Article

Can Antitrust Keep Up?: Competition policy in high-tech markets

Jonathan B. Baker

Not since the 1911 breakup of the Standard Oil trust has a government antitrust case attracted as much public attention as the Justice Department’s thus-far successful suit against high-tech giant Microsoft.

Although its ultimate outcome may not be known for several years, the Microsoft case has generated an intense nationwide debate among antitrust practitioners and business leaders about the relevance of antitrust to high technology. The critics’ central claim is that the pace of change in high tech is so rapid that antitrust, and the legal machinery within which it must operate, is too slow and potentially counterproductive. Those who defend the applicability of antitrust to high-tech firms and industries, including the federal enforcement agencies, instead see a realistic potential for promoting innovation and consumer welfare through competition policy.

Antitrust’s contemporary engagement with high-tech markets, most visibly in the Microsoft case, is based on an emerging intellectual consensus about the way competition works in those markets. That consensus has been shaped importantly by two documents issued during the Clinton administration’s first term-government antitrust guidelines on intellectual property licensing and a Federal Trade Commission staff report on competition policy in the high-tech global marketplace. The contemporary perspective on competition in high-tech markets is based on five core principles likely to withstand shifting political winds.

Intellectual Property

The first principle is that for antitrust purposes, intellectual property is just another form of property. A high-tech firm’s patents and copyrights are no different from-indeed, are often more critical assets than-its physical property, its plant and equipment. That it may be easier to misappropriate many forms of intellectual property than to steal physical products is no reason to make intellectual property more or less suspect under the antitrust laws than any other form of property.

Consider mergers. If Coke were to acquire Pepsi or if United Airlines were to acquire American, the transaction would surely trigger an antitrust investigation, because the products and services sold by the acquired firm are likely important substitutes to customers of the acquiring firm. The analysis is similar when the key assets of the firms are rights conferred by the intellectual property laws. Federal enforcement agencies are routinely skeptical, for example, when a firm that owns the rights to one drug for treating a disease seeks to acquire a firm with the rights to another drug with a similar use. Standard antitrust doctrines also apply to mergers among rivals in “innovation markets” in which a handful of identifiable firms engage in research and development on similar new products for the same buyers, even though no new products have yet been created.

This principle-intellectual property is just property-also applies to the antitrust review of intellectual property licensing arrangements. In a recent case, the FTC charged that two leading producers of the equipment used for laser vision correction, Summit and VISX, fixed prices under the guise of an agreement to share patents. This aspect of the case was settled in mid-1998, and the two firms recently announced substantial reductions in their per-procedure fees. Similarly, in 1997 the Justice Department worked with the video compression industry to structure a “patent pool” by which nine firms could share the twenty-seven patents needed to meet an international standard without risking harm to competition.

A firm can also sometimes harm competition by cutting off a past customer. This was made clear by a 1951 Supreme Court case involving a newspaper monopolist faced with competition for advertising revenues from a newly established radio station. The newspaper violated the antitrust laws when it refused, for no legitimate reason, to do business with advertisers that bought commercial time on the radio station. Here too, antitrust enforcement makes no distinction between intellectual and other property. Thus, in a case settled in 1999, the FTC challenged Intel for cutting off access to technical information about upcoming Intel microprocessor products that customers needed to design complementary products like personal computers. The FTC argued that Intel harmed competition when it cut off its customers to gain leverage in unrelated commercial disputes involving the scope of competing intellectual property rights.

Competition and Innovation

The second principle, that competition promotes innovation, recognizes that firms have a powerful incentive to gain a march on their rivals by cutting costs or being first with new products or product improvements. The primary counter-arguments, often associated with economist Joseph Schumpeter, are that a monopolist may have greater access to low-cost internal finance, and may be better able to take advantage of scale economies in research and development and to appropriate the full value of its new ideas. But the first two are potential advantages of all large firms, regardless of the extent to which they face competition. And the third, the monopolist’s possible advantage in appropriating the gains from innovation, could matter only when other means of assuring reasonable appropriability, such as intellectual property protection, are weak.

