Rear Windows: The software market is changing faster than the Clinton Administration can regulate it.

Mr. Reynolds is a senior fellow at the Hudson Institute and senior editor of the Institute's journal, American Outlook (www.hudson.org). This article is adapted from a study prepared for the National Center for Policy Analysis.

WHEN the Federal Trade Commission first began investigating Microsoft for antitrust violations in May 1990, Windows 3.0 did not yet exist and the Internet was an obscure academic curiosity. The Justice Department took over in 1993 and has changed complaints so frequently that we are apt to forget which shell hides the pea.

The court battle relies heavily on implausible gripes from Microsoft rivals. America Online, the Internet-service provider, feigned fear of Microsoft Network even though nearly ten times as many subscribers use AOL and its Compuserve arm to dial up the Internet as use MSN. A claim by the CEO of Oracle that Digital Equipment had to be "pressured" to abandon the bad idea of marketing $500 network computers was even more ludicrous. In any case, all the insinuations about secret deals that somehow never reached completion are beside the point. The government has not charged Microsoft with attempted collusion. And terms like "pressuring" and "bullying" have no meaning in antitrust law or economics.

Allegations of a "predatory pattern of conduct" are no clearer. Microsoft is said to have "preyed" on Netscape, maker of a popular website browser. (A browser is what you use to call up a site.) Microsoft has given away its own browser, Internet Explorer, as part of its operating systems, Windows 95 and Windows 98. The courts have tried to define "predatory" as shoving a competitor out of business without any efficiency gain. Yet having two choices of popular browser at a price of zero is clearly more efficient than having one practical choice (Netscape) at a higher price. And Netscape is stronger than ever, now that it has America Online's wealth and wide exposure. "Predatory pricing" means lowering prices temporarily, then putting them back up after competitors are dead. But nobody, least of all the prosecutors, is claiming that Microsoft ever intends to charge any price for its basic browser, much less a steep price.

The real complaint is that including Internet Explorer in Windows involves a tie-in sale, or "bundling." The worry seems to be that Microsoft might gain a monopoly of browser use and then use that monopoly to direct docile consumers toward the company's favored Internet-service provider (MSN), search engines (which direct you to websites that include the keywords you're looking for-Microsoft didn't have one before October), and destination sites (Microsoft Expedia).

This scenario is speculative, even fanciful. Indeed, the government's case against Microsoft is incoherent at every turn, a textbook example of antitrust law desperately chasing after a rationale. That case simply misunderstands both the economics and the technology of the market it is dealing with. If it reveals any anti-competitive behavior at all, it is on the part of Microsoft's rivals.

With Windows, Microsoft is said to have a monopoly of the operating-system "market." But this is an odd sort of market. Rival operating systems are either bundled with the computer (Apple, Sun, and IBM mainframes) or downloaded for free (Linux and Free BSD). News reports echo the Justice Department's claim that Windows is on 90 per cent of all computers. What the written complaint really said, however, is that Windows has recently been installed on 90 per cent of new "Intel-based" personal computers. The expression "personal computers" excludes business computers. And "Intel-based" excludes Apple, even though Apple's iMac has been one of the two best-selling computers.

In any event, the relevant "market" is not operating systems per se, nor browsers, but the market for all software services. An operating system like Windows or that of Apple is just a means to an end. It gets you into an application on your hard drive or network. Or it gets you to a browser, which then lets you use search engines to find websites.

Those accusing Microsoft of leveraging Windows have also been saying that programs written in JAVA, the programming language created by rival Sun Microsystems, will soon make operating systems like Windows irrelevant because these programs will run well on any operating system. Computer users would be able to use browsers to download JAVA programs and run those programs from a network on the Internet. Networks would replace the personal computer. Hence, the Justice Department's central theme-that Netscape's dominance of the browser trade, if left unchallenged, might have provided a viable alternative to any operating system. The Appeals Court, however, properly dismissed this idea as a "prognostication." In fact, neither JAVA nor Netscape is a substitute for an operating system, and neither are Netscape and JAVA combined. Netscape could have rendered an operating system unnecessary only if network computers had been a viable idea. When personal-computer prices came way down, the dream of $500 network computers soon became an illusion. People are not about to give up the independence of having their own word processor and other applications.

