Commentary

How Not to Define Business Monopoly

This article was published in the Washington Times, November 10, 2002.
The 1998 Microsoft antitrust case seems finally settled, but it is hard to be completely sure about private suits, the states and the European Union. One reason we cannot be entirely certain this foolishness is over is that the Appeals Court judges never took a close enough look at the original “findings of fact,” particularly the abuse of market share figures to prove monopoly. Since that label stuck, it still invites legal harassment.

Judge Thomas Penfield Jackson’s endlessly repeated claim that 95 percent of the world’s computers run on Windows, for example, required pretending that Apple, Sun Microsystems and Palm do not compete with Windows because their computers are not “Intel compatible.” That is like saying Campbell’s has 95 percent of the soup business if we just leave out Progresso because their soup styles are Italian, Healthy Choice because their soup is dietary, and all other soups because they are dry mixes, imported or frozen.

The technological speculations behind this case were as fanciful as these phony facts. The government claimed Microsoft feared Netscape in 1994-95 because they supposedly believed some sort of mysterious software tailored to require Netscape would run on any operating system, because Netscape was available for Mac and Linux.

In reality, the actual threat was always the Internet itself, which can often substitute for desktop software. Microsoft worried that Netscape’s portal had quickly become the gateway for nearly all eyeballs looking to get on the Web. Left unchallenged (by Yahoo as well as Microsoft), a Netscape monopoly might have excluded Microsoft from Internet content and services. Such issues get a bit technical, but are fully explained in a book I wrote last year, “The Microsoft Antitrust Appeal.”

Those who remain determined to call Microsoft a monopoly keep fabricating meaningless market share figures. The Washington Post recently published graphs purporting to show that Microsoft’s share of operating systems is 92.5 percent and its share of office suites is 93.9 percent.

But note that those figures were only for software “sold over the counter,” which is the exception not the rule. Most operating systems are purchased with a computer, not separately. Try buying a copy of Sun’s Solaris 8.0, Palm 5.0 or even Mac 10.2 at your local computer store. Companies that do not sell software over the counter are indeed likely to have a small share of the over-the-counter software sales.

Office suites too generally come with a new computer. In the case of Dell, Gateway, HP and many others, the new computers usually come with Corel Office Suite, not MS Office. A few come with Lotus Smart Suite (IBM and Polywell) or Sun’s Star Office (eMachines). In any case, monopoly means consumers have no choices, not that most choose MS Office or Windows. In hand-held computers, the fact most people choose Palm over Windows does not mean Palm has a monopoly.

Consider all the new non-Microsoft choices that have sprung up since the 1998 Microsoft trial ended:

First, Linux now runs about half of all Internet servers and is making inroads into desktops. Wal-Mart sells cheap personal computers that run Linux with a Windows-like look, including the blatant ripoff of “Lindows.”

Second, the line has blurred between a laptop computer and other portable devices. Best Buy runs ads describing a Palm Pilot with a folding keyboard and Internet access as the ultimate lightweight laptop. Other such laptop replacements have the keyboard built in. AlphaSmart’s Dana is a “Palm Powered laptop alternative.”

Nokia Communicator uses a Symbian operating system to handle word processing, spreadsheets, e-mail and more. The Linux-based Sharp Zaurus does all that on an even smaller scale. When the government added up the “relevant market” in which Windows laptops had to compete, however, it failed to include any of these handy devices. In fact, the “relevant market” did not even include Apple’s popular laptops.

Third, the distinction is vanishing between “personal” computers and “workstations” for science and business. IBM ads ask, “Is it a computer or a workstation? Who Cares?” Indeed, that distinction has become increasingly arbitrary as personal computers became more and more powerful. Although the government did add workstations to its count of “personal” computers running Windows, it totally excluded the many workstations produced by Sun Microsystems and Silicon Graphics. Why? Because those workstations, like Apple’s computers, do not use Intel microprocessors.

But to define computers out of existence because they use Motorola or SPARC microprocessors was an absurd definition of the market. Sun knows very well that both its Sun Blade desktop workstations and SunRay “thin clients” compete directly with Windows-based alternatives. Yet the “relevant market” that Microsoft is still accused of monopolizing has been somehow defined to completely exclude Sun, Apple and a rapidly growing array of lightweight devices powered by Palm, Linux and Symbian operating systems.

If we are going to accept the way the government and many reporters have defined Microsoft’s market share, then we might just as well claim Gillette has a huge market share in razors if we leave out electric shavers and exclude BIC because they are disposable. And we might as well claim Jell-o has a monopoly of cold desserts, if we leave out ice cream and pudding.

The government’s effort to label Microsoft as a monopoly was founded on a blend of technological fantasy and statistical fraud. It is terrific that the case is settled, but too bad the fraud continues.

Alan Reynolds is a senior fellow with the Cato Institute and a nationally syndicated columnist.