Moreover, the case for relying on patent monopolies rather than competition to generate new products and cost-saving process improvements now seems less potent than it once appeared. Innovation is a cumulative process. New ideas and new production processes are commonly refined and improved over time. Broad intellectual property rights can offer rich rewards to initial innovation, but they may well impede follow-on innovation. After all, the economic incentives to improve a patented idea or process are limited when the improvement can be used only in conjunction with the original patent, as that improvement can be sold to only one buyer, the owner of the original patent. For this reason, a policy of fostering competition among owners of intellectual property could be justified solely in terms of its benefits in encouraging innovation, even ignoring consumer benefits from limiting the exercise of monopoly power.

This principle encourages antitrust enforcers to promote competition in high-tech markets with little concern that doing so will hurt the goose that lays golden eggs. Accordingly, for example, the FTC commonly negotiates curative divestitures when it believes that mergers threaten the loss of independent research streams aiming for new drug patents.

Network Effects

The third core principle is that network effects increase the concern with anticompetitive exclusion. Network effects (also termed network externalities or demand-side scale economies) are the value to a buyer when some other buyer also purchases the product or service. They are often especially important in communications. A telephone has little value if it can be used to call only a handful of numbers, but great value if everyone is on the network. Similarly, software programs for word processing or spreadsheets have greater value to a firm if its suppliers and customers use similar software and new employees need not be trained to use it.

Markets with strong network effects are frequently characterized by “tipping.” A firm that is expected to become dominant, perhaps after it achieves a small advantage in early competition, may find that new buyers disproportionately select its product. The market may quickly establish such a firm as the winner of what is effectively a standard-setting competition. Such competition is typically winner-take-most or winner-take-all. Once a market tips, it is harder for a rival to compete, even with a better product. To dislodge an industry leader may require a dramatically improved new product that “leapfrogs” the market leader’s technology.

When network effects are important, antitrust enforcers must move rapidly once they identify a competitive problem. Network effects also increase the concern with exclusion because they help ensure that the policies protecting rivals from unjustified exclusion will not reduce the pace of industry innovation. This point is relevant to understanding the procompetitive potential of government enforcement against Microsoft.

The district court judge, Thomas Penfield Jackson, found that Microsoft dominated the market for operating systems for Intel-based personal computers. Yet Netscape’s browser created potential competition for Microsoft’s Windows operating system, particularly when used in conjunction with Sun’s Java programming language. Netscape’s new product promised to make it possible for software developers to write applications programs that would access the operating system through the browser, and thus enable computer users to continue to use their favorite software after switching operating systems.

In response, according to the district court, Microsoft protected its market power in operating systems by limiting Netscape’s ability to market its browser through the two best channels for reaching computer users: as software installed with new computers or as the browser installed when signing up with leading Internet service providers like America Online. Microsoft kept Netscape’s browser off of new computers by convincing PC makers to prefer Microsoft’s own browser, Internet Explorer (IE), through exclusive contracts and by integrating IE into Windows. Microsoft also contracted for near-exclusivity with AOL and other Internet service providers. None of these exclusionary acts had an efficiency justification: Judge Jackson found that they were aimed purely at excluding Netscape and thus at preventing a long-term threat to the continued market power of Microsoft’s Windows operating system.

On the facts found by the district court, antitrust enforcement has a strong potential for promoting innovation in operating systems (and in software, like browsers, that creates potential competition for operating systems). Preventing Microsoft from excluding actual and potential rivals by arbitrary means encourages software development by those rivals. Such support for new product development by Microsoft’s rivals probably means that the research and development effort rises industry-wide: Microsoft’s own incentives to innovate may be dulled, but only slightly. After all, the “prize” to Microsoft for successful innovation-dominance of the new browser market and continued dominance in operating systems-remains extremely high because of the power of network effects. For similar reasons, tough antitrust enforcement against the unjustified exclusionary practices of dominant firms can be expected in general to promote innovation in winner-take-most high-tech markets, considering the collective incentives of the dominant firm and its rivals.

Judge Jackson’s decision upholding the government’s key monopolization charge follows from the legal rule that a monopolist violates the antitrust laws when it excludes rivals through conduct with no legitimate business justification. This rule permits a court to find an antitrust violation without need for a costly and time-consuming inquiry into harm to competition under circumstances in which competitive harm is likely. There is little danger of impeding industry innovation by applying this legal rule to the conduct of dominant firms in high-tech markets where network effects are important and innovation competition is winner-take-most. In contrast, Microsoft reads the relevant Supreme Court precedents differently, contending that proof of harm to competition is required and the company should win because, it claims, the government failed to make such proof.

