Remarks before American Antitrust Institute, Conference: An Agenda for Antitrust in the 21st Century, National Press Club
There is wide agreement that the last decade or so has presented an unusually lively and challenging period for antitrust analysis. Among many reasons we can point to are deregulation and problems of transition to a free market (telecommunications and electricity production offer leading examples), developments in procedural cooperation and possible substantive convergence in response to the increasing globalization of competition and enforcement approaches, and priorities in addressing an unprecedented merger wave.
An additional challenge involves the application of established antitrust principles to the growing high-tech sector of the economy. It is that application of antitrust law to the new economy, and particularly the relationship between antitrust and intellectual property, that I will discuss today.
A. The New Economy Challenge.
Of late, we hear increasing concern that the century old Sherman Act cannot keep up with the more dynamic and fast-moving developments of the 21st Century. A New York Times editorial earlier this year frames the issue:
"Can, and should, laws designed to manage the emergence of industrial and natural resource monopolies in the late 19th and early 20th Centuries be applied to the technology and intellectual property giants of the 21st Century?"(2)
Advocates of a negative answer, according to a similar chain of logic, might argue that the First Amendment, a hundred years older than the Sherman Act, should not influence state regulation of media because the Internet and cable television are so different from distribution of Federalist pamphlets. We don't hear that argument made nearly as often. In both cases, the issue ought to be whether core principles are still valid, not when those principles were adopted. And just as free speech remains central to our democratic political system, competition - - and antitrust enforcement - - are fundamental to continuing economic growth.
B. Core Principles of Antitrust.
Antitrust is concerned primarily with cartels and the acquisition or maintenance of monopoly power by unacceptable means. Most (but not all) of the rest of antitrust is collateral to those two principles.(3) To illustrate:
1. Mergers in general, and the remarkable merger wave of the last eight years in particular, may be problematic when they are likely to lead to concentration that facilitates collusion, or lead directly to dominant market power. (4)
2. Predatory and exclusionary conduct are of most concern when they achieve, maintain or create a dangerous probability of monopoly power.(5)
3. Boycotts are most likely to be challenged when they contribute to maintaining a cartel or to monopoly power.(6)
4. Minimum resale price maintenance is treated aggressively because it may be a facilitating practice for cartel behavior among sellers or a direct cartel arrangement among dealers.(7)
There are exceptions of course to antitrust enforcement's focus on cartel and monopoly concerns. For example, the Robinson-Patman Act, and particularly older cases interpreting that statute, is concerned with the buying power of large chains unfairly diminishing commercial opportunities for small business. Antitrust rules against tie-in sales derive at least in part from an unwillingness to see consumers coerced into purchases they do not want to make.(8) Finally, there are some boycotts and refusals to deal that have anti-competitive effects, and are so lacking in business justification, that they are challenged without regard to a direct effect on cartel behavior or monopoly power.(9)
C. Core Antitrust Principles and the High-Tech Sector.
Assuming the core principles of antitrust are still valid, why should they be abandoned when applied to the high-tech sector of the economy? Three arguments have been advanced.
1. Durability of Market Power
It is often argued that the high-tech sector of the American economy is so dynamic that cartels and monopoly power will be short-lived. They will be defeated more quickly and efficiently by market forces such as new entry than by any band of bureaucrats. Moreover, government regulation is likely to make more, and longer lasting, mistakes than the market.(10)
While barriers to entry may often be lower in high-tech sectors, partly because successful entry so often depends on new ideas, those barriers can nevertheless be substantial. The systems designed to encourage and protect innovation - patents and copyrights - can and often are used to barricade a market against entry by new rivals.
