TO THE COMMENTS
OF RELPROMAX ANTITRUST INC.
Antitrust, Fall 2001, Articles and Features
Additionally, the case became a contest between Microsoft and the other lions of the information technology industry. Never before in the history of the Antitrust Division have competitors of a firm under prosecution invested so heavily in the outcome of one of our cases. By virtue of its control over the operating system, Microsoft enjoys distinct advantages in competition for adjacent markets. With huge amounts of money on the line, Microsoft's competitors hired scores of lawyers and lobbyists to press their views of the case and to insist that the remedy be crafted so as to advance their interests.
Finally, Microsoft was the first major antitrust case to be tried in the age of 24-hour news and, as such, came to be covered like the Super Bowl, complete with play-by-play reporting, color commentary, player interviews, and endless expert analysis. Antitrust lawyers and academicians suddenly became television personalities, providing blow-by-blow commentary on every development in the litigation, A nine-hour court of appeals arguments was broadcast live over the Internet, and, miraculously enough, ordinary people listened to lawyers debating the finer points of technology trying and market definition. The general public actually formed opinions about the appropriate outcome of an antitrust case.
Under the circumstances, it was very easy for the antitrust case itself to become mere substext -- i.e., to become completely submerged beneath the show business, pontification, and self-interested advocacy. Antitrust, after all, is about the application of technical legal principles to highly complex economic circumstances. The types of issues that actually matter in antitrust analysis are far less glamorous and intriguing than the issues usually discussed in connection with the Microsoft case. One bedrock principle of antitrust law--that a firm might lawfully obtain a monopoly and exercise the power it affords--is almost counter-intuitive to the American psyche and presents difficult line-drawing exercises even for the antitrust cognoscenti. Thus, care must be taken to distinguish between "Microsoft, the antitrust case" and "Microsoft, the public spectacle." The two have developed quite differently.
In this article, I address what I consider to be the real Microsoft case--the antitrust dispute fought out in the courts. In particular, I wish to refocus attention on the legal allegations charged in the complaint, how those allegations were resolved in the courts, and the remedies in the proposed Final Judgment presently undergoing Tunney Act review (the "proposed decreed"). That the Microsoft litigation that emerged from the court of appeals was a far narrower antitrust case than was originally brought seems to have gone largely unnoticed. Viewed purely as an antitrust matter, we believe the remedy presently under consideration by the district court fully and demonstrably resolves the monopoly maintenance finding that the court of appeals sustained.
Underlying the monopoly manintance claim was proof that Microsoft took actions, among other things, to discourage the development and deployment of rival Web browsers and Java technologies, in an effort to prevent them from becoming middleware threats to the operating system monopoly. The settlement extends the product definition to all manner of present and future middleware products, prohibits in the broadest terms the practices found to be unlawful, deprives Microsoft of the means of disciplining firms that might develop or deploy competing middleware products, and requires disclosure of proprietary interfaces and protocols so that rival firms may create middleware that can compete on a function-by-function basis with Microsoft's integrated products. The prescribed relief is supported by an unprecedented level of enforcement power, including a full-time, on-site, independent compliance team and the power to extend the decree in the event of serious violations. In strict antitrust terms, by any serious reading of the court of appeals ruling, the settlement represents a big win for the government.
The Real Microsoft Case
A dose of reality is in order. Microsoft occupies a critical place in the the worldwide computer industry. Its proprietary Windows operating system exists as the de facto standard for the entire realm of IBM-compatible peronal computers. In rough terms, operating systems perform certain basic computing functions--drawing a box, opening a file, executing print commands, etc. These functions are accessed by software running on the operating system through application programming interfaces (APIs). The operating system is said to "expose" APIs. To facilitate the development of programs that run on the Windows operating systems, Microsoft discloses some, but not all, of those APIs to the software development community. Tens of thousands of software programs are written to the Windows operating system, a fact that has become a seemingly self-perpetuating basis for Microsoft's monopoly over the operating system market. Because computer users want their operating system to be a stable platform for present and future applications programs, new operating systems, whatever their technological merits, have difficulty gaining patronage. This is a "applications badrrier to entry."
