Attached please find the Association for Competitive Technology's Tunney Act comments on the Microsoft settlement. A paper copy has been submitted by fax.
This e-mail is sent by a law firm and may contain information that is privileged or confidential. If you are not the intended recipient, please delete the e-mail and any attachments and notify us immediately.
COMMENTS OF THE ASSOCIATION FOR COMPETITIVE TECHNOLOGY
The purpose of a Tunney Act proceeding is to determine whether the settlement that the federal government has entered into is "within the reaches of the public interest." United States v. Microsoft Corp., 56 F.3d 1448, 1460 (D.C. Cir. 1995) (internal quotation marks and emphasis omitted). The RPFJ easily meets that forgiving standard. Indeed, as shown in detail below, this conclusion is easily established by measuring the RPFJ against four settled principles that govern relief in all antitrust cases, and by comparing the RPFJ to the radical "remedies" that have been proposed by the States that have refused to consent to the RPFJ ("Litigating States").
First, it is well settled that an antitrust remedy should be designed to protect consumers rather than advance the interests of competitors. The RPFJ will accomplish this goal. It prevents Microsoft from engaging in exclusionary or retaliatory tactics, as well as foreclosing a number of more specific paths to unfair competition. However, it is carefully crafted to ensure that Windows will remain available to consumers as a reliable operating platform.
By contrast, many of the Litigating States' proposals seem to have been designed by Microsoft's competitors. Indeed, the companies that will benefit most from the Litigating States' efforts are the same ones that have led the campaign to scuttle settlement efforts case and to impose far-reaching restrictions on Microsoft: AOL Time Warner, Sun Microsystems, Oracle, IBM, and Apple. As a prominent commentator recently noted, "Microsoft's enemies were largely responsible for instigating the lawsuit and were active behind the scenes in helping the government frame the charges and compile the evidence. Executives from Sun, AOL, Netscape and other companies testified against Microsoft." Fred Vogelstein, The Long Shadow of XP, Fortune, Nov. 12, 2001.
Each of these companies dominates a particular market that is distinct enough from Intel-compatible PCs not to be a part of this case, but related enough that Microsoft's rivals fear Microsoft's competition. For example, Sun Microsystems dominates the market for server operating systems, but its market share is being eroded by lower-cost alternatives from Linux and Windows. Why Competitors Are Largely Quiet on Microsoft Settlement, Associated Press, Nov. 15, 2001; Peter Burrows, Face-Off, Bus. Wk., Nov. 19, 2001, at 104. In asking for "must-carry" provisions for Java, limits on technical integration and the use of Microsoft middleware, and restrictions on Microsoft's investment in intellectual property, Sun seeks to maintain its stranglehold over the server marketplace. Similarly, Oracle enjoys a privileged position in the server database market but it, too, is facing stiff competition from lower-priced alternatives that are gaining increasing favor with reviewers and customers. As Oracle tries to move into different markets, such as e-mail, where consumers expect tighter integration, it will be unable to maintain its high prices unless Microsoft's capacity for product improvement is limited. Finally, "Microsoft and AOL are both dominant companies, orbiting in separate if overlapping domains. Yet both companies regard themselves as being on a collision course, as all forms of information and entertainment, including music and movies, are increasingly rendered in digital form." Steve Lohr, In AOL's Suit Against Microsoft, the Key Word Is Access, N.Y. Times, Jan. 24, 2001. An internal document makes clear that AOL is willing to take any necessary steps to gain control of the desktop, including even spreading false rumors about the stability of Windows XP. See http://www.betanews.com/aol.html.
Beyond these companies' own statements and court filings their views are parroted by various proxies. These include organizations that were specifically formed to hobble Microsoft, such as the misnamed Project to Promote Competition and Innovation in the Digital Age ("ProComp"), and existing trade organizations that these companies have recently joined and come to dominate, such as the Computer and Communications Industry Association ("CCIA") and the Software Information Industry Assocation ("SIIA"). The apparently high level of coordination between these groups and the Litigating States' counsel is ample reason for skepticism when examining some of the States' arguments.
The reality is that these rivals, both directly and through their proxies, are trying to use the courts to increase their own profits rather than consumer satisfaction. This is shown by the fact that, while they condemn Microsoft for integrating its products, they, too, are vying to bring integrated products to consumers. For example, Sun's SunONE initiative tries to offer the same level of integration as Microsoft's .Net service. See SunONE, Services on Demand vision, at http://www.sun.com/software/sunone/overview/vision/. Not surprisingly, Oracle shares this vision of a global network of centralized information and online services. It envisions an all-Oracle solution, advising businesses to "wage their own war on complexity" by turning to Oracle for "an integrated, complete software suite." AOL is likewise promoting its "AOL anywhere" strategy, which makes its popular services and features available to consumers anywhere, anytime through multiple platforms and mobile devices. Clearly, these companies do not believe that selling IT products piecemeal best meets consumers' needs, yet that is what they are trying to force Microsoft to do.
