Commentary: What A Microsoft Deal Should Look Like

The Microsoft trial isn't over yet, and the current talks between the company and government prosecutors could lead to a settlement. Still, the issue of what remedies should be employed to end alleged Microsoft abuses is being hotly debated. The state attorneys general, for example, reportedly favor forcing Microsoft to auction off rights to its crown jewels--the source code and user interface of the Windows operating system. The intention: to get this technology in the hands of others as well as Microsoft and create competition where there is none.

When the trial began, most antitrust economists viewed such a "structural" remedy, which would dramatically change the nature of Microsoft Corp., as far too radical. They believed that the eventual remedy would focus on "conduct"--that is, some tough package of restrictions on Microsoft's behavior, such as limiting what functions can be included in the operating system. That would be consistent with the past two decades of antitrust policy, which has concentrated primarily on policing bad conduct rather than on breaking up monopolies.

BAD PRECEDENT. But I am among those who are reluctantly coming to the conclusion that the economy, the computer industry, and perhaps even Microsoft's shareholders would be better off with a carefully constructed structural remedy. As the trial has unfolded, it has become clear that a conduct remedy strong enough to control Microsoft's behavior would have to be tightly supervised and monitored for years to come. Indeed, trustbusters wound up launching the current suit because Microsoft basically ignored the 1994 consent decree curbing its behavior.

But government supervision of Microsoft would set a bad precedent. Since Microsoft has its finger in virtually every high-tech pie, any monitoring of it would necessitate the regulation of its relations with other companies as well. The result would start us down a slippery and deadly slope to government micromanagement of the high-tech sector.

By contrast, a well-crafted structural remedy would have the virtue of minimizing the continual government interference, which might slow innovation. One of the guiding principles of the New Economy is that regulators should set ground rules, then step away. Indeed, freedom from heavy-handed regulation has been one of the key advantages the U.S. has over Japan and Europe.

There are other structural remedies besides the licensing idea. One would break Microsoft into two or three "Baby Bills," and then set them free to compete in all Microsoft's businesses. Another would horizontally divide the operating-systems and applications sectors. Such actions would follow precedents set by the breakups of both Standard Oil in 1911 and AT&T in 1984.

In Microsoft's case, the most logical approach would be to deal directly with its source of power: the Windows monopoly. Selling off the intellectual property rights to Windows would allow competitors, for example, to sell smaller versions of the operating system, or package them with other applications. Or Windows could be put on top of Linux, making it a far more user-friendly operating system. If Microsoft and its Windows rivals stick to a common programming interface, and they have every incentive to, they'll avoid dividing the standard that has served customers so well.

With a structural remedy in place, Bill Gates would be able to do what he says he most wants--integrate anything he chooses into the operating system. Given the ferociousness of Gates as a competitor, losing the Windows monopoly could very well unleash a powerful new groundswell of profitable software innovation by Microsoft. By contrast, effective conduct remedies hurt shareholders more by hobbling the company over the long run.

Getting a structural remedy through the appeals process won't be easy, given the lack of recent precedent. But if the alternative is permanent government supervision of Microsoft--and by extension, the rest of the industry--it's the right thing to do.