RETHINKING ANTITRUST POLICIES FOR THE NEW ECONOMY
JOEL I. KLEIN
Assistant Attorney General
U.S. Department of Justice
Haas/Berkeley New Economy Forum
Haas School of Business, University of California at Berkeley
Portola Valley, California
May 9, 2000
I am very pleased to be here today. I want to thank the Haas School and Dean Tyson for inviting me to address you about a subject that I've thought and even re-thought about a great deal over the past few years: the subject of this panel, "Rethinking Antitrust Policies for the New Economy."
My conclusion is that the core principles of antitrust reflected in the Sherman Act -- like other fundamental principles embodied in venerable texts like the Constitution and the Bill of Rights -- should not be changed in this new era. All of these charters state enduring rules that can and should be applied in new situations. The Framers of the Constitution surely could never have imagined electronic eavesdropping; but the Supreme Court had no trouble ruling that this form of invasion of privacy was subject to the Fourth Amendment.
Core antitrust principles have served our Nation, our citizens, and our economy extremely well in the more than a century since the Sherman Act was passed. And I expect that they will continue to do so in the 21st Century, during this period of remarkable technological progress and expansion.
The core principles of antitrust are actually what Adam Smith wrote about more than two centuries ago: that free and competitive markets result in maximum economic development, wealth creation, and consumer welfare, but that markets will not always remain free and competitive in the absence of effective government oversight. In the end, antitrust is all about market power -- which every business understandably wants -- and the limits on how it can be obtained, preserved, and extended.
The legitimate ways of acquiring and maintaining market power are essentially the same today as they were a hundred years ago; and the illegitimate ways are fundamentally the same as well. "Skill, foresight and industry" is the term that antitrust lawyers use to summarize the permissible means of acquiring market power. But I don't have to tell this audience that that simple phrase can capture a broad range of productive and profitable business activity, activity that has contributed so much to our Nation's economic strength. And market power can legally be maintained in the same way, through innovation and competition in the marketplace.
The illegitimate means of getting and keeping market power have changed little since Senator Sherman's day as well. I will describe them in detail in just a bit. They deter innovation and restrict consumer choice, and they are as illegitimate and illegal today as they were a hundred years ago.
Two important corollaries follow from all this: First, sound antitrust policy does not believe that big is bad or that success must be punished. Quite the contrary -- where success is the result of skill, foresight and industry, consumer welfare is enhanced. To be sure, there have been times when antitrust enforcement has appeared to take a different view. For example, during the 1960's, the Division sometimes disregarded sound, market-based antitrust analysis in favor of a big-is-per se-bad philosophy. But that view fell out of fashion decades ago, and there is little prospect of its revival.
And second, since we believe that free and competitive markets maximize innovation and consumer welfare, we tend to disfavor regulation generally and certainly as a way to remedy abuses of market power. Ongoing regulation is invariably inefficient, both because it under-deters anticompetitive behavior and because it can be exploited by opportunistic rivals to hamper procompetitive conduct. Thus, where possible, we seek structural, market-based solutions to serious competitive problems, because these solutions mean that consumers, not government agencies or existing monopolists, will get to chose when longstanding monopolies yield to innovative technologies and innovative business models.
In this regard, I've been reading a lot lately about this issue of regulation versus structural solutions -- as it affects a case of some interest to me. It's a case that is well known to many of you as well, I'm sure. It involved complex and wide-ranging antitrust claims, resulting in a trial that gained lots of attention, followed by a Justice Department proposal to break-up a major American corporation.
Here's what the Wall Street Journal had to say about the case on its editorial page:
"While the Justice Department can't promise any consumer benefits that might result from its suit to break up [the company], it is sure of one thing: This is the largest antitrust action ever filed. So much for the mentality of modern-day trustbusters. As long as they can tackle the biggest of all 'big businesses,' what is the difference whether the massive expenditure of federal money and effort is likely to cut anyone's . . . bills?"