Second, it is equally well settled that an antitrust remedy should be tailored to fit the conduct that has been found illegal. Here, the RPFJ carefully addresses each of the types of conduct that the Court of Appeals found illegal. It regulates the agreements that Microsoft can enter into and prevents Microsoft from retaliating against software or hardware distributors. The RPFJ also gives both computer manufacturers and consumers more choices in configuring their computers, and specifically enables them to turn off any Microsoft middleware and replace it with the middleware of their choice. And the RPFJ requires Microsoft to disclose technical information and license its intellectual property to those whose products interoperate with Windows.
To be sure, the RPFJ in some respects goes beyond the findings of illegal conduct affirmed by the Court of Appeals. Unfortunate as that may be, it should not deter the Court from adopting the RPFJ. As the District Court for the District of Columbia stated in another context:
United States v. American Tel. & Tel. Co., 552 F. Supp. 131, 151 (D.D.C. 1982), aff'd sub nom. Maryland v. United States, 460 U.S. 1001 (1983) (mem.).
Nevertheless, the vast majority of the RPFJ's provisions respond to the findings that were affirmed by the Court of Appeals. Virtually all of the proposals by the Litigating States, by contrast, address areas wholly outside the scope of this case, such as Microsoft's corporate acquisitions, the Office suite of programs and, of all things, Microsoft's conduct of its intellectual property litigation. The Litigating States' proposals should be rejected for that reason alone.
Third, any antitrust remedy should minimize "collateral damage" to third parties. Here, the RPFJ carefully avoids serious harm to other sectors of the information technology industry.
The Litigating States' proposals, by contrast, would inflict enormous damage on the rest of the industry. Perhaps most important, their proposals would fragment the Windows standard and, in turn, profoundly disrupt other businesses that rely upon it. The Litigating States' proposals would also weaken intellectual property protections, setting an unnerving precedent for any IT firm aspiring to lead its market, and slow the pace of research and development in the IT field.
Fourth, an antitrust remedy should be easy to administer, and not be regulatory. The Litigating States, in an effort to impose their concept of "competition" in the information technology industry, would create a court-run agency to supervise Microsoft's every move and to judge its every action. In contrast, the RPFJ would create a more independent, more limited supervisory body that would have full access to Microsoft information, but that would not combine the roles of prosecutor and judge. This too counsels strongly in favor of the RPFJ, and against the proposals advanced by the Litigating States.
The remainder of these Comments is organized as follows. Section I summarizes and explains in more detail the four principles that are pertinent to the District Court's determination of whether the RPFJ is "within the reaches of the public interest." Microsoft, 56 F.3d at 1460 (internal quotation marks and emphasis omitted). Section II applies these principles to the RPFJ and, for comparison purposes, to the proposals advanced by the Litigating States.
Antitrust law recognizes that competition gets its vigor from the urge to win. A desire to ensure that all competitors will do equally well makes robust competition impossible and sets the stage for price-fixing and similar behavior. Accordingly, antitrust law and antitrust remedies are designed to foster real competition, so that consumers and the wider economy can ultimately benefit. Thus, while competitors driven by their own urge to win may try to misuse antitrust law to further their own goals, government agencies and courts should be vigilant to ensure that their power is used in consumers' interests rather than competitors'.
The case law on remedies generally and antitrust remedies in particular supports the goal of harnessing competition. A close reading of that case law reveals four specific principles that promote that goal, and that are dispositive here.
Perhaps the most important principle of antitrust law is that any remedy must be designed to benefit consumers, not just punish the defendant or enhance the position of the its rivals. The law is clear that, in a civil antitrust case, any injunctive remedy must be, as its name suggests, remedial rather than punitive. E.g., United States v. E. I. Du Pont de Nemours & Co., 366 U.S. 316, 326 (1961); International Salt Co. v. United States, 332 U.S. 392, 401 (1947); Hecht Co. v. Bowles, 321 U.S. 321, 329 (1944).Moreover, as Judge Robert Bork has shown in his famous book, The Antitrust Paradox, the entire purpose of antitrust law is promotion of consumer welfare, not the protection or enhancement of competitors. Robert H. Bork, The Antitrust Paradox 51, 56-89 (1978); see also National Collegiate Athletic Ass'n v. Board of Regents, 468 U.S. 85, 107 (1984).
It follows that any remedy must have as its principal purpose the promotion of consumer welfare. And, as the District Court recently noted, the states have the burden of establishing "the efficacy of every element of the proposed relief" in achieving that objective. Hearing Tr., Sept. 28, 2001, at 8, United States v. Microsoft, No. 98-1232 (D.D.C.).