The Microsoft operating system monopoly, fortified by the applications barrier to entry, is very durable, and gives Microsoft distinct advantages in competing in downstream markets for applications software, hardware, and Internet services. These has never been any serious contention, however, that Microsoft obtained its operating system monopoly through unlawful means, but rather that it unlawfully maintained it. Thus, under our antitrust laws, the existence of the operating system monopoly is not subject to direct attack and was not challenged in this case.2
In 1998, the Antitrust Division, nineteen states, and the District of Columbia commenced their respective antitrust cases against Microsoft. 3 The complaints alleged: (1) that Microsoft engaged in a series of anticompetitive acts to maintain its monopoly position in the market for operating systems designed to run on Intel-compatible personal computers, in violation of Section 2 of the Sherman Act; (2) that Microsoft attempted to monopolize the Web browser market, also in violation of Section 2; (3) that Microsoft illegally tied its Web browser, Internet Explorer (IE), to its operating system, in violation of Section 1; and (4) that Microsoft entered into exclusive dealing arrangements that also violated Section 1. After a full trial on the merits, the district court sustained all claims save for exclusive dealing. 4 The court of appeals, however, narrowed the liability findings, sustaining only the monopoly maintenance claim, but with fewer anti-competitive actions than the district court had found. 5 The court of appeals reversed with respect to eight of the twenty anticompetitive acts the district court had sustained as elements of the monopoly maintenance claim. The court of appeals reversed the attempted monopolization and tying conclusions, remanding the latter claim for further proceedings under the more rigorous rule of reason standard.6
And, of course, the court of appeals vacated the Final Judgment that had set forth the break-up remedy and interim conduct remedies.7 In its own words, the court of appeals "drastically altered the District Court's conclusions on liability."8
The surviving portion of the case focused primarily on the so-called "browser war" and allegations that Microsoft engaged in various anticompetitive pratices to maintain its operating system monopoly. For all of its claimed technological prowess and industry foresight, Microsoft had fallen behind in the race to commercialize Web browser technology. An upstart company called Netscape had beaten the mighty Microsoft to the punch and quickly gained a respectable market share as the preferred technology for navigating the then-burgeoning Internet. More importantly, Netscape proponents were touting the prospect of a new world of Internet computing that would make operating systems less relevant, if not virtually irrelevant. Netscape touted its Web browser as a new category of software that came to be known as "middleware," a form of software that, itself, exposed a broad range of APIs to which software developers could write applications. Assuming that Netscape Navigator could become fairly ubiquitous, that large numbers of developers wrote programs to it, and that Netscape, itself, ran on multiple operating systems, operating systems other than Windows could become viable. In this sense, the browser was a "nascent" threat to the operating system monopoly.
Microsoft took this middleware threat seriously. The trial record disclosed a corporate preoccupation with thwarting Netscape and displacing the Netscape Navigator with IE as the prevailing Web browser in the industry. This campaign featured a host of strong-arm tactics aimed at various computer manufacturers, Internet access providers, and independent software developers. Even the decision to integrate its own browser into the operating system -- in effect, giving it away for free--had an element of impeding the growth of Netscape and once was described as taking away Netscape's oxygen. Microsoft went so far as to make it more difficult to remove its browser from Windows in the apparent recognition that computer manufacturers would fear that the presence of two Web browsers on the desktop would cause consumer confusion and prompt expensive service calls.
The government's case against Microsoft focused heavily on the browser war--and a relatively narrow aspect of that war--the consumer's experience the very first time he or she boots up a brand new computer. a consumer with some basic level of computer knowledge and a small amount of effort, could elect to use the Netscape browser. Just like any form of computer software, alternate browsers were available through download and other forms of retail distribution. The underlying assumption of the browser war, however, was that most consumers become wedded to the first products to which they are exposed and those the computer manufacturer makes it easiest to deploy. Thus, desktop placement and other forms of "first-sighting" were important market movers.
The district court based its monopoly maintenance liability finding on the government's proof that Microsoft had engaged in a series of exclusionary acts. Those acts involved, in sum, the integration of IE into Windows, while closing off access for Netscape; various dealing with original equipment manufacturers, independent software providers, Internet access providers, and Internet hardware providers; effort to thwart Java technologies; and a course of conduct as a whole.9 While Judge Jackson had sustained the government on most of the claims, the court of appeals was a bit more selective. It ruled that certain of Microsoft's practices--for example, Microsoft's practice of preventing computer manufacturers from substitution their own user interfaces for the Windows interface supplied by Microsoft--were justified and thus lawful. 10 The court of appeals also explicitly rejected the course of conduct theory.11 Under that theory, we, in effect, argued that the whole was greater than the sum of the parts--that Microsoft's individual practices, rather than being judged on their individual merits, should be evaluated as part of a grander scheme of monopolization, providing an independent basis for monopolization liability.
