Deja vu

Quite a story. But where does it leave us? And perhaps even more impor­tantly, where does it leave Microsoft? Quite possibly stronger than ever. Sure, Microsoft incurred bad publicity and monstrous legal fees.

Microsoft hit the ground running in late 2001, almost immediately after reaching its agreement with the DoJ. Before the year was out, Microsoft’s Windows XP evolved the translation frontier upward to subsume a media player and an instant messenger, and it launched several initiatives to reshape the Internet in its own image: Passport to ease e­commerce; the ambitious if somewhat amorphous.Net initiative that combines “a set of Microsoft software technologies for connecting your world of information, people, systems, and devices”;³⁷ and a “Hail­storm” of announcements about forthcoming products and services. The company made a major push into the server market, where it tried to leverage its way from the desktop into the back office—a computing environment that remains competitive. Microsoft also reportedly issued a series of threats: to impose a permanent cost disadvantage on corpo­rate customers who failed to upgrade to Windows XP—and its coun­terpart in the applications world, Office XP—on Microsoft’s schedule, and to tell consumers that all third-party software was incompatible with XP unless the developers gave Microsoft a copy of their source code. It’s unclear how many of these announcements and/or threats Microsoft ever carried out, but their combined effect demonstrates the extent to which the company felt unshackled by its agreement with DoJ.

But of all of these initiatives, one stands out. The Windows Media Player (WMP) “integrated” into Windows XP demonstrated just how much Microsoft had learned in the temple of network economics. Think back for a moment to what actually happened in the browser wars, and to why it was so important for Microsoft to win them. By 1995, Windows was secure as the translation frontier between human users and their desktop hardware. Netscape and Sun had created versions of their products that ran on top of various platforms—Windows, Mac OS, Unix, Linux, even DOS—and that made it pleasant and easy for people to use the Internet. All of a sudden, human users had two different reasons to talk to hardware, and two different interfaces through which to talk. When they wanted to do business- or office-like tasks, they’d ask Windows to translate. When they wanted to use the Internet, they’d ask Navigator to translate. People could suddenly talk to their hardware without a Microsoft translator. That made Microsoft nervous.

What made Microsoft even more nervous was the thought that some folks using Netscape Navigator’s interface on a “thin client” or an “Inter­net appliance” that wasn’t set up to do office work might want to access an Internet-based word processor or spreadsheet. If that happened, not only would people be able to talk to their hardware without Microsoft, but they’d be able to do office-like tasks without using either Microsoft’s Office suite or Microsoft’s platform. In other words, competition could reemerge in both halves of the translation chain; new applications closer to humans could sit atop Navigator, and new platforms or utilities closer to hardware could create parallel paths from interface to microchip. That threat—no matter how remote—was untenable, so Microsoft stopped it.

Perhaps the most negative consequence of Microsoft’s behavior in the browser wars was thus that it rendered this parallel translation frontier stillborn. And so today, instead of having two competing translation frontiers with enough different features to reveal which were best for which tasks, we have one monopoly frontier: Microsoft’s integrated Windows/Internet Explorer.

Microsoft integrated WMP into Windows for a similar set of reasons. Before WMP, most people used the industry open standard of MP3 to encode music files. But Microsoft created its own proprietary format, WMA. While WMP will play MP3 files, it will only rip WMA files. The average user who buys a new computer with WMP already installed is likely to develop a large collection of music files encoded in a format that only WMP can play—threatening competing products from Real Networks and Apple. Leveraging the platform outward, embracing and extending, wrapping a technology developed elsewhere into Windows and claiming it as Microsoft’s own—we’ve seen it all before. Microsoft learned its lessons well. Many of the initiatives launched with Windows XP do the same thing. Its integrated instant messenger threatened AOL and Yahoo!, and Passport was created to translate all of our shopping needs to the hardware.

And that was all before Kollar-Kotelly approved the agreement in November 2002. But as we’ve seen, antitrust inquiries rarely fade into oblivion. In June 2003, the remaining states, Microsoft, and various amici (scholarly “friends of the court”) filed their briefs appealing Kollar- Kotelly’s ruling. At about the same time, Microsoft settled its lawsuit with AOL—a suit that AOL had brought seeking compensation for the damage that Microsoft had inflicted upon its subsidiary, Netscape—and these two giants of the information sector promised to play nicely with each other. That agreement left some of AOL’s former playmates out in the cold; Real Networks’s stock dropped roughly ten percent the day after Microsoft and AOL announced their deal.

