IS MICROSOFT BAD? Is Microsoft a monopoly? In the loose sense of the word, yes. Is Microsoft bad for consumers and entrepreneurs? No: Microsoft's market dominance is good for us. And the U.S. Justice Department should limit its oversight of Microsoft to enforcing existing antitrust law. In the following pages we explain why and air the two most strident voices in the debate about Microsoft's influence: Microsoft chief technology officer Nathan Myhrvold and Wilson Sonsini Goodrich & Rosati lawyer Gary Reback.

By the editors of The Red Herring

Microsoft has a monopoly in personal computing operating systems (over 90 percent of all desktop PCs run Windows, and its share of server systems is still growing)--a monopoly achieved through its early, momentous decision to license rather than sell DOS to IBM as well as the desire of software developers and users for a common software platform. Given the "increasing returns" that accrue to a standard technology like DOS (discussed below), Microsoft's OS monopoly may have been unstoppable.

But the Justice Department and many of Microsoft's competitors believe that Microsoft uses its monopoly in operating systems to beat better, competing products in other markets like browsers or desktop applications. In an industry defined by persistent progress, they think a single, dominant company is intolerable and want the government to do something about it. They're wrong.

The monopoly Microsoft's decision to license the OS to IBM for its new "home computer" may have been the most foresightful decision in the history of American business. It left Microsoft free to license the standard to other computer manufacturers. Soon most software developers wrote only for DOS--and consumers wanted the operating system that ran the most software. In time, Microsoft would use DOS to create a new market for its graphical operating system, Windows. Not many people think Microsoft's OS monopoly is itself wrong; these days most information technologies are dominated by a single company. Sun has eliminated most of the Unix vendors; Oracle owns databases; Cisco, networking; IBM, mainframe software and systems; Intel, microprocessors. Consumers, and the industry, seem to like it that way.

Nor could Microsoft's OS monopoly be considered illegal by any reading of U.S. antitrust law. Section 2 of the Sherman Act states that "Every person who shall monopolize, or attempt to monopolize . . . any part of trade or commerce among the several States . . . shall be deemed guilty of a felony."

Courts have never found monopolies that were acquired through the development of a better product, superior business practices, or luck to be in violation of Section 2 of the Sherman Act: the law bans the act of monopolizing, not monopolies. But the Sherman Act does explicitly prohibit the use of one monopoly to create another. In 1995 the Justice Department forced Microsoft to sign a consent decree, wherein Microsoft agreed that it would not require computer manufacturers to buy other Microsoft products in order to license Windows. The evolution of personal computers into communications devices made Microsoft's OS monopoly a matter of renewed concern to the technology industry. When Microsoft began to tie its Web browser, Internet Explorer, into Windows and--according to Netscape--strong-armed manufacturers into including Explorer with new systems as the price of licensing Windows at competitive rates, competitors cried foul. Last October the Department of Justice took notice: Attorney General Janet Reno took the unusual step of calling Microsoft an illegal monopolist and slapped the company with a fine of $1 million a day until it complied with the 1995 consent decree.

Even though Microsoft probably could dig up $1 million a day from under office couch cushions, company officials in Redmond were vexed. They'd been bundling features like utilities, storage, games, and spreadsheets into the OS for years; that, they said, was the platforms business. During the launch of Windows 95, the Justice Department had questioned the legality of bundling the Microsoft Network, Microsoft's online service, with the new OS--and the case had been thrown out of federal court. How was folding the browser into Windows any different? Microsoft was rolling out new products with reasonable consistency and at reasonable prices. As Microsoft chief operating officer Bob Herbold wrote in an open letter to customers last November, if the price of automobiles had dropped as quickly as that of personal computers, the average car would cost $2 and go 600 miles on a thimbleful of gas.

If a monopoly is defined by the tendency to increase prices, how was Microsoft a monopoly?

The point of many returns The Feds and Microsoft's competitors have an answer: Microsoft is a new kind of monopoly, they say, and U.S. antitrust law--developed to fight railroad and steel monopolies at the end of the last century--needs to be rewritten. To support this argument, they call upon an economic theory called "increasing returns," an idea advanced by the economists Paul Romer of Stanford University and W. Brian Arthur, among others.