Thus, we found ourselves victorious on the monopoly maintenance claim, albeit with less underlying conduct to remedy. The attempted monopolization count was gone and, based on the court of appeals' decision and the need to move to the remedy phase as quickly as possible, we dropped the tying claim. Those two claims had been a direct assault on Microsoft's ability to compete outside of the operating system--in particular, its ability to integrate new functions into Windows. But the court of appeals had made it clear that, albeit with some limits, Microsoft could lawfully integrate new functions into the operating system and use the advantages flowing from its knowledge and design of the operating system to compete in downstream markets. What was left in the case was a series of individual practices directed against competing browser developers and others, which the court of appeals found to be unlawful because of their potential to protect the operating system monopoly. That was the conduct to be remedied; not the existence of the Microsoft operating system monopoly itself and not the prospect that Microsoft might come to dominate other downstream markets for reasons unrelated to its conduct protecting the operating system franchise.
Dropping the Structural Remedy. Having taken the helm of the Antitrust Division just weeks before the court of appeals decision, for me the prior history of the case had to be just that history. The court of appeals decision was the new reality. It set forth the rules of the game. Judge Jackson's order to separate Microsoft's operating system business from the appliations business had received a chilly reception, at best, in the court of appeals decision . 12 By admonishing that a structural remedy would only be appropriate if the government proved a more direct causal connection between Microsoft's exclusionary practices and maintenance of the operating system monopoly (a connection that Judge Jackson said he could not find), it was clear that a structural remedy was not favored. 13 Moreover, even in the absence of the court of appeals' cautionary language about a structural remedy, its adverse conclusions on the tying and attempted monopolization claims undercut any real basis for separating the operating system and applications businesses. After all, the court of appeals declined to hold that Microsoft could not lawfully integrate appliations functions into the operating system. Given what was left in the case--essentially a series of heavy-handed contracting practices with computer manufacturers and software developer, unlawful when undertaken to protect the operating system monopoly--a conduct remedy seemed all that could be secured, let alone justified.
Eliminating structure from the remedial picture was also an important tactical issue as we moved into a remedy hearing before a new judge. The new judge obviously would take into account the court of appeals not-so-subtle message regarding structural relief, and likely would have little patience for some quixotic effort to press for that remedy, even as a mere bargaining chip. Moreover, the question of structural relief would have greatly complicated the remedy phase of the case. Microsoft would have been entitled to raise a raft of issues regarding the impact of such relief. This would have substantially protracted the discovery phase of the remedy proceeding, eliminating any prospect for a quick resolution, either in the district court or otherwise. Even if we had persuaded the trial court to impose structural relief, we would have found ourselves right back before the court of appeals, which already had said that it would review structural relief with considerable skepticism and under stringent legal k standards. Finally, from my own personal perspective, the structural remedy that Judge Jackson had ordered, i.e., a break-up into two companies without ongoing line-of-business restrictions--might have been a hollow, short-lived solution. Without the types of continuing line-of-business restrictions found to be hopelessly complicated and regulatory in the AT&T decree, Microsoft easily could have kept its operating system monopoly and reasonabled its appliations businesses through acquistion and internal development in a matter of years, if not months. Presenting the case for line-of-business restrictions would have been yet another complication. Taking structural relief off the table at the outset of the remedy proceeding before the new judge enabled us to get favorable procedural rulings that were essential to moving quickly to a prompt resolution.
The Conduct Remedy
An antitrust remedy for a Section 2 violation must stop the offending conduct, prevent its recurrence, and restore competition. While prohibiting the exact conduct found to be unlawful in the court of appeals decision would be relatively simple work, addressing the broader question of monopoly maintenance was more difficult. Preventing recurrence must involve proactive steps to address conduct of a similar nature. Restoration requires prospective relief to create lost competition and may involve actions to disadvantage the antitrust offender and/or favor its rivals. The relief, however, must have its foundation in the offending conduct. The monopoly maintenance finding, as sustained by the court of appeals, was not a mandate to range broadly across the computer industry purporting to solve unproven problems unrelated to the middleware threat to the operating system. Indeed, Judge Kollar-Kotelly stated in open court that she expected the government's proposed remedy to reflect the fact that portions of the government's case had not been sustained.
At the outset of the remedies proceedings, we stated that our proposed relief would be modeled upon the interim conduct provisions of Judge Jackson's (now-vacated) Final Judgment. Those provisions were "interim" because they were intended to remain in place only during the year in which Microsoft was preparing to be broken into two companies. the government itself had proposed those interim remedies, and thus they provided a base from which to develop appropriate remedies. We were cognizant, however, that those remedies provided relief based on Judge Jackson's liability findings, including the attempted monopolization and tying claims and portions of the monopoly maintenance claim that were not upheld by the appellate court.