"Where is the problem that justifies risking possible damage to the efficiency of a vital part of the U.S. infrastructure; damage to the investments of innumerable small investors and pension fund beneficiaries; possible damage to an important research and development enterprise? If there is a problem that justifies all this we can't find it. Maybe it is because we prefer to deal in economics, rather than politics in such matters."
By now, you may have guessed that this is an editorial about the Department's monopoly maintenance case against AT&T, a 1974 editorial as a matter of fact; but if it sounds familiar, it is because the same charges have been leveled against the Department's lawsuit, and our proposed remedy, in the Microsoft case.
Then as now, the Department challenged illegal practices by a firm with monopoly power in a critical market, practices designed to maintain and extend the monopoly.
Then as now, the Department was criticized for challenging a technology leader and a critical part of the economic infrastructure.
Then as now, the Department sought a structural remedy because it is the most effective and efficient means of protecting and preserving competition.
And then as now, dire predictions were made that structural relief would kill the goose that laid the golden egg. One of my favorites is a Forbes Magazine article published the day after the AT&T divestiture took place: "For the consumer, costs will go up and service down . . . It's quite alarming, in fact, just how many top executives in the industry are predicting [this] . . . get used to it; it's going to get worse."
We now know, of course, that the divestiture in the AT&T case, far from making things worse, has unleashed unprecedented competition, innovation and consumer benefit. By separating the local telephone monopolies from other aspects of the telecommunications business, it has fostered the growth of the Internet, wireless communications, broadband services and fiber optics, and other extraordinary innovations that were unimaginable when the divestiture took place.
And it has also led to substantial competition in telephone services and significantly lower prices for consumers. Since divestiture, prices for long distance calls have fallen dramatically, while per capita use of long distance service has almost tripled -- an extraordinary output effect by any standard.
We believe that the proposed divestiture in the Microsoft case similarly would produce substantial innovation and competition in the software business. The district court found that Microsoft illegally maintained its operating system monopoly through a broad pattern of unlawful acts that crushed emerging threats to that monopoly posed by Netscape's browser, Sun's Java, and other cross-platform middleware technologies. We need to make sure that new technologies aren't subject to the same treatment in the future or, even worse, that innovators decide to avoid such technologies altogether for fear that they may meet the same fate if things don't change.
The central feature of our proposed remedy is splitting Microsoft into an operating systems company and an applications company. Unlike the AT&T case, where line-of-business restrictions remained on the local telephone companies, here the separated businesses would be entirely free to compete with each other in all lines of business. Each company would have the incentive to compete vigorously through developing and licensing products that compete with the other's core business.
For example, a separate applications company would have the incentive to develop the best possible office suite, not only for Windows, but also for other computing platforms like the Apple and Linux operating systems. Indeed, much like the browser was in 1995, before Microsoft commenced its illegal campaign, Office has the very real potential to be a cross-platform middleware threat to the dominance of the Windows monopoly.
Because Office is an enormously popular product -- with over 100 million copies in use around the world -- its availability on other operating systems would give those operating systems a real opportunity to compete against Windows. As these other computing platforms grow and proliferate, moreover, we would expect the Windows operating systems business to face real competition for the first time. And this is only one of several ways in which the proposed split is likely to facilitate competition. In toto, the result will be exciting and innovative new products, with more choices and lower prices for consumers.
Now, there are some who are suggesting that the reorganization will result in a loss of efficiency currently generated by the coexistence of the operating system business and the applications business under one roof. That argument is wrong as a matter of fact, and wrong as a matter of history as well. It is wrong as a matter of fact, since the two companies would be free to exchange technical information, as long as that information was also made available to third parties; and Microsoft has long claimed that it provided third-party applications developers all the information about its operating system that those developers could need to write their applications for Windows. If so, there should be no real loss of efficiency in the reorganization.