Another barrier to entry that appears to occur more frequently in high-tech sectors involves network effects. Network effects, also known as positive network externalities, arise when the value of a network increases with the number of its users. A single firm, perhaps because it is the first mover, becomes or threatens to become the only supplier of certain products or services, because of the value of compatibility or interoperability. Consumers are more likely to remain with the established network because of their sunk costs (sometimes referred to as "lock-in") and suppliers of complementary products will tailor those products to the established network and resist preparing products for would-be challengers. In that event, network dominance itself becomes a formidable barrier to entry. Also, high-tech industries are no different than others in the sense that "brand-name recognition," and reputation for reliability, can create virtually insurmountable advantages for incumbents. Finally, practices illegal under the antitrust laws such as price discrimination, exclusionary contracts, or intimidation tactics available only to very large firms can themselves impede entry by more efficient challengers.
None of this questions the more basic point that it is easier to become an applications programmer and grow from that base than to design, finance and construct a steel factory. But barriers to entry can and do exist in the high-tech sector, and companies can retain market advantages for decades or even longer. One cannot assume that the market invariably will succeed in dissipating entrenched market power in an acceptable time frame, or that superior products will displace inferior products that enjoy first-mover advantages.
2. Protecting Incentives to Innovate.
Another argument is that little (or perhaps a good deal of) market power is a good thing because it creates prospects for profits that create incentives to innovate which, long term, is more important to consumers than price competition. Because effects on incentives to innovate are hard to measure, government should pursue a cautious or perhaps even a hands-off policy.
The better view is that both antitrust, by protecting competition, and intellectual property, by rewarding innovation, create incentives to introduce new products. In addition, antitrust for the most part has no quarrel with the argument that market power is more acceptable if it is reasonably necessary to achieve efficiencies, including efficiencies connected with innovation. When modest anticompetitive effects of a transaction are significantly outweighed by the positive consumer welfare consequences of innovation, antitrust has historically struck the balance in favor of innovation. A few examples illustrate the point:
i. Research and development joint ventures are a classic example of arrangements that may lessen short-term competition to innovate, and yet may have long-term pro-competitive and pro-consumer effects. Since the passage of the Sherman Act in 1890, there has been exactly one federal government challenge to a research joint venture(11) - an example of caution in interfering with private arrangements to innovate that would be hard to beat.
ii. When competitors control patents that include legitimate conflicting claims, so that each patent holder is blocked from bringing a superior, non-infringing product to the market, the courts consistently have allowed cross-licenses, even when the cross-licenses incorporated agreements on price(12) or where the combination of blocking patents had dominant or even monopoly power.(13)
iii. Occasionally access to information about a monopolist's product is essential for manufacturers of collateral products or services to compete effectively in the market. Courts have fairly consistently held, however, that there is no obligation to pre-disclose, even by a monopolist, because, among other reasons, any such duty would tend to discourage aggressive competition in innovation.(14)
iv. Evidence of an intent to avoid unnecessary interference with incentives to innovate is found throughout the FTC/DOJ Antitrust Guidelines for Collaborations Among Competitors, issued early this year. For example, discussion of research and development collaboration is introduced with the following language:
"Most such agreements are pro-competitive, and they typically are analyzedunder the rule of reason. Through the combination of complementary assets, technology, or know-how, an R&D collaboration may enable participants more quickly or more efficiently to research and developnew or improved goods, services, or production processes."(15)
Undoubtedly, there have been instances where an excessive concern to protect rivalry has inappropriately outweighed regard for innovation - the rather formalistic "Nine No-No's" of the 1970s is probably an example(16) - but the overall trend in antitrust enforcement has been solicitous of innovation.
3. High-tech industries operate in ways that are unprecedented in industrial economies.
A third claim is that antitrust and economics have focused mostly on static price analysis, but dynamic innovation competition is different and will benefit consumers. For example, prices often fall rather than increase with market power in high-tech industries. Non high-tech lawyers and economists will never get it.
It may be that the conventional notion that competition is always a stimulant and monopoly is always a narcotic(17) does not apply in exactly the same way in high-tech industries, and perhaps new models of analysis are called for. It is true that successful high-tech companies are often aggressive in price and innovation, but competition is still important, if only because it is likely that consumers would be better off with two or three aggressive companies, assuming the market can support more than one, rather than a single dominant firm. The history of innovation since the monolithic AT&T was broken up is some evidence that innovation is more likely to thrive in the presence of competition than in its absence.