In addition, because those remedies had never been subjected to the rigors of an adversarial proceedings, and had been prepared without any meaningful consultation with Microsoft, certain practical issues had never been fully vetted. This was a real danger, because the remedies had to be fully capable of being put into practice, anything else would be an enforcement nightmare. The interim remedies had also been based upon a trial record developed largely in 1998 and 1999. The industry had changed significantly since then. Among other things, by most accounts, Microsoft had essentially won the browser war, relief to revive Netscape Navigator as a middleware threat may have too little, too late. In addition, the character of potential middleware platforms had largely changed. It is unclear whether another general middleware threat like the browser will ever again emerge. The middleware war of today appears to focus on more specialized types of software, such as instant messaging systems and media players-systems that might be platforms for families of related functions, but whose potential to be a platform for a broad range of application remains to be seen. Today, one would not necessarily predict that a software developer would write a financial services program, for example, to run on either a messaging system or a media player. Finally, and perhaps most importantly, the operating system world has changed. At the time the case was filed, the browser was designed to sit on top of the operating system like any other application that had to be "opened" manually. In the ensuing years, the technology evolved so that browser functions increasingly were integrated into the operating system and invoked automatically to create a more seamless user experience. This made middleware products easier for consumers to use, but placed greater demands on competitors to create products that could function as seamlessly as Microsoft's.
Enjoining the Unlawful Conduct. There was no question that any proposed Final Judgment had to include prohibitions to enjoin the offending conduct that the district court had found and the court of appeals had sustained. I must note, however, that the affirmative prohibitions contained in the proposed decree go considerably beyond the literal findings of the court of appeals. The decree broadly bans exclusive dealing, gives computer manufacturers extensive control of the desktop and initial boot sequence, and prohibits a broad range of retalitory conduct. There can be no question that Microsoft must fundamentally change the way in which it deals with computer manufacturers, Internet access providers, software developers, and others under the proposed decree.
The middleware definition was a very complex issue and would have been fought hard in a litigated remedy proceeding. The team has no accepted industry or technical meaning, and one might reasonably distinguish between products that might function as middleware (i.e., an intermediate software program between an operating system and an application) and products that actually might threaten the operating system monopoly. At the time of trial, the term middleware was used to describe software programs that exposed APIs. In today's world, by virtue of the extensive degree to which software programs interact with each other, a very broad range of programs--large and small, simple and complex--expose APIs. Obviously, all software that exposed APIs could not qualify as middleware for purposes of the case.
As middleware is defined in the proposed decree, it captures, in today's market, Internet browsers, e-mail client software, networked audio/video client software, and instant messaging software. On a going forward basis, it also provides guidelines for what types of software will be considered middlewares for purposes of the decree. These guidelines are critical because, while it is important that future middleware products be captured by the proposed decree, those products will not necessarily be readily identified as such.
Preventing Recurrence. Having addressed the basic prohibitions, the more formidable tasks of addressing recurrence, restoration, and enforcement confronted us. Because of the critical role Microsoft's Windows operating system plays within the computer industry, the company has at its disposal a broad array of potential means of projecting its will upon manufacturers and developers. Many in the computer industry believe that Microsoft can influence product development and deployment decisions by other firms through a variety of carrots and sticks, and there was evidence in the trial record that Microsoft had used a number of such tactics to impede the emergence of competing Web browsers. By the same rocken, not all forms of collaboration between Microsoft and other industry participants are anticompetitive, and some actually benefit competition and consumers. To cover the broad range of potential strategies Microsoft might deploy, we sought to address the recurrence issue through the broad concepts of non-discrimination and non-retaliation, rather than by simply enumerating a list of specific prohibitions that we can identify today.
Restoring Lost Competition. With regard to restoration, we had to begin from the premise that the middleware threat to the operating system monopoly was a nascent one. It is said that some 70,000 applications currently run on Windows, making the applications barrier to entry quite formidable. At the very peak of our advocacy in the case, unfortunately not even we could hypothesize a point at which so many applications would be written to middleware APIs that there would be a meaningful threat to the Windows operating system, and the fact was noted in the court of appeals decision. Thus, the task in the restoration aspect of the decreed was to restore the potentiality of middleware. As had been the case in the mediation before Judge Posner and in Judge Jackson's order, our preferred approach was to ensure that middleware developers had access to the technical information necessary to create middleware programs that could compete with Microsoft in a meaningful way. This would have been a difficult undertaking in the earliest phases of the case, and it became even more difficult with the evolution of operating systems toward more integrated and seamless functionality.