The theory of increasing returns holds that the company that is ahead gets farther ahead, and companies that are behind disappear altogether. Its fundamental insight can be summed up as "the more you get, the more you get." The theory consciously reverses classical economics, which says that a monopoly will increase prices as it sees its returns diminish. With increasing returns, companies don't have to raise prices to make a profit: they make a profit by selling more stuff at less cost. Increasing returns is supposedly characterized by "path dependence," or the tendency of consumers to keep using the same technology standard, even if there are better alternatives.

Sound familiar? Increasing returns says that, in high technology, a handful of monopolies are inevitable as companies hit upon a hot idea that, for whatever reason, becomes the standard. Latecomers don't have a chance.

The implication is that technology monopolies smother innovation. Too often, the theory goes, the little startup with the great idea is either absorbed or squashed by a company like Microsoft once the titan decides it wants that market. This belief is not uncommon in the technology industry. Andrew Schulman, a senior technical editor at the consultancy O'Reilly & Associates, former software developer, and longtime critic of Microsoft, says there's no mystery anymore to what it takes to succeed in high tech. "A new software business finds a niche that it can build up so that Microsoft will have to pay to take over what the company is doing instead of simply bypassing it," he says. "That's the business plan for independent software developers today."

With the theory of increasing returns as its philosophical justification, and the alleged violation of the consent decree as its excuse, the Justice Department has begun a major investigation of Microsoft. At our press time, the government was looking into whether Microsoft threatened to withdraw Compaq Computer's Windows 95 license when it heard that Compaq planned to place the Netscape Navigator icon instead of the Internet Explorer icon on the desktop of its Presario PCs. In December a federal court was sufficiently convinced by the Justice Department's preliminary findings that it barred Microsoft from requiring computer manufacturers that license Windows to buy the company's Web browser with it; the case will be decided in May.

But the Justice Department's investigation goes far beyond the company's browser business. The government is asking whether Microsoft is trying to become a monopolistic "gatekeeper" of Internet content by cutting deals with companies like Walt Disney, Paramount, and Time Warner to make some of their content available only on Internet Explorer­enabled sites; whether Microsoft forced the Santa Cruz Operation to include code with its Unix operating system that made it incompatible with Windows; whether Microsoft engineered incompatibilities into its version of Java so that it wouldn't work with non-Windows applications; and whether Microsoft is plotting to seek a "vig," or cut, on all electronic commerce transactions.

Borrowed times Does the theory have legs? Is Microsoft bad for innovation? Many, many people say yes.

Microsoft has never been known for creating superior technology. DOS, Windows, and Internet Explorer were all adopted from other companies or organizations. And while Microsoft has always improved its offerings (consider the rave reviews of Internet Explorer 4.0), the common perception is that it lacks the vision to develop anything really new. "They put a lot of money into R&D that doesn't lead to anything revolutionary," says developer Rick Ross, president of the Java Lobby.

Has Microsoft ever used its monopoly in operating systems to stamp out innovative competing technologies? Andrew Block, Sun Microsystems' manager of competitive analysis, can offer interested parties a lengthy list of great technologies--including the DR DOS operating system, the OpenDoc component architecture, the Objective-C programming language, X.500 network directories--that he claims were eliminated by Microsoft's predatory practices.

"These technologies had advantages over Microsoft's corresponding technologies, and you can argue that they failed because of their companies' inability to promote and execute upon them," Mr. Block says. "But in many of the cases, Microsoft announced its competing products years in advance, froze out the competitors from the market, and left people waiting for their products. By the time Microsoft delivered--or sometimes didn't deliver--the products, most of the competing companies had abandoned them."