D. Has the Antitrust - IP Balance Changed?
To summarize to this point, the history of the last 110 years has treated antitrust and intellectual property as complementary regimes, both designed to encourage innovation within appropriate limits. As a matter of policy, we are comfortable rewarding innovation through patents and copyrights so long as the compensation is not significantly in excess of that necessary to encourage investment in innovation, and the market power that results is not used to distort competition in, for example, related product or service areas. But because intellectual property is now a principal, if not the principal, barrier to new entry in high tech markets, we are also concerned that it be interpreted in a way that does not distort the traditional balance between intellectual property and antitrust.
I am concerned that recent cases, and particularly the Federal Circuit's opinion in Independent Service Organizations Antitrust Litigation("Xerox"),(18) have upset that traditional balance in a way that has disturbing implications for the future of antitrust in high-technology industries.
In Xerox, a group of independent service organizations (ISOs) in the business of servicing high-speed copiers and printers alleged that Xerox' refusal to sell or license replacement parts and diagnostic software to ISOs precluded them from competing effectively against Xerox in the service aftermarket. Obviously, the issue before the court was very similar to that considered by the Supreme Court in 1992 in Kodak,(19) where ISOs challenged Kodak policies designed to limit the availability of parts to ISOs and to make it more difficult for ISOs to compete with Kodak in servicing Kodak equipment. The Supreme Court concluded that Kodak's policy of selling replacement parts for its machines only to buyers of Kodak equipment who use Kodak service to repair the machine could be an illegal tie-in sale, as well as a monopolization or an attempt to monopolize the service and parts markets, and remanded for a trial on those issues. On remand, Kodak raised for the first time the argument that its parts enjoyed patent and copyright protection,(20) but the Ninth Circuit rejected that defense on evidence that it was a pretext.
In the subsequent Xerox case, the Federal Circuit explicitly rejected the Ninth Circuit's approach in Kodak, refusing to consider evidence that the purpose and effect of Xerox' actions were anti-competitive. Thus, the Federal Circuit wrote:
"We see no more reason to inquire into the subjective motivation of Xerox in refusing to sell or license its patented works than we found in evaluating the subjective motivation of the patentee in bringing suit to enforce that same right.
In the absence of any illegal tying, fraud in the Patent and Trademark Office, or sham litigation, the patent holder may enforce the statutory right to exclude others from making, using or selling the claimed invention free from liability under the antitrust laws. We therefore will not inquire into his subjective motivation for exerting his statutory rights, even though his refusal to sell or license his patent invention may have an anti-competitive effect, so long as that anti-competitive effect is not illegally extended beyond the statutory patent grant.(21)
Although Xerox' action was similar to the abrupt discontinuance of previous business relationship found by the Supreme Court to be illegal in Aspen Skiing,(22) and the plaintiffs were prepared to show the discontinuance was introduced for an anticompetitive purpose, the Federal Circuit nevertheless affirmed the lower court's grant of summary judgment against the ISOs. Moreover, it did so in sweeping language that exalts patent and copyright rights over other considerations and throws into doubt the validity of previous lines of authority that attempted to strike a balance between intellectual property and antitrust.(23)
The Federal Circuit concluded that a patentee can refuse to license or sell, and is immune under the antitrust laws for that refusal, unless one of the following conditions applies:
1. The patent was obtained by fraud on the PTO,
2. The suit to enforce the patent was "sham" - as that term was defined by the Supreme Court in Professional Real Estate Investors,(24) or
3. The patent was used as part of a tie-in strategy to extend market power beyond the legitimate confines of the patent grant.(25)
Putting aside questions about the logic of these three exceptions to absolute immunity for the patent holder,(26) these appear to be extremely narrow limits on a virtually unfettered right of a patent holder to refuse to deal in order to achieve an anticompetitive objective. For example, a claim that a patent was obtained by fraud on the PTO -- a so-called Walker Process claim(27) -- is more difficult to prove than almost any antitrust allegation because the Federal Circuit requires clear evidence that a patent applicant made knowing and willful misrepresentations that resulted in a patent that would not have issued in the absence of a misrepresentation.(28) With respect to sham litigation, the Supreme Court has held that the litigant has an absolute defense if the plaintiff had an objectively reasonable claim (regardless of any predatory motive), and an objectively reasonable claim is defined as one where "no reasonable litigant could realistically expect success on the merits."(29) Misuse of a patent to coerce purchasers to take an unpatented separate product, assuming the patent confers true market power, has long been thought to be an illegal tie,(30) but hardly exhausts the various situations in which a patent can be misused with anticompetitive effect.
