The Tenth Annual Seattle Computer Law Conference
I want to thank the Computer Law Conference for this opportunity to speak to you today and to share some general views on the application of the antitrust laws to the computer industry and other so-called high tech sectors.
I have some trepidation, as a newly minted representative of an antitrust enforcement agency, about delivering a maiden speech in a setting where people may have very strong feelings on antitrust issues, one way or the other. Under the circumstances, I want to assure you up front that I do not intend to state an opinion on AThe Case.@ There are a number of reasons. First, I am only one Commissioner out of five, with an equal voice, and any views I express here are only my own and not necessarily those of any other Commissioners or the Commission itself. Second, it would be both unseemly and unproductive to opine on a matter involving a sister agency that is now in a mediation process as part of a federal trial. Third, and equally important, I don=t know anything about the case other than what I read in newspapers and magazines. For reasons that will become apparent, it is increasingly difficult for people not directly involved in a particular litigation to have an educated view of the merits.
I do believe, however, that some basic issues presented in Microsoft - - whatever your bottom line - - are not unique. Even people who disagree on ultimate factual determinations may be able to agree on what the issues are and on the method of analysis.
I also want to advance another proposition, which may be somewhat more controversial, namely: that it is not necessary to develop new antitrust principles to deal with so-called Ahigh tech@ industries; what is required is a discriminating application of familiar principles to the special facts of a Ahigh tech@ environment. I will talk about this more in a few minutes.
You should also understand that while I am giving you my own personal view as one member of a collegial body that enforces the antitrust laws, it is a view formed by forty years of observing how prosecutors and courts approach these types of cases. This view would be no different if I gave this talk two months ago, before I assumed my present responsibilities. If I were counseling private clients I would say many of the same things that I intend to say here now.
The Need for Balancing
There will always be some conduct, like price fixing, market allocation, or bid rigging, that is considered so likely to be harmful to competition that it is almost always condemned without further inquiry into the precise anti-competitive harm or pro-competitive purpose. These types of agreements rarely have plausible pro-competitive justifications and are considered Aper se@ illegal. The Supreme Court has necessarily limited the application of this Aper se@ rule to conduct that is Aplainly@ or Amanifestly@ anti-competitive, for which no elaborate demonstration of anti-competitive harm is needed. But, at the same time, it may be necessary to analyze various factors to determine whether particular conduct fits into this category in the first place. The distinction between a legitimate joint venture and a naked cartel is not always clear cut.
At the opposite end, there are arrangements, like short-term exclusive dealing arrangements between companies with no market power, which are considered so likely to be harmless that they are almost never attacked. These arrangements are per se legal in fact if not in name. But, it is still necessary to consider some market facts before a counselor can safely draw that conclusion. And, of course, there is a significant and growing number of cases between the two extremes which require consideration of factors that may point in opposite directions. Ever since the Supreme Court announced the Rule of Reason almost 90 years ago in the Standard Oilcase, antitrust enforcement has often depended on an analysis that will balance the various beneficial and harmful aspects of business practices.
The Rule of Reason is applied to cases Awhere the economic impact of certain practices is not immediately obvious.@ The specific analysis that is undertaken has evolved over the century and is still evolving, as required by fundamental changes in the world economy and advances in economic understanding. The Rule of Reason applies in cases under a statute like Section 1 of the Sherman Act, which requires a combination of two or more actors, or Section 2 of the Sherman Act, which a single actor can violate alone. The balancing process tends to be somewhat different in these two kinds of cases, however. At the risk of over-generalization, I would say that Section 1 cases tend to involve balancing over space and Section 2 cases tend to involve balancing over time. Ideally, this is an objective balancing process, but in the real world, subjective elements tend to intrude. Even if people can agree on factual elements (no mean accomplishment), they may weigh them differently because of different value judgments.