A common complaint among software developers is that Microsoft's tools do more for Microsoft than they do for developers--and, ultimately, consumers, who end up with programs made from second-rate parts. For instance, ActiveX, Microsoft's component architecture, is hard to use, but it's even harder to write to Microsoft's Win32 APIs and not use ActiveX. And many developers feel that ActiveX may not be the best possible architecture with which to build distributed applications. Dennis Van Dusen, a vice president at the systems integration company Planmatics, likens the process to being forced to buy an entire stereo system from one manufacturer instead of being able to pick and choose components: "Microsoft restricts developers to working within its platform at all times, and when you work within that you nearly always have to use the tools developed for it." Mr. Schulman of O'Reilly is more acerbic: "Microsoft supports third-party developers the way a rope supports someone about to be hanged."

The executioner's song and dance So what? What if all these allegations are true? Should Microsoft be penalized, or broken up like AT&T? Probably not.

While the central insight of increasing returns--that companies that own a standard will go from strength to strength because a standard grows more valuable to everybody as it becomes more common--is demonstrably true, the implicit prescription--that such monopolies must be regulated out of existence because they lock consumers into buying possibly inferior products forever--is nonsense.

Instead of looking to a new economic theory, the Justice Department should accept a simpler explanation for Microsoft's success: Microsoft has succeeded by selling people what they want. To argue that its OSs have been inferior but have succeeded because of "path dependence" is to argue that technical excellence is the sole criterion by which people choose products.

But both users and developers wanted a common platform more than anything else. Users wanted to know that their applications and operating systems would work well together. Users who cared more about other things--like an elegant interface or swift I/O--could buy other products, like Sun Solaris or the Macintosh. Software developers wanted to sell to the largest possible market, and they wanted the economy of a single platform. Windows developer Mike Sax, president of Sax Software, says, "Because of Windows' single standard, you can test your products more easily; in Java and Unix you have to test on 25 different platforms. Having one dominant standard makes it easier to concentrate on being innovative, instead of concentrating your resources on cross-platform testing."

If Microsoft has succeeded in markets other than operating systems, it is not because consumers are "forced" to use products like Microsoft Office or Explorer. If people like Microsoft's products because they work well with Windows, that is because working well with Windows is overwhelmingly important to consumers. Nevertheless, when a product--like Microsoft Money--has failed to meet consumers' expectations, Microsoft's OS monopoly has been no help at all.

Nor does this monopoly free Microsoft from the pressure to innovate: it's not as though we're still using DOS 1.0. In technology, change is rapid and unpredictable, and increasing returns don't last forever: a monopoly must innovate to keep its market share. IBM's monopoly in mainframes did not guarantee success in personal computers, nor did Novell's early success in LANs mean that it would control wide area networks. As yet, we don't know whether users really want a monolithic, bundled operating system and browser: Microsoft's strategy may be too closed and proprietary for the Internet.

The U.S. government doesn't know either. Every attempt by the Justice Department to regulate the technology industry for the benefit of consumers has been comically misguided. In the '60s, after Justice decided that IBM should not make both mainframes and punch cards, minicomputers and new programming languages made mainframes and punch cards less important. In 1995, when Justice tried to regulate MSN, the Internet was making MSN irrelevant. Government should generally leave consumer choices to consumers (see Jason Pontin's column "The Limits of Government").

Even Mr. Arthur, the father of the theory of increasing returns, questions the utility of regulating Microsoft. While he welcomes scrutiny of the company, he worries that government intervention will stifle innovation. "What keeps the lights burning in Silicon Valley at 3 a.m. is the possibility of locking one of these markets," he says. "I would hope the government doesn't get too restrictive."

The Department of Justice should limit its investigation of Microsoft to asking whether Microsoft violated Section 2 of the Sherman Act. Microsoft should not have threatened to withdraw Compaq's Windows 95 license, and the federal court was right to bar Microsoft from bullying manufacturers: the company is depriving consumers of free choice. But the government should not tell Microsoft what features it may include in its operating system. Justice must remember that there is no precedent for punishing a monopoly that was honestly won through good business. If anything can be learned from the theory of increasing returns, it is this: monopolies happen naturally in technology because people want standards. And since none of the evils associated with classical monopolies--high prices, bad products, diminished entrepreneurialism--accompany the honestly won technology monopoly, the government should relax and let people have what they want. By and large, Microsoft is good for us.