The breakup of Microsoft proposed by the Justice Department and the attorneys general of 17 states would be a positive development for the software industry and for the public welfare. Curbing Microsoft's abuse of its monopoly will require a strong structural remedy, and the proposal to create two companies, one selling Microsoft's Windows operating system and the other selling applications such as Office and Internet Explorer, would promote competition in both fields. Despite the concerns that have been raised over such a dramatic penalty, a breakup of Microsoft seems to be the best of the available options and is more than justified by the trial record.

While a breakup is a harsh punishment, it is unclear whether anything weaker would effectively deter Microsoft from continuing its monopolistic practices. Conduct remedies that left Microsoft structurally intact but placed rules on its behavior would likely be evaded by the software giant; the antitrust case grew out of Microsoft's circumvention of a consent decree not to bundle Internet Explorer with Windows. Imposing permanent regulations on Microsoft's actions would thus require constant judicial monitoring (and constant appeals from Microsoft) to ensure compliance.

The government plan would place several restrictions on Microsoft's conduct for the next few years, such as mandating that Windows licensing fees be made public and that "unbundled" versions be available should Microsoft incorporate other software into Windows. Yet these short-term regulations will be far more effective accompanied by the breakup than would a system of conduct remedies alone.

On the other side, there is an additional fear that the breakup will not be strong enough. The portion of Microsoft controlling the operating system could retain a monopoly. Some suggest that a better solution would be to auction the Windows source code to several other companies, immediately creating competing systems. However, competing versions of Windows could be sufficiently similar to create customer confusion and remove any real advantage from the split. Furthermore, were competing versions of Windows introduced while Microsoft still held on to its applications software, it would not be long before Office became more compatible with Microsoft's version than the others and a near-monopoly were reestablished.

The goal of the breakup is not only to create multiple competing operating systems, but to reduce the possibility that future software companies will be able to build a similar monopoly. The suit was designed to prevent Microsoft domination of "middleware" like Java and Netscape. Middleware allows software developers to write programs that function equally across different operating systems, making an operating system itself less relevant. After the breakup, the applications portion of Microsoft would have incentives for compatibility with as many platforms as possible. Linux (a Unix-like operating system) would become a much more viable operating system if a Linux version of Microsoft Office was developed; a dependable supply of quality applications would be a boon to both Linux and Apple's MacOS. Only with equivalent applications on a variety of platforms can there be true operating-system competition.

The final concern is that two monopolies might be more predatory than one. While a unified Microsoft had an incentive not to lose customers by raising the prices on Office and Windows too high, the two parts of a divided Microsoft could instead face a "prisoner's dilemma" and both choose to raise prices for consumers. However, this view attributes too much strength to the Office monopoly and assumes the two "baby Bills" will not work against each other. Applications software is by nature harder to monopolize than operating systems: Word documents can be read and written by Corel's WordPerfect or Sun's StarOffice, and the possibilities for new entrants in the applications field would only rise after the breakup. Alternatives to Office software are far stronger than alternatives to Windows, and could have the potential to replace Office were the operating system a neutral platform.

The record of the case shows that Microsoft does not deserve a mere slap on the wrist and that improving competition in the software industry will require much more. A strong structural remedy is needed, and the government's proposal seems promising. If Judge Jackson considers the remedy proposals with the same prudence and wisdom that he has shown in earlier stages of the case, one can feel secure that Microsoft will pay an appropriate penalty for its actions.

DISSENT: Regulation is a Better Solution

The staff is too harsh in recommending that Microsoft be broken up. The computer industry is very different from a traditional industry. Technology monopolies are extremely tenuous. But, furthermore, they are necessary. In order for third parties to write new software, a standard platform must be established to write to. Otherwise small applications developers will suffer by being forced to take on the additional costs of writing for multiple platforms.

The staff believes that Sun's cross-platform Java programming language alleviates this problem. However it is not clear that writing in Java is any better a state of the industry than writing applications for Windows. Java is less powerful, and at least for the moment slower than traditional programming languages. This is not to say that the government should force the industry to adopt one platform over another, or break down the benefits of a standard.

A better solution would be to make Microsoft a slightly regulated monopoly. The possibility of a competitor breaking open the market should be left open, but unless that happens, the government should keep a close eye over Microsoft's shoulder to ensure that none of its practices block potential competition and that it does not take unfair advantage over other companies. Although the staff rejects this possibility due to likely evasions by Microsoft, these evasions would only ensue if the government were not persistent in upholding its standards.

--M. Ari Behar '02