Balancing Effects in Space
Under a traditional application of the rule in a Section 1 case, the plaintiff bears the initial burden of proving that there is an agreement and that this agreement has a substantial adverse effect on competition. The burden then shifts to the defendant to demonstrate the pro-competitive effects of the conduct. If the defendant can demonstrate that the conduct has pro-competitive effects, then the plaintiff must demonstrate that the conduct is not reasonably necessary to achieve the pro-competitive effects. Ultimately, the issue will be whether the anti-competitive effects of the conduct substantially outweigh its pro-competitive effect.
Typically, the opposing factors that must be weighed in this kind of case involve varied impacts in different sectors, like the balance between intramodal and intermodal competition addressed by the Supreme Court in Sylvania. This landmark case involved exclusive territorial agreements between a seller and its buyers. The effect of the exclusive territorial agreements was to restrict competition between sellers of the same brand (Aintramodal@ or Aintrabrand@ competition). The Supreme Court acknowledged this restriction, but recognized that vertical restrictions such as exclusive territories could have an overall pro-competitive effect, because the associated efficiencies would enhance competition in another arena.
For example, manufacturers might use territorial restrictions to give retailers an incentive to engage in promotional activities that increase sales of its product. Retailers are more likely to spend money promoting a manufacturer=s product when the retailer can be assured that a no-frills retailer, which does not incur similar expenditures, is not located just down the street (thus preventing so called Afree riding@). Consequently, the Supreme Court concluded in Sylvania that territorial restrictions may have the overall effect of helping the brand compete with other brands (Aintermodal@ or Ainterbrand@ competition). The Supreme Court balanced the adverse effects on intrabrand competition against the favorable effects on interbrand competition and concluded that the latter is more important, stating, Awhen interbrand competition exists. . . it provides a significant check on the exploitation of intrabrand market power because of the ability of consumers to substitute a different brand.@ I think of this as balancing effects in space.
Balancing Effects Over Time
Section 2 cases, which usually involve a single actor, are particularly likely to involve a balance of effects over time. It is possible that some conduct with a favorable impact on consumers in the short-term might have an unfavorable impact in the long-term -- or vice-versa. Antitrust law, quite properly, takes account of both possibilities.
Take the offense of attempted monopolization under Section 2. The Supreme Court has stated that it requires: A(1) that the defendant has engaged in predatory or anti-competitive conduct with (2) a specific intent to monopolize and (3) a dangerous probability of achieving monopoly power.@ Thus, there must be some kind of conduct that can be characterized as anti-competitive or predatory before the balancing act is triggered. The law does not address the existence of a monopoly; it addresses efforts to Amonopolize,@ or A[u]nfair methods of competition,@ which suggests something beyond developing a better mousetrap and offering it for sale. The definition of what is predatory or anti-competitive conduct for a monopolist or an aspiring monopolist however, is difficult, controversial and highly fact specific.
There is a vast body of case law where courts have tried to distinguish between predatory and non-predatory conduct. For example, how do you decide whether a deep discount is predatory or not? Discount pricing confers immediate benefits on consumers, and you don=t want to discourage it just because it makes competitors uncomfortable. But if a company prices below some measure of incremental cost and actually incurs out-of-pocket losses on its incremental sales, it might raise a legitimate question about how the company expects ultimately to benefit. Is there some risk that the conduct will so cripple competition that the company will enjoy monopoly returns in the future? For this reason, most courts today apply some kind of cost screen to help decide whether a particular discounted price is or is not predatory.
Exclusive dealing arrangements raise similar issues. They can be attacked under Section 1 but in some circumstances may be deemed predatory in a Section 2 case. When a seller and a buyer agree that the buyer will purchase all of a particular product exclusively from the seller, there are immediate and obvious benefits to the seller. For example, an exclusive dealing contract may assure efficient utilization of plant capacity, facilitate long-term planning on the basis of known costs, protect against fluctuations in price, reduce selling expenses, and protect a new entrant trying to establish a foothold against counterattacks of entrenched competitors. All of these efficiencies may be reflected in lower prices to the buyer, to the ultimate benefit of consumers.
