Federal trustbusters have spent 10 years “investigating” Microsoft, at the behest of competitors first Novell (which then owned Word Perfect and DR-DOS), followed by recent intense lobbying from AOL-Netscape, Sun Microsystems, Oracle, IBM and the venture capital firm Kleiner Perkins.
This has already been a long and expensive excursion, for taxpayers and Microsoft stockholders (which includes almost everyone with a pension). Yet we may never learn how it is, exactly, that Microsoft might have been accused of violating the notoriously enigmatic old antitrust laws. District Judge Thomas Penfield Jackson is postponing ruling on the legal issues in the hope that his chosen negotiator, Judge Richard Posner, can make that chore unnecessary by hammering out a settlement.
How could anyone presume to remedy a problem that has yet to be even clearly defined? Over the past decade, the antitrust complaints against Microsoft have changed as dramatically and frequently as the technology. Yet proposed remedies have remained the same to restrict enhancements to the operating system, expropriate intellectual property rights, or allow judges and attorneys to dismantle the company.
Of course, Microsoft’s competitors never bothered to wait for the firm to be formally accused of something before proposing “remedies.” These remedies were and are the whole point, and the ever-changing, amorphous accusations were merely a means to that end. Indeed, way back on April 9, 1998 five weeks before the Justice Department even issued the complaint a group described as “unnamed competitors” presented the press with a self-serving list of “remedies” designed to dismember or otherwise cripple Microsoft. Since then, the CEOs of such companies as Oracle and Sun Microsystems have become far less bashful, offering various suggestions about how best to tie Microsoft down.
Whatever the legal complaints might be, they are surely about Microsoft’s conduct and not about its organizational structure. Yet the biggest obstacle to any settlement has been the government’s suspicious insistence on “structural” remedies for decidedly non-structural complaints (such as bundling a browser with the operating system). The remedy reportedly favored by Justice is “horizontal” divestiture. That means one company would market only Windows. A second firm would get Microsoft’s numerous software applications. A third would get Microsoft’s relatively small Internet service, msn.com, and related web sites.
This “remedy” is amazingly irrelevant to any alleged offenses. The newly separated Windows company would still have exactly the same high market share for Windows 98, and the same modest share for Windows 2000 and Windows CE. The applications company would still have the most popular office suite, and would still struggle bravely against Quicken and TurboTax in the market for financial software and online services. The newly severed Internet company would become even punier and less effective in comparison with the AOL-CompuServe-Netscape-Time Warner behemoth.
If the problem is said to be a monopoly of Windows 98, creating a separate Windows company would do nothing to change that. In fact, any company or group of companies that sold nothing but Windows would be in a much stronger position to charge a higher price, since they would no longer need be concerned that doing so would shrink the market for applications.
Even if the judge tries to recycle the government’s quaint 1995 anxiety about Windows and Internet Explorer being marketed in the same package, divestiture does nothing about that either. A separated Windows would still have no choice but to include Internet browsing as do BeOS and Linux operating systems, the free Star Office suite from Sun Microsystems, Palm Pilots, Nokia phones, AOL-TV, and powerful new “game” computers from Sony, Sega, Nintendo and (barring antitrust obstacles) Microsoft.
The only rationale for breaking-up Microsoft along product lines comes from the software lobby, which claims intimate knowledge of Windows is what gives Microsoft an edge in applications (an obviously handy excuse for those selling second-rate software). Unfortunately, not one shred of evidence was presented at the trial to substantiate this paranoia.
In short, the government’s threatened “remedy” separating Windows from applications is entirely unrelated to any complaint aired at the trial. There is simply no link between the remedy and the problems ostensibly in need of a remedy. That means this “break-up” remedy itself could not possibly survive appeal, even if all alleged breaches of the law were somehow upheld. And that is why government trial lawyers and their accommodating judge would prefer to settle rather than have their bluff called and end up with egg on their face in Appeals Court which is what happened after far less ambitious previous antitrust rulings against Microsoft.
