radio Corp.27 The plaintiffs, United States television manufacturers, alleged that their Japanese rivals had conspired to charge below-cost prices (predatory pricing) in the United States in order to drive them out of business. After ten years of discovery, the district court granted summary judgment for defendants. The Third Circuit reversed, and the Supreme Court reversed the Third Circuit. The Court announced the following propositions: 1. Conduct equally consistent with permissible competition and illegal conspiracy does not, without more, support even an inference of conspiracy. 2. The plaintiff must present evidence which tends to exclude the possibility that the alleged conspirators acted independently. 3. The plaintiff must show that the inference of conspiracy is reasonable in light of competing inferences. The first two statements pretty much bury the unilateral contract theory. Whether it will be unearthed in some later case remains to be seen. The third statement is the most important, and has interesting implications. How does one show the degree of reasonableness of an inference of conspiracy? Consider the facts of Matsushita. There, the inference of a twenty-year conspiracy to suffer losses in order to ruin the plaintiffs seemed unreasonable for several reasons. First, two American manufacturers remained in the market, with a combined domestic market share of 40 percent. Second, nothing prevented new producers from entering the market. Third, powerful incentives to cheat existed on the side of the Japanese. Each producer had an incentive to free-ride on the predatory investments of the others, and to cheat on the cartel pricing that presumably would occur during the long-awaited period of recoupment. One interesting feature of Matsushita is that the Court used evidence on market structure to reject a Section 1 conspiracy claim. Suppose the plaintiffs in Matsushita were able to prove that the defendants’ prices were predatory and had introduced hard evidence of conspiracy, such as transcripts saying, “Let’s keep underpricing our TVs until we have driven the American makers out of business and achieved freedom to gouge the American consumer.” The result, of course, would have been a per-se violation of Section 1, irrespective of the number of American produc27

475 U.S. 574 (1986). For additional discussion of Matsushita, see Chapter 10.

II. Rejection of Unilateral Contract Theory

143

ers remaining or the ease of entry. Socony instructs courts to ignore evidence on market power after a demonstration of conspiracy. But Matsushita indicates that if only circumstantial evidence of agreement exists, courts must take market power evidence into account in assessing the plausibility of the conspiracy charge.

8 Facilitating Mechanisms

Facilitating mechanisms make collusion more effective by serving one or more of the following functions: establishing a method of monitoring compliance, enabling parties to enforce compliance, and hiding collusion from public view. The important cases have dealt with data dissemination plans. The central doctrinal issue is whether per se or rule of reason analysis applies to facilitating mechanisms, such as data dissemination plans. Start with the basic proposition from Socony that price fixing is per se illegal. It follows that a data dissemination plan violates Section 1 if it is a necessary and sufficient element of a price-fixing plan. In general, however, the link between the dissemination plan and the price-fixing charge presents a difficult question. A data dissemination plan may facilitate collusion, or it may facilitate other less harmful or even beneficial ends. In addition, there will not be clear evidence in these cases of an agreement on prices, for otherwise the existence of a data dissemination plan would be irrelevant. The plaintiff therefore must prove, more or less, that the dissemination plan was a sufficient element of a price-fixing conspiracy, so that it would be appropriate for a court to infer conspiracy from the existence and operation of the plan. Per se and rule of reason tests require different standards of proof when applied to facilitating mechanisms. A per-se rule dispenses with the plaintiff’s burden of establishing the inference of a conspiracy to restrain trade. A rule of reason test, on the other hand, requires the plaintiff to prove conspiracy. Because the plaintiff usually will not have clear evidence of agreement, he will have to rebut alternative justifications for 144

I. Data Dissemination Cases

145

the facilitating mechanism to show that the practice could serve only one purpose. This is a heavy burden. The cases in this chapter trace the development of the respective boundaries of the rule of reason and per-se standards. However, before discussing the cases, let us return briefly to an issue confronted in Chapter 4 (cartels). Recall that the difficult Section 1 cases can be divided into two categories: those in which the parties have an explicit agreement, but the available evidence falls short of demonstrating it (inference cases); and those without an explicit agreement, but with tacit collusion (conscious parallelism cases). The case types have a connection in the sense that a decision not to apply Section 1 of the Sherman Act to “conscious parallelism” cases would make it difficult to enforce the statute with respect to “inference” cases. A facilitating mechanism could support collusion in either type of case because it helps the parties detect cheating on the (explicit or tacit) collusive agreement. Thus, adoption of the per-se rule with respect to facilitating mechanisms enables Section 1 to be applied to tacit collusive agreements, whereas adoption of the rule of reason test effectively limits Section 1 to explicit collusive agreements. For these reasons, this chapter – like the previous one – really concerns the reach or scope of Section 1. We learned in the previous chapter that the Supreme Court eventually made it clear that parallel behavior alone generally would not, in the absence of other evidence, support a finding of an unlawful conspiracy. Later decisions suggested that the plaintiff must offer some evidence of collusion in addition to parallel conduct, in order to survive a summary judgment motion. All of the cases in this chapter fall in the parallelism-plus category; they involve parallel conduct in combination with some coordinating structure that could facilitate collusion.

i. data dissemination cases The Court leaned toward a per-se standard under Section 1 in its first data dissemination case. American Column & Lumber Co. v. United States1 involved a trade association made up of around four hundred firms supplying hardwood lumber to local mills and lumber yards. The firms making up the association produced about one-third of the industry’s national output. The association adopted a plan (hereafter “the Plan”) that required member firms to submit, among other things, price 1

257 U.S. 377 (1921).

146

Facilitating Mechanisms

lists, detailed daily sales and shipment reports (including invoice copies), monthly production and stock reports. The plan required the secretary of the association to distribute to members a monthly summary of each member’s production, subdivided by product grade; a weekly report of sales and shipments; a monthly inventory of the stock of each member; a monthly summary of each member’s price lists, showing prices asked and supplemented by reports of changes; and a monthly report on market conditions.2 At the time the government brought suit, 365 firms participated in the Plan. The Plan attempted to put into practice the “new competition” ideas of a lawyer named Arthur Jerome Eddy. Eddy had urged firms to share price and output information in order to encourage price stability. The officers of the Plan claimed price stability as the key benefit resulting from participation. While promoting the Plan they were careful to say that it did not require price fixing, though they also noted that members typically follow their most intelligent competitors. The Supreme Court held that the Plan violated the Sherman Act. The Court admitted that there was no explicit agreement to limit production or to fix prices, and the evidence of conspiracy was entirely circumstantial. Two findings bolstered the inference of conspiracy. First, the members shared information on prices and met monthly, and in some areas as often as weekly, to discuss the price and output reports. Thus, the members of the Plan had sufficient knowledge to carry out a pricefixing conspiracy. Second, given the absence of direct evidence of conspiracy, the Supreme Court found the missing ingredient supplied by the “disposition of men to follow their most intelligent competitors,”3 coupled with the pressures they may have felt to maintain the social contacts and reputations forged through participation in the Plan. The enforcement mechanism – that is, the penalty that ensured compliance with the alleged collusive agreement – came as a byproduct of the Plan’s benefits according to the Court. Members participated because of the fear of losing out on the additional revenue earned through participation, as well as the important social contacts and prestige that came from the cooperative effort. Having established the intent prong, the Court next laid the groundwork for the findings that the parties had the power to and actually did affect prices. The Court noted that in some of the monthly meetings, the 2 3

257 U.S. at 396. Id. at 399.

I. Data Dissemination Cases

147

officers urged members to cut back on production in order to keep prices stable. Finally, the Court said the Plan had a dramatic effect: prices for several grades of hardwood rose anywhere from 30 to 340 percent in 1919. The Court added that even given the high rate of inflation in 1919, it “could not but agree with the members of the ‘Plan’ themselves . . . in the conclusion that the united action . . . contributed greatly to this extraordinary price increase.”4 By now you should have noticed something wrong with the Court’s reasoning. Start with the unlawful purpose and conspiracy findings. The enforcement mechanism the Court identified consisted largely of the supposed benefits of the organization: price stability and contacts with fellow tradesmen. This is quite a slim reed on which to hang a conspiracy finding. Remember, cartels lack stability precisely because the benefits to cheating exceed the benefits to participating, provided the other members participate. But the Court in essence says that this cartel had an unusual degree of stability because the benefits of participating were too great to miss. This is inconsistent with the theory of cartels. Indeed, if the Court applied this reasoning consistently, the enforcement mechanism would appear in every conspiracy case, because each cartel member would lose out on the cartel profits if he chose to cheat. But he chooses to cheat precisely because the benefits of cheating exceed those of compliance. In essence, the Court relied on a finding that an enforcement mechanism existed to bolster its conclusion that the Plan operated as a conspiracy. In the absence of direct evidence of agreement, evidence of an enforcement mechanism can serve as a substitute.5 But in order to use enforcement as a substitute, there must be independent evidence of enforcement – meaning independent of the supposed benefits of the cartel. The Court’s reliance on the benefits of the cartel as a sufficient enforcement mechanism allowed the government to evade the requirement of proving conspiracy. What about the Plan officers’ statements? Certainly their promotional efforts suggest this was an effort to fix prices, with a corresponding

4 5

Id. at 409. See Richard A. Posner, Oligopoly and the Antitrust Laws: A Suggested Approach, 21 Stan. L. Rev. 1562–606 (1969) (discussed in Chapter 4). Indeed, one could argue that the Plan in American Column serves as an example of a “trigger mechanism” supporting collusion (see Chapter 4). If the parties’ discount rates are relatively low, the sanctions imposed on cheaters could be sufficient to deter price cuts. But in order to be useful in court, this theory needs to be distinguished from the simple claim that the parties did not cheat because they did not want to lose the benefits from collusion.

148

Facilitating Mechanisms

purpose and intent to violate Section 1. But one should question this evidence for several reasons. First, the officers of the Plan had every incentive to claim that it would be a success, since each received a salary from the association. Second, one has to wonder whether the officers themselves believed the Plan had such a dramatic effect on prices. The roughly 350 members of the Plan produced less than one third of the industry’s entire output. Even assuming that the officers had control over the pricing of the Plan participants, how could a group with such a small market share have the influence to bring about industry-wide price stability? The disposition of men to follow their most intelligent competitors does not supply the missing ingredient. That inclination does not reveal itself in a sufficiently consistent manner to serve as the single piece of evidence in support of a conspiracy finding. To sum up, the government’s attempt to prove conspiracy in American Column had several holes in it. By accepting such weak proof of conspiracy, the Court in American Column essentially adopted a perse standard for data dissemination plans. Holmes and Brandeis dissented. Holmes thought that there was nothing wrong with information dissemination standing alone, and suggested that it could enhance efficiency. Holmes’s reference to efforts to “conform to market conditions”6 suggested that he envisioned a competitive market in which some traders did not know the market price – that is, the equilibrium price at which demand equals supply – for the item they were selling.7 For a firm in a competitive market, knowledge of the market price would not lead to price fixing. The firm would take the market price as given and try to sell as much as it could at that price. Brandeis described the scheme as one that allowed small hardwood producers to counterbalance the informational advantages enjoyed by large dealers and producers, who had better information on prices. The large dealers could, under these conditions, take advantage of small producers. One could argue that the heavy burden a rule of reason test would impose on enforcement agencies with respect to information sharing arrangements makes a per-se rule more desirable. However, a per-se rule

6

7

257 U.S. at 412 (“A combination in unreasonable restraint of trade imports an attempt to override normal market conditions. An attempt to conform to them seems to me the most reasonable thing in the world . . .”). See Chapter 1 for discussion of market prices.

I. Data Dissemination Cases

149

creates a broad-brush approach to enforcement that tends to crush both bad and good information sharing agreements. The recurring tension between administrative enforcement concerns – specifically in avoiding the imposition of an impossible burden on antitrust enforcement officials – and the economic reasonableness of the doctrine, comes out through the price-fixing cases. The government’s burden under a rule of reason standard would be hard to meet. In addition, a rule of reason test (i.e., requiring proof of conspiracy) would render Section 1 ineffective against the tacit collusion facilitated by data dissemination plans. On the other hand, a lower standard of proof would effectively prohibit some potentially desirable information sharing practices. The economic reasonableness side prevailed in Maple Flooring Manufacturers Assn. v. United States,8 which adopted a rule of reason test. The Maple Flooring Manufacturers Association consisted of twentytwo corporations that produced, shipped, and sold 70 percent of the total output of maple, beech, and birch flooring. The association disseminated: (1) information detailing its members’ average costs, (2) a single base point freight rate booklet,9 (3) anonymous information on quantity and prices by reporting members and amount of stock on hand. All reports concerned past and completed transactions. Finally, the association held monthly meetings at which members discussed industry problems. The Court held that the association’s activities did not violate Section 1. More precisely, the Court held “only that trade associations which openly and fairly gather and disseminate information as to the cost of their product, volume 1 . . . , price on past transactions, . . . stocks, meet and discuss without attempting to fix prices do not violate Section 1.”10 The important question, of course, is how the Court distinguished the Maple Flooring plan from that in American Column. The Court held, first, that the Maple Flooring plan, unlike that in American Column, failed to establish price uniformity. Second, the evidence suggested that the members of the association did not use this plan to facilitate price fixing. Why? The price information disseminated dealt with past rather

8 9

10

268 U.S. 563 (1925). What is a single base point freight rate booklet? Let the basing point be X. Then all prices at locations other than X are equal to the base price of a good purchased at X plus the shipping charge. So if you buy the good at Y, then you pay base price at X + cost of shipping from X to Y. This is the price, whether or not the good was produced at and shipped from X. Because the plan seems at first glance to be strange, it has been a long-standing concern to antitrust lawyers. 268 U.S. at 586.

150

Facilitating Mechanisms

than current transactions. Moreover, the members generally did not discuss prices at monthly meetings of the association. Adopting the reasoning of Holmes’s dissent in American Column, the Court suggested that the plan enhanced efficiency by providing “a better understanding of economic laws and a more general ability to conform to them.”11 The rule of reason requires an analysis of the purpose, power, and effects of trade restraints. The Court concluded that the Maple Flooring plan had no effect, and that it purported not to fix prices, but instead to promote more intelligent competition. The Court said very little about the power of the parties to fix prices. However, the holding that the plan had no effect on prices implied that the parties lacked power to control the market price. The information sharing plan in Maple Flooring could have served to facilitate a tacit or explicit collusive agreement. If firms could observe the average costs of their competitors, then each knows that as long as its price exceeds the average cost of the least efficient firm in the market, they will all make a profit on every unit of output sold. For example, suppose firms A, B, and C have average cost $1, $1, and $2, respectively. The dissemination of information on average cost could serve as a signal to A and B to set price at $2. However, this may not be the result. The outcome in which A and B set price at $2 requires each of them to sacrifice a chance for the greater profits they could earn by charging less than $2 and selling more units. Further, we know from basic cartel theory that firms A and B each have a strong incentive to undercut the $2 price level. Indeed, we know from the model of Bertrand (price) competition (Chapter 4) that the equilibrium strategy for firms A and B is to set price at $1 (unless the firms face capacity constraints). Since a persuasive story can suggest both the tacit collusion and competition outcomes in this example, the choice between rule of reason and per-se standards in essence becomes a decision to allocate the burden of proof. The rule of reason places the burden on the prosecution to demonstrate agreement, which is quite difficult in the example given here and more generally in any case involving tacit collusion. The per-se rule removes this burden from the prosecution. One question remaining from American Column and Maple Flooring is precisely how a dissemination plan could be economically efficient. Isn’t the only purpose of such plans to fix prices? Suppose the market is competitive, but because of infrastructure inadequacies, some firms find 11

Id. at 584.

I. Data Dissemination Cases

151

it hard to discover the market price. Under the textbook model, a firm can sell as much as it wants at the competitive price. If it charges above the competitive price, it sells nothing. If it charges below the competitive price, it simply loses the revenue it could have made by selling at the competitive price. If the firm deals with sophisticated consumers or dealers, it should respond to its own lack of information about the market price by setting its price well above its average cost and hoping customers will purchase the good. But if dealers can costlessly supply market information, this approach is socially wasteful because it leads to fewer sales overall than if the price were set at the competitive level. The problem becomes even more complicated in the setting of a general price inflation, which describes the relevant periods for both American Column and Maple Flooring. Price inflation makes it harder for producers to discern changes in the market price of items they buy or sell. The producers cannot easily tell what portion of the price change to attribute to changes in market conditions (i.e., fundamentals) versus those that result from inflation. In other words, “real” price changes reflect shifts in demand or supply for a given item, and phantom changes occur because of general price inflation. The term phantom accurately describes the changes because a general inflation affects all prices equally, and therefore has no effect on the resource constraints binding on a producer. The danger of general inflation becomes apparent when producers mistakenly believe that changes resulting from general price inflation really reflect changes in market fundamentals, and thus alter production in response to an unreliable or uninformative price signal. Such responses are socially wasteful. It follows that dissemination of accurate information on the market price may enhance efficiency. But who would supply it? As Posner has noted,12 no single firm has an incentive to supply it to others. The trade association is a solution. By pooling information from diverse sources, the trade association can lower the cost of gathering information on market prices and conditions; and by making receipt conditional on the supply of information, the association can provide incentives for its members to supply information. Posner has argued that if a data dissemination plan serves simply as a method of discovering market price, it should reduce dispersion around 12

Richard A. Posner, Information and Antitrust: Reflections on the Gypsum and Engineers Decisions, 67 Geo. L. J. 1187 (1979).

152

Facilitating Mechanisms

the average price without altering the average price level.13 However, if, as Brandeis argued, the agreement prevented advantage-taking by better informed traders, then one should observe an increase in the average price level and a decline in inventories as a result of the plan. Why? The small producers who did not know the market price and therefore would have set price too high would now sell at the market price, getting rid of their unsold stock. The small producers who would have set price below the market price would raise it to the market price. The average transaction price increases because those who would have sold at a price below the market will sell at the market price and those who would not have sold at all will sell at the market price. Many years after Maple Flooring, the Court decided United States v. Container Corp. of America.14 The case involved eighteen firms that shipped 90 percent of the cardboard cartons in the Southeastern United States. The firms set up an information exchange whereby suppliers would provide each other information on price charged on their most recent sales to a particular customer. The government brought a civil antitrust suit, charging price fixing in violation of Section 1. The district court dismissed the complaint after a full bench trial, and the Supreme Court reversed the dismissal. The issue was whether the arrangement was likely to violate Section 1. The Court held that the information exchange violated Section 1. The surprising part of the Court’s opinion is the holding that the perse rule applied. Even though the evidence suggested that the defendants used the information to meet prices of competitors, the Court, citing Socony, held that mere “interference with the setting of price by free market forces is unlawful per se.”15 This is inconsistent with the rule of reason standard applied in Maple Flooring. However, the Court backpedaled immediately after suggesting that the per-se rule applied when it noted that “[p]rice information exchanged in some markets may have no effect,” but that this was an oligopolistic market.16 The inference antitrust commentators draw is that the opinion announces a more restrictive rule for price information exchanges in oligopolistic markets. It is hard to believe that the Court’s opinion

13 14 15 16

Id. at 1197. 393 U.S. 333 (1969). Id. at 337. Id.

I. Data Dissemination Cases

153

announced a general per-se rule, because it made no effort to overrule Maple Flooring. However, the precise nature of the rule implicit in Container is hard to discern from the opinion itself. In addition to the point that oligopolistic markets warrant different treatment, the Court noted that the exchange seemed to stabilize prices, and that demand for the product was inelastic. Thus, one could argue that Container holds that per-se analysis applies to price information exchange agreements operating in oligopolistic markets with inelastic demand. The suggestion that the per-se rule of Container applies to oligopolistic markets with inelastic demand makes sense from an economic perspective. Evidence of inelastic market demand tends to make the tacit collusion inference more plausible because a firm is less likely to undercut the collusive price if demand is inelastic. The reason is as follows. In general, the firm that undercuts the collusive price gains through two routes: by taking all of the new orders induced by the lower price, and by the shift of orders (that would have been placed at the collusive price) from other firms to the price-cutter. If demand is inelastic, few new orders will be induced by the price cut. However, the shifting of orders should still be large enough to provide an incentive to undercut the collusive price. Of course, the shifting of orders will be observed by other firms (particularly those firms that lose orders) and so the price-cutter’s decision to deviate from the agreement will be noticed by its competitors. If the fear of retaliation is sufficiently great, firms will be deterred from cutting price. In a concurring opinion, Justice Fortas declared that the opinion did not announce a new per-se rule. He argued that in light of evidence in the record that the agreement actually caused prices to stabilize at a higher level, the agreement reduced competition. That was sufficient in his view to find a violation of Section 1. Recall that Chicago Board of Trade requires nothing more than a showing under the rule of reason that the trade restraint reduced competition. Justice Marshall’s dissent relied on the framework of Maple Flooring, which, in light of the potential benefits of price information exchanges, requires the plaintiff to produce either proof of conspiracy or evidence of price effects in order to find a Section 1 violation. Note that competing inferences are possible here. One, along the lines of the Stigler article (Chapter 4), is that exchange of information made it easier to monitor an explicit or tacit collusive agreement and to detect cheating. It eliminated the need for the cartel participants to assess buyer honesty. The other inference is that the participants were competing, but wanted to

154

Facilitating Mechanisms

do so on the basis of solid evidence. If buyers lie about price quotes from other sellers, then it is both harder for firm members to monitor a collusive agreement and for competitors to know whether they should meet competition. In light of the structure of the market and the ease of entry, Marshall thought it unlikely that the plan could have served as a facilitating device for a price-fixing agreement. The government’s case was inadequate in Marshall’s view because of the lack of evidence of a price-fixing agreement, and the ambiguity of the evidence concerning effects. The Court clarified the holding of Container in United States v. United States Gypsum Co.,17 which involved a concentrated industry.18 The Gypsum Court adhered to Container’s treatment of price information exchanges in oligopolistic industries. Thus, Container and Gypsum have destroyed the simplicity of the Maple Flooring rule that the courts should treat price information sharing plans under the rule of reason. The Court has not stated the rule clearly. However, the Court seems to prefer application of the per-se rule in the case of an oligopolistic industry. In footnote 16 of Gypsum,19 the Court suggested that although the general standard with respect to information exchanges is the rule of reason, perse analysis may be appropriate when (1) the industry is oligopolistic, or (2) the exchange includes information on current prices, as in American Column. Although the law is not crystal clear, one thing is certain: Price information exchanges are not per-se violations of the Sherman Act.20

ii. basing point pricing and related practices Basing point pricing refers to a combination of pricing practices observed in certain industries, typically those with substantial shipping costs and homogeneous output. Consider the single base case. Suppose Pittsburgh is the basing point and a firm ships widgets to Texas, Illinois, and Florida. The shipping costs from Pittsburgh to these points are $100, $200, $300 respectively. The price of the widget, if sold at the plant door, is $1,000. Suppose there is another widget plant in Detroit. Then under a single base system, the Detroit firm would charge “Pittsburgh plus” prices, which means that the Detroit firm would charge $1,100 for 17 18

19 20

438 U.S. 422 (1978). Gypsum also establishes the rule that proof of specific intent is necessary in a criminal antitrust proceeding. Id. at 443. 438 U.S. at 441 n.16. United States v. Citizens & S. Nat. Bank, 422 U.S. 86, 113 (1975).

II. Basing Point Pricing and Related Practices

155

shipping a widget to Texas, $1,200 to Illinois, and $1,300 to Florida. In addition, the companies participating in such a plan usually prevent customers from purchasing “f.o.b. mill door” – that is, purchasing at the plant door and arranging their own shipping. The practice raises eyebrows because common sense would suggest that a Detroit company would charge a lower shipping price to an Illinois customer than would a company in Pittsburgh.

A. Basing Point Pricing Cases The Supreme Court examined basing point pricing in Federal Trade Commission v. Cement Institute.21 The Cement Institute was a trade association made up of seventy-four manufacturers of cement. The Institute employed a multiple basing point system of pricing. The FTC charged that this system facilitated a price-fixing agreement among the members, and thus was an unfair method of competition in violation of Section 5 of the FTC Act. The Commission found that the pricing system violated Section 5 of the FTC Act, and the Supreme Court held that the evidence was sufficient to support the Commission’s conclusion. There was no evidence of an explicit agreement to fix prices. In a footnote the Court stated that it is enough to warrant a finding of combination that the parties, with full information, accepted an invitation to participate in a scheme that restrained trade.22 The Court’s opinion notes that in spite of the conflicting testimony of experts, there were reasons to think that the unusually uniform price structure resulted from collusion. First, this was a concentrated industry, which implied that members of the Institute could more easily collude. Second, the institution of the basing point system seemed to coincide with an equalization of prices of cement at various locations. Third, the Institute did not allow producers to sell f.o.b. to buyers. Fourth, some evidence suggested an effort on the part of the Institute to punish cheaters. The Court pointed to boycotts of sellers who sold foreign cement below the delivered price of the domestic cement, and to evidence that the Institute punished price chiselers by establishing punitive basing points (a type of predatory pricing). 21 22

333 U.S. 683 (1948). In footnote 17 of its opinion the Court cited Interstate Circuit (see Chapter 7). Cement Institute is one of the “New Sherman Act cases” (see discussion in Chapter 7).

156

Facilitating Mechanisms

The evidence the Court relied on should be examined in light of some basic economics. First, the fact that the industry was concentrated can be used to support an inference of conspiracy, because concentration is virtually a necessary condition for the maintenance of collusion. However, concentration is not a sufficient condition for collusion. Concentration may simply reflect the superior efficiency of large firms,23 and in light of this one should have evidence that tends to exclude the efficiency thesis. Second, consider the evidence that the Institute did not allow producers to sell f.o.b. to buyers. Evidence of collusion? Suppose the Institute’s membership included many small producers who by themselves could not ship enough to take advantage of economies of transport, but combined with others, maintaining a policy of offering goods at delivered prices, could exploit economies of transport. It may have been desirable, under these assumptions, for the Institute to require members to adhere to this policy.24 A further reason for adopting such a policy as a group is that a small producer can more credibly make the claim, “I cannot sell f.o.b.” if it can say that it is simply complying with the Institute’s policy. Without the policy, large buyers would have considerable leverage in pushing a supplier to make concessions on its non-f.o.b. sales policy. Finally, any self-regulating scheme will likely have coalitions that disagree with one policy or another. A general ban on f.o.b. selling might be resisted by the largest sellers who can take advantage of economies of transport without such a policy or without having to rely on the commitment of other suppliers to ship their goods over the rails. In light of these considerations, it seems that the Institute’s policy of refusing to allow members to sell f.o.b. may not have come about as part of a plan to fix prices. At the least, the inference that the Institute’s policy could only have been adopted as part of a conspiracy to fix prices seems unwarranted. Third, the Court made much of the evidence suggesting that the prices were uniform down to several decimal places. The most exciting example, in footnote 15 of the opinion, involved an instance where several competitors delivered bids of exactly $3.2865854 per barrel of cement to the U.S. Engineer’s Office in Tucumcari, New Mexico, in 1936. 23

24

See Harold Demsetz, Two Systems of Belief about Monopoly, in Harvey J. Goldschmid, H. Michael Mann, and J. Fred Weston, editors, Industrial Concentration: The New Learning 164–84 (Boston: Little, Brown, and Company, 1974). David D. Haddock, Basing-Point Pricing: Competitive vs. Collusive Theories, 72 American Econ. Rev. 289–304 (June 1982).

II. Basing Point Pricing and Related Practices

157

How could such uniformity occur in the absence of an agreement to fix prices? Closer inspection reveals that the base price of the nearest mill was $2.10 and the land grant freight rate to which the government was entitled was $1.1865854.25 The surprising uniformity in this example came about as the result of the government’s special freight rate. Still, the question remains: why did the competitors merely meet rather than undercut the price of the firm with the best location? Doesn’t it suggest tacit collusion, a gentlemen’s agreement in which no firm undercut the price of the producer nearest to the point of delivery? This is a troubling question, and is probably the heart of the basing point pricing controversy. Inference of conspiracy is appropriate here only if the fact that disadvantaged competitors merely meet rather than undercut the price of the firm with the location advantage implies that they are tacitly conspiring to fix prices.26 The pieces of evidence that seem to be least consistent with competition are the findings that the Institute boycotted producers who sold foreign cement below domestic prices and that the Institute established punitive bases to punish firms that did not comply with the pricing plan. Even here some of the evidence warrants further questioning. First, as the Seventh Circuit noted in its opinion, the evidence in the record did not indicate that the instances of retaliation occurred pursuant to some policy of the Institute or all of its members; the evidence suggested that these instances involved either unilateral conduct or concerted action involving only a subset of Institute members.27 Second, to the extent that some of the actions of noncomplying members reduced the effectiveness of the agreement as a means of minimizing transport costs it is understandable that the Institute would try to discipline the noncomplying members. A court applying modern boycott doctrine might find some of the actions of the Institute permissible under the Sherman Act.28 25 26

27 28

Aetna Portland Cement Co. v. FTC, 157 F.2d 533, 567 (7th Cir. 1946). Could such a practice develop in a competitive market? The Seventh Circuit’s opinion in the case suggested how this could have emerged from competition. Consider two firms, A and B. Suppose their production costs are the same and A is nearest to a certain point of delivery X. If B undercut A on an order from a customer at point X, the information would reach A immediately, and A would cut its price to meet the competition. B would face the choice to meet A’s lower price or to stay outside of the market. The end result of this competitive process is a set of equal price quotations. Firms that could not meet A’s price would stay outside of the market supplying point X. See Aetna Portland Cement Co. v. FTC, 157 F.2d 533, 556–9 (7th Cir. 1946). Aetna Portland Cement Co. v. FTC, 157 F.2d 533, 545, 565 (7th Cir. 1946). See Chapter 9 (discussion of Northwest Wholesale Stationers).

158

Facilitating Mechanisms

The Supreme Court’s willingness in Cement Institute to accept some rather inconclusive evidence as proof of conspiracy should be put in perspective. Recall that the Court cited Interstate Circuit and repeated the proposition that acceptance of an invitation to participate in a plan that restrains trade could warrant a finding of conspiracy. In Chapter 7, I described this as the “unilateral contract theory” of inference, which received a great deal of support from the Court’s Interstate Circuit and American Tobacco opinions. However, the unilateral contract theory has not fared well since then. The Court’s language in more recent opinions rejects the unilateral contract theory.29 It is in this light – that is, of the modern inference case law – that we should view the Second Circuit’s more lenient treatment of a delivered pricing system in du Pont (Ethyl) v. FTC.30 At the time of the suit, the market for lead-based gasoline was declining, largely because of regulation, and was highly concentrated. Ethyl and du Pont were the two biggest suppliers in 1974 with 33 and 38 percent shares respectively. The FTC challenged the following practices of lead-based fuel producers: (1) sale of products only at delivered prices, (2) the giving by du Pont and Ethyl of additional notice of price increases beyond the thirty days required by their contracts, and (3) use by du Pont and Ethyl of a “most favored nation” clause binding them to charge each customer no more than it charged its most favored customer. The Court addressed the question whether the practices the four manufacturers of lead-based gasoline undertook violated Section 5 of the FTC Act. In administrative proceedings, the FTC had concluded that the practices violated Section 5. The Second Circuit reversed, holding that “[b]efore business conduct in an oligopolistic industry may be labeled ‘unfair’ within the meaning of Section 5 a minimum standard demands that, absent a tacit agreement, at least some indicia of oppressiveness must exist such as (1) evidence of anticompetitive intent or purpose on the part of the producer charged, or (2) the absence of an independent legitimate business reason for its conduct.”31 The Court was not persuaded that the practices themselves sufficiently caused the price uniformity and high profit margins the FTC reported. It is generally accepted now that delivered pricing plans and 29 30 31

See Chapter 7. 729 F.2d 128 (2d Cir. 1984). Id. at 139.

II. Basing Point Pricing and Related Practices

159

most favored nations clauses are not per se violations of the Sherman Act.

B. A Note on the “Most Favored Customer” Clause The most favored customer clause in du Pont (Ethyl) is a good example of a facilitating practice with potentially harmful consequences. How could such a clause be anticompetitive? Consider its effects. The clause reduces the incentive to cheat on a collusive price agreement by offering a secret discount to one of your customers. Since any discount you offer to one will have to be offered to all of your customers, and your coconspirators will see this immediately, you will be considerably less likely to offer a discount to anyone. Hence, the most favored customer clause helps maintain a stable price-fixing agreement. Indeed, by virtually eliminating the incentive to cheat, the most favored customer seems to provide a neat solution to the cartel instability problem analyzed in Chapter 4.32 Is there a defense for the most favored customer clause? One could offer a justification based on bargaining costs. Suppose the market of immediate purchasers of lead-based gasoline is concentrated, just as the seller’s market was in the du Pont (Ethyl) case. Suppose further that these purchasers are firms that sell the gasoline to consumers. Each firm would realize that it could get a substantial competitive advantage over its rival if it could bargain for a lower price from du Pont. However, bargaining is costly. The most favored customer clause may have the benefit of reducing the incentives of purchasing firms to invest too much in the bargaining process. Suppose we have two firms, A and B, negotiating with du Pont (see Figure 8.1). A and B are also competitors in the output market. If they both strike the same deal with du Pont, their unit costs are $2 each. If one strikes a “great deal” with du Pont by holding out, while the other settles early, the early settler has a unit cost of $6 (in part because he suffers a competitive disadvantage) and the hold-out firm has a unit cost of $1. If both firms hold out, they wind up with unit cost of $5, after taking

32

For a careful discussion of the most favored customer clause and several anticompetitive theories, see Jonathan B. Baker, Vertical Restraints with Horizontal Consequences: Competitive Effects of “Most-Favored-Customer” Clauses, 64 Antitrust Law Journal 517 (1996).

160

Facilitating Mechanisms

FirmA /FirmB Settle Hold Out

Settle

Hold Out

($2, $2)

(6, 1)

(1, 6)

(5, 5)

Figure 8.1

into account all of the costs of bargaining (including the costs of negotiating, search, and delay). This example has the same structure as the cartel instability example introduced in Chapter 4. If firm A holds out, it winds up with costs of either $1 or $5. If firm A settles early, it winds up with costs of either $2 or $6. Both firms A and B are better off in an outcome in which they settle early. However, the “dominant strategy” in this example is to hold out, which results in both having unit costs of $5. This example suggests that the most favored customer clause may be an efficient solution to a common strategic problem observed in “bilateral oligopoly” settings. If this is correct, then the appropriate standard for courts to apply is a rule-of-reason test, or one that requires the plaintiffs to offer further proof of conspiracy than the mere existence of a most favored customer clause.

iii. basing point pricing: economics Recall the example: Pittsburgh = basing point, Pittsburgh Price = $1,000 per widget. All steel companies, wherever located, charge $1,000 to Pittsburgh buyers. Suppose rail freight is $100 per ton to Chicago. Then all steel companies, wherever located, charge $1,100 to Chicago buyers. A simple competitive story can explain the development of such a scheme. Suppose the market in widgets is competitive with all of the producers in Pittsburgh and a new company opens in Detroit. Suppose, however, that even after the opening of the Detroit company, the Pittsburgh companies remain the established leaders and sell the lion’s share of widgets. Then competition will generate Pittsburgh-plus prices. The Detroit company will have to meet the Pittsburgh-plus prices in order to make any sales. Further, if the company is small relative to the market, it can sell all the widgets it makes at those prices. No incentive exists to undercut the Pittsburgh-plus schedule, especially if the Detroit company faces no competition from a local rival.

III. Basing Point Pricing: Economics

161

Figure 8.2

If producers spread out, Detroit could become another regional base for widget production. If Detroit producers become large enough relative to the market, they may eventually find that they can no longer sell all that they produce at the Pittsburgh-plus price. A decision to expand sales significantly may require a price reduction. We should then expect to see Detroit producers offering price reductions for customers located nearer to Detroit than to Pittsburgh. As demand increases and production spreads out, we should observe, in the limit, an outcome in which each good is distributed locally, provided that no company enjoys a cost advantage. On the other hand, if Detroit and other production sites remain small relative to the market, the Pittsburgh-plus system could continue. In Figure 8.2, the horizontal axis measures distance, and the vertical axis measures delivered cost per unit. If all producers have the same production costs, we should observe rational market divisions of the sort illustrated in Figure 8.2. In that figure, firm 1 makes all of its sales in region 1, firm 2 in region 2, and so on. However, some practices observed in delivered pricing systems have generated suspicions of price fixing. One is the refusal to offer f.o.b. prices. If the system facilitates price fixing, then the parties clearly will eliminate the f.o.b. option, for otherwise some buyers would purchase at the plant door and evade the schedule of fixed prices. F.o.b. sales would allow some sellers to cheat on the scheme. The delivered pricing system

162

Facilitating Mechanisms

prevents customers from unraveling the price schedule and makes it easer to monitor and enforce the price-fixing agreement. Two other practices that are typically mentioned as difficult to explain under a competitive model are cross-hauling (shipping the same good in opposite directions, or carrying a good across the catchment area of another producer) and freight absorption (absorbing transportation costs, as when a firm farther away from the point of delivery meets the delivered price of the firm nearest the point of delivery). David Haddock has offered a competitive model that explains many of the features of delivered pricing systems.33 According to Haddock, cross-hauling is the least difficult to explain. A producer will be reluctant to rely solely on the closest supplier for all of his requirements, for that would tie his fate too closely to that of the supplier. A policy of spreading supply sources in order to provide insurance against disruptions will likely result in some cross-hauling. Furthermore, to eliminate all cross-hauling requires extreme attention to detail. Suppose a shipper carries widgets from point A to point Z, and happens to transport more than the required number. Then it may be rational to return the excess to the plant, that is, to cross-haul. To eliminate all such instances would be costly. Figure 8.3, taken from Haddock’s article, shows how economies of transport generate shipments by firm 1 across the natural sales areas of firms 2 and 3. The concave curve shows the delivered cost per unit of firm 1’s product. Economies of transport cause the per-unit delivered cost for firm 1 to fall below that of firm 2 for some sites that are closer to firm 2 than to firm 1. The refusal to sell f.o.b. can be explained, according to Haddock, by economies of scale in production and in transport. In order to take advantage of these economies, a firm may find it necessary to refuse to permit customers to purchase f.o.b. The same argument explains instances of freight absorption. Suppose that because of elastic demand, the producer can expand sales significantly in some distant location by cutting the delivered price to that location. The increase in sales allows the producer to take advantage of economies of scale in production. Although the delivered price suggests freight absorption, the seller does not lose money overall. 33

David D. Haddock, Basing-Point Pricing: Competitive vs. Collusive Theories, 72 American Economic Review 289–304 (June 1982).

III. Basing Point Pricing: Economics

163

Figure 8.3

The big question is why firms would adopt this system. Could it result from any process other than coordination? The evolution of a system of basing point pricing is consistent with a model of competition (though not “perfect competition” as described in Chapter 1). Suppose a seller located in Texas faces competition from a production center, such as Pittsburgh. If price rises to PC in Figure 8.4, the Pittsburgh firms will ship to Texas. However, if price falls below PC, the Texas firm has the regional market to itself, which is consistent with the downward sloping demand curve for prices below PC. Thus, the demand curve for the Texas firm’s output has a “kinked” shape: horizontal at PC, and then downward sloping for prices below PC. The marginal revenue curve for the Texas firm has a jagged, lightning-bolt shape: horizontal at PC, vertical for prices below PC (see the vertical dashed line) until the vertical dashed line meets the dotted downward sloping line, at which point the dotted downward sloping line (the part of it below the horizontal axis) represents marginal revenue. If marginal cost cuts the marginal revenue curve somewhere in its vertical segment, the Texas firm will charge the Pittsburgh-plus price. Since marginal cost will equal marginal revenue, the Texas firm maximizes its profits by charging the Pittsburgh-plus price. The Texas firm may be able to make all of the local sales without undercutting the Pittsburgh-plus price, because it has the advantage of faster shipment to nearby customers.

164

Facilitating Mechanisms

P

PS PR PQ PC

Q Figure 8.4

Now consider destinations further from Pittsburgh than Texas. The price at which the Pittsburgh firms will ship to those destinations is given by PQ, PR, and PS. The Texas firm could ship to those destinations at a lower price, but if its marginal cost curve intersects the relevant marginal revenue curves in the vertical portion, it will charge Pittsburgh-plus prices to these other destinations.34 Although basing point pricing at first glance seems inconsistent with the textbook model of competition, it is consistent with competition under the conditions described in Haddock’s article (costly transportation, scale economies in production and transportation). A per-se rule would be inappropriate since the root cause of the phenomenon lies in market structure rather than the firms’ conduct. To the extent that a perse rule requires firms to alter practices that have evolved through competitive efforts, it may actually reduce competition. 34

For a simple illustration, see Christian G. Cabou, David D. Haddock, and Michele H. Thorne, An Analytic History of Delivered Price Litigation: Do Courts Properly Distinguish Rivalrous From Collusive Instances?, 30 Economic Inquiry 307–21 (April 1992).

III. Basing Point Pricing: Economics

165

This is a direct answer to Donald Turner and Richard Posner, who both urged a more aggressive approach in the area of conscious parallelism (Chapter 4). Posner wanted to see the conspiracy proof requirement weakened so that parallel conduct that exhibits stability would merit per-se treatment. Turner preferred to maintain the conspiracy proof requirement, largely because he could not see an alternative under which courts could state clearly the practices that were illegal under the Sherman Act. However, in spite of his relatively conservative approach, Turner saw basing point pricing as an easy case, a clear example of anticompetitive conduct. As Haddock’s analysis of basing point pricing demonstrates, easy antitrust cases may not be so easy after all. They may arise from competitive practices that are not well understood.

9 Boycotts

The question this chapter considers is when a concerted refusal to deal, or boycott, violates Section 1. There are really two questions involved. The first is the threshold question of whether an agreement exists. Without an agreement, there is no conspiracy and no violation of Section 1. If an agreement does exist, then the question turns to whether the agreement unreasonably restrains trade. Several issues examined in the cases below become relevant. Is the challenged practice an agreement to refuse to deal, or merely an agreement to share information on an actor or group of actors? Does rule of reason or per-se analysis apply? What does it mean to apply per se analysis? Would the per-se rule apply only after proof of an agreement to refuse to deal, or would it apply after evidence established that an agreement, of any sort (e.g., to share information), resulted in several members of a group refusing to deal with someone else? Boycott doctrine can be divided into three developmental stages. The first is pre-Socony, in which the Court applied the rule of reason. The second is post-Socony, in which the Court shifted toward a per-se rule. The third period is post-BMI/Sylvania, in which the Court adopted a hybrid rule of reason/per-se test. I trace the development of boycott doctrine in each of these periods below.

i. pre-socony The first Supreme Court case to deal with the questions set out above, Eastern States,1 involved a group of retail lumber dealers that boycotted 1

Eastern States Retail Lumber Dealers’ Assn v. United States, 234 U.S. 600 (1914).

166

II. Post-Socony

167

wholesalers who sold directly to consumers. The Court upheld a lower court finding of a conspiracy in violation of Section 1. As is true of ordinary price-fixing cases, the government must prove conspiracy when charging the defendants with boycotting in violation of Section 1. For boycott doctrine, the interesting part of Eastern States is the Court’s treatment of the retailers’ argument that direct-selling wholesalers infringed on their “exclusive right to trade.”2 The Court said that every retailer has the right to stop trading with any wholesaler, for any reason; but when they combine to do this, they conspire in a manner that hurts the public. Conspiring in this manner goes beyond the lawful, personal right to trade with whomever one wishes. Eastern States implies that a Section 1 boycott case requires proof of conspiracy. Why? Because every trader has an unconditional personal right to deal with whomever he wishes. It follows that if some supplier in a competitive market refuses to deal with you because he aims to drive you out of business in order to enhance his market power, you have no cause of action under the Sherman Act. The result might be different with a monopolistic supplier,3 as we will see in Chapter 10. Another important early case is Cement Manufacturers Protective Assn. v. United States.4 The facts at the core resulted from the “free option” arrangement between the contractors and the cement manufacturers. The arrangement fixed in advance the price of an order of cement for future delivery. The contractors could cancel an order if the spotmarket price dropped, thus allowing contractors to lower their costs on a construction project. Contractors took advantage of this plan in this way, but they discovered that they could profit in another way: by ordering cement from several manufacturers and demanding delivery from all suppliers if the spot price rose above the contract price. In response, the Cement Manufacturers Protective Association circulated among its members information on specific job contracts in order to monitor the contractors. The Court upheld this boycott as a legitimate approach to preventing abusive business conduct. The Court stressed the fact that the Cement Manufacturers Protective Association did not require members to refuse to supply cement to contractors who had ordered more than the necessary amount. Members remained free to meet the demands of such 2 3 4

Id. at 611. See United States v. Colgate & Co., 250 U.S. 300 (1919). 268 U.S. 588 (1925).

168

Boycotts

contractors or to refuse them, even though there would clearly be a strong incentive to refuse. The Court also relied on its conclusion that the contractors behaved fraudulently. Eastern States and Cement Manufacturers reveal two important strands of early boycott doctrine. First, a finding of conspiracy to boycott is for obvious reasons less likely if the defendant group leaves its members free to decide whether to deal with others. Second, the purpose of the boycott is important. The Court did not take the time to develop a theory of abusive practices, but the distinction is crucial to understanding the different outcomes in Eastern States and Cement Manufacturers. Absent the abusive practices matter, the cases are very much alike. In both, group members circulated a list of bad dealers, and defendants stopped dealing with them. There was no concrete evidence of agreement in either case. Cement Manufacturers also extends the Eastern States implication that conspiracy must be demonstrated. To say that the plaintiff must prove conspiracy raises a new set of questions. What sort of agreement must the plaintiff demonstrate? Is it enough to show an agreement to circulate a list of boycott targets? Cement Manufacturers holds, in effect, that the plaintiff must demonstrate that an agreement to boycott exists. That is the difference between the conclusions in Eastern States and in Cement Manufacturers. In Eastern States, the Court found an agreement to boycott. In Cement Manufacturers, the Court found agreement to circulate a list, but not agreement to boycott. Two early cases involving film distributors, Paramount Famous Lasky Corp. v. United States,5 and United States v. First National Pictures,6 attracted attention for suggesting a shift toward per-se analysis. Both involved agreements among film distributors that specified certain terms on which distributors would contract with exhibitors. In Paramount, the distributors required arbitration of disputes, and agreed to boycott any exhibitor who failed to comply with the requirement. In First National, the distributors required exhibitors to post cash deposits. The Court found both agreements unlawful because they restricted the freedom of distributors to compete. The Court said that the purposes of the boycotts were irrelevant. One could argue that the distributors in both cases were trying to prevent abusive practices, just like the defendants in Cement Manufac5 6

282 U.S. 30 (1930). 282 U.S. 44 (1930).

II. Post-Socony

169

turers. However, the motion picture cases can be distinguished from Cement Manufacturers on two grounds. First, there was evidence of agreement in the motion picture cases, and second, the need for collective action is not nearly as clear in the motion picture cases as in Cement Manufacturers. The two aspects are linked in their implications for a finding of conspiracy. Each film distributor could have decided on its own not to deal with an exhibitor who refused to submit disputes to arbitration or to post a bond to guarantee his offer. But in Cement Manufacturers, in order to have information on offending dealers, the defendants had to join forces in distributing information. Furthermore, the suggestion of an anticompetitive conspiracy is bolstered, in the motion picture cases, by the evidence of an agreement to boycott. If the refusals to deal were independent, the distributors would not need to agree on a boycott. However, with an anticompetitive conspiracy, agreement to boycott is necessary because each firm would otherwise have a strong incentive to deal with the boycott target. Thus, in spite of the difficulty of squaring the motion picture cases with earlier boycott decisions, one can argue that all of these cases are consistent with rule of reason analysis. The cases show that inference of an illegal conspiracy to boycott requires evidence on the purpose of the agreement and the intent of the parties. A rational inference rule would find an unlawful conspiracy more often when the evidence suggests that the boycott supports a price-fixing cartel. Thus, one can envision a twodimensional sliding scale in which the probability of an unlawful conspiracy increases as the evidence on purpose points more clearly toward restraining competition, and as the evidence on agreement points more clearly toward a concerted refusal to deal. Applying this approach, Cement Manufacturers is a case with weak evidence of an agreement to boycott, and the evidence on purpose suggests that the conduct was not part of a price-fixing plan. Eastern States, on the other hand, also lacked strong evidence of conspiracy, but the evidence on purpose suggests that the conduct was part of a price-fixing plan. Although the Court declared that purpose was irrelevant in the motion picture cases, the facts do not suggest that the refusals to deal in these cases were benign. This, coupled with concrete evidence of an agreement to boycott, makes the inference of unlawful conspiracy plausible in the motion picture cases. To sum up, the pre-Socony boycott cases are consistent with rule of reason analysis. That analysis requires evidence of (1) an agreement to boycott and (2) an anticompetitive purpose. Per-se analysis requires

170

Boycotts

evidence on only the first issue – whether there was an agreement to boycott. Although the Court never quite articulated the standard applied in the early cases, there is explicit consideration of purpose in some of them, and examination of purpose helps reconcile their conclusions.

ii. post-socony The crystallization of boycott doctrine began with Fashion Originators’ Guild of America v. FTC (FOGA),7 where the Court signaled a shift toward per-se analysis. It involved a boycott that aimed at eliminating copying competitors and promoting compliance with state unfair competition laws. Manufacturers of original dress designs sought to stop “style piracy” by refusing to sell to retailers who also sold garments copied from Guild members’ dresses. In 1936, Guild members sold more than 60 percent of all women’s garments wholesaling at $10.75 and above. The issue was whether FOGA’s boycott violated Section 5 of the FTC Act and Section 1 of the Sherman Act. The Court held that the boycott violated both statutes. The existence of a conspiracy to boycott was not an issue in FOGA. The defendants admitted that they had combined to suppress competition from style pirates. The circumstantial evidence of conspiracy was also strong. The Guild circulated a list of complying and noncomplying retailers, with white cards for the former and red cards for the latter. Thus, the knowledge basis necessary to carry out a conspiracy was present. Second, there was a monitoring and enforcement mechanism at hand. The Guild monitored the sales of members and imposed fines on those who sold dresses to offending retailers. Given the simplicity of the conspiracy issue, the bulk of the Court’s opinion in FOGA should be read as a discussion on the importance of purpose as an element of inquiry in boycott cases. The Court gave two reasons for holding that the defendants’ boycott violated the FTC and Sherman Acts. First, the boycott restricted output and suppressed competition. Second, the Court condemned the association’s efforts as those of an “extragovernmental agency” regulating trade. Much of the opinion concerns the defendants’ attempts to prove reasonableness. The Court responded to the defendants’ argument that they did not have monopoly power by stating that it was sufficient that the defen7

312 U.S. 457 (1941).

II. Post-Socony

171

dants’ activities tended toward monopolizing the market. Furthermore, the Court said, a finding of monopoly power was not necessary to uphold the FTC’s decision (the FTC had found the defendants guilty of violating Section 5 of the FTC Act) because the Commission had Congressional authorization to stop monopolization efforts in an early stage. The Court also responded to the defendants’ claim that the government showed neither a reduction in output nor price fixing. The Court said that the antitrust laws were concerned with other harms, as well as those listed by the defendants, which necessarily tend toward monopolization. The Court offered exclusion of competitors as one example. As a second example, the Court provided a cite from the Trans-Missouri opinion in which Justice Peckham stated that a combination that fixes prices in order to lower them violates the antitrust laws because it drives out of the market “worthy men” who have devoted their lives to a given trade. Lastly, the Court noted, again, that the FTC had the authority to stop monopolization in its incipiency. The defendants had attempted to prove before the FTC that their combination was reasonable – that it increased social welfare by restraining an abusive practice – but the Commission declined to hear the evidence. The Court held that the Commission did not err by declining to hear such evidence, “for the reasonableness of the methods pursued by the combination . . . is no more material than would be the reasonableness of the prices fixed by unlawful combination.”8 FOGA implies that purpose is irrelevant, given proof of a conspiracy to boycott. The Court relied largely on the reasoning of the price-fixing cases. The boycott restrained competition and thus violated the policy of the Sherman Act. Since the Court relied on the reasoning of the price-fixing cases, one should assume that the underlying theoretical argument is that of Trenton Potteries; that the administrative costs of distinguishing good from bad boycotts would outweigh the benefits. Put another way, the additional administrative costs under a rule of reason analysis outweigh the “false conviction” costs associated with the per-se rule. I noted another theory the Court offered as justification: the notion that the Guild had formed an extragovernmental trade regulation agency. The Court suggested that boycotts are per-se unlawful because they are a form of private justice. Although the Court did not explicitly overrule Cement Manufacturers, the holding and especially the reference 8

FOGA, 312 U.S. at 468.

172

Boycotts

to extragovernmental agencies are clearly inconsistent with Cement Manufacturers. Why might a group such as FOGA do what it did, and could one argue that it was economically reasonable? Think about the patent and copyright laws. They reduce output and price competition. However, the reason for these laws is that they preserve the incentive to innovate. If the patent and copyright laws enhance consumer welfare, then it is possible that FOGA’s activities did so too. One could argue that the defendants should have enforced their claims in each state under the unfair competition laws. However, the costs of suing individual retailers in as many as thirty states would have been prohibitive. FOGA happens to be one of the cases in which we observe a conflict between the consumer welfare standard and the goal of maximizing competition. The Court’s failure to deal directly with this problem can be taken as evidence of its unwillingness, in this period, to step outside of the narrowest interpretation of the boundaries of Chicago Board of Trade. This approach was later modified by the Sylvania and BMI decisions (see Chapters 6 and 13). One could defend the practices of the trade association as reasonable on the grounds provided in BMI, and, as we will see later in this chapter, modern boycott doctrine permits defendants to make such arguments.9 FOGA effectively closed the door to defendants who would argue that their boycott should be held reasonable because of some inherent conflict between unrestrained competition and consumer welfare. Although the Court announced a per-se rule, one could read it more narrowly as simply shutting off one type of reasonableness defense, one that Chicago Board of Trade had already shut off. Thus, even though FOGA clearly states a per-se rule, the decision can be squared with the rule of reason, and indeed one could argue that the reasonableness standard as then interpreted required the outcome in FOGA.

A. No Effect on Competition The argument that a court should hold a boycott reasonable because of some inherent conflict between competition and consumer welfare is 9

See the discussion of post-BMI boycott doctrine in this chapter. Commentators have suggested that current law does not provide a large enough shelter for firms that invest in new processes or products that can be imitated easily, see Thomas M. Jorde and David J. Teece, Rule of Reason Analysis of Horizontal Arrangements: Agreements Designed to Advance Innovation and Commercialize Technology, 61 Antitrust Law Journal 880 (1993) (arguing for an “innovation sensitive” rule of reason analysis in the boycott area).

II. Post-Socony

173

simply one of several reasonableness defenses available to defendants. An alternative reasonableness defense is that the boycott had no effect whatsoever on competition. The Court rejected that defense in Klor’s v. Broadway-Hale Stores,10 a decision that also articulated a per-se rule. The Klor’s corporation operated a retail store in San Francisco. BroadwayHale ran a chain of department stores, with one of them located next door to Klor’s. The stores competed in the sale of household appliances. Klor’s charged that Broadway-Hale and ten national manufacturers and their distributors agreed among themselves not to sell to Klor’s or to sell only at discriminatory prices. The defendants did not deny this, so the validity of the plaintiff’s claim was not at issue. The trial court and Ninth Circuit ruled in favor of the defendants’ motion for summary judgment on the theory that there was no public harm because the actions had not affected market price and output in a manner that would hurt consumers. The issue before the Court was whether (a) there was public harm and (b) did the defendants’ conduct violate the Sherman Act? The Court held that there was harm to the public, and that the defendants’ activities violated the Sherman Act. The interesting aspect of this case was the defendants’ argument that their activities did not violate the Sherman Act because the statute protects competition, not competitors. The defendants had submitted affidavits demonstrating that there were many other sellers of household appliances competing with Klor’s and Broadway-Hale in the same geographic market. Thus, closing down Klor’s could not have hurt competition. The Court held that boycotts such as this fall in the per-se category, and thus there is no need to prove public harm, which is assumed to exist because of the nature of the restraint. The Court also suggested that accepting the defendants’ argument would lead to a big reduction in enforcement efforts. A large number of antitrust disputes involve a single competitor as the primary victim. If the antitrust laws denied standing to any competitor excluded from the market as the result of a conspiracy among other firms, then there would be nothing in the law to prevent the conspirators from monopolizing the market by eliminating rivals one by one.11 This is apparently a per-se holding. However, Klor’s can be reconciled with the rule of reason, as it is applied today, because the defendants

10 11

359 U.S. 207 (1959). The Court cited International Salt Co. v. United States, 332 U.S. 392, 396 (1947), for the proposition that the Act outlaws restraints that tend to result in monopoly whether the tendency is “a creeping one” or “one that proceeds at full gallop.”

174

Boycotts

offered no procompetitive justification. Under the rule of reason, as interpreted in Chicago Board of Trade, courts cannot uphold a restraint of trade that has no procompetitive justification. As we noted in Chapter 6, courts have generally put the burden on the defendant to provide a procompetitive justification for an arrangement that restrains trade.

B. The Klor’s Paradox and Boycott Doctrine The problem at the heart of Klor’s is an interesting one that resurfaces occasionally in antitrust cases.12 An exclusionary plan implemented in an unconcentrated market will not have a significant effect on market price or output. It follows that the plan causes no significant antitrust injury. But this implies that in a market of one thousand suppliers, the perpetrators of the exclusionary plan can eliminate as many as eight hundred competitors without the threat of an antitrust suit. Unable to prove significant public harm, none of the eliminated rivals would have standing under the Sherman Act. The problem can be clarified with an example. Return to the Cournot competition model introduced in Chapter 1 and discussed in Chapter 4. Instead of two firms, suppose there are N firms choosing their output levels (under the assumption that their competitors will not alter their output in response). Let the market demand schedule be P = 4 -Q and let the total cost schedule (describing cost as function of output) for each firm be Ci = qi , i = 1, . . . , N . Note that the market quantity Q is the sum of the output levels of all of the firms – that is, Q = q1 + . . . + qN . The equilibrium price and output levels are13 P = (4 + N ) (N + 1) Q = 3N (N + 1) . 12

13

The most recent case in which the basic issue in Klor’s was revisited was Summit Health v. Pinhas, 500 U.S. 322 (1991). To see how market price and output are derived under the Cournot competition model with N firms, see Michael D. Intriligator, Mathematical Optimization and Economic Theory 207–9 (1971).

II. Post-Socony

175

Now let us consider the effects of a boycott that reduces the number of firms in the market. If initially there are one thousand firms in the market, the market price will be P0 = 1.003 and the market quantity will be Q0 = 2.997. If the boycotting group manages to eliminate two hundred firms, the market price rises to P1 = 1.004, and the market quantity falls to Q1 = 2.996. If the boycotting group manages to eliminate eight hundred firms, the market price rises to P2 = 1.015, and the market quantity falls to Q2 = 2.985. It follows that if the boycotting group eliminates as many eight hundred competitors, the market price increases by P2 P0 = .012, and the market quantity falls Q0 - Q2 = .012. In percentage terms, the market price rises by 1 percent, the market quantity falls by .4 percent. The total change in welfare, equivalently the loss in the consumer’s surplus implied by these changes, is .00007, or roughly one hundredth of a penny. No doubt the total welfare loss would be larger if the boycotting group could eliminate 900 or 990 of their 1000 initial competitors. However, the point of the example should be clear. Given that we start with a competitive market, it takes a long time before the boycotting group’s efforts have any significant impact on market price, total output, or consumer welfare. Let us label this problem the Klor’s paradox. A more general statement of the problem is as follows. Suppose there are N firms in a competitive widget market at this moment; and that in each period the number of new firms entering the industry is equal to the fraction s times N, the number of firms exiting the market is equal to the fraction t times N, and s > t, so that the industry is growing. Let a cartel of incumbent firms adopt a policy of eliminating a number of firms each period equal to the fraction q times N. As long as s - t ≥ q, the industry will not decline in size even though the rate of exit is artificially high due to the cartel’s boycott policy. Because the industry will not decline, the market will remain competitive in the long run. However, the number of firms eliminated by anticompetitive conduct approaches infinity in the long run also. Should the antitrust authorities stay on the sidelines? There are several ways to answer this problem. The first is to point out that even though the reduction in consumer welfare is not substantial, it is still a reduction. Even in the numerical example discussed above, the welfare loss of one hundredth of a penny is still an injury to consumers. Given this, the defendants’ conduct is socially harmful and should be prohibited. The antitrust laws effectively prohibit the conduct by authorizing injured parties to bring suit against the defendants.

176

Boycotts

A second argument for enforcement is that a nonenforcement policy could be a mistake. The cartel’s boycott may eliminate so many competitors that the industry turns into an oligopoly or a monopoly. This reasoning suggests that it would be better to enforce the law now rather than wait for the day when the market is no longer competitive. However, this argument is far from persuasive as a general matter. The problem is that it supports expending resources now to avoid an uncertain harm in the future. An alternative approach to the Klor’s paradox starts with the realization that the antitrust defendants’ argument is somewhat like a tort defendant arguing that he should not have to pay damages because a judgment against him will not affect the aggregate level of precaution. But the only way to regulate aggregate precaution is to enforce the law against every individual. Of course, there are important differences between the tort defendant and the antitrust defendant. The antitrust laws inquire into the anticompetitive intent of the defendant, while tort law is largely unconcerned with the defendant’s motives. A requirement that the antitrust plaintiff prove public harm serves a useful purpose in a regime in which motive is a relevant issue, because a finding that there was no public harm suggests that the defendant did not intend, or rationally could not have intended, to monopolize the market. Thus, if the defendants really do have a good efficiency defense for their boycott, but are unable to persuade the Court of its validity, the public harm requirement would serve as an added layer of protection for their activity. In sum, the only good argument that can be offered for a public harm requirement is based on legal error. Courts may not do a good job of distinguishing efficient from harmful boycotts. If so, a public harm requirement may be necessary in order to cut down the risk that efficient boycotts will be punished. However, as we will see later in this chapter, this argument has considerably less force today than it did at the time of the Klor’s decision. Boycott doctrine has come around to recognizing efficiency defenses. Given this change in the law, the case for a public harm requirement in boycott doctrine looks considerably weaker. Today the Klor’s paradox remains largely unresolved by the Supreme Court. Is proof of public harm a general requirement for a plaintiff to obtain standing under the antitrust laws? Or is it required of the plaintiff only in certain areas of antitrust law, only with respect to certain issues? The Court has not provided clear answers to these questions.

II. Post-Socony

177

The Court revisited the Klor’s paradox in Summit Health v. Pinhas.14 There, an eye surgeon was excluded from part of the Los Angeles eye surgery market by a conspiracy among his colleagues to deny him staff privileges at a hospital with which the doctors were affiliated. The defendants argued that the plaintiffs had not satisfied the interstate commerce requirement of the Sherman Act because the greater Los Angeles-area market for eye surgery could not have been materially affected by one exclusion. Four Supreme Court justices bought this argument, without discussing Klor’s in their dissent. However, the Klor’s paradox was at the heart of the dispute.

C. Associated Press FOGA and Klor’s both announced a per-se rule, but the decisions can be reconciled with the Chicago Board of Trade’s narrow reading of the boundaries of the rule of reason inquiry. Because of this, the decision that most clearly indicates the acceptance of a per-se rule is Associated Press v. United States.15 Associated Press (AP) is a cooperative, incorporated, nonprofit association that at the time of the government’s suit distributed news gathered by its employees, member newspapers, and foreign news agencies, to its twelve hundred newspaper members. The suit concerned the Association’s bylaws. The government sought to enjoin the practices required by AP’s bylaws, particularly the first, which allowed member newspapers to veto the membership application of a local rival. The district court granted the government’s motion for summary judgment and enjoined the practice required by the first bylaw. The district court enjoined the requirements of the second and third bylaws pending the abandonment of the membership restriction. The issue before the Supreme Court was whether the practices required by the AP bylaws violated the Sherman Act, and the Court held that they did. The Court’s opinion does not explain clearly whether it is a per-se or rule of reason holding. However, some conclusory statements early in the opinion indicate that it is a per-se holding. The Court noted that “the court below found, and we think correctly, that the By-Laws on their face, and without regard to past effect, constitute restraints of trade.”16 The 14 15 16

500 U.S. 322 (1991). 326 U.S. 1 (1945). Id. at 12.

178

Boycotts

Court also stated that it was irrelevant that AP had not yet achieved monopoly power because the agreement restrained competition, and that was sufficient to find a violation. The Court discussed each of the three prongs (purpose, power, and effects) of the reasonableness inquiry. Let us start with effects. The Court recited certain facts as evidence of anticompetitive harms. For example, in twenty-six cities, existing newspapers had contracts with AP, UP, and INS; and in these cities it was hard for a competitor to enter, because the competitor would “be required to pay the contract holders large sums to enter the field.”17 It is entirely fair to question whether these facts prove the existence of an anticompetitive effect. First, new entrants may not have needed access to AP, UP, or INS in order to compete. Second, how large were these twenty-six cities? Did they make up a small portion of the U.S. market? Could each city support more than one newspaper? The Court was equally unpersuasive in its analysis of purpose. The Court noted that this sort of combination appeared on a smaller scale within each newspaper; each reporter promises to provide news exclusively to the newspaper for which he or she works. The Court recognized that destroying this exclusivity relationship would destroy newspapers as independent brands.18 But the Court asserted that the AP combination was different: it was a large group that had combined for the purpose of restraining competition. The Court never explained how it reached its conclusion as to purpose, or the relevance of the size of the network. If the restrictions on news sharing were necessary for the survival of AP, then the Court’s decision suggests that large news networks violate the antitrust laws by their mere existence. What about the individual freedom to trade with whomever one wishes (remember Eastern States)? The agreement did not require local newspaper members to veto the application of a new entrant who wished to become a member of the network; the bylaws merely permitted the incumbent member to cast a veto. This is certainly consistent with individual freedom: it allowed each individual member to refuse to supply its services to a local rival. The Court’s analysis in Eastern States and Cement Manufacturers would seem to support this individual right to refuse to deal. 17 18

Id. at 13. This is a restriction on intrabrand competition in order to enhance intrabrand competition. I should note that the AP exclusivity agreement was challenged in a pre-Sherman Act restraint of trade case, and the exclusivity agreement was enforced. See Matthews v. Associated Press of New York, 136 N.Y. 333 (1893).