But there can be immediate harm to the seller=s competitors and long-term adverse effects on competition. Competitors are foreclosed for a specified period of time from marketing their products to the original seller=s particular customers. This may not matter if there are a lot of other customers to sell to besides the original seller=s customers. If a seller with exclusive arrangements has a very large share of the market and the period of foreclosure is sufficiently long, however, it can have the long-term effect of inhibiting competition not only from existing firms in the market, but from even more efficient firms in the future.
Consequently, determining whether exclusionary arrangements are ultimately anti-competitive may require the balancing of these short-term costs and efficiencies with the long-term effects. The Supreme Court examined exclusive dealing arrangements in the Tampa Electric case and stressed the need Ato weigh the probable effect of the contract on the relevant area of effective competition . . . and the probable immediate and future effects which pre-emption of that share of the market might have on effective competition [. . .]@
Tie-in sales are another example where a balance of present and future competitive effects may be significant. It is worthwhile to look at a few examples of tie-in sales, or product Abundling,@ because they also illustrate how some subjective factors can come into play. These are, of course, lively issues today but the issues will, I think, emerge more clearly if we use older examples which this audience may be better able to view in a dispassionate way.
Take a simple example. During World War II, when my father wanted to buy a bottle of scotch whiskey (which was then in short supply), he sometimes had to buy a couple of bottles of cheap wine that he didn=t really want. This was a classic tie-in sale and lot of people were viscerally indignant. I suspect examples of this kind may have contributed to judicial hostility at the time. But, there may really have been no economic harm. We don=t know anything about the competitive effects, and my father obviously thought he was better off with the whole package than without it, however indignant he may have been. The whole scenario was driven by regulatory evasion; the storekeeper could not simply charge more for the scarce whiskey because of wartime price controls. While regulatory evasion may supply the motive for particular conduct that may be otherwise inexplicable, it does not constitute independent evidence of competitive harm.
Consider a slightly different case of regulatory evasion. When oil supplies were interrupted in the 1970s, there were long lines at gasoline stations around the country. But, there were no lines at my local car wash, which sold gasoline and other auto-related products to its customers. The price of gasoline was controlled (which caused the long lines elsewhere), but the price of car washes was not, and this is how the facility was able to charge a market-clearing price for a package of products and services and keep the lines short.
Many people have a different visceral reaction to this example than they do to the liquor store example. It may be because there are some efficiencies associated with the way the car wash packaged its services. It is, after all, convenient for customers to take care of a number of auto-related needs at the same time, and it is significant that the car wash offered the same package after price controls were terminated. But, it is likely that there are other factors at work - - perhaps different feelings about wartime and peacetime price controls. In other words, some people=s conclusions about the ultimate competitive impact of regulatory evasion may be affected by their views about the impact of the regulations themselves. Regulatory evasion is an important issue today in some industries like telecommunications and energy, and these individualized reactions can be important because so many regulatory bodies are involved.
Take another example of packaging, where it is unclear whether the sale involves one product or two. When I was with General Motors, the company once made car radios standard equipment in certain car models, and wound up in litigation with distributors of high-end imported radios. Like complainants in the exclusive dealing example, these importers claimed they were almost entirely foreclosed from the opportunity to sell their products to purchasers of these models. And they were concerned that if the practices spread, it would ultimately put them out of business altogether. The lawsuit was settled when GM offered to extend a monetary credit to consumers who chose to delete the standard option, but when we were doing the research to defend it, we found that at one time such basic amenities as auto windshields and headlights were optional items as well! Presumably, a whole line of aftermarket suppliers disappeared when these items were made standard. Maybe you react differently to the example of windshields and headlights than you do to radios, but do you know exactly why?
