I. Introduction
Thank you for inviting me to address this conference on the application of the antitrust laws to high technology industries.(1) You picked a timely subject. Both the Federal Trade Commission ("Commission") and the Antitrust Division of the Department of Justice have recently brought a number of cases, in both the merger and nonmerger areas, involving firms in high tech markets. Although the Microsoft(2) and Intel(3) cases have engaged most of the public's attention, they are only the most visible examples of recent antitrust enforcement involving high tech products. The FTC has challenged creation of patent pools allegedly used to fix prices, alleged fraud in procuring patents(4), abuse of the standard setting process(5), mergers that would result in an anticompetitive accumulation of power over innovation markets(6), and other anticompetitive practices that would deny to consumers the important benefits of innovation.
We apply the antitrust laws with sensitivity to the special characteristics of high tech industries and of intellectual property, but also with the recognition that -- as in other industries--competition plays an important role in spurring innovation and in spreading the benefits of that innovation to consumers. This is not new. Our approach has roots that go back at least to the 1976 Antitrust Guide to International Operations in the Ford Administration and the 1988 Antitrust Guidelines for International Operations in the last year of the Reagan Administration, and is most recently embodied in the 1995 Antitrust Guidelines for the Licensing of Intellectual Property three years ago. It is also informed by the extensive hearings and a detailed report by the Commission staff two years ago on antitrust in the 21st century.(7)
That is not to say, of course, that enforcement in this area is entirely free from controversy. A few critics question whether the antitrust laws that were originally designed to apply to traditional manufacturing and distribution industries should be applied at all to competition in fast-moving industries where products often are quickly outmoded and market share may be ephemeral. Others express concern about the potential conflict between the antitrust laws and the laws that protect intellectual property. Are monopolies granted under our patent laws fundamentally in conflict with the "antimonopoly" focus of antitrust? Can intellectual property rights coexist with effective antitrust enforcement?
These are some of the issues I would like to discuss with you today.
Let me start with the notion that these are fast-moving industries in which today's technology is quickly outmoded, opening the way for new competitors to overturn the dominance of incumbents. If those generalizations were uniformly true of high tech markets, then surely antitrust enforcement would be less important. Except for price-fixing and other per se violations, antitrust could safely leave such markets alone, for any effort to create or exercise market power would quickly be corrected by market forces.
Of course, we know from experience that this caricature of high tech markets is accurate in some cases and inaccurate in others. For example, even in an innovation-driven market, dominance in one generation may enable a firm to gain exclusive control over critical inputs allowing monopoly power to be carried over from generation to generation regardless of the relative superiority or inferiority of the incumbent's later generation products. In addition, large sunk costs, high risks, and other entry barriers may mean that while product characteristics change rapidly, the identity of the dominant players may be unchanging for long periods of time.
While it is true that rapidly evolving technology may, in many circumstances, erode entrenched interests, there may also be countervailing tendencies that strengthen monopoly power. For instance, the networking effects present in many high tech industries can lead to a winner-take-all market with very limited opportunity for any firm to compete with the dominant network. Regulatory barriers, as in the need for FDA approval of pharmaceuticals, may mean that new entrants arrive only very slowly regardless of the sophistication of the underlying technology. And, of course, patents or other intellectual property may play a role -- not as something for antitrust to condemn, but as a fact of life in a particular market, like economies of scale or large sunk costs, that makes entry unlikely, slow, or insufficient.
What about the apparent conflict between intellectual property rights and our antitrust laws? There may in fact be a lot less conflict or even potential for conflict than the question implies. Remember first that intellectual property is a form of property. Through the years antitrust enforcement has shown respect for the property rights of factory owners in the land on which their factories sit, the bricks and mortar or steel and glass of which those factories are built, the expensive machinery that operates inside those factories, and the products that they produce for sale. But if the only two factory owners, the only two producers, in a particular market said to each other, "Hey, let's not compete; let's contribute our factories to a partnership that will sell the combined output at a single price set through a formula," our respect for their property rights wouldn't stop us for a second from enforcing the antitrust laws to the fullest extent. Our job is to prevent the use of that property in an anticompetitive and unlawful fashion.
Now, translate that scenario into the patent context, and you have the situation the Commission faced in its recent challenge to the patent pool between Summit Technology and VISX, the only two FDA-approved manufacturers of lasers used in photo refractive keratectomy ("PRK") to treat vision disorders. The Commission complaint charged that both companies had the intellectual property and other assets to enter the market as independent competitors, but instead formed a patent pool and used it to fix prices.