II. Post-Socony

179

In response to the individual freedom concern, the Court said that this combination went beyond the exercise of each individual member’s right to trade. As a combination it put competitors at a disadvantage, and then denied them the opportunity to correct or lessen that disadvantage. The Court noted that the members “surrendered” their freedom in order to comply with the restrictive bylaws.19 Of course, the only real restriction in the bylaws affecting AP members was the requirement that they supply news exclusively to the AP network. To describe this as surrendering freedom is to treat a contractual arrangement as coercive. But the subtle reference to coercion served its purpose in advancing the Court’s argument that individual members did not exercise their freedom in this plan but, rather, gave up their freedom in order to join a cartel that monopolized the market. The Court’s cursory analysis of market power was consistent with per-se analysis. The Court held that evidence that AP had not achieved a complete monopoly was irrelevant. The defendants argued that significant competition remained outside of the network; there were other news networks, and some newspapers survived without access to AP, even without access to any of the networks. The Court rejected this on the ground that complete absence of competition is not a prerequisite under Section 1.20 In difficult-to-follow reasoning, the Court suggested that the AP network inhibited competition by allowing members to gain a competitive advantage and drive out their rivals – indeed this was, according to the Court, the purpose of the plan. In any event, Socony held that complete absence of competition is not a necessary finding, so here the Court merely repeated the language of the price-fixing cases. The defendants also argued that they did not have market power, because access to the network was not indispensable. A newspaper could compete without access. It could read the news from the AP affiliate and print it the next day, with new facts gathered on its own. The Court held that the Sherman Act does not require indispensability, because to do so would make the statute a “dead letter.”21 Any remaining doubt as to whether this was a per-se holding was erased in a passage in which the Court noted that in this case, as in FOGA, the combination was in reality an extragovernmental agency set

19 20

21

Associated Press, 326 U.S. at 19. The Court noted, however, that 96 percent of morning newspapers had access to AP, 326 U.S. at 18, which presumably indicated full monopolization in one market. Id. at 18

180

Boycotts

up to regulate trade.22 The Court described the network as an artificial creation designed to force nonmembers out of business by a giving a decisive advantage to network members. The strongest doctrinal support for the Court’s theory came in Justice Douglas’s concurring opinion. He drew an analogy to US v. Terminal Railroad,23 a case in which Jay Gould and several other railroad owners had acquired all of the terminal facilities for railroads crossing the Mississippi River into St. Louis. The Court required the defendants to provide fair access to competitors. Douglas suggested that the proper solution in these cases may be to require the facility owner to provide fair access to competitors. Consider some of the weaknesses in the Court’s reasoning, coupled with Douglas’s argument. Let us start with the analogy to Terminal Railroad. Though persuasive initially, it is a poor analogy. The news network was a technological development that lowered the cost of providing a given quality level of reporting. It is in this sense that it provided its members a competitive advantage, as would any cost-reducing technological innovation.24 The facilities acquired in Terminal Railroad did not represent the development of a cost-reducing facility because they existed before the acquisition. The acquisition gave the proprietary consortium of railroads the power to charge discriminatory rates that could put competitors at a disadvantage.25 BMI and Sylvania (Chapter 6 and 13) can be used to make a good argument in favor of the bylaws requiring exclusive provision of news and permitting each member to veto the membership application of a local rival. The exclusivity requirement restricted intrabrand competition in order to enhance interbrand competition. The veto rule protected incentives to gather news at the local level. Imagine yourself as the owner of a small newspaper, and suppose a new local rival gains access to the network. The rival would not have to collect any news: It could let your paper do all of the work, and free-ride off the network.

22 23 24

25

Id. at 19. 224 U.S. 383 (1912). Suppose AP controlled 96 percent of the market. So what? Does this imply that there is no competition, that AP has monopoly power? No, it is perfectly consistent with AP being a more efficient producer than the rest. This is a special case of the more general “raising rivals’ costs” strategy emphasized in Thomas G. Krattenmaker and Steven C. Salop, Anticompetitive Exclusion: Raising Rivals’ Costs to Achieve Power over Price, 96 Yale L. J. 209 (1986).

III. Post-BMI/Sylvania

181

With this clarification of the purpose of the veto rule in mind, the Cement Manufacturers argument in favor of the defendants becomes clear. In the absence of a veto rule, the regional newspapers would face a type of abuse. Rivals would have an incentive to enter in order to skim the cream. The veto rule gave each individual member the freedom to choose whether to permit this to happen, and thus allowed each member to protect himself from a type of abuse, as did the arrangement upheld in Cement Manufacturers.

iii. post-bmi/sylvania In an abrupt shift from the rule of Associated Press, the Court reconciled its boycott decisions under a hybrid rule of reason/per-se test in Northwest Wholesale Stationers v. Pacific Stationery & Printing Co.26 Northwest Wholesale Stationers was a purchasing cooperative made up of about one hundred office supply retailers. The cooperative could purchase goods in large quantities, lowering the cost to the retailers. As the result of its rebate policy, members effectively paid a lower price for supplies than did nonmembers. Members also benefitted from the warehousing facilities the cooperative provided. Pacific Stationery was expelled after violating the bylaws of the cooperative. The Ninth Circuit held that the expulsion was within the per-se category because Northwest did not provide adequate procedural safeguards. The argument was drawn from Silver v. New York Stock Exchange,27 where the Court held that NYSE had violated the Sherman Act because it excluded a broker from access to its facilities without first providing a hearing. Silver is best described as a per-se opinion because the Court reached this conclusion despite the absence of evidence that the plaintiff-broker’s exclusion was designed to hurt competition. Of course, the Court actually narrowed the per-se rule in order to accommodate the Sherman Act with the federal policy favoring self-regulation in the securities industry. The doctrine of FOGA would have led to a perse finding whether or not the exchange provided procedural safeguards. In Northwest Wholesale Stationers, the issue was whether Northwest’s decision to expel Pacific should fall within the per-se category. This was broken into two questions: (1) whether the exclusion fell within the per-se category because Northwest provided no procedural protections 26 27

472 U.S. 284 (1985). 373 U.S. 341 (1963).

182

Boycotts

to Pacific, and (2) whether the mere expulsion of Pacific should fall within the per-se category. The Court said no to both questions. In justifying its answer to the first question, the Court distinguished Silver from the ordinary boycott case involving membership in a consortium. The national policy favoring self-regulation of the securities exchanges led the Silver Court to take a narrow approach toward application of the Sherman Act. Here there was no such policy. The absence of procedural safeguards is therefore irrelevant. With respect to the second question, the Court announced a new rule: Unless the cooperative possesses market power or exclusive access to an element essential to effective competition, the conclusion that expulsion merits application of the per-se standard is not warranted. This implies that the rule of reason is the default standard courts should apply in boycott cases. The Court relied on NCAA and BMI for this doctrine. Rules are often necessary in cooperative ventures; they often make the provision of new products possible. Rules must be enforced. Therefore, those who break them must be disciplined. Further, the Court’s reliance on BMI suggests that even if a cooperative possesses market power or exclusive access to an essential facility, rule of reason analysis may be appropriate where there are clear efficiency gains that result from the cooperative’s exclusionary policies. The Court suggested that market power or the possession of an essential facility are necessary conditions for application of the per-se rule, not sufficient conditions. The Court asserted that all of the previous boycott decisions are consistent with this proposition. Obviously, the claim that a new decision can be squared with previous ones provides some sense of continuity to the law. But the assertion is difficult to accept in view of Associated Press and FOGA, both cases where there were plausible efficiency gains resulting from the defendants’ exclusionary practices. In addition, the assertion invites litigants and future courts to use the old decisions as guides in defining the content of the new rule, a practice that can generate confusion when the old decisions are inconsistent with the new rule. The Court should have admitted that the rule of Northwest Wholesale Stationers would generate a different conclusion when applied to the facts of some of the earlier cases, particularly Associated Press. The Court applied the rule of Northwest Wholesale Stationers in Federal Trade Commission v. Indiana Federation of Dentists.28 Dental health 28

476 U.S. 447 (1986).

III. Post-BMI/Sylvania

183

insurers had attempted to control the costs of dental treatment by requesting dentists to submit x-rays and other information along with requests for payment. In 1976, a group of Indiana dentists formed the Indiana Federation of Dentists in order to pursue a policy of resisting insurers’ requests for x-rays. The FTC ruled that this violated Section 5 of the FTC Act. The Seventh Circuit held that the evidence did not support this conclusion. The issue was whether the findings of the FTC were sufficient to establish a violation of Section 1 of the Sherman Act.29 The Court held that they were. The Court’s rationale was largely an application of Northwest Wholesale Stationers. The Court first held that the rule of reason applied because the facts did not fall in the category of cases described in Northwest Wholesale Stationers which merit application of the per-se rule. The Court described that category as consisting of cases in which the boycotting group uses its market power to discourage a supplier or customer from doing business with a competitor (as in Klor’s and FOGA).30 The second reason for applying the reasonableness test was that this was a professional group, and therefore the defendants had a greater claim to rule of reason treatment.31 The rest of the Court’s opinion applies: Professional Engineers and Chicago Board of Trade (or NCAA) to the defendants’ arguments. The Court relied on Chicago Board of Trade to respond to the defendants’ claim that the FTC had not shown that they possessed market power or that their plan reduced consumer welfare. Neither of these defenses is permissible under the Sherman Act because Chicago Board of Trade places the burden on the defendant to come forth with a procompetitive justification for trade restraints. Professional Engineers was applied to the defendants’ argument that their plan was justifiable because it aimed to protect the quality of dental 29

30

31

Before going into the Court’s rationale, we should pause to note that we began with an FTC case and wound up with a Sherman Act case. How did that happen? This case reveals one of the enforcement techniques available to the FTC. The FTC does not have authority to enforce the Sherman Act; its authority is limited to the FTC Act and the Clayton Act. However, if the defendant appeals, the FTC may raise a Sherman Act charge, once the case finds its way into the federal courts. Thus, a defendant initially charged with a violation of Section 5 of the FTC Act, a statute with rather modest penalties, may find itself defending a Sherman Act charge in the court of appeals. Note that distributors in the motion picture cases would not be subject to the per-se rule under this description. See Professional Engineers (discussed in Chapter 6) and Goldfarb v. Virginia State Bar, 421 U.S. 773 (1975).

184

Boycotts

services. Professional Engineers rejected the claim that the intention to enhance product quality could serve as a justification for restraints on competition. How should we describe current boycott doctrine after Indiana Federation of Dentists? Proof of a violation of Section 1 requires evidence (1) that the parties agreed not to deal with another party, and (2) that the agreement unreasonably restrained trade. The first prong of the analysis has remained intact since the pre-Socony period. The cases require evidence of an agreement to boycott – evidence merely of an agreement to share information is not sufficient to prove a conspiracy to boycott. The second component of the proof standard has been radically altered in its different developmental phases. Today, the second component requires application of reasonableness analysis generally, and per-se analysis when the consortium has market power or access to some essential facility. Rule of reason analysis, however, respects the boundaries set by Chicago Board of Trade. Arguments suggesting an inherent conflict between competition and consumer welfare are generally impermissible. However, as long as the defendant does not claim that such an inherent conflict exists, the Court will show more sensitivity to the defendant’s economic reasonableness arguments, under BMI/Sylvania doctrine, than it did in Associated Press.

iv. conclusion This chapter has traced the development of boycott theory and doctrine. The core theoretical problem of boycott law was exposed by the Klor’s case. If a market is competitive initially, a boycott that eliminates only one firm will not have a significant effect on consumer welfare. The Court addressed this problem directly in Klor’s, where it held that proof of public harm is not a requirement under boycott law, and indirectly in Summit Health, where it held that the Sherman Act’s interstate commerce requirement was satisfied even though the defendants had eliminated only one eye surgeon in the Los Angeles market. The development of the doctrine, like that of price-fixing law, reveals the tension between the tendency of courts to gravitate toward economic reasonableness tests and the administrative concerns of enforcement agencies. The Supreme Court adopted a reasonableness standard in the early boycott cases under the Sherman Act. The Federal Trade Commission was successful in getting the Court to adopt a per-se standard in FOGA. However, the per-se standard proved difficult to maintain after

IV. Conclusion

185

the Court decided in BMI and in Sylvania to accept some efficiency defenses for collusive agreements. In Northwest Wholesale Stationers the Court abandoned the per-se test in favor of a reasonableness standard. The Court’s treatment of boycott doctrine reveals a pattern similar to that of the price-fixing cases: a rejection of the reasonableness test at the urging of enforcement agents, eventually leading to a point at which application of the per-se standard to new cases could not be defended, after which the Court begins a return to the reasonableness test.

10 Monopolization

The preceding chapters have all dealt with concerted action. In this chapter, I shift the focus to unilateral conduct, specifically the offense of monopolization, governed by Section 2 of the Sherman Act. This chapter describes the development of Section 2 doctrine and examines modern theories of monopolization.

i. development of section 2 doctrine Standard Oil Co. v. United States1 is cited largely for its statement of the rule of reason as the test that courts should apply in Sherman Act cases. Standard Oil of New Jersey controlled 80 percent or more of all refining capacity in the United States during the thirty years prior to its dissolution in 1911. It has been said that the company achieved its position through oppressive tactics including bribery of public officials, exclusive dealing, harassing lawsuits, and price cutting in order to eliminate small rivals.2 The circuit court ordered the company to rearrange itself into thirty-three independent companies, and the Supreme Court upheld the decree with only minor changes.

1 2

221 U.S. 1 (1911). See Ida Tarbell, The History of the Standard Oil Company (New York: Peter Smith, 1950). For arguments in defense of Standard Oil, see Dominick T. Armentano, Antitrust and Monopoly: Anatomy of a Policy Failure 59–70 (Holmes & Meier, 2d ed. 1990) (Standard Oil achieved its position through superior foresight, skill, and industry); John S. McGee, Predatory Price Cutting: The Standard Oil (N.J.) Case, 1 Journal of Law & Economics 141–8 (October 1958) (evidence suggests that Standard Oil did not use predatory price cutting to drive out competitors).

186

I. Development of Section 2 Doctrine

187

The issue before the Supreme Court was whether the acts Standard Oil used to gain monopoly power were illegal per se, or illegal only if unreasonable. The Court held that Section 2 condemned monopoly power when it was abused, or used unreasonably, as evidenced by trade practices that would violate Section 1 if carried out by two or more firms in combination.3 I will refer to this below as the abuse theory of monopolization. What is an example of abuse under the Standard Oil test? Suppose a group of firms refuses to deal, under circumstances in which the test of Northwest Wholesale Stationers (Chapter 9) implies that they have violated Section 1. This means that the firms have market power or access to a facility deemed essential for competition, and cannot show that their refusal to deal is efficient or enhances competition – say, by enabling the firms to bring a new product to the market. If, instead of a group, a single firm were involved as the defendant, and the same conditions obtained, then the refusal to deal would be an abuse of monopoly power in violation of Section 2. The abuse standard consists of two components. The first is a conductfocused test that looks largely to Section 1 case law to determine whether the defendant’s actions violate Section 2. The second is an intent-based component under which a court infers intent to monopolize (or specific intent) from the absence of credible efficiency justifications for the defendant’s conduct. The Court in Standard Oil noted that the firm’s conduct gave rise “in the absence of countervailing circumstances . . . to the prima facie presumption of intent and purpose to maintain the dominancy over the oil industry, not as a result of normal methods of industrial development, but by means of combination . . . with the purpose of excluding others from the trade.”4 The next important Section 2 case following Standard Oil was United States v. United States Steel Corp.5 Immediately after its formation, the company controlled 80 to 95 percent of domestic production of some

3

4 5

221 U.S. at 61 (“In other words, having by the 1st section forbidden all means of monopolizing trade, that is, unduly restraining it by means of every contract, combination, etc., the 2d section seeks, if possible, to make the prohibitions of the act all the more complete and perfect by embracing all attempts to reach the end prohibited by the 1st section, that is, restraints of trade, by any attempt to monopolize, or monopolization thereof, even although the acts by which such results are attempted to be brought about or are brought about be not embraced within the general enumeration of the 1st section.”) Id. at 75. 251 U.S. 417 (1920).

188

Monopolization

lines, but its overall share of iron and steel had declined to 50 percent by the time of the government’s suit. U.S. Steel reaffirmed the abuse theory initially set out in Standard Oil. Rejecting the government’s claim that the corporation’s size was “an abhorrence to the law,”6 the Court said that the “law does not make mere size an offense.”7 The Court held that U.S. Steel had not violated Section 2 for the following reasons. (1) The company’s market share and supporting testimony convinced the Court that U.S. Steel no longer had monopoly power.8 (2) U.S. Steel had not adopted the oppressive tactics observed in the other cases in which the courts ordered dissolution.9 The Court concluded that the corporation had grown through economies created by vertical integration and specialization within subsidiary units.10 (3) The finding that U.S. Steel lacked market power was further supported by evidence of pricing agreements between the company and its rivals, agreements the company had abandoned before the government’s suit.11 Commentators have suggested that U.S. Steel rendered Section 2 of the Sherman Act a dead letter over the 1920s, 1930s, and 1940s.12 One could argue that the abuse theory of Standard Oil required a finding of a Section 2 violation, since the company may have acquired and maintained its dominant position in part through the price-fixing conspiracies.13 However, this would be missing one of the essential components of the abuse standard: the inference of intent to monopolize. The Court concluded that U.S. Steel’s conduct did not fall in the same category as Standard Oil’s because it lacked similar instances of predation and oppression of rivals. The reasonable inference, therefore, is that U.S. Steel’s growth came through superior efficiency rather than competitionstifling tactics. 6 7 8 9 10

11 12

13

Id. at 450. Id. at 451. Id. at 442, 444. Id. at 455–6. Id. at 443–4. For an economic analysis of U.S. Steel’s conduct that provides support for the efficiency hypothesis, see Joseph C. Mullin and Wallace P. Mullin, United States Steel’s Acquisition of the Great Northern Ore Properties: Vertical Foreclosure or Efficient Contractual Governance? NBER Working Paper No. 5662, July 1996. 251 U.S. at 445. Phillip Areeda and Donald F. Turner, Antitrust Law: An Analysis of Principles and Their Applications, vol. IV, p. 6 (Boston: Little, Brown, and Company, 1980). Indeed, this was the argument of the dissent, see Standard Oil, 251 U.S. at 460–1. If this is plausible, it follows that the firm had abused its power in precisely the sense defined in Standard Oil. However, the majority took the view that U.S. Steel had no monopoly power to abuse.

I. Development of Section 2 Doctrine

189

A. Alcoa The abuse theory remained the standard under Section 2 until Judge Learned Hand wrote into law a more aggressive enforcement approach in United States v. Aluminum Co. of America.14 Judge Hand made several points that an economic theorist would find interesting today and probably persuasive. For example, in considering the relevant aluminum market, Judge Hand reasoned that the Court should exclude scrap or secondary aluminum as a competing source of supply, because Alcoa took into account the impact of secondary market aluminum on the price of virgin aluminum. Since the secondary aluminum was simply yesterday’s virgin aluminum, and because Alcoa was the only producer of virgin aluminum, Alcoa had an incentive to restrict output even more than the simple monopoly model would imply.15 Alcoa could not directly set the price of secondary market aluminum, because the market set that price. But by forecasting demand, and controlling supply, Alcoa could indirectly control secondary market prices. Thus, virgin and scrap aluminum were as much in competition as sheets of virgin produced by Alcoa in different plants on the same day. Judge Hand also argued that Alcoa expanded capacity in order to deter entry by rivals, a theory that economists have only recently explored.16 However, the end result of Hand’s reasoning suggests limits to the usefulness of economic theory as a source of antitrust doctrine. The more sophisticated theories require courts to make distinctions among business practices that are difficult to make in a consistent manner and with a high degree of accuracy. Error is nothing new to courts, and a realistic view must accept that it will occur. However, recognition of the potential for error requires a consideration of the relative costs of erroneous findings of guilt or liability (“false convictions”) and erroneous findings of innocence or nonliability (“false acquittals”). When the underlying conduct is procompetitive, lowering costs and prices in the long run, the costs of false convictions can be large. Alcoa obtained its dominance (sole domestic producer) in the production of virgin aluminum through a monopoly guaranteed by a 14 15 16

148 F.2d 416 (2d Cir. 1945). For further discussion, see Chapter 11. See A. Michael Spence, Entry, Investment and Oligopolistic Pricing, 8 Bell Journal of Economics 534–44 (1977); Avinash Dixit, The Role of Investment in Entry-Deterrence, 90 Economic Journal 95–106 (March 1980); Drew Fudenberg and Jean Tirole, The Fat-Cat Effect, The Puppy-Dog Ploy, and the Lean and Hungry Look, 74 Am. Econ. Rev. 361–6 (May 1984).

190

Monopolization

patent that expired in 1909. It maintained its position until 1912 through a series of cartel arrangements limiting aluminum imports, and other practices, such as buying up electric power through contracts requiring the source not to provide electricity to any other aluminum producer. Alcoa abandoned these arrangements under a consent decree in 1913. In spite of its abandonment of monopolistic practices, Alcoa remained the sole domestic producer when the government brought suit in 1937. The government charged that Alcoa’s position violated Section 2. Judge Hand held that Alcoa had violated Section 2: (1) its size prevented competition from taking place, and (2) it did not “passively acquire” its position. Specifically, Alcoa kept doubling its capacity before others entered the field, blocking entry. The core of Hand’s argument consists of three parts. The first is a syllogism: Price fixing by all firms in a market is illegal; hence when a monopolist sets its price, it violates the Sherman Act. The second concerned the definition of illegal monopoly status: Monopoly that falls into your lap is perfectly legal; and monopoly acquired through superior skill, foresight, and industry is also legal. The third part was an assertion that efficiency is simply one of the goals of the Sherman Act. Breaking up Alcoa may be inefficient, but the Sherman Act aims primarily to promote atomistic competition in economic and political spheres. Why was Alcoa’s monopoly status unlawful? It was not passively acquired, and Hand concluded that it did not result from superior skill, industry, or foresight. Alcoa actively pursued its power by expanding capacity more rapidly than warranted by demand conditions, thus removing profit opportunities for entrants. There are flaws in this argument. Note that the first proposition is inconsistent with the second. The Hand syllogism leaves no room for monopolies acquired passively or through efficiency. It follows that the first two propositions must be read together, with the second limiting the first to cases in which active acquisition is shown. Of course, since the contradiction between the first and second propositions must have been obvious to Hand, why did he state the first one? Probably to counter the notion, implicit in the abuse standard, that illegality under Section 2 requires unusually oppressive or predatory conduct. In other words, the first proposition was set forth in order the alter the starting premise of antitrust analysis under Section 2. The second argument, concerning the difference between lawful and unlawful monopoly, is a theoretically sensible distinction. But how in

I. Development of Section 2 Doctrine

191

practice can a court distinguish bad capacity expansions from those good capacity expansions made in order to meet demand? Furthermore, anticompetitive, preemptive expansion is rational only if the firm anticipates no future entry. But in the absence of entry barriers, exogenous shifts in technology or tastes can render the incumbent’s preemptive investment obsolete, opening the door to new entrants. Thus, any market with easy entry would seem to be one in which preemptive capacity expansion is an irrational strategy. It follows that antitrust courts could better address preemptive investment claims by adopting an “objective reasonableness” test requiring an examination of market structure, ease of entry, and the stability of technology and consumer demand in determining the plausibility of a plaintiff’s preemptive investment claim. The Court could dispose of objectively unreasonable claims early in a summary judgment ruling.17 The troubling matter with Hand’s monopolization test is the Court’s second-guessing of possibly competitive actions. Most of the entryblocking actions Judge Hand described could also serve as examples of superior skill, foresight, and industry. Which is the correct view, and how should one go about making these distinctions in a consistent manner? The question returns us to the public utility argument raised in TransMissouri: just as the framers of the Sherman Act did not intend federal court judges to review business records in order to determine the reasonableness of prices, they surely did not intend for courts to make decisions on the reasonableness of output decisions – for example, capacity expansions, product designs. Although Hand’s distinction between monopoly acquired lawfully (through luck or superior skill, foresight, and industry) and unlawfully is theoretically appropriate, its application by judges would seem to require courts to emulate public utility regulators. In particular, Hand’s test requires judges to make certain findings, such as the reasonableness of capacity expansion, in the absence of governing legal standards. Hand’s active-acquisition standard tends to push judges beyond their areas of competence and the boundaries of their authority under the Sherman Act. As a result, the test generates a high probability of error. Particularly worrisome are false convictions or findings of liability, because they discourage firms from making competitive price and output decisions. 17

This is based on the framework for predatory pricing claims established in Matsushita Electric Industrial Co. v. Zenith Radio Corp., 475 U.S. 574 (1986), discussed infra in the text accompanying notes 54 and 55.

192

Monopolization

One would ordinarily attempt to lower the probability of an erroneous conviction by requiring evidence of specific intent, that is, evidence of intent to destroy competition. Although the early Section 2 cases such as Standard Oil failed to provide a clear statement of the intent test under the statute, they clearly suggested a requirement on the plaintiff’s part to present evidence of specific intent. Moreover, the evidence emphasized by the early cases was objective evidence of intent, in the form of conduct that could not be defended on efficiency grounds. Under this test, Standard Oil was found guilty of violating Section 2 because its conduct suggested, to the Court’s eye, that its primary or sole aim in many transactions was to suppress competition. While an objective approach to specific intent reduces the likelihood of false convictions for monopolization, there is an alternative approach to intent evidence that fails in this respect: the subjective approach. Under the subjective approach to specific intent, courts analyze the statements of corporate officers and internal corporate memoranda for evidence of a desire to suppress competition. This sort of evidence is unreliable. Legally unsophisticated managers often send around memoranda urging employees to work harder in order to “crush” or “kill” the competition. These statements are unlikely to be correlated with either the firm’s likelihood of success or the social harmfulness of the firm’s conduct. Moreover, such evidence of intent is ubiquitous among legally unsophisticated firms, because all of them are run by people who would like to monopolize their industries but lack the sophistication to avoid putting their desires in print. Hand’s third core argument returns us to the debate over the purpose of the Sherman Act: to promote competition or to enhance consumer welfare? If we say that its purpose is to promote competition, then several questions follow. What kind of competition? Competition in which the best competitor wins, or a regime that encourages numerous competitors? Both “survival of the fittest” and “protection of small businesses” are notions consistent with the general goal of promoting competition. The “survival of the fittest” objective would lead us to give the greatest weight to efficiency as the desirable result of antitrust enforcement. However, Hand’s argument that efficiency should be treated as a second-order concern of the Sherman Act requires a conception of competition that seeks to maintain a steady state in which a large number of independent producers exist. Further, promoting competition in this static or end-state sense will often require limitations on competition.

I. Development of Section 2 Doctrine

193

Large firms must be prevented, under this objective, from driving small firms out, which requires a protectionist policy. Although the ultimate impact of Hand’s approach depends on how courts apply it, it is clear that Hand’s Alcoa test shifts the inquiry from a search for conduct that violates the Sherman Act (the abuse theory) to an examination of structure (specifically, market share and how it was acquired and maintained). However, a structural approach would be unworkable unless the intent requirement were also consistent with it. A specific intent requirement would allow many firms with monopoly power that had not engaged in predatory conduct to evade a conviction, as in the U.S. Steel case. Recognizing this, Hand declared that specific intent need not be shown in a Section 2 monopolization case because “no monopolist monopolizes unconscious of what he is doing.”18 In other words, the courts will presume specific intent. To sum up, Alcoa alters the conduct and intent standards established in Standard Oil in the following senses. First, Alcoa shifts the conduct standard from the abuse test to the active-acquisition test. Second, Alcoa shifts the intent test from a standard requiring proof of specific intent to monopolize, in the sense of showing that no credible efficiency justifications could account for the defendant’s actions, to a standard that presumes specific intent to monopolize. In practical terms, the shift in the conduct standard relieves courts from the constraint of finding a basis in Section 1 case law for condemning the conduct at issue under Section 2. This permits courts to focus entirely on the anticompetitive effects of the defendant’s conduct without having to pigeonhole the plaintiff’s theory into some standing doctrinal precedent. The jettisoning of the specific intent test, moreover, enables the Court to find a defendant liable even when there may be credible efficiency justifications for his conduct. The test empowers the Court to balance anticompetitive effects (e.g., entry barriers) against efficiency gains, and to hold the defendant liable whenever the former seem to outweigh the latter.

18

Alcoa, 148 F.2d at 436. The criminal law makes distinctions between general and specific intent. Specific intent refers to the defendant intended to commit crime that he is charged with. Alcoa would have exhibited only a general intent if it had planned and intended to do nothing more than expand capacity. Specific intent would have been evident if the company had intended to expand capacity in order to exclude rivals and gain monopoly power. On the distinction between specific and general intent, see generally Wayne R. LaFave and Austin W. Scott, Jr., Criminal Law 216–24 (West Publishing Co., 2d ed. 1986).

194

Monopolization

B. Post-Alcoa 1. Current Status. In assessing the current status of the Alcoa doctrine, it is important to note the different ways of interpreting it. Return to the three core arguments of the opinion: (1) that the mere setting of price by a monopolist is unlawful because it is equivalent to all of the firms in a market agreeing on a fixed price; (2) the distinction between monopoly lawfully acquired and maintained (through luck or superior skill, foresight, and industry) and unlawfully acquired and maintained; and (3) the proposition that efficiency is a second order concern under the Sherman Act. If one focuses on the second proposition, one could define unlawful maintenance as equivalent to the abuse standard and argue that the Alcoa doctrine is equivalent to that of Standard Oil. On the other hand, if one focuses on the first and third propositions, then the doctrine suggests a standard under which a large market share leads to a presumption of illegality and efficiency defenses are irrelevant; in effect, overturning key parts of the U.S. Steel decision. Today, many, if not most, federal courts reject substantial parts of the Alcoa doctrine, particularly the first and third propositions. Consider the language of Judge Kozinski in U.S. v. Syufy Enterprises,19 a monopolization case: The government trots out a shopworn argument we had thought long abandoned: that efficient, aggressive competition is itself a structural barrier to entry. The notion that the supplier of a good or service can monopolize the market simply by being efficient reached high tide in Judge Learned Hand’s opinion in [Alcoa]. In the intervening decades the wisdom of this notion has been questioned by just about everyone who has taken a close look at it. The antitrust laws protect competition, not competitors. We make it clear today, if it was not before, that an efficient, aggressive competitor is not the villain the antitrust laws are aiming at eliminating.

What clearly remains of Alcoa is the notion that the unlawful monopolization test is broader than the abuse standard; in the sense that some acts might not meet the Standard Oil definition of abuse, but nevertheless violate Section 2. In short, Hand’s general distinction between lawful and unlawful acquisition remains. Although courts often repeat Hand’s statement that specific intent is not required because “no monopolist monopolizes unconscious of what he is doing,” the actual case law has moved significantly toward the intent requirement implicit in 19

903 F.2d 659, 667 (9th Cir. 1990).

I. Development of Section 2 Doctrine

195

Standard Oil and other early decisions. I hope to make this clear in Part II below. 2. Development and Refinement of Alcoa Doctrine. The Supreme Court provided its stamp of approval to Judge Hand’s arguments in American Tobacco Co. v. United States.20 That is, it reinforced the structural approach to Section 2, and suggested that courts can apply it to Section 1 oligopoly cases. Certain propositions of Alcoa were quoted with approval: (1) The Hand syllogism: price fixing by competitors is like price fixing by a monopolist. (2) Progressive embracing of each new opportunity is exclusionary. (3) Violation of Section 2 requires power and intent, but not specific intent because “no monopolist monopolizes unconscious of what he is doing.” a. GRIFFITH and leverage theory. In United States v. Griffith21 the Supreme Court tried to state a clearer test for liability than that announced in Alcoa. The defendants were four affiliated corporations that owned all or part of theaters in eighty-five towns in the western United States. When the government filed its complaint, the defendants had a monopoly in fifty-three of those towns and in the remaining thirtytwo they were forced to compete. During the five years in which the defendants engaged in the activity giving rise to the complaint, the proportion of towns in which the defendants had a monopoly rose from 51 to 62 percent. The defendants purchased film distribution rights for all of their theaters as a block. For this, the government charged them with violating Sections 1 and 2 of the Sherman Act. The Court accomplished three things in Griffith: (1) it clarified the specific intent requirement, (2) it suggested an alternative to the ambiguous “nonpassive” acquisition test of Alcoa, and (3) it provided a specific theory of liability under Section 2. On specific intent, the Court noted that proof is not required where the evidence indicates that the monopolizing activities were successful – and here there were signs of success. Specific intent is relevant, under Griffith, only in those Section 2 cases with charges of an attempt or a conspiracy to monopolize. The Court suggested, as an alternative to the nonpassive acquisition test of Alcoa, that the monopolist act in an exclusionary manner. This is essentially what Judge Hand concluded had happened in Alcoa. 20 21

328 U.S. 781 (1946). 334 U.S. 100 (1948).

196

Monopolization

The specific theory of liability is sometimes referred to as the Griffith formula: use of monopoly power to foreclose competition, to gain a competitive advantage, or to destroy a competitor, is unlawful. The novel part of this formula is the “advantage” part. Thus, leveraging monopoly power, that is, using monopoly power in one market to gain an advantage in another, is a type of abuse or exclusion that violates Section 2. Since under the Griffith rule proof of specific intent is not required, a court applying the doctrine may conclude that a dominant firm has violated Section 2 when the competition-excluding effects of leverage seem to outweigh purported benefits to consumers. The leverage theory of Griffith has been criticized for double counting monopoly power.22 The argument is that a monopolist of widgets can earn its supracompetitive profit either by charging a monopoly price for widgets or by requiring consumers to purchase something else – for example, a gadget – with every purchase of a widget. Either way, consumers will not pay more than the maximum they are willing to pay. If the method of earning these profits does not allow the monopolist to price-discriminate, it will set a surcharge for the combined widget-gadget package that is equal to the monopoly surcharge that it would have sought on widgets alone. If the method does enable the monopolist to price-discriminate, then it gains additional profit. However, since the price-discriminating monopolist increases his output beyond the monopoly level, consumers may gain also. I will present a more detailed examination of leveraging theory later in this chapter. For the moment, the simple message I hope to get across is that the double counting critique is correct in certain cases but there are also cases in which leveraging theory is valid. b. UNITED SHOE and entry barriers. The cases have announced three standards applicable under Section 2. The first, the abuse standard, requires proof that the defendant engaged in activities that would violate Section 1 if engaged in by two or more parties. The second, the nonpassive acquisition test of Alcoa, condemns monopoly power that is not acquired through luck, superior foresight, industry, or skill. The third is the exclusion test of Griffith. The first is the narrowest test, in the sense that application of the abuse standard results in the lowest likelihood of finding of a violation of Section 2. The second is the most expansive, sug22

Director and Levi, Law and the Future: Trade Regulation, 51 Northwestern Univ. L. Rev. 281, 290, 292 (1956).

I. Development of Section 2 Doctrine

197

gesting that a Section 2 violation can occur through the mere acquisition of monopoly power in a manner that does not satisfy the Court. In United States v. United Shoe Machinery Corp.,23 the Court held that the defendant violated Section 2, under any of the three standards. United Shoe made shoe machinery and was the largest supplier in the United States. There were at least ten other domestic and foreign firms making competing products, but United met 75 to 85 percent of the demand. United, along with the others, leased its shoe machines. It also supplied repair service without separate charge. The issue was whether United Shoe’s position violated Section 2 of the Sherman Act. Judge Wyzanski held that it did: “The facts show that the defendant has, and exercises, (1) such overwhelming strength in the shoe machinery market that it controls the market, (2) this strength excludes some potential, and limits some actual, competition, and (3) this strength is not attributable solely to defendant’s ability, economies of scale, research, natural advantages, and adaptation to inevitable economic laws.” The specific practices that violated Section 2 were the leasing and free service provision. Judge Wyzanski held that they created barriers to entry. The leasing arrangement created an entry barrier because it locked in consumers, making it difficult for a rival to enter and compete for business. The lock-in was accomplished by setting a term as long as ten years on some machines, and requiring a charge for early returns. The return charge assessed against a lessee who returned the machine before expiration declined to insignificance as the lease approached expiration. There are two weaknesses in this argument. First, the leasing policy may have been an efficient response to incentive problems inherent in leasing these machines.24 If the contracts were efficient – and their widespread use in the industry suggests they were in fact efficient – then they lowered the cost of the services the machines provided. How might they have been efficient? (1) The provisions could have provided United Shoe a means of controlling opportunism.25 Suppose a machine has to be made to fit a particular use, and once so made is hard to switch to another use.

23 24

25

110 F. Supp. 295 (D. Mass. 1953). See Scott E. Masten and Edward A. Snyder, United States v. United Shoe Machinery: On the Merits, 36 Journal of Law & Econ. 33 (1993). Oliver E. Williamson, Transaction Cost Economics: The Governance of Contractual Relations, 22 J. L. & Econ. 233 (1979).

198

Monopolization

Then any company leasing such a machine has an incentive to reduce its payments on the lease.26 One way to control this is by making it costly to the lessee to terminate the lease. (2) The provisions may have attempted to discourage shoe manufacturers from free-riding on general (as opposed to machine-specific) information and training provided by United Shoe.27 Specifically, if United Shoe provided valuable training on the repair and handling of shoe machines to shoe manufacturers, they might have an incentive, after acquiring the training, to break their leases and substitute a less expensive machine. (3) The provisions may have been a response to depreciation costs. Suppose the machines were not specific to the user, and were therefore valuable in the secondary market;28 and suppose the secondary market discounts used machines heavily in their first years, just as the used-car market imposes large discounts on relatively new used cars.29 Under these conditions, a lessor would have an incentive to encourage the lessee to hold on to the item or face a significant charge for an early return. The second weakness in Judge Wyzanski’s argument is a basic inconsistency in its structure. Wyzanski concluded that the leasing policy created a barrier to entry, thus allowing United Shoe to earn monopoly profits. But profits attract entrants. Certainly lessees near the end of their terms, when the return charge falls to insignificance, would have a great incentive to switch to a lower-cost rival, and rivals a corresponding incentive to give them that option. The opportunity to underprice United Shoe for the business of those lessees would encourage firms to enter the shoe machinery industry.30 26

27 28

29

30

See, for example, Benjamin Klein, Robert C. Crawford, and Armen Alchian, Vertical Integration, Appropriable Rents, and the Competitive Contracting Process, 21 Journal of Law & Economics 297, 302 (1978). Masten and Snyder, supra note 24. Masten and Snyder note that the machines generally were not tailored specifically to the user’s needs, Masten and Snyder, supra note 24, at 47. Car buyers cannot assess quality at a brief inspection, so the expectation (in equilibrium) is that a relatively new used car has a substantial probability of being a “lemon.” Faced with large discounts, owners of high-quality cars are unwilling to sell at such a steep discount. The result is an equilibrium in which low-quality used cars make up a disproportionately large share of the market. See George A. Akerlof, The Market for “Lemons”: Quality Uncertainty and the Market Mechanism, 84 Q. J. Econ. 488–500 (1970). I am suggesting in the text that the same process could operate in the market for used machines. This must be distinguished from Posner’s criticism. Posner argued that it is implausible that the customers of United Shoe would cooperate with efforts to monopolize the market, unless of course they were adequately compensated, see Richard A. Posner, Antitrust Law: An Economic Perspective 203 (Chicago: University of Chicago Press,

I. Development of Section 2 Doctrine

199

Judge Wyzanski implicitly assumed: (1) that the number of lessees near the end of their lease terms was too small to support entry by rivals, or (2) that the return charges the lessees faced exceeded both the benefits from switching and the costs United Shoe suffered from having a machine returned early. The first assumption is questionable because a rival that could beat United’s terms should have been able to borrow sufficient funds to stay in business until it could capture enough of United’s customers to make a profit.31 In addition, such a rival would have an incentive to make a credible commitment to beat United’s terms in order to induce shoe manufacturers to terminate their contracts with United. The second assumption is also questionable. If the charge did not exceed the costs imposed on United Shoe by an early return, then it was simply fair compensation, and there is nothing anticompetitive about that. If the return charge did exceed the costs imposed on United, then it would have discouraged shoe manufacturers (who presumably would have been aware of the excessive return fees) from entering into a relationship with United in the first place. Furthermore, a return charge substantially larger than the cost United suffered from an early return would have been the sort of penalty that courts traditionally have been unwilling to enforce.32 Put another way, United could not credibly threaten to enforce such a penalty. The free provision of service blocked entry by creating a two-level hurdle. New rivals had to sell machines and service. Of course, new rivals had to build plants also. However, the problem with the two-level hurdle

31

32

1976). Aghion and Bolton have shown that even if customers are compensated, they will still favor a policy that tends to exclude entrants, see Phillipe Aghion and Patrick Bolton, Contracts as a Barrier to Entry, 77 Am. Econ. Rev. 388–401 (June 1987). My point is that if the leasing policy allowed United Shoe to make monopoly profits, these profits would have encouraged entry. There is the problem that customers of United Shoe might be reluctant to switch if they feared that too few customers would switch to enable the entrant to survive. For an exploration of entry barriers attributable to strategic considerations, see Eric B. Rasmusen, J. Mark Ramseyer, and John S. Wiley, Jr., Naked Exclusion, 81 Am. Econ. Rev. 1137–45 (December 1991). However, the customers of United Shoe were shoe manufacturers. The strategic and information-failure stories that might present plausible theories of exclusion in a market of ordinary consumers have to be considered less plausible in this setting. See, for example, Barry Wright v. ITT, 724 F.2d 227, 238 (1st Cir. 1983), where Judge Breyer makes this point in a passage discussing the lock-in effects of a contract. Again, given that the customers of United Shoe were shoe manufacturing firms (not unsophisticated consumers), it seems plausible that at least one of them would have been able to hire an attorney who could point out that courts are reluctant to enforce penalty provisions in contracts.

200

Monopolization

argument is this: Either the monopoly surcharge covered the cost of free service provision or it did not. If it did cover the cost of service, and provided a monopoly profit, a rival would have had an incentive to enter. The rival could have charged a slightly lower price, offered free service, and still earned a better-than-competitive rate of return. If the monopoly surcharge failed to cover the cost of service for a new entrant, then entry deterrence resulted not from the provision of free service, but the fact that United possessed an advantage in providing service, that is, that it was an unusually efficient service provider. But in that case the surcharge would have been a rent United Shoe earned on its special capacity, or experience-based skill, for providing service. Judge Wyzanski’s theory of entry barriers has received theoretical support in an influential article by Phillipe Aghion and Patrick Bolton.33 In the Aghion and Bolton analysis, an incumbent monopolist and a consumer design a contract in one time period (today) specifying price and quantity to be supplied in the following period (tomorrow), knowing that a rival producer may enter in the latter period. Aghion and Bolton show that the wealth-maximizing agreement for the two parties is one in which the monopolist charges a price lower than the ordinary monopoly price, coupled with a liquidated damages provision. The damages provision requires the consumer to pay a certain sum to the incumbent monopolist if he chooses to exit the relationship. Aghion and Bolton show that the penalty is excessive from a consumer welfare perspective. Put another way, the penalty discourages the consumer from switching even in cases in which the entrant is the low-cost producer. Why would a consumer sign a contract in which he forgoes the benefits of purchasing from a lower-cost entrant? The reason is that the incumbent monopolist and the consumer form a coalition in which they act together in the same manner as a monopsonist who purchases from the entrant. The standard monopoly-pricing result applies: they forgo some wealth-enhancing trades in order to appropriate part of the lowercost entrant’s rent. Although Scott Masten and Edward Snyder concluded that the contractual provisions in United Shoe were designed to enhance efficiency,34

33

34

See Aghion and Bolton, supra note 30. For a discussion of the implications of the Aghion-Bolton analysis for antitrust policy, see Joseph F. Brodley and Ching-to Albert Ma, Contract Penalties, Monopolizing Strategies, and Antitrust Policy, 45 Stan. L. Rev. 1161–1213 (May 1993). Masten and Snyder, supra note 24.

I. Development of Section 2 Doctrine

201

the general point of the Aghion-Bolton analysis remains: entry-deterring penalties may be a desirable arrangement for both the incumbent monopolist and its customers. However, it is also important to note some of the limitations on this result. The penalty provisions in the Aghion and Bolton theory do not deter all entrants. A rival firm with sufficiently low costs will enter and take customers away from the incumbent, in spite of the penalty. Indeed, the Aghion and Bolton theory is really one of monopoly pricing, rather than entry deterrence, and therein lies its most significant limitation. An outcome in which a firm earns consistent monopoly profits will attract potential entrants. The penalty lock-in theory relies on an assumption that certain potential entrants – specifically those whose costs are lower than the monopolist’s but not low enough to make entry profitable in the Aghion-Bolton equilibrium – will allow a monopolist to exclude them. But there is an alternative arrangement in which such a potential entrant, the consumer, and the incumbent monopolist can all be made better off. Given that the penalty provision is designed to appropriate part of the entrant’s rent, the entrant could offer the parties an even larger share of the rent, as a “bribe,” in exchange for an agreement on their part to trade with the entrant whenever the entrant is the low-cost producer.35 c. GRINNELL test. The “post-Alcoa” doctrine surveyed so far covers cases that are generally understood as providing further definition to the “active acquisition” standard announced in Alcoa. In addition to Griffith and United Shoe, United States v. Grinnell Corp.36 is an important part of the case law attempting to clarify and refine the Alcoa test. Grinnell states the general test that courts now apply in every Section 2 35

36

I am referring in the text to the case of a single supplier for whom the cost of production is random (the model examined in the Aghion and Bolton paper). The alternative scenario is one of several suppliers whose production costs are fixed. In this alternative story, entry deterrence doesn’t occur because the low cost firms simply enter and take the business away from the incumbent monopolist. Still, even in this case, all of the suppliers have an incentive to band together and offer a bribe that more than compensates the consumer for breaching the penalty provision or renegotiating the contract to remove the penalty provision. I am referring to a general feature of the monopolypricing problem. In the simple case in which a monopolist producer charges the monopoly price to a group of consumers, those consumers could make themselves better (and the monopolist too) by offering the monopolist a side-payment slightly greater than the monopolist’s profit, in exchange for a commitment by the monopolist to produce the competitive amount and charge the competitive price. These agreements are unlikely to be observed because it is difficult for consumers to combine for this purpose – in other words, transaction costs prevent this solution. 236 F. Supp. 244 (D.R.I. 1964), aff’d except as to decree, 384 U.S. 563 (1966).

202

Monopolization

case. Under the Grinnell test there are two essential elements of monopolization: (1) possession of monopoly power; and (2) willful acquisition as distinguished from growth as a consequence of superior product, business acumen, or historical accident. In Part II, I examine some important categories or types of Section 2 claim. We will see that courts have started to modify the Alcoa doctrine substantially, even as they continue to adhere to its formal statement as expressed in the Grinnell test.

ii. leveraging and essential facilities Most if not all of the modern Section 2 cases fall into one of three categories: leverage theory, essential facility theory, and predatory pricing. The leverage theory, introduced in our discussion of Griffith, involves claims that the monopolist has extended its monopoly power from the monopolized market into a related competitive market. The essential facility theory holds that a monopolist may violate Section 2 if it has access to a facility that provides a competitive advantage and denies access to competitors. The key question lurking beneath these cases is whether proof of specific intent is required in order to find the defendant guilty of violating Section 2. Consider, for example, the leverage theory. Is it sufficient to present evidence that the harms resulting from barriers to competition in the extension market probably outweigh purported consumer benefits? Or does the plaintiff have to show that there are no credible efficiency justifications for the defendant’s conduct? If the answer to the first question is yes, then the leveraging doctrine can be applied, consistent with Alcoa, independently of any requirement of proving specific intent. If the answer to the second question is yes, then a defendant can be held liable under the leveraging doctrine only if the evidence indicates a specific intent to monopolize. In this part, I will discuss both the legal doctrine and theory of leveraging, essential facilities, and predatory pricing, largely in the context of simple examples. I will argue that the modern case law leans toward requiring proof of specific intent to monopolize.

A. Leveraging and Essential Facility Cases One prominent example of the Griffith leveraging doctrine in application is Otter Tail Power Co. v. United States.37 Otter Tail refused to allow 37

410 U.S. 366 (1973).

II. Leveraging and Essential Facilities

203

municipalities to “wheel” power through its transmission lines and to sell wholesale power. This was a violation of Section 2. One can view Otter Tail either as a leverage or essential facility case. As a leverage case, the theory is that Otter Tail used its monopoly on transmission lines to gain an advantage in the competitive market for wholesale power. The essential facility theory is that the transmission lines were a facility essential to effective competition, and that Otter Tail’s denial of access excluded competitors from the relevant market. Berkey Photo v. Eastman Kodak Co.38 suggests limits to the leverage theory. Kodak developed a new camera and new film to accompany it. Berkey found itself at a disadvantage temporarily because it could not process the new film, only Kodak could. Berkey brought suit on the theory that Kodak had used its power in the camera production market to gain an advantage in the film processing market. The Second Circuit held that Kodak had not violated Section 2. Berkey Photo now stands for the proposition that a vertically integrated firm has no duty to predisclose information on a new product in order to give rivals a chance to prepare for its arrival. The Court agreed that it made sense to allow a firm to reap the rewards of its own investments in research and development by keeping news of its planned products secret. The most important leverage/essential facility case is Aspen Skiing Co. v. Aspen Highlands Skiing Corp. 472 U.S. 585 (1985). The Court held that the evidence supported a jury’s determination that the refusal by the owner of three of four Aspen ski mountains to continue a joint skiing ticket with another owner violated Section 2. The jury concluded that Aspen Ski Company’s (Aspen) refusal to deal reflected a deliberate effort to discourage customers from doing business with its smaller rival (Highlands). Thus, Aspen violated Section 2 because it scrapped the joint-marketing arrangement only to destroy a rival, a violation of the Griffith doctrine. The Court did not hold that a monopolist has a duty to cooperate with its smaller rival in a marketing arrangement. The jury concluded that the evidence supported the inference that the defendant intended to destroy a rival. The fact that Aspen failed to offer any credible procompetitive justification buttressed the jury’s conclusion. The justifications proffered by Aspen were, first, that it was difficult to monitor the accuracy of the method for allocating revenue between the two companies, and second, that it wanted to disassociate itself from the allegedly inferior services 38

603 F.2d 263 (2d Cir. 1979).

204

Monopolization

of Highlands. The jury rejected both justifications because they were inconsistent with the evidence. There are two controversial aspects to the Aspen decision. First, even though Aspen does not set forth a duty to deal, it does suggest that a large firm should think twice before it quits a cooperative agreement with a smaller rival. In addition, it raises the possibility that if you have market power, the law will condemn you no matter what you do. The litigants in Aspen had been sued earlier for price fixing by the Colorado Attorney General. Thus, entering a joint marketing arrangement exposes the parties to the risk of a Section 1 prosecution, and exiting exposes the larger party to a Section 2 suit. The second controversial aspect was the defendant’s decision not to challenge the finding that it had monopoly power.39 Aspen was a monopolist in Aspen, Colorado. But skiers around the United States consider many other places to ski, perhaps hundreds of other places. In this larger market, it is unlikely that Aspen Ski Company had more than a minuscule market share. However, there is an alternative theory that could explain Aspen’s conduct in terms of monopolization theory. Suppose the ski market can be divided into long term (week stays) and short term (one day stays) customers, where the former come from all around the country and the latter consists of local residents. Local customers may choose to ski Highlands for a short stay. By destroying Highlands, Aspen would be in a position to charge monopoly prices to short-stay customers.40 Indeed, the Court’s opinion makes sense as an application of Section 2 doctrine only if the defendant’s conduct is viewed as monopolization of the short-stay ski market. Although the controversial aspects of the Aspen Ski decision suggest that the Court made Section 2 doctrine more worrisome for dominant firms – by implying a duty to cooperate with rivals – the Court’s treatment of the intent issue signals a significant move away from the Alcoa doctrine. To be sure, the Aspen Ski opinion repeats Hand’s statement that specific intent is irrelevant.41 But the Court also made it clear that the defendant lost for one reason: it failed to bring forth any credible efficiency justification whatsoever. The Court also quoted approvingly from lower court jury

39

40 41

Areeda criticized the jury instructions for leading jury members to analyze an inappropriate market, see Phillip Areeda, Monopolization, Mergers, and Markets: A Century Past and the Future, 75 Cal. L. Rev. 959, 980 (1987). I am indebted to Franklin Fisher for this point. 472 U.S. at 602.

II. Leveraging and Essential Facilities

205

instructions that imposed a specific intent burden on the plaintiff.42 The implication is that if Aspen had brought forth a credible efficiency justification, it would have escaped liability. Put another way, the Court’s opinion implicitly rejects the theory, injected into the case law by Alcoa, that the defendant could be held liable even though there were substantial consumer benefits connected to its conduct on the ground that those benefits were outweighed by competitive harms.43 What does Aspen Ski imply for the essential facility theory of monopolization? The Supreme Court refused to explicitly consider the essential facility theory, even though it was a key part of the appellate court’s reasoning. The Court’s decision to treat this as a case of specific intent, and its refusal to even discuss the essential facility theory, imply that essential facility claims require proof of specific intent to monopolize. In other words, Aspen Ski suggests that the essential facility doctrine is a valid theory of Section 2 liability only to the extent it is consistent with and supported by proof of specific intent. In any event, one may well ask what is the leverage or essential facility theory in this case? The essential facility claim runs as follows: the joint marketing arrangement itself was an essential facility that Aspen denied to its smaller rival. The joint marketing arrangement did not lower the costs of service, but it did raise demand for the joint product, permitting Highlands to service the market for full-week visitors. Without access to the joint marketing arrangement, Highlands could only hope to attract short-stay local visitors, and even in this market they would fare poorly relative to Aspen. However, in some respects the essential facility theory does not fit well here. The marketing arrangement was joint, but Aspen’s size made it especially beneficial to Highlands. The joint marketing arrangement allowed Highlands to free-ride on Aspen’s development and marketing expenditures. Some of these expenditures were substantial: Aspen developed one of the three mountains it owned. These expenditures attracted a substantial number of new visitors to the four mountain area. The essential facility doctrine should not permit a smaller firm to remain in a parasitic relationship with a larger rival. 42 43

Id., at 595–6. For a debate on the importance of “balancing” versus “specific intent” tests in the Section 2 case law, see Steven C. Salop and R. Craig Romaine, Preserving Monopoly: Economic Analysis, Legal Standards, and Microsoft, 7 Geo. Mason L. Rev. 617 (1999); Ronald A. Cass and Keith N. Hylton, Preserving Competition: Economic Analysis, Legal Standards, and Microsoft, 8 Geo. Mason L. Rev. 1 (1999).

206

Monopolization

This suggests an alternative view of Highland’s claim that a mere auditing of revenues generated by use of the mountains could guarantee a mutually agreeable division of the total revenue. When the larger rival is responsible for the vast majority of business, allocation of revenue on the basis of use is not the same as allocating revenue on the basis of productivity. Suppose the four mountains generate $1 million of revenue annually, with each mountain seeing equal use. An allocation based on mountain use would guarantee Highlands $250,000 annually. But if Aspen alone is responsible for 90 percent of the total four-mountain revenue, Highlands should receive no more than $100,000. In Olympia Equip. Leasing Co. v. Western Union Tel. Co.,44 the Seventh Circuit recognized the potential for parasitism in the essential facility doctrine, and accordingly refused to find the defendant liable. In response to an order from the FCC, Western Union unbundled its provision of certain telecommunications equipment and service (it had previously offered a “bundle” in which the customer paid a single price for the equipment and the service), and then decided to exit the equipment provision business altogether. It instructed its sales force to inform customers of alternative providers of telecommunications equipment, and Olympia benefitted greatly from this service. Without ever hiring a single salesperson, Olympia was able to capture 20 percent of the relevant equipment market, all through the referrals from Western Union salespeople. In order to sell its supply of equipment more rapidly, so that it could complete the exit process, Western Union instructed its salespeople to concentrate on selling Western Union’s equipment. Olympia’s market share immediately plummeted to zero, and the firm brought a Section 2 claim against Western Union. Judge Posner, speaking for the Seventh Circuit, held that a monopolist has no duty to subsidize the efforts of a rival. Such a duty would reduce rather than enhance competition according to Posner. Why? A monopolist, realizing that it would be saddled with a duty to continue aiding a rival once it undertook any practice that benefitted the rival, would avoid all such practices. If Western Union had adopted that policy from the start, it would never have allowed its salespeople to inform its customers of the existence of rival equipment providers. The end result is that it would have taken longer for a competitive market in equipment to develop after Western Union’s exit (assuming, of course, that Western Union had a monopoly in the equipment market). 44

797 F.2d 370 (7th Cir. 1986).

II. Leveraging and Essential Facilities

207

B. The Essential Facility Doctrine Under the essential facility doctrine a firm or group of firms that owns a cost-reducing facility must be prepared to share access with competitors on reasonable terms. A denial of access may support a finding of a violation of Section 1 or Section 2. The Section 1 violation would be possible where a group of firms denies access, and the Section 2 violation covers the single firm case. In a Section 1 case, the legal standard is provided by Northwest Wholesale Stationers, which holds that unless a group has market power or access to an essential facility, rule of reason analysis applies to a denial of access. Thus, in the Section 1 context the essential facility doctrine provides an exception to the default rule-of-reason test. The status of the essential facility doctrine is uncertain. Several courts and commentators have applied the doctrine to various cases, but the Supreme Court has never recognized it as an independent theory of liability; divorced, that is, from an intent inquiry. The Court had a clear invitation to do so in Aspen Skiing, where the appellate court had analyzed the case as an essential facility dispute. However, the Court declined, choosing instead to treat it as a case in which the jury had found intent to destroy a competitor. Thus, in spite of the attention given to the doctrine by appellate courts and antitrust commentators, the Supreme Court has signaled that it does not view the essential facility doctrine as independent from an intent inquiry. More specifically, Aspen Skiing implies that an essential facility claim must be accompanied by proof of specific intent. Although courts often cite Terminal Railroad, the opinion in Gamco, Inc. v. Providence Fruit & Produce Bldg.45 provides the best statement of the essential facility doctrine. Gamco provides a concise statement of the theories supporting the doctrine. The basic theory is that parties violate the rule by gaining exclusive access to some facility and using it to ruin a competitor, or to put competitors at a disadvantage. However, there are three types of essential facility acquisition observed in the cases, and along with these types come three distinguishable theories. One is that the parties have combined to gain control of some portal to a substantial market and have used it to put rivals at a disadvantage. Terminal Railroad (Chapter 9), in which the defendants gained control of the rail terminal facilities crossing the Mississippi River into St. Louis, 45

194 F.2d 484 (1952).

208

Monopolization

illustrates this theory. The second theory is that the parties have combined to gain control of some public right and used it to their advantage in competition. An example is provided by American Tobacco Growers, Inc v. Neal,46 in which the state of Virginia permitted the defendants to regulate warehouse sales of tobacco. The defendants abused the selfregulatory scheme to deny competitors access to auctions. The third theory is that the parties have created a facility that either reduces cost or increases demand, and used it in an anticompetitive manner. The example is provided by Associated Press (Chapter 9), in which the government argued that the news-sharing network denied access to nonmembers for no other reason than to monopolize the market. The news-sharing network of Associated Press illustrates some of the limits of the essential facility theory. Suppose the cost of providing news, without access to the network, is $3 per unit of news. With access, news provision costs $2 per unit. Those with access to the network (the members) are able then to undercharge their competitors. They can charge $2.80 per unit, make an $.80 profit on each unit and still undercut nonmember rivals. What’s wrong with this? Nothing, really. This is in fact the way competition is supposed to work. Competition forces members of the news-sharing network to share their cost advantage with consumers. The network members undercut their nonmember rivals and at the same time pass on the cost savings to consumers. This example debunks the claim that facility members are able to shield themselves from competition from nonmembers. Facility members do compete against nonmembers, and that is why the members have to undercut the prices nonmembers charge. The development of a cost-reducing facility does not harm competition, as the Supreme Court suggested in Associated Press; it enhances competition. Taking this into account, the essential facility doctrine is still defensible, on the following grounds. Even though competition forces members to share cost savings with consumers, an even better result could be achieved if nonmembers received access. Providing access to nonmembers would enhance competition, with all firms subject to the lower ($2 per unit) cost level. Instead of paying $2.80, consumers would find a price perhaps as low as $2.10. But forcing members to share access may not enhance consumer welfare. One reason, noted in the earlier discussion of Associated 46

183 F.2d 869, 871–2 (4th Cir. 1950).

II. Leveraging and Essential Facilities

209

Press (Chapter 9), is that there may be an optimal number of facility members. Allowing additional members increases the risk of freeriding, reducing the benefits the facility provides. Specifically, permitting a local rival of a member newspaper to have unrestrained access to a news-sharing network could harm incentives to collect news at the local level. One newspaper might work hard to collect local news, while the other merely scooped it off the wire. Facility members should have the ability to set limits on membership in a manner that optimizes facility services. An open access rule is inconsistent with this goal. A second reason for not requiring open access is its effect on incentives to create an essential facility (of the cost-reducing or demandincreasing type). The economic rent earned by members may be necessary to compensate them for the sunk costs incurred in development. Open access, by introducing new members and increased competition among members, reduces the per-member rent generated by the facility, perhaps below the level necessary for recoupment of development costs. The alternative is to require new entrants to cover not only the cost of development, but also the rent reductions that result from increasing ownership by each additional member. But this approach presents administrative difficulties, and the entry fee could be prohibitive, returning the parties to the access-denial issue that spawned the litigation.47 A third problem is obvious: What prevents the firms from colluding after the imposition of access requirements for nonmembers? The essential facility doctrine requires open access, but what guarantees that the new membership will pass on the benefits of access to consumers? These problems suggest that courts should be wary of using the essential facility doctrine to force the members of a consortium to open the door to a rival. There is no reason to think that the welfare change resulting from an open access rule generally would be positive. An open access rule could discourage the development of cost-reducing facilities, or reduce the services provided by the facility, hurting consumers in the long run. However, there is an alternative theory that suggests that discouraging development is precisely the aim of the doctrine. Suppose the facility is acquired for the sole purpose of putting competitors at a 47

For a full discussion of the theoretical case for pricing access, see William J. Baumol and J. Gregory Sidak, Toward Competition in Local Telephony (1994).

210

Monopolization

disadvantage, perhaps by raising their costs,48 as the Court assumed in Terminal Railroad. Then a wise policy would aim to deter such acquisitions. But how could a court distinguish between own-cost-reducing facility acquisitions (acquisitions that reduce the acquirers’ costs) and rival-cost-increasing facility acquisitions? To simplify the remaining discussion, I will use the term “RCI facility” to describe a facility whose sole function is to increase the costs of rivals. One can make an argument for a conservative approach to essential facility claims. Ordinarily the prospects and incentives are not great for acquiring an RCI facility. In an open auction for such a facility, the winning bid should equal the expected profits derived from owning the facility. Thus, a firm could not, under these conditions, earn a monopoly profit from acquiring an RCI facility. In addition, changes in tastes and technology may render such an acquisition obsolete, as competitors find ways to avoid the disadvantages imposed by the facility. However, these arguments become less persuasive in certain cases. For example, when the firm acquires an RCI facility in the absence of a competitive auction, it could earn a monopoly profit as a result. It follows that a transfer of an RCI facility from the public domain, in the absence of an open auction for acquisition, should raise antitrust concerns. The upshot is that courts should distinguish cases involving the acquisition of an existing market portal from those involving the creation of a cost-reducing (or demand-increasing) facility. Since an access-sharing rule discourages acquisition and development efforts, courts should be quick to reject demands to share access to cost-reducing facilities, such as the news-sharing network in Associated Press (or, as an example of a demand-increasing facility, the joint-marketing arrangement in Aspen Skiing). Cases involving the acquisition of an existing market portal are more difficult, since both Terminal Railroad and American Tobacco Growers can be put in this category. In cases where the market for acquisition is competitive, the potential benefits of an access-sharing rule should be small, since firms will have weak incentives at best to acquire a facility for the purpose of raising the costs of rivals. But in the cases where the market for acquisition is not competitive, as is almost surely the case when the market portal involves some public or regulatory function, courts should stand ready to use the essential facility doctrine to 48

This theory of essential facilities is suggested in Thomas G. Krattenmaker and Steven C. Salop, Anticompetitive Exclusion: Raising Rivals’ Costs to Achieve Power over Price, 96 Yale Law Journal 227 (1986).