Today Microsoft is both better armed and better informed about the power of monopoly leveraging than it was a decade ago, and its oppo­nents are spent, disarmed, discouraged, and looking about for help wher­ever they may find it. One of these days, we’ll start calling the events to date “the first Microsoft trial,” because it’s almost inconceivable that there won’t be a second. Microsoft retains control of a bottleneck monopoly and a large staff now well versed in monopoly leveraging. The company will retain antitrust lawyers who make sure that it complies with the letter of Judge Kollar-Kotelly’s order while likely working hard to violate its spirit. And this, despite her explicit admonition that:

During this litigation, promises have been made on behalf of Microsoft that the company will change its predatory practices which have been part of its com­petitive strategy in order to comply with the remedial decree. The Court will hold Microsoft’s directors, particularly those who testified before this Court, responsible for implementing each provision of this remedial decree. Let it not be said of Microsoft that “a prince never lacks legitimate reasons to break his promise,” for this Court will exercise its full panoply of powers to ensure that the letter and spirit of this remedial decree are carried out.³⁸

But the temptation to be Machiavellian is just too great. The trial has done little either to reduce Microsoft’s capabilities or to change its incen­tives, and leveraging up to the legal limits constitutes highly rational behavior for a powerful monopolist.

In case you think I’m kidding about all those trials, I have three words for you: United Shoe Machinery.³⁹ Several competing manufacturers of shoemaking equipment merged to create this monolith of the manufac­turing age in 1899. Twelve years later the government realized that United Shoe Machinery was an abusive monopolist and went to court seeking its breakup. It took until 1918 to reach the Supreme Court, which refused the government’s request. Almost thirty years later the government tried again, and spent the better part of six years in court seeking its breakup—but achieved only limited behavioral relief. But 1953’s behavioral relief proved inadequate. The Supreme Court ordered United Shoe Machinery broken up in 1968—about fifty-six years after the government’s first complaint. On the United Shoe Machinery clock, then, we should break up Microsoft around 2050—unless antitrust law kicks into Internet time.

But whenever those future Microsoft trials occur, the new judge will face the same challenges that plagued Judges Jackson and Kollar-Kotelly. Structural remedies will continue to be tough to implement, and behav­ioral remedies will never promise more than unlikely prospects for effec­tiveness. And Microsoft is counting on that. Microsoft may accept minimal restrictions unlikely to have much of an impact on the way that it does business, but it will never accede to a fundamental change.

That inherent quandary facing market cops and courts alike reveals, once again, the relationship among technology, law, and economics— and stresses the importance of diagnosing problems correctly and of selecting appropriate tools before trying to fix them. The market cops detected a problem in the software markets; they traced that problem to Microsoft. But Microsoft’s anticompetitive actions were an effect; no one spent much time seeking the deeper cause. Microsoft behaved as it did because it could, and because it was rational to do so. It could because we gave it strong IP protection on its software without forcing it to promote human knowledge. The technology of compilation made it pos­sible to circulate object code while keeping the knowledge-laden source code secret. The law of copyright gave that object code immense value. Economic incentives simply dictated that Microsoft use its rights to max­imize corporate profits.

The market cops then tried to use antitrust remedies to fix the effect while leaving the cause untouched. It’s hardly surprising that they failed. Perhaps next time they’ll look beyond effect, to cause. The only way to fix the markets that Microsoft has broken—and will continue to break if it behaves rationally—is to change Microsoft’s incentives. The only way to change Microsoft’s incentives is to change its powers and the rights that underpin them. And the only way to change those rights is to realize that they are, at heart, IP rights, and that we need to seek advice from the priests of IP. Perhaps we should have paid more attention to their jeremiad.

And so, the trial, like the bubble, was fundamentally a tale of misdi­rection. Internet investors chased network growth without considering lock in. Market cops chased antitrust remedies without considering IP rights. Both tales were doomed to end poorly—at least for the general public. We need to rethink both of these stories in a new light—a light that shines more brightly over the open-source bazaar and the song of music. These stories demonstrate more clearly the ways that the infor­mation sector’s reduced transaction costs let consumers pay less and pro­ducers sell more—but that also render traditional distribution channels less lucrative, and thereby motivate traditional distributors to try to reim­pose the transaction costs.

That framework allows us to recast both the bubble and the trial. The bubble collapsed because distribution revenues could never materialize without lock in, and dot-com intermediaries could thus never become profitable. At trial, we learned that Microsoft’s violation of the principle of modularity (among its other transgressions) choked its rivals’ distribu­tion channels and thereby raised their transaction costs. Microsoft sub­verted its role as a software developer to its more lucrative role as a software distributor. The dot-coms lost because they had no transaction costs to impose; Microsoft won because its IP rights and network barri­ers to entry enabled it to reimpose transaction costs on everyone but itself.

That new framework also points toward the information sector’s future. The bubble’s New World paradigm gave way to more sober New Channel thinking, while Microsoft’s continued anticompetitive behavior may stir policymakers to consider the messages of IP reformists. And these reassessments will inform the key debate over industrial policy in an information age—the debate over transaction costs. Where do we want to reduce them to benefit consumers and producers? And where do we want to impose them to serve other important public-policy objectives?

Those questions will shape the future, as we continue our transition into the information age, and as the information sector continues to absorb large swaths of the economy. If we navigate this transition suc­cessfully, the information sector will make us rich one transaction at a time. And if we don’t, things may remain as they are: an information sector whose only inevitable monopolist is Microsoft.