One of the most interesting sets of issues facing policy makers today concerns the appropriate application of the antitrust laws to high-tech industries.
Most would agree that balanced antitrust enforcement - challenges to monopoly, cartels and practices that unfairly create or maintain monopoly power, facilitate cartels or otherwise restrain trade - has served the country well throughout the 20th Century. Indeed, after considerable ideological warfare about over-enforcement and under-enforcement of antitrust, there appears to have emerged in the most recent decade a rough consensus about appropriate levels of antitrust enforcement. Also the rest of the world seems to be following the American lead by initiating or augmenting competition enforcement regimes.
But should antitrust apply in comparable ways to high-tech industries, particularly those industries where the products or services are primarily the embodiment of intellectual property? Public attention has focused upon the Department of Justice's proceedings against Microsoft and the Federal Trade Commission's challenge to behavior by Intel, but "high-tech" applies well beyond the computer hardware and software industries. Comparable problems arise in applying the antitrust laws to industries as diverse as pharmaceuticals, biotechnology, communications, international credit and finance, and various delivery systems for entertainment and news.
The importance of these industries to the economy, and limited antitrust experience and precedent with regard to some uses of intellectual property, suggests that there ought to be careful antitrust attention to be certain that critical economic growth is not compromised by the abuse of private market power. On the other hand, it leaves open the question of whether antitrust principles, developed primarily in the context of smokestack industries, should apply comparably and with equal force to new problems that emerge in connection with high-tech industries. Some would argue that it is dangerous to undermine investment incentives in these dynamic industries, and therefore that antitrust should have no role at all.(2) A less confrontational approach suggests that because of the robust pace of innovation in high-tech industries, government should not intervene "unless certain that doing so will benefit consumers and the economy."(3) A variation on that theme argues that litigation lags are likely to be longer than the time it would take for competitive forces to reconfigure an industry, and therefore that antitrust should intervene only in the most unusual circumstances.(4)Even under these last two approaches, antitrust would have a modest role.
If antitrust enforcement is to play a constructive role, I believe it is essential to acknowledge that high-tech industries are different and enforcement must take those differences into account. That does not mean there will not be old-fashioned predatory strategies or cartel behavior from time to time in high-tech industries; indeed I am confident that does occur. But more subtle problems of antitrust enforcement must adjust to the special circumstances of high-tech industries. That was the conclusion of the FTC's Report on Global and High-Tech Competition,(5) published in 1996 after a very extensive set of hearings on those subjects.
What are the differences that call for adjusted antitrust treatment in the high-tech sector? Among others, they include the following:
1. Technical Issues. Many high-tech industries involve questions that are challenging for lawyers and judges who typically lack a technical background. For example, defining relevant markets, i.e., the process of identifying those firms that compete so closely with other firms that they can substantially influence the exercise of market power, is difficult enough under any circumstances. But it can become far more difficult in high-tech industries such as biotechnology, where products that might curtail the market power of a dominant incumbent firm are not in existence yet, and will not reach the market for several years;(6) or in the cable industry where the essential question is when satellite transmission will become a real competitive force in the cable market.(7) Similar problems arise with respect to telecommunications, a sector of the market where many believe competition for local operating companies will eventually be offered by electric utilities through their existing grid and electricity wires into the home. Each of these issues raises questions in the realm of science and technology that often will be difficult to address.
2. Speed of Market Transition. New generations of products, undermining existing market power, appear more frequently in high-tech than in mature industries. In the first one-half of the 20th Century, firms in steel, oil and aluminum remained dominant for generations, but that is often not the case in many high-tech industries. An often-cited example involves IBM, which probably was a dominant firm if not a monopolist in certain markets when the government initiated its case in 1969, but which had lost monopoly power in many of these markets 13 years later when the case was abandoned.(8)
3. Need for Collaborative Activities. In high-tech industries, joint research and development is often essential to share the risks of innovation and to combine complementary technologies. Collaboration and later coordination on standards may be essential to allow products to work at all. As a result, some have suggested that antitrust must abandon its entrenched skepticism of cooperative arrangements and allow more leeway in high-tech markets.
