I strongly support the antitrust laws and believe that the general principles established by our antitrust laws are of great and abiding importance toward ordering a competitive economy.
One of the most important developments of modern times is globalization--the increase in international competitiveness.1 An increase in competitiveness-in competition-of course, is the objective of antitrust law. Thus, we might think of the increase in global competitiveness as an alternative to antitrust enforcement. The more general the competition and the more serious the global competition, the less we need to rely on antitrust prosecution to maintain competitive conditions in American markets. As a consequence, our antitrust policies should focus, as Chairman Pitofsky mentioned,y more on securing and maintaining ease of entry into American markets than on refining our economic analysis of the potential effects of very specific industrial practices-and surely more than on fine-tuning industrial structure.
Perhaps more importantly, in many modern contexts we must think about antitrust enforcement in new and different ways. We must adapt our interpretation of the antitrust laws to novel industrial situations and to modern industries that operate much differently from the industries that our antitrust laws have regulated in the past. We must recognize that high tech industries have a much different base and operate differently than the manufacturing industries with which we have had the predominant experience in antitrust enforcement over the last one hundred years. Compare, for example, a prototypical manufacturing industry such as steel to a modern informationbased industry such as computer software. At its creation in the early years of this century, U.S. Steel had a market share of roughly ninety percent.3 Competition, as we would expect, eroded that market share, but did so very slowly, at a rate of roughly ten percent per decade.4 Therefore, fifty years later, even U.S. Steel still remained by far the dominant firm in the American steel industry. Quite in contrast, in the computer software industry in the years since 1980, a period of less than two decades, we have seen three separate instances of industrial dominance-IBM, Apple, and now Microsoft.5 Moreover, these revolutions have been complete, not partial, as in the case of U.S. Steel; that is, there remain very few people who still use DOS or Macintosh exclusively. They are period pieces and of little market significance in the modern software market.
Because market power in high tech industries stems from a different source than market power in manufacturing, we must address antitrust enforcement differently. Market power in manufacturing derives from the physical possession of property: from the possession of a dominant set of plants such as by U.S. Steel, or from the possession of dominant natural resources such as by Alcoa.6 These physical resources are not easily duplicated so that it takes a substantial period of time-any decades--to develop competitive alternatives. In contrast, market power in high tech industries derives from information, or more importantly, from the configuration of information through a network.7 It is not simply that idea A is superior to idea B, but that membership in network A provides a broader set of advantages (for the moment) than membership in network B.
Computer software is appropriately viewed as a language. It is not surprising that a language such as Windows commands a very high market share. There are many benefits from mutual communication in a language. The benefits of language do not derive from the individual efficiency of any one component of the language. Thus, we are not going to understand the sources of the market share attending a language by asking, as we do in other antitrust contexts, whether one component of that language, such as Internet Explorer, is more efficient than or superior to an alternative potential component, such as Netscape's Navigator. …