It is also possible to tweak this example with purely hypothetical variations. Would it matter if 95% of the customers previously ordered these models with the GM radio option? Would it matter if there were a history of complaints about changes in vehicle dimensions or connections that made it difficult for outside manufacturers to keep up and produce products that fit? Would it matter if dealers had been discouraged from offering imported radios? Would it matter whether the interface between radio and car was simple or complex? Would it matter if there was a technological breakthrough that enabled the auto company to efficiently combine some of the electronic elements of a radio with other electronic elements in the car?
What if a product is Abundled@ with repair and maintenance products and services, for which there is also a lively aftermarket? Autos are already sold with a replacement part (the spare tire). What if other replacement parts were made standard? What if lifetime warranties were standard?
There are myriad other possibilities, and I doubt that your reaction to each of these variations is entirely dependant on your assessment of present consumer benefits. The use of these examples from other industries may help you to understand how outsiders look at some problems in the high-tech industries with which you are most directly concerned. There is no rigorous way (comparable to the cost tests for price predation) to distinguish between a pro-competitive combination of features in a single product and an anti-competitive tie of two separate products. As an antitrust counselor, I used to suggest to business people in a variety of client companies that if you arguably have some market power, it is useful to distinguish between conduct primarily designed to do something for customers as opposed to conduct primarily designed to do something to competitors. Most business conduct is a combination of both, and it is not so easy for an outside tribunal to get an accurate answer.
Remember that identification of Apredatory@ conduct is just the beginning of the inquiry. It is also necessary to define product markets in order to determine whether a company has market power, and market definition is not a precise scientific exercise. In addition, the Supreme Court treats the issue of Aintent@ as a separate element to be proved, and it is difficult to figure out the subjective Aintent@ of a large institution with many actors. (It is true, however, that the expressed Aintent@ of individual actors may illuminate how they, presumably as industry experts, view the likely competitive effects.)
Beyond all this, it is ultimately necessary to make an objective assessment of overall effects on consumer welfare, and the signposts may point in opposite directions. Conduct may simultaneously reduce competition in a way that hurts consumers and increase efficiency in a way that helps consumers. And, to complicate matters further, the relative importance of these effects today may differ from their effects tomorrow. And, as I hope the examples illustrated, subjective value judgments may intrude. This may explain why I think it is rash for me, or anyone else with an equally limited knowledge of the facts, to critique complicated cases from the sidelines or predict judicial outcomes.
But an equally important point for this particular forum is that there is nothing unusual or revolutionary about an antitrust analysis that balances short-term and long-term effects. It has been going on for a long time. The next question, however, is whether there is, nevertheless, something special about Ahigh-tech@ industries that makes this balancing process inappropriate.
Are High-Tech Industries Unique?
It has been argued that predictions of future competitive harm are particularly speculative in Ahigh-tech@ industries. It is argued that barriers to entry are low; the most important productive assets are not buildings and machines, or even trade names, but the brains of talented and highly mobile people. It is also claimed that the pace of innovation is rapid; today=s entrenched positions can be overthrown tomorrow.
There is obviously much to be said for these arguments. But, some of the same arguments could be made about yesterday=s high-tech industries. Henry Ford catapulted from backyard tinkerer to industry leader in a remarkably short period of time. Moreover, high-tech industries are not the only ones characterized by easy entry and rapid change. Publishing, entertainment and fashion are also examples of large industries that can be turned upside-down overnight. I remember that a highly respected treatise on antitrust, published in 1959 by Carl Kaysen and Donald Turner, cited Life magazine as an example of a product with an unassailable market position. A few years later, Life had disappeared, because of the impact of a high-tech product called television.
The experts can also be wrong in the other direction; sometimes dramatic changes are forecast that do not happen at all. During World War II, it was widely believed that traffic congestion would no longer be a problem in the postwar world because personal helicopters could take advantage of three-dimensional airspace. We also thought that household pests - - like flies and mosquitos - - would disappear when DDT became widely available. It didn=t work out that way, of course, because helicopters proved to be much more expensive and complex machines than we thought, and DDT had some seriously unpleasant side effects - - and pests mutated around it anyhow.