4. Barriers to Entry. Because competition in high-tech industries so often depends upon the implementation of ideas, and ideas have little respect for geographic borders or entrenched market power, many predict that existing market power will be transitory. Put another way, competitive problems that may occur in high-tech industries are said to be "self-correcting" through the rapid and seemingly perpetual introduction of new products.
5. Output and Price Effects. The traditional profile of a monopolist is of a firm that will curtail output in order to raise price. But that model often does not hold in high-tech markets. Partly because the front-end investment in new forms of technology is so great, and the marginal cost of additional copies or products modest, high-tech firms often price aggressively at the outset to achieve dominant market positions and ultimately to take advantage of economies of scale. There may also be unusually substantial learning efficiencies, sometimes described as an "experience curve" that result in production costs declining along with increased output. Also, the orthodox idea that monopoly is a narcotic and that the reward of monopoly is to enjoy the "quiet life"(9) hardly describes many of the aggressive, dynamic, innovative high-tech firms in today's American economy.
6. Network Efficiencies. Finally, and most perplexing, there is the question of how to deal with network efficiencies. These efficiencies occur when the value of a product or service is positively correlated with the number of individuals who use the product or service. This can occur directly when a product's value is determined by the number of users in the network - for example, fax machines. It also can occur indirectly where a product achieves dominance and producers of essential complementary products (for example, application software firms that write programs for a dominant platform) overwhelmingly devote their resources in a way that is useable only with the dominant system. On the one hand, such networks are efficient and occasionally inevitable; on the other hand, they increase the likelihood that one firm, by achieving a critical mass, will dominate a market or retain market power for an extended period of time.
Because of these and other differences, some would argue that allowing firms with market power in high-tech industries to exploit that market power is a sensible price to pay for a vibrant innovative economy. Vast returns to the innovator, in that view, are thought to be worth it.
But even if one acknowledges the validity of these and other differences between high-tech and more mature industries, the essential question remains: should antitrust's conventional role as enforcer of competitive principles be abandoned because it is unnecessary or counter-productive, or should basic antitrust principles apply in a modified form, after taking due notice of different characteristics of the high-tech sector of the economy?
I believe antitrust should - indeed must - continue to apply. None of the "high-tech differences" justifies a complete or even substantial exemption.
While it is true that the technology of high-tech industries is often complex, that should be a reason for antitrust officials and judges to understand the technology, not to abandon the field. Speed of change in the high-tech sector is indeed impressive, but that often occurs at the margin of the industry; large firms can still obtain and maintain dominant positions, and extract monopoly rents, for extended periods of time for reasons or through practices unrelated to the enhancement of consumer welfare.(10) Collaborative efforts are often essential, but antitrust traditionally has taken a very lenient view of research and development cooperation and standard setting across the economy,(11) and, in almost all cases, production cooperation as well. There are important situations in which barriers to entry in high-tech industries are low - indeed where new entrants can leap-frog incumbent technology and quickly come to dominate a market - but that is not the whole story. Because high-tech industries so often are based on intellectual property, patent and copyright laws can insulate a dominant firm (sometimes legitimately; sometimes not) behind the barricade of hundreds and even thousands of those forms of protection. And, as I will discuss shortly, a network, once it achieves dominance through network efficiencies, can preclude competition for extended periods.
Finally, while it is true that aggressive high-tech firms often (though not always) enter the market at low prices, and many high-tech firms are admirably innovative and hardly can be said to seek the "quiet life," the question remains whether two or more firms vying for the consumer's allegiance would offer even lower prices or higher levels of innovation. A second concern is that the high-tech firm, once it gains a monopoly or dominant position, will gradually raise price as the market matures. There is simply no evidence at this point that this nation's preference for competition over monopoly, where competition is feasible and monopoly is achieved or maintained through unnecessarily exclusionary practices, should be abandoned in the high-tech sector.