Rather than compete on price, the two firms agreed to charge a $250 licensing fee that was paid into the pool each time laser eye surgery was performed using either firm's equipment. The proceeds of the pool were split according to a formula. The result was that prices were far higher than they would have been if the two firms had been competing with each other, as each had originally planned. As the popularity of the procedure has grown, so has the extent of consumer injury, estimated at some $30 million per year in 1997.
There are, no doubt, differences between real property and its intellectual counterpart. For example, with tangible property, it's often fairly straightforward to determine whether the parties stand in a competitive or complementary relationship. If they sell in the same market and to the same customers, and if their customers treat their products as substitutes, that's a pretty good clue that they are competitors. This issue is sometimes less straightforward with intellectual property, because when you scrutinize the complex licensing relationships between the two parties, sometimes their relationship is purely complementary and not at all competitive. For example, if they each have blocking patents, so that neither could lawfully produce a product without a license from the other, the relationship is vertical. Note, by the way, that once again this is not so different from tangible property. If two factory owners are buying physical inputs from each other, closer investigation might reveal that neither could have a final product without the cooperation of the other. It may turn out that such a relationship also is best described as complementary rather than competitive. The basic rule, enunciated in the Intellectual Property Guidelines, is that a relationship is horizontal, i.e., the parties are competitors, if they would have been actual or likely potential competitors absent a licensing relationship between them.
But, as the Summit case shows, it would be a mistake to conclude that antitrust should stay out of the way because the relationship involved intellectual property. In that case, the evidence showed that the companies were true competitors. Each had the intellectual property and other assets to enter the market independently. Thus, their relationship was competitive rather than complementary, and the pool eliminated substantial competition that could otherwise have occurred. Now, it is true that the parties could argue that, notwithstanding the underlying merits, they really were afraid of litigation, and the pool was a way to avoid that litigation. But that argument goes too far. Once it is shown that the patents are not completely blocking and the parties could have competed with each other, there is an anticompetitive effect to be weighed in the rule of reason balance. Concerns about avoiding litigation, however real, are an efficiency to be weighed against that effect. In that weighing, of course, we need to take into account whether the restraint was reasonably tailored to accomplishing the efficiency, or whether instead there were practical and substantially less restrictive alternatives. In this case, a simple cross-license at no or low royalties would have avoided litigation without the substantial harm to competition that resulted from the price floor set by the pool.
To remedy this harm, the Commission accepted for public comment a consent order settling the patent pooling part of the case.(8) As I discuss later, a separate charge of fraudulent patent acquisition against one of the parties -- VISX -- continues in administrative litigation.
The approach the Commission took to the Summit case provides an interesting contrast with intellectual property cases from the early and middle parts of this century. Early in this century, the Supreme Court exempted from antitrust scrutiny a pooling arrangement that amounted to outright price-fixing, with no transfer of technology or other efficiency at all.(9) Later on, the courts swung in the other direction to condemn intellectual property licensing arrangements, as when the Supreme Court struck down a patent pooling agreement on price-fixing grounds, even though the patents apparently were blocking.(10)
Similar problems infected the antitrust treatment of intellectual property outside the patent pooling area. For example, the courts frequently inferred market power from the existence of a patent or copyright, without weighing the significance of substitutes for the patented technology or copyrighted material.(11) And the government's infamous "Nine No-No's" articulated a highly restrictive view of permissible licensing practices.(12)
Fortunately, enforcement of the antitrust laws no longer begins with the assumption that restrictive use of intellectual property is necessarily anticompetitive. Current enforcement instead starts with three basic assumptions about intellectual property: first, intellectual property is comparable to other forms of property, so that ownership provides the same rights and responsibilities; second, the existence of intellectual property does not automatically mean that the owner has market power; and third, the licensing of intellectual property may often be necessary in order for the owner efficiently to combine complementary factors of production, and thus may be procompetitive.(13)
II. The Role of Antitrust Enforcement
This is good news for intellectual property owners, because it clarifies the relationship between antitrust and intellectual property. Does this mean that intellectual property can never be the foundation of a monopoly? Of course not. Nor does it mean that acquisitions of intellectual property, or contracts relating to such property, are necessarily any less (or more) anticompetitive than similar transactions involving tangible property. Indeed, a number of recent FTC actions involving intellectual property help make the point.