When CB radios in cars became popular in the 1970s, sales trends extrapolated forward suggested that CB radios would be ordered by the vast majority of customers in the near future. It didn=t happen, presumably because not many people in cars really want to talk to other people in cars, and because cell phones were better for all purposes.
A highly visible example of unduly optimistic future predictions was John Galbraith=s book, The Affluent Society, published in 1958. The basic message of the book was that this country=s economy was so efficient and so productive that we inevitably would accumulate immense wealth, and that we could now begin to devote substantial resources to public purposes. It was one of the most influential books on economics ever written and its glowing predictions undoubtedly influenced the explosion of government regulation in the late 1960s and early 1970s. And, of course, it was spectacularly wrong - - or at least it was wrong until wealth creation came from an unforeseen direction, so well represented here.
We now hear equally optimistic predictions about the dramatic changes that high-tech industries will bring to our lives - - particularly in the areas of communications and medicine. The suggestion is that antitrust intervention is inappropriate and unnecessary because any market dominance is bound to be transitory. I sincerely hope that these things will come true. But I=ve lived long enough to be skeptical.
Some arguments about the special status of high-tech industries point in a different direction. Theorists speculate that some high-tech industries are particularly susceptible to monopolization and ultimate competitive stagnation because of something called Anetwork effects.@ The idea is that some facilities are linked together in such a way that they become progressively more valuable to customers, as more and more people use them. A telephone exchange would be an obvious example because the value of a telephone progressively increases as more and more people are on line. This argument reminds me of older and more familiar claims that certain industries tend toward monopoly conditions because of ever increasing economies of scale. Only this time, the emphasis is on the shape and shifting of the demand curve rather than the shape of the supply curve. If the effect is real, it means that new entrants will have a very hard time catching up, and it also means that arguably predatory conduct may have a particularly detrimental long term effect.
This is not a speech about network effects, and I don=t want to dwell on the arguments pro and con. The only point, for present purposes, is that some people think there may be special factors in some high-tech areas that will ultimately lead to greater than normal market rigidity, and that, therefore, we need to be more - - rather than less - - concerned about market practices that have future competitive effects.
I personally tend to be somewhat skeptical of claims on both sides. I am not confident that the explosive changes we have witnessed in this century can be extrapolated indefinitely into the future, and I am equally wary of claims that particular markets -- high-tech or not - - can be locked up for long. Market forces are immensely powerful, and the tidal waves of the future are likely to approach from unexpected directions. At the same time, I don=t think that people who make or enforce laws can simply refuse to think about the future just because it is so hard to predict.
Let me give you an example of a well-established body of law - - outside antitrust but interacting with it - - that is largely based on an inherently speculative vision of the future. I refer to the patent system which protects intellectual property. I am sure this is important to a great many of you, and I am also sure that most of you know a great deal more about it than I do. But you may not have reflected on the fact that this system involves a similar balance between short-term and long-term effects. A patent grants a legal monopoly for 20 years and this monopoly, like any other, may have an immediate adverse effect on consumers. The patent monopolist has full freedom to exploit the patented technology or suppress it. The patent monopolist can, if it wishes, reserve sole rights to sell the product for whatever the market will bear, or require other sellers to pay a royalty -- so that it profits from higher prices both on its own sales and the competitors= sales.
But, patent rights are firmly embedded in our law in part because of the belief that they will have the beneficial long-term effect of stimulating innovation, not only in the particular industry involved but across the entire economy. This is an article of faith that I personally believe to be well-founded, but we have to recognize that it is based on highly speculative predictions. The patent regime, then, can be justified by the same intellectual process that supports some antitrust cases based on predation -- only in the case of patents, present consumer harm is balanced against the prospect of future consumer benefits, instead of the other way around. In that sense, patent protection is the flip side of Microsoft.