The argument is also wrong as a matter of history. The opponents of the AT&T remedy made the very same claim, arguing that the divestiture would imperil the efficiency of the telephone network; and that argument has surely failed the test of time.
Now, let me move away from this specific example of "the more things change, the more they stay the same" to the more general point about antitrust enforcement that I referenced at the beginning of my remarks. While technology changes, and that of course affects the particulars of our analysis, antitrust enforcement remains remarkably constant in its application of the core principles that have proven to be effective in protecting and preserving competitive markets while maximizing innovation and assuring low prices for consumers. These principles, as I noted earlier, have to do with market power and separating the legitimate, procompetitive ways it is acquired and preserved, from the illegitimate, anticompetitive ways.
Let me reiterate the fundamental point: businesses want market power -- i.e., the ability to make more than normal, competitive profits. It's good for the business, good for its employees, and good for its shareholders. And a rational, procompetitive system of antitrust laws must seek to ensure that the way business gets that market power is good for consumers as well.
To take an obvious example of a good way of acquiring and protecting market power, one from outside the antitrust arena, though by no means inconsistent with it, let's look at patent law. Here is an example where we grant statutory protections that tend to create and protect market power. That is why drugs cost so much -- absent patent protection, once a drug is created, it could be duplicated and readily sold at a small fraction of its patent-protected-price. The rationale behind patents, and the market power they establish and protect, is that, in the absence of patent protection and the returns it generates, no one would spend the money on R&D necessary to develop the drug in the first place. In short, we create a legally imposed barrier to entry -- intellectual property (or IP) protection -- in order to ensure that innovation is encouraged. One can argue, as many do, whether the period of IP protection is too long or too short to stimulate a desirable level of overall R&D, but the basic principle that, absent some IP protection, innovation would be harmed is clearly correct.
The next point I want to note here is that market power is not a unitary thing: there is market power and there is market power. Lots of businesses enjoy at least some market power, but very few enjoy monopoly power over any significant period of time. Brand loyalty or a first-mover advantage, for example, may give a business the ability to charge prices a bit above the competitive level, but in the absence of stronger barriers to entry than just brand loyalty or a simple first-mover advantage, the magnitude of these supracompetitive profits are likely to be quite modest.
This point, in turn, is key to understanding a fundamental market dynamic that animates antitrust analysis, i.e., the strength of barriers to entry is ultimately what determines how much market power a business will be able to sustain and exploit. At the same time, and somewhat paradoxically, the more a business exploits such power, the more potential competitors want a piece of the action. In short, supracompetitive profits, like well-known movie stars, draw a crowd; businesses, just like the bank robber, Willie Sutton, want to be where the money is.
And, in fact, as it turns out, because of the powerful incentives of the marketplace, it's quite rare that we see strong barriers to entry enduring for long periods of time. That is especially true in the absence of illegal business practices that augment the natural barriers that exist, a point that I want to come back to in a moment because it is at the heart of what antitrust enforcement is all about. But this view about the strength of entry barriers, at least in certain critical industries, has not always been widely shared. On the contrary, there have been quite a few times in our history when entry barriers to particular markets were thought to be so strong, we concluded that the market was a so-called "natural monopoly" and that we had no choice but to regulate it. Indeed, not so long ago, that was the case with respect to surface and air transportation, telephones, and energy (and as to the latter two, still is the case to some degree even today).
But now, with increasing confidence and conviction, we in America (and much of the world as well) have been won over to the view that, in the absence of illegal practices, technology will ultimately be able to erode almost any barrier to entry. Consequently, for several decades now, we have wisely adopted a national policy that favors deregulation and market forces instead of regulation.
This is not to suggest that market forces cannot generate strong barriers to entry. They can, especially in markets characterized by a so-called positive feed-back loop, either from scale economies or from what economists call "network effects." What this fancy jargon means is something we all tend to understand intuitively: in certain circumstances, nothing succeeds like success. A network effect occurs when the more a business sells of a particular product or service, the more people want it because its increasing adoption increases its value to the next user. A classic example, of course, is the telephone: the more people on a given network, the more value the network has to potential users, making it easier to get the next customer, and so on. Indeed, once a network gets a sufficiently large number of customers, it becomes almost impossible for a new entrant without access to the network to successfully challenge its dominance.