API disclosure apparently had been a very difficult obstacle to resolution of the case at every stage. Press reports of the mediation before Judge Posner indicated that the scope of API disclosure was the sticking point that doomed earlier efforts to settle. There had never been any allegation in the case that Windows was an essential facility, the proprietary technology for which had to be openly shared in the industry. Further, failure to disclose APIs sufficient to create interoperable software was not a critical underpinning of the case; after all, we argued successfully that the 70,000 applications written to Windows contributed so the strength of the monopoly. That is not to say that disclosure had not been used selectively by Microsoft, the government showed that Microsoft had sought agreements with middleware developers to refrain from developing competing products in exchange for preferential access to proprietary technical information. Such agreements, however, would have been facially unlawful, without regards to Microsoft having some inherent legal obligation to disclose APIs. Technical support, in that sense, was merely consideration for an otherwise unlawful agreement.
There is a duality in Microsoft's position on API disclosure. The company frequently argues that it has strong incentives to make broad disclosures of APIs and other technical information because it wants developers to write applications programs to its operating system. Nonetheless, the company is forced to acknowledge that many important Windows functions rely upon undisclosed APIs. Moreover, it seemed fairly clear that Microsoft's reluctance to disclose those APIs was based upon more than a concern about the burden of administering the disclosure regiment. The simple fact is that certain of the undisclosed APIs relate to cutting-edge functions and applications, many of them integrated deeply into the operating system, that could be meaningful points of differentiation between Windows middleware products and those offered by third parties. Thus, we believed that we were on strong legal ground in seeking to require Microsoft to disclose its undisclosed APIs as a means of restoring competition in the middleware industry.
The decree's provisions for API disclosure will be critical in enabling independent developers to match Windows' functionality in their middleware products. Simply stated, if Microsoft middleware products rely on an API, that API must be disclosed. API disclosure of this type has become all the more critical as the operating system has evolved to include more deeply integrated functions that are invoked automatically in connection with other applications. But for detailed API disclosure, many middleware functions offered in third-party software would be more difficult to implement than Microsoft's integrated functions. Independent middleware vendors also will be aided by the decree's provisions that require Microsoft to create and preserve "default" settings, such that certain of Microsoft's integrated middleware functions will not be able to over-ride the selection of a third-party middleware product. This is yet another respect in which changes in the technology required us to go beyond the relief contemplated in Judge Jackson's order. By giving middleware developers the means of creating fully competitive products, requiring the creating of add/delete functionality that will make it easier for computer manufacturers and users to replace Microsoft middleware functionality with independently developed middleware, and making it absolutely clear that computer manufacturers can, in fact, replace Microsoft middleware, the decree will do as much as possible to restore the nascent threat to the operating system monopoly that browsers once represented.
The disclosure of communications protocols for servers was another matter entirely. Very clearly this was a by-product of the negotiated resolution of the case that would have been highly contentious and by no means a certainty had the remedy been fully litigated. In recent years, Microsoft has moved aggressively into the network server market, competing with firms, such as IBM, Novell, and other. Servers tend to communicate through fairly standard communications protocols. Nonetheless, all server manufacturers have implemented their own proprietary communications technologies, working in tandem with the standard protocols, that each server line employs to differentiate itself from other brands ad models in terms of features and performance. In network computing, these proprietary technologies are implemented through software code residing at both ends of the communications link, i.e., on the desktop operating system or "client" and on the server itself. Anyone who has ever worked in a network computing environment it familiar with the process of the network administrator loading new software in the form of annoying updates, downloads and overnight system changes. While other server manufacturers must load all of their client software independently, Microsoft ships some of its client-server interoperability technology on Windows itself.
Competing server manufacturers complain that Microsoft has an inherent advantage in server competition in that it has a superior knowledge of the operating system and possesses the ability to embed within the operating system secret technology that gives its servers more features and better performance than servers offered by other. Although these claims have never been fully discovered or litigated, it seems intuitively correct that Microsoft would behave in this manner. Yet there is a clear body of case law that a vertically integrated firm, even one that is a monopolist, might rationally and lawfully gain a competitive advantage in one market from its strength or monopoly power in another. 14 In other words, there is no legal rule of antitrust law requiring distinct businesses within a single firm to act as though they are complete strangers to each other.
The connection between the "secret sauce" argument in the server market and the monopoly maintenance claim in this case may not be obvious. To establish the connection on even the theoretical level, one must hypothesize a number of circumstances concerning the future economic course of the server market. To use the Berkeley School "econo-phrase," Microsoft's practices of embedding secret technology in the Windows operating system would have to cause the server market to "tip" in its favor, such that competing servers would no longer provide a platform threat to the desktop operating system.
Further, this issue developed primarily since the trial--it was barely raised and never litigated in the proceedings before the district court. Indeed, the word, "server" never appeared in the complain and was mentioned only a handful of times in the findings of fact--in those instances mostly in the context of discussing the browser as a component in network computing. In fact, the most cogent statement of the server theory appears in a series of proposed findings of fact we submitted that actually were not adopted by Judge Jackson. 15 Undoubtedly, Microsoft would have argued strenuously in the remedy proceeding that server issues had never been litigated in the case and provided no basis for relief.