a. Anticompetitive Acquisitions
A long line of antitrust cases holds that acquisitions of patents can be used to acquire or maintain monopoly power.(14) In several recent merger cases, the Commission considered the acquisitions of patents and related technology where the merging firms were either the only two, or two of only a few, firms capable of innovating in high tech markets. In such situations, the acquisition would lead to almost certain anticompetitive effects. As Areeda and Hovenkamp note, "the clearest case [of exclusionary conduct] would be the acquisition of an equivalent patent covering the only known economic alternative to the monopolist's product or process. Such an acquisition forecloses potential competition by rivals who might otherwise have access to that patent. Even the acquisition of one out of several equivalent patents might have exclusionary effects."(15)
Attacking anticompetitive acquisitions of intellectual property in high tech industries involves the application of traditional antitrust concepts in an area that we, and others, have recently come to understand as vitally important to a competitive American economy. In industries where the main focus of competition is the development of new technologies rather than price competition, that competitive rivalry must be protected. If too much of the ability to innovate in a relevant market is accumulated in one entity, and substitutes are lacking, competition may suffer.
This approach to the acquisition of intellectual property rights is reflected in the Commission's enforcement decisions involving mergers of pharmaceutical firms. In Glaxo, the Commission alleged harm to innovation markets where the merging parties -- Glaxo and Burroughs Wellcome -- were the two firms farthest along in developing an oral drug to treat migraine attacks. Current migraine drugs were available only in injectable form and were not sufficiently substitutable to be included in the relevant market. Both Glaxo and the acquired firm, Wellcome, competed to develop the new drugs, and the expectation was that the drugs would compete with each other after they were developed. Barriers to entry, based on the necessity of completing the FDA approval process were high. The complaint alleged that after the merger Glaxo could unilaterally reduce output in the relevant market by decreasing the number of research and development efforts to develop an oral drug. It would have the incentive to do so because the remaining research effort would presumably produce a monopoly product until some third firm could complete the FDA approval process many years hence.
The consent order settling this case required the divestiture of Wellcome's worldwide research and development assets for non-injectable drugs. Divestiture as a remedy in innovation markets requires special care because the success of research and development efforts often depends on a complex array of expertise and sustained knowledge. It may be necessary to require on-going obligations beyond divestiture to assure that the purchaser has some probability of successful completion of the research effort. In Glaxo, for example, the order imposed significant obligations on Glaxo to assist the acquirer in its efforts to continue the research and development effort successfully. Glaxo had to provide information, technical assistance, and advice to the acquirer about the R&D efforts, including consultation with and training by Glaxo employees knowledgeable about the project.(16) We count the divestiture as a success in this case since both Glaxo and the acquirer of its intellectual property now have oral migraine drugs on the market. With the required assistance from Glaxo, the acquiring firm, Zeneca, received complete FDA approval in only 15 months.
In Ciba-Geigy/Sandoz, another pharmaceutical merger case, the Commission alleged a market for the development of gene therapy products, despite the fact that there were no such current products approved by the FDA. The complaint noted that the first products would not be available until the year 2000, but that the market could grow to $45 billion by the year 2010. The technology at issue involves the treatment of disease through manipulation of genetic material and insertion or reinsertion of this material into a patient's cells. Although there were many firms doing pioneering research into gene therapies for various disease states, the merging firms were two of only a few entities with the intellectual property rights and other assets necessary for commercialization of such therapies. The firms' combined position in gene therapy research was so dominant that other firms doing research in this area needed to enter into joint ventures or contract with either Ciba-Geigy or Sandoz in order to have any hope of commercializing their own research efforts. Competition between the two firms made possible such ventures or contracts on reasonable terms. Without competition, the combined entity could appropriate much of the value of other firms' research, leading to a substantial decrease in such research. In addition, there was direct competition between the two companies with respect to specific therapeutic products.