Two things I want to emphasize here about these kinds of positive feedback situations: first, they existed in the old economy, just as they do in the new. We had an old-economy case against AT&T, for example, where market power was derived in this fashion. And our new-economy case against Microsoft relies on this notion as well.
Like the telephone system, the Windows operating system at issue in the Microsoft case also benefits from a positive feedback loop. People select an operating system based largely on the number of applications available to run on that operating system, and people who develop applications want to develop them for the most popular operating system, since that is the way to sell the most applications. As a result, a dominant position in operating systems reinforces itself because the applications developers write to your operating system and then more new computer buyers want your operating system because desirable applications are available to run on it.
The second point to understand about these positive feedback loops is that there's nothing illegal or even undesirable about them: they are an outgrowth of market forces and consumer choice and, so far as the antitrust laws are concerned, businesses which have the skill and foresight to understand and take advantage of those forces are entitled to enjoy the fruits of their efforts.
In both AT&T and Microsoft, antitrust enforcement became an issue not because of the acquisition of market power but because of how that power was protected and/or expanded. This is a fundamental point to understanding the future of antitrust enforcement and so, in the time that remains, I would like to expand on it briefly.
As I have noted, we in America have chosen, wisely in my view, to reject an effort to regulate all monopolies; instead, we generally put our faith in the ingenuity of the market -- entrepreneurs and innovators -- to erode barriers to entry and protect consumer welfare. But if monopoly power, once had, can be used to protect and extend itself, our reliance on the market will be frustrated and consumers will be hurt. Unlike positive-feedback-loops, which are a natural and inevitable market phenomenon, abuse of market power is anticompetitive and harmful; it means that a monopoly position has prevented innovation and entrepreneurship that would strengthen the economy and increase consumer welfare.
What's interesting in this regard -- and this is why I say that the new economy is fundamentally no different from the old when it comes to antitrust enforcement -- is that the anticompetitive techniques used to protect and extend monopoly power in the new economy are essentially no different from those used throughout history. Put a bit differently, while technology changes, human nature, as Adam Smith taught us long ago, does not. There are, to paraphrase Paul Simon, only so many ways to illegally hurt your competitor.
In our business, there are generally about a half-dozen or so of these techniques and they are used in the new economy in much the same way that they were used in the old. Let me first mention the basic techniques and then illustrate their application by referring to cases involving the new and old economies, mentioning for illustrative purposes three that are currently in court. The basic techniques -- apart from good old fashioned collusion in which potential competitors agree not to compete -- typically involve cutting off competitors' access to important suppliers and markets, inducing rivals not to compete, using tying to force customers to purchase other products, and engaging in predatory tactics to raise rivals costs or cut their revenues without a real business justification. Basically, these are the time-tested tricks of the monopolist's trade.
Let's take a quick look at several of them. First, there are the traditional anticompetitive distribution techniques: intimidating or coercing distributors who need your monopoly product, either informally or through formal exclusionary contractual arrangements. These kinds of practices are as old as the antitrust laws themselves and rest on the sound premise that the use of market power to restrict distribution of competing products can only injure consumers. That point is at the heart of our complaint in the Dentsply case, involves exclusive dealing contracts with distributors of artificial teeth to dental labs. It was also a key issue in the Microsoft case where the judge found that Microsoft repeatedly intimidated OEMs who wanted to distribute competitors's products and used exclusionary contracts with Internet Access and Content Providers to limit their distribution of the Netscape browser.