II. Leveraging and Essential Facilities

211

require access-sharing. In other words, in a case such as American Tobacco Growers, there is a strong argument for viewing the essential facility doctrine as a rule that protects consumer welfare.49

C. The Leverage Debate The “double counting” critique of the leverage theory runs as follows. The leverage argument double counts monopoly power because it assumes that the monopolist increases its market power through leveraging, when in reality the monopolist has only one source of market power, which is not affected by leveraging.50 Otter Tail illustrates the key issues in the leverage debate. Suppose the transmission lines permit Otter Tail to supply power at a cost of $2 per unit. Rivals who do not have access to the lines would incur a cost of $3 per unit to supply power to Otter Tail’s customers. Access to the lines permits Otter Tail to charge $2.80, undercutting the rivals and locking in customers. The leverage theory suggests that Otter Tail uses its monopoly of transmission lines to gain a monopoly in the market for wholesale power. In this sense, Otter Tail increases or extends its monopoly power in transmission into a related market. The theory may not be correct. Otter Tail could permit a rival to use its lines and simply charge the rival $.80 per unit of power transmitted. Otter Tail would then earn the same profit as when it had exclusive access to the lines. If Otter Tail could easily arrange this option, it would follow that leveraging – specifically, denial of transmission line access to competitors – does not enhance the company’s monopoly power. The “monopoly profit” (technically, rent is the correct term here) has little to do with the decision to deny access to competitors. The analysis so far ignores transaction costs. There are many reasons that a “wheeling” contract – that is, an agreement permitting a rival to

49

50

For a development of this theory, and a critique of the essential facility doctrine, see Keith N. Hylton, Economic Rents and Essential Facilities, 1991 Brigham Young University Law Review 1243. See Director and Levi, supra note 22, Ward S. Bowman, Jr., Tying Arrangements and the Leverage Problem, 67 Yale Law Journal 19, 21 (1957). For an insightful criticism of the leverage thesis, see Oliver E. Williamson, Book Review, 83 Yale Law Journal 647 (1974) (reviewing Ward S. Bowman, Jr., Patent and Antitrust Law: A Legal and Economic Appraisal (1973)). See also Louis Kaplow, Extension of Monopoly Power Through Leverage, 85 Columbia Law Review 515 (1985).

212

Monopolization

supply power through Otter Tail’s lines – may be unattractive. For instance, in order to ensure receipt of $.80 per unit of power, Otter Tail must be able to monitor the competitor’s use of its lines, which may be prohibitively costly. The alternative is to require the competitor to pay an up-front fee for access, which could eliminate the need for monitoring. But this is also problematic. If the plan permits the competitor to share access to the lines, then after the competitor pays the up-front fee it is a sunk cost from his perspective. The competitor has an incentive, then, to increase its share of the revenue by underpricing Otter Tail. Foreseeing this, Otter Tail would charge an up-front fee that compensates for the risk of being driven from the market, which may be a prohibitive amount. The upshot is that because of transaction costs, Otter Tail may not be indifferent as between retaining exclusive access and sharing its lines under a wheeling contract. If this is true, then it is fair to say that denial of access does permit Otter Tail to exploit its position in a way that it otherwise could not.51 Hence, there is something, after all, to the notion that leverage permits a firm to exploit its power in ways it could not, absent leveraging. Of course, whether this is socially undesirable is a separate question.

iii. predatory pricing The Griffith formula holds that a monopolist violates Section 2 by attempting to destroy a rival. One method of destroying a rival is to undercut its price, taking all of the business. Of course, that is how competition works typically; and instead of losing all of their business to a price cutter, rivals respond with matching price cuts. If one firm is more efficient than the other, it may take all of the business after it cuts price to a level below the unit cost of its less efficient rival. The rival, vanquished, exits the market, and consumers gain to the extent the cost advantage of the surviving firm is shared with them in the form of a lower price. In this version of the competitive process, consumer welfare is clearly enhanced, so there should be no violation of the Sherman Act. But suppose the firms are equally efficient, and one lowers its price below its own unit cost (and that of its rival) in order to drive the rival from the market. Suppose the rival, unable to meets its costs, exits the market. Since such a price reduction is unsustainable in the long run, the price-cutting firm eventually will have to raise its price above its own 51

See Williamson, supra note 50, at 654–9.

III. Predatory Pricing

213

unit cost. If the vanquished rival has exited and is not ready to return, the sole remaining firm will raise its price to the monopoly level. This is the classical theory of predation, a simple and old notion of monopolization.52 Although the theory is as old as economic activity itself, predatory pricing has existed as a legal claim for a relatively brief period, probably dating from the 1911 Standard Oil decision.53 Common law courts did not recognize predatory pricing as a valid claim.54 Since there was no antitrust law before 1890, plaintiffs brought common law predatory pricing claims as tort actions. As a general rule in tort law, a seller could not be held liable for underpricing a rival and thereby putting it out of business. The rationale was that the benefits to consumers generally outweighed the harm to the victim.55 One exception to the common law rule existed for the case in which an entrepreneur organized a market or fair, and someone else established a competing fair nearby, or on the most traveled route leading to it.56 The likely rationale for this exception is that the establishment of such a market was a costly process and the benefits widely dispersed. It probably involved considerable effort in advertising. A competitor could free-ride on this investment by establishing a competing market, reducing the return to the original entrepreneur. If this activity were rampant, there would be little incentive for entrepreneurs to set up these occasional markets, which would harm the public. Predatory pricing consists of two periods of action. First, there is the predatory campaign, a period in which the predatory firm cuts price, perhaps below its own unit cost. The predatory campaign continues until it drives rivals out of business. Second, there is the period of recoupment over which the predator raises price to the monopoly level.

A. Predation Cases The significant problem with predation doctrine is the potential for “false convictions” under Section 2 (and by this I refer to false findings of 52

53

54

55 56

For the most up-to-date and complete discussion of predatory pricing theory, see Patrick Bolton, Joseph F. Brodley, and Michael H. Riordan, Predatory Pricing: Strategic Theory and Legal Policy, 88 Georgetown L. J. 2239 (2000). John S. McGee, Predatory Price Cutting: The Standard Oil (N.J.) Case, 1 J. Law & Econ. 137 (October 1958). See, for example, Keeble v. Hickeringill, 11 East 574, 103 Eng. Rep. 1127 (K.B. 1706 or 1707). See Oliver Wendell Holmes, Jr., The Common Law 144–5 (1881). William Blackstone, 3 Commentaries on the Laws of England 218–19 (1765).

214

Monopolization

liability as well). This is a serious problem for several reasons. First, the costs of false convictions are probably substantial, since erroneous findings of predation discourage price competition, the central aim of the Sherman Act. Second, the costs of false convictions, unlike those of false acquittals, are unlikely to be constrained over time by market forces. A false acquittal leaves the market in its natural state, which means that the threat of entry continues to serve as a constraint on monopoly pricing. A false conviction, on the other hand, discourages competitive conduct, both for incumbent firms and potential entrants. Third, the problem of false convictions suggests limits to the usefulness of economic theory as a source of antitrust doctrine. Economists have demonstrated ways in which common business practices (e.g., price cutting by the most efficient firm) may in fact be predatory. But courts are not well equipped to competently analyze all of the predatory strategies identified in the economics literature. False convictions are therefore possible and likely, if courts try to design antitrust doctrine so that it embodies all of the predation possibilities suggested by theory. Courts have shown increasing concern over the past two decades for the false convictions problem in predatory pricing litigation, culminating in the Supreme Court’s opinion in Brooke Group Ltd. v. Brown & Williamson Tobacco Corp.,57 a decision that has all but put an end to predatory pricing litigation.58 Since Brooke Group is an extension of several important cases, I will trace the path leading up to it rather than focus on it alone. Barry Wright Corp. v. ITT Grinnell Corp.,59 a First Circuit opinion on exclusionary practices generally, is a natural starting point in the path leading to Brooke Group. It provides the best discussion in the case law of the false convictions problem, which is the analytical core of modern predatory pricing doctrine. The First Circuit held that the Sherman Act does not make unlawful prices that exceed both incremental (marginal) and average costs. The plaintiff had urged the court to adopt the rule the Ninth Circuit had been applying to predation claims. Under this alternative rule, a price cut where price exceeds average cost would be deemed predatory if the plaintiff could prove by clear and convincing evidence that the cutter 57 58 59

509 U.S. 209 (1993). Bolton, Brodley, and Riordan, supra note 52. 724 F.2d 227 (1st Cir. 1983).

III. Predatory Pricing

215

had a predatory intent. Most of the Barry Wright opinion is an examination of the benefits of a bright line rule versus the Ninth Circuit’s intent rule. The underlying dispute involved an effort by the ITT Grinnell corporation, a maker of nuclear plant pipe systems, to develop an alternative source of supply for a type of shock absorber called a snubber. For many years, Grinnell had received its requirements from Pacific Scientific Company. In order to reduce its dependence on Pacific, Grinnell entered into a contract with Barry Wright under which the latter would attempt to produce and supply snubbers to Grinnell. Barry Wright experienced production delays, and Pacific responded to Grinnell’s efforts by offering sharp price discounts. Grinnell terminated the agreement, and Barry Wright brought suit against Pacific for predation. Pacific had sufficient market power to satisfy the first element of the Grinnell test (see United States v. Grinnell Corp., in Part I.B.2.c). Thus, citing United Shoe, the Court framed the issue as whether Pacific’s actions increased its monopoly power by some exclusionary method. The bulk of the Court’s discussion focused on the relative merits of the alternative standards of proof, not on the specifics of the defendant’s pricing. Judge Breyer provided three arguments for rejecting the Ninth Circuit’s intent-based standard. First, false convictions are likely because of the difficulty in distinguishing competitive behavior from predation. Second, the possibility of error under the intent-based standard would lead lawyers to advise their clients not to cut prices. Third, the possibility of error brings forth frivolous claims. The basic theory is that false convictions are likely and costly, because they encourage rent-seeking litigation and discourage competition. Breyer also held that the Court’s decision in Barry Wright would have been the same even if the First Circuit had adopted the Ninth Circuit’s intent rule. The reason is that Pacific’s costs declined as output increased. Thus, the price cuts offered to Grinnell reflected an effort to expand output in order to exploit scale economies in production, not to put Barry Wright out of business. Because of this, clear and convincing evidence of intent to injure was lacking. Consistent with its effort to get away from rules requiring an examination of intent, the Court suggested that evidence of subjective intent should not receive much weight in predatory pricing cases. The reason is that all business people attempt to increase market share, and given this it is easy to misinterpret internal memoranda expressing a desire to “get more business” at the expense of a rival.

216

Monopolization

The Supreme Court first showed its concern for the false convictions problem in Matsushita Electric Industrial Co. v. Zenith Radio Corp.60 Since Matsushita develops the core of the reasoning adopted in Brooke Group, it is arguably the most important Supreme Court decision on predatory pricing and unquestionably a watershed case in predation doctrine. The plaintiffs charged that a group of Japanese television manufacturers had conspired over a period of twenty years to drive them out of business by selling televisions at below-cost prices in the United States. The trial court dismissed the case after ten years of discovery and the Supreme Court affirmed. Matsushita announces an objective reasonableness test for predation claims. The necessary conditions for successful predation are: (1) the defendants have some prospect of achieving monopoly; (2) monopoly pricing must not result in quick entry by new competitors; and (3) the firms must be able to maintain monopoly power long enough to recoup their losses and harvest additional gain. If any of these claims seems implausible in light of the evidence, the Court should dismiss the predation claim as speculative.61 The claim in Matsushita failed to meet the objective reasonableness threshold on several grounds. Entry seemed to be easy, and no evidence suggested that American television manufacturers could not reenter if the Japanese manufacturers raised price. There was insufficient evidence of success: after twenty years of the alleged predatory conspiracy, Zenith and RCA remained sharing 40 percent of the market. Finally, there was little reason to believe in the stability of the alleged Japanese cartel. Each firm had an incentive to let the others bear the cost of the predation campaign, and each firm would have an incentive to undercut the monopoly price during the recoupment period. The first appeals court opinion to apply the objective reasonableness test of Matsushita was the Seventh Circuit’s A.A. Poultry Farms, Inc. v. Rose Acre Farms, Inc.,62 another opinion stressing the false convictions problem. The plaintiffs were seven independent egg processors – firms that buy, pack, and ship eggs from farmers. They competed with Rose Acre, which was both a producer and a processor (integrated). The plain-

60 61

62

475 U.S. 574 (1986). For an early recommendation of such a test, see Paul J. Joskow and Alvin K. Klevorick, A Framework for Analyzing Predatory Pricing, 89(2) Yale Law Journal 213, 227–8 (December 1979). 881 F.2d 1396 (7th Cir. 1989).

III. Predatory Pricing

217

tiffs introduced evidence that Rose Acre sold some of its output at below cost prices,63 and also evidence of specific intent. The jury returned a verdict of $9.3 million in damages and the trial judge granted a judgment n.o.v. The issue before the Seventh Circuit was whether, assuming that the plaintiffs’ evidence on costs and subjective intent was valid, a reasonable jury could conclude that predation in violation of Section 2 had taken place. The court said that no reasonable jury could find a violation because the evidence did not satisfy the objective reasonableness requirement. Judge Easterbrook, speaking for the Seventh Circuit, identified three general approaches to examining predation: (1) comparing costs and price, (2) examining intent, and (3) the structural approach of Matsushita. The Court argued that the structural approach is more reliable, in the sense of being less error-prone, than the other two. Easterbrook noted that the cost-price comparison approach suffers from two drawbacks. The first is that below-cost pricing is sometimes lawful. Specifically, courts have held below-cost pricing lawful when used as a promotional device to enter a new market,64 and when used by a firm in a declining industry to defray part of the fixed costs of operating.65 The second drawback, noted earlier, is the difficulty of obtaining accurate measures of average and marginal cost. On the intent approach, the Court argued the drawbacks were even more severe, given the limited value of subjective intent evidence and the time consumed in efforts to discover memoranda suggesting bad intent. Expressing concern for the risk of false convictions, the Court held that subjective intent is irrelevant in predation cases. 63

64

65

At trial, plaintiffs produced an expert witness who testified that the prices for RAs specials were consistently below average cost and in one year below average variable cost. Buffalo Courier-Express, Inc. v. Buffalo Evening News, Inc., 601 F.2d 48 (2d Cir. 1979). What about the common practice of designating certain items as “loss leaders,” which are sold at a loss in order to attract customers into the store? See Hiland Dairy, Inc. v. Kroger Co., 402 F.2d 968, 975 (8th Cir. 1968) (A firm may sell a particular item as a “loss leader,” that is at below cost, in order to attract customers into its store and sell other items at higher prices to recoup the losses); United States v. Milk Drivers & Dairy Employees Union, Local No. 471, 153 F. Supp. 803 (D.Minn. 1957). Pacific Engineering & Production Co. v. Kerr-McGee Corp., 551 F.2d 790 (10th Cir. 1977). To see why this makes sense, note that Average Total Cost (ATC) = Average Fixed Cost (AFC) + Average Variable Cost (AVC). Let price = p and quantity = Q. Thus, a firm’s profit is equal to pQ - (AVC)Q - FC. As long as price exceeds AVC the firm has an incentive to produce a positive output, because the fixed costs are incurred even when output is set at zero. Hence, a firm in a declining industry that cannot cover its total costs will still prefer to price at a level that covers AVC.

218

Monopolization

The Court held that the predation claim was not objectively reasonable because Rose Acre only had a small share of the relevant egg market. In the national market, Rose Acre had a 1 percent share. In regional markets, it had a roughly 23 percent share. In what the Court defined as the relevant market, something larger than regional and smaller than national (the firm sold as far away as five hundred miles), Rose Acre had a below 20 percent share. Because this is well below the 33 percent level which Judge Hand (Alcoa, see Chapter 11) deemed insufficient to find monopoly power, the Court held that the market was too competitive for successful predation to occur. Although the Supreme Court has never addressed the issue directly, an alternative approach to deciding predation claims is through examination of antitrust injury. Viewed from this perspective, the Court’s decision in Atlantic Richfield Co. (ARCO) v. USA Petroleum Co.66 has implications for predation disputes. ARCO involved a vertical, maximum price fixing agreement between ARCO and its dealers. USA Petroleum and other independent gasoline retailers were hurt by the loss in sales they suffered as a result of the low prices ARCO-brand dealers charged. USA Petroleum charged that ARCO and its dealers had engaged in price fixing, in violation of Section 1, and attempted monopolization, in violation of Section 2. The Section 2 claim was later withdrawn with prejudice. The district court granted summary judgment for ARCO on the Section 1 claim because USA Petroleum had not shown that ARCO dealers had charged predatory prices,and without such a showing could not satisfy the antitrust injury requirement of Clayton Act Section 4. The Court further held that USA Petroleum could not show that the prices were predatory, because the predation claim failed the objective reasonableness test of Matsushita. The Supreme Court accepted this reasoning. Taking as given the finding that the prices charged were not predatory, the Court held that USA Petroleum failed to show antitrust injury because they had not alleged that the low prices hurt consumers. As a rule, the Court announced, “[a]lthough a vertical, maximum price-fixing agreement is unlawful under section 1 of the Sherman Act, it does not cause a competitor antitrust injury unless it results in predatory pricing.”67 One can choose to read the decision in ARCO narrowly as a Section 1 case. However, given the concern over false convictions reflected in the Court decisions, it would seem irresponsible to ignore ARCO’s implica66 67

495 U.S. 328 (1990). Id. at 339–40.

III. Predatory Pricing

219

tions for predatory pricing law. The key implication is that if standing (antitrust injury) requires evidence that the price cuts were predatory, then the objective reasonableness test of Matsushita has become a requirement for standing. And if so, evidence showing subjective intent to destroy a rival is irrelevant in the case of a plaintiff who cannot meet the objective threshold. The Court embraced a cost-price comparison rule in Brooke Group Ltd. v. Brown & Williamson Tobacco Corp.68 Indeed, Brooke Group combines the price-cost filter and the objective reasonableness test of Matsushita. To survive a motion for summary judgment, a plaintiff in a predation action must show (1) that the prices complained of are below an appropriate measure of its rival’s costs (the Court did not decide whether that measure should be average cost or average variable cost), and (2) the defendant had a dangerous probability of recouping its investment in below-cost prices. The first requirement brings all of the federal courts in line with the price-cost rule adopted by the First Circuit in Barry Wright and by several other courts. The second requirement is a restatement of the Matsushita doctrine. The plaintiff in Brooke Group, Liggett, lost its predation suit even though it was able to satisfy the first element of the test by showing that Brown & Williamson’s prices for generic cigarettes were below average variable costs. Liggett could not clear the second hurdle, proof of a dangerous probability of recoupment, for many of the same reasons stressed in the Matsushita opinion. The plaintiff’s theory posited that Brown & Williamson conducted a prolonged predatory campaign, in order to recoup its losses later after the plaintiff had been beaten into submission to a tacit price-fixing scheme. The Court, citing Matsushita, argued that the incentives to defect during the recoupment period from such a collusive scheme would be so strong that doubts about the evidence would have to be resolved in favor of the defendant. The Court noted that the evidence did not indicate that prices rose to supracompetitive levels after the predation period, or that the parties were behaving in a manner consistent with tacit collusion.

B. Predation Theory Antitrust commentators have recognized the problem of false convictions in the predation area for a long time. The most influential effort to 68

509 U.S. 209 (1993).

220

Monopolization

Figure 10.1 Areeda and Turner Analysis

state bright line rules that would minimize false conviction costs is the Areeda and Turner prescription. 1. Areeda and Turner. Areeda and Turner divided the typical cost curve diagram of elementary economics into several regions (see Figure 10.1). In the first region, A, price is greater than or equal to short-run marginal cost and greater than (or equal to) short-run average cost. In region B, price is greater than marginal cost, less than average cost, and greater than average variable cost. In C, price is less than marginal cost and less than average cost. Areeda and Turner state that price cuts in areas A and B should be deemed nonpredatory, while those in area C should be deemed predatory. Although Areeda and Turner did not explicitly adopt an error-cost framework, it is easy to translate their argument into statements about the relative costs of false convictions and false acquittals in the regions shown in Figure 10.1. Consider region A. Areeda and Turner argue that price cuts in this region will not eliminate equally efficient rivals, and will result in more efficient use of resources, since price will be closer to marginal cost. Hence, in their view, false convictions predominate in area A, which implies price cuts in A should be shielded from Sherman Act litigation.

III. Predatory Pricing

221

Consider region B. Areeda and Turner argue that price cuts in this area lead to welfare-enhancing output increases since price exceeds marginal cost. Although these price cuts may eliminate equally efficient rivals, they conclude that is an unavoidable risk, given that any rule protecting equally efficient rivals would also protect less efficient rivals. Moreover, the administrative costs of determining some other price level (e.g., “cost-minimizing” price level) above which the dominant firm should keep its price would be extremely high. It follows that costs associated with false acquittals are less than those associated with false convictions, and so a bright line rule protecting price competition is appropriate. The remaining region is where price is less than marginal cost and less than average cost. Consumers gain from these cuts, but what they gain is a pure transfer from the price cutter. Since price cuts in this region move price away from marginal cost, leading to inefficient overproduction, they are socially costly in terms of resource allocation. Moreover, such price cuts may eliminate rivals of equal or even superior efficiency. Because the ratio of false acquittal to false conviction costs is higher in this region, Areeda and Turner argue that these price cuts should be deemed predatory. Areeda and Turner suggested, as a solution to the problem of measuring marginal cost in the single output case, that courts use average variable cost as a proxy for marginal cost. Of course, this suggestion is harder to implement in the multiple output case, in which it is difficult to define or to measure marginal cost. Suppose, for example, a firm produces haircombs and hairbrushes. How should one measure total output? Assuming an appropriate measure of total output, how should one allocate fixed costs? 2. Limit Pricing and Predation. A number of authors have noted that the Areeda and Turner test fails to capture at least one type of predation: limit pricing. Limit pricing refers to the strategy of setting price at a level that will make it unprofitable for a rival to enter, provided that the incumbent holds output fixed.69 The incumbent monopolist could easily satisfy the Areeda and Turner conditions (indicating

69

In du Pont (Titanium), 96 F.T.C. 653 (1980), the FTC held that limit pricing is a reasonable business strategy that does not violate the antitrust laws, provided that there is no evidence of below cost pricing or that the defendant’s strategy depended on an ability to recoup losses in a later period.

222

Monopolization

Figure 10.2

nonpredation) before entry, and if the threat to hold output fixed is credible, entry will never occur. Figure 10.2 illustrates the limit pricing story. Suppose the monopolist’s costs are as shown, with AC representing average cost. Suppose the entrant would have the same average cost schedule as the incumbent monopolist, which suggests that qe is the minimum that the entrant needs to produce in order to compete against the incumbent monopolist. In the equilibrium illustrated in Figure 10.2, if the monopolist retains output at qo after entry, total output will be at least qo + qe, which is too high for any of the firms to break even. Realizing this, the potential entrant stays on the sidelines. The problem with this strategy is that it is not rational ex post, that is, after entry. Once the rival has entered the market, the rational move is to accommodate by reducing output or to enter into a collusive agreement if possible. The rival is capable of seeing this; hence the threat to maintain output at the preentry level is not credible. Can such a threat be made credible? One response is that the incumbent monopolist makes irreversible investments. Suppose plant size is

III. Predatory Pricing

223

chosen strategically.70 Then the incumbent monopolist would choose a size that precommits it to maintaining a high level of output even after entry occurs. Precommitment occurs because the monopolist would lose so much from having excess capacity that the rational decision would be to maintain output at the preentry level. Would it be rational for an incumbent monopolist to choose plant size in order to strategically ward off entry? Judge Hand suggested that a monopolist would adopt such a strategy in Alcoa. Recall that Hand held that Alcoa had anticipated demand and invested in capacity in order to deny market opportunities to new rivals. However, preemptive capacity investment would not be a good strategy in a world with rapidly changing technology and tastes. Such changes could make capacity investments obsolete and open new routes of entry for potential rivals. 3. Uncertainty and Counterstrategies. The point that technology and tastes change, making predatory investments obsolete, is an example of a general problem for predation theory: uncertainty. Predation requires suffering losses today in order to reap rewards in the future. But markets change. For predation to be rational, the expected reward would have to be sufficiently high to compensate for the risk associated with such a strategy. Even if tastes and technology are stable, it is difficult for a firm to predict when a rival will finally exit the market. If it is clear that in terms of economic fundamentals, the rival can compete against the monopolist, then the rival should be able to borrow funds that would allow it to withstand a lengthy predatory campaign. Investors presumably would realize that the rival has good long-term prospects, in spite of losses suffered during the campaign.71 Furthermore, the monopolist cannot be sure that other rivals will not appear once it has driven one from the market. There is also the problem of counterstrategies.72 Think about predation as analogous to the cops and robbers games played by children.

70

71

72

See Oliver Williamson, Predatory Pricing: A Strategic and Welfare Analysis, 87 Yale L.J. 284 (1977). But see Bolton, Brodley, and Riordan, supra note 52. They argue that investors will often not be constrained by incentive-based contracts to demand that lenders meet certain revenue targets. Given these constraints, investors will often not be able to keep funding the rival until it survives the predatory campaign. Frank Easterbrook, Predatory Strategies and Counterstrategies, 48 U. Chic. Law Review 263 (1981). On the general theory of strategies and counterstrategies, see Thomas C. Schelling, The Strategy of Conflict (1960).

224

Monopolization

Children adopt sophisticated strategies, such as playing possum, in order to fool playmates. Businesses are also capable of adopting such strategies; the rival targeted by a predator could play “dead” by shutting down production early, and wait for the monopolist to raise its price. In a market in which entry and exit are easy, rivals could enter opportunistically when the monopolist raises price, and exit when the monopolist lowers price. Another strategy is to inform consumers directly. Informed, rational purchasers, such as other businesses, may prefer to do business with a firm that promises to keep its price within a certain range over a long period, rather than risk finding themselves dependent on a single firm that might raise its price to the monopoly level. The rival could enter into long-term contracts with consumers, or secure the business of a sufficient number of consumers to weather a predatory campaign. This is not a speculative argument: Recall that in Barry Wright, the plaintiff, who complained of price cutting by the incumbent monopolist, had been supported with the help of the largest customer in the mechanical snubber market, ITT Grinnell, to serve as an alternative source of supply. Predation theory implies two regularities. One is that successful predation should be difficult to carry out in a contestable market,73 or one in which entry is easy. Rivals will exit when prices fall and enter when prices rise. The other regularity is that successful predation should be difficult in a market that includes large, sophisticated purchasers because they will have incentives to maintain a diversified portfolio of supply sources. 4. Predation and Reputation. In spite of the uncertainties with respect to tastes and technology, and the counterstrategies available to consumers and potential rivals, some have argued that the reason a dominant firm would nevertheless engage in predation is to establish a reputation as a predator. After establishing the reputation, no further work need be done, as the reputation itself deters entry in future periods and in other markets. The reputation argument also provides an answer to the claim that it is unlikely that a monopolist will recoup the losses

73

Contestability refers here to a market that is subject to hit and run entry, which is not necessarily one in which initial entry costs are low. See Elizabeth Bailey and John Panzar, The Contestability of Airline Markets during the Transition to Deregulation, 44 Law and Contemporary Problems 125–45 (1981). See generally William Baumol, John Panzar, and Robert Willig, Contestable Markets and the Theory of Industry Structure (1982).

III. Predatory Pricing

225

suffered during a predatory campaign. If one predatory campaign establishes a reputation that deters future entrants in several markets, or in future periods, then predation could be a profitable strategy even though the prospects for immediate recoupment appear to be dim. Research in game theory has shown the conditions under which a monopolist would predate in order to establish a reputation. Suppose the incumbent will exist as a monopolist for N time periods. In period N + 1, the market becomes competitive, so the incumbent firm no longer enjoys monopoly power. Suppose a rival enters in period N. Would the incumbent cut price in a predatory manner? No. To do so would lead to losses suffered in period N that could not be recouped in period N + 1. So the rational strategy is to accommodate entry in period N. Reputation as a predator is worthless to the incumbent in the last period of existence as a monopolist. Consider period N - 1. The rival enters. Should the incumbent cut price predatorily? That would involve suffering losses in period N - 1 in order to establish a reputation as a predator in period N. But we know already that the threat to predate in period N is not credible. Hence, the rational strategy is to accommodate in period N - 1. I could continue to roll backward through the periods. The end result is that the monopolist has no incentive to predate in order to establish a reputation in any period, 1 through N. I have just presented a brief version of an argument known as the chain-store paradox, due to Reinhard Selten.74 The basic proposition is that if the game has a finite length, then predation for the purpose of establishing a reputation as a predator is irrational. The proposition relies on the assumptions that the parties are rational and informed, and that the game has finite length. If either of these assumptions were dropped, then the proposition would not necessarily hold.75 It follows 74 75

Reinhard Selten, The Chain-Store Paradox, 9 Theory and Decision 127–59 (1978). If the finite length assumption is dropped, then the “Folk Theorem” applies. According to the theorem, an infinitely repeated game can have a wide variety of (subgame perfect) equilibria if the players’ discount rates are not too high. See Drew Fudenberg and Eric Maskin, The Folk Theorem in Repeated Games with Discounting or with Incomplete Information, 54 Econometrica 533 (1986). On the importance of the rationality and perfect information assumptions, see David M. Kreps and Robert Wilson, Reputation and Imperfect Information, 27 Journal of Economic Theory 253–79 (1982); Paul R. Milgrom and John Roberts, Predation, Reputation, and Entry Deterrence, 27 Journal of Economic Theory 280–312 (1982). The literature in this area has grown in complexity. The more recent examinations of predatory pricing theory have been experimental. See Yun Joo Jung, John H. Kagel and Dan Levin, On

226

Monopolization

from this that rational predation is in no sense ruled out by predationgame theory, much of which is devoted to exploring scenarios in which the chain-store paradox assumptions do not hold. Because the chain-store paradox depends on several assumptions, antitrust commentators have been reluctant to spell out its implications for predatory pricing law – the one exception being Frank Easterbrook.76 The key question is whether the assumptions underlying the paradox accord with reality. One can make an argument that the chain-store paradox assumptions are not far from reality – though this is admittedly a question for empirical research. The rationality assumption is an accurate description of firms, certainly the large ones. They employ consultants, accountants, and lawyers to aid them in making rational decisions. Perhaps the finite period assumption is easiest to question. However, that assumption is consistent with facts in many cases. Because of changes in technology, monopoly status will exist only for a predictably finite period in many cases. It is difficult to point to an example of a dominant firm whose market position is unlikely to be contested at any time in the foreseeable future. The future demise of monopoly status is often obvious long before it occurs. Take for example, local telephone service. The potential for optical fiber and wireless systems to provide competition has been clear for a long time, and the erosion in the monopoly status of the local telephone service provider has been under way for several years.77 5. An Alternative Theory. In light of the foregoing, the reader may wonder whether I think there is any theory of predation that clearly suggests the practice is sufficiently harmful to social welfare to warrant the costs of false convictions and the administrative costs of enforcement. There is such a theory in the common law of business torts.78 Recall that

76

77

78

the Existence of Predatory Pricing: An Experimental Study of Reputation and Entry Deterrence in the Chain-Store Game, 25 RAND Journal of Economics 72–93 (1994); R. M. Isaac and V. L. Smith, In Search of Predatory Pricing, 93 Journal of Political Economy 320–45 (1985). Easterbrook, supra note 72, has gone furthest in spelling out the implications of the chain-store paradox. Easterbrook suggests that the reputation argument may not provide a good theory of predatory pricing. On the implications of wireless communication for competition in the telecommunications industry, see Peter Haynes, The End of the Line: A Survey of Telecommunications, v. 329 (n. 7834), The Economist (October 23, 1993). William Blackstone, 3 Commentaries on the Laws of England 218–19 (1765–9).

III. Predatory Pricing

227

common law courts recognized a predation claim when an entrepreneur had set up a temporary market, and a rival set up a competing market nearby. More generally, the theory recognized by common law courts was that competition could sometimes take a parasitic form, in which one firm free-rides on the investments of another. The problem is defining and generalizing the common law theory so that courts can use it to distinguish parasitic from classical predation. The temporary market example suggests that the key features of the common law theory are: (1) the incumbent incurs significant set-up costs or significant sunk investments, (2) some period of monopoly is needed in order to recoup the set-up costs, and (3) the rival enters and undercuts during the period of recoupment, and (4) the rival free-rides in some sense on the incumbent’s investments.79 Absent the first feature there is little reason to offer protection to the incumbent, as it loses nothing if a competitor enters during the period of recoupment. The second feature seems necessary because if the incumbent could recoup its entry costs in the first minute of operation, or by selling assets while exiting, then no argument exists for providing protection. The fourth requirement is a method of distinguishing claims that merit protection from those that do not. The parasitic predation theory outlined here shares features with the theory of patent protection. Under these conditions, predation is harmful for the same reasons copyright or patent infringement is harmful; it destroys incentives to enter into productive activities. Are there any predation claims that could be justified on the basis of the theory offered here? Probably, though cases with the required facts should be observed in infringement actions under trademark, copyright, and patent law, and in tort suits for unfair competition. The alternative theory suggested here encourages plaintiffs in these cases to sue under the Sherman Act if the defendant has monopoly power or a dangerous probability of acquiring it. The parasitic predation theory also suggests, unlike the classical one, that the plaintiff should be permitted to sue even when the price-cutting rival has kept its price above its own average cost. 79

Alternatively, the rival engages in “cream-skimming” competition. On cream-skimming competition generally, see Breaking Up Bell: Essays on Industrial Organization and Regulation 61–94 (David Evans ed., 1983). The cream-skimming entrant is not the most efficient firm. It underprices the incumbent only by evading regulatory price or service requirements.

228

Monopolization

One might argue that the United States v. Microsoft80 case is an example of parasitic predation. After Netscape introduced its internet browser, Microsoft created a competing browser that it gave away free along with the Microsoft computer operating system software. Thus, the Microsoft case seems to satisfy the first three components of the parasitic predation definition. The questionable part of this argument involves the fourth component. The Microsoft case does not involve credible allegations of free-riding. Unless it could be shown that Microsoft free-rode on some investment by Netscape, the parasitic predation theory would be inappropriate. Perhaps one could argue that the idea of a browser itself was stolen by Microsoft, but this type of claim is better left to patent law.

iv. conclusion This chapter has traced the development of Section 2 doctrine and provided critiques of some theories of liability under Section 2, such as the leveraging and predatory pricing theories. The tension between economic reasonableness and administrative concerns emphasized in previous chapters is revealed in the case law here and has influenced the doctrine’s development. Alcoa’s theory of liability requires courts to determine the economic reasonableness of various expansion strategies adopted by firms. Since this type of review is outside of the traditional areas of judicial expertise, the probability of an erroneous decision is higher than in the usual case. An erroneous finding of liability in this context is worrisome because it works against the procompetitive goals of the statute. As courts have recognized this problem, they have shown increasing reluctance to apply a reasonableness standard that attempts to balance anticompetitive effects against efficiency gains. With the Aspen Skiing decision the Court suggested a return to the specific intent test implicit in the earliest Section 2 cases. Under that test, the plaintiff must show that the sole motivation behind the defendant’s conduct was the elimination of competition, which relieves courts of the balancing of competing explanations invited by Alcoa. The recoupment test of Matsushita imposes a similar requirement in predatory pricing 80

United States v. Microsoft Corp., 1998–2 Trade Cas. (CCH) ¶ 77,261 (D.D.C. 1998); see also United States v. Microsoft Corp., 147 F.3d 935 (D.C. Cir. 1998). For the Findings of Fact, see No. 98-1232 (D.D.C. 1998) (issued Nov. 5, 1999), ·www.usdoj.gov/atr/cases/f3800/msjudgex.htmÒ.

IV. Conclusion

229

cases. Because the recoupment test assumes the price-cutting firm would not have conducted a predatory campaign if it could not have expected to recoup its losses, it effectively requires the plaintiff to prove by objective evidence that the sole motivation behind the defendant’s conduct was the destruction of competition. In other words, the recoupment test is a specific intent requirement. It remains to be seen whether the modern trend toward the specific intent requirement will become a permanent feature of Section 2 case law.

11 Power

What is monopoly power? The economic definition of monopoly is simple: sole producer. That is not a useful concept for antitrust law, because there are few sole producers in the American economy. Antitrust courts have defined monopoly power as the power to control price or to exclude competition. In this chapter, I will focus on the first part of this definition.

i. measuring market power There are three methods of measuring monopoly power. (1) Market Share. The traditional method of measuring monopoly power, and the method commonly used in courts today, is to examine market share. This requires definition and identification of the relevant market. Given a definition of the relevant market, a firm’s market share is equal to the firm’s sales volume in the relevant market divided by total sales (by all firms) in that market. Market share estimates can vary substantially depending on the definition of the relevant market. Areeda’s text offers the following example: suppose there are ninety-nine producers of pleasure boats and one producer of canoes.1 If the relevant market is canoes, then the canoe producer has a 100 percent market share. If the relevant market is pleasure boats, then the canoe producer has a 1 percent market share. 1

Philip Areeda and Louis Kaplow, Antitrust Analysis: Problems, Text, Cases 572–3 (4th ed. 1988).

230

I. Measuring Market Power

231

It should come as no surprise that parties spend vast sums of money in litigation in efforts to get the Court to accept their definition of the relevant market. Determination of the relevant market seems to have dictated the outcome in several important antitrust cases.2 (2) Profit Margins. Another method of testing for monopoly power is to look for evidence of large profits. In many Section 2 cases, courts have cited large profits as evidence of monopoly power. This method is at least as unsatisfying as the market share approach, for several reasons. First, economic theory holds that a monopolist makes profits in excess of opportunity costs even in the long run. In other words, a monopolist makes positive economic profits in the long run. “Monopoly profits” are simply economic profits. However, it is difficult to measure economic profit, and courts typically make feeble efforts at best. The standard approach is to examine accounting profits. Some courts compare these to an industry average or to profit margins in similar industries. The use of accounting profits can be misleading. Accounting profits do not generally equal monopoly profits; indeed, there is no reason to expect them to be the same. In the long-run equilibrium of a perfectly competitive industry, each firm with productive assets whose opportunity cost is not counted in the standard accounting framework will make a positive accounting profit and zero economic profits. Economic profit is generally less than accounting profit because economic profit is equal to accounting profit less compensation for risk-taking and the “nonaccounted” market value of resources used in production. Thus, if a firm has an unusually competent set of managers, the market value of the management team, appropriately amortized, should be subtracted from the firm’s income statement to arrive at an estimate of economic profit. The same goes for all resources used in production whose market value in the best alternative use is not reflected in the expenditure portion of a firm’s income statements. In addition, the firm should make some allowance for risk-taking. Returns should be higher in riskier industries, all else equal. The mere existence of a positive accounting profit does not imply that the firm has monopoly power.3 Further, accounting profit itself is an

2

3

For two examples, see Aspen Skiing (Chapter 10) and Philadelphia National Bank (Chapter 15). I should note, however, that there is a literature on the persistence of high accounting profits – finding a significant degree of persistence. See, generally, Dennis C. Mueller, Profits in the Long Run (Cambridge: Cambridge University Press, 1986).

232

Power

unreliable measure judged on its own terms. It depends heavily on the method of accounting chosen – most important, the method of depreciation.4 Firms routinely make decisions whether to declare income in Period 1 as opposed to Period 2 on the basis of tax, stock market, or other considerations that have little to do with the underlying business fundamentals. To see the potential role of risk-taking in assessing profit margins, consider the Cellophane opinion,5 where the Court noted that cellophane could be and had been displaced by other flexible wrapping papers in some markets. The Court also noted that du Pont had reported high profit margins on cellophane production in several years.6 However, some features of the market in which du Pont operated suggest that high profits should have been anticipated. Cellophane was the result of extensive research efforts by du Pont’s chemists,7 and could be displaced in the market of end uses by several flexible wrapping papers. Given these conditions, du Pont’s efforts should be compared to prospecting for oil or gold. Some attempts will pay off fantastically, while most will fail. How should one account for risk in such an environment? (3) Constraints on Pricing. Disappointment with the first two approaches has led some economists to propose more direct methods of examining power over price. Most of these new methods aim to get directly at some assessment of a firm’s ability to raise price without being constrained by competitors. The most widely accepted approach is now embodied in the Justice Department Merger Guidelines. The approach takes pricing constraints into account in determining the relevant market, and works as follows. Start with a narrow definition of the market, one in which the firm presumably has a high market share. Now work through the following experiment. What would happen if a monopolist in the chosen market 4

5 6

7

See Franklin M. Fisher and John J. McGowan, On the Misuse of Accounting Rates of Return to Infer Monopoly Profits, 73 American Economic Review 82 (March 1983). United States v. E.I. du Pont (Cellophane), 351 U.S. 477 (1956). The Court referred to the government’s evidence that du Pont Cellophane had earned an average after-tax return of 15.9 percent on total cellophane operating investment over the period 1937–47, which the Court described as “liberal,” see du Pont (Cellophane), 351 U.S. at 404. For an economic history, see George W. Stocking and Willard F. Mueller, The Cellophane Case and the New Competition, 45 American Econ. Rev. 29–63 (March 1955), pp. 32–57. A thorough discussion of the case is provided in Eugene M. Singer, Antitrust Economics: Selected Legal Cases and Economic Model 54–61 (Englewood Cliffs, N.J: Prentice Hall, 1968).

I. Measuring Market Power

233

imposed a small but significant and nontransitory increase in price? In particular, suppose – as the Department typically hypothesizes – the increase is 5 percent for one year? If so many consumers would switch to the products of firms operating in a different market that the price increase would be unprofitable, then the Department expands the market definition to include the firms operating in that other market. Repeat the experiment until one reaches a market in which the hypothetical monopolist could profitably raise price by 5 percent for one year. The Department chooses that final market as the initial candidate for the relevant market. The relevant market is the smallest market that satisfies this test. Consider the canoes-pleasure boats example discussed earlier. Suppose our aim is to measure the market share of a dominant canoe producer. Under the Justice Department Guidelines, one would start with canoes as the initial market. If a hypothetical monopolist of canoes could impose a 5 percent price increase for one year, and find it profitable, then the process ends there. Canoes would be the relevant market used in measuring the market share of the dominant canoe producer. Suppose the next largest market definition is “pleasure boats less than twenty feet in length.” If as a result of the 5 percent price increase, so many potential canoe purchasers decided to purchase a small pleasure boat that the canoe producer could not profitably maintain the new price, then the market definition should be expanded to include both canoes and small pleasure boats. The next step is to consider what would happen if a hypothetical monopolist within this market imposed a 5 percent price increase. Would consumers switch to pleasure boats of twenty to thirty feet in length? The approach just described is a simple version of the test in the Justice Department Guidelines. There are several issues that must be taken into account in applying the Guidelines approach. The first goes to its usefulness. Return to the canoes-pleasure boats example. If the canoe market has been monopolized, then the market price of a canoe is, of course, the monopoly price. Since a monopolist always produces where demand is elastic, a price increase will result in a loss in revenue, and will (by hypothesis) be unprofitable to the monopoly canoe producer. In this case, the Guidelines approach might suggest that the relevant market is larger than the canoe market, which is wrong.8 If the 8

See Franklin M. Fisher, Horizontal Mergers: Triage and Treatment, Journal of Economic Perspectives (No. 2), pp. 23–40 (Fall 1987), at 29.

234

Power

market initially examined really is monopolized, some care must be taken to avoid this mistake. Putting this concern to the side, there are other difficulties in assessing the likelihood that consumers will switch after a 5 percent increase in the price of a product. One is the extent to which there are perceived quality differences between the product of the monopolist and the proposed substitute. Quality differences sometimes prevent consumers from substituting away from a high-priced product. Consider, for example, the experience of the American auto industry over the 1980s. Although the federal government tried to protect the domestic car industry by adopting tariff and import-restricting measures that raised the prices of Japanese cars, there were many consumers who were unwilling to switch to American-made cars because of their belief that they were of inferior quality. Another problem is that of networks, observed in the software industry. Purchasing one product may be desirable because it permits you to link into a network of similar products. In the software industry, a large firm such as Microsoft can ensure that its products become industry standards. The desire to have the ability to communicate and work with others makes it costly to switch away from the industry standard after an increase in price. Long-term contracts present another barrier to consumer switching and may force the analyst to consider a period longer than the one year suggested in the exercise above. If consumers are locked into threeyear contracts, a price increase will not generate a stampede toward the cheaper substitute. One wants to consider as well historical evidence on the interchangeability of the products, since the evidence may suggest some hard-to-discern reasons that consumers stick with a particular product. A second set of concerns falls under the category of “supply-side substitution.” The discussion so far has considered the behavior of consumers switching between products. A substantial price increase by a monopolist also may lead to entry by potential competitors. Of course, several factors may make supply-side substitution unlikely. For example, the cost of entry may be high, or there may be large sunk costs (i.e., irretrievable costs) connected to an effort to enter into a line of production. One could generate other considerations, and the Justice Department Guidelines discuss the important ones. However, the key point is that the analysis of pricing constraints is a fact-sensitive, tedious process.

II. Determinants of Market Power

235

ii. determinants of market power One way of numerically measuring a firm’s power to set price above the competitive level is to use the Lerner index. The Lerner index is L = ( p - mc) p , where p = price and mc = marginal cost.9 There are two significant drawbacks to this measure. First, it is hard to measure marginal cost. Second, even with an accurate measure of marginal cost, the index would overstate the extent to which price exceeds the competitive level. The reason the index overstates the monopoly surcharge is that mc is measured at the monopoly output level, which is usually less than mc measured at the competitive output level. To avoid overstating the monopoly surcharge, one should (if possible) use mc evaluated at the competitive output level. William Landes and Richard Posner took advantage of the relationships between elasticity measures and price to arrive at a surprisingly useful version of the Lerner index.10 One basic result of the theory of monopoly is that a firm with market power maximizes profit when p(1 - 1 edj ) = mc Equivalently,

( p - mc) p = 1 edi Assume the firm has a large market share, si, and that there is a competitive fringe, denoted by the subscript j. One can show that edi = edm s i + esj (1 - s i ) s i Thus, it follows that

( p - mc) p = s i [edm + esj(1 - s i )] This equation tells us that there are three determinants of market power.

9

10

Note that because the Lerner index measures the markup over marginal cost, it is a measure of exercised market power rather than potential market power. Under the assumption that a firm will exploit its potential power, this distinction should not be important. William M. Landes and Richard A. Posner, Market Power in Antitrust Cases, 94 Harvard Law Review 937 (1981). For an earlier article deriving similar measures of market power, see David Encaoua and Alexis Jacquemin, Degree of Monopoly, Indices of Concentration and Threat of Entry, 21 International Economic Review 87–105 (February 1980).

236

Power

1. Market share, si: The greater is market share, the greater is market power. 2. Market demand elasticity, edm: The greater is the market demand elasticity, the greater will be the elasticity of demand facing the individual firm, and the smaller will be the individual firm’s market power. 3. Fringe supply elasticity, esj: The larger is the fringe supply elasticity, the larger is the elasticity of demand facing the individual firm, and the smaller will be the firm’s market power. An additional consideration is ease of entry. It seems appropriate to treat ease of entry as another factor determining the elasticity of fringe supply. But that assumes entrants enter on a small scale. If entry occurs on a large scale, with sufficiently large probability, then this approach would understate the price-restraining effect of entry. In general, we can identify three determinants of market power: market share, demand-side substitution, and supply-side substitution. The Landes-Posner formula indicates that there may be cases in which market share is great, but because demand- and supply-side substitutability is also great, the firm’s market power is negligible. The key implication of the Landes-Posner formula is that proper identification of the relevant market is irrelevant. With a too narrowly chosen market, the firm’s share is “too high,” but the market power formula will also include elasticity measures that are “too high,” because they come from an excessively narrow market. The high market share estimate and high elasticity estimates will offset each other. To see this, note that as you consider larger markets as candidates for “the relevant market,” three things will happen, which taken together suggest the Lerner index remains roughly constant: (1) the firm’s market share, si, will fall; (2) there will be fewer substitutes for the market commodity (e.g., as you move from “Ford cars” to “cars” you will find fewer substitutes), and this implies that edm will fall; (3) there will be fewer supply side substitution possibilities, so esj falls. Janusz Ordover, Alan Sykes, and Robert Willig introduced an extension of the version of the Lerner index developed by Landes and Posner.11 Using the Cournot model of oligopoly (see Chapter 1), they showed that 11

See Janusz A. Ordover, Alan O. Sykes, and Robert D. Willig, Herfindahl Concentration, Rivalry, and Mergers, 95 Harvard Law Review 1857 (1982). Suppose the firm has market power in both the input and output markets. Suppose there is one input, l, with respect to which the firm has monopsony power. In this case, the modified Lerner index is equal

III. Substitutability and the Relevant Market

237

( p - mc) p = s i (1 + k i ) [edm + esj(1 - s i )] , where ki is a measure of interdependence in competition. If ki = 1, then the firms are following the same policy with respect to price and quantity. In this case, the Lerner index understates the firm’s measure of market power by 100 percent. If ki = -1, then the firms are following contrary policies. Any reduction in output by the large firm is met by an immediate expansion in output (or reduction in price) by its competitors. In this case, the correct measure of market power is zero, whatever the dominant firm’s market share. The Ordover, Sykes, and Willig formula demonstrates how competitive culture can be incorporated into the assessment of market power. How useful are these formulas? It is difficult to obtain accurate measures of market demand and fringe supply elasticities. As commentators have noted, there are many real world problems that get in the way of using these formulas. One is product differentiation. No doubt a revised formula could take this into account, but to apply the new formula would require some measure of the degree of substitutability among brands. Another problem is dealing with foreign competition, and the influence of tariffs and quotas on the supply responses of these manufacturers. One could generate a long list of factors not incorporated in the formulas that ought to receive weight in any effort to estimate of market power. The formulas presented in this section may be useful more for their implications for the qualitative rather than quantitative assessment of market power. The formulas identify the major determinants, and indicate how the courts should use them in translating a market share estimate into a statement about market power.

iii. substitutability and the relevant market: cellophane The Supreme Court confronted some of the issues raised in the preceding discussion in the Cellophane case.12 The market in flexible wrapping

12

to (ed + el)/(ed(el + 1), which can easily be shown to be greater than the standard Lerner index. The dual-market-power measure can be used to reach conclusions similar to those in the Landes and Posner analysis. See Keith N. Hylton and Mark Lasser, Measuring Market Power When the Firm Has Power in the Input and Output Markets, in Economic Inputs, Legal Outputs: The Role of Economists in Modern Antitrust, Fred S. McChesney, ed., 131–9 (Wiley, 1998). United States v. E.I. du Pont De Nemours & Co. (Cellophane), 351 U.S. 377 (1956).

238

Power

paper was highly differentiated. During the relevant period, du Pont produced almost 75 percent of cellophane sold in the United States, but cellophane constituted less than 20 percent of all flexible packaging material sales. The lower court found that the relevant market included all flexible wrapping materials, and that du Pont could not be a monopolist. The issue before the Court was whether the defendant had market power. The Court held that assessing market power in a market with differentiated products requires consideration of evidence on cross-elasticities of demand. The reason is straightforward: every producer of a differentiated product has a monopoly in the market of its own product. It would be absurd to hold that power over the pricing of one’s own differentiated product constitutes monopoly power. The cross-elasticity of demand is the percentage change in the quantity of A demanded that results from a 1 percent change in the price of B. Elasticity is a number of interest to economists because it provides a unit-free measure of the responsiveness of demand. A measure of the slope of the demand curve would not serve this purpose because slope is measured in the terms of the units of the commodity of interest. The slope measure can vary considerably depending on whether tons, pounds, ounces, or some other measure describes the volume of output. Elasticity avoids these problems. Elasticity is easy to define in words, but not easy to calculate. First, calculating the cross-elasticity requires proper specification and identification of the demand for the product of interest. Identification of the demand schedule is an issue because it is not easy to distinguish demand from supply shifts in a scatter-plot of price-quantity data. If all of the variables that influence demand also influence supply, it will be impossible to empirically distinguish demand from supply shifts. Assuming proper specification and identification of the demand schedule, the next question is precisely what elasticity measure is of interest. Economists use the term elasticity in two senses. One refers to the precise number; it measures the responsiveness of demand at a certain point in price-quantity space. Economists call this the measure of “pointelasticity” of demand. The other sense of the term refers generally to price-responsiveness along the relevant range of a demand schedule. If the courts rely on the former sense of the term elasticity in measuring price responsiveness, then a significant problem gets in the way of interpreting the measure. A basic result of the theory of monopoly is that the monopolist always produces at an elastic portion – that is, a point at

IV. Multimarket Monopoly and the Relevant Market

239

which the elasticity of demand exceeds one – of the demand schedule. Thus, a relatively high point elasticity of demand could reflect the fact that the firm is pricing at the monopoly level. An analyst who interpreted a relatively high cross-elasticity (cross-point-elasticity) estimate as evidence of lack of market power could be wrong, a problem known in the literature as the Cellophane fallacy. The fallacy is an issue in all efforts to use evidence on the price-responsiveness of consumers to make inferences on market power. For example, the Justice Department Guidelines approach to determining the relevant market, which requires an assessment of the responses of consumers to a “small but significant and nontransitory” price increase, is vulnerable to this criticism. Did the Cellophane opinion require lower courts to gather estimates of the cross-elasticity? No. The Court itself relied on historical evidence of functional interchange between cellophane and other types of flexible wrapping material, and found that in view of this evidence, cellophane fit into the broader market of flexible wrapping material. Historical evidence on functional interchange provides some information on the relevant cross-elasticity. Because historical evidence is cheap to gather, and not subject to the interpretive problems of point elasticities, it is probably the most efficient method of assessing cross-elasticity evidence.

iv. multimarket monopoly and the relevant market: alcoa Judge Learned Hand ran into a problem in the course of defining the relevant market in Alcoa:13 What products should courts consider in competition with the output of a monopolist when that monopolist makes more than one item? Alcoa produced two types of aluminum: virgin and fabricated. Virgin refers to aluminum produced and pressed into easily transported shapes that another manufacturer could fabricate into some intermediate or end product. Fabricated refers to aluminum that Alcoa itself fabricated into intermediate and end use products. Alcoa was the sole domestic producer of virgin aluminum. As an additional feature complicating matters, the aluminum Alcoa sold today would return to the market later as scrap aluminum, to compete against its future production. 13

United States v. Aluminum Co. of America, 148 F.2d 416 (1945) (discussed in Chapter 10).

240

Power

Judge Hand had to decide whether to include secondary and fabricated aluminum in the relevant market. Excluding these markets would result in a market definition that took virgin aluminum produced by Alcoa, and that imported into the United States as the relevant market. Consider the virgin versus fabricated question first. Were these products in competition? To answer this, consider Alcoa’s incentives. Alcoa would produce an additional unit of virgin aluminum only if its contribution on the margin to revenue, taking into account the diversion into fabrication, exceeded its contribution on the margin to cost. The company would choose the fraction diverted to fabrication to equalize the marginal contribution to total profits of each type of aluminum. Thus, if an additional unit of fabricated aluminum yielded a higher profit on the margin than would an additional unit of virgin aluminum, Alcoa would divert more of its production into fabricated. Of course, the marginal profit gained by producing virgin is itself a function of its value to other fabricators of intermediate and end use products. If Alcoa could maximize its profits by selling only virgin aluminum, presumably it would have done so. Under these conditions, it may not be appropriate to treat competitors in the market for fabricated aluminum as real competitors. Many of them existed only because Alcoa, in effect, allowed them to exist by meeting their demand. However, some of the supply of fabricated aluminum was created through the use of imported virgin aluminum. Thus, it would be incorrect to say that Alcoa’s production of virgin aluminum served as the sole source of supply for the fabricated market. Consider the market in secondary aluminum. Knowing that aluminum produced today would return in the form of scrap, and put downward pressure on prices tomorrow, Alcoa had an incentive to further restrain today’s production. How strong was this incentive? Let MR1 be the addition to this year’s revenue that results from a small increase in this year’s output.14 Let MR2 be the reduction in next year’s revenue that results from an increase in this year’s output, MR2 < 0 (because an increase today leads to an increase in the secondary supply in the next year). If r is the annual rate of interest, Alcoa would expand this year’s output until MR 1 + MR 2 (1 + r) = MC 14

This discussion follows Jean Tirole, The Theory of Industrial Organization 79–80 (MIT Press, 1989).

IV. Multimarket Monopoly and the Relevant Market

241

The existence of secondary market aluminum has two effects on Alcoa’s production decision. First, since secondary market aluminum is available today, the demand for Alcoa’s current output is smaller than it would otherwise be. This is equivalent to the case of a firm with market power that faces fringe competition, which we examined earlier in this chapter. Since the existence of fringe competition reduces the firm’s ability to set price above the competitive level, the “current period effect” suggests that price will be closer to the competitive level. Second, since aluminum produced this year reduces next year’s prices, Alcoa has an incentive to restrain this year’s production further, which suggests Alcoa will set a higher price, other things being equal, than it would in the absence of such a second period effect. Note that if the interest rate were extremely large, the second period effect could become negligible. Jean Tirole shows that in the long run the firm’s price-cost markup (Lerner index) is lower than “the one chosen by a monopolist in an industry without [a secondary market],”15 which implies consumers benefit in the long run from the existence of secondary market aluminum. However, the Lerner index remains positive (implying market power) unless the secondary market grows so large relative to the primary market that the monopolist’s current output has a negligible effect on the market price.16 Since the market was growing over the period of examination of Alcoa’s conduct, the firm probably retained market power. What does all of this suggest? First, it indicates that the existence of secondary market aluminum reduced Alcoa’s power to set price above the competitive level. However, Alcoa probably did retain some power to set price above the competitive level, and in light of this, it may not be correct to treat the entire supply of scrap aluminum as competing. Alcoa had chosen its most profitable output in light of the constraints placed by a predictable supply of secondary aluminum. The reasoning here provides some support to Judge Hand’s conclusions in Alcoa. Hand treated imported virgin aluminum as the only competing source and used Alcoa’s total virgin aluminum production as the measure of Alcoa’s output. Using this definition of the relevant market, he found that Alcoa’s share was above 90 percent. Because it is likely that Alcoa retained some market power in spite of the existence of a secondary market, this discussion provides mild support for Hand’s 15 16

Id. at 80. Id.

242

Power

decision to exclude secondary market aluminum as a competing source of supply. However, Hand’s decision to ignore all competition in the fabricated aluminum market seems questionable. The treatment of secondary market aluminum in Hand’s Alcoa opinion raises the closely related problem of measuring market power when the monopolist produces a durable good. The question was examined by Ronald Coase in an influential article titled “Durability and Monopoly.”17 Coase concluded product durability reduces the market power of the monopolist, and in certain cases could lead to an equilibrium in which the monopolist prices competitively. Since this conclusion is in stark contrast to Judge Hand’s in Alcoa, it is worthwhile to consider the reasoning behind it. In Coase’s model, the monopolist produces a durable product and consumers have a choice to purchase the product today or in a later period. Once a consumer has purchased the product, he no longer needs it, so his demand is no longer reflected in the market demand schedule. Thus, if all high-valuing consumers buy the product early, the monopolist will have an incentive to lower the price to make additional sales in later time periods. However, if consumers knew this at the start, they would postpone their purchases to later time periods, at the reduced price. The larger the share of consumers willing to postpone payment, and the larger the number of periods, the greater the monopolist’s incentive to set price at the competitive level. The aluminum market examined in Alcoa differs in important ways from the market for a durable product. In Coase’s durable-good problem, high-valuing consumers drop out of the market after making an initial purchase. In the aluminum market, high-valuing consumers return to the market regularly. However, there are some features shared by the two markets. In Judge Hand’s view of Alcoa, the product’s durability meant that scrap aluminum existed in the secondary market in later time periods as an alternative to virgin aluminum. Hand’s view did not assume, it seems, that consumers would change the timing of their purchases in order to take advantage of the secondary market supply. Under Hand’s view, it may be appropriate to discount secondary market aluminum as a competing source of supply, especially if, as was true in Alcoa, 17

Journal of Law & Economics, vol. 1, 143–9 (April 1972). For a more recent examination, see Jeremy Bulow, Durable-Goods Monopolists, 90 J. Pol. Econ. 314–32 (April 1982); Dennis Carlton and Robert Gertner, Market Power and Mergers in Durable-Goods Industries, 32 J. Law & Econ. S203 (1989); John Shepard Wiley, Jr., Eric Rasmusen, J. Mark Ramseyer, The Leasing Monopolist, 37 UCLA L. Rev. 693–731 (1990).

V. Measuring Power

243

the primary market is growing. However, if one assumes, following Coase, that some consumers strategically alter the timing of their purchases, then Coase’s analysis applies to Alcoa, and it probably would be incorrect to exclude secondary market aluminum entirely as a competing source of supply even in the case in which the primary market is growing rapidly.

v. measuring power: guidelines Suppose a court has accurately determined the relevant market, say by following the approach of the Justice Department Guidelines. What market share is required to establish “monopoly power”? No one has provided a rigorous answer to this question. The courts have developed rough guidelines. The most important guidelines were stated in the Alcoa opinion. There, Judge Hand said that a 33 percent market share was insufficient to find monopoly power, 66 percent was possible though doubtful, and 90 percent or more was sufficient.18 Judge Wyzanski held that United Shoe had monopoly power when its market share was 75 percent.19 Antitrust courts often refer to these findings as benchmarks for assessing monopoly power. In light of the discussion in this chapter of the determinants of monopoly power, the tendency of courts to adhere to numerical guidelines established in earlier Section 2 cases is disappointing. If there is any message this chapter should convey, it is that market share by itself says very little about the degree of market power possessed by a firm.20 Recall that the three general determinants of market power are market share, demand-side substitutability, and supply-side substitutability. An analysis of market power that fails to take all three determinants into account is bound to be incomplete. 18 19

20

United States v. Aluminum Co. of America (Alcoa), 148 F.2d 416, 424 (1945). United States v. United Shoe Machinery Corp., 110 F.Supp. 295 (D.Mass. 1953), aff’d per curiam, 347 U.S. 521 (1954). Again, I should note that there is a literature showing that profitability is correlated with market share, see Mueller, supra note 3. However, market power, as defined in this chapter, depends on more than just market share.

12 Attempts

Because Socony makes questions of power and success irrelevant under Section 1, given proof of a conspiracy, the concept of attempt comes into play only under Section 2. An attempt to monopolize is a distinct violation of the Sherman Act, and has a distinct set of doctrines. The key difference between an attempt and a monopolization charge is that in the former, the defendant either did not succeed or the Court does not find significant evidence of success. Consider, for example, the monopolization charge in Griffith. The defendant in Griffith had not achieved monopoly status. However, the government charged the defendant with monopolization because of the evidence of success: the percentage of towns in which the Griffith circuit enjoyed a local monopoly had increased roughly 50 to 60 percent over the period in which the defendants had adopted their practice of block-purchasing films. Had there been no such evidence of success, the appropriate charge would have been attempt to monopolize.

i. the

SWIFT

formula and modern doctrine

The doctrinal formula applied in the area of attempts comes from Justice Holmes’s opinion in Swift & Co. v. United States.1 Justice Holmes described an attempt as conduct that closely approaches but does not quite attain complete monopolization plus a wrongful intent to monopolize. Thus, attempt requires proof of specific intent plus a dangerous probability of success. The cases below put flesh on this skeletal formula. 1

196 U.S. 375 (1905).

244

I. The Swift Formula and Modern Doctrine

245

Specifically, they tell us how one proves specific intent, and what the dangerous probability element requires. Only a few years after Swift, Holmes wrote another opinion on an attempt to monopolize charge, United States v. Winslow.2 Winslow dealt with a Sherman Act challenge to the merger that created the United Shoe Company. The companies that merged each had large (greater than 50 percent) market shares in their individual markets. However, the machines they produced did not compete against each other. The Court held that the merger did not violate the Sherman Act. Holmes referred to the merger as “an effort after greater efficiency,”3 and went on to say: [W]e can see no greater objection to one corporation manufacturing 70 percent of three noncompeting groups of patented machines collectively used for making a single product than to three corporations making the same proportion of one group each. The disintegration aimed at by the statute does not extend to reducing all manufacture to isolated units of the lowest degree. It is as lawful for one corporation to make every part of a steam engine, and to put the machine together, as it would be for one to make the boilers and another to make the wheels. Until the one intent is nearer accomplishment than it is by such a juxtaposition alone, no intent could raise the conduct to the dignity of an attempt.4

Thus, Winslow holds that a merger of noncompeting manufacturers, whatever their intent, falls short of presenting a dangerous probability of success. Although Winslow is a little noticed opinion, it contains an important message that – as we will see below – the Supreme Court has only recently come around to restating clearly. Winslow implies that the attempts test requires both specific intent and an objectively dangerous probability of success. Holmes suggested that he was unwilling to allow plaintiffs to satisfy the attempts test by substituting evidence on intent for ambiguous evidence on the probability of success. The modern doctrine of attempts begins with Lorain Journal Co. v. United States.5 From 1933 to 1948, Lorain Journal enjoyed a local monopoly in advertising and news dissemination. In 1948, WEOL, a radio station, received its license. Lorain Journal refused to accept advertising from companies using WEOL. This forced advertisers to choose, and many of them were reluctant to give up newspaper advertising in order

2 3 4 5

227 U.S. 202 (1913). Id. at 217. Id. at 218. 342 U.S. 143 (1951).

246

Attempts

to use the radio. The lower court found that Lorain’s conduct was an attempt to regain its pre-1948 monopoly of news and advertising. The Supreme Court had to decide whether Lorain’s conduct constituted an attempt to monopolize. The Court held that it did. The Court did not speak with clarity or in great detail about the specific intent requirement. It noted that if several local newspapers had combined to do what Lorain Journal did, they would have violated Section 1. Since Lorain Journal was a monopolist, the Court concluded, its activity violated Section 2. This is the “abuse standard” announced in Standard Oil, and provides one theory of what the courts require to prove specific intent. However, the Court’s finding of specific intent seemed to rely largely on the absence of a credible procompetitive justification.6 The finding of a dangerous probability of success followed from the fact that Lorain Journal had market power in the local advertising market and all signs suggested that if not for the injunction the Court issued, WEOL would have suffered losses until it had to leave the market. The defendant argued that it had a right to deal with whomever it wished. Recall that the Court had recognized such a right – a “personal” right to deal – in Eastern States (see Chapter 9). However, the Court said that this right is absolute only “in the absence of any purpose to create or maintain a monopoly.”7 This rule – that in the absence of an intention to create or maintain a monopoly one has the right to deal with whomever one wishes – is now known as the Colgate doctrine.8 Here, the defendant’s argument was rejected because the Journal’s actions were all part of an effort to maintain a monopoly. On the two major issues addressed in Lorain Journal – the evidentiary requirements of the specific intent test and the extent to which a 6

7

8

The defendant’s justification appears in footnote 8 of the opinion. The footnote explains that the Journal saw its scheme as part of an implicit contract to protect business from foreign (i.e., outside Lorain) competition. The Journal said that it refused advertisers who competed with Lorain businesses, in order to protect itself in a manner consistent with the overall plan. This justification flies in the face of the Sherman Act. A policy of protectionism – that is, protection from competition, for no other reason than to retain local market power – is not a justification that the Sherman Act recognizes, see NCAA v. Bd. of Regents of University of Okla., 468 U.S. 85, 115–17 (1984) (discussed in Chapter 5). “In the absence of any purpose to create or maintain a monopoly, the act does not restrict the long recognized right of a trader or manufacturer engaged in an entirely private business, freely to exercise his own independent discretion as to parties with whom he will deal.” United States v. Colgate & Co., 250 U.S. 300, 307 (1919). For a detailed discussion of the Colgate doctrine, see Chapter 13.