Notwithstanding any difficulties the government might have faced in pursuing server relief in the remedies proceeding, the potential competitive consequences in this area are substantial. If it is true, as many seem to believe, that serverbased applications might be the next important threat to the operating system monopoly, this would be an area in which it is important to stretch for relief, whether or not the issue was fully fleshed-out in the liability phase of the case. The proposed decree, therefore, requires Microsoft to disclose communications protocols embedded in the operating system, but preserves for the company the ability to deploy proprietary communications technology provided separately. Though this does not fully negate Microsoft's inherent advantages as the designer of the operating system, it requires Microsoft to compete as other server manufacturers do by separately providing client-side communications technology.
Enforcement. It has been suggested that no "conduct" remedy could be effective because Microsoft is a defiant company that cannot be trusted to comply with an antitrust decree. Our practice with regard to enforcement, however, is never influenced by the extent to which we "trust" a defendant. Rather, the decree must stand on its own as an enforcement vehicle. The settlement in this case contains some of the most stringent enforcement provisions ever contained in any modern consent decree. In addition to the ordinary prosecutorial access powers, backed up by civil and criminal contempt authority, this decree has two more aggressive features. First, it requires an independent, full-time, on-site compliance team--complete with its own staff and the power to hire consultants--that will monitor compliance with the decree, report violations to the Department, and attempt to resolve technical disputes under the disclosure provisions. The compliance team will have complete access to Microsoft's source code, records, facilities, and personnel. Its dispute resolution responsibilities reflect the recognition that the market will benefit from rapid, consensual resolution of issues, more so than litigation under the Department's contempt powers. The dispute resolution process complements, but does not supplant, ordinary methods of enforcement. Complainants may bring their inquiries directly to the Department if they choose.
The decree also contains a provision under which the team may be extended by up to two years in the event that the court finds serious, systematic violations. Assuming that Microsoft will want to get our from under the decree's affirmative obligations and restrictions as soon as possible, the prospect that it might face an extension of the decree should deter violations and provide an extra incentive to comply. We opted for this three-part compliance regime, not because of the assumption that Microsoft would act in bad faith, but rather because of the inherent complexity of the decree, which seeks to address Microsoft's interactions with firms throughout the computer industry. Under the circumstances, ongoing supervision backed up by tough penalties was in order.
Reaction to the Settlement. In our view, the proposed decree fully remedies the violations upheld in the court of appeals decision. No conceivable remedy could guarantee that middleware products will emerge as a palpable threat to the operating system monopoly. However, the decree, if entered, creates an environment in which Microsoft will have to compete on the merits and will be prevented from using anticompetitive means to impede the emergence of competing middleware products. In the short term, computer manufacturers will be able to offer consumers choices with respect to popular middleware programs free of interference from Microsoft. To the extent that the decree prevents Microsoft from unduly influencing the natural propagation of middleware technologies, it will have accomplished an important purpose, whether or nor any particular middleware product ultimately grows into a meaningful threat to the operating system.
Not surprisingly, the settlement has its critics. Some of the criticism, however, heartfelt and passionate, can be dismissed as failing to reflect a real appreciation for the underlying law. Generalized claims that Microsoft's operating system monopoly should have been taken away as "punishment," or that the government should have obtained monetary relief, will not be seriously debated in antitrust circles. The remaining critics tend to fall into two categories. The loudest and most vocal critics, understandably, are Microsoft's competitors--some of which hoped the case would bring about a wholesale emasculation of Microsoft, while providing their own companies specific strategic, technological, and financial advantages. A second group, which again includes competitors, raises questions as to whether a remedy of this scope and complexity reasonably can be expected to correct the unlawful conduct in which Microsoft has been found to have engaged. Questions of that nature are entirely appropriate and we hope will be raised and addressed in the Tunney Act process.
The Microsoft case is unique in the level at which competitors have sought to assist in and influence the government's liability and remedial determinations. From the very outset of the litigation, virtually every significant player in the computer industry, directly or through coalitions, hired a team of antitrust lawyers (and sometimes lobbyists) to advance their views of the case. Support from competitors was extremely helpful to us during the liability phase, in that it gave the government ready access to discovery materials, technical expertise and other assistance. This support also was helpful in conceptualizing the remedy and testing the likely competitive effects of various remedial approaches.