This legal dispute, about the rules governing a high-tech monopolist’s conduct, will now be refereed by an appeals court that, in an earlier decision involving Microsoft, questioned the application of antitrust laws to high-tech markets. The appeals court instead emphasized how hard it is for courts to determine whether new product designs harm or promote competition, as well as the danger of chilling dominant firm innovation by subjecting product integration to judicial review. Moreover, another federal appeals court with specialized jurisdiction over patent cases has recently announced, in a case involving Xerox’s relationship with independent firms servicing its copiers, that a patent holder’s refusal to license or sell its products was close to immune from antitrust liability, regardless of anticompetitive effect. Although that conclusion has not been reached by federal appeals courts with general jurisdiction, this decision too questions the emerging intellectual consensus on the way competition works in high-technology markets and the appropriate role for antitrust enforcement. In consequence, when these two legal disputes end up in the Supreme Court, as seems likely, their resolution may well have strong implications for the application of the antitrust laws to high-technology industries that go beyond the fortunes of the two firms, Microsoft and Xerox, or the software and copier industries.

Rapid Information Exchange

The fourth key principle underlying contemporary antitrust enforcement in high-tech markets is that rapid information exchange need not create a perfectly competitive marketplace. There is much to like in Bill Gates’ powerful vision of “frictionless competition.” What could be more procompetitive than the instant and universal exchange of enormous amounts of market and product information? Information exchange can dramatically reduce a buyer’s transactions costs of search and help firms make better production and pricing decisions. In addition, Internet access and advertising could reduce the sunk costs of entry, also making markets more competitive. For these reasons, the rapid information exchange made possible by the Internet will likely in general be a strongly competitive force, helping buyers obtain better and cheaper products.

But not always. Entry in the world of electronic commerce will not necessarily be as easy as going online to create a web site. The entrant must establish a reputation, both for high-quality products and for fair business practices. Furthermore, rapid information exchange could lead to higher-than-competitive prices if it gives sellers a better way to reach an anticompetitive consensus on prices and market shares or if it makes it easy for price-fixing rivals to detect cheating on that consensus.

For these reasons, antitrust enforcers are taking a close look at business-to-business (B2B) marketplaces, to make sure that they are vehicles for lowering the transactions costs of exchange and not ways of fixing prices. They are also scrutinizing B2B exchanges to see whether they could be exploited by their members to obtain or preserve market power by keeping out entrants or aggressively competitive rivals. These concerns about exclusion may be greatest when most of the major firms in an industry own or control a B2B exchange (as opposed to merely participating in it). Widespread industry control of an exchange could also limit competition among B2B marketplaces by making it practically impossible for a rival B2B site, offering better or lower-cost services, to attract sellers and buyers. Notwithstanding these possibilities for anticompetitive mischief in the electronic marketplace, most B2B collaborations have strong potential for lowering costs of doing business and promoting competition, so will likely pass antitrust muster, many with little or no scrutiny by federal enforcers.

Business Conduct and Innovation

The final core principle shaping antitrust’s perspective on competition in the high-tech sector is that most business conduct involving innovation by high-tech firms is procompetitive or competitively benign. To be sure, antitrust enforcers pay attention to emerging markets. When today’s Old Economy industries like oil refining, steelmaking, and aluminum production were high-tech, they too were an appropriate focus for federal antitrust activity. But relative to the scope of high-tech markets in the economy, both yesterday and today, antitrust enforcement in the high-tech sector has been both measured and infrequent. Indeed, the antitrust enforcement agencies and the courts have long recognized the importance of innovation, the procompetitive benefits of most R&D collaborations, and the strong innovation record of many firms with large market shares. For every Microsoft, there are many other market leaders-like Cisco, Oracle, or Intel (which settled with the FTC)-whose market conduct does not seem to attract antitrust litigation. In short, the contemporary perspective on competition and antitrust enforcement in high-technology markets does not supplant the market. Rather, it supports the market by enlisting competition in the service of promoting innovation.

Antitrust is not dead in this high-tech age, nor should it be. Competition is vital to ensuring the continued rapid pace of innovation, and antitrust remains as essential as it always has been to preserving competition.

More