That leaves us with what I have already described as the most perplexing question in this area: what to do about firms that achieve dominant and almost unassailable market positions legally, but maintain those positions for extended periods as a result of network efficiencies. While such situations occur across the economy, they do seem to occur more frequently in high-tech markets characterized by rivalry based on research and innovation. I have elected to discuss this difficult issue not because I believe it will often be the proper subject of antitrust review, but rather because it raises in the most stark way the conflicting values at work in this emerging area of the law.
As noted earlier, there are instances where it may be efficient for a single firm to become the only supplier of related products or services. Once that occurs, 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 them to the established network and resist preparing products for would-be challengers. When that occurs, market power, far from being transitory, can be retained by a dominant firm for an extended period of time.
Network monopolies or near monopolies are not unassailable. Occasionally, a new technology can completely leap-frog the market position of the incumbent network and quickly attract customers and suppliers. Sega's success with a 16-bit video game, quickly surpassing Nintendo's 8-bit format, demonstates this point. Challengers may invest heavily in a loss leader strategy, subsidizing efforts to challenge a dominant firm, and can win at least a portion of the market. But there are situations presently, and surely others will arise, where a network obtains market power legally, triumphs over all others, and is far more efficient in marketing its products or services than any existing or likely challengers.
There are substantial costs to a "hands-off" policy. Once a network monopoly is in place, it often is a simple matter for the monopolist to exclude would-be challengers. A network monopolist may have the ability to monopolize successive generations of product, or complementary products or services, simply by adopting a policy of allowing only those products manufactured by it to connect with the existing network. One result may be that over time, the best products or services may not win out. Also, potential competitors, recognizing the enormous difficulties of challenging an incumbent network monopolist, may lose incentives to compete.
In that situation, does antitrust have any constructive role to play?
One possible approach, consistent with traditional if not recent antitrust principles, would be to break up the network. Assuming that is not feasible, or that diseconomies (and resultant injury to consumer welfare) are substantial, another possible approach is to mandate access to the network for all qualified suppliers, customers and competitors on fair, reasonable and non-discriminatory terms. That is the sort of remedy that has been imposed occasionally on bottleneck monopolies and would be supported by some (but by no means extensive) precedent.(12)
Why or why not impose mandatory access?
Two questions that need to be addressed are, first, what are the pros and cons of a mandated access remedy and, second, assuming some sort of remedy is appropriate for some but not all network situations, what is the rule of law that distinguishes between beneficial and benign network positions on the one hand, and counter-productive network situations on the other? These are issues that in one form or another are before the Commission in several investigations and cases. For that reason, and also because learning in this area is an ongoing process, I will limit myself in the remainder of this paper to a review of factors pro and con with respect to a mandated access remedy, and of possible rules and standards, but will offer no firm conclusions. Let me examine some of the arguments for and against required access.
First, denial of mandatory access may generate incentives whereby applicants would be encouraged to either set up their own competitive network, or join smaller rival networks, that might challenge the dominant network firm. But that makes no sense in those extreme network situations in which an incumbent passes a "tipping point" - i.e., a degree of penetration that leads inevitably to a single network dominating the field. In that event, there will be no viable alternative networks that can be organized or joined.
Second, profits derived from dominating a network may be a fair return to innovators who achieve dominance. Certainly, it is a factor that rivals would take into account when they compete for the dominant position, and would encourage innovators to compete aggressively. Mandating access to late-comers or "free-riders" seems to reward the passive and less energetic at the expense of those who pioneered a field. The best response to that - and I am not sure it is a complete answer - is that few would recommend royalty free mandated access. The late-comers can be charged a royalty, including a premium that takes into account the risks the network incumbent took in achieving its dominant position.
Third, mandating access introduces all sorts of coordination and fairness
problems that would not otherwise exist. Mandatory access is hardly self-executing. For example, some third-party - - administrator, regulator, judge - - would have to supervise on a continuing basis the entire arrangement to see that access is not denied on unfair or discriminatory terms. That concern is less weighty where the network is controlled by a joint venture, since the participants in the joint venture would already have worked out many coordination problems likely to arise. It could be a serious problem when the network is controlled by a single dominant firm.