In theory, however, there is an entirely different way to break up the company along vertical lines slicing Microsoft into three supposedly identical “Baby Bills.” The cute phrase hopes to create an illusion of similarity with the breakup of AT&T, but that is absurd. The Baby Bells were regional monopolies that were not allowed to compete with each other nor with AT&T’s long distance service. Nobody is proposing that those who live in the East will be forced to buy Microsoft products from “Bill Atlantic,” or that those living in the South will have to deal with “Bill South.” The analogy is fatuous. And far from being an “alternative to regulation,” as some now pretend, the break-up of MaBell resulted in one judge having to rule on nearly everything any phone company did, for years and years.
Another minor problem with the Baby Bills approach is that it is literally impossible. Microsoft’s main value is organizational and intellectual capital teams of uniquely knowledgeable people, not buildings and machines. There is no way to equally divide the immeasurably superior talents of particular managers and researchers, nor the unknowable future value specific patents and contracts. If anyone tried to break Microsoft into three “equal” parts, one would surely be best, one worst, and one mediocre. And the sum of the parts would be considerably less valuable than the synergy of the whole. This is not like carving a cake, or picking the members of a high school football team.
When Standard Oil was broken up, that just meant Standard Ohio got the refineries in that state, Standard Indiana got the Indiana operations, and soon. Microsoft is almost entirely in one place. Would divorcing applications researchers from operating systems researchers mean putting up a fence between them and cutting-off all e-mail? Would employees of a severed portion of Microsoft be expected to relocate to some other city or state? It is all just a ridiculous idea, reflecting excess arrogance and inadequate reflection.
Since dividing Microsoft horizontally is irrelevant and dividing it into equal pieces is impossible, trying to do both must be impossibly irrelevant. Yet a former Federal Trade Commission economist, Thomas Lenard, has dreamed up a “hybrid” scheme that would slice and dice Microsoft from both directions.
After separating Windows from everything else, Mr. Lenard would then have three companies offering the three existing versions of Windows. What’s the point? Either the three versions of Windows 98 would remain essentially the same or they would become significantly different. If they remain the same, this would be an extremely painful way to accomplish nothing. If they instead differentiated their products to gain market share, then the various version of Windows could end up as incompatible as the many versions of UNIX including Linux, Sun’s Solaris and Apple’s forthcoming OS-X system.
Actually, one version of Windows would surely become superior by offering new features (voice recognition) or better performance (a faster browser, fewer crashes). Software reviewers would then declare that version the best, and the other two would go out of business. No computer manufacturer would dare install a second-rate version of an otherwise similar operating system, and no software developer would bother to write software for a second-best version of Windows.
The disconnect between dozens of petty legal complaints and these sweeping “remedies” is nothing new. Antitrust is a big game with big stakes but no predictable rules. Remedies are often approached as an afterthought, within cautious hubris, for reasons eloquently explained by the mediating judge, Richard Poser, in his 1976 classic, “Antitrust Law: An Economic Perspective”:
“The initiative in the [Antitrust] Division lies with the trial lawyers along with the execution, the theorizing, the design of remedies, and virtually every other aspect of the enforcement process. Trial lawyers tend to combative rather than reflective, and the division’s trial lawyers … tend to be young or mediocre, or to be zealots.”
Asking a team of lawyers to reassemble one of the two most valuable companies the world (is Cisco next?) has ever seen would be very serious business indeed. But it was never a serious “remedy” for any describable problem. Irrelevant and unworkable plans for cutting Microsoft into various pieces popped up without their proponents feeling obliged to explain how such proposals might be expected to “remedy” even a single violation of antitrust law, actual or imagined. These “remedies” have a life of their own.
Proposals to smash or otherwise hobble Microsoft preceded the trial, and even the complaint, although we have not yet been told which laws may have been broken, and how. Perhaps this case was never really about what Microsoft might have done, but about what might be done to Microsoft.
Alan Reynolds is director of economic research at the Hudson Institute and senior editor of the Institute’s quarterly, American Outlook.