More important than the Xerox result itself, questions arise as to what the Federal Circuit's approach portends - i.e., an approach that seems to exalt protection of intellectual property rights - with respect to continuing validity in the Federal Circuit of the long-standing balance between antitrust and intellectual property. Let me be clear that I have no quarrel with the fundamental rule that a patent holder has no obligation to license or sell in the first instance. A patent holder is not under any general obligation to create competition against itself within the scope of its patent.(31) But what will the rules be when the patent holder conditions the availability of its patented products or inventions on terms that affect competition? The Xerox opinion could be read to say that the invocation of intellectual property rights settles the matter, except in the three narrow situations described in the opinion, regardless of the effect of the refusal to deal on competition or the importance of the refusal to deal to protect incentives to innovate. That should not be the way these issues are addressed.
Assume in each of the following situations that the patent reflects substantial market power. How would the approach of the Federal Circuit play out:
1. Suppose a patent holder refuses to sell except on condition that the purchaser not buy from a potential competitor. Assuming monopoly power, that could be a violation under Lorain Journal.(32) Would a patent be an absolute defense because the product incorporates a patented element?
2. Suppose an inventor licensed an important process patent to five firms in a traditional manufacturing sector. One of the firms is a price-cutter. If the inventor terminates the license, would Xerox preclude any investigation into motive, or into the possibility that termination, though unilateral, resulted from anticipation of what the other licensees wanted?(33)
3. Suppose two firms have entered into a patent pooling agreement in which each firm retains veto power over the selection of its partner's licensees. Would a unilateral refusal to license, designed to reduce competition below levels that would exist in the absence of the pooling agreement, be protected activity, even if the pooling agreement led to the refusal?(34)
4. Suppose a patent holder knowingly misinformed a standard setting organization that it had no patents in a particular area, and as a result the organization developed a standard that required use of the patent holder's patent. If the patent holder refused to license, or would license only at exorbitant rates, would that be protected activity?(35)
Some might believe that these concerns about the sweeping language of a particular decision are unwarranted or premature. I hope so. But consider the recent District Court decision in Townshend v. Rockwell Int'l Corp., (36) where the owners of basic patents underlying the 56k modem technology sued for patent infringement. The defendant, Rockwell, launched antitrust counterclaims alleging that the patents on which the suit was based were invalid, the technology under the patents had been adopted as part of an industry standard through fraud on a trade association and its members, and the patents were made available to competitors only on condition that they cross-license their technology to the patent holder. The District Court dismissed the three antitrust claims, concluding in part that "[b]ecause a patent owner has the legal right to refuse to license his or her patent on any terms, the existence of a predicate condition to a license agreement cannot violate the antitrust laws."(37) Among the authorities cited for that and other conclusions is the Federal Circuit's Xerox opinion.(38)
It is important for people concerned about incentives to produce intellectual property and people concerned about antitrust to engage constructively. The Federal Circuit's overall attentiveness to preserving incentives to innovate makes sense, particularly in an economy that depends more and more on innovation to advance consumer welfare. Nevertheless, the broader implications of the Xerox decision are troubling. Traditionally, cases at the intersection between intellectual property and antitrust have been analyzed by examining the impact on economic incentives to innovate and balancing them against anti-competitive effects. As I tried to point out earlier, that balance usually (though perhaps not always) has been accomplished with great respect for and concern about protecting incentives to innovate. The newer approach, focusing only on a statutory right to exclude, appears to depart from that tradition.