Nonetheless, the government still had to make judgments based upon its own assessment of the marketplace, rather than solely or even predominantly from the perspective of competitors. Once liability had been established, a sense of entitlement set in among some firms that had rendered assistance. Firms in the industry became very aggressive in pressing for all manner of relief, whether or not it had anything to do with the antitrust liability that had been established.16 The specter of the Microsoft operating system monopoly overhangs every level of the information technology industry and very clearly imposes serious challenges upon firms seeking to compete in adjacent markets. One can easily hypothesize any number of governmental actions constraining Microsoft or bestowing advantages upon its rivals that might be said to "level the playing field" upon which Microsoft must compete with other firms. The Antitrust Division, however, has no mandate to "regulate" competition divorced from remedying specific antitrust violations. Thus, from a law enforcement perspective, the relief had to be tailored to proven violations; it could not be a laundry list of unrelated requirements competitors might find useful. In this regards, it is useful to consider some of the principal concerns expressed publicly by competitors and others, in light of both the court of appeals decision and the proposed decree.
Probably the single most widely, publicly expressed criticism relates to the questions of product bundling and commingling of software code. Firms in adjacent markets very clearly wanted the decree to restrict, if not prohibit, Microsoft from competing outside of the operating system market. That is one of the reasons that the structural remedy was so popular among competitors. Rules prohibiting Microsoft from integrating products into the operating system would benefit competitors. The court of appeals, however, refused to establish a rule of antitrust law that outlaws Microsoft's integration of new functions or products into the operating system. We had to accept that we had lost on that issue in the context of four separate claims; the contempt proceeding under the 1995 decree, and in the monopoly leveraging, tying, and attempted monopolization claims of this case. Given the manner in which the courts have treated this issue, it is surprising that there has been any suggestion that the settlement is deficient because it does not restrict bundling.
Equally surprising is the public criticism that the proposed decree should have required Microsoft to sell a la carte versions of Windows with the middleware priced separately. As an initial matter, this would have been a remedy most appropriate for the tying claim, rather than for monopoly maintenance. Even more fundamentally, where is the consumer benefit in this relief proposal? Computer manufacturer tout the fact that their products include the Windows operating system, and consumers reasonably expect that a Windows operating system will include the Windows features. The proposed decree provides the computer manufacturer the option of featuring alternative middleware products, which consumers might accept or reject as they see fit. We saw no public benefit in depriving consumers of that choice.
Similar issues have been raised with regard to the commingling of code. At trial, the government challenged Microsoft's actions to prevent removal of its browser--in particular, eliminating the browser from the list of programs that could be removed by the computer manufacturer or end user with the add/remove function, and making it impossible to remove the Microsoft browser without removing related operating system code and thereby losting necessary operating system function. Our proposed decree addresses both issues by requiring that Microsoft redesign Windows to include an effective add/remove function for all Microsoft middleware products and to permit competing middleware to take on a default status that will override middleware functions Microsoft has integrated into the operating system. In tandem, these provisions provide an unimpeded choice to select an alternative middleware product. We have never read the court of appeals decision with regard to commingling to be an attempt by the court to articulate an affirmative rule of software design under which all commingling would have been prohibited. This aspect of the court's opinion very clearly addressed the subject of consumer choice, which has been fully addressed in the proposed decree. Those seeking a broad ban on commingling appear to be seeking to reduce consumer choice, not increase it.
Public critics also express concern that the proposed decree does not address Microsoft's incursions into new markets and services, such as e-commerce, Internet services, or content. They contend that, left unabated, Microsoft will unfairly gain control over those markets from its base in operating systems. Given Microsoft's market power in operating systems, its movement into adjacent markets will almost always merit careful antitrust scrutiny. That said, the rulings in this case provided no basis for attempted monopolization or monopoly leveraging remedies that would proscribe Microsoft actions that do not have the purpose and effect of protecting the operating system monopoly. Once again, with the monopoly maintenance claim as the only surviving basis for relief, the remedy has to focus on middleware or middleware-type treats to the operating system, nor the prospect that Microsoft might come to dominate other markets.
It has been reported that critics also claim that the proposed decree is riddled with "loopholes," particularly with respect to carve-outs from the exclusive dealing and non-retaliation provisions. Those carve-outs, which are stated in phraseology borrowed from well-established Supreme Court joint venture law, afford Microsoft a limited ability to engage in collaborative conduct that would be plainly lawful under the established precedents. We can understand why some would prefer that Microsoft be prevented from engaging in any collaborative conduct. Among other things, such relief would constrain Microsoft in its ability to more into new technologies, products, and markets, isolating the company into a world of internal product development. But we never alleged that all of the scores of types of agreements into which Microsoft routinely enters contributed to illegal maintenance of the operating system monopoly. Indeed, many developers actually benefit from their collaborations with Microsoft and the prospect that their products might be commercialized through, or with, Microsoft. Consumers benefit to the extent that such collaborations bring such new products and services to the market. A flat prohibition would have prevented such collaborations without regard to their potential procompetitive benefits. It might also have prevented Microsoft, for example, from offering a promotional allowance in connection with a new product to one developer unless it provided the allowance to all. Requiring Microsoft to provide those allowances to firms that did not even sell the product would have been nonsensical. For these reasons, an absolute ban on collaborative activity would have been overkill under the circumstances of this case.