Finally, there is a sort of overarching philosophical concern that mandated access is a form of "confiscation"of property, to be avoided in most or all circumstances within any legal system that respects intellectual property and other property rights.
Arrayed against these considerable and formidable arguments is the simple fact that a dominant network, by exploiting its position, cannot only perpetuate its domination over substantial period of time, but can extract very substantial monopoly rents from consumers. There also should be a concern that by excluding from the network all except the dominant firm or its partners, potential participants who themselves may be innovators and who might carry technology beyond the level of the initial network arrangement are denied an opportunity to innovate so as to improve or actually surpass the initial network arrangement.
My own tentative view in light of these pros and cons is that antitrust enforcers should proceed cautiously in breaking up or mandating access to an existing network, even when that network is dominant. For the time being, until we know more about the origins and significance of network efficiencies, antitrust should probably concentrate in most situations on its traditional role of ensuring that companies achieve network monopolies through legitimate competitive conduct, and that they maintain the network dominance only through superior skill, foresight and industry - not by unnecessarily exclusionary conduct. That is particularly true when the network derives from intellectual property, a concept that has traditionally influenced antitrust policy which recognizes the wisdom of encouraging innovation.
Assuming a cautious approach is appropriate, there remains the problem of identifying which are the unusual situations in which antitrust enforcement is required. Chapter 9 of the Commission's report on its 1996 hearings on Competition Policy in High-Tech and Global Markets sets out three possible positions, and it seems to me those serve well to define the possible approaches. I do not endorse any of these but cite them only to illuminate a discussion.
Access is "essential." A cautious approach would be to reject any possibility of mandatory access except where it is "essential" to the existence of competition. If applicants for access can plausibly invent around the network monopoly, establish their own competitive networks, or join other networks that may not be equivalent but are acceptable alternatives to the dominant network, that arguably might eliminate any consideration of court-ordered access. There is some precedent that would point to a more aggressive approach. In the Supreme Court's 1945 decision in Associated Press,(13) an early "essential facilities" case with overtones of network efficiencies, the Court rejected the standard of mandating access only if it is "essential," requiring access even when the network confers only a significant competitive advantage and there was no good business reason to exclude. Some have criticized the Supreme Court's analysis in Associated Press as unnecessarily sweeping in view of the facts of the case.(14)
Purpose and intent. A second approach would be to ask the controllers of access to the network why they have refused access on reasonable terms. If they could assert a plausible efficiency, and if there were no evidence of a primary purpose to exclude, then access would not be required. In effect, the burden of explaining exclusionary behavior would shift to the party or parties controlling the network. Incidentally, the efficiency probably would have to be substantial, though not necessarily greater in consumer value than the anticompetitive effect.(15) Balance. In the end, focusing on any single factor - "essentiality," intent, efficiency and, of course, the exclusionary effect of denying access - may simply not be enough for a full evaluation of the issue. Professor Steven Salop of Georgetown Law School framed the issue with particular clarity in a recent, as-yet-unpublished paper about antitrust and monopoly power.(16) He raised the question about how to deal with a situation in which exclusion of competition might allow sellers to raise price 50 cents on a product, but mandating access might deprive consumers of 5 cents per unit in efficiencies. The issue could as well be put the other way - i.e., 5 cents increase in price because of lost competition vs. 50 cents of lost efficiencies if access is required by law. Either way the question properly raised is whether the right result can be reached by looking at any one factor and one factor alone.
Proper application in the antitrust laws to high-tech industries raises some of the most challenging questions that we have seen in a long time. As with any adjustment to new facts or proposed law, a cautious approach is called for. But abandoning antitrust principles in this growing and increasingly important sector of the economy seems like the wrong direction to go.
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.