An approach that starts from the point that a patent holder does not have to sell or license to anyone, and proceeds from that unchallenged assumption to the rule that it therefore can condition its sales or licenses in any way it sees fit, (with tie-in sales as the sole antitrust exception), would be an unwise and unfortunate departure from the traditional approach in this area. I question whether there is reason to believe any such interpretation is necessary to encourage the innovation process.
1. Chairman of the United States Federal Trade Commission. The views expressed are my own and do not necessarily reflect the views of the Commission or other Commissioners. I want to thank Michael McFalls for his valuable contributions to this paper.
2. New York Times, April 28, 2000.
3. In emphasizing avoidance of cartels and containment of illegal monopoly power - concepts derived from economic literature - I don't mean to suggest that the reasons for hostility to cartels and illegal monopolies are exclusively economic, or that the antitrust laws be interpreted solely with economic goals in mind. On the contrary, concentrated market power can impair individual and business freedom and, depending on the sector of the economy in which it occurs, can on occasion threaten democratic values that require dispersion of economic power. See Robert Pitofsky, The Political Content of Antitrust, 127 U. Pa. L. Rev. 1051 (1979).
4. U.S. Department of Justice and Federal Trade Commission, Horizontal Merger Guidelines § 0.1 (April 2, 1992).
5. Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985); Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447 (1993).
6. Northwest Wholesale Stationers, Inc. v. Pacific Stationery & Printing Co., 472 U.S. 284 (1985).
7. Continental T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36, 51 n.18 (1977).
8. Northern Pac. Ry. v. United States, 356 U.S. 1, 6 (1958); Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 12 (1984).
9. Klor's Inc. v. Broadway-Hale Stores, 359 U.S. 207 (1959); Federal Trade Commission v. Indiana Federation of Dentists, 476 U.S. 447 (1986).
10. Frank H. Easterbrook, The Limits of Antitrust, 63 Tex. L. Rev. 1, 15 (1984) ("[T]he economic system corrects monopoly more readily than it corrects judicial errors. There is no automatic way to expunge mistaken decisions of the Supreme Court. A practice once condemned is likely to stay condemned, no matter its benefits.")
11. Automobile Mfrs. Ass'n v. United States, 307 F. Supp. 617 (C. D. Cal. 1969), aff'd sub. nom. City of New York v. United States, 397 U.S. 248 (1970).
12. See Standard Oil Co. (Indiana) v. United States, 283 U.S. 163 (1931).
13. See Baker-Cammack Hosiery Mills v. Davis Co., 181 F.2d 550, 569-71 (4th Cir.), cert. denied, 340 U.S. 824 (1950).
14. Berkey Photo, Inc. v. Eastman Kodak Co., 603 F. 2d 263, 281 (2d Cir. 1979) ("If a firm that has engaged in the risks and expenses of research and development were required in all circumstances to share with its rivals the benefits of those endeavors, this incentive [to innovate] would very likely be vitiated. Withholding from others advance knowledge of one's new products, therefore, ordinarily constitutes valid competitive conduct.") cert. denied, 444 U.S. 1093 (1980).
15. FTC/DOJ Antitrust Guidelines for Collaborations Among Competitors, § 3.31 (a).
16. For a description of the Nine No-No's, see Willard K. Tom & Joshua A. Newberg, Antitrust and Intellectual Property: From Separate Spheres to Unified Field, 66 Antitrust L. J. 167, 178-184 (1998).
17. United States v. Alum. Co. of Am., 148 F. 2d 416, 427 (2d Cir. 1945) ("Many people believe that possession of unchallenged economic power deadens initiative and depresses energy; that immunity from competition is a narcotic, and rivalry is a stimulant, to industrial progress; that the spurs of constant stress is necessary to counteract an inevitable disposition to let well enough alone.")