I. The Swift Formula and Modern Doctrine

247

nonmonopolist is free to refuse to deal with others – the law pretty much remains as the opinion states. The specific intent prong of the attempts test remains a vague part of antitrust law. The cases generally suggest that the absence of a credible procompetitive justification is sufficient to establish specific intent to monopolize. In Union Leader Corp. v. Newspapers of New England,9 one of the leading cases, Judge Wyzanski referred to the presence of “unfair tactics” as conclusive evidence of the defendant’s specific intent to monopolize. However, Wyzanski limited this standard to cases in which “it is inevitable that only one competitor can survive,”10 or, in other words, natural monopoly settings. The unfair tactics taken as conclusive evidence of specific intent in Union Leader probably fall short of what most courts today would require to prove specific intent to monopolize. For example, one of the “unfair tactics” which conclusively demonstrated illegal intent was an effort on the part of the Union Leader Corporation to get local advertisers to give all of their advertising to the Union Leader Corporation’s newspaper (the Haverhill Journal). However, an effort to persuade advertisers to give all of their business to your newspaper is not the same as the refusal to deal in Lorain Journal.11 The practice of trying to get all of the business in a market is similar to that of negotiating a series of requirements or exclusivity contracts, which courts generally treat under a rule of reason test.12 In the background of these decisions, shifts in Section 2 monopolization doctrine (i.e., ordinary monopolization as opposed to attempt) have influenced the doctrine of attempts. The monopolization standard moved away from the ambiguous nonpassive acquisition standard of Alcoa toward a test of exclusionary intent suggested in Griffith. Thus, specific intent to monopolize, in the attempts context, now means specific intent to exclude competitors in a manner that is likely to result in monopoly power. As for the Colgate doctrine, the cases involving refusals to deal can be explained by applying the doctrine in a straightforward manner. Generally, a firm that does not have market power and refuses to deal with

9

10 11

12

180 F.Supp. 125 (D. Mass. 1959), aff’d in part and rev’d in part, 284 F.2d 582 (1st Cir. 1960), cert. denied, 365 U.S. 833 (1961). 180 F. Supp. at 148. Why? The evidence suggested that the refusal to deal in Lorain Journal was an effort by a firm with market power to block entry by a new rival, and the case did not suggest any efficiency justifications. See Barry Wright Corp. v. ITT Grinnell Corp., 724 F.2d 227, 236–39 (1st Cir. 1983) (discussion of reasonableness of requirements contracts).

248

Attempts

another need not worry of liability under an attempt to monopolize charge. Conversely, a firm that has monopoly power and refuses to deal with another cannot rely on the Colgate doctrine for protection.

ii. dangerous probability requirement The other significant prong of the attempts test requires proof of a dangerous probability of success. Since the Swift decision, the dangerous probability test requires some examination of market power. That inquiry can be divided into two issues. First, is it necessary to determine the relevant market, as in ordinary monopolization cases? The alternative is a lower standard that would permit a finding of a Section 2 violation even though the market had been defined narrowly, for example, in a manner that did not appropriately take into account demand- and supply-side substitution.13 The second question is whether the plaintiff must show that the defendant has market power, in the traditional sense of having a large market share. With respect to the first issue, the relevant market, courts have taken three approaches. One, an extreme position, holds that the definition of the market is unimportant, and that specific intent is the only relevant question. Another suggests that market definition is necessary, but that a narrow definition is permissible. The third approach requires proper definition of the relevant market, taking into account all supply- and demand-side substitution issues, as monopolization cases generally require. In Lessig v. Tidewater Oil Co.,14 the Ninth Circuit suggested that it would take the first approach, although it retreated from that position in later decisions holding that the courts must take market power into account in determining whether the defendant had specific intent.15 No other court has suggested that the proper definition of the relevant market is irrelevant, but some have permitted a narrow definition. 13 14 15

See Chapter 11. 327 F.2d 459 (9th Cir. 1964). See Philip Areeda and Louis Kaplow, Antitrust Analysis 618 (4th ed. 1988). The issue here arises in introductory criminal law courses and it is usually framed as follows: should possibility be taken into account in determining whether a criminal defendant had specific intent to commit a crime? Suppose a man walks into my classroom with a banana in his hand and says, “Hold your hands up, this is a stickup!” He may have had a specific intent to rob my class, but it was impossible. However, one could argue that he did not have a specific intent to rob my class because he must have known that it was impossible to credibly threaten harm with a banana.

II. Dangerous Probability Requirement

249

The case for proper determination of the relevant market in Section 2 litigation has been set out in Chapter 11. In general, proper determination of the relevant market guarantees that the courts will make findings of Section 2 violations only in cases where there is a credible claim of public harm. It is difficult to make a credible claim of public harm when the market has been defined too narrowly in an economic sense. If, for example, the government charged a manufacturer of a differentiated product with monopolization even though that product competed against one hundred other products with similar uses, it would be difficult to argue that the manufacturer’s pricing caused significant harm to consumers. Is there an argument for adopting a narrow definition of the market in attempt cases? Yes, and it runs as follows: because attempted monopolization requires a finding of specific intent, these cases will often involve conduct that has no procompetitive justification. And since the defendant’s actions are either (1) harmful to the public, or (2) harmful to a particular victim without providing a potential benefit to the public, the courts take only a small risk in adopting a narrow definition of the market. With respect to the second issue, the necessity of finding market power, the cases have suggested two approaches. Lorain Journal indicated that, ordinarily, a finding of a dangerous probability requires some proof that the defendant has market power, with market power demonstrated by the fact that the defendant has a high market share (say, greater than 65 percent). The exception is the case of a market with natural monopoly features, where one producer will eventually meet the demand for the entire market. The natural monopoly exception to the requirement of a substantial market share was established in Judge Wyzanski’s Union Leader opinion. Union Leader involved an effort by one newspaper to enter into competition with another. The Court found the entrant guilty of an attempt to monopolize even though the entrant had not acquired monopoly power, and had at best a marginal market share. The crucial feature was that the market was too small to support two newspapers. In such a market, success would have led to monopoly status, which is true generally of any market that has the features of natural monopoly. Thus, in a market where because of size or other characteristics unfair tactics would allow one firm to leapfrog over the other toward monopoly, the Court may find an attempt in violation of Section 2 even though the defendant does not have market power in the traditional sense.

250

Attempts

The Supreme Court clarified the dangerous probability test in Spectrum Sports Inc. v. McQuillan.16 The Court held that the dangerous probability element of the attempts test requires some examination of market power in the relevant market.17 The language of the Court in Spectrum Sports sends a clear signal that lower courts must make a proper determination of the relevant market in attempt cases. Narrow definitions of the market are inconsistent with the standard of Spectrum Sports. What does “some examination” of market power require? In the general case, the standard clearly requires proof of market power in the relevant market – consistent with the Court’s treatment in Lorain Journal. Since the Court did not reject the result in Union Leader, it seems safe to assume that the test can also be satisfied by proof that because of natural monopoly features, a dangerous probability of success exists even though the defendant does not have monopoly power according to traditional criteria.18 It follows that in the cases that lack natural monopoly features, the dangerous probability test now imposes the same evidentiary burdens on plaintiffs as does the monopolization test under Section 2.19 Above all, the clearest message from Spectrum Sports is a rejection of the theory expressed in Lessig that specific intent evidence could be used as a substitute for objective evidence of a dangerous probability of success. Put another way, Spectrum Sports rejects the notion that courts can trade off evidence on intent for evidence on objective dangerousness. This message was implicit in Holmes’s opinion in Winslow, discussed earlier in this chapter. Indeed, one could argue that Spectrum Sports merely restates a position the Supreme Court had taken in the Winslow opinion. The more important question is whether this is the right decision. Recall that there is a plausible argument for the Lessig theory: that acts 16 17

18

19

506 U.S. 447 (1993). Id. at 457. The Court had made this statement earlier in Walker Process Equipment v. Food Machinery & Chemical Corp., 382 U.S. 172, 177 (1965). One might argue that the natural monopoly exception is inconsistent with the Court’s rejection of an entirely intent-based test in Spectrum Sports. However, the “some examination” requirement of Spectrum Sports falls short of requiring proof of market power in every attempt case. “Some examination” requires the plaintiff to prove that the likelihood of success is high in the market in which the predatory behavior is occurring. Clearly, that likelihood can be high, even though the defendant has a small market share when the market is a natural monopoly. United States v. Grinnell Corp., 384 U.S. 563, 576–80 (1966) (see Chapter 10 for discussion of Grinnell monopolization test).

II. Dangerous Probability Requirement

251

showing specific intent generally have no procompetitive justification, so we should be prepared to condemn them whether or not the defendant has market power. The problem with this argument, the Court noted, is that it assumes a level of certainty that courts seldom have. The specific intent requirement, though probably superior on accuracy grounds to alternative monopolization tests (see Chapter 10), is still capable of being applied erroneously. Given this possibility, the objective dangerousness requirement upheld in Spectrum Sports should be understood as providing an additional measure that reduces the likelihood of a false conviction. But is it more desirable to reduce false convictions for attempted monopolization than false acquittals? The rationale in Spectrum Sports does not answer this question, but I think there is a good case to be made that it is. False convictions in attempted monopolization cases against defendants who do not have market power are quite likely to chill competition from aggressive new entrants. In view of the importance of entry in constraining oligopolistic pricing, the law should be reluctant to let this happen.

13 Vertical Restraints

This chapter examines the law and economics of vertical restraints, though the primary emphasis is on the law. In particular, I examine agreements covering resale prices, exclusivity, and marketing territories. Section 1 of the Sherman Act applies to these arrangements. However, for reasons that I will try to make clear below, the Section 1 doctrine governing vertical restraints is a separate and peculiar set of rules. The focus in this chapter is not so much on the economics of vertical restraints, a subject that has spawned a large literature,1 but on the underlying tensions and contradictions in the law governing vertical restraints.

i. resale price maintenance A. Law and Justifications Resale price maintenance refers to the efforts of a manufacturer to restrict the range of prices charged by a retailer of the manufacturer’s product. Typically, the manufacturer seeks to restrain price cutting. 1

For excellent (though slightly dated) surveys, see Frank Matthewson and Ralph Winter, The Law and Economics of Vertical Restraints 109–47, in The Law and Economics of Competition Policy (Frank Matthewson, Michael Trebilcock, and Michael Walker, eds., Vancouver, British Columbia: The Fraser Institute, 1990); F. M Scherer and David Ross, Industrial Market Structure and Economic Performance 541–69 (Boston: Houghton Mifflin Company, 3d ed. 1990). See also Andy C. M. Chen and Keith N. Hylton, Procompetitive Theories of Vertical Control, 50 Hastings L. J. 573 (1999).

252

I. Resale Price Maintenance

253

Why? The theories examined in this chapter offer several justifications, but before we consider them, we should pause to consider a problem: if the minimum resale price a manufacturer requires does not constrain the retailer, then it has no effect; and if it does constrain the retailer, then it will encourage him to turn away from the manufacturer’s product. Our first problem is figuring out why a manufacturer would want to constrain a retailer of its product. Assume the manufacturer deals with a large number of retailers. To the extent that a resale price restriction affects the manufacturer, it will do so only through altering the manufacturer’s revenue. Therefore, the manufacturer should prefer a restriction that maximizes its revenue. Lower prices at the retail level enhance the manufacturer’s revenue, other things being equal. That is, once the manufacturer sells the good to the retailer, it would prefer to see the retailer charge the lowest possible price, since that maximizes the quantity sold. Indeed, once the manufacturer has sold his output to retailers, he would be happy to see them drive the resale price to zero. This reasoning supports the notion that resale price maintenance plans are in fact price-fixing conspiracies. What kind of price-fixing conspiracy? One theory is that a cartel among manufacturers of an item may find that a resale price maintenance plan makes it easier to monitor compliance with the cartel’s price level.2 If retailers were free to cut prices, it would be difficult to determine whether price cuts resulted from cheating by a member of the cartel, or from the actions of a particular retailer. Under this theory, resale price maintenance is really a facilitating mechanism (see Chapter 8) for collusion among manufacturers. An alternative anticompetitive theory is that resale price maintenance permits retailers to fix prices among themselves, with the manufacturer serving as coordinator and enforcer of the cartel.3 Of course, it is not clear why a manufacturer would willingly participate in a price-fixing plan among retailers. There are other anticompetitive theories of resale price maintenance, and there are procompetitive theories as well. Since the economic literature on this subject is expanding, I will not try to provide an exhaustive treatment of the various theories. I will review the best known theories below and discuss them in connection with the law.

2

3

See, for example, Thomas W. Gilligan, The Competitive Effect of Resale Price Maintenance, 17 Rand J. Econ. 544, 546 (1986). See, for example, Herbert Hovenkamp, Federal Antitrust Policy, §11.2b, which discusses this as part of a broader survey of anticompetitive theories of resale price maintenance.

254

Vertical Restraints

1. Dr. Miles. Federal antitrust law in this area begins with Dr. Miles Medical Co. v. John D. Park & Sons Co.4 Dr. Miles manufactured medicines, using a secret formula, and entered into contracts with wholesalers and retailers that required compliance with minimum prices. Dr. Miles sued a wholesaler on the ground that the wholesaler obtained the medicine at “cut prices” by inducing others to breach their price agreements. The Court held that a manufacturer who sells his medicine to a wholesaler is not entitled to restrict its resale through interference with the purchaser’s pricing. “Having sold its product at prices satisfactory to itself, the public is entitled to whatever advantage may be derived from competition in the subsequent traffic.”5 This is understood as a per-se condemnation of retail price maintenance. The Court addressed two arguments from Dr. Miles: (a) that the restrictions are valid because they relate to goods manufactured under a secret process, and (b) that a manufacturer is entitled to control the prices on all sales of his or her own products. Dr. Miles’s first argument drew on an analogy with patent-antitrust law and the antitrust law governing the licensing of trade secrets. A patent holder may license another manufacturer to produce the patented item and may put restrictions on the price the licensee charges.6 Similarly, Fowle v. Park7 held that the manufacturer may include minimum price provisions in the contract licensing another to use his trade secrets in production. Dr. Miles’s argument ran as follows: a patent holder who restricts the prices the licensee charges is restricting the prices charged by someone who manufactures and sells the product. Shouldn’t it follow from this that a manufacturer who uses a secret process can restrict the prices charged by someone else who merely sells the product? The Court answered that in the cases of trade secrets and patent licensing, the manufacturer provides the process to the other party, while in this case Dr. Miles did not share the process, only the finished product, and the evidence did not suggest that the dealers would discover the secret manufacturing process as a consequence of selling the product. This distinction fails to provide any sense of an economic justification for the decision, but one can be provided. If I license another manufac4 5 6

7

220 U.S. 373 (1911). Id. at 409. See Bement v. National Harrow Co., 186 U.S. 70 (1902). The Court ruled that a patentee may enforce minimum price clauses in its licensing agreements, and this remains the rule regarding price restrictions in patent licensing contracts. 131 U.S. 88 (1889).

I. Resale Price Maintenance

255

turer to make a product that I have patented, I face the risk that the new manufacturer will have lower production costs than I do. He could undercut my price and take the market away. Of course, if the licensee has lower production costs, then it would be efficient for the licensee to be the sole producer. But patent law does not aim for this result; otherwise, the government would award patents and auction the right to produce to the highest bidder. A patent allows the holder to exclude others from producing the item covered by the patent. Why, then, would a patent holder wish to remain a producer rather than cede the entire production enterprise to a lower cost licensee? It is not hard to generate reasons. One is the difficulty of enforcing any agreement to split rents after the patent holder has exited the market. If the patent holder charges an up-front fee, there would be no need to monitor the agreement. But an up-front fee may be prohibitively expensive. If, on the other hand, the patent holder charges a fee that depends on the amount produced by the licensee, then the agreement has to be monitored. Monitoring is costly and could be difficult. Moreover, how can the patent holder maintain a credible threat to revoke the license if the licensee is the sole producer? As an alternative, the patent holder could remain in the market as a producer, and charge either an up-front fee to the licensee or a per-unit fee, and allow the licensee to price in the manner that he desires. But the licensee may undercut the patent holder, leaving the holder with no market. A large up-front fee merely increases the risk that a licensee will undercut the patent holder ex post, and a per-unit fee brings costly monitoring problems. Thus, in order to take advantage of the efficiencies of licensing production to a lower-cost producer and simultaneously ensure its survival as a producer of its own patented product, a patent holder may want to restrict the pricing decisions of a licensee. From an ex post perspective, this is not the most efficient arrangement. The most efficient arrangement (ex post) would allow the low-cost licensee to fill the entire market demand.8 However, from an ex ante perspective this is consistent with the aim of patent law, which is to encourage inventors by providing a

8

Indeed, the best arrangement would be to promise exclusive control to the inventor, and then deny it ex post, allowing any rival to produce the good. That would encourage inventors and still allow the good to be distributed at the lowest cost. Of course, the problem with this arrangement is that it could be carried out only once. Inventors would soon learn that patents are worthless.

256

Vertical Restraints

period in which they control through the power of exclusion the production and sale of the patented item. The economic case for permitting price restrictions in patent licenses does not apply in the case of a dealership agreement. The patent holder who wishes to protect his initial investment can do so by charging a price to the dealer that guarantees the patent holder the necessary rent (i.e., revenue sufficient to recoup development costs). The only reason for placing restrictions in order to protect the patent holder’s incentives would be if it were impossible, for some reason, for the patent holder to protect himself by charging an adequate price to the dealer. But if the patent holder cannot charge an adequate price to the dealer, then how is it that he can charge an adequate price to the consuming public? Further, how can the dealer “steal the market” from the patentee, and what incentive would the dealer have to try this? The dealer, by hypothesis, cannot steal the patentee’s market, and should obviously have no incentive to ruin the patentee. Now consider the second argument of Dr. Miles. The Court responded that a general restraint on alienation is ordinarily invalid. It then noted exceptions; for example, the sale of goodwill, or a trade secret. Dr. Miles didn’t fall in any of the exceptions. The Court asked what justification Dr. Miles offered for the restraint, and found only an anticompetitive justification – a desire to fix prices and to prevent competition. At least this was the Court’s view of the assertion in Dr. Miles’s complaint that a standard retail price was desirable and that sales at less than the fixed prices resulted in confusion and damage. The Court said that the advantage of established retail prices primarily concerns the dealers. The enlarged profits . . . go to them and not to the complainant. It is through the inability of the favored dealers to realize these profits, on account of the described competition, that the complainant works out its alleged injury.9

In other words, the Court adopted the theory in Dr. Miles that retail price maintenance schemes are essentially cartels designed to favor a group of dealers, with the manufacturer serving as coordinator and enforcer. The Court’s interpretation of Dr. Miles’s justification brought the contracts within the class condemned by the Sherman Act. The treatment of this last issue is the most troubling part of the opinion. As Holmes noted in his dissent, the case was before the Court 9

220 U.S. at 407.

I. Resale Price Maintenance

257

on demurrer – that is, on a finding by the lower court that the plaintiff had no cause of action, taking his factual allegations as valid. Thus, the majority held that if the Court accepted the plaintiff’s allegations as true, the suit fails because the resale agreements fall within the class condemned by the Sherman Act. But this was the issue. The Court had not, before Dr. Miles, considered the application of the Sherman Act to resale price agreements. While the Court was correct in distinguishing such an agreement from a covenant not to compete ancillary to a sale of goodwill, it is also fair to say that the agreement is not the same as a naked horizontal price-fixing agreement. It was therefore the kind of case that ordinarily merits some consideration of reasonableness justifications. In any event, the law has been settled since Dr. Miles. Over the same period, the reasonableness justifications for resale price maintenance have filtered their way to the Court from various directions. We will see below that the Court has essentially accepted all of the reasonableness justifications that might have been asserted by Dr. Miles. However, the rule of Dr. Miles remains the law. How long this rather strange state of affairs will continue is an open question. 2. Justifying Resale Price Maintenance. Let us consider the arguments for and against the rule of Dr. Miles. The Court in Dr. Miles accepted the theory that retail price maintenance agreements are essentially dealer cartels coordinated by the manufacturer. Of course, one problem with this theory is figuring out why a manufacturer would have an incentive to go along with such a cartel. After all, once the manufacturer sells its product to the retailers, it would prefer to see them resell the product at the lowest possible price in order to maximize sales. However, suppose the dealers know a great deal more than the manufacturer about local demand conditions. In this case, it may be a rational strategy for the manufacturer to set up an arrangement that enables the dealers to exploit local demand conditions to get the best price. In exchange, the manufacturer charges a higher wholesale price to the dealers, which effectively allows it to take a share of the local monopoly profits of the dealers. This theory suggests that retail price maintenance agreements may be profitable to the manufacturer and harmful to consumers. There are two sets of arguments against the Dr. Miles rule. First, there are good legal arguments for allowing resale price agreements under the Sherman Act. The manufacturer could retain ownership of the goods by expanding vertically into retail, or by consigning the goods to dealers. In either of these cases, the rule of Dr. Miles would not apply. Why the

258

Vertical Restraints

antitrust laws should be concerned about resale prices in the nonagency context and totally unconcerned in the agency context is hard to see. What constrains the manufacturer from charging a sky-high price? Competition from other goods. At least before the Dr. Miles opinion, nothing in the legislative history or the case law suggested that Congress intended the tentacles of the Sherman Act to promote competition in every nook and cranny of the economy. Holmes made this point in his dissent, and noted very sketchily that benefits might exist for both the manufacturer and the public from restricting price competition among retailers. Second, there is a good reasonableness justification for RPM: the prevention of “free riding” by price-cutting dealers.10 Suppose, for example, dealers need to provide presale advice on how to use the product, or how to avoid injuring yourself with the product. Suppose retailer A provides this information and competes with retailer B, across the street. B’s best response is not to provide information, and to undercut A on price. Customers will go to A for presale information, and to B to purchase. Under these conditions A could not last long as a retailer. Do customers really get the good more cheaply from B? Only if A continues to provide information services. Once A leaves the market, the portion of the cost of the good borne by A will now shift to consumers. They purchase from B without presale advice. As a result, they make bad decisions, get injured by the good, or use it an unproductive way. Because these are real costs associated with the good’s consumption, there is no reason to believe that consumers get the good more cheaply from B if they purchase it without presale advice. One counterargument holds that if the provision of presale service is important, the manufacturer could deter free riding by charging a higher price for the good to dealers who refuse to provide presale service. But this generates a costly monitoring and enforcement problem for the manufacturer. It is difficult for the manufacturer to know with certainty in advance whether a given dealer will provide presale service. If a dealer promises to provide presale service and the manufacturer later discovers that the dealer did not provide the service, threatening to charge a higher price for the item in the future may not be an effective method

10

Lester G. Telser, Why Should Manufacturers Want Fair Trade? 3 J. Law & Econ. 86–105 (October 1960); Victor Goldberg, The Free Rider Problem, Imperfect Pricing and the Economics of Retailing Services, 79 Northwestern Univ. L. Rev. 736–57 (November 1984).

I. Resale Price Maintenance

259

of disciplining the dealer.11 In addition, if dealers who refuse to provide presale service compete with dealers who provide such service, charging a higher price to the former group may violate the Robinson-Patman Act. The free-riding argument can be extended in a number of directions. Consider a dealer who has a reputation for selling high-quality goods. The dealer’s decision to carry the good is a certification of quality.12 But then other dealers who do not certify quality (i.e., have not invested in a reputation for carrying high-quality merchandise) may undercut the high-quality seller unless the manufacturer can restrain price competition. A desire on the manufacturer’s part to maintain a high degree of geographical dispersion may justify RPM.13 For goods that require relatively frequent postsale service, maintaining several dealers capable of providing service after sale may maximize the manufacturer’s customer base. However, a dealer who skimps on pre- and postsale service will be able to offer lower prices. Consumers may be willing to drive long distances to purchase a car or computer at a cheap price. However, they would feel inconvenienced by having to drive long distances for every service problem that arose. An awareness on the part of consumers that it may be difficult to find high-quality, postsale service locally would reduce their demand for the good. The dealers themselves might realize this, but the incentive to undercut could remain a powerful force, preventing the adoption of the best arrangement. As these examples suggest, there are plausible efficiency justifications for RPM. I have focused on versions of the free-rider argument, but the more general problem is the divergence between the incentives of dealers and those of manufacturers.14 RPM may serve as an 11

12

13

14

Why? One problem is that the threat to charge a higher price in the future may not be credible. After all, both the dealer and the manufacturer know that the manufacturer would prefer that the dealer provide presale service, which is costly. Charging a higher price is a penalty that inflicts pain on both the dealer and the manufacturer. Also, consider that if the probability of detection is less than one, the manufacturer may need to impose a harsh penalty in order to induce compliance, which only makes the credibility problem more severe. Howard P. Marvel and Stephen McCafferty, Resale Price Maintenance and Quality Certification, 15 RAND J. Econ. 346–59 (Autumn 1984). J. R. Gould and L. E. Preston, Resale Price Maintenance and Retail Outlets, 32 Economica 302 (1965); Patricia B. Reagan, Resale Price Maintenance: A Reexamination of the Outlet Hypothesis, 9 Res. Law & Econ. 1 (1986). Indeed, the literature on RPM has expanded well beyond the free rider justification. See G. F. Mathewson and R. A. Winter, An Economic Theory of Vertical Restraints, 15

260

Vertical Restraints

effective method of aligning the incentives of dealers with those of manufacturers. If RPM is such a sensible way of selling goods that require presale service, why doesn’t it appear in all markets that provide such service? Consider, for example, the market for automobiles. The typical relationship between a car manufacturer and dealer is an exclusive territorial assignment. One does not see rigorously enforced resale price maintenance agreements as the norm. Indeed, one sees the opposite: transaction prices for many cars are established through one-on-one haggling between the buyer and the dealer. We see this even though presale service is quite important in the car market. Consumers need a great deal of advice about what car to purchase and the various add-on features. What explains this? While an answer is beyond the scope of this discussion, one reason for the convention observed in the car market is that the consumers in that market have different informational needs. Some spend a lot of time studying cars and need little advice, while others need to have their hands held by the dealer from the time they walk into the store until they leave. A resale price maintenance scheme would effectively require informed consumers to subsidize the information purchases of uninformed consumers. The information charge, which operates as a tax on informed consumers, could drive some informed consumers away from the car market, or toward cars that were sold without presale service. Perhaps the presence of a large number of informed buyers has led car dealers toward a system that enables the informed consumer to avoid paying for information. More generally, resale price maintenance may not be the best way to finance the provision of presale service to customers in all settings.15 The joint costs (i.e., costs to consumers and manufacturers) of a resale price

15

RAND J. Econ. 27–38 (Spring 1984); Richard E. Romano, Double Moral Hazard and Resale Price Maintenance, 25 RAND J. Econ. 455–66 (Autumn 1994); Greg Shaffer, Slotting Allowances and Resale Price Maintenance: A Comparison of Facilitating Practices, 22 RAND J. Econ. 120–35 (Spring 1991); Benjamin F. Blair and Tracy R. Lewis, Optimal Retail Contracts with Asymmetric Information and Moral Hazard, 25 RAND J. Econ. 284–96 (Summer 1994); Martin K. Perry and David Besanko, Resale Price Maintenance and Manufacturer Competition for Exclusive Dealerships, 39 J. Industrial Econ. 517–44 (September 1991); Benjamin Klein and Kevin M. Murphy, Vertical Restraints as Contract Enforcement Mechanisms, 31 J. Law & Econ. 265–97 (1988). For a diagrammatic analysis of the welfare implications of resale price maintenance, see F. M. Scherer and David Ross, Industrial Market Structure and Economic Performance 542–8 (Boston: Houghton Mifflin Company, 3d ed. 1990).

I. Resale Price Maintenance

261

maintenance plan may outweigh the benefits, and where this is the case, one should expect to observe alternative methods of financing the provision of presale service.

B. Scope I noted that a patent holder can restrict the prices that a licensee of the patented product charges. Suppose the patent holder tries to restrict purchasers of the good from reselling at certain prices? In Bauer & Cie v. O’Donnell,16 the Court held that such an agreement violates the Sherman Act. Alternatively, one can say that the patent holder’s rights do not include control over the trade of a patented item. The economic justification for this holding is the same as that offered above for the Court’s holding on the secret-process issue in Dr. Miles: the patent holder sells the product, not the process, to a purchaser. Presumably the patent holder could protect the stream of rents the patent guarantees by charging an appropriate fee to the purchaser of the patented item. If the purchaser chooses to resell at a loss, his choice to do so has no implications for the rights protected by patent law. Note that this implies that patent law does not allow a producer to monopolize the trade in his patented product. It allows him to prevent others from producing it, but that is the limit of his entitlement.17 Bauer & Cie makes clear that the existence of a patent provides no general exception to the rule against RPM. The existence of a copyright also fails to provide an exception: in Bobbs-Merrill Co. v. Straus,18 the Court held that a retailer who had sold a copyrighted book for less than the minimum price specified by the publisher did not infringe the copyright, in spite of notice that the publisher deemed such a sale an infringement.19 16 17

18 19

229 U.S. 1 (1913). No doubt the patent holder could earn more money by controlling the trade of a patented product. If the good can be sold in the secondary market, then a patent holder who had the right to control trade in after-markets could prevent used products from competing against new. However, the law does not give patent holders this power, even though it would be an additional spur to research and development. 210 U.S. 339 (1908). In Straus v. Victor Talking Machine Co., 243 U.S. 490 (1917), a license notice attached to each patented machine stated restrictions on retail prices. The facts combined those of Bauer & Cie (patented product) and Bobbs-Merrill (general notice). The Court ruled the restriction invalid.

262

Vertical Restraints

In Albrecht v. The Herald Co.,20 the Court extended the rule against vertical price fixing from minimum to maximum prices.21 However, in State Oil Company v. Khan22 the Supreme Court overruled Albrecht, holding that rule of reason analysis applies to maximum resale price maintenance. Because the manufacturer and the consumer both have an interest in low retail markups, one should expect to observe maximum resale price agreements in industries where dealers may have market power. By prohibiting such agreements, the relatively short-lived rule of Albrecht worked against the interests of consumers.23

ii. vertical nonprice restraints In this section, I consider nonprice restrictions on retailers. An example is a requirement that a dealer sell only to customers within a specified geographic market. Topco24 held that in the context of horizontal relationships, courts should treat territorial division agreements the same as price-fixing agreements. Should this presumption of illegality also apply to nonprice restrictions in the context of vertical relationships?

A. Exclusivity Agreements One common form of nonprice restriction in the vertical context is an exclusivity agreement, that is, where a manufacturer appoints a single 20 21

22 23

24

390 U.S. 145 (1968). The Albrecht decision was questionable when handed down because the facts suggested that the manufacturer (publisher) was attempting to prevent a local distributor from taking advantage of an exclusivity agreement by charging monopoly prices. 522 U.S. 3 (1997). Before the Khan decision, there was at least one warning that Albrecht would be jettisoned. The Court narrowed the scope of the rule against vertical maximum price restrictions in Atlantic Richfield Co. v. USA Petroleum Co. (ARCO), 495 U.S. 328 (1990). USA Petroleum claimed that a maximum price-fixing agreement between ARCO and its dealers caused it to lose a substantial amount of business. The Court held that a rival dealer complaining of a vertical maximum price agreement between a manufacturer and its own dealers could not satisfy the antitrust injury requirement unless it could show that the maximum price agreement resulted in predatory price levels. Although ARCO did not clearly address the facts of Albrecht, the Court did suggest in a number of passages that a plaintiff who complained of maximum price fixing would have to show one of the following: (1) that the prices were “fixed too low for the dealer to furnish services . . . for which consumers desire and for which they are willing to pay,” (2) that the prices limited “the ability of small dealers to engage in nonprice competition,” (3) or that the “scheme tends to acquire all of the attributes of an arrangement fixing minimum prices.” United States v. Topco Associates, 405 U.S. 596 (1972). For discussion, see Chapter 6.

II. Vertical Nonprice Restraints

263

dealer to handle its product.25 The typical complainant in such a case is a would-be dealer in the manufacturer’s product. The terminated dealer has various Section 1 and Section 2 claims available. The Section 2 theories are: (1) that the appointed dealer has monopolized or attempted to monopolize the retail market, or (2) that the manufacturer and appointed dealer conspired to monopolize the retail market in the relevant product. The Section 1 theories are: (1) that the agreement to terminate one dealer in order to give an exclusive dealership to another constitutes a concerted refusal to deal, and (2) that the contract itself is an unreasonable restraint of trade. The difference between the Section 1 theories is slight; the first suggests that the parties have agreed to boycott one competitor, while the second points to the agreement itself as an unreasonable restraint. The Section 2 claim must meet the hurdle established in Cellophane.26 If the manufacturer faces competition from makers of similar items, it is likely that the manufacturer’s product has close substitutes. Assuming substitutability, a manufacturer of a single brand probably would not possess sufficient market power to warrant a Section 2 violation. Consider, for example, the producer of one type of car, who arranges with one dealer to supply the car in a certain geographic market, at the same time excluding other potential dealers in that market. Unless the specific car produced would not be considered a substitute for other cars, which is unlikely, a finding of market power would be hard to defend. The same observation disposes of the attempt to monopolize claim. Spectrum Sports27 held that the dangerous probability component of the attempts test requires market power in most cases (those not presenting natural monopoly conditions), and it does not appear here for the reasons already given. The conspiracy to monopolize theory fails also. Colgate28 permits a manufacturer not possessing or attempting to acquire monopoly power to deal with whomever it wishes. However, one might argue that because 25

26

27

28

For an economic analysis of exclusive dealing in the agency context, see B. Douglas Bernheim and Michael D. Whinston, Exclusive Dealing, 106 J. Pol. Econ. 64–103 (1998). Bernheim and Whinston’s conclusions for the most part suggest that exclusive dealing in the agency setting will not serve as a method of achieving monopolization through foreclosure. United States v. E. I. duPont de Nemours & Co. (Cellophane), 351 U.S. 377 (1956). For discussion, see Chapter 11. Spectrum Sports, Inc. v. McQuillen, 506 U.S. 447 884 (1993). For discussion, see Chapter 12. United States v. Colgate & Co., 250 U.S. 300 (1919); for discussion see Chapter 12.

264

Vertical Restraints

conspiracy to monopolize does not require proof of market power, the only question is whether the parties possessed a specific intent to monopolize. In general, courts analyze exclusivity agreements under the rule of reason. Thus, unless the facts suggested that the conduct of the manufacturer and dealer served no potentially procompetitive purpose, a court probably would not find a specific intent to monopolize. The Supreme Court has recognized many of the economic reasonableness arguments that parties could assert to defend an exclusive dealership arrangement.29 Recall that the alternative theory is that the parties conspired in violation of Section 1. The boycott claim requires application of Northwest Wholesale Stationers. Because the parties do not have market power, (again Cellophane) the per-se rule does not apply. Consider instead the claim that the contract itself is unreasonable. One could cite Joint-Traffic30 (or Addyston Pipe) for the conclusion that such a contract does not violate the Sherman Act. The Court held in Joint-Traffic that covenants not to compete ancillary to some legitimate business transaction do not fall within the set of contracts that restrain trade. An exclusive dealership agreement is a type of competition-limiting contract analogous in many respects to the agreement in Mitchel v. Reynolds.31 Many cases have considered the reasonableness of exclusive dealership agreements, and the courts have generally upheld them under the rule of reason.32

B. Territorial Restrictions The other important form of nonprice vertical restraint is the standard territorial allocation scheme. Continental T.V. v. GTE Sylvania,33 a decision that revolutionized antitrust doctrine, set the doctrine in this area. Sylvania terminated Continental, one of its retailers, after Continental

29 30

31 32

33

See Sylvania, this chapter. United States v. Joint-Traffic Association, 171 U.S. 505 (1898). For discussion, see Chapter 5. 1 P.Wms. 181, 24 Eng. Rep. 347 (K.B. 1711). See, for example, Packard Motor Car Co. v. Webster Motor Car Co., 243 F.2d 418, 420 (D.C.Cir.), cert denied, 355 U.S. 822 (1957); Fuchs Sugars & Syrups, Inc. v. Amstar Corp., 602 F.2d 1025 (2d Cir. 1979), cert. denied 444 U.S. 917 (1980); Oreck Corp. v. Whirlpool Corp., 579 F.2d 126 (2d Cir. 1978); Aladdin Oil Co. v. Texaco, Inc., 603 F.2d 1107 (5th Cir. 1979). 433 U.S. 36 (1977).

II. Vertical Nonprice Restraints

265

began an unauthorized expansion into the markets of other retailers. Continental claimed that Sylvania had violated Section 1 of the Sherman Act by entering into and enforcing franchise agreements that prohibited the sale of Sylvania products other than from specified locations. The trial court, applying United States v. Arnold, Schwinn & Co.,34 concluded that Continental had violated Section 1. The Ninth Circuit reversed after making technical distinctions between the facts in Schwinn and those of Sylvania. The issue, in Sylvania, was whether Sylvania’s restriction on retail locations was a per-se violation of Section 1. The Court held that rule of reason analysis should apply. The Court began by analyzing the case under the rule of Schwinn, which was as follows: a. Under the Sherman Act, it is unreasonable without more for a manufacturer to seek to restrict and confine areas or persons with whom an article may be traded after the manufacturer has parted with dominion over it. b. But, the rule of reason governs when the manufacturer retains title, dominion, and risk with respect to the product, and the position and function of the dealer in question are, in fact, indistinguishable from those of an agent or salesman of the manufacturer. Pause for a moment to ask why courts had developed this rule. Why not hold all restrictions on resale per se unlawful, whether or not the manufacturer retains title? After all, isn’t the inference of conspiracy even stronger when the manufacturer retains title and the dealer acts as an agent? The answer is that the pure agency situation involves a single entity for antitrust purposes. Long before the Copperweld decision,35 courts had accepted as a settled rule that Section 1 of the Sherman Act could not be applied to an agreement among employees within a single firm.36 In light of this, it seems that the rule of reason that applies to resale restrictions (price or territory) where the manufacturer retains title is not

34 35

36

388 U.S. 365 (1967). Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752 (1984); for discussion, see Chapter 4. In other words, such agreements are per se legal. See the discussion of this issue in the Supreme Court’s Joint-Traffic decision (Chapter 5).

266

Vertical Restraints

like the general balancing test described in most antitrust cases. Rather, the rule governing resale restrictions distinguishes cases in which the manufacturer retains title, dominion, and risk from cases in which the manufacturer has set up a standard retail distribution network. There are cases in which there is some ambiguity as to whether the dealer is an independent retailer or an agent of the manufacturer. The rule of reason applied in this area is a test used not to determine the economic reasonableness of an agreement between a manufacturer and agent, but to distinguish independent retailer situations from agency situations. This must be true, in spite of the words used to describe the test, because once a court makes the determination that the relationship is one of agency, it must treat the agreement as operating within a single entity. Let us return to Sylvania. It was undisputed that title to the televisions had passed from Sylvania to Continental. Thus, under the Schwinn rule, Sylvania’s restrictions violated Section 1, unless the restraints fell outside of Schwinn’s broad prohibition. The Court saw no point in making technical distinctions between the facts in Schwinn and in this case. This was a violation of Section 1 under Schwinn. However, the Court decided to reconsider Schwinn. After presenting an economic analysis of restrictions that prohibit intrabrand competition in order to promote interbrand competition, the Court overruled Schwinn. The economic analysis repeated the arguments presented earlier in this chapter: that such restrictions encourage the provision of pre- and postsale service and information. The restrictions hinder competition among dealers within a given brand, but to the extent that they enhance the brand’s value, they enhance competition among brands.37 The revolutionary part of Sylvania is the implication that horizontal restraints may be deemed reasonable if they enable a consortium of producers to more effectively differentiate their product from those of others. Although one could read Sylvania narrowly by restricting its holding to vertical relationships, the Supreme Court has taken a broader view of the doctrine. The Court has not held explicitly that Sylvania overrules Topco. However, the Court has applied the reasoning of Sylvania 37

One could argue that efforts to differentiate a product may result in the dissemination of false information, or efforts to distort consumer demand. These efforts may result in a reduction in real interbrand competition, see Warren S. Grimes, Spiff, Polish, and Consumer Demand Quality: Vertical Price Restraints Revisited, 80 California L. Rev. 815 (1992). However, this can also be said of advertising generally. Although advertising results in the distribution of some false information and gimmickry, it is generally thought to be a procompetitive force.

III. Manufacturer Retains Title

267

when analyzing horizontal restraints in NCAA38 and in Indiana Federation of Dentists.39 Thus, it is virtually settled that the Sylvania doctrine applies to horizontal restraints as well as vertical restraints. Though sometimes cited as evidence of Topco’s continuing validity, Palmer v. BRG of Georgia40 can be reconciled with the arguments of this section. Harcourt Brace Jovanich Legal and Professional Publications (HBJ) and BRG of Georgia competed in the provision of bar review courses in Georgia between 1976 and 1980. There were no other providers of bar review courses in Georgia over the period. In 1980, HBJ and BRG entered into an agreement in which HBJ licensed BRG to use its material and promised not to teach bar review courses in Georgia, while BRG promised not to enter into any of HBJ’s markets outside of Georgia. Immediately after the agreement, the price of a bar review course in Georgia more than doubled. It is hard to make the case that this agreement enhanced interbrand competition; indeed, it completely shut down competition between the only existing Georgia brands, HBJ and BRG. Since the horizontal agreement seemed not to provide a noticeable enhancement of interbrand competition, the rule of Topco should apply under the analysis of this section, which is what the Court held.41

iii. manufacturer retains title Dr. Miles applied the rule against price fixing to sales by a manufacturer to a dealer on the ground that the resale price restraint interfered with the retailer’s right to sell his or her property. But if the goods are never the property of the retailer, does Dr. Miles still apply? Given the reasoning of Dr. Miles, it seems that it shouldn’t apply. In United States v. General Electric Co.,42 the Court held that if the manufacturer both retains title and bears the substantial risks of ownership, the antitrust laws do not prevent it from fixing retail prices. 38

39

40 41

42

National Collegiate Athletic Ass’n v. Board of Regents of the University of Oklahoma, 468 U.S. 85 (1984). For discussion, see Chapter 6. Federal Trade Commission v. Indiana Federation of Dentists, 476 U.S. 447 (1986); for discussion, see Chapter 9. 498 U.S. 46 (1990). One could argue that if entry is easy, there is little to worry about; any attempt by BRG to raise its price to a monopoly level would attract a new competitor. However, the standard approach to the rule of reason under the Sherman Act has focused on evidence of an increase in competition. In the absence of such evidence, Chicago Board of Trade requires finding a violation of the Act. 272 U.S. 476 (1926).

268

Vertical Restraints

Is this a sensible distinction? Probably not, if judged by consumer welfare. On consumer welfare grounds, resale price maintenance poses the problem that it may facilitate price fixing among retailers or manufacturers. This problem does not disappear once the retailers become part of an agency relationship. As noted earlier, the requirement of a distinction between agency and independent dealer cases comes from some well accepted understandings of the scope of the Sherman Act, not from any reference to consumer welfare. In the standard agency relationship, the manufacturer and seller are a single entity for antitrust purposes, so it follows that Section 1 does not apply. The rule of reason applied in the agency setting should be viewed as a test that aims to distinguish agency from independent dealer relationships. Such a test is necessary, or at least useful, because in some instances it is unclear whether the relationship is an agency relationship or a retail distribution network. The classic agency relationship is that between a manufacturer and an employee-salesperson. An agreement requiring the employee-salesperson to sell widgets at a price no less than $3 per widget does not violate Section 1, because only one entity is involved. The other extreme is of a retail distribution network, such as in Dr. Miles. The manufacturer sells the item directly to retailers or directly to wholesale distributors, who then sell the item to retailers. The rule of Dr. Miles applies to resale price agreements in this context. The space between these extremes includes arrangements that bear some of the features of both. The seller serves as an agent in some respects, yet in others the arrangement resembles a retail network. This perspective on the distinction between agency and independent dealer cases helps reconcile the Court’s decision in Simpson v. Union Oil Co.43 with the rule of General Electric. The Court ruled that Union Oil violated the Sherman Act by fixing the price its service station consignees charged to retail customers. The consignment arrangement was similar in many respects to that in General Electric, where the Court found no violation of Section 1 because the relationship came close to one of pure agency. However, the Court noted that Simpson, unlike the dealer in General Electric, carried property and liability insurance, and bore substantial risks of loss. The Court’s decision might be defended on the ground that the arrangement was so close to that of a retail distrib43

377 U.S. 13 (1964).

III. Manufacturer Retains Title

269

ution system that it may have seemed inappropriate to apply the rule of General Electric. Simpson illustrates the difficulties the Court faces under current law. On one end of the spectrum, a per-se illegality rule applies to resale price agreements in retail networks. On the other, a per-se legality rule governs the agency relationship. The need to determine whether the case is on one end or the other will inevitably generate seemingly contradictory decisions. The appearance of contradictions is not the only undesirable byproduct of the Dr. Miles regime. The Court’s efforts to distinguish agency and independent dealer cases raise questions about the competence of courts to make these distinctions in a sensible way. In determining whether a case falls in the agency context, the Court examines title, control, and allocation of risk. However, it is difficult to determine how risk is being allocated between a manufacturer and a retailer. Under a consignment relationship, the retailer pays the manufacturer a fee only for the goods sold, and returns the unsold portion. While on its face this seems to shield the retailer from market risk, the manufacturer can always structure the contract in a way that shifts market risk to the seller. Consider an example. Suppose manufacturer M sells retailer R five widgets. The widgets sell for $1 each. With probability –12 the retailer will sell 0, with probability –12 it will sell all five. The retailer has an expected revenue of $2.50. If the manufacturer bears the risk that not all of the goods will be sold, then the retailer accepts the goods, returns his fee for each good sold, and returns the unsold goods (ignoring shipping costs). If 2.5 widgets sell on average, the retailer returns on average 2.5 times the per widget fee to the manufacturer and 2.5 unsold widgets to the manufacturer. It would be easy, in this case, for the market risk to be transferred to the retailer. Let the retailer pay 2.5 times the per widget fee at the start of the contract period. Now if the retailer sells all five, it keeps the excess revenue; if it sells none, it suffers the loss. It is not necessary for the parties to set out the terms of this transaction in a special contract. The manufacturer could shift part, or all, of the risk by instituting a monthly fixed fee in exchange for a reduction in the fee charged per widget sold by the retailer. In this hypothetical consignment contract, the risk is transferred to the retailer even though the manufacturer retains title. Observations on certain features of a vertical relationship, such as whether the manufacturer or the retailer carries insurance, may fail to convey precise information on the allocation of risk between the parties.

270

Vertical Restraints

The prices agreed on by the parties may include implicit compensation for risk even though other external signs fail to indicate how the risk is being allocated.

iv. agreement A. Colgate Doctrine United States v. Colgate & Co.,44 involved a resale price maintenance plan in which evidence of agreement was never presented. The trial court interpreted the indictment as charging Colgate only with refusing to sell to a dealer, for the purpose of procuring adherence to a resale price fixed by Colgate. That is, the trial court did not interpret the indictment as charging Colgate with enforcing a resale price agreement. The Supreme Court considered itself bound by that interpretation, and the result was the celebrated Colgate doctrine: that each manufacturer is free to deal with whomever it wishes as long as its decision is not part of a plan to create or maintain a monopoly. There is one problem with the Colgate doctrine: it makes Dr. Miles almost unenforceable. In light of Colgate, a manufacturer that can afford good legal advice should be able to avoid getting caught in the Dr. Miles trap. How? Simply announce your policy and terminate dealers who do not comply. Is this a bad result? Probably not, because any other result would quickly become a generalized duty to deal in vertical relationships. Any manufacturer that terminated a dealer would risk facing a charge of violating the antitrust law. There are three legal rules that courts could adopt in this setting. One is a rule requiring courts to subject any refusal to deal by a single firm to scrutiny on antitrust grounds. The second is to make all refusals to deal per se lawful. The third is the compromise doctrine of Colgate, which applies a per-se legality rule when the defendant does not have market power, and subjects refusals to deal to antitrust scrutiny when the defendant has market power. Which rule is preferable? The first rule would create a “duty to deal” or, equivalently, a requirement that the manufacturer terminate only for cause. This would reduce competition, because the best way for a manufacturer to avoid a lawsuit from a terminated dealer would be to: (1) vertically integrate forward into the dealership market or (2) enter into long-term dealership 44

250 U.S. 300 (1919).

IV. Agreement

271

arrangements. In addition, a general duty to deal would cause manufacturers to screen their potential dealers to avoid entering into retailing agreements with likely price cutters. The greater incentive to screen would work against the interests of new entrants, who would not have developed a reputation for cooperating. The immediate effects are entirely analogous to a requirement that employers terminate only for cause, or that apartment owners seek court orders to evict tenants who fail to make rent payments. Such requirements encourage the party burdened with the task of proving a good reason for termination to shift from short-term to long-term relationships, from relationships with transients to relationships with repeat customers. To the extent long-term contracts reduce the number of parties competing to serve as retailers for a particular manufacturer, such a shift would be inconsistent with the policies of the Sherman Act. The impact could be harmful to consumers, because each retailer who had developed a reputation for cooperating with manufacturers would be in a position to demand a premium from the manufacturer. This would lead to an increase in the difference between the retailer’s price and the price received by the manufacturer, reducing consumer welfare. In addition, a duty to deal would encourage sham litigation. In many cases the reason for termination will not be clear. Every dealer terminated for failing to provide service will have an incentive to sue on antitrust grounds for treble damages. This danger would provide an additional motive for manufacturers to seek out cooperative retailers. Recall that the second potential rule is per-se legality. If entry is easy, consumers would not suffer harm in the long run under this rule, for entry should occur until economic profits are driven to zero (see Chapter 1). Under these conditions, monopoly profits secured through a refusal to deal would be short-lived. The interesting question is whether a different rule should apply where entry is difficult, or in order to minimize the short-term harm to consumers that results from anticompetitive refusals to deal. It is a question, I should note, because it is not clear that a duty to deal would enhance consumer welfare even in the case in which entry is difficult. For example, if an incumbent electric power firm has incurred substantial sunk costs in developing a transmission infrastructure, consumers may not benefit in the long run from a requirement that the incumbent firm allow rivals to use its transmission lines.45 45

See the discussion of the essential facility cases in Chapter 10.

272

Vertical Restraints

Replacing Colgate with a per-se legality rule would remove the potential for antitrust to improve consumer welfare in settings where a firm’s refusal to deal enhances or protects its market power, thus harming consumers, and entry does not occur rapidly enough to wipe out the profits from such action. For example, a per-se legality rule would result in inaction in cases like Lorain Journal46 in which consumers probably were made worse off by the defendant newspaper’s refusal to deal with some advertisers. In view of this, the Colgate rule is potentially preferable on consumer welfare grounds to a per-se legality rule. However, whether Colgate is preferable to a per-se legality rule in operation depends on whether courts are good at distinguishing anticompetitive from competitive refusals to deal.

B. Dr. Miles/Colgate Boundary Consider the following three hypothetical cases:47 (1) Manufacturer’s contract with each retailer specifies that the retailer will adhere to prices established by Manufacturer. (2) The contracts are at will and are silent as to price; M tells retailers that they should only accept the contract if they agree to adhere to specified retail prices and indicates that they will be terminated if they fail to comply. Manufacturer puts pressure on noncomplying retailers and terminates noncomplying retailers who fail to fall back into line after demands. Manufacturer continues retailers who respond favorably to such pressure. (3) The contracts are at will and silent as to price; Manufacturer announces retail prices and that it will terminate all who do not comply. Conclusion: (1) clearly violates Dr. Miles, (3) is permissible under Colgate, and (2) is the substantial ambiguous area that the cases are concerned with. It should be clear that the Dr. Miles/Colgate boundaries can generate some difficult-to-justify results. Imagine a manufacturer who must decide whether to terminate or merely warn a price-cutting dealer to stop cutting prices. Under the Colgate doctrine, the manufacturer may clearly terminate, and would have no obligation to justify the termination absent evidence that it was part of a common scheme to maintain resale prices of a certain level. Suppose the manufacturer is soft-hearted and decides to warn the dealer to stop cutting price. Then he has crossed the line 46 47

Lorain Journal Co. v. United States, 342 U.S. 143 (1951). For discussion, see Chapter 12. See Phillip Areeda and Louis Kaplow, Antitrust Analysis: Problems, Text, Cases 683–4 (4th ed. 1986).

IV. Agreement

273

from Colgate and moved into the area between the boundaries of Dr. Miles and Colgate. One case that illustrates the legal problems faced by such a manufacturer is United States v. Parke, Davis & Co.48 The government sought to enjoin an alleged conspiracy between Parke Davis and its wholesalers and retailers. Parke Davis sought to obtain compliance with its resale price maintenance provisions by bargaining with retailers. The Court concluded that if a manufacturer does not wish to rely on individual selfinterest to bring about voluntary compliance, and takes action to achieve compliance, the retailer’s compliance is not then a matter of individual free choice. In this case, the Colgate doctrine no longer protects the manufacturer. Under Parke Davis, taking action to achieve compliance may be sufficient to establish an agreement that violates the rule against vertical price fixing. Consider a manufacturer who, rather than simply terminating the dealer, enlists the aid of others in encouraging compliance. Albrecht v. The Herald Co.,49 implies that this may be sufficient to incur antitrust liability. In Albrecht, the Court found a conspiracy between the seller and agents hired to compete against a retailer who violated maximum pricefixing requirements. The inescapable conclusion is that the Dr. Miles/Colgate doctrinal boundaries create a set of perverse incentives, a nice-guys-finish-last regime. The manufacturer who must decide whether to terminate a pricecutting dealer is encouraged by the law to terminate quickly without negotiation. Consider the advice that should be offered to the dealer. The dealer’s best strategy is to manufacture evidence of an agreement on prices. Try to get something in writing that looks like an agreement with respect to price. For example, write a letter back to the manufacturer that sets out the details of a phone conversation and concludes with “I accept your proposal to keep my prices at or above level X, because I do not want to be terminated.” With this evidence in the files, the dealer is in a good position to collect treble damages if the manufacturer terminates it for price cutting. In fact, if its expected damages from termination are substantial, the dealer may have an incentive to cut its price in order to collect the treble damages jackpot at the end of a lawsuit. It seems that the Supreme Court is aware of the unsatisfying and indefensible results generated by the Dr. Miles/Colgate doctrinal boundaries. 48 49

362 U.S. 29 (1960). 390 U.S. 145 (1968).

274

Vertical Restraints

They would never have arisen if the rule of reason had been adopted in this area, and the Court now seems to be moving toward rule of reason analysis. However, the importation of rule of reason considerations has occurred through an indirect route. The Court has not overturned the per-se rule of Dr. Miles. Instead, the Court has altered the standard of proof in a manner that makes it more difficult for plaintiffs to take advantage of the per-se rule. The reason for raising the burden on plaintiffs is to allow the manufacturer leeway to exploit efficiencies described in Sylvania. The first sign of an effort to limit the scope of Dr. Miles was the Court’s decision in Monsanto Co. v. Spray-Rite Service Corp.,50 though the limitation on Dr. Miles was slight. Monsanto manufactures chemical products. Spray-Rite served as a wholesale distributor of agricultural chemicals made by Monsanto, from 1955 to 1968. In 1968, Monsanto declined to renew Spray-Rite’s distributorship. Spray-Rite sued under Section 1 of the Sherman Act. It claimed that Monsanto and some of its distributors conspired to fix resale prices of Monsanto herbicides. Monsanto denied the allegations of conspiracy, and asserted that Spray-Rite had been terminated for its failure to provide services in promoting sales. The jury finding that presented an issue was that of a conspiracy to fix resale prices. The Seventh Circuit’s opinion affirming this portion of the findings held that an antitrust plaintiff can survive a motion for summary judgment if he shows that a manufacturer terminated a price cutter in response to or following complaints by other distributors. The issue in Monsanto, narrow and technical, can be framed as follows: Does the standard of proof under Section 1 allow an antitrust plaintiff in a vertical price-fixing case to survive a motion for summary judgment if it shows only that a manufacturer terminated a price-cutting distributor in response to or following complaints by other distributors? The Court held that something more than evidence of complaints is necessary. The plaintiff must produce evidence that tends to exclude the possibility that the manufacturer and nonterminated dealers acted independently. The reason is easy to understand in light of Sylvania and Colgate. Complaints about price cutters will always occur in these cases, and manufacturers rely on information from distributors in coordinating sales efforts. If an inference of agreement may be based on such ambiguous

50

465 U.S. 752 (1984).

IV. Agreement

275

evidence, the Court reasoned, there is a danger that the doctrines of Sylvania and Colgate will be eroded. It should be clear that the doctrine of Sylvania would be significantly eroded if the Court had adopted the Seventh Circuit’s decision in Monsanto, but the doctrine of Colgate would be eroded in a relatively slight and indirect manner. The plaintiffs in this case were alleging conspiracy from the start. The Colgate defense does not apply to conspiracies. The plaintiff’s position therefore threatened only indirect erosion of Colgate. The manufacturer’s freedom to act unilaterally would be restricted because of cases with unclear evidence of conspiracy. Given the risk in this situation of a finding of conspiracy, the manufacturer would have an incentive to avoid termination even though Colgate protects its conduct. Does this case provide some measure of assurance to the manufacturer that efficient terminations, of the sort described in Sylvania, will not become the subject of antitrust litigation? Not much. In spite of the language, the Court upheld Monsanto’s damage judgment on the ground that the jury’s finding of conspiracy was reasonable in light of the evidence. A glance at the evidence immediately demonstrates the basis for wariness on the part of manufacturers. The evidence had some similarities to that in Parke Davis: efforts to cajole, persuade, and bully dealers to maintain prices. The case exemplifies the paradoxical result that manufacturers who try to encourage compliance rather than respond immediately with termination receive harsher treatment under the law – in the sense of having to pay 300 percent of the dealer’s loss rather than nothing. The next step in the direction of limiting the scope of the per-se rule of Dr. Miles was the decision in Business Electronics v. Sharp Electronics.51 Business Electronics and Hartwell were retailers for Sharp in the Houston area. Sharp published a list of minimum retail prices, but its written dealership agreements with BE and Hartwell did not obligate either to observe them. Hartwell, after complaining about Business Electronic’s price cutting, gave Sharp the ultimatum that it would terminate its dealership unless Sharp ended its relationship with Business Electronics. Sharp responded by terminating Business Electronics. Business Electronics brought suit under Section 1. The Fifth Circuit reversed the trial court and remanded for a new trial. It held that to render illegal per se a vertical agreement between a manufacturer and a dealer to 51

485 U.S. 717 (1988).

276

Vertical Restraints

terminate a second dealer, the first dealer must expressly or impliedly agree to set its price at some level, though not a specific one. Like Monsanto, the issue in Business Electronics is narrow and technical: Is a vertical agreement between a manufacturer and a dealer to terminate another dealer per se illegal only if there is an express or implied agreement to set resale prices at some level? The Court defended its affirmative answer as an implication of Monsanto and Sylvania. The Court reasoned that a per-se rule should not extend to agreements to terminate in response to price cutting because this would damage Sylvania. Almost any decision to terminate could be challenged as based on an effort to control price cutting – especially in a case in which one of the competing retailers has complained – and almost all of them would be challenged on this ground. The Court also argued that a per-se rule coupled with charges supported by complaints from other dealers would trample over Monsanto, because the complaints would serve as the Court’s evidence of agreement. Is Colgate threatened, or violated in some sense, by the plaintiff’s argument in Business Electronics? Apparently not, because this charge was for agreeing to terminate; in other words, concerted action is at the foundation of this claim, not unilateral action. But Colgate is implicated in an indirect manner. If the Court had held that evidence of an agreement to terminate price cutters is sufficient to warrant condemnation under the per-se rule, then even purely unilateral termination decisions would be chilled because of the fear that some evidence may exist that would lead a jury to find a conspiracy. The weak link in the Court’s argument is its assumption that Monsanto is not enough. The Court could have accepted a per-se rule extending to agreements to terminate in response to price cutting, and said that Monsanto requires that the plaintiff produce evidence that tends to exclude the possibility of independent behavior and the possibility that the agreement to terminate was based on something other than price (e.g., lack of service). This amounts to stating Monsanto in a more expansive way and leaving more to the jury’s discretion. The problem with this approach is that it draws on a rather inflexible and narrow view of the efficiency argument in Sylvania. Under this approach, the jury would listen to arguments over whether the termination occurred because of price cutting or because of poor service. It would decide which was the more important explanation. But every case would probably involve some of both. Furthermore, and more important, the efficiency argument for RPM is that it encourages competition

IV. Agreement

277

along other dimensions. Under this view, it is a sensible policy and requires termination to enforce. It should be sufficient to the manufacturer under this rationale that one of its retailers has failed to follow the RPM plan. If a retailer is not complying with the price floor, then it is probably competing in some prohibited way. The argument for making room for Sylvania can be supported with an additional economic justification. Termination in response to price cutting may be efficient even if service provision is unimportant. Suppose the dealers serve as points of information for geographically dispersed consumers. The manufacturer maximizes revenue by spreading stores over the map rather than having one location. Price cutting among the retailers may maximize the profits of a given retailer, while hurting the manufacturer by reducing the number of profitable points of operation. In sum, the holding in Business Electronics is consistent with a view that RPM may be a sensible policy, whether or not the manufacturer can prove its effectiveness in terms of enhancing presale service. This suggests that the per-se rule should have a limited scope. Business Electronics limits the scope of the per-se rule by requiring proof of agreement on some price level. Does Business Electronics effectively eliminate the per-se rule? No. One important fact is that there was no evidence of the type in Parke Davis. There was no evidence of efforts by the manufacturer to persuade retailers to raise their prices. Had there been such evidence, then inference of agreement may have been permissible under Monsanto, and Sharp may have fallen under the per-se rule. Indeed, with this factor taken into account it is possible to specify the precise sense in which the holding in Business Electronics limits the scope of the per-se rule. Business Electronics implies that in the absence of hard evidence of an agreement on a price level, circumstantial evidence of the sort observed in Parke Davis is necessary to survive a summary judgment motion, though it may not be sufficient as suggested by Monsanto. There is an unavoidable conflict between the per-se rule of Dr. Miles and the efficiency arguments the Supreme Court accepted in Sylvania. The tension may last for some time, with the Court issuing opinions that draw artificial distinctions in order to maintain the doctrinal boundaries of Dr. Miles and Colgate. However, the boundaries will eventually have to give. Dr. Miles, coupled with the small set of decisions buttressing it such as Parke Davis and Albrecht, exists today as a formalist doctrine required by precedent but otherwise unconnected to the economic reasonableness arguments accepted by the Court. The economic theories

278

Vertical Restraints

that give some measure of permanence or stability to antitrust rules no longer stand behind the doctrinal edifice. The theoretical arguments support Sylvania. One has to think that it is only a matter of time before a strong breeze causes the current edifice to crumble.

14 Tying and Exclusive Dealing

i. introduction Tying occurs when a seller of a product conditions the sale of the product on the buyer’s agreement to purchase some other product. For example, the seller of a copying machine (tying product) conditions the sale of the machine on the buyer’s agreement to buy all of the copying paper (tied product) from the seller of the machine. It should be clear that coercive tying requires market power. One of a dozen grocery stores in a community could not force consumers to purchase a pound of flour with every box of toothpicks.

A. Anticompetitive Theory of Tying and Chicago School Critique What’s wrong with tying? The thought that has informed antitrust attacks is the leverage theory; that is, the notion that a firm with monopoly power in one market can expand its power into another market.1 The potential harm to consumers, under this theory, results because the seller of the tying product extends monopoly power in the tying product into the market for the tied product. We have seen that the leverage argument is not airtight (see Chapter 10). Consider, for example, the argument that the Otter Tail Power Company leveraged its monopoly of transmission lines by forcing consumers to buy power from it as well.2 With low contracting costs, Otter 1

2

See, for example, Louis Kaplow, Extension of Monopoly Power Through Leverage, 85 Columbia Law Review 515 (1985). Otter Tail Power Co. v. U.S., 440 U.S. 366 (1973); discussed in Chapter 10.