Many competitors and others in the industry strongly favored a remedy that required Microsoft to disclose its source code in its entirety. There is within the computer science community, particularly among devotees of other platforms, a generalized notion that Microsoft unfairly foists inferior software upon the public by maintaining the protected proprietary quality of its operating system. Proponents of that view believe that innovation would accelerate, and Microsoft itself would be forced to compete on the merits, if Windows could be transformed into an open source code platform.17
While competitor demands for access to source code are often stated in terms of interoperability, there is no question that such access would benefit competitors in other ways, as well. Access to the source code would provide competitors free access to Microsoft's programming approaches and innovations, allowing firms to imitate, copy, or clone products and services Microsoft offers.
The proposed decree assures disclosure without providing an easy way for competitors or even hackers to misappropriate intellectual property and trade secrets for themselves. To the extent that the secure facility provided in Judge Jackson's order was to ensure adequate disclosure and documentation under the decree, we elected the effective but less intrusive approach of providing the compliance team, a group of computer science experts, with access to the source code. If a developer believes that there has been a failure of the disclosure regime, the compliance team will be able to determine quickly whether Microsoft has met its obligations under the decree. If it has failed, it may either cure the problem or risk contempt sanctions. There being no basis under the court of appeals decision to order Microsoft to disclose its intellectual property to undermine the operating system monopoly, it would have been a significant stretch to require open access simply to police a restorative provision. Our proposed decree provides an effective mechanism for developers to receive source code assistance, consistent with our remedial goals, without the danger of misappropriation of Microsoft's intellectual property.
With regard to compulsory licensing to computer manufacturers, again, the court of appeals ruling is a constraining factor. Judge Jackson ordered this disclosure upon the assumption that the tying and attempted monopolization decisions had been sustained, providing a basis for separating middleware products from the operating system. Those findings, of course, were reversed by the court of appeals, calling into question whether computer manufacturers could lawfully reconfigure Windows by stripping out Microsoft middleware products in their entirety. Moreover, one would have to question the practicality of having computer manufacturers redesigning Windows for their own purposes. As it is, the proposed decree provides the computer manufacturer considerable control over the Windows desktop at initial boot-up and beyond. Requiring Microsoft to design effective add/delete and default functions into new versions of Windows would seem far superior to inviting computer manufacturers to create do-it-yourself versions. Last, a compulsory licensing regime would have raised many of the same security concerns as the secure facility.
In sum, it is understandable why competitors would want Microsoft to unbundle its integrated products, refrain from all collaborative activity, and widely disseminate its proprietary intellectual property. Those requirements, however, are largely beyond the scope of the court of appeals decision or otherwise do not advance public goals. The antitrust laws protect competition, not competitors. We believe that the relief we have negotiated fully addresses the legitimate public goals of prohibiting Microsoft's unlawful conduct and restoring competition.
6 Id. at 96. The Court also held that although if it pursued the tying claim the tying government had the burden to show an anticompetitive effect in the browser market, it was "precluded from arguing any theory of harm that depends on a precise definition of browsers or barriers to entry ...." Id. at 95.
16 Indeed, the number of requests for meetings with me immediately after my nomination but before my confirmation became so daunting that I adopted the posture of refusing to meet personality with my third parties in the Microsoft case, and I have maintained that posture to this day. Third parties had complete access to the staff and my Deputy throughout the process, and I was willing to reed any written submission that a third party mit
17 For example, in a recent Washington Post article, Sun Microsystems is reported to have argued that all of Microsoft's formats and all of its communications protocols should be "turned into open, published standards," U.S. Settlement Leaves Microsoft More Entrenched, WASH. Post. Nov. 9, 2001, at E1.
Charles A. James is the Assistant Attorney General in charge of the Antitrust Division of the U.S. Department of Justice. The Antitrust Division commenced its case against Microsoft on May 18, 1998. A coalition of 19 states and the District of Columbia, with New York as the lead plaintiff, filed their lawsuit on the same day, and the two cases were consolidated and prosecuted jointly. Nice estates, including New York, have joined with the United States in propesing a settlement. The views expressed herein are the author's and relate to the Antitrust Division's case and settlement and do not purport to represent the views of any state plaintiff.