You may be disappointed because there is so much Aon the one hand - on the other hand@ rhetoric in this talk. But, that=s just a reflection of the fact that antitrust cases are necessarily dependent on individual facts. Even people with an understanding of the technologies involved - - and that may be no small matter - - differ on the inferences to be drawn.
But, even if the facts are unique, the methods of analysis are not. Arguments about Apredation@ in high-tech industries have historic parallels in simpler settings. In particular, I have tried to show that the process of balancing present and future effects is familiar in antitrust and elsewhere; there is nothing novel about it.
I have also tried to show why I am not persuaded that so-called Ahigh-tech@ industries require special rules. There may be special facts that need to be taken into account, and predictions may be inherently more difficult, but somehow we have to make them. Maybe, we need to approach these issues with an antitrust version of the Heisenberg uncertainty principle: the less confident we are about ultimate competitive effects, the more confident we have to be that the conduct is really predatory, and vice-versa.
Finally, I recognize the argument that when the scales are close to evenly balanced, or the future is inherently unpredictable, the most prudent course for a government authority may be to stand aside and do nothing. This suggestion opens up a vast number of economic and political questions that I can=t begin to address today -- including, not least, the degree of prosecutorial discretion available to people who have sworn to uphold the laws of this country. But, you need to understand that no single individual or no single regulatory agency can effectuate a laissez-faire policy. Legislatures, state governments, private parties and ultimately juries and courts have a lot to do with shaping antitrust rules. The consequences of abdication by one or both federal antitrust agencies, which some people have vehemently advocated, might be very different from what they expect. In fact, they might be the most unpredictable elements in the whole controversy.
. Broadcast Music, Inc. v. CBS, 441 U.S. 1, 8 (1979).
. See, e.g., Antitrust Guidelines for Collaborations Among Competitors, Issued in Draft by the Federal Trade Commission and the U.S. Department of Justice, Oct. 1, 1999, ' 3.2.
. Standard Oil Co. v. U.S., 221 U.S. 1, 58 (1911).
. FTC v. Indiana Fed=n of Dentists, 476 U.S. 447, 459 (1986).
. Practices that violate either of these statutes may also be A[u]nfair methods of competition@ under Section 5 of the Federal Trade Commission Act.
. Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977).
. Id. at 52.
. Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447, 456 (1993).
. A company is legally entitled to acquire or maintain monopoly power solely Aas a consequence of a superior product, business acumen, or historic accident.@ U.S. v. Grinnell Corp., 384 U.S. 563, 570-71.
. E.g., William Inglis & Sons Baking Co. v. ITT Continental Baking Co., 668 F.2d 1014, 1036 (9th Cir. 1981); Tri State Rubbish, Inc. v. Waste Management, Inc., 998 F.2d 1073, 1080 (1st Cir. 1993); Brooke Group v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 222 (1993); see also Phillip Areeda & Donald F. Turner, Predatory Pricing and Related Practices Under Section 2 of the Sherman Act, 88 Harv. L. Rev. 697 (1975).
. Tampa Electric Co. v. Nashville Coal Co., 365 U.S. 320 (1961).
. Id. at 329 [emphasis added].
. Carl Kaysen and Donald F. Turner, Antitrust Policy: An Economic and Legal Analysis (1959).
. John Galbraith, The Affluent Society (1958).
. This example demonstrates that utopian predictions, as well as doomsday predictions, can stimulate government regulation.
. See David A. Balto, Networks and Exclusivity: Antitrust Analysis To Promote Network Competition, George Mason Law Review, Volume 7, Spring, 1999.
. The notion of allowing patent holders to earn economic rents long antedates the antitrust statutes themselves. In Article I, Section 8 of the Constitution, Congress is granted the power A[t]o promote the progress of science and useful acts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.@