279

280

Tying and Exclusive Dealing

Tail could have collected a monopoly rent by charging a sufficiently high “wheeling fee” (the rental fee charged by an owner of a power line to a power supplier). Thus, Otter Tail could have fully exploited its monopoly of transmission lines without requiring consumers who needed access to the company’s lines to also purchase power from the company. More generally, a set of arguments called the “Chicago School” showed that tying probably is not an effective method of leveraging monopoly power.3 Two basic results from this literature cover the cases of independent goods and complementary goods. For independent goods, the demand for one is independent of the price of the other. Consider the case of independent goods sold in fixed proportions, so that consumers always buy one X (or two Xs) along with any purchase of Y. If the firm has market power in Y, and X is sold in a competitive market, it cannot charge more than the monopoly price for Y plus the competitive price for X without suffering a loss in profits.4 Since the firm cannot charge more than the competitive price for X, it could only lose money from adopting a strategy of tying X to Y, because some consumers who would be willing to purchase X alone would be unwilling or unable to purchase the combined XY bundle. Now consider the case of complementary goods X and Y. As the name suggests, complementary goods tend to be complements in consumption or use, such as computers and monitors. For two complementary goods, the demand for each one depends on the sum of the two goods’ prices. Suppose the seller has market power in Y, the profit maximizing price of the combination of X and Y is $100, and the production cost is $25 per unit for both X and Y. On the assumption that consumers purchase the two goods together, the seller should be indifferent between charging $50 for each product, $25 for X and $75 for Y, or any other combination of prices that sums to $100 (the profit-maximizing price). Now suppose a rival appears who can profitably sell X at a price of only $10. Would the seller have an incentive to tie X and Y in order to prevent the rival from making any sales? The answer is no. Given that the profitmaximizing price for the XY bundle is $100, the monopolist in Y can improve his position by letting the rival enter and sell X at $10, while 3

4

For a discussion of the tying literature, see Keith N. Hylton and Michael Salinger, Tying Law and Policy: A Decision-Theoretic Approach, 69 Antitrust Law Journal 122 (2001). Richard Schmalensee, Commodity Bundling by Single-Product Monopolies, 25 J. Law & Econ. 67–71 (April, 1982) (showing that in the case of fixed proportions and independent demand curves a firm cannot do better than charging monopoly price for tying good and competitive price for tied good). See also the appendix to this chapter.

I. Introduction

281

raising his price for Y to $90. In this scenario, the monopolist earns a profit of $65 per unit. Before the rival’s entry, the monopolist’s profit was $50 per bundled unit. In neither of these cases does the firm actually increase its monopoly power, or its ability to exploit its monopoly power, through tying. Of course, the results depend on underlying assumptions. In the Otter Tail example, if one assumes that transaction costs would prevent the power company from negotiating and enforcing a profitable wheeling contract, then Otter Tail’s tying requirement would have the effect of expanding its monopoly power by allowing the firm to use it in a way that it could not if the tie-in were prohibited.

B. Anticompetitive Theory after the Chicago School Generally, tying can harm consumers by enabling a firm to increase or maintain its monopoly power, or to better exploit its monopoly power. The difficult problem for economists has been explaining how this is accomplished. The Chicago School arguments have altered the theoretical landscape so that while it was once assumed that tying is anticompetitive, the default viewpoint is now the Chicago position. As a result, the anticompetitive theory of tying has increasingly become situationspecific and dependent on special assumptions. The most successful and broadly applicable anticompetitive theory, due to Michael Whinston, is that tying deters entry in the tied product market by enabling the monopolist to commit itself to maintaining sales in that market, thereby weakening the resolve of potential competitors.5 Potential competitors stay out of the tied product market under Whinston’s theory because they know that if they enter, the monopolist will cut price to a level that makes it unprofitable for them, in order to maintain its sales of the tied bundle. Whinston’s theory provides a rigorous account of how tying can be used to increase the overall market power of a firm that has a dominant position in the tying product market. Whinston’s analysis shows that if the market for the tied good is not competitive, tying may be an effective method of increasing profits.6 5

6

See Michael D. Whinston, Tying, Foreclosure, and Exclusion, 80 American Economic Review 837–59 (1990). Whinston’s article provides the foundation for “post Chicago” analyses of tying. See also Jose Carbajo et al., A Strategic Motivation for Commodity Bundling, 38 J. Ind. Econ. 283 (1990). On the importance of commitment (or precommitment), see the discussion of the “chain-store paradox” in Chapter 10. See Whinston, supra note 5.

282

Tying and Exclusive Dealing

In the complementary goods setting, Whinston’s analysis shows that the monopolist in the primary good may have an incentive to tie in order to block competition from a rival in the primary good (recall, that in the absence of such a competitive threat the monopolist has no incentive to block competition from a cheaper rival in the secondary good). An alternative, broadly applicable theory of anticompetitive tying, due to Richard Craswell, argues that tying may enhance a firm’s market power by making it difficult for consumers to determine the difference between the implicit price of the tied good and its cost.7 If tying prevents consumers from gaining access to information on the competitive price of the tied good, the monopolist in the tying good will have a freer hand to set the implicit price of the tied good above the competitive level without losing too many sales.8 There are many theories that explain how tying could enhance a firm’s ability to exploit its market power. Transaction costs suggest one set of arguments that could explain how tying can enhance market power. Return to the Otter Tail example discussed above. Because transaction costs may make it difficult for the power company to exploit its power through a wheeling arrangement, the company may need to tie provision of power to its lines in order to fully exploit its power. Another theory that explains how tying enables a firm to better exploit its power posits that tying facilitates price discrimination. Return to the copying machine example. The seller would like to charge each machine purchaser the maximum price that purchaser is willing to pay. But there are two obstacles to this: obtaining information on willingness to pay is difficult, and arbitrage on the part of buyers can render a price discrimination plan ineffective. Specifically, a buyer who credibly asserts that his demand is low would purchase the item and resell it to a buyer whose demand is high. Competition among low-value buyers would drive the price down to marginal cost, which is the equilibrium price in a competitive market. Tying, under the price discrimination theory, is therefore seen as a method of getting around the information and 7

8

Richard Craswell, Tying Arrangements in Competitive Markets: The Consumer Protection Issues, 62 B. U. L. Rev. 661 (1982). Another theory is that tying is an effective method of restricting the substitution possibilities for purchasers of the tied product, though this is a theory of tying as a method of exploiting existing market power rather than increasing overall market power. Roger D. Blair and David L. Kaserman, Vertical Integration, Tying, and Antitrust Policy, 68 American Economic Review 397–402 (June 1978).

I. Introduction

283

enforcement problems that stand in the way of price discrimination.9 By tying the copying paper to the machine, and incorporating the monopoly surcharge into the price for paper, the monopolist can in effect sort buyers into different categories and apply a different surcharge to each according to willingness to pay.10 One might ask whether the copying machine example is really a case of price discrimination. The copying machine seller charges the purchaser more only when that purchaser uses more of the relevant goods. Price discrimination generally means selling the same good at a different price. At first glance, the copying machine example does not fit within the standard definition of price discrimination. However, to see how this example involves price discrimination one should assume the monopolist uses the paper as a counting device, that is, as a way of measuring demand. Yet another way in which tying may enable firms to exploit market power is by facilitating collusion among several firms. For example, suppose several copying machine manufacturers are colluding with respect to price in the market for the tied product, paper. In this case, if they require machine purchasers to also purchase paper, this will hinder the ability of consumers to take advantage of cheating by cartel members, and enhance the ability of cartel members to discover cheating when it occurs.

C. Efficiency and Other Considerations Even if we accept the theories suggesting that tying is anticompetitive, we cannot determine whether tying reduces consumer welfare without also taking the efficiency justifications into account. Tying could be an effort to provide convenience to consumers (reducing their search and

9

10

Ward Bowman, Tying Arrangements and the Leverage Problem, 67 Yale Law Journal 19, 20 (1957); M. L. Burstein, A Theory of Full-Line Forcing, 55 Nw. U. L. Rev. 62 (1960); William James Adams and Janet L. Yellen, Commodity Bundling and the Burden of Monopoly, 90 Quarterly Journal of Economics 475–98 (1976); Schmalensee, supra note 4. One might ask whether the copying machine example is really a case of price discrimination. The copying machine seller actually charges the purchaser more when and only when that purchaser uses more of the relevant goods. Price discrimination generally means selling the same good at a different price. At first glance, the copying machine example does not fit within the standard definition of price discrimination. However, to see how this example involves price discrimination one should assume the monopolist uses the paper as a counting device, that is, as a way of measuring demand.

284

Tying and Exclusive Dealing

transaction costs) or to protect the manufacturer’s reputation.11 Or the tying arrangement could be an effort to exploit economies of scope and scale in production or distribution.12 Any of these would suggest that both consumers and producers could be made better off by tying. I will return to the efficiency defenses later in this chapter. We should also keep in mind that some of the tying we observe may not be attributable to either anticompetitive or efficiency motivations.13 Consider, for example, a health club: suppose the board of directors decides the club should offer free babysitting to members. Of course, this is a form of tying; every club member pays for babysitting regardless of his or her need for it, and the result is that childless members subsidize members with children. This may reflect a decision on the part of the club’s directors to position the club to meet the demands of a specific segment of their market (e.g., families). On the other hand, it may simply reflect the arbitrary preferences of a majority of board members. This last example takes us into one of the more troubling issues in this area. To what extent is a manufacturer free to determine the shape of the product that he or she wishes to sell? Alternatively, how far are the tentacles of the Sherman Act supposed to reach in promoting competition? Must a producer alter the design of its product merely to enhance its substitutability with other products? Of course, product bundling typically aims to avoid this: firms often bundle in order to differentiate their products from others. The question at bottom is whether the tying decision is an attempt to create a new product with combined attributes that consumers can buy at a lower price or with less search effort, or whether the tie-in is an attempt to create or further exploit a monopoly position.

ii. early cases Early efforts to control tie-ins involved attempts by the owner of a patented product to tie a second, unpatented product. One case is 11

12

13

For example, in the franchise context, each franchisee has an incentive to free ride on the franchisor’s reputation for quality. Tying may be used as a method of preventing this type of opportunistic behavior. See Benjamin Klein and Lester F. Saft, The Law and Economics of Franchise Tying Contracts, 28 Journal of Law & Economics 345–61 (1985). See, for example, Roger D. Blair and Jeffrey Finci, The Individual Coercion Doctrine and Tying Arrangements: An Economic Analysis, 10 Florida State University Law Review 531, 549–50 (1983); Roger D. Blair and David L. Kaserman, Antitrust Economics 381–407 (Homewood, IL; Richard D. Irwin, Inc., 1985). I have made no attempt to provide an exhaustive statement of the potential reasons for tying. For an excellent statement of the theories, see Blair and Finci, supra note 12, at 544–55.

II. Early Cases

285

Motion Picture Patents Co. v. Universal Film Manufacturing Co.,14 in which the holder of a patent on a motion picture projector sold it on condition that it be used only to project motion picture films embodying inventions protected by certain other patents (that had expired). The patent holder brought an infringement suit against a purchaser who refused to comply with the condition. The Court denied the patent holder’s claim for infringement, and today the law is settled that tying provisions in a patent licensing agreement constitute misuse of the patent. The patent holder in Motion Picture Patents did not need to control the use of the machine in order to protect rents the patent assigned to him. The rents protected by patent law are those due to the exclusive control of production. As long as the defendant did not interfere with the patent holder’s exclusive control of production, no claim of patent infringement could (or should) stand. In any event, what is important about these early cases is that the Court worried about the leveraging of monopoly power, and this argument has played a central role in tying suits under the Clayton and Sherman Acts. Section 3 of the Clayton Act applies to tying. Section 3 makes it unlawful to LEASE or SELL GOODS, WARES, MERCHANDISE, MACHINERY, SUPPLIES, AND OTHER COMMODITIES, whether patented or unpatented, on the condition that the lessee or purchaser shall not use the goods of a competitor of the lessor or seller, where the effect of such a condition may be to substantially lessen competition or tend to create a monopoly.15

One of the early Section 3 Clayton Act cases is United Shoe Machinery Corp. v. United States,16 in which the Court affirmed an injunction under Section 3 against the conditions on which United Shoe leased its shoe making machines. United Shoe’s contract did not explicitly forbid the shoe manufacturer from using the machines of a competitor in combination with United Shoe’s machines. However, the contract permitted United Shoe to cancel the lease agreement if the lessee failed to comply with provisions that effectively prohibited the use of competitors’ machines. For example, several provisions of the lease agreement required the lessee to use the leased machine only on shoes on which other operations had been performed with the aid of United Shoe machines. 14 15 16

243 U.S. 502 (1917). 15 U.S.C. §14 (1988). 258 U.S. 451 (1922).

286

Tying and Exclusive Dealing

Another early Section 3 Clayton Act case is IBM v United States.17 IBM leased its machines on the condition that users purchase IBM punch cards. The Court held that this violated Section 3 of the Clayton Act. IBM argued that it tied use of the cards to the lease of its machine in order to protect its goodwill. The Court rejected this on the ground that IBM could have employed the less restrictive alternative of warning consumers and conditioning the lease on use of cards meeting the company’s specifications. The Court did not spell out the legal standard under Clayton Act Section 3 in these early cases. IBM suggested that the standard was an abbreviated form of rule of reason analysis. Under such an analysis the Court would examine: (1) whether the defendant has market power in the tying product, (2) the effects of the tie-in on the market for the tied product (does it force consumers to pay monopoly prices in the tied product market?), and (3) the availability of less restrictive alternatives.18

iii. development of per-se rule A. International Salt and the Per-Se Rule The first case to state the legal standard under Clayton Act Section 3 was International Salt Co. v. United States.19 International Salt tied the purchase of salt to the lease of two patented machines, the Lixator and the Saltomat.20 The issue was whether the tying provisions violated Section 1 of the Sherman Act and Section 3 of the Clayton Act. The Court held that they did. The Court found that International Salt had monopoly power in the markets for the tying products because they were patented (there was no analysis of market share). The Court also found that a not insubstantial amount of business ($500,000) was foreclosed in the tied market as a result of the tie-in. The Court held that this was all that was required to find liability under Clayton Act Section 3. The per-se rule operates from this point, because it “is unreasonable, per se, to foreclose com17 18 19 20

298 U.S. 131 (1936). See Ernest Gellhorn, Antitrust Law and Economics 316–17 (1986). 332 U.S. 392 (1947). “[T]he ‘Lixator’ dissolves rock salt into a brine used in various industrial processes . . . the ‘Saltomat’ injects salt, in tablet form, into canned products during the canning process.” Id. at 394.

III. Development of Per-Se Rule

287

petitors from any substantial market.”21 Although International Salt was far from obtaining monopoly power in the market for salt, the Court held this irrelevant because the Clayton Act forbids agreements that tend to create a monopoly, “and it is immaterial that the tendency is a creeping one rather than one that proceeds at full gallop.”22 International Salt argued that since the lease contracts required it to repair and maintain the machines, it should be permitted to require lessees to comply with use specifications. The salt that the company required its lessees to use was purer than many grades on the market, and therefore posed less risk of damaging the machines. The Court responded, as it did in IBM, that the company had the less restrictive alternative available of requiring lessees to use salt of a certain quality, without specifying the source. This is a good point to return to the efficiency arguments for tying.23 As I noted earlier, there are two that seem relevant to many of the cases: the goodwill defense and the economies of scale defense. The goodwill defense claims that the manufacturer’s tying is an attempt to protect its reputation for quality. The economies of scale defense claims that tying is necessary in order to exploit economies in joint production or marketing. Courts generally have rejected the goodwill defense for the reason given in International Salt: that there is the less restrictive alternative of instructing consumers to use the tying product only with complementary products that meet certain quality specifications. But there are cases in which this less restrictive alternative may not be sufficient. For example, suppose the good is leased, like the shoe machines in United Shoe or the salt processing machines in International Salt. Suppose further that the long-term value of the machines can be reduced by using them in certain ways that have the advantage of increasing value or saving money in the short term. A long-term owner probably would not sacrifice the longterm value of the machine for short-term gain. However, a lessee who uses the product for a short period may be willing to make such a sacrifice. The lessee considers the value of the machine over a much shorter horizon than the owner.

21 22 23

Id. at 396. Id. For a clear discussion of the issues, see Ferguson, Tying Arrangements and Reciprocity: An Economic Analysis, 30 Law & Contemporary Problems 552 (1965); Blair and Finci, supra note 12.

288

Tying and Exclusive Dealing

In addition to this “horizon problem,” there is another incentive problem that appears frequently in the franchising context. The “franchising problem,” occurs because franchisees have an incentive to free ride on the franchise’s brand capital. A franchisee could cut his costs by using cheap inputs, while externalizing the adverse effect on the franchise’s brand to all of the other franchise units. If the brand-devaluation loss suffered by the cost-cutting franchisee is relatively small (because it is spread over one hundred other units), then the franchisee has an incentive to reduce the quality of his inputs – since he gets the entire gain from cutting his costs and externalizes most of the loss to the other units. In order to control this incentive, franchisors often tie inputs to the franchise trademark by requiring franchisees to purchase inputs from the franchisor. In a setting where either the horizon problem or the franchising problem is observed, a tying arrangement may be the only reliable way for the seller (lessor) to constrain the buyer (lessee). The seller could issue clear specifications regarding the quality of complementary products, but if the buyer has a strong incentive to ignore those specifications, the seller’s efforts may be futile unless he imposes a tie-in. The tie-in may be necessary in order to protect the seller’s reputation, because it improves the seller’s ability to monitor buyers and to enforce contractual requirements. Although International Salt rejects the goodwill defense, an interesting exercise is to try to find some grounding for it in current antitrust doctrine. Probably the best source of doctrinal support comes from Sylvania.24 The Sylvania doctrine permits restrictions on intrabrand competition in order to enhance interbrand competition. However, it can also justify the application of a reasonableness inquiry when the purpose of the trade restraint is to differentiate a product and therefore make it a more effective competitor.25 An alternative efficiency argument for tying is that it permits the manufacturer to take advantage of economies of scale and scope in production. Suppose two items can be produced more cheaply together, at large volumes, than under separate production processes. Then the

24

25

Continental T.V. Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977); see Chapter 13 for discussion. This was the interpretation of the Sylvania doctrine accepted in NCAA v. Board of Regents of the University of Oklahoma, 468 U.S. 85, 103 (1984). For discussion of the NCAA opinion, see Chapter 6.

III. Development of Per-Se Rule

289

manufacturer could exploit the scope and scale economies by tying the two together at sale. Of course, if there are such economies, one could argue that the manufacturer should prove it by selling the two together at a cheaper price, without the tie-in.26 The problem with this is that if a competitor can meet the price of one of the items, the manufacturer may not be able to exploit scale and scope advantages unless there is a tie-in. To find doctrinal support for this argument, one could turn to BMI. The copyright clearing house agencies make a new product available by centralizing enforcement and royalty collection, thereby reducing transaction costs that ordinarily prevent much of this activity. However, centralization works in the BMI example because there are scale economies in enforcement and collection. When a manufacturer ties in order to take advantage of scale economies, its conduct is arguably indistinguishable from that of the defendants in BMI. Returning to the problem of International Salt, has anyone tried to find out why the firm tied salt purchases to sales of its processing machines? The reader is referred to John Peterman, The International Salt Case.27 Peterman argues that the evidence is inconsistent with the claim that the defendants used the tying arrangement to charge discriminatory prices. First, recall that the price discrimination theory assumes that the reason the seller ties is to prevent buyers from rendering price discrimination ineffective through arbitrage. However, according to Peterman, International Salt leased Lixators with the requirement that lessee purchase salt only in areas of the country in which it sold salt. In other areas (where IS did not sell salt) it sold Lixators. Thus, any buyer who felt put upon by the tie-in lease could buy the machine in another location. Second, the price discrimination thesis assumes that the monopoly surcharge is incorporated into the price of the tied product. In the International Salt case, however, there was a “meeting competition” clause in the salt requirements contract, which suggests that the salt was not sold at above market prices. What was International Salt doing? Peterman argues that the company was probably taking advantage of economies in distribution of salt. The requirements contract may have saved International Salt money by providing a group that would demand salt at regular intervals. Under this theory, International Salt lowered the price of its Lixator as long it reduced costs of distributing salt (to the point where on the margin it 26 27

See Ferguson, supra note 23. 22 J. Law & Econ. 351 (1979).

290

Tying and Exclusive Dealing

equaled the reduction in profit from the leasing of Lixators) more than it reduced profits from selling Lixators. Why tie? Because the firm could not meet the economies in distribution unless it could be sure that the demand for salt was there. The tie-in therefore allowed International Salt to lower the price of its machine while selling salt at the market price. The evidence supports this because the lease price was lower than the appropriately amortized sales price of the machine.

B. Articulation of the Per-Se Standard International Salt held that per-se analysis applies in the area of tying, but the Court did not articulate the precise standard. Furthermore, the earlier cases left in doubt whether the same standard applied under the Sherman and Clayton Acts. The Court resolved these issues in Northern Pacific Railway Co. v. United States.28 The defendant railroad had inserted clauses into contracts to sell and to lease land located near its rail lines that required purchasers and lessees of land to ship goods produced on the land over Northern Pacific’s rails, provided that Northern Pacific’s rates were not above those of competing carriers. Because the tying product was land, which is not a “commodity,”29 Clayton Act Section 3 did not apply. The issue, then, was whether the preferential routing contracts violated Section 1 of the Sherman Act. The Court held that they did. The Court declared that the rule of International Salt is a per-se rule that applies whenever a party has sufficient economic power with respect to the tying product to appreciably restrain free competition in the market for the tied product and a “not insubstantial” amount of interstate commerce is affected.30 In this case, the facts “established beyond any genuine question that the defendant possessed substantial economic power by virtue of its extensive land holdings.”31 Furthermore, the Court held that a not insubstantial amount of interstate commerce had been affected. 28 29

30 31

356 U.S. 1 (1958). Section 3 of the Clayton Act applies only to commodities, which means some type of tangible property that can be leased or sold; see, for example, Satellite Television Ass’n Resources, Inc. v. Continental Cablevision of Va., Inc., 586 F.Supp. 973, 975 (E.D.Va. 1982), aff’d, 714 F.2d 351 (4th Cir. 1983), cert. denied, 465 U.S. 1027 (1984). 356 U.S. at 6. Id. at 7.

III. Development of Per-Se Rule

291

The Court noted that: (1) the defendant’s land was “strategically located within economic distance of transportation facilities,”32 (2) the defendants offered no procompetitive justification, and (3) the preferential routing clauses conferred no benefit on the land purchasers because there was no evidence that they got the land at a lower price because of the agreement.33 The market power component of the test of Northern Pacific is more lenient than that suggested by International Salt. International Salt was found to have monopoly power in the tying product market, while there was no such finding in Northern Pacific. The Court said that monopoly power is not a requirement for application of the per-se rule; all that is required is “sufficient economic power to impose an appreciable restraint” on competition in the tied product market.34 Northern Pacific implicitly holds that the standard of liability under the Sherman and Clayton Acts is the same. As the dissent in Northern Pacific noted, the Court had examined tying under the Sherman Act (i.e., apart from the Clayton Act) in the Times-Picayune case, and had suggested that when the facts are examined solely under the Sherman Act, rule of reason analysis applies; which implied the necessity of a finding of monopoly power in the tying product market. The fundamental issue in Northern Pacific was whether the defendant’s efforts were designed to enhance monopoly power or an attempt to exploit economies in a manner that would benefit consumers as well as the seller. The monopoly exploitation story is straightforward: Northern Pacific received the land for free and tried to use it to enhance its business by selling it on the condition that purchasers use the railroad’s services. If this were all to the case, one would have to say that the Court reached the right conclusion. However, the contracts allowed buyers to use other railroads when price or service reasons justified switching. This suggests that if the railroads were competing, Northern Pacific could not have earned a monopoly profit on its rail services as a result of the tiein. If it charged a monopoly price on the rail services, the land purchasers would have switched to another carrier. An alternative argument supporting the monopolization theory is that the tying arrangement allowed the railroads to monitor the pricing of 32 33 34

Id. at 8. Id. at 7–8. Id. at 10.

292

Tying and Exclusive Dealing

other carriers.35 Thus, if there was a tacit or explicit pricing agreement, the contract facilitated Northern Pacific’s efforts to monitor compliance. One can offer an efficiency justification for the tying arrangement in Northern Pacific. Because of high fixed costs, railroads experience reductions in average cost as the volume of service increases. Northern Pacific may have used the tying contract to ensure a certain volume of traffic, which would lower the cost of rail service. Under this theory, the tie-in would fail to offer anything concrete (e.g., a lower sales price) to land purchasers at the moment of the transaction.36 On the other hand, purchasers would be indifferent to the provision, operating on the assumption that they could switch carriers if price or service conditions justified. However, to the extent that land purchasers used Northern Pacific’s rails, the railroad would be able to offer lower rates. This argument suggests that rule of reason analysis might have avoided a finding that the contract violated the Sherman Act. The difference between the market power requirement under the Section 2 monopolization standard and that of Northern Pacific is best illustrated by United States v. Loew’s.37 The case involved a block booking arrangement under which the license of one or more feature films to a television station was conditioned on acceptance of a block containing some less desirable films. The Court held that this tie-in violated Section 1 of the Sherman Act, repeating the per-se test of Northern Pacific. Sufficient economic power followed from the fact that each movie was copyrighted and therefore “unique.” It is unlikely that the Court would reach this conclusion in a Section 2 case. Cellophane requires consideration of substitutability in determining the relevant market. It should be clear that one movie, although copyrighted and unique, is highly substitutable with other movies.

C. Jerrold and the Goodwill Defense The goodwill justification – that is, the argument that tying is a means of protecting a manufacturer’s reputation – appears prominently in many of the cases. The Court rejected it in IBM and in International Salt, and 35

36

37

See F. Cummings and W. Ruhter, The Northern Pacific Case, 22 Journal of Law & Econ. 329 (1979). The Court noted “[w]hile they got the land they wanted by yielding their freedom to deal with competing carriers, the defendant makes no claim that it came any cheaper than if the restrictive clauses had been omitted.” Northern Pacific, 356 U.S. at 8. 371 U.S. 38 (1962).

III. Development of Per-Se Rule

293

courts have continued along this line. The most significant innovation in this regard is the treatment in United States v. Jerrold Electronics Corp.,38 which, while relying on the Northern Pacific standard, carved out an exception for the goodwill defense. Jerrold sold antenna systems with the requirement that the purchaser let Jerrold and only Jerrold install and service the system. Many of its contracts provided for the exclusive use of Jerrold equipment whenever extra capacity was needed, and forbade installation of extra equipment without Jerrold’s approval. Jerrold sold a majority of new antenna systems from 1950 to mid-1954, and was the leader in sales of community television antenna systems. The government claimed that this violated Section 1 of the Sherman Act and parts of this violated Section 3 of the Clayton Act. (The Sherman Act applied to restrictions on services and commodities, the Clayton Act to restrictions on commodities.) The district court held that Jerrold’s position in the tying product market justified the conclusion that the tying arrangements were unlawful per se under Northern Pacific. However, in view of what the Court considered to be “unique circumstances” and the fact that it could examine the industry in historical detail easily,39 the Court considered the reasonableness justifications of the defendant. The Court noted that the restrictions were imposed when Jerrold was entering a market that was still in its infancy. Because of significant upfront costs in setting up an antenna system, the restrictions were a reasonable way of minimizing the risk of product failure and ensuring that customers would pay for installation (payment was contingent on successful operation under Jerrold’s contracts). Although Jerrold could have informed customers of the need to comply with rules guaranteeing successful operation, this would have put it in the position of suffering injuries to reputation because some customers failed to follow Jerrold’s instructions. The court concluded that Jerrold’s policy was reasonable at its inception, given the cost of the reputational damage that would have otherwise occurred, but that it violated the antitrust laws some time after that. The Court interpreted Northern Pacific as putting the burden on Jerrold to prove that the restrictions were reasonable throughout their use, and held that Jerrold failed to meet this burden.

38 39

187 F. Supp 545 (E.D. Pa. 1960), aff’d per curiam, 365 U.S. 567 (1961). Note that this suggests that under the utilitarian test of Trenton Potteries, rule of reason analysis may be justifiable. The administrative costs of undertaking an economic analysis of the effects of tying provisions are relatively small in this setting.

294

Tying and Exclusive Dealing

The government had also charged that the full system sales violated Clayton Act Section 3, as well as Sherman Act Section 1, because they tied different commodities. Jerrold argued that the full system sales were not a combination of commodities but a single commodity. The Court noted that other manufacturers sold parts of full systems, the number of pieces in Jerrold’s systems varied considerably, Jerrold charged the customer separately for each item in a system, and Jerrold allowed customers to substitute antennas and cable from other sources. In light of these factors, the Court said that the evidence suggested Jerrold had tied distinct products. However, the Court held that the full system sale was reasonable at inception, given Jerrold’s policy of servicing the equipment it installed. As the industry grew, the reasons for Jerrold’s policy disappeared and the tie-in became unreasonable. The single product inquiry suggested by Jerrold, and followed by many courts today, involves four questions: (1) do other sellers sell parts of the bundled item separately, (2) is there variation in the size and composition of the bundled item, (3) are the items priced separately, and (4) are there substitutes for some of the components in the bundle? A yes to all four questions often leads to the conclusion that the seller has tied separate products. To say that Jerrold applies rule of reason analysis would be incorrect. Rule of reason analysis starts with a presumption of reasonableness and requires proof of unreasonableness on the plaintiff’s part. The district court in Jerrold applied the standard in reverse, starting on the basis of Northern Pacific with a presumption of unreasonableness and allowing the defendant to prove reasonableness. This is a standard that is about as difficult on the defendant as a per-se unreasonableness rule. However, there are some benefits. The reverse rule of reason allows the court to hear the efficiency arguments and to make some findings of reasonableness. In Jerrold, the Court found that the restrictions were reasonable at inception, and this finding now stands as an exception to the per-se tying rule. The exception applies when the manufacturer uses tying to facilitate entry into a new industry. I have so far avoided any discussion of the efficiency argument for bundling parts of Jerrold’s system together. The system parts were all standardized in this case, like the shoe machines in United Shoe (Chapter 10). Moreover, again like United Shoe, the systems were custom-made, in the sense that the type of equipment and the size of the overall system were varied to meet the needs of each operator. In this setting, there is a plausible free rider concern that might justify a tying policy. If Jerrold

IV. Tension Between Rule of Reason Arguments and Per-Se Rule

295

did not tie the parts together, an operator might have an incentive to procure a cheaper alternative for one or more parts of the Jerrold system. If, as has been suggested in the case of United Shoe,40 Jerrold invested a substantial amount in the design of a system and in educating operators, this incentive to substitute might work to deny Jerrold a competitive return on its initial investments in design and education.

iv. tension between rule of reason arguments and per-se rule As Jerrold and other decisions have made clear, courts are aware of the rule of reason arguments that support tying arrangements.41 A per-se rule creates difficulties. A court faces the choice of making exceptions that dilute the utility of a per-se rule, or declaring an arrangement outside of the scope of the Clayton Act. Jerrold creates one exception to the per-se rule, for new products. The broadest class of tying contracts considered outside the scope of the tying prohibition consists of “single products.” Right and left shoes, for example, are typically tied together. No one considers bringing an antitrust claim against shoe sellers for the simple reason that a court would hold that a pair of shoes is a single product.42 The strain created by the simultaneous existence of plausible reasonableness justifications and a per-se illegality rule has had the effect of shifting the Court in the direction of a reasonableness inquiry. Some signs of that shift are observable in modern tying decisions. The first of those is United States Steel Corp. v. Fortner Enterprises [II].43 U.S. Steel offered to finance the cost of acquiring and developing land provided that the buyer (developer) purchased prefabricated houses 40

41

42

43

See Scott E. Masten and Edward A. Snyder, United States v. United Shoe Machinery: On the Merits, 36 J. Law & Econ. 33 (1993). See, for example, Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 2, 41 (1984) (O’Connor, J., concurring) (stating several economic justifications for tying). On the single product exception, see, for example, Automatic Radio Mfg. Co. v. Ford Motor Co., 242 F. Supp. 852, 857 (D. Mass. 1965) (rejecting argument that Ford dashboards and car radios were separate products); Montgomery County Association of Realtors, Inc. v. Realty Photo Master Corp., 993 F.2d 1538 (Table), 1993–1 Trade Cases (CCH) P70,239, 1993 WL 169046 (4th Cir., May 20, 1993) (unpublished opinion holding, among other things, that summary of real estate specifications and picture of house constitute a single product for purposes of tying analysis); see also Service & Training Inc. v. Data General Corp., 963 F.2d 680 (1992); Faulkner Advertising Assoc., Inc. v. Nissan Motor Corp. in U.S.A., 905 F.2d 769 (1990). 429 U.S. 610 (1977).

296

Tying and Exclusive Dealing

from it. Thus, in order to get the financing deal offered by U.S. Steel, you had to buy houses from it. The tying product was credit, the tied product, prefabricated housing. Although U.S. Steel’s tie-in affected only a small fraction of the total market for prefabricated housing, the Court held that the dollar volume easily satisfied the quantitative substantiality test of Northern Pacific. The remaining question was whether defendants had sufficient economic power in the tying product market to appreciably restrain competition in the tied product market. The Court concluded that the plaintiff had not shown that the unique character of the financing supported the conclusion that U.S. Steel had the kind of power in the tying product market required in order to find a per-se violation. Although the most important lower court finding was that the uniqueness of the financing arrangement suggested that U.S. Steel had “sufficient economic power,” the Court considered and rejected three other theories that might have supported the finding of a Sherman Act violation. (1) The fact that the credit arm of U.S. Steel was controlled by one of the nation’s largest corporations did not prove, in the Court’s view, that it had a special advantage in borrowing funds that was not available to other lending institutions. (2) Leveraging and price discrimination theories were not persuasive to the Court because U.S. Steel provided credit with each home. Thus, it had to “sell” a credit package in order to sell a home, which suggests that it could not have been using its power in the credit market to expand into the housing market or to sort buyers by willingness to pay.44 (3) The fact that houses were priced above market was not sufficient evidence to the Court. Rather, the important question is whether the total package, credit plus housing, exceeded the market price. As for the uniqueness finding, the Court concluded that the only unique part was the price (100 percent financing at low interest rates). In short, it was a good deal. But this is not the kind of uniqueness that gives rise to a Sherman Act violation. 44

Each credit package was connected to a plot of land, which was then tied to a prefabricated house. This is distinguishable from the case in which the monopolist ties an item whose quantity is variable. For example, a copy machine manufacturer who ties paper to the sale of a machine might be able to use the tie as a method of transferring the monopoly surcharge from the machine to the paper. Incorporating the monopoly surcharge in the price of paper would enable the manufacturer to price discriminate. Consumers with a greater demand for paper would pay a larger share of the monopoly surcharge than consumers with low demand for paper. The theory that tying is a form of price discrimination is considerably less plausible when the tied product is sold in a fixed proportion to the tying product.

IV. Tension Between Rule of Reason Arguments and Per-Se Rule

297

Although the Fortner II Court continued to rely on the standard of Northern Pacific, some statements in the opinion intended to clarify that standard actually alter it significantly. It appears under the language of Fortner II that a monopoly or dominant position is necessary to show sufficient economic power unless the product has some aspect of uniqueness.45 Uniqueness, generally, is some feature of the product that makes it difficult for the buyer to switch to the product of a competitor. The Court suggested that a showing of uniqueness could be based on any one of three factors: (1) a legal constraint, such as a patent, restricting competition in the tying product market; (2) a physical advantage, such as the location of the land parcels in Northern Pacific; or (3) a cost advantage over competitors in the tying market. The second decision suggesting movement in the direction of a reasonableness inquiry is Jefferson Parish Hospital District No. 2 v. Hyde.46 East Jefferson Hospital in New Orleans required patients wanting an operation to use the hospital’s anesthesiologists. The Court was unanimous in the view that this tying arrangement did not violate the Sherman Act.47 However, it gave different reasons. The majority opinion relied on the per-se rule of Northern Pacific and concluded that the evidence did not suggest that the hospital had sufficient market power in the tying product and that there was a significant impact in the tied product market. The insufficient power finding was based on the fact that the hospital attracted only 30 percent of the potential patients in the relevant geographic market. The conclusion that the quantitative substantiality prong of the test had not been satisfied was based on the fact that few patients were actually being “forced” by the tying arrangement to use an anesthesiologist that they did not want to use. One preliminary question is why this is not simply a case of someone asking for something that the market does not provide, as in the case of a shoe store customer who wants to buy only the right or only the left shoe. Doesn’t the fact that no one asks for surgery without anesthesia imply that surgery plus anesthesia is a single product? The majority held that the two were separate products because there “was ample . . . testimony that patients or surgeons often request specific anesthesiologists to come to a hospital.”48 Thus, Jefferson Parish adopts an empirical single 45 46 47

48

429 U.S. at 616–21. 466 U.S. 2 (1984). Note that only the Sherman Act is implicated because the tying arrangement involves services rather than commodities. 466 U.S. at 22.

298

Tying and Exclusive Dealing

product test, one that relies on evidence concerning the actual behavior of consumers.49 The central issue in Jefferson Parish was whether the tying arrangement led to “forcing.” The Court held that if such forcing were present, sheer market power could not be the reason because the hospital’s share of area patients was only 30 percent. The alternative reason, relied on by the appeals court, was that lack of information on the part of consumers made comparison shopping ineffective as a method of controlling monopoly pricing. The Court held that rather than implying a Sherman Act violation, this argument implied that the tying arrangement had no effect on consumer choice. If consumers failed to shop around, then the decision to close the anesthesiology department would do little to force them to make decisions that they would not ordinarily make. The tie-in of anesthesiological and hospital services did not foreclose competition “on the merits” in the market for anesthesiological services. This is a questionable argument. While consumers were probably not being forced to buy the services of anesthesiologists they would otherwise not patronize, it is still possible that opening the anesthesiology department of East Jefferson Hospital would exert some restraint on the pricing freedom of hospital anesthesiologists. Consumer shopping is not the only activity that constrains the pricing freedom of physicians. Outside anesthesiologists might be willing to charge a lower fee, and insurers and HMOs might eventually direct their customers toward the price-cutting anesthesiologists. Although consumers lack information, there are other parties in the payment chain who have enough information to make informed price-quality tradeoffs. Once this argument is considered, the sense in which this case is analogous to Sylvania becomes important. An unyielding application of the procompetition approach would suggest that this tying arrangement violates the Sherman Act. However, there are service-related reasons for closing the anesthesiology department at a hospital. Justice O’Connor noted these in her concurring opinion, and they were key to the district court’s finding of no violation. The closed department permitted the partnership that provided anesthesiological services to work twenty-four hours, coordinate schedules, maintain equipment, and invest in new tech-

49

Specifically, the single product test looks to the conduct of firms in the competitive fringe that offer the tying and tied goods (or services). Courts will find a “single product” if tying is the standard practice within the competitive fringe. For an opinion that breaks from this approach, see United States v. Microsoft Corp., 253 F. 3d 34 (D.C. Cir. 2001).

IV. Tension Between Rule of Reason Arguments and Per-Se Rule

299

nology.50 One could argue that the closure restricted intrabrand competition and enhanced interbrand competition. If, in spite of its relatively small share of the land market, Northern Pacific Railroad could be found guilty of violating the Sherman Act because of the uniqueness or strategic location of the land it sold, then it does not seem a hard stretch to conclude that a hospital anesthesiology department has sufficient grasp of its patients to meet the economic power requirement of Northern Pacific. That the Court did not find sufficient power suggests that the inquiry has been shifted, as in Fortner II, toward a monopoly power requirement. More precisely, the Court’s analysis adheres to the Fortner II test requiring proof of either market power or uniqueness based on a cost advantage. This standard would not permit a plaintiff to base a showing of either uniqueness or market power on proof that consumers were too uninformed to engage in intelligent comparison shopping. Four of the concurring Justices urged the Court to abandon the test of Northern Pacific and return to a rule of reason approach in tying cases analyzed under the Sherman Act. The majority rejected this recommendation, but their treatment of the market power question suggests that there is not a great difference between the reasonableness standard and the per-se test as applied in Jefferson Parish. The gradual movement toward the rule of reason observed in Fortner II and Jefferson Parish was thrown into doubt by the Court’s decision in Eastman Kodak v. Image Technicolor Services.51 Eastman Kodak produced copying machines and provided parts and service for the machines. Independent Service Organizations (ISOs) had also entered the market for service. They purchased parts from Kodak and from independent original equipment manufacturers who made parts under contract from Kodak. They sold the parts and service to Kodak equipment owners. The dispute arose because Kodak attempted to eliminate the independent service market by adopting a policy of selling parts only to equipment owners who relied on Kodak for service or who self-serviced their equipment. One of the ISOs brought suit on the theory that Kodak 50

51

These are important features because in the absence of closure, independent anesthesiologists might have an incentive to free ride off the investments of the incumbent group. It is interesting that the empirical evidence does not suggest that exclusive service contracts are used by hospitals to suppress competition, see William J. Lynk and Michael A. Morrisey, The Economic Basis of Hyde: Are Market Power and Hospital Exclusive Contracts Related?, 30(2) Journal of Law & Economics 399 (1987). 504 U.S. 451 (1992).

300

Tying and Exclusive Dealing

had tied service (tied product) to parts (tying product) in violation of the Sherman Act. The issue was whether a manufacturer of equipment and parts is entitled to a legal presumption that it lacks market power in the parts market if the equipment market is competitive. The Court refused to establish such a presumption. The Court had no trouble finding evidence of sufficient economic power in the parts market to justify a jury trial, and the quantitative substantiality test was conceded. Because Kodak was virtually a monopolist with respect to its own parts, this finding was not a surprise. However, in order for there to be economic power, in any relevant sense, it must be the case that Kodak could increase the price of service without causing such a decline in the equipment market that the service price increase would not be desirable. The Court held that this was in theory possible and plausible because of imperfect information and switching costs.52 Buyers may not be sophisticated enough to evaluate the total pricequality package at the period before purchase of a Kodak machine – that is, to do “life-cycle” pricing. After purchase, it is costly to switch to another machine. As a result, machine purchasers could find themselves locked in, and to some extent at the mercy of Kodak. At bottom, then, the Court said that a firm may lack market power in the traditional sense, and yet because of consumer ignorance or switching costs, could use tying as a means of collecting a monopoly profit. This seems inconsistent with Jefferson Parish. Jefferson Parish says that in spite of consumer ignorance that renders comparison shopping infeasible, market power or uniqueness based on cost advantage must be demonstrated. One could try to escape the impression of a contradiction by saying that demand-side substitution was ineffective in Jefferson Parish, while it was effective in Eastman Kodak – if anything too effective. But this is a questionable argument, especially today. HMOs and insurance companies provide demand-side substitution in the health market. As I suggested above, the rule of reason argument of Sylvania provides probably the only solid justification for the result in Jefferson Parish. An alternative attempt at reconciliation is to note that the Eastman Kodak holding, unlike that of Jefferson Parish, applies to a summary judgment. But this is unpersuasive also. Eastman Kodak opens the door for a plaintiff to try to prove in court that sufficient economic 52

The Court cited Craswell, supra note 7. For further discussion, see the appendix to this chapter.

V. Technological Tying

301

power, in the sense of Northern Pacific, resulted from the inability of consumers to engage in intelligent comparison shopping. But it is precisely this argument that the Court rejected in Jefferson Parish. Justice Scalia’s dissent argues that if Kodak had tied the parts and service to the initial sale of equipment, there would be no finding of a violation of Section 1. The reason is that the original equipment market was competitive. How, then, can a different result be justified? The majority’s justification seems to be that although the equipment market was competitive, Kodak had sufficient economic power because consumers were uninformed and locked in after purchase. In the years following Eastman Kodak, appellate courts have had to reconcile its language with that of Jefferson Parish. It appears now that courts hold that the key to Eastman Kodak was the firm’s mid-stream change in policy. In other words, if Eastman Kodak had imposed the tie-in at the very start, there would have been no violation of Section 1 – consistent with Scalia’s analysis.53 If a firm changes its policy by imposing a part-service tie-in, it may violate Section 1.

v. technological tying Given the legal restrictions on product tying, one might expect some manufacturers to simply bolt two products together and call the result a single product in order to avoid the per-se rule. Courts are aware of this incentive and plaintiffs have brought tying claims against sellers who have technologically integrated products. However, the legal standard governing technological integration differs substantially from the tying law discussed up to this point. While a complicated per-se doctrine applies to contractual tying (“if you buy A you must also buy B”), courts apply a test that pretty much presumes legality in the area of technological integration (physically combining A and B into an integrated product). Courts have held that in order to succeed in a technological tying claim, a plaintiff must show that the defendant integrated two products for the sole purpose of hampering competition rather than to produce some additional utility to consumers. The prevailing legal standard was set out in Response of Carolina, Inc. v. Leasco Response, Inc.,54 which 53

54

See, for example, Metzler v. Bear Automotive Service Equipment Co., 19 F. Supp. 2d 1345, 1356–8 (S.D. Fla. 1998) (reviewing Eastman Kodak and later appellate decisions). 537 F.2d 1307 (5th Cir. 1976).

302

Tying and Exclusive Dealing

requires proof that the integration was solely “for the purpose of tying the products, rather than to achieve some technologically beneficial result.”55 So, for example, a car manufacturer that bolts a radio onto the dashboard would not fall under the safe harbor of Leasco if it could be shown that the bolting offers consumers no additional utility beyond what they could achieve on their own by purchasing the car and radio separately. The Leasco standard puts a high burden of proof on plaintiffs in technological tying cases. Indeed, the standard is in some respects a version of the specific intent test described in Chapter 10. By requiring the plaintiff to show that the defendant integrated the two products for the sole purpose of tying them, the Leasco standard forces plaintiffs to present evidence that would justify the inference that the defendant intended to harm competition. This is obviously a more difficult standard for plaintiffs than the test governing contractual tying claims. The difference between contractual and technological tying standards has created a problem that probably will grow in importance over time. A seller who can choose between technological integration and contractual tying has an incentive to choose technological integration. Some have argued that this is especially important in the area of computer software, where the difference between contractual tying and technological integration is trivial.56 The existence of the more lenient legal standard for technological integration suggests that software sellers can, in effect, evade application of the per-se rule to their tying decisions. The more interesting question is what this evasion problem implies for the law. Does it imply that the contractual tying case law should be applied to software integration, as the district court held in the famous Microsoft III case?57 Or does this imply that the contractual tying case law should be revised to match the technological integration standard? 55 56

57

Id., at 1330. Brief of Professor Lawrence Lessig as Amicus Curiae, United States v. Microsoft Corp., 87 F. Supp. 30 (D.D.C. 2000) (No. 98-1233) (filed Feb.1, 2000). United States v. Microsoft Corp, 87 F. Supp. 2d 30 (D.D.C. 2000) (Conclusions of Law). For the appeals court decision, see United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001). The appeals court decided to apply rule of reason analysis to the bundling of a software application and an operating system.

VI. Exclusive Dealing

303

vi. exclusive dealing A. Introduction Clayton Act Section 3 applies to exclusive dealing arrangements, where, for example, one seller agrees to meet the requirements of a buyer at a fixed price. Why should antitrust law be concerned with this at all? The argument that appears in the cases is that the contracts foreclose the market from competitors. Note that this is foreclosure of a vertical relationship: it blocks competitors from dealing with a retailer or supplier. Of course, every exclusive vertical relationship, whether arranged through contract or through integration, forecloses someone else from the same relationship. So what is the problem? On closer examination it seems that two theories appear in the cases. One is that exclusive dealing creates barriers to entry. By foreclosing a substantial share of the market, they force competitors to enter at two levels. The second argument, stressed more in the academic literature than in the cases, is that exclusive dealing contract raise the costs of competitors by forcing them to go elsewhere, to presumably less desirable sources, to meet their input requirements. As in much of the preceding discussion, I will focus on the law here, and take up economic issues only when they arise in connection with the cases. The reason for this approach is that the anticompetitive and procompetitive theories of exclusive dealing are so numerous that it is difficult to state a useful set of conclusions about the consumer welfare implications.58

B. Law Because the language of Clayton Act Section 3 applies equally to tying and exclusive dealing arrangements, the statement of a per-se standard in International Salt had the immediate implication that a per-se standard also applied to exclusive dealing arrangements. That issue was examined in Standard Oil Co. of California (Standard Stations) v. United

58

On the theoretical issues, see B. Douglas Bernheim and Michael D. Whinston, Exclusive Dealing, J. Pol. Econ., vol. 106, pp. 64–103 (1998); Jean Tirole, The Theory of Industrial Organization 185–86 (Cambridge, Mass: MIT Press, 1988); Howard P. Marvel, Exclusive Dealing, J. Law & Econ., vol. 25, pp. 1–25 (April 1982).

304

Tying and Exclusive Dealing

States.59 Standard Oil was the largest seller of gasoline in the western United States, with roughly 23 percent of total gallons in the western market. It entered into exclusive dealing arrangements with 16 percent of the independent retail gasoline outlets in the area. The sales of these independent outlets amounted to $58 million, which constituted 7 percent of the total retail gasoline sales in the area. The issue before the Court was whether the requirement, under Section 3 of the Clayton Act, of showing an actual or potential lessening of competition or tendency to establish a monopoly is met by showing that the “quantitative substantiality” test of International Salt is satisfied. In other words, is “not insubstantial” foreclosure resulting from an exclusive dealing arrangement sufficient to find a violation of the Clayton Act? If the test of International Salt were applied to this case, violation would follow immediately from the fact that 7 percent of or $58 million of business in the retail gasoline market had been foreclosed by the exclusivity contracts. Although the Court affirmed the lower court’s finding of a violation (which was based on the legal conclusion that the required showing under Section 3 is met by proof of quantitative substantiality), it held that economic effects in the foreclosed market must also be considered in an exclusive dealing case. Why? The Court distinguished tying arrangements and exclusive dealing agreements on the ground that the former seldom have procompetitive or efficiency justifications while the latter often do.60 Exclusivity contracts assure stable supply, afford insurance against price increases, enable long-term planning, and reduce transaction costs such as the cost of storing goods. From the seller’s point of view, they may reduce selling expenses by switching to large infrequent orders of a given size, and provide insurance against price movements. For this reason, the Court held that the standards of proof should differ even though the wording of the relevant part of the Clayton Act is the same. Something more than mere quantitative substantiality in the foreclosed market must be shown in order to meet the standard of proof in an exclusive dealing case. 59 60

337 U.S. 293 (1949). Of course, this proposition is hard to defend on economic grounds. In the previous part of this chapter, we examined several efficiency justifications for tying arrangements. And it is difficult to say whether exclusive dealing is on balance more likely to be procompetitive than is tying. For an economic analysis of exclusive dealing, see B. Douglas Bernheim and Michael D. Whinston, supra note 58.

VI. Exclusive Dealing

305

The Court suggested some of the matters that should be considered in addition to the amount of foreclosure: (1) Whether competition has flourished in spite of the contracts, (2) Whether the contracts are unusually long, (3) Whether the defendant is a recent entrant into the market, (4) Whether the defendant has market power. Of course, a requirement that courts carefully examine market power and economic effects in the foreclosed market would be equivalent to adopting a rule of reason standard. This the Court rejected on the basis of the legislative history of the Clayton Act. Recall that the Clayton Act was a reaction to the Court’s 1911 Standard Oil decision, which held that only unreasonable restraints of trade violate the Sherman Act. The purpose of the Clayton Act was to avoid the reasonableness inquiry in certain areas by announcing bright line rules and preempting monopolization by certain practices before it reached an advanced stage. In a confusing passage, the Court concluded that “the qualifying clause of section 3 is satisfied by proof that competition has been foreclosed in a substantial share of . . . commerce affected.”61 However, this passage must be read in light of the facts, given the preceding discussion. Key to the Court’s conclusion were the facts that Standard had a dominant position, the market was concentrated, and that the contracts were used by Standard’s major competitors. Given this, the contracts created a “clog on competition” that had a much greater impact on competition, when considered in combination, than the 7 percent foreclosure by Standard’s contracts suggested. The Standard Stations opinion reveals an unsteady, almost schizophrenic, treatment of the legal standard governing exclusivity contracts. The Court is aware of the economic benefits that it cites in support of such contracts. The clear potential for economic benefits suggests that a reasonableness inquiry should govern. However, the legislative history of the Clayton Act stands as a barrier. Until the Clayton Act is repealed, any holding that rule of reason analysis applies will have to be regarded with some suspicion, unless the Court rejects its holding in Northern Pacific that the legal standards under the Sherman and Clayton Acts are the same (which would permit the Court to adopt a rule of reason test under the Sherman Act and something different under the Clayton Act). In spite of this, the Supreme Court held that rule of reason analysis applies to exclusivity contracts in Tampa Electric Co. v. Nashville Coal 61

337 U.S. at 314.

306

Tying and Exclusive Dealing

Co.62 A coal supplier asked the Court to rule that its agreement to fill an electric utility’s total coal requirements for twenty years was illegal under Section 3 of the Clayton Act, and that the contract was therefore unenforceable. The trial court had granted summary judgment in favor of the supplier. The Supreme Court reversed, and suggested that the trial court should have upheld the contract. The reasons were as follows. First, the share foreclosed in the relevant market was less than 1 percent. More attention was given to definition of the relevant market than in the earlier decisions. Here the Court defined the relevant market as the region in which sellers compete for the buyer’s business. Second, the Court held that requirements contracts are not illegal per se. Third, the Court noted several features distinguishing this case from earlier decisions; that the party responsible for the foreclosing did not have a dominant position in market, exclusivity was not an industry-wide practice, and there was no tying involved. In addition, the Court noted the economic benefits of requirements contracts. Courts have cited Tampa Electric for the proposition that rule of reason analysis applies to exclusivity agreements.63 This conclusion, although clearly right on economic grounds, should be viewed as somewhat infirm on legal grounds. The Clayton Act still stands on the books. As long as that is true, courts are bound to take into consideration the theory and purpose of that statute, as recognized by earlier courts. That theory, it is well known, rejects the rule of reason in the areas enumerated by the Act. Uncertainty regarding the proper standard is reflected in the Supreme Court’s decision in Federal Trade Commission v. Brown Shoe Co.64 Brown owned retail shoe stores and had entered into agreements with other stores that required them to concentrate on Brown shoes in return for Brown’s provision of low-cost insurance, merchandising records, architectural plans, and the services of a Brown field representative. The FTC found the stores that were parties to the agreement were effectively denied to Brown’s competitors, that such stores amounted to about 1 percent of the nation’s shoe stores, and that this foreclosure was

62 63

64

364 U.S. 320 (1961). See, for example, Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 2 (1984) (O’Connor, J., concurring). 384 U.S. 316 (1966).

Appendix

307

“significant.” The Commission held that Brown violated Section 5 of the FTC Act. The Supreme Court upheld the Commission on the ground that the FTC has broad powers to declare trade practices unfair. One may read Brown Shoe as simply reflecting the long-recognized power of the FTC to declare acts unfair even when it is doubtful that they would violate either the Clayton or the Sherman Act. But the Court also said that Brown’s acts were inconsistent with the policies under Section 1 of the Sherman Act and Section 3 of the Clayton Act.65 How should one interpret this in light of Tampa Electric? Perhaps the only way to make sense of FTC v. Brown Shoe is to note that the Clayton Act itself has been understood as authorizing the courts to take an aggressive approach toward finding violations, and that such an approach is necessarily inconsistent with a reasonableness inquiry.

appendix The point of this note is to illustrate a few of the difficulties in the anticompetitive theory of tying. Suppose a firm possesses a monopoly in the market for widgets (tying product). Let the price of a widget be pw, and let the quantity of widgets be qw. The tied product is called a gadget. Let the price of a gadget be pg and the quantity of gadgets be qg.66 Assume the market for gadgets is competitive. Suppose consumers desire both widgets and gadgets, and the goods are sold by the monopolist in the proportional relationship qg = j(qw)qw. I will also assume that the market demand curves are independent, in the sense that the overall market demand for widgets does not depend on the price of gadgets and vice versa. Let the total cost of producing widgets be cw(qw), and the total cost of producing gadgets be cg(qg). The profit of the monopolist is pw q w + p g q g - cw (q w ) - c g (q g ). 65 66

Id. at 320. The price of the tied gadget, in this analysis, can be understood as the implicit price in settings where the tying monopolist does not state a separate price for the tied product. In many of the tying cases, the defendant ties separate products together while at the same time listing a separate price for each one – see, for example, Jerrold. In other cases, the consumer can only guess at the implicit price of each item of the bundle.

308

Tying and Exclusive Dealing

If the monopolist sets the price for the tied good above the competitive level, consumers may perceive the monopolistic surcharge on the tied good as a tax on their purchases of the monopoly good. The surcharge can be expressed as pg - cg( qˆ g)/ qˆ g, where the latter term is the average cost of gadgets evaluated at the level that minimizes average cost, qˆ g (which is the competitive price level). If consumers are aware of the surcharge, their bids for the monopoly good fall from pw to pw [pg - cg( qˆ g)/ qˆ g]j(qw). Substituting, the monopolist’s profit is g g È c g (q ) c g (qˆ ) ˘ pw q w - cw (q w ) - q g Í g g qˆ ˙˚ Î q

which is bounded above by pwqw - cw(qw), the profit of the monopolist in the absence of tying. Thus, if we assume the demand curves for the tying and tied goods are independent and that the market for the tied good is competitive, a policy of tying cannot enhance the monopolist’s profits, as long as the consumers correctly perceive the difference between the price the monopolist sets for the tied good and its competitive price. If consumers do not desire the tied good, the result is even more pessimistic, from the perspective of the tying monopolist. When consumers do not desire the tied good, the entire price for the tied good (not merely the monopoly surcharge) is viewed as a tax on the purchase of the tying good. Given this, tying merely transfers demand from the tying good to the tied good. If the tied good is costly to produce, tying can only reduce the monopolist’s profits. These conclusions summarize a substantial part of the “Chicago School” critique of the anticompetitive theory of tying. The core of the Chicago School critique is that tying cannot be used to enhance monopoly profits when the market for the tied good is competitive. The analysis here also suggests that one needs to consider cases in which the tied product market is not competitive in order to find plausible theories of anticompetitive tying. Whinston has shown that when the tied product market is not competitive, tying can be used to deter entry in the tied product market, thus enhancing the overall market power of the tyinggood monopolist.67

67

Whinston, supra note 5.

Appendix

309

In addition to the entry deterrence theory, the analysis here suggests another way in which tying may enhance a firm’s market power. If, as Craswell has explained,68 tying can be used to obscure the competitive price of the tied component, then a monopolist may be able to enhance its overall power through tying. If consumers are unable to determine the competitive price of the tied product, they may underestimate the size of the tax effectively imposed on them by the tie-in. Under this information theory, tying may enhance the firm’s profits from sales to consumers who desire both products, though the firm must forgo sales to some consumers who do not desire the tied good. To see this in terms of the analysis presented above, consumers may be confused about the taxing effect of a tie-in when they have difficulty assessing the difference between the implicit charge for the tied good and its competitive level. This is more likely when the monopolist does not offer separately and charge a separate price for the tied good; the functional relationship between sales of the tied and tying goods is difficult to estimate (i.e., consumers cannot easily determine j(qw)); and the tied good market is not competitive. The last two conditions may have been present in the Eastman Kodak case, where the long-term nature of the parts-service tie made it difficult for purchasers to know the functional relationship between sales of the two, and the absence of a competitive service market after the tie-in was imposed made it difficult to determine the competitive price of service. However, to say that the information theory is more plausible in these settings does not imply that it is a persuasive version of the anticompetitive tying theory. If consumers find it difficult to estimate the implicit tax imposed by the tie-in, they may overestimate rather than underestimate the tax. If consumers tend to overestimate the implicit tax, tying would clearly hurt the monopolist. Moreover, if the tie-in damages the reputation of the seller, it may be self-defeating. Though the value of the information theory is ultimately an empirical question, one can use it to reconcile the holdings of Jefferson Parish and Eastman Kodak. If, as the Court held in Jefferson Parish, consumers are too uninformed to determine the implicit tax due to a tie-in, it is unlikely that tying could be used to enhance a firm’s market power. The reason is that consumers are likely to purchase the tying product and the most convenient complement whether or not the seller imposes a tie-in. For example, even if East Jefferson Hospital allowed patients to bring in 68

Craswell, supra note 7.

310

Tying and Exclusive Dealing

their own anesthesiologists, very few would have done so and the vast majority would have relied on the anesthesiologists provided by the hospital. On the other hand, in Eastman Kodak, where the firm imposed a tie-in on the parts/service after-market as a change in its selling policy, there is a theoretically plausible account in which the firm made the change in order to exploit the difficulty consumers would have, after the change, in assessing the difference between the price charged for service and its competitive level.

15 Horizontal Mergers

This chapter discusses the economics of mergers and traces the development of horizontal merger doctrine. A horizontal merger may yield efficiency gains by cutting production or marketing costs, and at the same time may reduce society’s wealth to the extent that the merged entity sets price above the competitive level. An economic reasonableness standard would assess these welfare tradeoffs in determining whether a horizontal merger should be deemed anticompetitive. However, as we have seen in other areas of antitrust, a reasonableness test places heavy administrative burdens on enforcement agencies. The tension between administrative and economic reasonableness concerns is particularly noticeable in the development of horizontal merger doctrine. After announcing a truncated reasonableness test in its 1962 Brown Shoe opinion,1 the Supreme Court moved away from it in subsequent decisions, toward a standard that put primary emphasis on market structure. This drift continued until 1974, when the Court signaled a return to the Brown Shoe standard in its General Dynamics2 decision.

i. reasons for merging and implications for law Why is antitrust law concerned about mergers? Two reasons appear in the literature. (1) Concentration, which may lead to tacit or explicit collusion. (2) Foreclosure; a forward vertical merger may foreclose retail 1 2

Brown Shoe Co. v. United States, 370 U.S. 294 (1962). United States v. General Dynamic Corp., 415 U.S. 486 (1974).