2. Barro, "Why the Antitrust Cops Should Lay Off High-Tech," BusinessWeek, August 17, 1998, p. 20.
3. Priest, The Law and Economics of U.S. v. Microsoft, AEI Newsletter, August 1998.
4. Teece and Coleman, The Meaning of Monopoly: Antitrust Analysis in High-Technology Industries, 43 Antitrust Bull. 801, 843, 846 (1998).
5. Anticipating the 21st Century: Competition Policy in the New High-Tech Global Marketplace (May 1996).
6. For example, see Ciba-Geigy Ltd., Dkt. No. C-3735 (March 24, 1997).
7. See Time Warner Inc., Dkt. No. C-3709 (February 3, 1997).
8. See, e.g., United States v. IBM, 1997-1 Trade Cas. (CCH) ¶ 71,786 (S.D.N.Y. 1997) ("It has been established beyond any real question that, whereas IBM formerly had a great degree of market power in an antitrust sense, that market power has been substantially diminished, and is continuing to diminish, to the point of its disappearance in the sense of a threat of antitrust violation."), aff'd 1998 U.S. App. Lexis 32577 (2d Cir. Dec. 30, 1998); ILC Peripherals Leasing Corp.v. IBM, 458 F.Supp. 423, 430-31 (N.D. Cal.1978) (granting directed verdict to IBM where declining market share showed that IBM lacked monopoly power), aff'd per curiam sub nom. Memorex Corp.v. IBM, 636 F.2d 1188 (9th Cir. 1980).
9. United States v. Aluminum Co. of Am., 148 F 2d 416, 427 (2d Cir.1945) ("Many people believe that possession of unchallenged economic power deadens initiative, discourages thrift and depresses energy; that immunity from competition is a narcotic, and rivalry is a stimulant, to industrial progress; that the spur of constant stress is necessary to counteract an inevitable disposition to let well enough alone.")
10. See Allen-Mylan, Inc. v. IBM, 33 F.3d 194, 211 (3d Cir. 1994) ("IBM also contends that price reduction and product improvement are inconsistent with the existence of monopoly power. But rapid technological progress may provide a climate favorable to increased concentration of market power rather than the opposite. Moreover, a decline in price does not necessarily imply an absence of monopoly power; a fair profit might have been made at even lower cost to users.")
11. For example, federal enforcement has brought exactly one case against a research and development joint venture in the 109 years that the Sherman Act has been the law. See United States v. Automobile Manufacturers Ass'n, 1969 Trade Cas.(CCH) ¶ 72,902 (C.D. Cal. 1969), modified sub nom. United States v. Motor Vehicle Manufacturers Ass'n, 1982-83 Trade Cas. (CCH) ¶ ¶ 65,088, 65,175 (C.D. Cal. 1982).
12. United States v. Terminal R.R. Ass'n, 224 U.S. 383 (1912); Associated Press v. United States, 326 U.S. 1 (1945); Otter Tail Power Co. v. United States, 410 U.S. 366 (1973).
13. Associated Press v. United States, 326 U.S. 1 (1945).
14. In Asssociated Press, the bylaws of an association of major newspapers barred members from supplying news to competing newspapers, clearly with an anticompetitive intent, so that requiring access to an "essential facility" may not have been necessary to the decision. That case and others dealing with access questions are analyzed in Boudin, Antitrust Doctrine and the Sway of Metaphor, 75 Geo. L.J. 395 (1986). See also Areeda, Essential Facilities: An Epithet in Need of Limiting Principles, 58 Antitrust L.J. 841 (1990).
15. This may have been the approach of the D.C. Court of Appeals in dismissing the government's charge that Microsoft had violated an earlier order by integrating its browser technology into its Windows software. The Court appeared to find that Microsoft's behavior could not have been a violation because "plausible benefits" followed upon its integration of the browser. See United States v. Microsoft Corp., 147 F. 3d 935, 951-52 (D.C. Cir. 1998).
16. Salop, Preserving Monopoly: Economic Analysis, Legal Standards and Microsoft, ____Geo. Mason U. L. Rev.___ (1999).