The remedy in this case was designed to protect competition both in the particular products in which the two firms competed and also the broader market for gene therapy research and development. For the specific therapeutic products of the two firms, the order requires the licensing of certain key intellectual property rights held by the combined firm, and also requires that an acceptable buyer be identified "up front." Rhone Poulenc Rorer was identified as the licensee before the order was accepted by the Commission. For the broader gene therapy research and development market, the order required the companies to grant gene therapy researchers non-exclusive licenses to certain essential gene therapy technologies that would otherwise have provided a bottleneck to the research of others.
b. Abuse of Standard Setting Process
Standard setting is a collaborative activity found in many high technology industries that has significant procompetitive potential. Where network effects are prevalent, consumers often benefit from widespread adoption of a standard. Efforts to select a single standard therefore can enhance the innovation and efficiency of an entire industry to the benefit of consumers. However, abuse of the standard setting procedure can have anticompetitive effects, as the Commission found in the Dell Computer case.(17)
Dell involved a standard designed for the Video Electronics Standards Association ("VESA") for a local bus to transfer instructions between a computer's CPU and peripherals. There would be considerable efficiency-enhancing potential in a product that would let computer and peripheral manufacturers know how to make products compatible with one another. The agreement on the standard was premised on representations by the participants that no firm held intellectual property rights that might block others from developing towards the standard. The anticompetitive potential of the standard setting activity surfaced when Dell Computer alleged that the new standard infringed on its patent, despite twice certifying, along with other members of the Association, that it had no intellectual property conflicts. Dell made its claim only after the bus was successful. Its claim for royalties, if successful, could have given it effective control of the standard. If Dell had provided information on its patent claim up front, the participants could have made an informed choice on using the Dell technology. Because Dell instead resorted to its patent ambush, its actions were anticompetitive.(18)
Dell's belated assertion of patent ownership in this case enabled it to exercise market power. The Commission's complaint alleged that further industry acceptance of the new standard was hindered and that uncertainty about the acceptance of the design standard raised the cost of implementing the new design. Other firms avoided using the new bus because they were concerned that the patent dispute would reduce its acceptance as an industry standard. In addition, willingness to participate in industry standard setting efforts was chilled.
The consent order requires that Dell refrain from enforcing its patent against any computer manufacturer using the new design in its products. In addition, Dell is prohibited from comparable behavior in its future standard setting involvements.
c. Fraudulent Procurement
Fraudulent procurement of intellectual property almost always causes anticompetitive effects. This is not a new area of the law for the Commission. Over thirty years ago, the Commission successfully challenged a cross-licensing arrangement between American Cyanamid and Pfizer for the sale of tetracycline.(19) The Commission found that the patents had been procured by fraudulent behavior that included suppressing of material information and misrepresentation of material facts, and that the subsequent cross-licensing and sale of the drug constituted illegal monopolization in violation of Section 5 of the FTC Act.
In attacking the fraudulent procurement of patents, of course, we are not attacking the patent system. Just the opposite. Our actions in these cases help protect the system in which patent examiners must rely on truthful representations by inventors seeking patents. Where that standard is not met, and a potential monopoly is wrongfully awarded, antitrust enforcement is appropriate.
The Commission's laser eye surgery case involves such a charge. The complaint charges that VISX obtained one of its key patents through a pattern of fraud and inequitable conduct, and asks that VISX be enjoined from enforcing its patent. More specifically, the complaint alleges that VISX intentionally withheld from the Patent and Trademark Office highly material "prior art" that might prove that the claimed invention was not patentable because it was already known to others in the field. Although Summit and VISX have agreed to abandon the patent pool as part of their settlement with the Commission, the charge of fraudulent patent acquisition against VISX continues in administrative litigation.
d. Maintaining Monopoly Power in a High Technology Industry
Finally, no discussion of monopoly power in a high tech industry would be complete without a mention of the Intel case. Because the matter is in litigation and the trial has not yet commenced, I'm going to limit myself to the allegations of the complaint and to general principles that do not depend on the evidence in that particular case. The complaint alleges that Intel retaliated commercially against customers who had patents that they either sought to enforce against Intel or refused to license royalty-free to Intel. This conduct, according to the complaint, tends to maintain Intel's monopoly power by, among other things, reducing the competitive threat posed by the existence of important technology not under Intel's control or available to it.
The Commission's complaint, filed in June of this year, has been the subject of some considerable debate, some informed, some less so. Any intelligent discussion of the case and the proper role of antitrust enforcement in high technology requires a basic understanding of what the case involves and what it does not. Let me try to clear up two pretty basic mischaracterizations.
First, some have suggested that a case such as the Commission's action against Intel represents an attack on intellectual property rights. Nothing could be further from the truth. Rather, the complaint addresses allegations that Intel has used commercial retaliation and the threat of retaliation to force others to surrender their patent rights and deprive those third parties of a forum to adjudicate those rights. In other words, an important issue presented by the case is whether a monopolist can use its power to deprive others of their intellectual property rights.