311

312

Horizontal Mergers

outlets to other manufacturers, or a backward vertical merger may reduce the number of suppliers to a given manufacturer. The concentration issue forces us to return to the conscious parallelism debate of Chapter 3. The notion that high levels of concentration are likely to lead to high prices is generally accepted and seems to be an operating assumption of enforcement agencies.3 However, the link between concentration and collusion has not been proved conclusively.4 Foreclosure is a potential problem because it may result in cost increases to rivals or reduce incentives to enter. As we have seen (Chapter 10), neither of these arguments is uncontested. Why do firms merge? A few prominent reasons are as follows. One, and perhaps the oldest explanation, is that firms merge in order to take advantage of economies of scale or scope (or “synergy”).5 An equally ancient theory is that firms merge in order to gain monopoly power. A third reason is to enter a new market, by purchasing a business as a going concern rather than acquiring assets or building equipment and structures from the ground up. A fourth is that firms merge in order to create an internal capital market;6 because a firm (it is argued) can more easily cross-subsidize risky portions of its operations and lower the cost of capital through the creation of an internal market for finance.7 A fifth 3

4

5

6

7

The theory that high concentration and profit levels are positively correlated is a prediction of the structure-conduct-performance model of Bain, see Joe S. Bain, Industrial Organization (1959). For more recent evidence on the concentration-profits relationship, see Leonard J. Weiss, The Concentration-Profits Relationship and Anti-Trust, in Harvey J. Goldschmid, H. Michael Mann, and J. Fred Weston, editors, Industrial Concentration: The New Learning (Boston: Little, Brown, and Company, 1974); David J. Ravenscraft, Structure-Profit Relationships at the Line-of-Business Level, 55 Review of Economics & Statistics 22–31 (February 1983); Michael Smirlock, Thomas Gilligan, and William Marshall, Tobin’s q and the Structure-Performance Relationship, American Economic Review 1051–60 (December 1984). Harold Demsetz has presented empirical evidence that is inconsistent with the concentration-generates-collusion hypothesis. See Harold Demsetz, Two Systems of Belief about Monopoly, in Harvey J. Goldschmid, H. Michael Mann, and J. Fred Weston, editors, Industrial Concentration: The New Learning 164–84 (Boston: Little, Brown, and Company, 1974); see also Harold Demsetz, Industry Structure, Market Rivalry, and Public Policy, 16 Journal of Law & Economics 1 (April 1973). For example, Richard Brealey and Stewart Myers, Principles of Corporate Finance 821 (McGraw-Hill, 4th ed. 1991). Id. at 827–8; Oliver E. Williamson, Corporate Control and Business Behavior: An Inquiry into the Effects of Organization Form on Enterprise Behavior (Englewood Cliffs, NJ: Prentice Hall, 1970). Of course, if the purpose of the merger is simply to “cross-subsidize,” then there is no net gain resulting from the merger. Bondholders pay a lower interest rate, but stockholders increase their liability, see Brealey and Myers, supra note 5, at 827. A better theory

I. Reasons for Merging and Implications for Law

313

reason is to take advantage of tax law; for example, one firm has losses that can offset the taxable income of another in a way that reduces the joint tax bill by an amount that exceeds the transaction costs of the merger.8 A sixth reason is diversification.9 Of course, investors can diversify on their own, but if the firm diversifies it provides some protection to the firm-specific human capital of managers, and to the extent the bankruptcy process reduces society’s wealth, this type of merger may be desirable. Managerial goals make up another broad category of reasons. One view sees these goals as pernicious in effect: managers seek the higher salaries, greater prestige, and perquisites that accompany leadership roles in large corporations.10 The larger the firm, the greater the pay and benefits generally. An alternative managerial goal that causes firms to acquire others is to displace incompetent management (liberating owners),11 or to free competent management from incompetent owners (liberating managers). Yet another theory is that mergers allow managers to transfer rents from various fixed claimants, such as debt holders and employees, to the shareholders.12 As this brief survey suggests, mergers occur for good and bad reasons. In some cases there will be efficiency gains, in others not. In light of the range of theories, one should be surprised to find that the evidence on mergers provides support to only one theory. As it happens, the evidence is mixed. For example, for the United States, accounting profit data indicate that average or below average performing firms tend to acquire above average performing firms.13 Thus,

8 9 10

11

12

13

would hold that the internal capital market is better able to monitor and assess the validity of certain claims of debtors within the internal market than is possible in the “external” capital market. Brealey and Myers, supra note 5, at 822–3. Id. at 824–5. I refer generally to the agency cost literature, see Jensen and Meckling, Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure, 3 Journal of Finance 305 (1976); Oliver E. Williamson, The Economics of Discretionary Behavior (1964). Henry G. Manne, Mergers and the Market for Corporate Control, 73 Journal of Political Economy 110–20 (April 1965). Andrei Shleifer and Lawrence Summers, Breach of Trust in Hostile Takeovers, in Corporate Takeovers: Causes and Consequences (A. Auerbach ed. 1988). It should be noted that this theory is inconsistent with the notion that shares of stock are priced efficiently. If managers transferred rents in a manner that reduced productive efficiency (e.g., by discouraging investment in firm-specific skills), then such transfers would impoverish shareholders in the long run. See Dennis C. Mueller, The Determinants and Effects of Mergers 296 (Oelgeschlager, Gunn, & Hain: 1980); id., The Effects of Mergers, in Economic Analysis and Antitrust

314

Horizontal Mergers

the evidence provides little direct support for the claim that the typical merger occurs because the managers at one firm realize that efficiency gains would result from ousting the less competent managers of another. Of course, this does not prove that efficiency-enhancing management substitutions do not take place; it simply suggests that the sample of observations is varied. Moreover, the results are inconclusive because they rely on accounting data, which are subject to measurement and reporting choices, and fail to reflect the economic variable of most importance, “economic profitability,” that is, the extent to which revenue exceeds the opportunity cost of resources used. What about sales or profitability following the merger? The data suggest that mergers do not positively influence these variables.14 Perhaps the most consistent finding relates to shareholder returns. Acquiring firm shareholders suffer a fall in value, while acquired firm shareholders experience a substantial gain.15 Because acquiring firms are usually larger than acquired firms, the net wealth effect suggested by this finding is either negative or neutral. The empirical literature on mergers is large and growing, but perhaps all it reveals is the complicated state of the world. The suggestion that mergers are wealth reducing is open to question, because it is almost impossible to provide a reliable assessment of the long-term effect of a merger. One problem is determining the baseline to which comparisons should be made. How can one say that a merger reduced the wealth of acquiring firm shareholders without having accurate projections of the future income of the acquiring firm – without the acquired entity? In addition, if a significant portion of shareholders are uninformed, short-run effects on stock prices may not accurately reflect changes in fundamental value.16 Finally, what seems to hold on

14

15

16

Law 307 (Terry Calvani and John Seigfried, 2d ed., 1988). Mueller noted that acquiring firms were no more profitable than comparably sized firms in their industry; see Mueller, The Determinants and Effects of Mergers, at 296. However, two earlier studies cited by Mueller indicate that acquiring firms in the 1960s were less profitable than comparable nonmerging firms, J. F. Weston and S. K. Mansinghka, Tests of the Efficiency Performance in Conglomerate Firms, 26 Journal of Finance 916 (1971); R. W. Melicher and D. F. Rush, Evidence on the Acquisition-Related Performance of Conglomerate Firms, 29 Journal of Finance 1941 (1974). See, for example, Mueller, The Effects of Mergers, in Economic Analysis and Antitrust Law 307–8 (Terry Calvani and John Seigfried, 2d ed., 1988). See, for example, id. at 310–16; Peter Dodd and Richard Ruback, Tender Offers and Stockholder Returns: An Empirical Analysis, 5 Journal of Financial Economics 105–37 (June 1980). This is certainly implied by research on stock market volatility. See Robert J. Shiller, Do Stock Prices Move too Much to be Justified by Subsequent Changes in Dividends? 71

I. Reasons for Merging and Implications for Law

315

average may not tell us much about the effects of mergers in particular settings. In light of the range of theories of the merger process, one question that recurs in various guises is the extent to which antitrust courts should permit defendants to rely on efficiency defenses in merger cases. The Supreme Court decisions have been conservative, showing reluctance to apply a full-blown rule of reason standard. Oliver Williamson has argued that efficiency defenses should receive a full and fair hearing.17 Williamson’s argument is based on the “welfare tradeoff” model.18 The key implication of the model is that, under reasonable assumptions, a modest efficiency gain will outweigh the increase in deadweight loss resulting from a merger that enhances the monopoly power of the merged firm. Suppose two firms merge creating a new firm that is a monopolist. Suppose further that before the merger, the two firms were earning competitive returns (i.e., price was equal to average cost for both firms). If the monopoly creates efficiencies that allow the merged firm to reduce total cost, then the efficiency gain can be represented by the area DGBH in Figure 15.1. The increase in deadweight loss is given by the triangle EGC. Williamson has argued that (1) under the assumption that the threat of entry from geographically remote rivals constrains the merged entity’s ability to increase price and (2) under reasonable assumptions about the elasticity of demand, the gain in efficiency is likely to exceed the increase in deadweight loss.19 In particular, he shows that for price increases up to 10 percent, and demand elasticities as high as 2, relatively small cost reductions (less than 2 percent) are sufficient to outweigh the harmful effects of the price increase. It may be incorrect to infer from Williamson’s analysis the general proposition that efficiency gains from cost-reducing mergers outweigh the potential deadweight costs.20 In the appendix, I present an

17

18

19

20

American Economic Review 421–36 (June 1981); id., Stock Prices and Social Dynamics, Brookings Papers on Economic Activity 457–98 (Fall 1984). Oliver E. Williamson, Economies as an Antitrust Defense Revisited, 125 U. Pa. L. Rev. 699, 703–13 (1977). Oliver E. Williamson, Economies as an Antitrust Defense: The Welfare Tradeoffs, 58 Am. Econ. Rev. 18 (1968). Williamson, supra note 17. Williamson also argues that it is proper to assume that the threat of entry from geographically remote rivals constrains the merged entity from charging the monopoly price. For a careful reexamination of Williamson’s argument that reaches different conclusions, see Michael E. DePrano and Jeffrey B. Nugent, Economies as an Antitrust Defense:

316

Horizontal Mergers

Figure 15.1

examination of the welfare tradeoff question in a simple model. I examine the conditions under which the welfare gain exceeds the deadweight loss, assuming the merged entity charges the monopoly price. I find that when the point elasticity of demand for the monopolist’s product is equal to 2, a 21 percent cost reduction is necessary for the efficiency gain to outweigh the deadweight loss.21 When the point elasticity is 3, an 11 percent cost reduction is sufficient. Of course, the assumptions I adopt differ from Williamson’s, and it may be that Williamson’s are more representative of the real world. However, it seems fair to conclude that the general proposition that efficiency gains from cost-reducing mergers exceed associated deadweight costs requires either (1) the assumption that the merged entity is unable to fully exploit its monopoly power (e.g., Williamson’s assumption that geographically remote rivals constrain the merged firm’s pricing), or (2) the assumption that the merged firm has little potential monopoly power to exploit.

21

Comment, 59 Am. Econ. Rev. 947–53 (1969). In particular, DePrano and Nugent show that it is not clear, as a general matter, that only modest efficiency gains are needed to offset the deadweight loss due to monopolization. My discussion in the appendix provides a more transparent version of the analysis in DePrano and Nugent, supra note 20.

II. Horizontal Merger Law

317

The problems with providing a full and fair hearing to the efficiency defenses are as follows. First, enforcement is made more difficult because courts have to consider all of the reasonableness arguments. This is an old problem, and it appears in all areas of antitrust enforcement. The experience is sobering. Under the rule of reason applied before the Supreme Court adopted a per-se ban on trade restraints in Trans-Missouri, the government fared poorly because of the difficulty in proving economic unreasonableness. In addition, the Sherman Act enforcement regime was not comparable to the regime of contract disputes heard in common law courts. In the regime of contract disputes, the litigants were parties to the agreement, and therefore knew something about its purpose and effect. The government has considerably less information. The rule of reason might be assumed to provide no advantage to either party in the regime of contract disputes. But under the Sherman Act, a rule of reason test places an enormous burden on the enforcement agency (see Chapter 5). Nothing I have said so far suggests that on social welfare grounds the government should face an easier burden of proof. However, the relentless pressures of the enforcement agent have proven difficult for the Supreme Court to withstand. Thus, as a positive matter, one should predict that over the long run, an antimerger statute that the government must enforce will retain substantial portions of per-se liability. The second problem with allowing a full-blown reasonableness inquiry is suggested by legislative history. The Clayton Act was Congress’s reaction to the 1911 Standard Oil decision, which declared that the rule of reason applies to all questions under the Sherman Act. It would fly in the face of that legislative history to apply a full-blown rule of reason standard under the Clayton Act.

ii. horizontal merger law In Northern Securities Company v. United States,22 decided in 1904, the Supreme Court said that all mergers between directly competing firms constitute a combination in restraint of trade and therefore violate Section 1. The approach reflected the state of Section 1 doctrine at the time. The per-se rule of Northern Securities was abandoned in 1911 with the Standard Oil decision. The rule of reason was applied to a merger in United States v. United States Steel Corp.,23 where the Court held that the 22 23

193 U.S. 197 (1904). 251 U.S. 417 (1920).

318

Horizontal Mergers

consolidation of most of the industry into one firm possessing 80 to 90 percent of the market in some lines did not violate either Section 1 or Section 2 of the Sherman Act. U.S. Steel was decided after enactment of Section 7 of the Clayton Act, so one might wonder why the Court did not rely on the Clayton Act in its decision. The reason is that Clayton Act Section 7 banned only mergers effected through stock transfers, and later court decisions further limited its scope.24 The acquisitions leading to the buildup of the U.S. Steel Corporation were largely asset acquisitions (i.e., mergers effected through asset acquisitions). The restrictions on Section 7, the government’s loss in the United States v. Columbia Steel Co.25 case, and a 1948 FTC report suggesting that giant corporations would soon take over the country led Congress to amend Section 7 in 1950. Section 7 now covers all types of acquisition.

A. Rule of Reason and Incipiency Doctrine The first important Supreme Court decision in the new life of Section 7 was Brown Shoe Co. v. United States.26 The acquisition of Kinney by 24

25

26

In Thatcher Mfg. Co. v. FTC, 272 U.S. 554, 560–1 (1926), the Supreme Court held that the FTC could not restrict the transfer of assets following a stock acquisition when the transfer occurred before the filing of the FTC’s complaint. If the FTC was able to file the § 7 complaint before the stock transfer it could bar the acquiring company from transferring the stock or assets of the acquired company to a subsidiary, FTC v. Western Meat Co., 272 U.S. 554, 561–3 (1926). The Court later ruled that § 7 did not apply to stock acquisitions in which the acquired company was on the verge of insolvency, International Shoe Co. v. FTC, 380 U.S. 291 (1930). The Court found that the two companies were competitors, but not “substantial competitors.” It also found that the acquired company was about to go bankrupt and had no other potential buyers. Certain other asset transfers arising out of stock transfers were also removed from the scope of § 7. In Arrow-Hart & Hegeman Electric Co. v. FTC, 291 U.S. 587 (1933), the FTC brought a complaint against a holding company for its ownership of two manufacturing companies, which were competitors. After the FTC’s complaint was filed, two new holding companies were created, each owning only one of the manufacturing companies. Soon after that, all four companies (the two holding companies and the two manufacturing companies) were merged into a single company. The Court held that the FTC had no authority to enjoin the new holding company because § 7 applies only to stock ownership transfers, not to mergers pursuant to state laws,291 U.S. at 595.The new holding company was outside the scope of the FTC’s action because it never owned the manufacturing companies’ stock. 334 U.S. 495 (1948). The Court held that a merger between two producers of steel fabrications that created a 24 percent share firm in the western fabrication market (combining 13 and 11 percent firms) did not substantially lessen competition. The reasons were that the western fabrication market was expanding and the evidence suggested that some of the firms were competing on a national scale. 370 U.S. 294 (1962).

II. Horizontal Merger Law

319

Brown resulted in a horizontal merger at both the manufacturing and retail levels. The district court held that the merger was too insignificant at the manufacturing level to violate Section 7, and the government did not appeal. However, Brown contested the district court’s finding that the merger of retail outlets could substantially lessen competition. The two would control 7.2 percent of all shoe stores and 2.3 percent of all retail shoe outlets. The issue was whether the merger violated Section 7 of the Clayton Act – that is, whether it may tend to lessen competition substantially. While finding a violation, the Court announced something close to a rule of reason approach. Under the test of Brown Shoe, the following questions are relevant. Is the market concentrated? Has there been a trend toward concentration? To what extent have markets been foreclosed? Is entry easy? The Court noted that statistics on market share do not conclusively indicate anticompetitive effects. After articulating a rule of reason standard, the Court set it aside and ignored it for the remainder of the opinion. The Court held that the record supported the district court’s finding that the relevant geographic market was the set of cities of over ten thousand in which both Brown and Kinney operated retail stores. In this geographic market the Court considered the probable effect of the merger. In 118 cities falling within the geographic market definition, the Brown and Kinney combined market share exceeded 5 percent in one of the product lines (men’s, women’s, and children’s shoes).27 Although Tampa Electric28 would suggest that 5 percent foreclosure is insignificant, the Court found that this level of foreclosure was evidence of a substantial lessening of competition. The reason was that “in an industry as fragmented as shoe retailing, the control of substantial shares of the trade in a city may have an important effect on competition.”29 In addition, “[i]f a merger achieving 5 percent control were approved, we might be required to approve future merger efforts.”30 The Court also cited evidence of a historical trend toward concentration. This is the Court’s first statement of the incipiency doctrine, which holds that a horizontal merger may violate Section 7 even though the share of the market absorbed is relatively small, because the merger is

27 28

29 30

Id. at 343. Tampa Electric Corp. v. Nashville Coal Co., 365 U.S. 320 (1961); for discussion, see Chapter 14. Brown Shoe, 370 U.S. at 343. Id. at 343–44.

320

Horizontal Mergers

an indicator of a movement toward concentration. On theoretical grounds, the incipiency theory is supported by the notion that mergers are difficult to undo, and thus it is wiser to block them before monopolization results rather than wait for monopolization to occur before acting. On legislative grounds, the theory is supported by the notion that the 1914 legislative package (the FTC and Clayton Acts) was designed to set by bright line rules and promote aggressive enforcement. Brown Shoe does not formally say that the “incipiency test” applies to mergers, but this is the test the Court applied. In support of the theory the Court pointed to a trend toward concentration in the shoe industry. Finally, the Court discussed mitigating factors: (1) business failure, (2) inadequate resources that may have prevented one of the firms from maintaining its competitive position, and (3) the need for combination to enable small companies to enter into more meaningful competition with larger ones. One may be inclined to ask whether there is any difference between the first and second mitigating factors. The difference is this: the first excuses mergers in which one of the firms showed signs of imminent failure. The second excuse applies to firms that cannot make investments in technology, services, or appearance and as a result, cannot present themselves as a significant competitive force. The list of mitigating factors does not include economies of scale, though one might be able to push a scale economies argument under one of the mitigating factors. However, in general Brown Shoe leaves little room for efficiency defenses. Indeed, one of the justifications the Court offered for its holding is that Congress, in enacting the Clayton Act, spoke warmly of the desirability of retaining numerous small, independent businesses, even if that resulted in “occasional higher costs and prices.”31 On technical grounds, the most troubling feature of Brown Shoe is the treatment of the relevant geographic market. The Court’s approval of the lower court’s choice of cities with population over ten thousand implies that a relatively minor suburb of a large city would count as a contained geographic market. This is unrealistically narrow. A shopper in Skokie, Illinois, who finds prices too high in that city could easily drive to Chicago or to one of the many nearby suburbs to compare prices. A city as small as ten thousand would make a contained geographic market only if it were not surrounded by another city or shopping area for several miles. 31

Brown Shoe, 370 U.S. at 344.

II. Horizontal Merger Law

321

The incipiency doctrine itself is highly questionable. It is based on the simple proposition that, other things being equal, a large number of relatively small competitors is preferable to high levels of concentration. But in the presence of scale economies, this proposition is invalid. If firms merge in order to take advantage of scale economies, the result in a competitive market will be an increase in competition, in the sense that prices will fall, and an increase in concentration. True, it is difficult and costly to undo a merger, and this reasoning has supported the notion that the incipiency doctrine saves judicial resources. The flaw in this argument is that it compels us to forgo a certain gain today in order to avoid a very uncertain, quite speculative harm in the future. The ultimate aim of antitrust law is to benefit consumers. To make the minimization of administrative costs an overriding concern is to lose sight of the ultimate end of the statute.

B. Toward an Incipiency Doctrine: Mergers in Concentrated Markets The second major horizontal merger decision was United States v. Philadelphia National Bank (hereafter PNB),32 which involved the acquisition of Girard Bank by Philadelphia National Bank. The Court applied Section 7 to enjoin the merger of the second and third largest commercial banks in the Philadelphia area. As in Brown Shoe, the Court took a two step approach to defining the relevant market: first, defining the relevant product market, then defining the relevant geographic market. Commercial bank services constituted the relevant product market, and the relevant geographic market included the city of Philadelphia and three contiguous counties. Also as in Brown Shoe, the Court’s choice of the relevant geographic market was important to the outcome and highly questionable. The market for bank services could be divided into three segments. One, the largest geographically, is the country, and includes large corporations that search nationwide for the best terms on a loan. A second geographic market is the “region,” which includes the state of Pennsylvania and nearby states. This is the market that regional companies would consider in their shopping for loans. A third geographic market is local, and is made up of the city of Philadelphia plus the three contiguous counties. This is the market in which consumers shop when comparing retail 32

374 U.S. 321 (1963).

322

Horizontal Mergers

banking services and fees. The Court’s choice of the local market was implicitly a decision to treat shopping by retail customers as the most important area of competition. The issue in PNB was whether the merger between banks would lessen competition substantially in the relevant market, and therefore violate Section 7. The Court held that it did, and in the course announced a new test. A merger that “(1) produces a firm controlling an undue percentage share of the relevant market, and (2) results in a significant increase in the concentration of firms in that market, is so inherently likely to lessen competition substantially that it must be enjoined in the absence of evidence clearly showing that the merger is not likely to have such anticompetitive effects.”33 This test declares a threshold, though a vague one, beyond which mergers are presumptively illegal and therefore do not require evaluation of economic impact. What is an “undue percentage”? What is a “significant increase in concentration”? These ambiguities are partially clarified by the facts in PNB: (1) the merger created a firm having 30 percent of the relevant market and (2) the shares of the two largest firms would jump from 44 to 59 percent. Furthermore, the Court noted there was no evidence of easy entry to offset these concerns. The Court suggests that evidence of easy entry could be used to offset the findings under the test. What about evidence that there is vigorous competition in the market, in spite of the high level of concentration? In theory, the Court ought to take such evidence into account. Recall that the price-marginal cost mark up formula of Ordover, Sykes, and Willig (Chapter 11)34 suggests that even in a market with high concentration, price will remain at the competitive level if there is vigorous competition among firms. On the other hand, if there is a noncompetitive culture of parallel pricing, the concentration levels will understate the degree of monopoly pricing. Although in theory courts should take evidence on competitive culture into account, PNB holds that such a defense is impermissible. The Court held that testimony by the officers of rival banks that vigorous competition existed was too unreliable to be counted as countervailing evidence. Given the number of markets in which banks compete, one obvious defense is that the merger purported to enhance competition at the 33 34

374 U.S. at 301. Janusz A. Ordover,Alan O. Sykes,and Robert D.Willig,Herfindahl Concentration,Rivalry, and Mergers, 95 Harvard Law Review 1857–64 (1982). For discussion, see Chapter 11.

II. Horizontal Merger Law

323

national level, where large New York City banks dominated. The Court held that the argument that a merger restricts competition in one market but enhances it in another is impermissible as a defense. The one exception to this is the third mitigating factor of Brown Shoe, which permits a merger if two small companies combine in order to compete against a larger one in their market. Here, the Court said that the merger involved two large companies who combined ostensibly to compete more effectively against even larger competitors outside of their market. This characterization is not persuasive in the end because it depends on the Court’s choice of the relevant geographic market. If the Court had expanded its definition of the relevant market to include the country, then the merger in PNB would have fit the description of an effort by two small competitors to compete against a larger competitor in their market. A third argument the defendants offered was that Philadelphia National merged with Girard in order to acquire suburban branches. In other words, the merger was really an effort by Philadelphia National to enter into a new market. The Court rejected this defense because of the existence of a preferable alternative, the construction of a new network of branches. Thus, although a common motivation for mergers is the desire to enter into a new market cheaply by acquiring existing facilities, PNB suggests that this justification is generally impermissible as a defense. The formal tests announced in Brown Shoe and PNB can be reconciled as follows: (1) Brown Shoe announces a rule of reason approach for unconcentrated markets; (2) PNB announces a threshold beyond which a horizontal merger is presumptively illegal, which applies in concentrated markets. The interesting feature of both cases is that although the Court announces a formal test, the incipiency doctrine plays a crucial role in the rationales for the decisions. The doctrine in practice therefore differs from the formal tests. One might justifiably append the following third prong to the statement reconciling the two opinions: (3) in any case with evidence of a trend toward concentration, the Court will find a violation on the basis of the incipiency test. PNB suggests that in cases involving a sufficiently high concentration level, courts will not accept evidence showing the existence of a competitive culture as a defense under Section 7. The holding in United States v. Aluminum Co. of America (Rome Cable)35 suggested the converse 35

377 U.S. 271 (1964).

324

Horizontal Mergers

proposition, that the courts could use evidence on competitive culture (i.e., the lack of one) to find a violation of Section 7 when the concentration threshold conditions of PNB are not satisfied. Alcoa made aluminum and aluminum conductor. Rome made copper conductor and aluminum conductor (and purchased aluminum from Alcoa). Alcoa acquired Rome. The merger was part vertical, part horizontal. It was horizontal in the aluminum conductor market. In the market for “bare and insulated aluminum conductor” Alcoa had a 27.8 share and Rome had a 1.3 share. This was held a “significant increase in concentration,” under the test of PNB because the market was concentrated, there was a trend toward concentration, and Rome was an aggressive competitor. As the Court’s rationale suggests, the incipiency doctrine played a role in the decision. The Court’s decision in United States v. Continental Can Co.36 suggested yet another wrinkle in the application of the PNB test. The Court enjoined the merger of the second largest can producer with the third largest maker of glass containers. In the combined “bottle or can” market, the can producer would have had a 22 percent share and the bottle producer a 3 percent share, implying a merged firm with at least a 25 percent share. Although the increase in concentration was not nearly as large as that in PNB, the Court’s decision was based in part on the observation that the bottle and can industries were relatively concentrated. The two largest can producers were responsible for 70 percent of sales, the three largest bottle makers 55 percent. One interesting question raised by Continental Can is what determines the relevant product market in a merger across competing industries. If a maker of airplanes merges with a maker of bicycles, the horizontal market power considerations will be minimal, since consumers rarely choose between airplanes and bicycles. At least with respect to users of the end product, the relevant question is whether they view the items as substitutes. The Court noted in Continental Can that the evidence indicated that bottles and cans were viewed by end users – producers of soft drinks, fruit juices, milk, and so on – as substitutes. As in Cellophane, there was historical evidence of substitution between products by end users. A more sophisticated approach would follow the 1984 Justice Department Merger Guidelines.37 Recall that the Guidelines method for iden36 37

378 U.S. 441 (1964). (1984) U.S. Department of Justice 1984 Merger Guidelines, 49 Fed. Reg. 26,823 (1984).

II. Horizontal Merger Law

325

tifying the relevant market requires an analysis of the effects of a significant, nontransitory price increase by a hypothetical monopolist (see Chapter 11). Thus, suppose a monopolist gained control of bottle production and raised prices by 5 percent for one year. If so many end users switched to cans that the price increase would be unprofitable, then bottles and cans would be considered within the same market. This exercise probably would have supported the Court’s conclusion in Continental Can that bottles and cans were in the same market. However, there are other methods of storing liquids – for example, plastic containers, boxes lined with flexible wrapping material (“boxed drinks”). If one were to take these into account in performing the Guidelines’ exercise, the likely conclusion would be that plastic containers, boxes, bottles, and cans all belong in the same market. And the market shares of the merging firms in Continental Can would have been small within this larger market.

C. High Point of Incipiency Doctrine The high point of the incipiency doctrine was United States v. Von’s Grocery Co.38 Von’s was the third largest grocery store chain in the Los Angeles market, and had 4.7 percent of all sales. It acquired Shopping Bag, the sixth largest retailer, and together they made up 7.5 percent of the market. The Court held that this merger violated Section 7 of the Clayton Act. The reason offered was the Court’s concern that the data revealed a trend toward concentration. The number of owners operating single grocery stores in the area decreased from 5,365 in 1950 to 3,590 in 1963. This high point in the incipiency doctrine coincides with the low point of the Supreme Court’s merger jurisprudence. One has to think that something is deeply wrong when a justice says in a published opinion that the only consistency he can see in the cases is that the government always wins.39 Is this the only consistency? No. The decision in Von’s was a foreseeable step along a logical progression starting with Brown Shoe. Indeed, Von’s follows immediately from Brown Shoe, or one might say that the holding of Von’s was implicit in Brown Shoe. Recall that the Brown Shoe Court justified the finding of a violation of Section 7, even though the degree of foreclosure was 38 39

384 U.S. 270 (1966). Id. at 301 (J. Stewart, dissenting).

326

Horizontal Mergers

roughly 5 percent, on the ground that approval in that case might require the Court to approve additional mergers. Obviously, if 5 percent foreclosure is evidence of a substantial lessening of competition, then the 7.5 percent of Von’s is also sufficient. Although it is understandable that someone could read Von’s as a flight from the path of the earlier cases, it is not. It reveals a flaw that was present at the start: the incipiency doctrine. Merger jurisprudence cannot be defended on economic reasonableness grounds until the Court places sharp limits on the scope of the incipiency doctrine, or abandons it altogether. This is a good time to return to a recurrent theme in this text. Reasonableness justifications are by no means necessary for a court to hand down conclusions on the legality of an individual’s conduct. In theory, a court could hold that the legal test applicable to mergers is the incipiency test, and reject any requirement that the doctrine be justifiable according to some economic reasonableness criterion. This is an option that has always been before the Supreme Court, and has been the source of recurring controversy in antitrust law. The Court confronted a similar choice when its Trans-Missouri opinion rejected the common law rule of reason as the legal standard under Section 1. The Court had the option then of stating the per-se rule as the legal test, and simply admitting that it was indefensible on economic reasonableness grounds. However, the Court eventually adopted the theory of Taft in his celebrated Addyston Pipe opinion. According to that theory, the common law rule of reason itself applied a per-se unreasonableness test to certain agreements. Thus, the Court’s solution to the first crisis of this sort in antitrust was to find a reconciliation between the common law test and the new statutory standard. This reconciliation eventually led the Court to incorporate economic reasonableness tests in more areas of antitrust law. Von’s confronted the Court with a minor crisis similar to that generated by Trans-Missouri. To remain with the incipiency test ultimately would require the abandonment of economic reasonableness justifications in merger law. The Court abandoned the incipiency test.

D. Rejection of Incipiency Doctrine The Court signaled its inevitable movement away from the incipiency doctrine in United States v. General Dynamics Corp.40 The Court affirmed 40

415 U.S. 486 (1974). For a change, the government did not win.

II. Horizontal Merger Law

327

the lower court’s approval under Section 7 of a merger between two coal producers. The market share statistics indicated that the market exhibited “undue concentration” and showed a trend toward concentration. While the merger did not produce the significant increase in concentration observed in PNB, it did not have a trivial impact on concentration. The issue was whether the lower court erred in finding that other factors affecting the coal industry indicated that this merger did not lead to a substantial lessening of competition. The Court held that the lower court was justified in taking other factors into account. The Court noted that Brown Shoe set forth a rule of reason test. Thus, as a matter of law, the lower court could not have been in error because it was simply applying the test of Brown Shoe. As a further justification for the lower court’s decision not to rely on statistical evidence alone, the Court distinguished the facts in General Dynamics from those of Von’s. In the grocery market, evidence on past production levels could be taken as valid predictors of future production or foreclosure levels. The coal market was different because much of the production was committed through long-term supply contracts with electric utilities. Uncommitted reserves provided a more reliable indication of a coal producer’s ability to compete for new contracts. The acquired company had relatively little in the way of uncommitted reserves. The remaining question was whether the evidence supported the lower court’s finding. The lower court found that the acquired firm’s coal reserves, combined with long-term contract commitments, indicated that it was not a significant competitive force in the market, and therefore its disappearance would not substantially lessen competition. The government argued that the lower court had essentially applied the “failing company defense,” and should have been overruled because it did so improperly. The failing company defense requires: (1) high probability of business failure and (2) no other prospective purchaser. The acquisition could not meet the requirements of the failing company defense because the acquired company was not failing and there was no proof that the acquirer was the only available buyer. But the lower court was not applying the failing company defense, the Court said, it was merely holding that the acquired firm was not a significant competitive force. While the failing company defense generally assumes a negative effect on competition, in this case there was not a negative effect because there was no lessening of competition. The government also argued that the lower court should not have relied on data on existing coal reserves. The company could go out and acquire more at any time. The Court sidestepped the thrust of this

328

Horizontal Mergers

argument by relying on the lower court finding that acquisition of new reserves was infeasible. A better response would have been to note the inconsistency between this argument and the government’s position that the statistics alone dispose of the case. If a company could acquire new reserves or learn to deep-mine old sites that had been strip-mined, then the obvious implication is that entry is easy. In a market with easy entry, statistics showing undue concentration are not reliable evidence of a lessening of competition.

E. Merger Guidelines Although the 1984 Justice Department Merger Guidelines,41 reissued jointly with the FTC in 1992,42 do not bind future administrations, it is worthwhile to consider them, especially since they have not been repudiated by a successor administration. The “real law” that applies to many potential defendants is a combination of the doctrine announced by the courts and the enforcement policies of the Justice Department. The guidelines do not bind private plaintiffs, but a great deal of antitrust litigation has been on the coattails of government action. It follows that the guidelines provide an important source of relevant law. The “numerical guidelines” are the most significant innovation. The Merger Guidelines set up three regimes: low concentration, moderate concentration, and high concentration. Concentration is measured by the Hirschman-Herfindahl index (HHI). Suppose there are N firms in a market and the market share of each firm is si (i = 1, . . . , N). The HHI is HHI = s21 + s22 + s 23 + . . . + s 2N. The low concentration regime is one in which the postmerger HHI is less than one thousand. In this case, challenge is unlikely. The moderate regime is characterized by a postmerger HHI between one thousand and eighteen hundred. Here an increase in the HHI, resulting from the merger, that is less than one hundred points is presumed lawful – challenge will occur only in unusual circumstances. If the postmerger HHI increases by more than one hundred the challenge is likely unless other factors, such as ease of entry, suggest that the merger is unlikely to lessen 41 42

(1984) U.S. Department of Justice 1984 Merger Guidelines, 49 Fed. Reg. 26,823 (1984). (1992) U.S. Department of Justice & F.T.C. 1992 Horizontal Merger Guidelines, 57 Fed. Reg. 26,823 (1992).

II. Horizontal Merger Law

329

competition. The high concentration case is characterized by an HHI in excess of eighteen hundred. Here the merger is presumed lawful if the increase is less than fifty. If the increase is more than fifty, then challenge is likely unless other factors suggest that the harm to competition is not great. If the increase is more than one hundred, then challenge is very likely. There is a fast method of applying the test in the case of a merger between two firms. Let A be the market share of firm A and let B be the market share of firm B. Then the increase in the HHI resulting from the merger is 2AB.43 Apply the test by comparing 2AB with the “increase thresholds” detailed by the guidelines. The other factors that might justify challenge even though the numerical thresholds are not crossed are mentioned in several parts of the guidelines. They are for the most part intuitive. For example, if the firms have previously been involved in a conspiracy (recently), or if entry is unusually difficult, or if one of the firms was a significant competitor. A few less intuitive factors are also correlated with the likelihood of successful collusion: for example, the industry’s product is homogeneous, information on the transaction prices charged by competitors is readily available, orders are frequent and small relative to the size of the market.44 The guidelines also include a number of similar concerns that should be taken into account in determining the significance of market share statistics. The remaining contributions are as follows: (1) The leading firm proviso: the department is likely to challenge the merger of any firm with a market share of at least 1 percent with the leading firm in the market, provided the leading firm has a market share that is at least 35 percent. (2) Ease of entry: challenges are unlikely where entry is easy. (3) Efficiencies: the department will consider efficiencies in deciding whether to challenge a merger. Consider the following example. Suppose there are sixty-three firms in the widget industry: sixty with 1 percent each, two with 5 percent, and one with 30 percent. Suppose the two firms with 5 percent merge. The increase in the HHI is then 2(5)(5) = 50. The premerger HHI is

43

44

Recall that the HHI is the sum of the squares of the market shares. Thus, if A and B merge, the square of the sum of market shares is (A + B)2. The increase in the HHI resulting from the merger is therefore (A + B)2 - A2 - B2 = 2AB. Suppose three firms merge, A, B, and C. Then the increase in the HHI is given by 2(AB + AC + BC). These factors were originally suggested in George Stigler’s study, see Chapter 4.

330

Horizontal Mergers

60 + 5 2 + 5 2 + 30 2 = 1010. The postmerger HHI is therefore 1060. Since this is in the moderate concentration range and the increase is less than one hundred points, the government is unlikely to challenge unless other factors suggest that the merger is likely to lessen competition substantially. Suppose instead, that the 30 percent firm merges with one of the 5 percent firms. Then the increase in the HHI is 2(30)(5) = 300. The postmerger HHI is 1,310. Since the increase is greater than one hundred, challenge is likely unless factors such as ease of entry suggest that there is no threat to competition. To see the extent to which the restraints affect merger decisions, note that in a moderately concentrated market, any merger between two firms with market shares greater than 7 percent will cross the numerical threshold at which enforcement becomes likely.

iii. conclusion This chapter has reviewed the development of horizontal merger doctrine. As is true of Sherman Act Section 1 doctrine, the tension between economic reasonableness and administrative concerns has influenced the development of merger law under the Clayton Act. In view of the difficult proof burdens under an economic reasonableness standard, enforcement agencies have pushed for structurally triggered per-se tests, like that adopted in Philadelphia National Bank. Courts, on the other hand, have tended in spite of the legislative history of the Clayton Act toward the application of an economic reasonableness test to mergers.

appendix Mergers: The Welfare Tradeoffs In this note I examine whether the efficiency gain resulting from a costreducing merger is likely to outweigh the deadweight loss from enhanced monopoly power. Let the inverse demand curve for the industry’s output be given by p = A - bq where p is the product price and q is total output. Suppose that when both firms operated under competition, the total cost of the industry’s output was

Appendix

331

TCcomp = c0 q Under monopoly, the total cost is given by TCcomp = c1q Because the monopoly is most efficient at producing, c1 < c0 . Under these assumptions it is straightforward to show that the monopoly level of output is qmon = ( A - c1 ) 2b and the monopoly price is pmon = ( A + c1 ) 2. The competitive total output and price levels are qcomp = ( A - c0 ) b pcomp = c0 . The increase in deadweight loss is equal to D = ( pmon - pcomp )(qcomp - qmon ) 2. The increase in efficiency is E = qmon (c0 - c1 ). The efficiency gain exceeds the deadweight loss if E > D. Recall that a monopolist always produces where the elasticity of demand, e, exceeds one. This is satisfied because e = ( A + c1 ) ( A - c1 ) > 1. Substituting terms, and after some algebra, one can show that E > D if and only if:

[e - (c0 c1 )(e - 1)] > 2(e - 1)(c0 c1 - 1). 2

Since the ratio c0 /c1 is a measure of the cost reduction, this condition allows us to determine whether the efficiency gain exceeds the deadweight loss for given cost reductions and specific elasticities of demand. Consider e = 3/2, and let z = c0 /c1. In this case E > D if and only if

(3 - z) > 4(z - 1). 2

332

Horizontal Mergers

It happens that the larger root of this quadratic provides a solution that is infeasible in terms of the model because it requires A < c0. Therefore the only solution that should be considered is the smaller root, which is z* = 1.536. It follows that in the case where the demand elasticity is 1.5, if the cost reduction is at least 35 percent, the efficiency gain will exceed the deadweight loss (i.e., E > D). For e = 2, the small root is z* = 1.27, which suggests that the cost reduction must be at least 21 percent. For e = 3, the increase in efficiency outweighs the increase in deadweight costs if the cost reduction is not less than 11 percent. It follows that in order to have single-digit cost reductions that are sufficient for the efficiency gain to exceed the deadweight loss, the demand elasticity must be greater than or equal to 4. Put another way, the merged entity’s Lerner index – the markup over marginal cost – must be less than or equal to 25 percent.45 45

These results are similar to those in DePrano and Nugent, supra note 20.

16 Mergers, Vertical and Conglomerate

This chapter examines the law governing nonhorizontal mergers. In particular, it traces the development and provides a critical overview of vertical and conglomerate merger doctrine. The key economic difference between horizontal and nonhorizontal mergers is that nonhorizontal mergers do not result in a reduction of competition. This suggests that the “welfare tradeoff” – that is, the tradeoff between deadweight costs due to reduced competition and efficiency gains due to cost reductions – will fall in favor of efficiency gains in the nonhorizontal merger context. Consistent with this intuition, the law has tended toward a rule of reason approach, though even in this case we observe a truncated test that limits the availability of some efficiency defenses.

i. vertical mergers The policy arguments about vertical mergers can be grouped into two categories: procompetitive theories and anticompetitive theories. I will review the arguments under these headings, and then trace the development of vertical merger doctrine.

A. Procompetitive Theories The antitrust issues raised by vertical mergers are similar to those in exclusive dealing arrangements. This should not be surprising because a vertical merger is simply a type of exclusive dealing contract. The difference is that in the case of a merger, the contractual relationship is expected to last longer, and the “contract” is not as detailed. Indeed, 333

334

Mergers, Vertical and Conglomerate

in an influential article titled “The Nature of the Firm,”1 Ronald H. Coase argued that a firm “becomes larger as additional transactions (which could be exchange transactions coordinated though the price mechanism) are organized by the entrepreneur,”2 and that “a point must be reached where the costs of organizing an extra transaction within the firm are equal to the costs involved in carrying out the transaction in the open market.”3 Coase’s theory suggests that vertical mergers occur because organizing the supply of services within the firm rather than procuring them from the market can achieve savings in some instances. Since Coase’s early contribution, the literature on vertical integration has developed into two broad strands of procompetitive or efficiency theory: the transaction cost model and the successive monopoly model.4 The transaction cost model sees vertical mergers as a means of controlling opportunism that arises after contracts are formed. I will present one version of this theory in Part I.E of this chapter. The successive monopoly model views vertical mergers as a response to the large welfare losses that can result when two firms with monopoly power are in a vertical relationship. Suppose A, a monopolist in the manufacture of widgets, sells its product to B, a monopolist in the distribution of widgets. First, A sets a monopoly markup on its widget, and then sells it to B, who then applies its own monopoly markup to the (already marked up) widget. The result of this successive monopoly pricing is to shrink markets and raise prices substantially further than would be observed if a single firm owned both the manufacturing and retail parts of the widget business. Recognizing that it is losing a lot of revenue as a result of the distributor’s markup, the widget manufacturer would have an incentive to vertically integrate, by taking over the widget distributor. Under a vertically integrated structure, one monopoly markup would be applied. Although consumers would prefer perfect competition, they are unquestionably better off under a single integrated monopoly than under successive monopoly. Thus, vertical integration serves both the interests of the vertically integrating firm and consumers.5 1 2 3 4

5

4 Economica (n.s.) 386–405 (1937). Id. at 393. Id. at 394. For a survey of efficiency theories of vertical integration, see Andy C. M. Chen and Keith N. Hylton, Procompetitive Theories of Vertical Control, 50 Hastings L. J. 573 (1999). This understanding of vertical integration was first advanced in Joseph J. Spengler, Vertical Integration and Antitrust Policy, 58 J. Pol. Econ. 347 (1950).

I. Vertical Mergers

335

B. Anticompetitive Theories The antitrust concerns generated by vertical mergers are familiar: (1) a vertical merger may put potential competitors at a disadvantage by raising the costs of entry (entry deterrence), (2) a vertical merger may put existing competitors at a disadvantage by raising their costs (raising rivals’ costs). The classic illustration of both concerns is when a manufacturer vertically integrates forward and “locks up” the desirable retailers or retail locations. This was the theory implicit in FTC v. Brown Shoe (discussed in Chapter 13).6 The classic example of anticompetitive foreclosure through backward integration is when a producer locks up the available supply of some production input. Consider, for example, Alcoa’s decision to purchase electricity on the condition that the seller refrain from selling electric power to any other producer of aluminum (see Chapter 10);7 or consider the acquisition of terminal facilities in U.S. v. Terminal Railroad (see Chapter 9).8 C. Development of Doctrine The statutory interpretation issues examined in the discussion of horizontal mergers (Chapter 15) reappear in the material on vertical mergers. The first of such issues to note is that Section 7’s scope of application to vertical mergers was at least as restricted and narrow, before the 1950 amendments, as its scope of application to horizontal mergers. Before the 1950 amendments, Section 7 referred to a lessening of competition between the acquiring and acquired companies.9 That language implied that the acquiring and acquired companies must have been in competition, which implied, in turn, that Section 7 did not apply to vertical mergers. The modern history of vertical merger law begins in 1957, with the Supreme Court’s surprising decision in United States v. E.I. du Pont de 6 7 8 9

384 U.S. 316 (1966). See U.S. v. Aluminum Co. of America (Alcoa), 148 F.2d 416, 422 (2d Cir. 1945). United States v. Terminal R.R. Assn., 224 U.S. 383 (1912). No corporation engaged in commerce shall acquire, directly or indirectly, the whole or any part of the stock or other share capital of another corporation engaged also in commerce, where the effect of such acquisition may be to substantially lessen competition between the corporation whose stock is so acquired and the corporation making the acquisition, or to restrain such commerce in any section or community or tend to create a monopoly of any line of commerce. Clayton Act, ch. 323 § 7, 38 Stat. 730, 731 (1914)(codified at 15 U.S.C. §18 (1988)).

336

Mergers, Vertical and Conglomerate

Nemours & Co. (General Motors)10 that unamended Section 7 did apply to vertical mergers. The Court held that the acquisition of part of GM by du Pont violated Section 7. Du Pont purchased 23 percent of the stock of GM. Du Pont sold finishes and fabrics to GM, which used them to paint the cars and cover the seats. Even though GM’s purchases constituted a very small share of the market for industrial finishes and of the market for industrial fabrics, the Court designated the relevant market as automotive finishes and fabrics. This settled the foreclosure issue, given that GM’s market share in 1957 was roughly 50 percent and GM bought roughly half of its needs from du Pont. Although GM’s share of purchases amounted to only 3.5 percent of industrial finishes, it constituted 24 percent of automotive finishes. The definition of the relevant market obviously played an important role in determining the outcome of this case.11 By approving the choice of automotive finishes, the Court implicitly assumed that any seller of automotive finishes could not divert production to some other industrial use. Did du Pont put its rivals at a competitive disadvantage by locking up the biggest customer in the automotive finishes market? Assume, for the moment, that du Pont’s action did create an entry barrier for potential rivals and raise the costs of existing rivals. How could this have occurred? Two ways come to mind. If economies of scale were important in the automotive finishes market, du Pont’s acquisition of GM could have made it impossible for competing paint makers to reach a competitive scale. Alternatively, if the market for automotive finishes could be divided into high-demand and low-demand customers, with GM alone in the first category, and GM’s rivals in the second, then du Pont’s purchase may have left its rivals fighting in a weak residual market. However, this thought exercise reveals weaknesses in the entry barrier theory. GM was a major car maker. Could one producer of automotive finishes force GM to purchase its paint at monopolistic prices and thereby restrict competition from other paint makers? If du Pont did not sell at monopolistic prices – that is, if du Pont sold to GM at competitive prices – then why should we be concerned with the identity of the firm that sells paint to GM? In addition, if one assumes that the appropriate 10 11

353 U.S. 586 (1957). Although decided in 1957, the relevant facts occurred before 1950. Recall that this was true of the early horizontal merger cases, such as Brown Shoe, Philadelphia National Bank, and Alcoa (Rome Cable) (all discussed in Chapter 15).

I. Vertical Mergers

337

choice for the relevant market is industrial finishes and fabrics, foreclosure of 3 percent could hardly pose a substantial barrier to entry or competitive burden for du Pont’s rivals. Another argument for blocking the merger is that it may have facilitated collusion among makers of automotive paints by eliminating competition for GM’s account.12 There are two versions of this theory. One is that du Pont benefitted at GM’s expense, selling paint at inflated prices to GM and other firms; the other is that GM was the main beneficiary because the collusion by automotive paint makers forced GM’s rivals to pay inflated prices in the automotive paint market. Consider the theory that the transaction hurt GM and benefitted du Pont. One question to consider is whether the du Pont-GM deal really occurred at arm’s length. If the contract provided net gains to both sides, then there probably was no anticompetitive motive because du Pont would have benefitted from GM purchasing finishes and fabrics at the best price/quality combination available. On the other hand, if du Pont exercised its influence as an owner purely in its own interests (i.e., selfdealing), then the merger may have benefitted du Pont at GM’s expense. That is, du Pont may have gained more after the merger by selling paint at inflated prices than it lost by engaging in conduct that reduced the market value of GM. But it is doubtful that du Pont had a sufficiently large ownership share (at 23 percent) to exercise such a degree of control over a company as large as GM. And surely owners of the other 77 percent of GM stock, anticipating the risk of self-dealing by du Pont, would have been on the lookout for this. Given the implausibility of a firm GM’s size becoming the victim of a self-dealing transaction, the most plausible anticompetitive theory is that the merger permitted GM to gain an advantage over its rivals. The transaction may have raised the costs of GM’s rivals by facilitating collusion within the automotive paint market. Under this theory, GM would act as enforcer or “cartel ringmaster”13 in a price-fixing cartel among the 12

13

See, for example, U.S. Department of Justice Merger Guidelines, 49 Fed. Reg. 26,823, 26,836 §4.222 (1984) (If upstream firms view sales to a particular buyer as sufficiently important, they may deviate from the terms of a collusive agreement in an effort to secure that business, thereby disrupting the operation of the agreement. The merger of such a buyer with an upstream firm may eliminate that rivalry, making it easier for the upstream firms to collude effectively). Krattenmaker and Salop have described this general approach as the “cartel ringmaster” model of indirect foreclosure. See Thomas Krattenmaker and Steven Salop, Anticompetitive Exclusion: Raising Rivals’ Costs to Achieve Power over Price, 96 Yale L. J. 209, 240 (1986).

338

Mergers, Vertical and Conglomerate

paint makers. However, if the relevant product market extended beyond automotive finishes, this theory would be hard to accept. A collusive price structure within the automotive finishes market would have encouraged industrial paint makers to enter the automotive paint market. Perhaps the best method of determining the relevant market in General Motors is to follow by the 1984 Justice Department Merger Guidelines.14 In defining the market, the guidelines approach would ask whether a hypothetical monopolist in the market for automotive finishes could keep its price 5 percent above the competitive level within the market for automotive finishes for a substantial period without causing other makers of industrial paint to enter that market, and buyers to switch to the new entrants. If such a strategy would be successful, then the market for automotive finishes should be deemed a separate market. This exercise should take into account supply- and demand-side substitution issues discussed in Chapter 11.15 While General Motors dealt with unamended Section 7, the Supreme Court first dealt with amended Section 7 in Brown Shoe Co. v. United States.16 The Court said that in order to show a violation of Section 7 in a vertical merger case, the plaintiff must show that the merger lessens competition substantially in an area of effective competition. This requires an examination of the merger’s effects in each economically significant product market in order to determine whether there is a substantial lessening of competition. What happens if there are four significant lines of commerce, and the foreclosure level is high in only one of them? Is this evidence of a substantial lessening of competition? Much of the Court’s opinion on the vertical aspects of the merger in Brown Shoe addresses this question. Let’s start with the easy extremes. Suppose there are four lines of commerce, and the share of the market foreclosed in each is near 100 percent. Then the merger violates the Sherman Act, and a fortiori violates the Clayton Act. On the other hand, if the foreclosure in each of the four lines is minimal then there is no violation, because competition has not lessened substantially as a result of the merger.

14 15

16

See Chapter 11 for discussion of market definition under the guidelines. In theory, the definition of the relevant market should not matter. If it is defined too narrowly, supply- and demand-side substitution effects will imply that market power is weak. See Chapter 11. 370 U.S. 294 (1962).

I. Vertical Mergers

339

The Court said that the intermediate case requires an examination of the degree of foreclosure and “economic and historical” factors that shed light on the “nature and purpose” of the arrangement.17 On the nature and purpose, the Court noted that vertical mergers are analogous to exclusive dealing arrangements, and therefore plaintiffs must demonstrate a significantly greater level of foreclosure to prove a violation than a tying case would require. The Court suggested that the efficiency justifications for exclusive dealing recognized in the Standard Stations opinion (see Chapter 14) could be used to defend vertical mergers. The most important efficiency justification cited by the Court is the claim that the merger’s purpose was to insure the supply of a vital input or the market for a seller’s output. On the other hand, the Court said that to the extent firms use vertical integration as a method of tying, the likelihood of a Clayton Act violation increases. Applying these arguments to the facts in Brown Shoe, the Court held that the evidence suggested that the Brown Shoe Company had intended to increase its sales of shoes from Kinney retail outlets, which the Court viewed as equivalent to a tying arrangement. This finding raised the likelihood of a Clayton Act violation. On economic and historical factors, the Court listed the following: (1) whether the acquired company is failing, (2) whether two small companies are merging in order to compete against a larger company, and (3) evidence of a trend toward vertical integration. Brown and Kinney could not rely on the first two factors. The Court found a violation of Section 7 largely on the basis of its conclusion that the merger seemed analogous to a tying arrangement and evidence suggesting a trend toward greater vertical integration in the shoe industry. We have discussed the incipiency doctrine, which holds roughly that because of evidence of a trend toward concentration, a court may hold that a merger violates Section 7 even though it does not result in significant concentration levels. In the case of a vertical merger, the incipiency doctrine would require the Court to examine evidence suggesting a trend toward vertical integration. Brown Shoe implicitly adopts the incipiency test in the area of vertical mergers. The foreclosure level in Brown Shoe was minimal, Kinney sold 1.6 percent of all shoes. If this were the result of an exclusive dealing contract, the Court would have 17

In such cases, it becomes necessary to undertake an examination of various economic and historical factors, such as the nature and purpose of the arrangement, to determine whether it is of the type Congress sought to proscribe. Brown Shoe, 377 U.S. at 328.

340

Mergers, Vertical and Conglomerate

upheld it under Tampa Electic (Chapter 14), because of the insignificant foreclosure level. What is the legal test in the area of vertical mergers? Formally, the test announced in Brown Shoe is a form of rule of reason analysis. A detailed description of the test would recognize that it distinguishes three types of merger. The first is where the level of foreclosure is high, approaching monopolization. We know from Alcoa that 90 percent foreclosure is sufficient to fall in this category. In this case, Brown Shoe suggests that a per-se illegality rule applies. The second type of merger is where foreclosure is trivial, and there is no violation in this case because the merger has no economically significant effect. Third is the intermediate foreclosure case, and it is here that a truncated rule of reason test applies. I refer to the rule-of-reason test as truncated because Brown Shoe suggests that only a limited set of efficiency defenses are available, specifically those suggested by the Court’s comparison of vertical mergers and exclusive dealing contracts (the insurance argument) and the one suggested by the second mitigating-factor defense (that the merger enabled the firms to compete against a larger firm). To sum up, the Court in Brown Shoe set out a complicated rule of reason inquiry for vertical mergers, and then applied the much simpler incipiency test. Does the doctrine as the courts apply it today differ substantially from its formal statement? I doubt that there is a significant difference between the doctrine-in-application and its formal statement today. General Dynamics (Chapter 15) pretty much repudiates the incipiency test.18 After rejecting incipiency as the predominant concern in the area of horizontal mergers, the Court surely would not continue to let it dominate in the analysis of vertical mergers.

D. Some Rule of Reason Considerations Rule of reason analysis in the area of vertical mergers requires consideration of a number of questions. Perhaps the best guide to the relevant questions and presumptions is the Vertical Merger Guidelines issued by the Justice Department. The Department’s guidelines note three antitrust concerns implicated by a vertical merger: (1) the potential for increased entry barriers, (2) the potential for facilitating collusion, and (3) the potential for evasion of rate regulation. 18

The Court in General Dynamics distinguished Von’s on the ground that market share statistics were less useful in the coal supply market, see Chapter 15. But the holding of General Dynamics makes clear that market share statistics are insufficient to prove a violation of Section 7.

I. Vertical Mergers

341

According to the Guidelines, three necessary conditions must exist before the potential harm resulting from entry barriers justifies enforcement. First, the degree of vertical integration must be so extensive (across the market) that the potential entrant must enter at two levels. What does this mean? Consider the acquisition by Brown of Kinney’s retail outlets. If all or almost all of the shoe producers have acquired retail shoe outlets, so that very few or no retail outlets remain, then unless new independent retailers enter the market, a potential entrant in the shoe manufacturing market will have to acquire a retail network in order to sell the shoes. Second, the requirement of two level entry must make entry significantly more difficult than at the primary level alone. Again, consider this in light of the Brown Shoe facts. This means that if a potential entrant in the shoe manufacturing industry must also enter into the shoe retail market, it will find it very costly to acquire a shoe retail network. Third, the structure of the primary market must be so conducive to noncompetitive performance that the increased difficulty of entry is likely to affect its performance. In terms of the shoe market example discussed so far, this would require a high level of concentration in the shoe manufacturing industry. In general, the guidelines require an HHI of at least eighteen hundred. The Justice Department considers a market with an HHI above eighteen hundred to be highly concentrated. The guidelines suggest that enforcement may be justified even if the HHI is below eighteen hundred if certain other factors are present: if firms in the market have previously been found to have engaged in collusion, if the product is homogeneous (suggesting collusion is easier), if detailed information on transaction prices and quantities is available to competing firms, if firms require customers to accept “delivered prices,” if the acquired firm was an aggressive competitor, and other considerations. The guidelines also make clear that the courts should take efficiencies into account in determining whether a vertical merger violates the Clayton Act. The guidelines note that an efficiencies defense for a vertical merger should receive more weight than in the area of horizontal mergers.

E. Vertical Mergers and Transaction Costs I have tried along the way to present the teachings of transaction cost economics (see, for example, Chapters 1, 6, 10, and 14). Here I will note its implications for vertical mergers. Recall that Ronald Coase

342

Mergers, Vertical and Conglomerate

hypothesized that firms increase in size when the cost of organizing services within the firm is less than that of procuring them from the market. The costs of procuring things from the market – finding suppliers, determining prices, writing contracts – are typically described in the economics literature as transaction costs. There are several types of transaction cost observed in market transactions. One is the simple cost of contract writing; the time and effort required to draft agreements, or the cost of hiring lawyers to write contracts. Another set of costs are generated by concerns about opportunism. Contracting parties incur costs in devising ways to minimize the risk that one of the parties will take advantage of the other in a period of vulnerability. The classic case is where the seller simply refuses to perform after the buyer has paid the full price. In this section, I will focus on potential opportunism as a source of transaction costs. One theory of vertical integration is that it appears in relationships that require contracting over specialized or transaction-specific assets. In these relationships, a bilateral monopoly arises after the parties enter into a contractual relationship. The specialized assets generate a stream of rents that the two parties to the contract must allocate between themselves. The allocation of bargaining power between the parties will determine the final allocation of rents. An efficient allocation of rents, however, is one that would fully compensate each party for its share of investment in transaction-specific assets. When contracts have to be renewed or renegotiated, the bargaining power of each party is determined by that party’s outside opportunities. There is no reason to assume that each party will use its bargaining power in a manner that leads to an efficient allocation of transaction-specific rents. One party may try to take advantage of a change in conditions in order to grab a larger share of the rent. Vertical integration minimizes the incentive of one party to expropriate the other’s share of rent by (1) dampening the influence of changes in outside opportunities on bargaining positions and (2) bringing about a convergence in the interests of the parties. Consider an example. Suppose newspaper publisher P leases a printing press from owner O.19 If after entering into the relationship, the publisher cannot easily turn to another printing press owner to lease printing 19

This example is taken from Benjamin Klein, Robert G. Crawford, and Armen Alchian, Vertical Integration, Appropriable Rents, and the Competitive Contracting Process, 21 J. L. & Econ. 297, 298–302 (1978).

I. Vertical Mergers

343

services, the owner of the press may be in a good position to demand a higher lease payment. Conversely, if the owner of the press cannot easily turn around and lease the press to someone else, P may be in a good position to bargain down the lease charge. The latter scenario is likely if the printing press is designed to meet the specifications of the publisher, and is unlikely to meet the specifications of any other publisher in the area. Suppose the press owner, O, observes an increase in the demand for newspapers, and then demands an increase in the lease payment from the publisher, P. This may sound unfair, but it is not. The opportunity cost of leasing a machine to P has increased, and O is justifiably seeking compensation for the higher opportunity cost that he incurs. Asking for a larger lease payment is unfair only if O seeks to gain more than compensation for the increase in his opportunity cost. Suppose, however, that the demand only for P’s newspaper increases, because P has invested in the quality of its workforce. In this case the opportunity cost to O has not changed. Yet O may have an incentive to press for an increase in the lease payment. In the latter scenario, P will realize what is happening and reject the proposed increase in the lease charge. The haggling that follows can be costly; not unlike union-sector bargaining, where strikes can put both parties in a position far worse than if they had agreed to the first contract proposal. Long-term contracts and vertical integration can do nothing to minimize or alter the changes in outside opportunities that generate opportunistic bargaining. However, by sewing up the terms of the relationship over a period, these methods of contracting dampen the influence of momentary changes in the parties’ bargaining positions. Vertical integration, in particular, also reduces opportunistic bargaining by bringing about a convergence of interests. If the printing press owner is able to hold up the publisher for the entire value of the publisher’s quality investments, then the publisher will have little incentive to make future quality investments. The printing press owner may not recognize the harm this does to its long-term interests. In the vertically integrated unit, this incentive divergence is less likely.20 Within the vertically integrated

20

This is analogous to the argument that merging upstream and downstream producers should minimize the likelihood that the upstream producer will pollute the water, thereby imposing costs on the downstream firm. If the costs are all borne within the unit, the “unitized” firm will minimize the total costs associated with pollution.

344

Mergers, Vertical and Conglomerate

printing-press/publisher unit, the press owner would be deterred from taking any action that reduced the market value of the integrated unit. The transaction cost theory has implications for other areas of antitrust. Recall that in Chapter 6, the theory provided an explanation for some price-fixing agreements. Chapter 10 suggested that the theory could explain certain “lock-in” contract provisions, such as the leasing contract in United Shoe.

F. Summing Up Like horizontal merger doctrine, vertical merger doctrine has moved from a focus on market share statistics to a detailed examination of economic reasonableness issues. However, the rule of reason test that applies to vertical mergers is still truncated in the sense that it permits only a limited set of efficiency arguments as defenses. In view of the welfare tradeoff analysis (Chapter 15), this is overly conservative. Since the anticompetitive potential is smaller in the context of vertical mergers than in that of horizontal mergers, the social cost of taking efficiency arguments into account is also smaller.

ii. conglomerate mergers Conglomerate mergers are neither horizontal nor vertical mergers. Consider, for example, the merger in the Clorox opinion. Procter & Gamble Corporation sought to extend its line of products by adding Clorox bleach to the line. Clorox is an example of one general class of conglomerate merger: where one firm uses an acquisition in order to extend its range of products. Another example is the case of two firms that sell the same product in entirely different geographic markets. If the product has a short shelf life and the geographic markets are sufficiently far apart, the two firms will not be in competition. A merger between two such firms would be a conglomerate merger.

A. Potential Competition Theory The product extension merger, as in Clorox, does not lessen competition. Procter & Gamble did not compete against Clorox before the two firms merged, and they obviously did not compete after the merger. In light of this, what is the problem with conglomerate mergers? The theory at the foundation of the conglomerate merger cases is the potential

II. Conglomerate Mergers

345

competition doctrine. In the Clorox case, by merging with Clorox, Procter removes itself as a potential independent entrant. There are two potential harms to consumer welfare: (1) as a potential entrant, Procter put competitive pressure on Clorox, and the consumer benefits from this are eliminated by the merger; (2) as a later independent entrant, Procter would actually compete against Clorox, and the consumer benefits from this are eliminated by the merger. Both harms reflect the loss of distinct benefits that consumers receive. One is the benefit from “actual potential competition,” which is the benefit consumers receive today from Procter’s existence on the sidelines, waiting in the wings. The second benefit is one realized in the future, though its value may be measured today in present value terms. It is the benefit to consumers from competition after Procter enters independently and competes against Clorox. We may label the present value of this benefit the “potential competition benefit.”The conglomerate merger opinions usually offer some combination of potential and actual potential competition arguments to support the decision. It may be helpful in reading the opinions to determine whether a given argument is a potential or potential actual competition argument. There is a problem with the potential competition doctrine. It is not unreasonable to believe that at least some of the mergers designed to enter new markets are made with an intent to provide fiercer competition. Many of these mergers, if allowed, would benefit consumers by leading to lower prices and higher quality. For example, in the Clorox case, Procter & Gamble sought to combine its strength in the advertising and marketing areas with Clorox’s expertise in production and distribution. If the combined strengths produced a stronger firm, the merger could have enhanced competition in the liquid bleach market, leading to lower prices and higher quantities of liquid bleach consumed. In light of this, the complaint from the potential competition corner is that the chosen path may not provide the greatest benefit to consumers. An alternative type of entry – independent or “de novo” entry – could provide greater benefits to consumers. Once we see that this is the nature of the complaint, it becomes clear that it is speculative. In particular, it raises the following problems: 1. The potential competition doctrine proves too much by suggesting that no merger should be permitted. If the Clorox merger is not good enough, then why should we approve any particular merger when it is possible to find an alternative method of entry that could provide greater benefits to consumers?

346

Mergers, Vertical and Conglomerate

2. The potential competition doctrine may be wrong as an empirical proposition. We have no evidence that independent entry provides a greater benefit to consumers than that provided by a successful merger. 3. By implicitly imposing a duty to maximize consumer benefits, the potential competition doctrine breaks away from most of antitrust doctrine by imposing positive rather than negative injunctions – that is, commands to “do good” rather than to “avoid doing harm.” In this sense, the potential competition doctrine is analogous to a duty-to-rescue rule. And like a rescue requirement, it runs the risk of keeping people off the beaches altogether. Firms often use the acquisition of an established firm as a means of entry because it is cheaper than entering a market through internal expansion. Given this, making firms expand internally or acquire smaller firms (how much smaller?) will raise the cost of entry and thereby deter some acquisitions that would have benefitted consumers. The paradox of potential competition doctrine is that enforcement may lead to a reduction in potential competition. How? By eliminating a cheap method of entry, the doctrine raises the cost of entry. This in turn reduces the probability of entry and the benefits provided by potential competition. Given these uncertainties, it seems clear that a potential competition challenge should not be permitted unless: (1) the market is concentrated and (2) the number of potential entrants is small. If the market in which entry takes place is concentrated, then at least one can argue plausibly that the benefits from independent entry would have been significant. If the number of potential entrants is small, then, again, one can argue plausibly that the loss of one potential entrant is important. If either of these conditions is not satisfied (e.g., there are many potential entrants) then the gain from forcing a merger to proceed through internal expansion rather than acquisition is likely to be small. The Supreme Court’s views on conglomerate mergers and Section 7 were expressed in Federal Trade Commission v. Procter & Gamble Co. (Clorox).21 Procter was a diversified manufacturer of detergents and other high turnover household products. At the time of this suit, Procter did not make bleach. Clorox made liquid bleach, and was the largest 21

386 U.S. 568 (1967).

II. Conglomerate Mergers

347

producer with almost 49 percent of the market. Procter acquired the assets of Clorox. The Court upheld the Federal Trade Commission’s finding that the acquisition violated Section 7 of the Clayton Act. The Court held that competition in the liquid bleach market was lessened as a result of the acquisition because it: (1) raised entry barriers and discouraged rivals from competing aggressively, and (2) eliminated the potential competition of Procter. The FTC advanced these theories, and the Court’s opinion largely restates them. Entry barriers would rise because Procter would use its large advertising budget to promote Clorox aggressively. The theory the Court accepted is that advertising by an incumbent creates an entry barrier, because it forces rivals to match the incumbent’s advertising and promotion expenditures. Further, advertising creates brand names that consumers are reluctant to substitute away from. It would be impossible for an entrant to match Procter’s advertising expenditures because Procter’s size (as then the country’s largest advertiser) allowed it to obtain substantial volume discounts. The ability to take advantage of volume discounts might be an efficiency argument in favor of the merger; however, the Court cited Brown Shoe for the proposition that an efficiency defense is impermissible. An alternative entry barrier theory mentioned by the Court is that Procter could use its leverage as a large producer of consumer products to persuade retailers to give Clorox preferred shelf space. This discourages entry because a new rival would be unable to compete with Clorox in this respect. Competition would be discouraged because potential entrants would fear retaliation by Procter. Procter, the FTC had argued, would use its resources to subsidize a predatory pricing campaign by Clorox. Similarly, Procter would use its resources to discipline existing competitors. In addition, Procter’s ability to obtain preferred shelf space for Clorox would further discourage rivals from competing aggressively. The removal of Procter as a potential entrant substantially lessened competition because of the concentration in the market for liquid bleach. The top six companies accounted for 80 percent of national sales in 1957. The remaining 20 percent was divided among two hundred liquid bleach producers.22 The lack of evidence that Procter had any plans to enter 22

And there were some areas of the county in which concentration levels were higher. “Because of high shipping costs and low sales price, it was not feasible to ship the product

348

Mergers, Vertical and Conglomerate

independently was not important, because Procter was the most likely entrant. Procter sold complementary products and was in the process of diversifying. Moreover, entry was relatively easy in the liquid bleach market. The Sixth Circuit had heard all of these arguments and rejected them on the ground that they were speculative.23 The appeals court had also relied on postmerger evidence indicating that other producers in the industry were doing quite well after the merger and that Clorox’s market share did not significantly change in the four years following the merger.24 The Supreme Court took a different view on the speculativeness issue, arguing that it was not a problem in this case in light of the Clayton Act’s goal of arresting monopolization in its incipiency. Speculativeness aside, there is a significant weakness in the Court’s opinion. If it was easy to enter the bleach-making business, as the Court noted, how could entry barriers have been a serious problem? Furthermore, if entry was easy, why should the elimination of Proctor as a potential entrant matter? The Court never addressed these questions directly, but an answer to the last is implicit in the Court’s opinion: Procter’s elimination mattered because Procter was the most important potential entrant. But this raises another question. Does the Clayton Act require the most powerful entrant to either wait on the sidelines or to enter independently? The Court’s concession that entry was easy makes it harder to accept the FTC’s entry barrier theories. The Commission’s predation theory is also questionable. Is predation rational in a market with easy entry? Under the objective reasonableness standard of Matsushita (Chapter 10), the courts would have to regard the predatory pricing arguments of Clorox as speculative, and therefore insufficient to survive a motion for summary judgment.