Yet the stillborn network computer continues to haunt the Microsoft antitrust suit. That is because it is the only credible way in which Netscape's browser might have appeared to be a serious alternative to Windows. Even in this sense, the notion was illogical. The brand name on the browser has nothing to do with what can be downloaded from millions of websites-including JAVA applications that would run on any operating system. Internet Explorer could do the job, as could browsers made by Sun or Oracle.

Eventually, the distinction between operating systems and browsers may vanish in favor of a single system that operates online or off. That is the direction in which Windows 98 is aiming, but Microsoft rivals are also headed there. The Wall Street Journal reported, April 1, that IBM and Sun are working together to build a new JAVA-based operating system. That could be a real threat to Windows. Software sales of IBM were nearly $13 billion in 1997, slightly larger than Microsoft's. Sun too is nearly as large as Microsoft ($9.8 billion). Any new IBM-Sun operating system will surely "bundle" a browser from Sun, Lotus, or Netscape. The new Macintosh system cannot be counted out either, and Apple has bigger plans in the works.

The Internet itself is also a serious threat to selling any new operating system. For web surfers, an elegant operating system is irrelevant: Windows 3.1 (bought years ago) or Linux (which is free) are quite adequate for surfing purposes. But Microsoft and Apple cannot make money if consumers stop trading up to their latest software. So, those companies are compelled to continually add new and better features to their latest bundles. Apple, for example, is marketing upgrades to its operating system on the basis of some novel features for searching the web.

Meanwhile, there is no meaningful "browser market," much less one that could be monopolized. Monopoly means the ability to charge a higher price, without concern about existing or potential competitors. It is not possible for anyone to sustain a monopoly of browsers and then use that monopoly to charge a higher price. Entry into the browser business is too easy. The first browser, Mosaic, was designed by a lone grad student. Today, there are at least 35 browsers for Windows alone, according to www.browserwatch.com. Basic browsers have usually been free. Netscape was the rare exception, for a while. The Department of Justice did not object to Netscape's charging a sizable fee for its basic browser before 1998, but does object to Microsoft's offering a similar product for free. Justice did not object when Netscape accounted for more than 80 per cent of browser use a couple of years ago, but does object that Microsoft now has half as large a share. These curious distinctions seem to reflect politics, not economics.

In early 1998, Netscape began giving away its source code to encourage the creation of more browsers. Microsoft also encourages others to build customized variations of Internet Explorer. Companies with customized browsers use them to advertise favored websites, as Netscape and Microsoft do. America Online uses an Explorer-based browser to promote Preview Travel (not Microsoft's Expedia) and Digital City (not Microsoft's Sidewalk). Airline credit-card companies also push their own online travel services, offering frequent-flier miles. And NBC-TV ads ask us to use their SNAP! search engine.

Such competition in online marketing is hot and getting hotter. The largest computer manufacturers will soon be using special keys or icons to direct customers to favored sites. Search engines already promote commercial sites with the subtlety of billboards in Times Square. So do websites themselves, with banners, pop-up windows, and recommended links.

To suggest that Microsoft could possibly have any sort of monopoly over this frenetic Internet marketing is preposterous. So is fretting about any "unfair" monopoly over browsers. If computer magazines ever announce that a new version of Netscape or some other browser is superior to Explorer, serious Internet users will defect from Explorer as quickly as they dropped Mosaic for Netscape.

A RED HERRING 

Another intellectual prop for the Microsoft case is the "new" economic theory of "network externalities" and "lock-in." Computer users naturally gravitate toward a common operating system, for example, because it makes it easier to communicate with other computers, and to exchange files. Whatever relevance this concept may have for operating systems or word processors, it has nothing to do with browsers. Smith's capacity to make use of the Internet is not even slightly affected by the fact that Jones uses a different browser, because web developers take care to ensure that their pages work well with all significant browsers. Windows browsers are easy to use and equally easy to discard in favor of another. Network externalities are a red herring.