A second common, anticompetitive distributional practice involves tying two products together -- once again, a violation as old as the antitrust laws themselves. Tying allows a firm to use its market power in one product to force consumers to take a second product and thus often makes it harder for the firm's competitors to distribute their products. To be sure, a tying case can present complex factual issues about whether there are one or two products at issue, which in turn can raise important questions about potential integrative efficiencies that might result from a "tie." But distributional efficiencies -- i.e., simply putting two products together -- are no defense to tying. That was true in the 1930s when a unanimous Supreme Court ruled that IBM's decision to tie calculating cards to its calculator was unlawful and that was also true under the District court's opinion in Microsoft involving the tying of Microsoft's browser to its monopoly operating system.
Since a lot of discussion has focussed on the tying issue in Microsoft, let me emphasize that ties in the software industry, especially where, as in our case, the tied product (e.g., browsers) could undermine the monopoly position of the tying product (e.g. operating systems) can have particularly strong anticompetitive effects. In this regard, we need look no further than the remarks of Microsoft's Chief Operating Officer of Microsoft when he was asked in 1998 how small software companies could compete on products that Microsoft plans to fold into its operating system. His reply: these smaller rivals had three possible paths -- they could fight a losing battle, they could produce a successful product and then sell to Microsoft or another large company, or they could "not go into business to begin with because, hey, if you're a betting person, you know which way it's going to go." It's hard to think of a greater deterrent to innovation.
The next set of traditional antitrust violations involve what we call predatory, as distinguished from exclusionary, practices. Here we're talking about a business incurring expenditures that would be profitable only if they will defeat a competitor and then allow the business to recoup the short-term costs of the action through the long-term preservation of monopoly profits. And here again, these practices were used in the old economy as well as the new, a point readily demonstrated by the fact that this issue is at the heart of our American Airlines case and was key in Microsoft as well.
In the American Airlines case, we charged that, when faced with new entrants in Dallas, American incurred great expense -- by saturating the relevant city-to-city markets where the new entrant had started service (e.g., Dallas/Wichita) and lowering prices substantially -- in order to drive the new entrant from the market. The essence of the case is our claim that American would never have engaged in these practices had it not known that it could eliminate new entrants and then recoup its short-term losses by enjoying monopoly profits in the future. As American's CEO said to his colleagues at the time, "if you're not going to get them out [of the market], then [there is] no point to diminish [our] profit."
Moving next to the new economy, the facts of the Microsoft case provide an especially powerful example of this predatory technique. There, the judge found that Microsoft had spent hundreds of millions of dollars to develop and distribute Internet Explorer, not just for Windows but for Internet Access Providers and even for Apple. Microsoft did this, the court further found, even though it internally described IE as a "no-revenue product" and knew that, standing on its own, Microsoft's IE business strategy made no sense. After all, it's hard to sustain a business plan by paying millions of dollars to induce others to distribute a no-revenue product, especially one that cost hundreds of millions to develop. What made this strategy even more perplexing is that, according to Microsoft's own documents, "browser market share" -- share of this no-revenue product -- was seen a "priority number 1" within the corporation.
The reason this otherwise irrational business strategy made sense, of course, is that, as the district court found, Microsoft was protecting its monopoly profits in Windows by making sure that Netscape's browser did not obtain sufficient market share to create a platform that could ultimately erode Windows' dominance -- a fear that Bill Gates highlighted at the outset of Microsoft's anticompetitive campaign by noting that, if Netscape wasn't stopped, its browser would be able to "commoditize" the operating system.
I could give other examples of anticompetitive practices in the new economy -- like withholding technical information that competitors need to compete -- which were also observed in the old economy. But by now I think you get my basic point: when it comes to antitrust enforcement, the new, new thing isn't so new after all.
So let me conclude by highlighting two points. First, the focus of antitrust enforcement tomorrow, as it was yesterday, will remain on preventing the traditional anticompetitive techniques that businesses with market power have long used to maintain and extend that power. And second, given my first point, in the new economy as in the old, businesses with market power should have little problem in ordering their affairs in a way that keeps them free from antitrust difficulties.