B. Subsequent Potential Competition Decisions In Bendix Corp.25 the FTC sought to expand the potential competition doctrine by applying it to a case in which the acquiring firm was not likely

23 24 25

more than 300 miles from its point of manufacture.” Id. at 570. Most manufacturers were limited to competition within a single region because they had but one plant. Thus, Clorox’s seven principal competitors did no business in New England, the mid-Atlantic states, or metropolitan New York, where Clorox’s share was 56 percent, 72 percent, and 64 percent respectively. Id. 383 U.S. at 576. Id. 77 F.T.C. 731 (1970).

II. Conglomerate Mergers

349

to enter the acquired firm’s market through internal expansion. The FTC ruled that the merger violated Section 7, even though the acquiring firm was not likely to enter through internal expansion, because the acquiring firm would have been a likely entrant through acquisition of a smaller “toehold” firm. Other factors that seemed important were that Bendix had acquired the third largest competitor after considering the acquisition of several smaller competitors, Bendix would have acquired one of the smaller firms if the option of acquiring the third largest firm had been foreclosed, and the market in which Bendix entered had a four firm concentration ratio of 80 percent. The Supreme Court has not addressed the FTC’s toehold doctrine. In United States v. Falstaff Brewing Corp.,26 the Court expanded the meaning of potential competitor. The district court dismissed the suit because it found that the acquiring firm had decided not to enter the market except by acquisition of a significant competitor. The Supreme Court reversed and said that in spite of the acquiring firm’s intent, the other firms may have perceived the acquiring firm as a potential independent entrant, and its elimination therefore may have lessened competition.

C. Summary of Critique I have mentioned three criticisms of the potential competition doctrine. One is that the doctrine is speculative. As an empirical proposition, we cannot be sure that independent entry enhances competition more effectively than entry through the acquisition of an existing competitor. In the same vein, the potential competition doctrine creates a difficult linedrawing problem for courts. It is settled that Clayton Act Section 7 does not establish a per-se prohibition of conglomerate mergers. But a rigorous application of the potential competition theory would prohibit conglomerate mergers, because the theory holds that independent or “toehold” entry is always superior. If the courts do not interpret the statute as banning conglomerate mergers altogether, then precisely where should the line be drawn distinguishing a merger that lessens competition from one that does not? A second criticism of the potential competition doctrine is that it does not have a firm basis in the statute. There are two ways to understand the doctrine. One is as a positive injunction; that the Clayton Act requires parties to merge in a manner that maximizes benefits to 26

410 U.S. 526 (1973).

350

Mergers, Vertical and Conglomerate

consumers. The alternative requires the assumption that any conglomerate merger involving the acquisition of an existing significant competitor necessarily reduces competition. If this assumption is accepted, it follows that the Clayton Act requires the finding of a violation whenever a large firm acquires an existing significant competitor. The first justification – the positive injunction argument – has no basis in the purpose or end of the statute. There is nothing in the legislative history or the text of the Clayton Act that suggests the statute prohibits all mergers that fail to maximize, in a global sense, the welfare of consumers. The statute refers to a “lessening of competition,” not to a failure to maximize competition. The second justification is based on an unproven and probably false assumption. There is no evidence that conglomerate mergers involving the acquisition of a significant competitor harm competition. The third criticism of the potential competition doctrine is that it may be counterproductive in the long run. Firms choose to enter a market in the most effective way, and at the least cost given a certain level of effectiveness. In some cases the least costly method of entry is the acquisition of an existing competitor. However, if this route of entry is sealed off, then entry is necessarily more expensive. Entry that is more expensive is less likely to occur. Firms operating in an industry in which the probability of entry has been reduced will have less incentive to compete.

D. Enforcement Guidelines The speculativeness problem in potential competition doctrine has been lessened to some extent by court decisions, such as Bendix, that seem to require some finding of the feasibility and likelihood under the circumstances of independent or toehold entry, and that independent or toehold entry actually would enhance competition. Still, there is an uncomfortable level of uncertainty. The 1984 Justice Department Guidelines attempted to respond to the notice problems and speculative nature of the potential competition doctrine. Under the 1984 Conglomerate Merger guidelines, the Department is unlikely to find enforcement justifiable unless the acquired firm’s market is above eighteen hundred HHI.27 Challenge is not considered justifiable if entry is easy in the acquired firm’s market, or if the acquiring firm is 27

U.S. Department of Justice Merger Guidelines, 49 Fed. Reg. 26,823, 26,834 at §4.131 (1984).

III. Concluding Remarks

351

one of many firms of comparable size that could enter into the acquired firm’s market. In addition, the government will take efficiencies into account. Would the government have brought Clorox today? The HHI in Clorox is well above the Guidelines’ threshold. However, entry and efficiency considerations might lead to a decision not to sue.

iii. concluding remarks This chapter has traced the development and provided a critique of vertical and conglomerate merger doctrine. As is true of horizontal merger doctrine, the Court has developed a rule of reason test for nonhorizontal merger cases that limits the range of efficiency defenses. I have suggested in previous chapters that it is helpful to think of antitrust legal standards in terms of the potential for error and the costs associated with error. Compared to horizontal mergers, the cost of a false aquittal is relatively small in the case of nonhorizontal mergers. In addition, given Coase’s insight on the reasons for nonhorizontal mergers – to take within the firm transactions that would otherwise be arranged through the market – it would seem that the cost of a false conviction is relatively large. These considerations weigh in favor of giving greater consideration to efficiency defenses in nonhorizontal merger cases.

17 Antitrust and the State

I follow the standard practice by taking up the current topic, antitrust and the state, after presenting a substantial part of Section 1 and Section 2 doctrines. There is no good defense for this other than convenience. The extent to which the antitrust laws apply to private actors who seek the state’s aid in erecting competitive barriers is a subject of the highest priority. It is worthy of a course on its own. This is an important topic for a simple reason: the market takes care of itself in the long run. In the absence of entry barriers, successful pricefixing conspiracies attract entrants. Competition from entrants drives economic profits to zero in the long run. Similarly, the operations of the market tend to foil efforts to corner the supply of a key production input. Attempts to corner an input lead to increases in its price. As the input becomes more scarce, competing bidders will offer higher prices to get access to it. A would-be monopolist who tries to achieve dominance through this route would have to beat the bids of competitors, and the bid that has the highest likelihood of winning equals the expected profits of the monopoly. But it would be absurd to call it success when a monopoly is achieved in this fashion. The state is potentially the best friend of the would-be monopolist. The state can erect and enforce entry barriers. The state can enact legislation that hampers the ability of competitors to vie for crucial inputs or the business of big customers. The antitrust cases have many examples of private parties who enlisted or attempted to enlist the aid of the state in supporting an anticompetitive scheme. In other examples, 352

Antitrust and the State

353

antitrust defendants have attempted to justify their conduct by claiming that a state or the federal government authorized it.1 The first Supreme Court decision on Section 1, Trans-Missouri,2 involved a state action claim. The railroad cartel argued that the Interstate Commerce Commission regulated them, that the Commission had implicitly approved their rate-making bureau, and that it followed that they had an exemption from the Sherman Act. The Supreme Court rejected this argument, noting that the Interstate Commerce Act did not explicitly authorize the formation of rate-making bureaus. The Court heard the argument again in Socony, where the defendants argued that they carried out their practices at the behest of officials of the National Industrial Recovery Administration.3 The Court responded that acquiescence by federal officials does not guarantee antitrust immunity.4 In general, the Court has required explicit authorization when a party claims that his conduct is exempt from the Sherman Act because it is regulated by the federal government. If the courts require explicit authorization when a party claims that the Sherman Act conflicts with the regulatory framework of some other federal statute, it would seem that repeals by implication should be even more disfavored at the state level. However, there is a conflicting interest: federalism. If the commerce is purely intrastate, the Sherman Act, read literally, does not apply.5 And our government has long had a policy that some areas exist where states remain free to adopt different methods of governing. In this chapter, I consider defendants who claim immunity from the Sherman Act because the state explicitly authorizes their actions or because they receive some tacit or implicit authorization because of the nature of their acts. The former category is governed by the Parker v. Brown doctrine,6 and the latter by Noerr-Pennington doctrine.7 I will start with Noerr-Pennington doctrine. 1 2 3 4 5

6 7

Robert Bork, The Antitrust Paradox: A Policy at War with Itself 347–64 (1978). United States v. Trans-Missouri Freight Ass’n, 166 U.S. 290 (1897). United States v. Socony Vacuum Oil Co., 310 U.S. 150, 172 (1940). Id. at 226–7. See, for example, Summit Health, L.T.D. v. Pinhas, 111 S.Ct 1842, 1849 (1991) (“[The Sherman Act] does not prohibit all conspiracies using instrumentalities of commerce that Congress could regulate. Nor does it prohibit all conspiracies that have sufficient constitutional ‘nexus’ to interstate commerce to be regulated. It prohibits only those conspiracies that are ‘in restraint of trade or commerce among the several states’.”) (Scalia,J.,dissenting.) Parker v. Brown, 317 U.S. 341 (1943). Eastern Railroad Presidents Conference v. Noerr Motor Freight, Inc., 365 U.S. 127 (1961); United Mine Workers of America v. Pennington, 381 U.S. 657 (1965).

354

Antitrust and the State

Because the state-action antitrust immunity doctrines are so complicated, the bulk of this chapter is devoted to making sense of the case law. The concluding section presents a normative view of the Parker and Noerr immunity doctrines.

i.

NOERR - PENNINGTON

doctrine

Eastern Railroad Presidents Conference v. Noerr Motor Freight8 dealt with a claim that the defendants were immune because of the nature of their acts. Twenty-four Eastern railroads and an association of railroad presidents conducted a publicity campaign against truckers designed to foster the adoption or retention of laws that would hurt the trucking business. The railroads and truckers competed in the long-distance freight business. The truckers filed a complaint charging the defendants with conspiring to restrain trade and to monopolize the long distance freight business in violation of Sections 1 and 2 of the Sherman Act. Reversing both the trial court and the Third Circuit, the Court held that the railroads’ activity did not violate the Sherman Act. The Court’s reasoning began with the basic principle that mere attempts to influence the passage or enforcement of laws do not violate the Sherman Act. That principle, in turn, draws on several other propositions. The first is the Parker doctrine, which holds that a restraint of trade that results from valid governmental action is immune from the Sherman Act.9 Thus, if a party attempts to influence government to take an action that restrains trade, and the government, following accepted constitutionally valid procedures takes that action, the end result does not violate the Sherman Act, however harmful the effect on competition. It follows that the attempt to influence government does not itself constitute a violation, since the end result of those efforts is not a violation. Immunity in this case derives from the more basic Parker immunity, and in this sense can appropriately be described as derivative immunity. If two or more persons combine to influence government, the result is the same as long as they follow procedural norms. The reason is that although the combination satisfies the basic components of conspiracy doctrine – agreement and intentionality, duality, mere attempt whether

8 9

365 U.S. 127 (1961). Parker v. Brown, 317 U.S. 341, 352 (Here the state command . . . is not rendered unlawful by the Sherman Act since, in view of the latter’s words and history, it must be taken to be a prohibition of individual not state action).

I. Noerr-Pennington Doctrine

355

successful or not – the end result of the conspiracy is not a violation of the law. The Court noted that a holding that efforts to influence valid governmental action violate the Sherman Act would impair the power of government to take actions that have the effect of restraining trade. Government needs to receive information from people in order to function in a representative democracy. The First Amendment protects the right of individuals to petition government. The lower courts in the Noerr litigation had acknowledged these points. The novel question was whether other factors present in the Noerr case took the defendants’ actions out of the derivative immunity zone. The first factor was an intent, which the district court found, to harm competition; specifically to destroy the truckers as competitors. The Court held that this factor alone does not weaken the derivative immunity rule. It is all too common that individuals attempt to influence government in order to give themselves an advantage over competitors. There would be little left of the derivative immunity rule if it depended on the intent of the parties attempting to influence government. The second factor was the “third-party” publicity technique, whereby the public relations experts hired by the railroads used public opinion writers, reporters, and other apparently disinterested individuals to make their case to the public and to the legislature. The Court held that although the practice may be unethical, a finding that such techniques were used does not weaken or create an exception to the derivative immunity rule. Does it matter whether the defendants aimed the publicity campaign toward the public or toward legislators? The Court implicitly held that it does matter. Derivative immunity applies to efforts to influence the legislative process, not to efforts to influence consumers at large. Although the truckers argued that the campaign in part tried to influence their customers, the Court rejected this argument because the evidence suggested that the campaign targeted the government. The Court announced one exception to the derivative immunity rule, the sham exception. If the publicity campaign were ostensibly directed toward government but in reality was nothing more than an effort to interfere directly with the business relationships of a competitor, then derivative immunity does not apply. If one constructed a model of the derivative immunity umbrella described in Noerr, it would have the following shape. First, to receive protection from the immunity rule, one must attempt to influence valid

356

Antitrust and the State

governmental action. Although the opinion is not clear on this question, it does not suggest that efforts to influence government to act in a procedurally invalid manner are immune. Parker immunity applies to valid governmental action. Noerr immunity derives from this source alone. Invalid governmental action should, it seems, be impermissible as a source of derivative immunity. Second, the manner of influencing government is irrelevant as long as the effort is to influence procedurally valid governmental action. The Court does not discuss in Noerr whether the substantive validity of the action matters. One would think there is little scope for substantive validity to assert itself in this area. Substantive constitutionality is a doctrine that the Supreme Court abandoned at the dawn of the New Deal, and remains in only small areas such as reproductive privacy matters. Thus, if I pay a minister of a church to approach a legislator to influence his or her position on some legislative matter, the unethical features of my contract with the minister are irrelevant insofar as Noerr immunity is concerned. Unless I intend the minister to influence the legislator to act in a procedurally invalid manner, derivative immunity applies. On the other hand, if I pay the minister to give a sermon to his congregation warning of the dangers of damnation for consuming goods produced by my rival, then I have influenced business relationships directly, and fallen outside of the Noerr immunity zone. Whether there is a Sherman Act violation is an altogether different question, but Noerr immunity should not be available as a shield. Noerr suggests, somewhat surprisingly, that the conspiratorial nature of my efforts are irrelevant so long as I aim to influence valid governmental action. If I combine with two of my competitors and enter into an explicit contract binding the three of us to direct our efforts toward destroying another competitor through influencing government, derivative immunity still applies. As one would anticipate, the important question remaining after Noerr is precisely what the sham exception means. The Court has not answered this question completely, but provided a big part of the answer in California Motor Transport Co. v. Trucking Unlimited.10 A combination of trucking firms in California opposed all applicants for operating rights in the state. The Court held that the activity fell within the sham exception of Noerr and therefore was not protected. 10

404 U.S. 508 (1972).

I. Noerr-Pennington Doctrine

357

In justifying its decision, the Court began by noting the similarity between this case and Noerr. The principle recognized in Noerr applies to efforts to use established channels in government agencies and in the courts to influence government to take a particular action. The mere fact that these parties had tried to use government agencies and courts to bring about an anticompetitive result was not enough to move them out of the zone of derivative immunity. What mattered here to the Court was that the defendants used the resources of the trucking firms to “harass and deter respondents in their use of administrative and judicial proceedings so as to deny them free and unlimited access to those tribunals.”11 In another passage, the Court noted that the conspirators were not trying to use the courts and agencies to influence public officials to take a particular action, but were charged with attempting to “bar their competitors from meaningful access to adjudicatory tribunals and so to usurp that decision-making process.”12 The Court then distinguished the use of unethical practices in the political process from their use in judicial proceedings. Although unethical practices typically do not result in sanctions when employed in efforts to gain influence in the legislative process, the courts will not permit them in the judicial process. The Court mentioned several examples of unethical practices that may violate the antitrust laws: (1) use of a fraudulently obtained patent to exclude a competitor, (2) conspiracy with a licensing authority to eliminate a competitor, (3) bribery of a public purchasing agent.13 Thus, Noerr immunity does not protect unethical practices when used in the judicial process. In California Motor, the allegations suggested that Noerr immunity did not apply because a pattern of fraud had been used to block access to a fair hearing in the courts and agencies. While the distinction between legislative and judicial processes is a clear one, the Court in California Motor did little to explain why the distinction makes sense. Why should such a distinction exist? There are several important differences. First, the notion of valid governmental action differs in the two regimes. In the legislative process, there are welldefined notions of procedurally valid action, though the boundaries that define substantively valid action are somewhat loose. It is understood in 11 12 13

Id. at 511. Id. at 512. Id. at 512–13.

358

Antitrust and the State

the legislative process that there will be losers and winners, and that “justice” or “fairness” will not necessarily determine who falls into which category. On the other hand, while hard to define precisely, there has long been a strong sense in the judicial process that the results should accord with justice. An appellate court decision explainable only by the fact that the prevailing party was able to win the backing of a majority of the judges on the panel satisfies no one. Once this principle is accepted, it follows that Parker immunity must mean different things in the legislative and judicial spheres. This must be the case because valid governmental action means different things in the two spheres. In the legislative process, procedurally valid governmental action is typically sufficient to invoke Parker immunity, and Noerr derivative immunity follows for efforts that result in such action. Valid governmental action is a narrower concept and requires meeting additional burdens in the judicial process. Hence, the very source of derivative immunity is narrower in the judicial process than in the legislative process. The second important distinction is that valid action in the judicial sphere is predicated on the assumption that the parties have acted in good faith. If the parties have acted in bad faith, one cannot be sure that the result of the judicial process is supported by the relevant evidence and is therefore valid. Legislative results are never overturned because one of the parties acted unethically in the lobbying process, provided that the legislators themselves complied with the law. On the other hand, courts sometimes overturn the results of the judicial process, because of bad faith on the part of one of the litigants, as in the case of a government prosecutor that withholds exculpatory evidence in his possession.14 Further, abuse of the judicial process can result in sanctions against a bad faith litigant.15 A related distinction is that the political process is open and participants know full well the incentives to exaggerate. The claims of those attempting to influence legislation are appropriately discounted. Often, some faction is waiting to capitalize on the demonstrably false claims of another. Because the process is open to the public, certain extremely unethical approaches to influencing legislators, such as offering bribes, are likely to be exposed. In this environment, it is unlikely that unethical publicity efforts will really be the sole cause of a given legislative 14 15

For example, Demjanjuk v. Petrovsky, 10 F. 2d 328 (6th Cir. 1993). For a discussion, see Chambers v. Nasco, Inc., 111 S.Ct. 2123, 2128 (1991).

I. Noerr-Pennington Doctrine

359

action. There will be other motivations and influences, whose impacts will be difficult to disentangle. To sum up, two distinctions seem important. First, the source of derivative immunity is narrower in the judicial process, because valid action is itself a narrower concept. Second, in the judicial sphere, valid action by decision makers, judge or jury, cannot be viewed as independent of the behavior of litigants, whereas in the legislative sphere the validity of a decision is not strongly dependent on the lobbying efforts of the interested parties. These distinctions suggest that the scope of Noerr immunity should be narrower in the judicial sphere than in the legislative sphere.

A. The Scope of Noerr Immunity in Legislative and Judicial Processes The cases following California Motor fall into two categories: “legislative sham” and “judicial sham.” Legislative sham cases clarify the scope of Noerr immunity in the legislative sphere, and judicial sham cases clarify Noerr immunity in the judicial sphere. Missouri v. National Organization for Women16 dealt with the scope of Noerr immunity in the legislative sphere. The National Organization for Women (NOW) refused to hold conventions in states that did not ratify the Equal Rights Amendment (ERA). Missouri sought to enjoin NOW. The Supreme Court held that the Sherman Act does not prohibit a boycott in a “noncompetitive” political arena for the purpose of influencing legislation. NOW ultimately sought to secure passage of the ERA. To the extent that they directed their publicity efforts toward this result, they fell squarely under the protection of Noerr. It also follows from Noerr that if an effort directed toward the legislature has the incidental effect of discouraging private parties from dealing with the target, the publicity campaign is still protected by Noerr. The interesting problem was that NOW had tried directly to influence other groups to boycott Missouri; the evidence clearly indicated this. The Noerr opinion does not say that derivative immunity applies to such activity. Indeed, the Court’s reliance in Noerr on the lower court’s findings that the railroads had directed their campaign at legislators implied that derivative immunity would not apply to efforts to influence private parties. 16

620 F. 2d 1301 (8th Cir.), cert. denied, 449 U.S. 842 (1980).

360

Antitrust and the State

How can one reconcile these cases? Start by looking closely at Noerr. The sham exception applies to publicity campaigns ostensibly directed toward government that in reality try to directly influence the business relationships of a competitor. Some relevant distinctions follow immediately. First, this was a publicity campaign directed toward government, not toward the business interests of a competitor. To see the distinction, return to my example of the minister. If my rival paid the minister to warn his congregation that they will go to hell if they consume my product, then there is nothing here directed toward the government, so he has no claim to derivative immunity under Noerr. NOW’s conduct is distinguishable because they had little concern over Missouri’s success in attracting conventioneers generally; they wanted Missouri to concede on the legislative matter. Second, and closely related, Missouri and NOW were not competitors in the market for conventions, or anything. The sham exception assumes the parties involved compete against each other in some way relevant to antitrust law. This contains an additional assumption that some conflict of economic interests makes the parties competitors in a natural sense. Here, there was no such conflict; NOW had no profit motive behind its conduct. Even if the Court had held that derivative immunity did not apply, it would not necessarily follow that NOW had violated the Sherman Act. There are boycott cases – for example, in the area of professional sports17 – in which courts have taken ethical and noneconomic considerations into account in assessing conduct under the Sherman Act. In addition, the courts have relied on First Amendment concerns as reasons to avoid a finding of a Sherman Act violation. In NAACP v. Claiborne Hardware Co.18 the Court held that the First Amendment protected the nonviolent elements of a boycott of white merchants organized by the National Association for the Advancement of Colored People and designed to make local government and business leaders comply with a list of demands for racial equality. Thus, even if Noerr did not immunize NOW’s boycott, Claiborne Hardware suggests that the First Amendment would have protected it; and at the very least, the Court would have taken First Amendment concerns into account in applying the rule of reason. 17

18

For example, Molinas v. National Basketball Association, 190 F. Supp. 241 (S.D.N.Y., 1961). 458 U.S. 886 (1982).

I. Noerr-Pennington Doctrine

361

The categories legislative-sham and judicial-sham are to some extent arbitrary, in the sense that one can generate examples that do not readily fit into either of them. Suppose a government regulates prices or quantities on the basis of the recommendations of a panel of experts drawn from the regulated industry. Without a more detailed description, it is hard to say whether this is an example of lawmaking through the legislative or the judicial process. More important than the label, then, is some grasp of the reasons a distinction should be made between legislative and judicial processes when considering the scope of Noerr immunity. Those reasons, not the label, should determine the outcome. The case that best illustrates the problem of separating substance from form is Allied Tube & Conduit Corp. v. Indian Head, Inc.19 Allied Tube, a steel conduit manufacturer, attempted to prevent Indian Head, a manufacturer of polyvinyl chloride conduit, from gaining approval by the association that writes the National Electrical Code (NEC). Approval by the NEC is important because many states adopt the NEC with no alterations, and private associations such as Underwriters Laboratory look for NEC approval. Allied Tube recruited more than one hundred individuals to register and participate in the annual meeting of the National Fire Protection Association, which designs the NEC through a process called “consensus standard making.”20 The end result was determined by majority voting. Allied Tube instructed its recruits on how to vote and used hand signals and walkie-talkies to communicate voting strategies to its supporters. In the end, the polyvinyl chloride conduit missed approval by only four votes, which suggests that the voting of the recruits was absolutely necessary for the result. The Court held that Allied Tube’s activities violated Section 1. Specifically, the Court held that “where . . . an economically interested party exercises decision-making authority in formulating a product standard for a private association that comprises market participants, that party enjoys no Noerr immunity from any antitrust liability flowing from the effect the standard has of its own force in the marketplace.”21 Because the NFPA is a private association, the result may seem unsurprising. However, the district court had decided in favor of Allied Tube on the grounds that (1) NFPA was a quasi-legislative body and (2) that 19 20 21

486 U.S. 492 (1988). Id. at 494. Id. at 509–10.

362

Antitrust and the State

Allied Tube’s efforts were really efforts to influence government action. The former theory implies that decisions of the NFPA enjoy some protection under the Parker doctrine, and thus derivative immunity follows solely from the special relationship between the state and the NFPA. The latter theory implies that derivative immunity applies to efforts to influence the NFPA even if it is accepted that NFPA is not a quasi-legislative body. The Court rejected both theories. The Court began its rationale by noting that the rule of reason generally applies to product standard-setting by private associations. This is because the setting of design and performance standards can generate procompetitive benefits. The Court did not detail precisely what those benefits were, but it is not hard to suggest some of them. Standards provide quality signals to consumers. To the extent that these signals help consumers distinguish well from poorly designed products, they enhance competition and product diversity. The Court rejected the theory that derivative immunity applies because the NFPA is a quasi-legislative body for two reasons. First, the NFPA had no government authorization to set standards. Second, the NFPA was composed of individuals who had incentives to restrain trade. The Court rejected the theory that derivative immunity applies because efforts to influence the NFPA were in reality efforts to influence government for the simple reason that it proved too much, by suggesting that any private group that restrains trade in anticipation of the government enforcing their restraints would have protection by derivative immunity. Thus, a group that fixes price in anticipation of the government adopting the agreed-on level as a minimum price regulation would receive immunity under this theory. Similarly, bribery that resulted in a minimum price regulation would also be immunized. The Court said that nothing in the case law suggests that Noerr immunity would extend to these instances. These examples show that the limits on derivative immunity depend on the means used to influence governmental action. The Court has never fully articulated the limits. However, two propositions might explain these limits: (1) Parker immunity applies to valid governmental action, (2) Noerr immunity applies only to efforts to influence valid governmental action. The difference between these propositions is slight. The first implies that Parker immunity does not apply even to governmental actors when they violate procedural norms. The second implies that even if governmental actors can claim a shield under Parker, the parties who attempt to influence them may not be able to hide behind Noerr.

I. Noerr-Pennington Doctrine

363

Valid governmental action is more than just a function of the result, it depends on the procedure followed. The bribery example is one of clear procedural invalidity. The example of a group that fixes price in order to influence regulation is procedurally invalid if the government relies on the assumption that the private group is impartial or acting for the public’s benefit. The activities of the NFPA fell within the doubtful zone illustrated by the price regulation example. The procedural invalidity was apparent in the facts: Allied Tube had rounded up a group, including some not very well-informed individuals, to attend the NFPA meeting simply to cast votes in favor of its position. The state governments, on the other hand, relied on the assumption that the standards promulgated by the NFPA reflected the considered judgment of industry experts. The second reason the derivative immunity argument in Allied Tube should have failed, as it did, is that the conduct of the defendants more closely resembled those of the defendants in California Motor than the conduct examined in Noerr.The standard-setting process would ultimately influence government, and one could say aimed to influence government, but it was not carried out in the open manner typical of a publicity campaign. The channel of influence is better analogized to that of an administrative agency. The decisions of such agencies often become law, and in this sense are legislative. However, unethical behavior at this level of government does not receive as much protection under Noerr as it would in the legislative process. If one accepts the argument that this case is similar to California Motor, then the reason for applying the sham exception is clear. Here, the defendants had “usurped [the] decision-making process” of the NFPA, and thereby denied their competitors a fair hearing.22 The notion of procedural invalidity as a basis for finding an exception to Noerr immunity was further developed in FTC v. Superior Court Trial Lawyers Ass’n.23 In addition, Trial Lawyers illustrates the limits of the Court’s holding in NOW. A group of Washington, D.C., attorneys who represented poor defendants in criminal cases boycotted assignments until the city council 22

23

In Sessions Tank Liners Inc. v. Joor Manufacturing Inc., 17 F. 3d 295, 299 (9th Cir. 1994) the 9th Circuit introduced the additional requirement that the plaintiff prove “the stigma of banning the plaintiff’s product from a uniform code caused independent marketplace harm to the plaintiff in jurisdictions that permitted the use of the plaintiff’s product.” The requirement seems to be entirely foreign to the theories underlying Parker and Noerr. 493 U.S. 411 (1990).

364

Antitrust and the State

passed legislation raising their hourly fees from $30 to $35. The boycott occurred in 1983 and the fee cap of $30 had been set in 1970. Given the high rates of inflation over the late 1970s, the trial lawyers had suffered a substantial loss in the value of their compensation. The FTC held that the boycott violated Section 5 of the FTC Act, and one of the questions addressed by the Court was whether Noerr immunity applied to the defendants. The Court held that they were not immune. The reason is easy to see when the facts are compared to those in NOW. Although the trial lawyers were attempting to influence legislation, it was not the kind of social legislation at issue in NOW. The defendants in NOW obviously had an economic interest in the ERA, but that interest took a back seat to general notions of equality of opportunity. In other words, the pecuniary interests of the defendants were relatively remote in NOW. The legislation sought by the trial lawyers, on the other hand, would raise their wages, and the purported social benefits were of second-order importance. The method the defendants used was also important to the Court. The trial lawyers used an anticompetitive conspiracy to force the legislature to raise their pay. If they had accomplished the same result by lobbying, there would have been no violation of the antitrust laws. The Court failed to explain precisely what separates valid from invalid procedures, and the question remains unanswered. However, the holding clearly implies that boycotting in order to force the legislature to pay a higher wage is an invalid approach, and is therefore not protected by Noerr. Reversing the appeals court, the Court held that the lawyers’ boycott was per se unlawful in spite of the First Amendment concerns implicated by the case. Because of the First Amendment issues, the appeals court had applied rule-of-reason rather than per-se analysis. The Supreme Court held that the First Amendment could not be cited as a basis for creating an exception to the per-se rule against price fixing. The Court argued that a First Amendment-based exception to the per-se rule would create an enormous gap that almost every boycotting cartel could exploit in order to evade punishment. On economic grounds one should question the decision in Trial Lawyers. Without restrictions on entry into the field, it would be difficult for the trial lawyers to raise their pay beyond the level that compensates them for the skills developed on the job. If their pay reached the monopoly level, other lawyers would enter and, because their presence would require wider spreading of assignments, drive average compensation down to the competitive level. Indeed, it is easy to see how this process

I. Noerr-Pennington Doctrine

365

would work in this case. The Court noted that although there were roughly one hundred “Criminal Justice Act” (CJA) regulars, more than twelve hundred lawyers had registered for appointments as counsel to poor defendants. It seems, then, that there were many lawyers ready to shift into CJA work if the wage reached a supracompetitive level. The view that the trial lawyers should have taken what they got and not complained is not persuasive. Lawyers who specialize in a field develop informational capital specific to the field of specialization. If the state refuses to raise the pay of lawyers who represent poor criminals, it may be difficult for the experienced lawyers in that group to shift into a new field. The Court’s refusal to forsake per-se analysis in view of the First Amendment concerns is also questionable. The Sherman Act was not designed to stamp out all forms of boycott, including those motivated primarily by political concerns. That is the central message of NOW. In every case, the courts must determine whether the allegedly anticompetitive behavior actually is of the sort that the Sherman Act aims to suppress. Since the Sherman Act was not designed to suppress political expression, or attempts to influence legislation, it is appropriate for courts to take the expressive elements into account in determining whether the statute applies. One way to do this is to take the First Amendment issues into account in the process of applying the rule of reason.24 Finally, I should note that this is one example where the Sherman Act leads to a harsh result that would have been unlikely under the pre-1890 common law regime (see Chapter 2). The early criminal conspiracy prosecutions of workers involved violence and other clearly unlawful conduct. Peaceful combinations of workers were almost never 24

This argument may seem inconsistent with our understanding of the boundaries set by Chicago Board of Trade. Recall that under Chicago Board of Trade, the rule of reason is confined to a consideration of competitive conditions. First Amendment issues fall outside of that area. How can one reconcile the Chicago Board of Trade view of the rule of reason with a consideration of First Amendment issues? The answer, I think, is that one should approach each question through two steps. First, one should determine whether the Sherman Act applies, and second, whether the behavior at issue violates the Act. Under this approach, it is no violation of the Chicago Board of Trade doctrine to consider First Amendment issues in the first stage. In borderline cases, where it is uncertain whether the statute applies, one presumably should carry over some of the issues examined in the first stage into the second stage analysis. The alternative is to hold that the statute does not apply in every borderline case. But that would be undesirable because it would create precisely the gaping hole the Court was keen to avoid creating in Trial Lawyers.

366

Antitrust and the State

prosecuted. Indeed, Justice White’s dissent in Trans-Missouri warned that adoption of per-se analysis would result in prosecutions under the Sherman Act in cases where workers combined to raise their wages. This is what we have in Trial Lawyers. This is a good place to take stock of the lessons learned from the Noerr immunity cases. They have suggested that immunity applies to efforts to influence governmental action unless those efforts fall within the sham exception. However, this description is insufficiently precise. The exceptions to Noerr immunity cannot all be characterized as various types of sham. The exceptions to Noerr immunity can be put in two categories. The first involves direct efforts to influence the business relationships of a competitor. This was the original description of the sham exception in Noerr. The reason efforts to influence business relationships do not fall within the zone of Noerr immunity is that they cannot properly be characterized as efforts to influence legislation or governmental action. It is important also to stress the word competitor in the statement of the sham exception. If the defendant’s actions are not motivated by a desire to restrain competition in order to enhance its profits, then they fall outside of the scope of the Sherman Act and the Court is likely to find that Noerr immunity applies. For example, the competitionrestraining efforts of the defendants in NOW received immunity under Noerr because it was apparent that they were not driven by the profit motive. NOW had attempted to directly influence the business relationships of convention centers in Missouri, but because they were in no antitrust-relevant sense in competition with the convention centers, their boycott did not raise the concerns that motivated passage of the Sherman Act. One can view this as a question of specific intent, or one of the boundaries of rule of reason analysis, but at bottom these cases reflect a judgment that the underlying activity is not of the sort the Sherman Act aims to suppress. The second category of exceptions to Noerr immunity is made up of instances of procedural invalidity. Noerr immunity derives from the Parker immunity that applies to valid governmental action. It follows that with procedurally invalid governmental action, Parker immunity may not exist, and neither may Noerr. Although the Court has never explained precisely what kinds of procedural invalidity remove a party from the zone of Noerr immunity, it has indicated that procedural invalidity is a basis for denying Noerr immunity. Allied Tube discusses two examples of procedurally invalid actions that fall outside of Noerr immu-

I. Noerr-Pennington Doctrine

367

nity: (1) a group that forms a cartel in the hope of influencing a governmental body to adopt the fixed price as a regulated price, and (2) where a party influences legislation through bribery.25 This neat theoretical structure connecting Parker and Noerr immunity was almost ruined by the Supreme Court’s opinion in City of Columbia v. Omni Outdoor Advertising,26 which took an extremely narrow view of the notion of procedural invalidity as a basis for denying Noerr immunity. Columbia Outdoor Advertising used its connections with the City’s (Columbia, South Carolina) administration and the City Council to push for local ordinances restricting erection of new billboards. The laws gave Columbia Outdoor Advertising (COA) an advantage because it owned the vast majority of billboards already in use. Omni brought suit against COA and the City claiming the two had conspired to hinder its ability to compete. The Court, speaking through Justice Scalia, held that both defendants had immunity. The first issue in City of Columbia was whether Parker immunity applied to the City. The Court held that Parker immunity applied because the state had delegated authority to regulate billboards to the City of Columbia. One could argue that the City had exceeded the scope of its authority, but the Court rejected this approach on the ground that Parker immunity must be interpreted liberally (so long as it is the state that is acting) in order to realize the benefits of federalism. If federal antitrust law reviewed all of the regulatory actions of states, there would be little scope for variation and experimentation at the state level. The appeals court had held that Parker immunity did not apply because the City had acted in conspiracy with COA. The Court held that there is no conspiracy exception to the Parker doctrine. The Court said that in almost every instance of state regulation, there are winners and losers, and it will always be possible for the losers to claim that the state acted in collusion with the winners. The end result would be topdown regulation through the Sherman Act, again violating the federalism principle.27 The Court briefly considered whether procedural invalidity such as bribery or some other violation of state or federal law could serve as a 25 26 27

Allied Tube and Conduit Corp. v. Indian Head, Inc., 486 U.S. 492 (1988). 499 U.S. 365 (1991). The facts suggested that the conspiracy claim was more than speculative. The officers of COA were politically connected fellows who funded the campaigns of the mayor and some of the council members. City of Columbia, 499 U.S. at 367–71. The Court never considered whether the evidence in this case indicated the existence of a conspiracy.

368

Antitrust and the State

basis for limiting Parker immunity. Surprisingly, the Court rejected this argument. The reason was that procedural invalidity has, according to Justice Scalia, nothing to do with the ends of the Sherman Act. This is a surprising statement because the Parker doctrine generally is said to protect valid governmental action, and this implies that some actions by governmental agents may not be immune because of some procedural invalidity. The Court’s holding suggests that Parker immunity does not depend, strongly at least, on the validity of the government’s action. The Court recognized only one exception to Parker immunity: where the state acts as a market participant. Suppose, for example, that the state owns a grocery store, but has not monopolized the grocery store market. Suppose the state provides certain advantages that make it difficult for rival stores to compete. Then the state may be held liable under the market participant exception. The second issue in City of Columbia was whether Noerr immunity applied to COA. Justice Scalia began by defining the sham exception as encompassing “situations in which persons use the governmental process – as opposed to the outcome of that process – as an anticompetitive weapon. A classic example is the filing of frivolous objections to the license application of a competitor with no expectation of achieving denial of the license but simply in order to impose expense and delay. A sham situation involves a defendant whose activities are not genuinely aimed at procuring favorable government action at all.”28 With this definition in hand, the Court could easily conclude that the sham exception did not apply. COA had used the outcome, not the process, to gain a competitive advantage. The plaintiff had urged a conspiracy exception to Noerr immunity. The Court rejected this on the ground that its application would eventually lead to a violation of the federalism principle. The conclusion is interesting because in California Motor, the Court referred to conspiracy with a licensing authority as an example of unethical conduct that might violate the Sherman Act. Here, the Court’s holding implies that conspiracy with a licensing authority violates the Sherman Act if that authority is analogous to a court or administrative agency. If the licensing process is as open as the legislative process, then a violation is doubtful. The Court relied on the definition of the sham exception stated in California Motor. However, California Motor dealt with a judicial sham, not a legislative sham. The decision established the norm that the sham 28

Id. at 366.

I. Noerr-Pennington Doctrine

369

exception should have greater scope in the judicial sphere than in the legislative sphere. It is one thing to say this, and quite another to take the test developed for judicial sham cases and apply it to a legislative sham. However, Scalia framed the question in a way that could lead to only one answer: Noerr immunity could not apply. The California Motor test raises the interesting question whether it is possible for a legislative sham to satisfy its requirements. Is it possible for someone to act in the legislative process in a way that effectively denies access to that forum to a competitor? Is there some way to use the legislative process in a way that disadvantages a competitor? The Court did not provide an example. Perhaps bribery of legislators might serve as one, because it denies others access to the paid-off legislators. But the Court’s discussion of the absence of a procedural invalidity exception to Parker immunity suggests that it might hold that bribery is also irrelevant. In the end, one has the impression that little remains of the notion of procedural invalidity as an exception to Noerr immunity. The sham exception, as originally defined in Noerr, applies to those cases in which a firm attempts to interfere directly with the business relationships of a competitor and at the same time, tries to present the effort as an appeal to the legislature. That exception remains intact after City of Columbia. However, the earlier immunity opinions had suggested that the category of exceptions to Noerr immunity also includes cases of procedural invalidity, such as bribery. This set of possible exceptions has been erased by City of Columbia. Are there any types of procedural invalidity that remain intact as bases for finding an exception to Noerr immunity? Yes. The Trial Lawyers boycott failed to receive protection under Noerr in part because of the Court’s conclusion that the lawyers had used improper means. City of Columbia does not, apparently, upset this implication of Trial Lawyers. What remains of the procedural invalidity theory will be clarified in future cases. The Court will either ignore the language on procedural invalidity in City of Columbia, or it will leave us with the sham category as the only set of exceptions to Noerr immunity.

B. Sham Lawsuits Suppose A, the incumbent, brings a patent infringement suit against B, the entrant. Suppose further that A has a slight though nontrivial probability of winning, and the damages A expects to receive are very small.

370

Antitrust and the State

The real reason A is bringing the suit is to saddle B with litigation expenses, and to raise suspicions among B’s business relationships. The ultimate aim is to make entry difficult for B. Is this lawsuit a sham in the judicial sphere? On purely economic grounds, this lawsuit is an anticompetitive sham. Firm A aims to make entry difficult, not to enforce its rights under the patent laws. There is language in California Motor suggesting that the lawsuit also satisfies the legal definition of sham. Specifically, California Motor referred to allegations in the complaint that the defendants had “instituted the proceedings and actions . . . with or without probable cause, and regardless of the merits of the cases.”29 However, California Motor also stressed the fact, in that case, that the defendants’ pattern of baseless claims served, in effect, to deny their competitors access to the relevant agencies and courts.30 What precisely characterizes a sham in the judicial sphere was not set out clearly in California Motor. Reading it expansively, the opinion suggests that any lawsuit that aims primarily to exclude a competitor from the market in order to enhance the market power of the incumbent is a sham. Thus, even if the plaintiff has a valid claim, if the real motive of the lawsuit is anticompetitive exclusion, the plaintiff may lose Noerr immunity. Alternatively, reading California Motor in a narrow sense leads to the conclusion that a lawsuit is a sham only when the action itself is brought in or conducted in bad faith. California Motor refers to an intent to “discourage and ultimately prevent” the target from “invoking the processes of the administrative agencies and courts.”31 Discouragement would seem to require some kind of bad faith on the part of the complainant. For if the plaintiff brings the claim in good faith, it cannot discourage the opposing party from invoking the judicial process. A party who brings or conducts an action in bad faith and who also prevails, however, would quite plausibly discourage the losing party from relying on the courts for protection. The narrow reading of California Motor implies a definition of judicial sham that excludes a number of lawsuits that would easily be considered shams on economic grounds. Which reading, the expansive or narrow one, is correct? 29 30 31

California Motor Transport Co. v. Trucking Unlimited, 404 U.S. 508, 512 (1972). Id. at 513. Id. at 512.

II. Parker Doctrine

371

The Court chose the narrow reading in Professional Real Estate Investors, Inc. v. Columbia Pictures Industries,32 where it held that an objectively reasonable lawsuit cannot be a sham. In order to be a sham, the suit must be objectively baseless. The Court announced a two-part test: First, the lawsuit must be objectively baseless, second, the court should determine whether the lawsuit attempts to interfere directly with the business relationships of a competitor. If the answer to both of these questions is yes, the lawsuit is a sham. The Court held that the copyright claim at issue in Professional Real Estate Investors was objectively reasonable, and therefore could not be deemed a sham. Although the Court defended its holding as one required by precedent, there is a good policy argument in favor of the narrow reading of California Motor adopted in Professional Real Estate Investors. The “indifference as to outcome” standard that would serve as the alternative to the objective baselessness requirement would require courts to examine the subjective intent of the litigant. Subjective intent is hard to discern. Mistakes in this area are likely. The mere chance that a court would erroneously hold that a plaintiff brought a sham lawsuit, and is therefore liable for treble damages, would deter antitrust plaintiffs from filing suit. Legitimate claims that might give rise to a sham counterclaim would be discouraged, and Noerr’s protection of the right to invoke the judicial process would be compromised.

ii.

PARKER

doctrine

Federal antitrust law does not apply to the state as an actor. The theory courts and commentators most frequently offer for this is that immunity helps secure the benefits of federalism. With immunity, the states have greater freedom to experiment. Successful experiments will be copied. People will move from unsuccessful states to successful ones.33 The effects of federalism on incentives to experiment are actually somewhat more complicated than this simple theory suggests. Federalism may discourage experimentation where one state can free-ride off the investments of another.34 But no one has seriously questioned the 32 33

34

508 U.S. 49 (1993). Frank H. Easterbrook, Antitrust and the Economics of Federalism, 26 J. L. & Econ. 23 (1983). Susan Rose-Ackerman, Does Federalism Matter? Political Choice in a Federal Republic, 89 J. Pol. Econ. 152–65 (1981).

372

Antitrust and the State

notion that federalism is a desirable administrative system. Hence, I will set the desirability question to the side, and take it as given that federalism is a desirable property of government. The modern doctrine of state action immunity – that is, the doctrine of Parker v. Brown – can be summarized as follows: there can be no immunity without (1) adequate and active public supervision and (2) a clear state purpose to displace competition. The Court established the active supervision requirement in California Retail Liquor Dealers Assn v. Midcal Aluminum.35 In this case, a California statute required wine dealers to charge the resale prices specified by their supplier. The Court held that the state statute was preempted by federal antitrust law because the statute had the effect of a resale price agreement but did not regulate the resulting resale price. Midcal illustrates the close relationship between state action and preemption questions. A liberal approach toward immunity would be equivalent to permitting the state to preempt the Sherman Act every time it decides to regulate some aspect of an area of business. If the allegedly anticompetitive behavior is in some sense required by or a direct result of the state’s regulatory program, then a liberal approach toward immunity would require the Sherman Act to yield to the state’s preferences. In addition to the supposed benefits of federalism, there is a second policy argument made in favor of a liberal approach toward immunity – that is, a broad immunity doctrine. If the active and adequate supervision prong of the Parker doctrine were enforced rigorously, then it would be clear to the state actors that the way to be sure of being immune from antitrust prosecution is to control all aspects of a business. In other words, the Parker doctrine encourages Soviet-style “command and control” regulation, a type that experience has shown to be disastrous. The Supreme Court clarified its position on these issues in Federal Trade Commission v. Ticor Title Insurance Company.36 Several states had authorized title insurers to establish rating bureaus, for the purpose of pooling information and establishing uniform rates for their members. The FTC charged that the insurers in several states had used the rating bureaus to maintain collusive price agreements on title search and examination services. The FTC did not challenge the practice of setting uniform rates for insurance against the risk of loss, because the 35 36

445 U.S. 97 (1980). 504 U.S. 621 (1992).

II. Parker Doctrine

373

McCarran-Ferguson Act exempts the “business of insurance” from federal antitrust regulation. Although title insurance practices in thirteen states were challenged in the initial FTC complaint, only four states remained when the case reached the Supreme Court. In these four states – Connecticut, Wisconsin, Arizona, and Montana – title insurance rates received state approval through a method called the “negative option.” Under this approach, the insurer filed proposed rates, and if the state did not object to the filed rates by a certain date, usually within thirty days, they were considered approved and became effective. The defendant title insurers claimed that they were immune under the Parker doctrine. Since the rating bureaus had been licensed and authorized by the state to set uniform prices, the only issue before the Court was whether the activity of the bureaus received adequate and active supervision from the state. The Court held that they had not, that the negative option did not satisfy the requirement of active state supervision. It is important to note that the Administrative Law Judge had found that the four state programs under examination were even more relaxed than the description of the negative option implies. There was little evidence that state regulators had seriously reviewed the rates filed by the bureaus – and in two of the states, Wisconsin and Montana, the regulators never received all of the information essential to the review process. It may seem that the Court’s conclusion is obviously right. How, after all, could the negative option programs, as implemented in the states in which the practices were challenged, have satisfied the active supervision requirement? Perhaps the only way to answer this question is to consider the other side of the issue. If the negative option plan is inadequate, then where should the line be drawn between active and inactive state supervision? Should the FTC have the authority to examine the details of a state regulatory program in order to determine whether state regulators have complied with its notion of active supervision? Suppose an activist FTC confronts a program that was poorly funded by the state because it is not a major priority within the state’s government? We arrive once again to federalism. If the government of Connecticut decides that its regulatory priorities do not permit it to staff and fund the state insurance department at a level that would enable it to engage in a rigorous review of the filings of insurance rating bureaus, then why should the FTC have the power to reverse that decision? Not all states

374

Antitrust and the State

have the money to meet the FTC’s tastes for regulation. Will the FTC fund the additional burdens it pushes on to the states? These questions force a hasty retreat to the position that states should be largely free to set their own priorities. Is there a policy argument in favor of the active supervision requirement? The Court offered one in Ticor. If state action immunity is granted liberally, as is implied by a weak active supervision requirement, then every decision by the state to regulate prices or quantities immediately displaces federal antitrust law. The state may not wish to do this. For example, Connecticut may license rating bureaus to share information and establish uniform rates on the theory that the established rates will reflect the general risk of loss observed in the population of customers, plus some allowance for a competitive rate of return. However, Connecticut may not, and probably does not, intend for the rating bureaus to serve as shields that protect price-fixing schemes from federal antitrust suits. Federal antitrust law arguably increases Connecticut’s freedom if it allows the state to adopt the former policy without also simultaneously adopting the latter. In one passage the Court referred to broad state action immunity as creating an “attractive nuisance”37 in the regulatory area. The argument seems to be this. If state regulation provides broad protection from liability under the Sherman Act, then industries in which collusion is sustainable will have an interest in setting up some regulatory program that is capable of providing protection. The active state supervision requirement raises the cost to this type of actor. If such an actor seeks immunity, it will also have to accept heavy-handed regulation. Thus, the active supervision requirement may be justified as a deterrent to rent-seeking efforts to establish antitrust shields under state law. It is difficult to find a middle ground. However, the absolutely nonexistent supervision that would give states the fullest freedom to adopt different approaches to regulation is not necessarily defensible on consumer welfare grounds. The reason is that it encourages private efforts to obtain immunity through the erection of undermanned, understaffed regulatory programs. At some point, the welfare costs of such activity start to outweigh the experimentation benefits provided by federalism. Where this point occurs is hard to say; however, the notion that such a point exists seems plausible. It is the role of courts to examine each case individually in order to find rules that maximize the freedom 37

Id. at 637.

III. Error Costs and Immunity Doctrines

375

of states to experiment and at the same time control incentives to obtain immunity for anticompetitive purposes. What about municipalities? The state action doctrine applies to cities and townships as well. Lafayette v. Louisiana Power & Light Co.,38 held that a municipality’s allegedly anticompetitive behavior in operating its local electric distribution system was subject to the antitrust laws unless directed or authorized by the state in connection with a state policy to substitute regulation or monopoly for competition. In Town of Hallie v. City of Eau Claire,39 the Court elaborated on its state authorization requirement and also held that the requirement of “active state supervision” (see Midcal) did not apply to authorized municipal action. On state authorization, Hallie held that although the municipality must be acting in accordance with a clearly articulated state policy to displace competition, there is no requirement that the state statute authorizing the municipality’s activity describe in detail everything the municipality may do. As long as the municipality’s actions are clearly “contemplated” or allowed for in the statutes authorizing the activity, immunity applies.

iii. some final comments: error costs and immunity doctrines To this point, I have tried to state the Noerr and Parker immunity doctrines, which happens to be a difficult task. Are the doctrines defensible on economic grounds? Let us return to the error cost analysis introduced in Chapter 6 and applied in other chapters. A more expansive zone of antitrust immunity raises the likelihood of an erroneous finding of innocence, or a “false acquittal.” Narrowing the zone of immunity, on the other hand, raises the likelihood of an erroneous finding of guilt, or a “false conviction.” In considering whether the legal standard should be made more or less favorable to the defendant, we should consider the relative costs of false acquittals and false convictions. Before we can say anything about the relative costs of false acquittals and false convictions in antitrust immunity doctrine, we need a set of assumptions on the nature of regulation. This is important because our premises will determine our assessment of the error costs associated with different legal standards. For example, if we believe that the state 38 39

435 U.S. 389 (1978). 471 U.S. 34 (1985).

376

Antitrust and the State

generally acts infallibly for the purpose of enhancing social welfare, then we should prefer that the state enjoy a great deal of freedom to regulate without subjecting the beneficiaries of regulation to antitrust liability. On the other hand, if we believe that state regulation, instead of being motivated by public interest concerns, is often motivated by the rent-seeking efforts of private interest groups, then we should prefer to narrow the zone in which the beneficiaries of regulation can escape antitrust liability. This is the familiar conflict between “public interest” and “public choice” views of regulation, introduced in Chapter 2. In that chapter, I examined the enactment of the Sherman Act from a public choice perspective – that is, from a perspective that views regulation as often the product of interest group pressure. I will apply that perspective here. From the public choice perspective, regulatory programs should be understood in part as products sold by politicians to interest groups. According to this view a city adopts a “living wage” ordinance, requiring all firms doing business with the city to pay wages no less than twice the federal minimum, not only because some members of the city council believe workers should earn more, but also because unionized firms have demanded this protection in return for their campaign contributions. Noerr immunity protects the unionized firms from antitrust suits brought by nonunionized rivals, who understand the anticompetitive motivations behind the campaign for a living wage. This example suggests that in the area of state action immunity the costs of false acquittals probably exceed those of false convictions. Why? A broad zone of Noerr immunity encourages firms to seek protection from competition through various state regulatory initiatives, and this is costly to society. A narrow zone of Noerr immunity, on the other hand, discourages private actors from seeking protection through regulation. One might argue that this is costly, too, to the extent that it effectively silences a class of petitioners. However, the class that is typically silenced is a small and concentrated group, the competitive beneficiaries of a particular regulation. Their voices may be replaced by those of other noncompetitive beneficiaries or disinterested parties. For example, if firms are fearful of using their campaign contributions to foster competitionrestricting legislation, their lobbying activity may be replaced by that of other parties who have no immediate pecuniary interest in the legislation. In view of the foregoing, it seems that the better argument, on economic grounds, favors restricting the zones in which Noerr and Parker

III. Error Costs and Immunity Doctrines

377

immunity apply. In particular, as the Supreme Court suggested in its Ticor decision, it may be helpful to make it costly for private actors to seek protection through regulation. I use the term protection here in a broad sense to cover both the erection of barriers to competition and the facilitation of collusion. If it is costly for firms to seek such protection through regulation, they will seek it less often. The activestate-supervision requirement upheld in Ticor has the benefit of making it costly for firms to seek protection through state regulation. With respect to Noerr immunity, the preferable approach, given that the First Amendment permits firms to petition legislators, is to raise the cost (or the price) of some particularly effective methods of securing protection from competition. This suggests an altogether different approach from that taken by the Court in City of Columbia. Rather than minimizing the number of exceptions to Noerr, the Court should recognize procedural invalidity as a basis for creating exceptions to Noerr immunity. This would have the additional benefit of making Noerr doctrine more consistent with its original description as derivative of Parker immunity.

Index

A.A. Poultry Farms, Inc. v. Rose Acre Farms, Inc., 216–18 abusive business, 167, 171 conduct-focused test of, 187, 188, 188n13, 193 intent to monopolize for, 187, 188, 193, 194 refusal to deal by, 187, 247 Section 2 violation for, 188, 190 access. See also essential facility local association, 180 problems of open, 208–11 up-front fee for, 212 acquittals, false, 130–1, 189, 214, 220, 351, 375 administrative concerns courts errors and, 130–1 economic reasonableness v., xiii–xv, 92, 106–7, 116, 171 per-se rule and, 122 Administrative Procedure Act, 48 adverse selection, 5–6, 24 Aghion, Phillipe, 199n30, 200–1 agreement(s), 28, 75, 132–3, 222 to boycott, 166, 168, 169–70, 184 cheating on collusive, 145, 153–4

Clayton Act against monopolization by tying, 287 dealership, 256, 260, 264 exclusivity, 262–4 inference doctrine development for, 132–40, 145 inferred versus not, 141 Interstate circuit and, 133–8 noncompete, 33–4, 100, 103, 257, 264 resale price, 269 termination of (vertical restraints), 270–1, 272–3, 275–6 unilateral contract theory for, 138–40, 158 unilateral contract theory rejection for, 140–3, 158 vertical restraint, 270–8 vertical restraints and exclusivity, 262–4 airline industry, as example of parallel pricing, 73 Akerlof, George, 24–5 Albrecht v. The Herald Co., 123n17, 262, 262n21, 262n23, 273, 277 379

380

Index

Alcoa. See United States v. Aluminum Co. of America (Alcoa) Allied Tube & Conduit Corp. v. Indian Head, Inc., 361–3, 366 allocation horizontal, 126 territorial, 118, 119 Ambook Enterprises v. Time, 141n26 American Column & Lumber Co. v. United States, 145–9, 149–52, 154 American Tobacco Growers, Inc. v. Neal, 208, 210, 211 American Tobacco v. United States, 77, 139–40, 158, 195 anticompetitive theory tying, after Chicago School and, 281–3, 288 tying and difficulty with, 307–10 tying, Chicago School and, 279–81 vertical mergers and, 335 Antitrust Division. See Justice Department antitrust law. See also Clayton Act; Sherman Act Antitrust Paradox (Bork), xii Appalachian Coals v. United States, 107–9, 111 apprehension enforcement and probability of, 50, 51 Areeda, Philip, 220–1, 230, 248n15 Arizona v. Maricopa County Medical Society, 122–5, 122n17, 126n23 Arrow-Hart & Hegman Electric Co. v. FTC, 318n24 ASCAP, 120–2 Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 203–6, 207, 210, 228, 231n2

Associated Press v. United States, 177–81, 182, 184, 208–10 association dissemination of information and, 146–9 as extragovernmental agency, 170, 171, 179–80 new competition practice of, 146 price fixing in, 106, 118–19, 121 price-setting cooperative, 120 product standard setting by private, 361–2 protection of access for local, 180 restraint of trade and bylaws of, 177–8 veto in, 178, 180, 181 Atlantic Richfield Co. (ARCO) v. USA Petroleum Co., 218, 262n23 atomism, 40, 190 perfect competition and, 4, 8 attempt, at monopolization Clayton Act and, 287 dangerous probability requirement of, 244–5, 246, 248–51, 263 definition of, 244 evidence, objective, in, 245, 250 evidence of unfair tactics in, 247, 249 exclusion of competitors with, 247 right/refusal to deal in, 246–8, 246n7, 263–4 specific intent of, 245, 246, 247–51 the Swift formula and modern doctrine of, 244–8 violation of, 244 attorneys’ fees, and deterrence, 58 author, opinion of, xi Automatic Radio Mfg. Co, v. Ford Motor Co., 295n42

Index Bain, Joe, 16 bargaining, 159–60, 342–3 barriers, 23. See also entry artificial (privately created), 15–16 natural barrier to, 16 Barry Wright Corp. v. ITT Grinnell Corp., xv, 199n32, 214–15, 219, 224 basing point pricing, 81 cases, 155–9 competition and, 162–5 cross-hauling and, 162 definition of, 154–5 delivered pricing system and, 158–9, 161–2 economics, 160–5 and kinked demand curve, 163 freight absorption and, 162 most favored customer clause and, 159–60 per-se rule not, 158–9, 164 Bauer & Cie v. O’Donnell, 261 Becker, Gary, 44 Bendix Corp., 348–9, 350 benefit, potential competition, 345 Berkey Photo v. Eastman Kodak Co., 203 Bertrand, Joseph, 76–7, 76n15, 83–5, 150 Blair, Roger, xvi blanket license arrangement, 120–1 block booking, 292 BMI. See Broadcast Music Inc. v. Columbia Broadcasting System Bobbs-Merrill Co. v. Strauss, 261, 261n19 Bolton, Patrick, 199n30, 200–1 boycotts, 29, 155, 157, 166, 168, 169–70 agreement to, 166, 168, 169–70, 184, 264

381

Associate Press and, 177–81 competition effected by, 170–1, 174–5 conspiracy’s proof/rule of reason for, 166, 167, 168, 169–70 direct-selling wholesalers, retailers and, 167 efficiency defenses for, 176 eliminating of firms by, 174–6 exclusionary plan and, 173–4, 177, 180 exclusions, rule of reason and, 181–2 free option arrangement, members and, 167–8 Klor’s paradox, boycott doctrine and, 174–7 no effect on competition and, 172–4 in noncompetitive political arena, 359, 360, 363–4, 365, 366, 369 per-se rule and, 166, 169–70, 171, 173, 177, 179–82, 184, 185 post-BMI/Sylvania and, 172, 181–5 post-Socony, 170–7 pre-Socony, 166–70 price-fixing, 169, 171 proof versus conspiracy to, 171 public harm and, 173–6, 184 restriction of distributors freedom prohibited in, 168–9 rule of reason and, 166–70, 173–4, 182, 183, 184–5 brand loyalty, 73 brand name, 25–6 Brandeis, Justice, 148, 152 Breyer, Judge, 215 Breyer, Justice, xv bribes, 8, 358, 367, 369 bright line rule, 215, 220–1, 320

382

Index

Broadcast Music Inc. (BMI) v. Columbia Broadcasting System, 120–2, 124–5, 128–9, 166, 172, 182, 184, 289 Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., 214, 219 Brown Shoe Co. v. United States, 311, 318–20, 321, 322, 323, 325, 327, 338–41 Brunswick Corp. v. Pueblo Bowl-OMat, 63 Buffalo Courier-Express, Inc. v. Buffalo Evening News, Inc., 217n64 Bureau of Competition, 48, 50n10 Bureau of Consumer Protection, 48 Bureau of Economics, 50n10 Burger, Justice, 119 business. See also corporation anticipation by, 76 capacity constrained, 115 failure of competitor, 320, 327, 339 internal capital market for, 312, 312n7 lack of resources of, 320 larger relationship to smaller, 204, 320 material reserves for, 327–8 protection of small, 40–2, 192–3, 320–1, 323, 339 right/refusal to deal in, 246–8, 246n7, 263–4, 270 risk taking compensation for, 231–2, 269–70, 293, 295 self-dealing in, 337 size, 188, 190, 193, 204 vertically integrated, 203 Business Electronics v. Sharp Electronics, 275–7

buyers lying by, 72 California Motor Transport Co. v. Trucking Unlimited, 356–7, 363, 368–71 California Retail Liquor Dealers Assn v. Midcal Aluminum, 372, 374 California v. American Stores Co., 55n24 California. v. ARC, 63n40 cartels, 18n13, 99n17. See also boycotts; collusion basic theory of, 68–71 collusion and, 69–73 conscious parallelism for, 73–89 conspiracy doctrine, conscious parallelism and, 75, 87, 89 cooperative elements of, 122 enforcement, 70–1, 87–8, 136, 146, 337 inferring conspiracy for, 87–9 influence government for regulated price by, 367 informal exchange relationships in, 71 instability of, 68–9, 76n14, 159, 160, 216 interdependence theory and, 76–7, 81–2, 86, 87 intraenterprise conspiracy and, 74, 78–80 lying by buyers of, 72 market share of, 72 monitoring of, 71–3, 86, 87–8, 136 output versus price agreements with, 72 price-fixing, 17–18, 62, 68–9, 80n27, 87, 107–8, 253, 256, 338 sales, 107–8

Index stable, 71n3, 73 casebook, drawback of, xi CBS, 120–1, 128 Cellophane. See United States v. E.I. du Pont De Nemours & Co. (Cellophane) Cement Institute. See Federal Trade Commission v. Cement Institute Cement Manufacturers Protective Assn. v. United States, 167–9, 171–2, 178, 181 Central Shade Roller Co. v. Cushman, 101n20 chain-store paradox, 225–6, 226n76 Chamberlain, Edward, 22, 75n12, 76 Chicago Board of Trade v. United States, 104–7, 108, 109, 109n37, 111–12, 116, 117, 119, 126–8, 129, 153, 172, 174, 183, 184, 267n41, 365n24 Chicago School, 280–1, 308 City of Columbia v. Omni Outdoor Advertising, 367, 369, 377 Clayton Act, 30, 39–40 amendments to, 318, 335, 338 against attempt of monopolization, 287, 304, 307 commodities affected by, 290, 290n29, 293, 294 consumer protection with, 320, 349–50 exclusive dealing and, 303, 304 fine/punishment of, 47, 48–9, 52 horizontal mergers and, 318, 320 injunctions under, 55, 55n24, 56 injury requirement of, 218 private actions, treble damages and, 58, 60 protection of small businesses with, 320

383

rule of reason and, 305, 317 Section 7, competition substantially reduced and, 56–7, 318, 319–20, 322, 323, 325, 330, 335, 338, 339, 347, 350 Sherman Act and, 47, 54 stock transfers and, 318, 335, 335n9, 336 tying and, 285–6, 290, 293, 294 vertical mergers and, 335–6 Cline v. Frink Dairy Co, 31 Clorox. See Federal Trade Commission v. Proctor & Gamble Co. (Clorox), 344–8 Coase, Ronald, 8, 242–3, 333, 341–2 Coke, Edward, 36 collusion, 68–9, 222 avoidance of per-se rule in, 117 brand loyalty and, 73 cartels and, 69–73 cheating on agreement of, 145, 153–4 circumstantial evidence for, 73, 133 concentration and, 156, 312 enforcement of, 70–1, 87–8 evidence and, 135 explicit versus tacit, 145 lying by buyers in, 72 monitoring of, 72–3, 153 one-shot versus repeat-player, 70–1, 96 retail price maintenance, 253 stable versus instability of, 69–70, 71n3, 135 tacit, 73–4, 75, 76–7, 85, 86–8, 88n39, 140, 145, 149, 153, 219 and trigger strategy, 77 tying facilitating of, 283 vertical mergers facilitation of, 337–8, 337n12, 340

