Commentary: Microsoft: Put Down The Stopwatch, Your Honor

Let's get this straight. U.S. District Court Judge Thomas Penfield Jackson needed 21 months to complete the first phase of the Microsoft Corp. trial: determining whether the company broke antitrust laws. But now he wants to wrap up the remedy portion of the case--and, it would appear from a May 24 hearing--put his stamp of approval on a breakup plan in a matter of weeks.

SPEEDY SCHEDULE. Aren't we going a little quickly here? Whether Microsoft should be split apart is a much tougher question than whether it broke the law in the first place. And yet in the hearing, Jackson set a breathtakingly speedy schedule for imposing a remedy. Justice is supposed to submit a final plan by May 26; Microsoft must respond within 48 hours; and shortly thereafter the judge will issue his final ruling.

That means we could see a judicially-ordered breakup of the world's premier software manufacturer without anything close to a detailed analysis of several key issues: Will the post-breakup companies be economically viable? How can we be sure these two sister monopolies won't jointly raise prices? What would be the impact on innovation? The judge "is just saying, `Let's just dive into the water and not worry about whether there are any rocks at the bottom,"' says George Mason University antitrust professor Ernest Gellhorn.

Why would Judge Jackson, who has generally done a great job managing this unprecedented case, act so rashly? In part, because he wants to prove that the wheels of justice can move on Internet time. The main critique of prior monopolization lawsuits is that by the time courts got around to issuing their rulings, the cases were usually irrelevant. Exhibit A is the government's antitrust suit against IBM--which dragged on well after the company had lost much of its market power. To prevent such an embarrassment, Jackson has taken several creative steps to speed up the Microsoft lawsuit. His goal: to ensure that any punishments imposed on the software giant took effect while the Windows operating system was still an important platform.

No doubt, such haste made sense when the government was still thinking in terms of behavioral remedies. But now that Justice and the states want to restructure the software industry, the case for a legal wind-sprint is much weaker. Instead of imposing weak restrictions on Microsoft's business practices, the trustbusters are attempting social engineering that will alter the computing landscape for a decade or more. This isn't something that needs to be rushed.

The government's proposal raises complex issues that weren't briefed at trial, but it isn't at all clear it's the best solution. Many people think the plan doesn't go far enough--and that the company should be severed into three or four companies. Others argue any breakup would be disastrous, since it could temporarily undermine software standards, which have allowed the industry to flower. Justice antitrust chief Joel I. Klein argues that his proposal is the most practical remedy. But even some economists who support his case are worried that dividing Microsoft into an operating-systems monopoly and an applications monopoly could lead to "double marginalization"--the two Baby Bills would each charge a hefty markup, and future software prices would increase.

Before any appeals court can make a legitimate assessment of whether a breakup makes sense, it will need a more complete record than Judge Jackson appears likely to deliver. That could increase the odds he'll eventually be reversed. And the public deserves a better investigation, too. "If [breakup] is going to be taken seriously, then it needs to be discussed in some detail," says Nicholas Economides, a professor of economics at New York University's Stern School of Business.

This doesn't mean that Jackson should delay the case forever. Just that the stakes are awfully high here. It's better to make the right decision a few months later, than the wrong one a few months sooner.

Before it's here, it's on the Bloomberg Terminal.