18. 203 F.3d 1322 (Fed.Cir. 2000)
19. Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992).
20. On remand, the ISOs dropped their tying claims against Kodak, contending that Kodak's unilateral refusal to sell replacement parts violated Section 2 of the Sherman Act. See Image Technical Servs., Inc. v. Eastman Kodak Co., 125 F. 3d 1195, 1201 (9th Cir. 1997).
21. 203 F. 3d at 1327-28.
22. Aspen Skiing Co.v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985).
23. For cases discussing the balance, see SCM Corp. v. Xerox Corp., 645 F. 2d 1195 (2d Cir.1981); United States v. Studiengesellschaft Kohle, m.b.h., 670 F.2d 1122 (D.C. Cir. 1981)
See also Tom & Newberg, 66 Antitrust L. J. at 173-175 (describing evolution of balance between antitrust and intellectual property); Louis Kaplow, The Patent Antitrust Intersection: A Reappraisal, 97 Harv. L. Rev. 1815 (1984).
24. Professional Real Estate Investors, Inc. v. Columbia Pictures Industries, 508 U.S. 49 (1993).
25. Plaintiffs also challenged Xerox' policy of refusing to sell copyrighted manuals. The Federal Circuit treated the policy as a unilateral refusal to sell intellectual property, and reached much the same conclusions as those relating to patent issues. The copyright portion of the opinion raises policy issues that are similar to the patent issues and, at least in the context of this paper, does not require separate discussion.
26. For example, if the plaintiff proved that the defendant had attempted to enforce an invalid patent, as part of a Walker Process or sham litigation claim, then the defendant would have no lawful patent to license and permissible limits of its refusal to do so would be irrevelant.
27. Walker Process Equipment Inc. v. Food Machinery and Chemical Corp., 382 U.S. 172 (1965).
28. See Nobelpharma AB v. Implant Innovations, Inc., 141 F. 3d 1059, 1070-71 (Fed. Cir. 1998). See also James B. Kobak, Intellectual Property, Competition Law and Hidden Choices Between Original and Sequential Innovation, 3 Va. J.L. & Tech. 6 (1998).
29. See Professional Real Estate Investors, at supra 508 U.S. at 60.
30. See, e.g., Int'l Salt Co. v. United States, 332 U. S. 392 (1947).
31. See, e.g., SCM Corp. v. Xerox Corp., 645 F. 2d 1195, 1206 (2d Cir. 1981) ("Where a patent has been lawfully acquired, subsequent conduct permissible under the patent laws cannot trigger liability under the antitrust laws.")
32. Lorain Journal Co. v. United States, 342 U.S. 143, l55 (1951) (holding that a local newspaper attempted to monopolize the market for local advertising by refusing to deal with advertisers who dealt with a new local radio station).
33. See, e.g., Mannington Mills, Inc.v. Congoleum Industries Inc. 610 F.2d 1059, 1073 (3d Cir.1979) (noting that "the patent system has no interest in permitting the patentee's monopoly to be used as a screen for the maintenance of a horizontal cartel at the licensee level"); International Wood Processors v. Powder Dry Inc., 792 F.2d 416 (4th Cir. 1986).
34. See, e.g., United States v. Krasnov 143 F. Supp.184, 201-02 (E.D.Pa. 1956), aff'd per curiam, 355 U.S. 5 (1957); United States v. Besser Manufacturing Co. 96 F. Supp. 304, 311 (E.D. Mich. 1951), aff'd, 343 U.S. 444 (1952).
35. These issues arose in a recent enforcement action against Dell Computer. See Dell Computer Corp., C-3658 (May 20, 1996) (consent order).
36. 2000 U.s. Dist. LEXIS 5070 (N.D. Cal. March 28, 2000).
37. Id at *26.
38. Id. at *23 ("Given that a patent holder is permitted under the antitrust laws to completely exclude others from practicing his or her technology," 3 Com's proposed licensing terms were not antitrust violations.)