384

Index

Combination Laws, 35, 53 Commerce Clause, 39 common law adjudication, xiv background for Sherman Act, 30–7 toward certainty, 105–6 criminal conspiracy in, 31, 35, 35n29, 37 economic reasonableness and, xiv, 92–3, 98–100, 102–3, 107, 109, 109n37, 112 market interference and, 31, 32–3 of monopoly, 31, 35–7 per-se illegality and, 80n27, 90, 104, 107, 112 predatory pricing and, 213, 227 restraint of trade and, 31, 33–4, 37, 91–4, 100, 100n20, 104, 318, 319–20, 322 ruinous competition and, 94–8, 99 rule of reason of, 116 Sherman Act abandonment of, 93–4, 103–4 Sherman Act case against standard of, 101–4 Sherman Act (Section 1) v., 90– 104 Sherman Act’s validity crisis and, 98–101, 326 theory, xiii competition, 55. See also exclusion absence of, 125–6, 179 actual potential, 345 anti, 79, 158–9, 162, 165, 169, 175, 176, 178, 191, 193, 253, 279–83, 307–10, 335, 368 basing point pricing and, 162–5 boycotts and, 170–5 concentration influence on, 319–25, 327

consumer welfare v., 172–3 cream-skimming, 227n79 destroying of, 55, 196, 203, 207, 215, 229 dominant strategy and, 69, 69n2, 96, 135, 160, 221, 306 essential facility, cost-reducing facility and, 208, 210 fair access for, 180 foreclosure on, 286–7, 303, 304, 305–7, 311–12, 319, 325–6, 327, 335, 337–40 foreign, 237 fringe, 241 geography limitations for, 347n22 intrabrand, 120, 178, 178n18, 180, 266, 267, 288 local versus national, 322–3 market interference statutes to suppress, 32–3 maximizing of, 350 meeting, 289 mergers (nonhorizontal) and nonreduction of, 333 mergers and reduction of, 56–7, 57n29, 95, 311–12, 319, 325–6, 327, 335, 338 monitoring of competitor and, 212 monopoly as absence of, 1 monopoly up to levels of, 14–15 over competitors, 173 potential competition theory and, 344–9 predatory pricing versus pricing for, 214–18 price restraints/increase and, 232– 4 price-cutting, 95, 115, 150, 153, 212–13, 242, 281

Index price fixing to reduce intrabrand then enhance interbrand, 118–19, 122 product quality not grounds for restraint on, 125–7, 184 product extension mergers noneffect on, 344–5 with professional services, 29n7 professionalism, no competitive bidding and, 125–6 regulating versus restraint of, 104–5, 112, 119 restraint of trade against, 104–5, 111, 116, 118–19, 127–8, 153, 169, 170, 171, 173, 178, 197, 290, 318, 319–20, 322, 325, 366 retaliation against, 347 risk taking compensation and, 231–2, 269–70, 293, 295 ruinous, 94–8, 99 Sherman Act and, 32–3, 116 small business protection and, 40–2, 320, 323, 339 survival of the fittest and, 192 tying confuses consumer and, 309–10 undercutting costs and, 212–13, 215 unstable, 115 competition, perfect atomism and, 4, 8 conscious parallelism and, 81 consumer and, 334 definition of, 4–8 homogeneous products for, 8 infinitely elastic demand curve in, 10, 21 long run competitive equilibrium, zero economic profit and, 9–12 market behavior not in, 26

385

mobility for, 6–7 no third-party effects on, 7–8 perfect information for, 5–6, 24 profit maximization in, 10–11, 37 specialization, cost curve and, 11 U-shaped long run average cost curve in, 11–12 complementary goods, and tying, 280 complexity, as factor in inferring agreement, 134, 135, 136 concentration and collusion, 156, 312 competition influenced by, 319–25, 327 horizontal mergers and, 311–12, 319–25 insignificant, 327 measurements/regimes of, 328–30, 341, 350–1 percentage of, 322, 324, 327, 329, 330 profit improvement from, 312, 312n3 in relevant market, 320–2 vertical mergers and, 339–40 conglomerate mergers acquiring existing firms in, 346, 350 Clayton Act, competition substantially reduced and, 347 competition, acquisition of competitor and, 350 consumer benefit and, 345–6, 349–50 definition of, 344 efficiency defense and, 347 enforcement guidelines for, 350–1 enforcement, reduction of competition and, 346 entry of, 345, 346, 347–8, 349, 350

386

Index

conglomerate mergers (cont.) leverage with, 347 per-se rule not for, 349 potential competition critique and, 349–50 potential competition expansion for, 348–9 potential competition limits for, 345–6 potential competition theory structure for, 344–5 potential competition’s entry for, 345–8 for predatory pricing, 347 product extension mergers of, 344–5 Connally v. General Construction Co., 31 conscious parallelism anticompetitive conduct inferred in, 79 collusion, concentration in mergers and, 312 conspiracy doctrine strained for, 75, 80n27, 81, 140–1, 165 conspiracy statute for, 77–8 Cournot and Bertrand models on, 83–5 inferring conspiracy and, 87–9, 165 interdependence theory and, 76–7, 81–2, 86, 87 intraenterprise conspiracy and, 74, 78–80 introduction/definition of, 73–5 nonexplicit agreement and, 140, 141, 145 oligopoly power, 82, 86 parallel behavior, independence and, 81 perfect competition and, 81

Posner on, 85–9, 165 Rahl on, 74–80, 85, 87–9 real conspiracy and, 81 Turner on, 80–3, 85, 87–9, 165 consensus standard making, 361 consigning, 257–8, 268, 269 conspiracy, 74, 273, 329 burden of proof with, 150, 167, 168, 169 common law and criminal, 31, 35, 35n29, 37, 111 complexity of, 134, 135, 136 conduct versus structure for, 79, 89 conscious parallelism and, 75, 77–8, 80n27, 81 current price information and, 146, 147–8, 154 duality for, 28, 28n4, 75, 79 evidence for, 73, 89, 133 government access barred from others by, 357 independence against, 134, 135–8, 140, 142 inference of, 87–9, 142, 145, 146, 156, 157, 157n26, 169 informal agreement of, 139 innocent acts of, 139 interdependence and, 81–2, 87 intraenterprise, 74, 78–80 irrelevance of harm for, 28, 75 market power for, 29, 29n5, 111, 143 Parker and no, 367 (three) proof necessary for, 28, 75, 139, 149 restraint of trade and criminal, 31, 53–4 shared information for, 134–5, 136, 146, 147, 148–50

Index Sherman Act and, 27–9, 31, 35, 35n29, 37, 75, 78, 79, 132, 133, 140, 167 states and, 354–5 unilateral conduct for, 29 consumer(s) antitrust standing of almost, 61 Clayton Act protection of, 320, 349–50 comparison shopping and, 298, 300, 301, 309 cost-reducing facility and, 208 gross loss to, 49 leasing and lock-in of, 197–8, 344 low retail for, 262 marginal and inframarginal, 3 monopolist and, 12 monopoly overcharge suffered by, 61–2 monopoly punishment and awards to, 44–7, 49, 58–9, 58n32 most favored, 159 perfect competition and, 334 perfect information and, 5 potential competition theory and, 345–6, 349–50 predatory pricing and informed, 224 price increase and switching of, 233–4 price-responsiveness of, 238–9 product durability and, 242 responding to tastes of, 20, 41 selling to, 3, 4n2 sophisticated, 151 supply, demand and, 2–3, 2n1 surplus, 3, 4n2 trusts versus, 38 tying and, 280, 282, 283–4, 291, 298, 300, 301, 308, 309–10

387

undercutting of prices effect on, 212–13, 221 vertical merger serving of, 334 wealth, monopoly, society and, 12–13, 20, 44 welfare/protection of, 40–2, 43–4, 51, 55, 171, 172–3, 175–6, 192–3, 200, 208–9, 211, 257, 268, 271–2, 311, 317, 330–2, 345–6, 349–50, 376 wholesalers direct to, 133, 167 Continental T.V. v. GTE Sylvania, 120, 128, 129, 166, 172, 184, 185, 264–7, 274–8, 288–9, 298, 300 contracts competition-limiting, 264 exclusivity, 247 illegal, 100 long-term, 234, 343 opportunism and long-term, 23–4 rejection of unilateral, 140–3 rule of reason for disputes of, 317 transaction, 342–3 tying and exclusive dealing, 30 unilateral, 87, 138–40, 158 wheeling, 211–12, 280, 281, 282 convictions, false, 130–1, 171 attempt at monopolization and, 251 bright line rules and, 220–1 competition, monopoly and, 189, 191–2 predatory practices and, 213–15, 217, 218, 219–20 cooperative, 181–2, 204 Copperweld Corp. v. Independence Tube Corp., 28n4, 79–80, 265–6 copyright, 20, 120, 121, 125, 172, 261, 289 core theory, 115–16

388

Index

corporation, 94 subsidiary relationship to, 79–80 cost(s) average, 16–17, 21, 46, 94, 97, 150, 214, 217, 220–1, 222, 292 declining, 215, 292 defraying operating, 217n65 of error, xv high-fixed, 16, 94–5, 115, 292 marginal, 16–17, 22, 46, 94–5, 163–4, 214, 217, 220–1, 235, 322 monopoly rents into, 14 opportunity, 18 pricing and, 94–6 products related to, 11, 44 switching, 300 transaction, 22–4, 282, 334, 341–4 undercutting (below), 212–13, 216, 217, 219, 220–1 cost curve marginal versus average, 16–17, 21, 46, 220–1 U-shaped long run average, 11–12 cost-reducing facility, 180, 207, 208, 210 cost-reducing merger, 315, 330 Cournot, Augustin, 22, 76, 83–5, 174, 236 court(s) administrative concerns of, xiii–xiv appeals, 48 derivative immunity narrower in, 359 economic reasonableness not assessed by, xiii–xiv errors of, 130–1, 171, 189, 191–2, 213–15, 217, 219, 251, 375–7 legislative process, immunity and, 357–69 opinion, xi

per-se standards of, xiv, xv, 80n27 price regulation by, 82 summary judgment of, 141–2, 141n26, 173, 177, 219, 274, 306, 348 “validity crisis” of, xiv, 98–9 Craswell, Richard, 282 credit, 296, 296n44 Criminal Fines Improvement Act, 49 criminal violations imprisonment and, 49, 49n7, 51–3 mental state, specific intent and, 53, 54–5 previous, 54 restraint of trade and, 31, 35, 35n29, 37, 53, 111–12 cross-hauling, 162 customer, most favored, 159 damages for actual injury, 61–2 decoupling proposal of, 60 optimal, 61n38 treble, 49, 58–60, 64–7 damages. See also consumers; fines dangerous probability of success, 244–5, 246, 248–51, 250n18, 263 Darcy v. Allen (The Case of Monopolies), 36–7 data dissemination, 144–54 association, 146–9, 177–80, 204 average cost, 150 and inflation, 151 interference of, 152–3 market price found from, 151–2 monitoring of, 153–4, 170 per-se rule and, 144–5, 148–9, 152–3, 154 rule of reason and, 144–5, 148–9 deadweight loss, 12–13, 15, 18–19, 333

Index efficiency with gain greater than, 315, 330–2 optimal fine, monopoly transfer and, 44–7, 46n5 welfare gain exceeds, 316, 330–2 deal duty to, 270–1 right/refusal to, 187, 246–8, 246n7, 263–4, 270, 271, 272 dealer agency versus independent, 268–9 dealership agreement and, 256, 260, 264 free-riding prevention by pricecutting, 258–9 local demand and, 257 long-term arrangements or vertical integration of, 270–1 termination of, 270–1, 272–3, 275–6 demand, 5, 15, 24, 46, 163, 189 cross-elasticity of, 238–9 elasticity of, 10, 20–1, 315 point elasticity of, 238, 316 supply and, 2, 2n1, 7–8 demand-side substitution, 236, 248, 300, 338 Demsetz, Harold, 16 derivative immunity, 354–5, 358 judicial process for, 359 legislature and, 359 limits to, 362–3 quasi-legislative (private association) not for, 362 sham exception to, 355, 356–7, 360 valid governmental action of, 356 dissipation, ex ante rent, 14 dominant strategy, 69, 69n2, 96, 135, 160, 196, 221, 306 Douglas, Justice, 180

389

Dr. Miles Medical Co. v. John D. Park & Sons Co., 254–8, 261, 267, 268, 272–8 Du Pont, 232 du Pont (Titanium), 221n69 du Pont (Ethyl) v. FTC, 158–9 duality, for conspiracy, 28, 28n4, 75, 79 duopoly model, 22, 76, 83–4, 84n37 Dyer’s Case, The, 33 Easterbrook, Judge Frank, 216–17, 226, 226n76 Eastern Railroad Presidents Conference v. Noerr Motor Freight, Inc., 353–66, 353n7, 368–9, 375–7 Eastern States Retail Lumber Dealers’ Assn v. United States, 133, 166–8, 178, 246 Eastman Kodak v. Image Technicolor Services, 299, 309–10 economic profit definition of, 9 entry and exit with, 9–10 zero, 9–10 economics, xi, xii conscious parallelism and, 21–2 definitions for, 1–9 econometrics of monopoly transfer and, 46–7 information, 24–6 and its limits as useful theory of antitrust law, 189, 214, 277–8 perfect competition vs. monopoly and, 9–21 reasonableness concept for, xiii–xiv transaction cost, 22–4 economies of scale mergers for, 312, 320–1, 336 tying for, 287, 288–90, 291

390

Index

Eddy, Arthur Jerome, 146 efficiency boycotts and defenses of, 176 conglomerate mergers and, 347 deadweight loss gain and gain of, 315, 330–2 horizontal mergers using defenses of, 315–17, 341, 344 from information, 151 of long-run competitive equilibrium, 12 monopolization and, 190, 192, 193, 194, 196–7, 202, 204–5, 212, 245 tying and, 283–4, 287, 288–9, 292, 294–5 vertical mergers and, 334, 340, 341, 344 elasticity demand and cross, 238–9 demand and point, 238, 316 (infinite) demand of, 10, 20–1, 315 fringe supply, 236 market demand, 235–6, 238–9 employees, antitrust standing of almost, 61 employment antitrust standing and loss of, 61–2 protection from competition and small business, 40, 40n43 enforcement cartel, 70–1, 87–8, 136, 146, 337 inference of agreement and evidence of, 86–8, 147 enforcement (agencies) agency enforcers of, 47–9 apprehension probability, fines and, 50 civil prosecution from, 48–9, 52 conglomerate mergers and, 350–1 early, 18n13

equity proceedings for, 55–7 fairness of, 52–5 fine/penalties (as optimal) for, 44–7, 49–52, 64 horizontal mergers guidelines for, 328–30 imprisonment, criminal violations and, 49, 49n7, 51–3 information sharing and, 148 optimal theory of, 43–7 per-se standards of, xiv, xv private antitrust suits for, 57–64 reasonableness of defendant’s conduct, punishment and, 50–3 reasonableness of test and burden to, xiii–xiv, 311, 317, 341 reduction of, 173 social welfare and, 172–3, 175–6, 311, 376 specific intent and, 53, 54–5 specificity desirable for, 52, 54 treble damages for, 49, 58–60, 64–7 underdeterrence vs. overdeterrence of, 43, 51, 59–60, 130 engrossing, 32, 32n18, 32n20 entry advertising as barrier to, 347 artificial barrier to, 15–16, 189, 190, 191, 193, 196–201, 198, 281, 303, 309, 329, 335, 336, 340–1, 347, 349, 350, 352 conglomerate mergers and, 345, 346, 347–8, 349, 350 easy, 22, 36, 103, 224, 236, 267n41, 329, 348 exclusion of, 57, 303 exit and, 9–10 geographically remote rivals constraint on, 316–17

Index government (state), 352–3 independent (de novo), 345, 346, 348, 349, 350 innovation-profit, 19–20 leasing and, 198 limit pricing and, 221–2 mergers and, 329, 335, 336, 340–1 monopoly, 15–16, 190, 191, 216 natural barrier to, 16 penalty provision in leaving and, 200–1, 201n35 potential, 348 potential competition theory and alternatives to, 345, 346, 347–8 price-fixing and new, 97 profits and, 12, 37 supply-side substitution, 234, 338 toehold, 349, 350 two level hurdle to, 199–200, 303, 341 tying and, 281, 303, 309 United Shoe and barriers to, 196–201 equilibrium competitive, 9, 21, 83–5, 150 Cournot, 83–5 joint-profit maximizing, 83, 135 long run, 9–12 market, 2–3, 5 price, 3, 10, 46n5, 77, 174 Stackelberg leader outcome, 83–5 unilateral/structurally determined outcome for, 85 equity proceedings, 55–7 essential facility cases, 57, 202–6 consumer welfare and, 208–9, 211 cost-reducing facility for, 207, 208, 210

391

demand-increasing facility for, 207, 208, 210 discouraging development of, 209–10 doctrine, 207–11 economic rent for, 209 rival-cost-increasing facility for, 210 types of, 207–8 ethical concerns price-fixing and, 123–4 evidence, 89, 134, 135, 141, 155–7, 267n41. See also intent analysis of, 155–8, 169 circumstantial, 73, 133, 139, 141–2, 146 clear (objective), 214, 215, 245, 250, 274, 300 competitive (culture), 322, 324 historical, 239, 293, 294 inconclusive, 158, 275 inconsistency with, 204 of specific intent, 192, 247 of unfair tactics, 247 exclusion, 57, 78, 173–4, 177, 180, 181–2, 195, 196, 214–15, 247 exclusive dealing Clayton Act application to, 303 entry barrier with, 303 law for, 303–7 market power and, 305 per-se rule for, 303 rule of reason for, 305–6 tying v., 304–5, 304n60 vertical mergers similar to, 333, 340 exclusive dealing. See also tying and exclusive dealing exemptions, 91–2, 92n4. See also immunity exit, entry and, 9–10

392

Index

facilitating mechanisms basing point pricing, economics and, 160–4 basing point pricing, related practices and, 154–60 data dissemination and, 144–54 per-se rule and, 144–5, 148–9, 152–3 retail price maintenance as, 253 rule of reason and, 144–5, 148–9 fair-minded compared to specific intent compared to reasonable, 52–3 Fashion Originators’ Guild of America v. FTC (FOGA), 170–2, 181, 182, 183, 184 Faulkner Advertising Assoc., Inc. v. Nissan Motor Corp., 295n42 FCC, 206 Federal Trade Commission (FTC) Act of, 39–40 Bureaus of, 48, 50n10 business review (merger) letter of, 57 limit pricing and, 221n69 preliminary injunction of, 55 prosecuting violations of Clayton/FTC Act by, 39–40, 47, 48–9, 155, 158, 170–1, 183, 306–7, 320, 364 Federal Trade Commission v. Brown Shoe Co., 306–7, 335 Federal Trade Commission v. Cement Institute, 155–8 Federal Trade Commission v. Indiana Federation of Dentists, 182–4, 267 Federal Trade Commission v. Proctor & Gamble Co. (Clorox), 344–8, 351

Federal Trade Commission v. Ticor Title Insurance Company, 372–4, 377 federalism, 353, 367, 368, 373–4 benefits to, 371–2 fine(s) Criminal Fines Improvement Act and, 49 maximum v. optimal, 49–50 optimal, 44–7, 58, 58n32, 64–7 private damages v. optimal, 58–9, 58n32 Sherman Act, 27n1, 43n1 state’s cost of, 47 fine(s). See also punishment First Amendment, 355, 360, 364, 365, 365n24, 377 Fisher, Frank, xvi f.o.b. prices inference of conspiracy and refusal for, 156, 161–2 FOGA. See Fashion Originators’ Guild of America v. FTC (FOGA), 170–2, 181 foreclosure, 23–4 degree of, 338–40 insignificant, 340 lock ups for, 335 mergers and, 311–12, 319, 325–6, 327, 335, 337–40 tying and, 286–7, 303, 304, 305–7 forestalling, 32, 32n18, 32n20 Fortas, Justice, 153 Fortner II. See United States Steel Corp. v. Fortner Enterprises (II) Fowle v. Park, 254 franchise, free-riding, 284n11, 288 free trade, 4, 41–2 freedom, individual, 168, 178–9, 181 free-riding, 209

Index franchise, 284n11, 288 predatory pricing and, 213, 227 resale price maintenance (RPM) and, 258–9 state, 371 freight absorption, and basing point pricing, 162 FTC. See Federal Trade Commission (FTC) FTC v. Superior Court Trail Lawyers Ass’n, 363–6, 369 Gamco, Inc. v. Providence Fruit & Produce Bldg., 207 game theory, 75 interdependence theory and, 76–7 predatory pricing and, 225–6 Prisoner’s Dilemma and, 68–9, 68n1. See also dominant strategy geographic market, relevant, 320–2, 323, 344 Goldfarb v. Virginia State Bar, 123–4 goodwill defense, and tying, 286, 287, 288, 292–5 government. See also regulation; state derivative immunity and, 354–6, 357, 358, 359 entry barriers of, 352–3 explicit authorization of, 353 immunity to Sherman Act from action of, 354 irrelevance of influencing of, 356 legislative versus judicial processes for action of, 357–9, 376 rate-making bureaus of, 353, 372– 4 unethical conspiracy to bar others access to, 357, 363, 368 Grady, Mark, 114

393

Haddock, David, 162, 164–5 Hand, Learned, 189, 190–3, 194, 195, 218, 223, 239–43 Hanover Shoe rule, 63 Harberger, Arnold, 18 harm anticompetitive, 178 conspiracy doctrine and irrelevance of, 28, 75 no public, 173–4 social, 175–6, 184, 249, 345–6 Hearn v. Griffin, 101n20 Herriman v. Menzies, 101n20 Hirschman-Herfindahl index (HHI), 328–30, 341, 350–1 Holmes, Oliver Wendell, 105, 148, 150, 224–45, 256, 258 horizon problem, and argument for tying, 287 horizontal mergers acquiring above average firms in, 313–14 business reasons for, 311–12, 320 and competitive culture evidence, 322, 324 concentration issue for, 311–12, 319–25, 327 definition of, 311 for diversification, 313 economic reasonableness and, 326 for economy of scale, 312, 320–1 efficiency defenses for, 315–17, 340, 344 entry into new market by, 323 failing company defense for, 320, 327 foreclosure (on competition) by, 311–12, 319, 325–6, 327 (merger) guidelines for, 328–30

394

Index

horizontal mergers (cont.) incipiency doctrine high point and, 325–6 incipiency doctrine, mergers in concentrated markets and, 321–5 incipiency doctrine rejection of, 326–8, 340 law, 317–30 managerial goals for, 313 mitigating factors for, 320 monopoly power and, 312, 316 reasons for, 311–17 rule of reason, incipiency doctrine and, 318–21 sales/profits following, 314–15 small market share, violation of Clayton and, 319–20 for smaller companies (to compete with larger ones), 320, 323 social welfare versus, 311, 317 welfare tradeoffs of, 315–17, 330–2 horizontal restraints, 128, 266–7 hornbook, xi–xii theoretical, xii IBM v. United States, 286, 287, 292 illegality per-se, 80n27, 90, 104, 107, 112, 116, 118, 133, 152, 171, 269, 276, 295, 340 Illinois Brick rule, 63 immunity, (state), 353–4. See also derivative immunity; sham exception as attractive nuisance, 374 derivative, 354–6, 357, 358, 359 error costs and, 375–7 exceptions to Noerr doctrine of, 366–7, 369, 377

exceptions to Parker doctrine of, 366–7 influence on valid governmental action for, 356, 362–3 liberal approach to, 372 municipalities and, 375 from nature of acts, 354 Noerr doctrine of, 354–69 Parker doctrine of, 356, 358, 362, 366, 367–9, 371–5 procedural invalidity as exception to, 366–8, 369, 377 sham exception of, 355, 356–7, 360, 366, 369 source of, 356 valid governmental action for, 356, 362–3, 366, 368 incentives, 10, 20, 118, 142, 181, 198, 240, 273 hold-up litigation and, 56 information and, 5 leasing need of, 197–8, 199 incipiency doctrine, 348 definition of, 319–20 high point of, 325–6 horizontal mergers in concentrated markets and, 321–5 rejection of, 326–8, 340 rule of reason, horizontal mergers and, 318–21 vertical mergers and, 339–40 independence, as factor in inferring agreement, 134, 135–8, 140, 142 independent goods, and tying, 280 independent versus unilateral conduct, 136–7 individual arrangements, 121, 128 inference doctrine, 133–40

Index inflation, and data dissemination plans, 151 information, 21. See also data dissemination current prices and, 146, 147, 154 economics, 24–6 efficiency from, 151 imperfect, 24, 300 inferring agreement and shared, 133, 134–5, 136, 146–7, 148–50, 168, 170, 178, 184 interference with market of shared, 152–3 market, 149–51, 298 network from, 234 perfect, 5–6, 24 privately held, 5n4 as public good, 5 resale price maintenance and charge for, 260 theory, 309 tying and lack of, 282, 298 injunction, preliminary, 55–6 injury actual, 61–2 antitrust, 218–19 derived, 62–3 passing on problem and, 63 speculative components of, 61 innovation, 41, 114 profits from, 19–20 spill over benefits from, 20 insurance, 123–5, 123n19, 123n20, 372–3 adverse selection in, 5–6 intent bright line versus Ninth Circuit predation test of, 215 clear evidence versus, 214–15 nonspecific, 195, 204–5

395

specific, 53, 54–5, 187, 192, 193, 194–5, 205, 207, 228–9, 245, 246, 247–51, 248, 248n15, 302, 366 specific versus general, 193n18 subjective evidence of, 371 subjective versus objective evidence of, 192, 215, 219 as substitute to evidence of market power, 250 interdependence theory, 22, 76–7, 81–2, 86, 87 interest groups, 38, 38n37, 39 interlocking directorates, 30 International Business Machines Corp v. United States, 26n28 International Salt Co. v. United States, 286–92, 303–4 International Shoe Co. v. FTC, 318n24 Interstate Circuit v. United States, 77, 133–8, 137n10, 141, 158 interstate commerce, 29, 91n1, 290 Interstate Commerce Commission (ICC), 91–2, 97–8 intrabrand competition, 120, 178, 178n18, 180, 266, 267, 288 invalidity, per se, 103 Jefferson Parish Hospital No. 2 v. Hyde, 297–301, 309–10 joint ventures, 119 judges, preliminary injunctions and, 56 jurisdiction, 29–30 Justice Department Antitrust Division enforcement of, 47–8, 49n7, 97–8 Conglomerate Merger Guidelines, 350–1 guidelines of Antitrust Division of, 48, 53, 239, 243

396

Index

Justice Department (cont.) Merger Guidelines, 232, 233, 234, 239, 324–5, 328, 337n12, 338, 340–1 Kaplow, Louis, xii, 248n15 Kiefer-Stewart Co. v. Joseph E. Seagram & Sons, 117, 123n17 King v. Journey-men Taylors of Cambridge, 35n29 kinked demand curve, and basing point pricing, 163 Klor’s paradox, 30, 174–7 Klor’s v. Broadway-Hale Stores, 173, 183, 184 Kozinski, Judge, 194 Lafayette v. Louisiana Power & Light Co., 375 Landes, William, 44, 235–6 lawsuit(s) actual injury and, 61–2 and attorneys’ fees, 58 bad faith, 59 Clayton Act reason for, 30, 48–9, 52 governmental, 55–7, 102–3 hold-up litigation, 56 by Justice Department and FTC, 47, 48–9, 52 optimal fines and, 46 private, 57–64, 102–3 sham, 271, 369–71 standing doctrine for, 50–64 treble damages of, 49, 58–60, 64–7 Leasco. See Response of Carolina, Inc. v. Leasco Response, Inc. leasing, 197–9, 343 barrier to entry for, 198 incentive problems for, 197–8, 199

quality of product versus source for, 286–7 return fees for, 199 short-term versus long-term interest (horizon problem) in, 287 tying and, 285–7, 289–90 legal error, as justification for public harm requirement, 176, 251. See also convictions, false legal evolution, xiii legislative process, immunity and, 357–69, 376 nonconspiracy for, 368 quasi-, 361–2 lemons problem, 24–5, 198n29 Lerner index, 235–7, 236n11, 241 Lessig v. Tidewater Oil Co., 248, 250 Letwin, William, 36 leveraging with conglomerate mergers, 347 debate on, 211–12 definition of, 202 double counting monopoly power of, 196, 211 essential facility and, 202–6 Griffith and theory of, 195–6 limits to, 203 tying with, 279–80, 285, 296 liability erroneous or false, 130, 213–14, 228 essential facility doctrine and, 207 intraenterprise conspiracy and, 79 marginal, 51, 66 monopolization and specific intent, 195–6, 228 per-se (strict), 50, 51 private, 52 licensee, 114, 255

Index licenses, 140–1 blanket, 120, 121 patent, 255–6, 261, 285 per use, 120, 121 trade secret, 254 limit pricing, 221–3 lock in, 197–8, 344 Lorain Journal Co. v. United States, 245–7, 250, 272 loss leader, 217n64 manufacturer(s) agency versus independent dealer and, 268–9 compliance, vertical restraints and, 273 dealer long-term arrangements or vertical integration with, 270–1 retaining of title by, 267–70 termination of dealer by, 270–1, 272–3, 275–6 vertical, 257–8, 268 manufacturer(s). See also resale price maintenance Maple Flooring Manufactures Assn. v. United States, 149–54 Maricopa. See Arizona v. Maricopa County Medical Society market. See also entry concentration, 320–5, 327 contestable, 224 contribution to wealth, 3–4, 4n2 Cournot competition and direct and indirect output effect on, 83 demand, 174 false convictions constraining, 214 inelastic, 153 information, 149–51, 298 interference, 31, 32–3

397

irrelevance of definition of relevant, 236 monopoly of, 102, 322, 324, 329 multi-, 239–43 opportunism offset by concessions in, 23 price, 76, 83, 108, 150, 151–2, 290, 296 quantity, 174–5 relevant, 218, 230–1, 233, 236, 237–43, 248, 249, 250, 292, 306, 325, 336–7, 338n15 relevant geographic, 320–2, 323 relevant product, 324, 338 secondary, 240–3 self-correction of, 352 share, 55, 72, 218, 230–1, 235–6, 243, 249, 318n25, 319–20, 325, 329 small proportion of, 111 unconcentrated, 323 market behavior perfect competition not in, 26 market demand, elasticity, 235–6, 238–9 market equilibrium, 2–3, 5 market power, 129, 179, 182, 187, 188, 207, 215, 230–43, 324. See also monopoly power conspiracy and, 29, 29n5, 111, 143 dangerous probability of success and, 248, 263 equipment versus parts control of, 300, 301 exclusive dealing and, 305 exclusivity (geographic) agreement and, 263 guidelines for, 218, 243, 249, 318n25 in relevant market, 250 rule of reason and, 183

398

Index

market power (cont.) specific intent and, 248, 249 tying and, 279, 282, 282n8, 286, 290– 1, 292, 296, 297–8, 300, 308, 309 marketing, joint agreement for, 203–4, 205 Marshall, Justice, 153–4 Masten, Scott, 200 Matsushita Electric Industrial Co. v. Zenith Radio Corp., 142–3, 216–17, 219, 228, 348 McCarran-Ferguson Act, 92, 373 McLean Trucking v. United States, 91 mercantilism, 4 mergers. See also conglomerate mergers; horizontal mergers; vertical mergers attempt to monopolize and, 245 competition and, 56–7, 57n29, 95 concentration issue for, 311–12, 319–20, 321–5, 327, 339 FTC business review letter for, 57 monopoly from, 30 noncompeting manufacturers for, 245 preliminary injunctions and, 55–6 punishment for, 50, 51, 54 relevant market and, 306, 325, 336–7, 338n15 Missouri v. National Organizational for Women (NOW), 359–60, 364–6 Mitchel v. Reynolds, 33–4, 92, 98, 101, 101n21, 103, 127–8, 264 mobility perfection competition and, 6–7 monopolization. See also attempt, at monopolization; tying abuse theory of, 187–8, 190, 193, 194, 196, 246

active or nonpassive acquisition standard test of, 190–1, 195, 196, 201–2 Alcoa and, 189–93, 201 conduct-focused test of, 187, 188, 188n13, 193 double counting monopoly power and, 21, 196 efficiency, (superiority) and, 190, 192, 193, 194, 196–7, 202, 204–5, 212, 245 essential facility cases, leveraging and, 202–6 essential facility doctrine and, 207–11 exclusionary manner of, 195, 196, 214–15, 247 Griffith test, 201–2, 212 intent of, 187, 188, 192, 193, 194–5, 204–5, 207, 215, 217, 228–9 legal versus illegal, 188, 190–1, 194 leveraging debate for, 196, 211–12 lower price in, 200–1, 242 not unconscious of its doing, 194–5 from patent, 189–90 post-Alcoa and, 194–202 predation theory and, 219–29 predatory pricing cases and, 212–19 preemptive capacity (of plant) investment in, 223 rival-cost-increasing (RCI) facility for, 208, 210 Sherman Act section 2 doctrine development against, 186–202 specific theory of liability of, 195–6 structure-focused test of, 193 subjective versus objective intent of, 192, 215, 219, 229

Index tests for, 190–3, 194 willful acquisition of, 202 monopoly. See also abusive business acquisition process for, 19 basics of, 12–15, 37 bilateral, 23, 342 common law of, 31, 35–7 corporate mergers tending towards, 30 deadweight loss (large/small) of, 12–13, 18–19, 330–2 definition of, 1 entry, 15–16, 190, 191, 216 ex ante rent dissipation, 14, 18 expanding output of, 15 inadequacies of criticism of, 19–21 incumbent, 222–3, 227 of market, 102 markets, 82 markup, 334 natural, 16–17, 249, 250, 250n18 no duty to subsidize rival by, 206 optimal fines, deadweight loss and transfer, 44–7, 49 positive economic profits for, 231 price, 62, 83, 85, 196, 201, 322 price discriminating by, 14–15, 196, 282–3, 289 price setting (fixing) of, 4, 194, 195 profits, 12, 18, 57, 231–2, 271, 281, 308, 309, 352 punishment, consumer and, 44–7, 49 relevant market, Alcoa and multimarket, 239–43 rents, 121, 211, 280 short run, 20 single seller for, 1, 4 size not rule for, 188, 190

399

sources of waste from, 19, 19n16 stability of, 15–18 successive pricing, 334 supplier, 167 surcharge, 235 tending towards, 171 wealth, customer, society and, 12–13, 20, 44 monopoly power, 82, 170–1, 178, 188, 196, 202, 204, 211, 247 attempt at monopolization, market share and, 248–9 Cellophane case and relevant, 232, 237–9 (non)constraints on pricing for, 232–4 cross-elasticity of demand and, 238–9 determinants of, 231, 235–7 fringe supply elasticity (supplyside substitution) and, 236 guidelines for, 218, 243, 249 Lerner index for, 235–7, 236n11, 241 market demand elasticity (demandside substitution) of, 235–6 market share for, 230–1, 235–6 maximizing profit with, 235 measuring of, 230–4 from mergers, 312, 316 multimarket monopoly, Alcoa and relevant, 239–43 price increase and, 233–4, 241 product durability and, 242 profit margins for, 231–2 restraining production with, 240–1 secondary market and, 240–3 tying and, 279, 281, 282, 285, 291–2, 297, 299, 309

400

Index

Monsanto Co. v. Spray-Rite Service Corp., 274–6, 277 Montgomery County Association of Realtors, Inc. v. Realty Photo Master Corp., 295n42 Motion Picture Patents Co. v. Universal Film Manufacturing Co., 285 municipalities, 375 Nash v. United States, 31, 52–3, 78, 88 National Collegiate Athletic Association (NCAA) v. University of Oklahoma, 128–9, 182, 183, 246n6, 267 National Society of Professional Engineers v. United States, 125–8, 127n25, 129, 183–4 NCAAP v. Claibourn Hardware Co., 360 negative option, 373 negligence, 53, 59, 78n21, 105 noncompete agreements, 33–4, 100, 103, 257, 264 Northern Pacific Railway Co. v. United States, 290–4, 296–7, 299, 301, 305 Northern Securities Company v. United States, 317 Northwest Wholesale Stationers v. Pacific Stationary & Printing Co., 181, 182, 183, 187, 207, 264 NOW. See Missouri v. National Organizational for Women (NOW) O’Connor, Justice, 298 oligopoly, 74–5, 136–7, 152–3, 176 bilateral, 160 consciously parallel, 82, 86

industry (market) of, 21, 152–3, 154, 158 power v. prices of, 82, 86, 236 Olympic Equip. Leasing Co. v. Western Union Tel. Co., 206 Ontario Salt Co. v. Merchants Salt Co., 101n20 opportunism, bargaining, 343 Ordover, Janusz, 236–7, 322 organized crime, and cartels, 71 Otter Tail Power Co. v. United States, 202–3, 211–12, 279–81, 282 output, 189, 191, 215 restraining, 240–1 ownership, common, 79 Pacific Engineering & Production Co. v. Kerr-McGee Corp., 217n65 Palmer v. BRG of Georgia, 267 parallel pricing, 75, 77, 82, 322 parallelism. See also conscious parallelism “plus” something, 141, 145 Paramount Famous Lasky Corp. v. United States, 168–9 Parker v. Brown, 353, 354, 356, 358, 362, 366, 367–9, 371–7 partnerships, 94 passing on, 63 patent(s), 20, 36–7, 297 licenses, 255–6, 261, 285 resale price maintenance and, 254–6, 261, 261n17 tying in with, 284–95 Peckham, Justice, 92, 171 penalty provision, 200–1, 201n35 penalty. See fines per-se rule, 80n27 for agency relationship, 269 basing point pricing not, 164

Index blanket license arrangement and, 120–1 boycotts, post-BMI/Sylvania and, 181–2, 184–5 boycotts, post-Socony and, 166, 171, 173, 177, 179–80 boycotts, pre-Socony and, 169– 70 conglomerate mergers not with, 349 data dissemination, oligopolistic industry and, 144–5, 148–9, 152–3, 154 delivered pricing plans not, 158–9, 164 enforcement and, xiv, xv exceptions to price fixing and, 122–3, 126 for exclusive dealing, 303, 304 facilitating mechanisms and, 144–5, 148–9, 152–3 First Amendment exception, price fixing and, 364, 365 for foreclosure on competitors, 286–7 goodwill defense, tying and, 287, 292–5 hybrid of rule of reason and, 166, 181 illegality, 80n27, 90, 104, 107, 112, 116, 118, 133, 152, 171, 269, 271, 295, 340 interbrand competition as not, 120 and its utilitarian justification, 116, 117–25, 131 mergers and, 317, 340 nonburden of inferring conspiracy under, 144–5, 150 for price fixing, 104, 107, 112, 116, 117–19, 122, 133

401

public service exceptions to, 123 reduction of administrative costs by, 122 for refusals to deal, 270, 271, 272 resale price agreements, retail networks and, 269 rule of reason versus, 129–31, 295–301 self regulation, securities industry and, 181, 182 Sherman Act and, 117, 120, 326 standard of proof for (Dr. Miles), 274, 275 termination and, 276, 277 trade restraints and, 317 tying and standard of, 286–90, 297, 302 tying, International Salt and, 286–90 tying, Jerrold, goodwill defense and, 287, 292–5 tying, rule of reason and, 295–301 unreasonableness, 294 per use license, 121 Peterman, John, 289 plaintiff informed, 53–4, 66–7 uninformed, 65–6 PNB. See United States v. Philadelphia National Bank (PNB) Posner, Judge Richard A., 74, 85–7, 151–2, 165, 198n30, 206, 235–6 potential competition theory actual potential competition and, 345 consumer benefit and, 345–6, 349–50 duty-to-rescue rule for, 346 enforcement, reduction of competition and, 346

402

Index

potential competition theory (cont.) entry alternatives and, 345, 346, 347–8 expansion of, 348–9 fiercer competition with, 345 limitations of, 345–6 potential competition benefit and, 345 power bargaining, 342–3 market, 129, 179, 182, 187, 188, 207, 215, 230–43, 263, 279, 282, 282n8, 286, 290–1, 292, 296, 297–8, 300, 305, 308, 309, 324 measuring, 243 monopoly, 82, 170–1, 178, 188, 196, 202, 204, 211, 218, 230–43, 247, 279, 281, 282, 291–2, 295, 297, 299, 309 oligopoly, 82, 86 tying and sufficient economic, 292, 296–7, 300–1 predatory pricing, 70–1, 142, 155, 212, 228–9, 250n18 alternative theory of, 226–8 Areeda and Turner analysis of, 220–1 bright line rule and, 215, 220–1 bright line versus Ninth Circuit intent in, 215 campaign of, 213, 216, 224–5, 227, 229 cases, 213–19 chain-store paradox and, 225–6, 226n76 clear evidence versus intent of, 214–15, 229 common law and, 213, 227 competitive pricing versus, 214–15, 217–18

conglomerate mergers for, 347 cost-price comparison rule in, 215, 219 costs/pricing examined in, 217 counterstrategies against, 223–4 difficulty for, 224 easy entry and, 224 free-riding of rival in, 213, 227 intent examined in, 217 limit pricing and, 221–3 objective reasonableness test for, 216–18, 219 parasitic versus classical, 227 price cutting as, 214–15, 216, 219, 221, 221n69 recoupment from, 213, 216, 219, 221n68, 224–5, 227, 228–9 reputation’s benefit for, 224–6 theory, 219–29 uncertainty and, 223–4 undercutting costs by, 212–13, 216 price (pricing). See also basing point pricing as average cost under perfect competition, 10 average level, 152 base-point, 81 competitive, 3, 62, 77, 150–1, 191, 201n35, 282, 308 competitive bidding and, 125–7 costs and, 94–6, 220–1 cutting (competition), 95, 115, 150, 153, 157, 157n26, 212–14, 215, 220–1, 242, 274, 276, 281 cutting (termination), 274, 276–7 delivered, 158–9, 161–2 depressed, 108, 110 discrimination, 14, 26, 30, 173, 196, 282–3, 283n10, 289–90, 296 (new) entry and agreed, 897

Index equilibrium, 3, 10, 46n5, 77, 174 exceeding costs, 214, 216, 217 increase, 233–4, 241, 258–9, 259n11, 337 inflation, 151, 337 joint-profit maximizing, 83, 85, 86, 135, 136 leadership model, 81–2, 86, 87 life-cycle, 300 limit, 221–3 maintenance, 82 market, 76, 83, 108, 150, 151–2, 290, 296 monopoly, 62, 83, 85, 196, 201, 322, 334 nonuniformity, past of, 149–50 oligopoly, 82, 86 predatory (below-cost), 70–1, 142, 155, 212–29 (unified) raising of, 139n17 regulation, 82 (federal) regulation of, 97–8 restraint in, 315–16 spot, 167 stability, 146, 147 successive monopoly, 334 taker, 10, 21 uniform, 156–7, 158 wars, 95, 97 price fixing. See also collusion by association, 106, 118–19, 121 boycotts and, 169 cartels, 17–18, 62, 68–9, 80n27, 87, 107–8, 169, 253, 256, 337–8 case for, 115–16 charge without control, 28 common law and, 99–100 definition of, 117 delivered pricing system and, 158–9, 161–2

403 desirable, 113–14 empty core problems for, 115 ethical concerns/justifications for, 123–4 by firms controlling industry, 106–7, 188 First Amendment and, 364 (naked) horizontal, 257 illegality/outlawing of, 22, 68, 106, 107, 112, 190 manufacturer’s retail, 267–9 with maximum fees, 123, 124–5, 262 with minimum fees, 123–4 monopoly, 4, 194, 195 most favored customer and, 159–60 noncompete and, 100 parallel pricing versus, 75, 77, 82, 322 per-se illegality of, 104, 107, 112, 116, 117, 118, 122, 133 per-se rule and, 116, 117–25 per-se rule and exceptions to, 122–3, 126 plan versus, 147–8 price wars and agreement of, 96–7 punishment/fines, 50, 51, 53 to reduce intrabrand then enhance interbrand competition, 118–19, 122 regulated price for, 367 (maximum) resale, 117 resale price maintenance and, 253, 256, 274 retailers, 253 rule of reason, (regulation) in, 104–6, 117 social harm of, 131 term, 118

404

Index

price fixing (cont.) territorial allocation schemes and, 118, 119, 126 trade restrained as, 155 uniformity of prices and, 156–7 vertical maximum, 118, 218, 262, 262n23, 273 vertical nonprice restrictions and, 119 price setting, cooperative association, 120 Prisoner’s Dilemma, 68–9, 68n1, 69n2, 76n14, 96, 135 procedural invalidity, 366–8, 369, 377 procompetitive theories vertical mergers and, 333–4 producer(s) marginal and inframarginal, 3 supply, demand and, 2–3, 2n1 surplus, 3 product(s) costs related to, 11, 44 differentiated, 128 durability, 242 homogeneous, 8 new product criterion and, 125 quality, competitive bidding and, 125–7, 184 service of, 199–200, 258–9, 260–1, 299–301 single tying with, 26, 297–8, 298n49 single unit versus components of, 293–5, 295n42 substitutes for unique, 129 substitution of, 324 uniqueness of, 296–7, 299, 300 production. See output Professional Engineers. See National Society of Professional Engineers v. United States

Professional Real Estate Investors, Inc. v. Columbia Pictures Industries, 371 professionals, 29n7, 123–8, 183 profits accounting, 9 accounting versus economic, 231 competitive, 315 concentration improving of, 312, 312n3 economic, 9–10, 12 entry and, 12, 37 in excess of opportunity costs for, 18 innovation, 19–20 margins, 231–2 maximization, 10–11, 22, 36, 94–5, 99n17, 135, 163, 277, 280–1, 315 mergers and, 314 monopoly, 12, 18, 57, 231–2, 271, 281, 308, 309, 352 short term, 10, 12, 20 promotions loss leader for, 217, 217n64 protection, 129 market mechanism versus societal, 40–2 of small businesses, 40–2, 192–3, 320–1, 323, 339 social welfare, 40–2, 43–4, 51, 55, 171, 172–3, 175–6, 192–3, 200, 208–9, 211, 257, 268, 271–2, 311, 317, 330–2, 345–6, 376 public choice, 37, 376 public good, 5 public harm, 30, 173, 249 public interest, 37–9, 376 publicity third-party technique of, 355, 359–60, 363

Index punishment apprehension likelihood and, 50 costless, 44, 44n4, 50 price fixing, 50, 51, 53 state’s cost for, 47 quality competition restraints not due to, 125–7, 184 low versus high, 24–6, 234, 259 Rahl, James, 22, 74–5, 76, 77–8, 80, 85, 87–9 reasonableness. See also rule of reason of defendant’s conduct, punishment and, 50–1, 52–3 objective, 216–18, 219 proof, without price evaluation, 92–4 reasonableness, economic administrative concerns v., xiii–xv, 92, 106–7, 116, 171, 228, 311, 317, 330 common law and, xiv, 92–3, 98–100, 102–3, 107, 109, 112 defensible, xiii–xiv, 92–3, 108–9, 120–1 economic theory for, xv–xvi mergers and, 326, 339, 344 noncompetition clauses and, 100, 103 objective, 191 purpose, power, effects of, 178 resale price maintenance and, 257 Sherman Act and, 92–4, 98–9, 106–7, 109n37, 111–12 recoupment, and predatory pricing, 216, 229. See also reasonableness, objective

405

Reed-Bulwinkle Act, 98 regrating, 32, 32n18 regulation conservation, 110 federal, 91, 97–8, 97n12, 353 state, 92, 92n4, 110 rent-seeking, expenditures, 13 reputation, 26 predatory pricing for, 224–6 tying for, 287–8, 292–5 res ipsa loquitur, as approach to inferring conspiracy, 78 resale price maintenance (RPM) agreement, not monopoly and, 270 consigning, vertical manufacturers and, 257–8 dealers, local demand and, 257 dealership agreement and, 256 definition of, 252–3 Dr. Miles and other cases for, 254–7 free-riding prevention by pricecutting dealers and, 258–9 geographical dispersion in, 259 higher price threat and, 258–9 information charge with, 260 justification of, 257–61, 276–7 like cartels, 253, 256 lowest price for, 247, 253 presale service and, 258–9, 260–1 as price-fixing conspiracy, 253 quality and, 259 scope of, 261–2 Response of Carolina, Inc. v. Leasco Response, Inc., 301–2 retailer. See resale price maintenance revenue, marginal, 22, 163–4 risk compensation for taking, 231–2, 269–70, 293, 295 continuum of, 6

406

Index

Robinson-Patman Act, 47 Royal Crown Co. v. The Coca-Cola Company, 55 ruinous competition, 96–8, 99 high fixed costs and, 94–5 rule of reason, 34, 53, 90, 100, 102, 120. See also reasonableness; reasonableness, economic analysis, 150, 169–70, 171, 294 argument against, 102–4, 106, 107, 109 boundary pressures on, 125–9 boycotts and, 166–70, 173–4, 182, 183, 184–5 burden of proof with, 144–5, 150, 167, 168 Clayton Act and, 305, 317 common law, 116 contract disputes for, 317 data dissemination and, 144–5, 148–9 denial of access for essential facility and, 207 for exclusivity agreements, 264, 305–6 exclusivity contracts and, 247 First Amendment and, 364, 365, 365n24 higher standards of profession and, 123 horizontal mergers, incipiency doctrine and, 318–21 horizontal restraints and, 128, 266–7 hybrid of per-se rule and, 166, 181 information and, 148, 149–50, 154 judiciary’s limited experience and, 124 least restrictive alternative not required by, 121

market power and, 183 merits of, 104–6, 109 modern, 116 most favored customer and, 160 per-se rule versus, 129–31, 295–301 for private associations, 262 procompetitive justifications of, 124, 174, 183–4, 203–4, 246, 246n6, 247, 291 professions and, 124, 126–7, 183 quick look application of, 129n30 reverse of, 294 Sherman Act and, 106, 126, 291, 292, 317, 326 for territorial restrictions, 265–6 trade restraints and, 317 tying and, 291, 292, 293, 294, 295, 297, 299, 300, 302n57 tying, per-se rule and, 295–301 for vertical mergers, 333, 340–1, 344 sales cartels, 107–8 pyramiding, 108 referrals, 206 Scalia, Justice, 301, 367, 368 Schumpeter, Joseph, 19–20 Schwinn. See United States v. Arnold, Schwinn & Co. section 2. See Sherman Act, section 2 self-dealing, 337 sellers, 3 Selten, Reinhard, 225 service free provision of, 199–200 parts and, 299–301, 310 resale price maintenance and, 258–9, 260–1 tying of, 299–301

Index Service & Training Inc. v. Data General Corp., 295n42 Sessions Tank Liners Inc. v. Joor Manufacturing Inc., 366n22 sham cases judicial, 359, 361, 369, 370 legislative, 359, 361, 369 sham exception, 355, 356–7 as anticompetitive weapon, 368 business relationships of competitor influenced in, 366, 369 competing against each other in, 360 publicity campaigns and, 360 sham lawsuits, 369–71 anticompetitive, 370 test for, 371 shareholders (stock), 314–15, 318n24 Sherman Act, xiv, 17, 20 boycotts and, 365 Clayton Act and, 47, 54 common law background for, 30–7 competition, market interference and, 32–3, 116, 119, 214 conspiracy aspect of, 27–9, 31, 35, 35n29, 37, 75, 78, 79, 146 debate concerning goals of, 40–1 development of passage of, 37–9 exclusionary conduct and, 78, 174, 214 explicit federal authorization of, 353 federal control without adequate state supervision and, 372 federalism and, 367 fines/enforcement for, 27n1, 43n1, 47, 49, 54, 74, 79, 97–8 immunity (from governmental action) to, 354, 354n9

407

interdependence and, 81 jurisdiction of, 29–30, 55 New Sherman Act and, 132–3, 141 patents and, 261 against price fixing, 22, 68, 218 public choice view for, 37, 376 public interest view for, 37–9, 376 as punishing conduct versus structure, 79, 89 resale price maintenance violation of, 256–7 restraint of trade and, 31, 33–4, 37, 39, 91–4, 100–2, 100n20, 116, 290 rule of reason for, 291, 292, 317, 326 Sherman Act, Section 1 case against common law standard and, 101–4 common law abandoned by, 93–4, 103–4 common law versus, 90–104 conscious parallelism and, 145 conspiracy and, 27–8, 78–9, 132, 133, 140, 167, 170 enforcement/penalties of, 47, 49, 54, 139 exclusivity agreements and, 263 mergers and, 317–18 per-se rule and, 117, 120, 326 price fixing and, 218–19 reasonableness and, 92–4, 98–100, 106–7, 109n37, 111, 317, 330 ruinous competition and, 94–8, 99 rule of reason and, 106, 126, 326 for territorial restrictions, 265, 266 as test for violation of section 2, 187, 187n3 tying and, 293

408

Index

Sherman Act, Section 1 (cont.) unreasonable restraint of trade and, 101–2, 290 validity crisis and its resolution for, 98–101 Sherman Act, Section 2 abuse theory and, 186–8, 190, 194, 196, 247 Alcoa and, 188–93 denial of access of essential facility and, 207 development of, 186–8 exclusivity agreements and, 263 Griffith test for, 201–2, 212 joint marketing, large firm and, 204 monopolization attempt and, 218, 222, 292 post-Alcoa and, 194–202 public harm and, 175–6, 184, 249 Section 1 as test for violation of, 187, 187n3 specific intent for, 187, 192, 193, 194, 195, 197, 228 side payments, 8. See also bribes Silver vs. New York Stock Exchange, 181, 182 Simpson v. Union Oil Co., 268–9 single product, and tying, 293–5, 295n42, 297–8, 298n49 single-supplier, 1 Smith, Adam, 4 Snyder, Edward, 200 social benefits, 129, 171 social harm marginal, 51, 66 price fixing for, 131 social optimum, 3, 5 social welfare, 40–2, 43–4, 51, 55, 171, 172–3, 175–6, 192–3, 200, 208–9,

211, 257, 268, 271–2, 311, 317, 330–2, 334, 345–6, 376 society waste of wealth of, 12–13, 20, 43 Socony. See United States v. SoconyVacuum Oil Co. specialization perfect competition, cost curve and, 11 specificity, and criminal law, 52, 54 Spectrum Sports Inc. v. McQuillan, 250–1, 263 stability, 15–18, 115, 146, 147, 148 Standard Oil Co. of California (Standard Stations) v. United States, 303–5, 317 Standard Oil Co. v. United States, 39, 52, 101–2, 105, 186–8, 192, 194, 195, 212, 246, 305 Standard Stations. See Standard Oil Co. of California (Standard Stations) v. United States standing, law on, 60 actual injury and, 61–2 derived injury and, 62–3 and optimal deterrence, 61–2 proving antitrust injury and, 63–4 state action antitrust immunity, 353–4 attractive nuisance of immunity of, 374 enforcement of antitrust regulations hesitancy by, 39 error costs, immunity and, 375–7 federal control without adequate supervision of, 372 free-riding, 371 immunity, error costs and, 375–7 immunity exception as market participant, 368

Index immunity to Sherman Act from action of, 354, 354n9 municipalities and immunity of, 375 Noerr immunity and, 359–69 Noerr-Pennington doctrine for, 353–9 Parker immunity doctrine for, 354, 356, 358, 362, 366, 367–9, 371–5 punishment and, 47 rate-making bureaus of, 353, 372–4 sham lawsuits and, 369–71 third-party publicity technique against, 355, 359–60, 363 unfair competition laws, 170 winners and losers with, 367 State Oil Company v. Khan, 117–18, 123n17, 262 state. See also immunity Stigler, George, 71–3, 86, 153 Strauss v. Victor Talking Machine Co., 261n19 summary judgment, 141–2, 141n26, 173, 177, 219, 274, 306, 348 Summit Health, Ltd. v. Pinhas, 29, 184 Summit Health v. Pinhas, 177 supply, 12 demand and, 2, 2n1, 7–8 supply-side substitution, 234, 236, 248, 338 surplus consumer, 3 producer, 3 total, 3 Swift & Co. v. United States, 244–5 Swift formula, 244–8 Sykes, Alan, 236–7, 322 syllogism, of Judge Learned Hand in Alcoa, 190, 195

409

Sylvania. See Continental T.V. v. GTE Sylvania Taft, Judge, 99–101, 104, 112, 326 Tampa Electric Co. v. Nashville Coal Co., 305–6, 319 tariffs, 38, 38nn36–7, 234, 237 tax, 313 for overproduction, 8 taxicab medallion, as ex ante rent dissipation, 13 termination, 270–3, 275 in regards to price cutting, 276–7 territorial allocation, 118, 119, 126 territorial restrictions, 263, 267, 289 resale, 265–6 rule of reason for, 265–6 textbooks economics, xi, xii, xiii law, xi–xii, xiii Thatcher Mfg. Co. v. FTC, 318n24 Theatre Enterprises v. Paramount Film Distributing Corp., 132–3, 140 Theory of Industrial Organization, The (Tirole), xii third-party (effects) perfect competition and no, 7–8 publicity by, 355, 359–60, 363 Ticor. See Federal Trade Commission v. Ticor Title Insurance Company Times-Picayune case, 291 Tirole, Jean, xii, 75n11, 76nn13–14, 241 toehold doctrine, 349, 350 tort law, 30, 53, 58–9, 176, 213, 226–7 Town of Hallie v. City of Eau Claire, 375

410

Index

trade competition and restraint of, 104–5, 111, 112, 116, 118–19, 127–8, 153, 169, 170, 171, 173, 177–8, 197, 290, 318, 319–20, 322, 325, 366 free, 4, 41–2 individual freedom of, 168, 178–9, 181 regulating, 170 restraint of, 31, 33–4, 39, 80, 91–4, 100, 100n20, 129, 133, 155, 317 states and restraint of, 354 unreasonable restraint of, 39, 53, 101–2 trade secrets, 254 Trail Lawyers. See FTC v. Superior Court Trail Lawyers Ass’n transaction costs economics, 22–4 models, 334 tying, 282 types of, 342 vertical mergers and, 334, 341–4 transaction-specific assets, 342 transportation. See basing point pricing treble damages, 49, 58–60 optimality of, 64–7 plaintiff informed and, 66–7 plaintiff uninformed and, 65–6 trusts, 38 Tullock, Gordon, 18, 80–3, 85 Turner, Donald F., 74, 80–3, 85, 87–9, 136n9, 165, 220–1 tying, 26 anticompetitive theory, after Chicago school and, 281–3 anticompetitive theory, Chicago school critique and, 279–81, 308

anticompetitive theory difficulty for, 307–10, 307n66 Clayton Act and, 285–6 collusion facilitated through, 283 consumers and, 280, 282, 283–4, 291, 298, 300, 301, 308, 309 contractual versus technological, 302 definition for, 279 early cases for, 284–6 economies of scale defense for, 287, 288–90, 291 efficiency and other constraints with, 283–4, 287, 288–9, 292, 294–5 entry deterrence of, 281, 303, 309 exclusive dealing versus, 304–5, 304n60 forcing of, 298 franchise problem, 284n11, 288 goodwill defense and, 287, 288, 292–5 horizon problem, 287–8 information costs and, 300 leasing and, 285–6, 289–90 leverage for, 279–80, 285, 296 market/monopoly power and, 279, 281, 282, 282n8, 285, 286, 290–2, 296, 297–8, 299, 300, 308, 309 per-se rule, International Salt and, 286–90 per-se rule, Jerrold, goodwill defense and, 287, 292–5 per-se rule standard for, 290–2, 297, 302 price discrimination for, 282–3, 283n10, 289–90, 296 quantitative substantially test of, 290, 296, 297, 304 for reputation, 287–8

Index rule of reason and, 291, 292, 293, 294, 295, 297, 299, 300, 302n57 rule of reason, per-se rule standard and, 295–301 single product, 293–5, 295n42, 297–8, 298n49 specific intent in, 302 switching costs and, 300 “taxing” effect on desired product by, 308–9 technological, 301–2 timing of, 293–4, 301 transaction costs and, 282 uniqueness of product with, 296–7, 299, 300 vertical mergers equivalent to, 339 tying and exclusive dealing, 30, 51, 54. See also exclusive dealing Union Leader Corp. v. Newspapers of New England, 247, 249, 250 unions, 61062 United Mine Workers of America v. Pennington, 353, 353n7 United States Steel Corp. v. Fortner Enterprises (II), 295–7, 299 United States v. Addyston Pipe & Steel Co., 28n3, 31, 99, 102, 104, 112, 119, 264, 326 United States v. Aluminum Co. of America (Alcoa), 189–93, 204–5, 223, 228 current status of doctrine of, 194–5 multimarket monopoly, relevant market and, 239–43 refinement of doctrine of, 195–6 United States v. Aluminum Co of America (Rome Cable), 323–4 United States v. Arnold, Schwinn & Co., 265, 266

411

United States v. Brown University, 109n37, 127n25 United States v. Colgate & Co., 167, 247–8, 263–4, 270–8 United States v. Columbia Steel Co., 318 United States v. Container Corp. of America, 152–4 United States v. Continental Can Co., 324–5 United States v. E.C. Knight Co., 29, 90n1 United States v. E.I. du Pont De Nemours & Co. (Cellophane), 232, 232n6, 237–9, 263, 264, 292, 324 United States v. E.I. du Pont De Nemours & Co. (General Motors), 335–8 United States v. Falstaff Brewing Corp., 349 United States v. First National Pictures, 168–9 United States v. General Dynamic Corp., 311, 326–8, 340 United States v. General Electric Co., 267 United States v. Griffith, 195–6, 201–2, 203, 212, 244, 247 United States v. Jellico Mountain Coal & Coke Co., 90n1 United States v. Jerrold Electronics Corp., 293–5, 307n66 United States. v. Joint Traffic Assn., 94, 98, 264 United States v. Loew’s, 292 United States v. Microsoft, 228, 302 United States v. Milk Drives & Dairy Employees Union, 217n64

412

Index

United States v. Parke, Davis & Co., 273, 275, 277 United States v. Philadelphia National Bank (PNB), 321–4, 327 United States v. Sealy, 118 United States v. Socony-Vacuum Oil Co., 109–10, 112, 121, 143, 144, 152, 166, 179, 244, 353 United States v. Topco Associates, 118–20, 122, 126, 262, 266, 267 United States v. Trans-Missouri Freight Assn., xiv, 40, 90–4, 97–8, 100, 101, 102, 105, 106–7, 112, 171, 191, 317, 326, 366 United States v. Trenton Potteries Co., 106–8, 109, 111–12, 116, 122, 126n23, 171, 293n39, 353 United States v. United Shoe Machinery Corp., 196–201, 215, 285, 287, 294–5 United States v. United States Gypsum Co., 54, 154 United States v. United States Steel Corp., 187–8, 193, 194, 317–18 United States v. Von’s Grocery Co., 325–7 United States v. Winslow, 245, 250 U.S. v. Grinnell Corp., 29n5, 215 U.S. v. Syufy Enterprise, 194 U.S. v. Terminal Railroad, 180, 207, 210, 335 utilitarian case for per-se rule, 116, 117–25, 131, 293n39 vertical mergers anticompetitive theories and, 335 Clayton Act (amended) and, 338 Clayton Act (unamended) and, 335–6

collusion facilitated from, 337–8, 337n12 competition substantially reduced by, 338–9 concentration and, 339 consumers served in, 334 development of, 335–40 economic and historical factors and, 339, 344 for economy of scale, 336 efficiency defense for, 334, 340, 341, 344 entry impairment with, 335, 336, 340–1 exclusive dealing similar to, 333, 340 failing company defense for, 320, 327 foreclosure (of competition) by, 311–12, 335, 337, 338–40 incipiency doctrine and, 339–40 locks up of retail by, 335 nature and purpose of, 339 opportunistic bargaining reduced in integration of, 343–4 per-se rule and, 317, 340 procompetitive theories for, 333–4 rule of reason consideration for, 333, 339–41, 344 for smaller companies (to compete with larger ones), 339 successive monopoly model for, 334 transaction costs models and, 334, 341–4 tying equivalent to, 339 welfare tradeoffs of, 333 vertical restraints agreement, 270–8 Colgate doctrine and, 270–2

Index compliance of manufacturer and, 273 Dr. Miles/Colgate boundary and, 272–8 duty to deal in, 270–1 exclusivity agreements and, 262–4 law, justifications and, 252–61 manufacturer retains title and, 267–71 nonprice, 262–7 resale price maintenance of, 252–62 territorial restrictions of, 263, 264–7 veto, 178, 180, 181 victim, identifiable versus unidentifiable and antitrust standing, 61, 61n38 waste, monopoly, 19, 19n16 wealth determining of antitrust and, 43

413

effects, 41, 41n44 market’s contribution to, 3–4, 4n2 monopoly, consumer, society and, 12–13, 20, 44 welfare. See social welfare welfare tradeoff model, 315–17, 330–2, 333 wheeling contracts, 211–12, 280, 281, 282 Whinston, Michael, 281–2, 308 White, Justice, 39, 52, 94, 101–2, 366 wholesaler, 133, 167, 181, 211, 254, 268, 273 Wickens v. Evans, 101n20 Williamson, OIiver, 23, 315–16 Willig, Robert, 236–7, 322 Wyzanski, Judge, 197–200, 243, 247, 248 Yellow Cab v. U.S., 79