Now it is true that one of the means of retaliation alleged in the complaint is Intel withholding of technical information from customers about its products. But it is important to remember, as the complaint states, that the information in question was information Intel ordinarily gave to its customers to design computer systems incorporating Intel microprocessors, not information that could be used to help microprocessor competitors. This involves no technology exchange or patent cross-license. It is the technical "how to" information computer manufacturers need to make a microprocessor work in their products. Without this information, the cut-off customers were at a real competitive disadvantage. Withholding the "how to" was the functional equivalent of withholding the microprocessors themselves. The complaint alleges that the exclusionary behavior ended only when these firms agreed to license their inventions to Intel.
Actions of this sort are particularly threatening and exclusionary because of the importance of patent protection to innovative technologies in their incubation period. In a market where the patent system is the only protection that breakthrough technologies have to protect the fruits of their research from the reaches of the monopolist, such actions send a powerful message that it is not worth the time, effort and expense to innovate in order to compete with the dominant firm. This inevitably tends to reinforce or maintain the dominant firm's monopoly power.
A second, and related, concern some have expressed is that the Commission's action seeks to force compulsory licensing of Intel's patents to its competitors. Even a cursory reading of the Commission's complaint and proposed order shows that suggestion to be seriously misleading. If the Commission prevails, all Intel will be prevented from doing is, in the words of Judge Posner, "retaliat[ing] against customers who have the temerity to compete with him by cutting such customers off."(20) The order would prohibit unjustified discrimination between similarly situated customers and not allow a monopolist to cut off customers solely because they are also competitors. Where, however, Intel had a legitimate business reason -- such as evidence of misuse or misappropriation of its inventions, the company would be free to protect its rights. The case -- and the relief sought -- focus on exclusionary conduct by an alleged monopolist that injures rivals and lacks a legitimate business justification.
Some have argued that the FTC's case cannot succeed unless the evidence shows that the monopolist's action somehow had a further effect on prices or output. That argument seriously misstates the law, which clearly prohibits monopolists from engaging in activities that entrench or preserve their monopoly position. As Justice Scalia had observed, the antitrust laws require that we examine the actions of a monopolist "through a special lens: behavior that might otherwise not be of concern to the antitrust laws ... can take on exclusionary connotations when practiced by a monopolist."(21) By definition, a monopolist charges a monopoly price and produces just enough output to maintain that price. When a monopolist improperly maintains that monopoly position, there is no immediately measurable increase in anticompetitive effects. Prices do not go up and output does not fall. It is competition that would have brought change and dynamism. Thus, in cases like Lorain Journal(22) and Otter Tail,(23) courts have struck down exclusionary conduct even though there was no demonstrable effect on price or output. The Supreme Court noted in Lorain Journal that "section 2 is violated as much by the prevention of competition as by its destruction."(24)
The leading antitrust treatise agrees. Areeda and Hovenkamp argue that the courts concerned about monopoly power do not and should not demand "a clear and genuine chain of causation from exclusionary act to the presence of monopoly."(25) The burden is too high, with the result that monopoly power will wrongfully be allowed to persist. Rather, they argue the monopolist should be faulted when it attempts "conduct other than competition on the merits, or other than restraints reasonably 'necessary' to competition on the merits, that reasonably appears capable of making a significant contribution to creating or maintaining monopoly power."(26)
Having addressed some of the specifics, let me return to the basic policy question. Does it make sense for antitrust to play a significant role in preventing the entrenchment of monopoly power in a high tech industry? My answer is yes. Where a monopolist's conduct is directed at chilling independent and competing innovation in markets where competition is defined by innovation, the argument for antitrust action is stronger, not weaker. It is usually very difficult to undo the effects of a scheme to monopolize after it has occurred. It would be even more difficult to undo a successful attempt to stifle innovation. If the antitrust laws are concerned with ensuring a continued pace of innovation in high technology industries, then it is particularly appropriate that the enforcement agencies be able to stop practices that are of the type that are likely to stifle innovation and thus cement monopoly power before the harm actually occurs.