In short, the only economic concept that is actually related to the core complaint is not collusion, predation, or network externalities, but "tying" or "bundling." A tie-in sale means forcing customers to buy one thing if they want another. If you want an Apple operating system, you have to buy an Apple computer. If you want an Apple computer, you have to accept the Apple operating system. For Microsoft to market Windows and Internet Explorer together is not at all comparable. It is not a tie-in sale because it is not any kind of sale. Internet Explorer is free. You don't have to buy Windows 98 to get Internet Explorer. Just log on to microsoft.com and download it for your old version of Windows, or for Mac. Conversely, Netscape's browser works well with every system, including Windows 98.

The closest the Microsoft suit has come to any tangible example of tying is an accusation that Microsoft threatened not to make an Apple-compatible version of MS Office unless Apple adopted Internet Explorer (Apple users are, of course, free to switch). Whatever merit that accusation may have, it has nothing to do with bundling Internet Explorer with Windows. Apple computers do not use Windows.

HEADED TOWARD ZERO

How can Microsoft be said to be forcing us to buy Explorer for free? How does the easy availability of Explorer make it any harder than it was before to download Netscape Navigator? Trying to make some sense of all this, I searched the index of Robert Bork's opus, The Antitrust Paradox, and looked up "tying." Judge Bork (who is working for Netscape) cites only two experts on this key topic, the first of whom is M. L. Burstein, in a classic 1960 study. Meyer Burstein holds Chicago doctorates in law and economics, taught at a dozen universities, and authored a dozen books. He seemed like the perfect person to clear things up.

So, I pleaded for his comments. Burstein seized on an observation from Scott McNealy, chairman of Sun, that the marginal cost of selling another unit of software is close to zero. Floppy discs are dirt cheap and computer downloads even cheaper. Burstein pointed out that near-zero marginal costs makes bundled marketing of software inherently efficient, even inevitable. Whenever marginal cost is nearly zero, competition should and does tend to drive prices toward zero. Mass-market software can command a significant price only briefly, while it is unique. Efficiency therefore requires lump-sum pricing for competing "bundles" of software services. Since competition drives the price of each separate software feature down toward zero, the only way a software company can make money is by offering increasingly attractive bundles of applications and features. This is why previously separate word processing, spreadsheet, and database applications are now bundled into much cheaper "Office Suites," and why many previously separate "utilities" have been bundled into packages.

Wendy Goldman Rohm's book, The Microsoft File, inadvertently reveals the real reason many Microsoft rivals prefer to compete in the courtroom. She navely echoes complaints that "Gates historically has slashed prices or given products away" and that "Microsoft offered predatory low prices."

This would not be the first time that antitrust law was deployed to protect consumers from terrific products that are given away or sold at low prices. Robert Bork's book noted, "Antitrust law has never clearly defined what it means by predation. . . . Unsophisticated theories of predation abound, leading to drastic overestimations of its likelihood." By contrast, he observes, a very effective "predatory strategy consists of misuse of the courts or other agencies of the government."

Could Microsoft's rivals be using antitrust agencies and courts for "delaying the appearance of new competition"? The next network version of Windows, Windows 2000, threatens to "bundle" many business-software features that other firms have heretofore sold separately for steep fees. Unlike Windows 98, which is aimed at home and small-business computers, Windows 2000 could take business away from such business-software giants as Oracle, Sun, and Novell.

In a packaged-software market estimated by BusinessWeek at $133 billion (not counting downloaded software), Microsoft's sales have been relatively small, less than $13 billion in 1997. If Microsoft could be prevented from including a browser with Windows 98, this precedent might later be used to thwart or stall what features can be included in Windows 2000. The big issue is not Netscape, but who gets to define the Microsoft product-the company or its rivals.