III. Conclusion
Applying the antitrust laws to high tech industries is important in order to secure for consumers the benefits of the innovation that drives economic growth. Neither the rate of innovation in those industries nor the pervasiveness of intellectual property rights is an argument for antitrust enforcers to withdraw. Instead, these are characteristics to be taken into account, as the special characteristics of every industry and every market are taken into account in antitrust analysis. Through careful, mainstream, but vigorous law enforcement, we can ensure that competition will remain as a spur to innovation and a boon to consumers.
1. These remarks are my own and do not necessarily reflect the views of the Commission or any individual Commissioner.
2. United States v. Microsoft Corp., Civ. Ac. No. 98-1232 (D. D.C. May 18, 1998) (complaint); New York v. Microsoft Corp., Civ. Ac. No. 98-1233 (D. D.C. May 18, 1998) (complaint).
3. Intel Corp., FTC Dkt. No. 9288 (June 8, 1998) (complaint).
4. VISX, Inc. and Summit Technology, Inc., FTC Dkt. 9286 (Aug. 21, 1998) (proposed consent order).
5. Dell Computer Corp., C-3658 (May 20, 1996) (consent order) (Commissioner Azcuenaga dissenting).
6. Glaxo plc, C-3586 (June 14, 1995) (consent order); Hoechst AG, C-3629 (Dec. 5, 1995) (consent order), Novartis AG, C-3725 (April 10, 1997) (consent order).
7. Anticipating the 21st Century: Competition Policy in the New High-Tech, Global Marketplace, A Report by the Federal Trade Commission Staff (May 1996).
8. VISX, Inc. and Summit Technology, Inc., FTC Dkt. 9286 (Aug. 21, 1998) (proposed consent order).
9. E. Bement & Sons v. National Harrow Co., 186 U.S. 70 (1902).
10. United States v. Line Material Co., 333 U.S. 287 (1948).
11. Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100 (1969); Crown Dye & Tool v. Nye Tool & Mach. Works, 261 U.S. 24 (1923).
12. The "Nine No-Nos" were intellectual licensing practices that were sure to attract the attention of the Justice Department in the 1970s.See Bruce B. Wilson, Patent and Know-How License Agreements: Field of Use, Territorial, Price and Quantity Restrictions, Address Before the Fourth New England Antitrust Conference (Nov. 6, 1970).
13. Intellectual Property Guidelines, § 2.
14. See, e.g., Great Lakes Chemical Corp., 103 F.T.C. 467 (1984); United States v. Singer Manufacturing Co., 374 U.S. 174 (1963);Kobe, Inc. v. Dempsey Pump Co., 198 F.2d 416 (10th Cir.), cert. denied, 344 U.S. 837 (1952).
15. P. Areeda and H. Hovenkamp, III Antitrust Law ¶ 707b at 175 (1996).
16. See P. Areeda and H. Hovenkamp, III Antitrust Law ¶ 707i at 184 (1996) (advocating "divestiture of sufficient assets to create viable new forms with free access to the monopolist's then-existing technology . . . . where an acquisition, or a series of acquisitions, has probably made a substantial contribution to monopoly power").
17. Dell Computer Corp., C-3658 (May 20, 1996) (consent order) (Commissioner Azcuenaga dissenting).
18. The Commission's position in this case is entirely consistent with the condemnation found in the patent law for surprise assertions of patents, which can lead to equitable estoppel against the party asserting infringement. See, e.g., Wang Labs. v. Mitsubishi Elecs. America, Inc., 103 F.3d 1571 (Fed. Cir. 1997); Hewlett-Packard Co. v. Pitney Bowes Corp., 1998 U.S. Dist. Lexis 10936 (D. Ore. Mar. 23, 1998).
19. American Cyanamid Co., 72 F.T.C. 623 (1967), aff'd sub nom. Charles Pfizer & Co. v. FTC, 401 F.2d 574 (6th Cir. 1968), cert. denied, 394 U.S. 920 (1969).
20. Olympia Equipment Leasing Co. v. Western Union Telegraph Co., 797 F.2d 370, 377 (7th Cir. 1986).
21. Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451, 488 (1992) (J. Scalia, dissenting).
22. Lorain Journal Co. v. United States, 342 U.S. 143 (1951).
23. Otter Tail Power Co. v. United States, 410 U.S. 366 (1973).
24. Lorain Journal, 342 U.S. at 154, quoting United States v. Griffith, 334 U.S. 100, 107 (1947).
25. P. Areeda and H. Hovenkamp, III Antitrust Law ¶ 651 at 77 (1996).
26. Id. at 78.