For two reasons, it is doubtful that any remedy at all is needed to protect consumers in this case. First, it appears that the particular conduct at issue in this case has never harmed consumers in any meaningful sense. The government's own witness, Professor Frank Fisher of MIT, testified during the trial that the narrow conduct found unlawful by the Court of Appeals had not harmed consumers at all. When asked by plaintiffs' counsel whether that conduct had harmed consumers, Fisher replied: "[O]n balance, I would think the answer was no, up to this point." Trial Tr., Morning of Jan. 12, 1999, at 29 (Fisher), United States v. Microsoft Corp., 87 F.Supp.2d 30 (D.D.C. 2000), aff'd in part, rev'd in part, 253 F.3d 34 (D.C. Cir.), cert. denied, 122 S.Ct. 350 (2001).
If Microsoft's conduct did not harm consumers even "on balance" it is difficult to see how any remedy is now needed to protect them. But if any remedy is needed, the Court must be careful not to risk harming consumers by adopting remedy proposals such as those advanced by the Litigating States remedies which, to paraphrase Abraham Lincoln, are "of the competitors, by the competitors, and for the competitors."
Second, even if Microsoft's conduct could have harmed consumers in some way, any such risk has now abated. This entire case is premised on the assertion that Microsoft enjoys market power by virtue of the fact that a high percentage of IBM-compatible PCs use Windows as their operating system. Whether or not that was true when the case was tried, such knowledgeable industry observers as Sun's president have effectively conceded that whatever market power Windows might once have given Microsoft is now virtually a thing of the past.
For example, in his January 3, 1999 interview on 60 Minutes, Scott McNealy rejected Leslie Stahl's suggestion that with its Java software, Sun now "ha[d] a chance to make Windows obsolete." Instead, McNealy retorted, "Windows is obsolete, [and] we have a chance to show the world that it is." 60 Minutes (CBS Television Broadcast, Jan. 3, 1999).
McNealy elaborated this theme in a subsequent Wall Street Journal op-ed piece, which appeared more than two years ago. He asserted that, because of the growth of the Internet, "[a] few years from now, savvy managers won't be buying many, if any, computers. They won't buy or build anywhere near as much software either. They'll just rent resources from a service provider," primarily over the Internet. Scott McNealy, Why We Don't Want You to Buy Our Software, Wall St. J., Sept. 1, 1999, at A26.
McNealy's predictions are already being borne out. A recent article assessed the changes in the operating system market. It noted that "Microsoft's main markets are maturing and the entire ground under its empire is shifting. Market researchers expect PC sales worldwide to drop [in 2001] and at best to stagnate in 2002. What is more, software is increasingly a service delivered over the Internet, meaning that operating systems are no longer central." Microsoft: Extending Its Tentacles, The Economist, Oct. 20-26, 2001, at 59. Thus, whatever market power Microsoft now possesses is rapidly being eroded, or is already effectively gone.
In short, because Microsoft's present market power is limited at best and will be further eroded in the near future, there is no need for antitrust remedies. See also William E. Kovacic, Designing Antitrust Remedies for Dominant Firm Misconduct, 31 Conn. L. Rev. 1285, 1314 (1999) (explaining that rapid technological change can indicate the instability of market power, and therefore to the need for milder remedies). At a minimum, any antitrust remedy must take into account the dramatic decline in any market power Microsoft might previously have enjoyed, and be limited accordingly.
Another principle that must guide the analysis of any proposed antitrust remedy is that it must be no broader than necessary to address the conduct that has been found illegal. As with all injunctive relief, the "substantive prerequisites for obtaining an equitable remedy as well as the general availability of injunctive relief . . . depend on traditional principles of equity jurisdiction." Grupo Mexicano de Desarollo, S.A., Inc. v. Alliance Bond Fund, 527 U.S. 308, 319 (1999) (quoting 11A C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure § 2941, at 31 (2d ed. 1995)). And one of these "traditional principles of equity jurisdiction," id., is that an injunction should be no more burdensome than necessary to prevent a recurring violation of the law. See generally Madsen v. Womens Health Center, 512 U.S. 753, 765 & n. 3 (1994), and cases cited therein.
This is as true in antitrust as in other areas of the law. For example, in the Lorain Journal case, which Robert Bork believes is the closest to this one, the Court noted that, "[w]hile the decree should anticipate probabilities of the future, it is equally important that it ... not impose unnecessary restrictions." 342 U.S. at 156. The Court of Appeals recognized this principle when it instructed the District Court that any remedy "should be tailored to fit the wrong creating the occasion for the remedy," Microsoft, 253 F.3d at 107, i.e., that it should be focused on "the conduct [the court] has found to be unlawful" and should be limited to provisions that are "required to rectify [Microsoft's] monopoly maintenance violation," id. at 104, 105.(1)
Consistent with these principles, since at least 1911 it has been the law in antitrust cases that "ordinarily ... [an] adequate measure of relief would result from restraining the doing of such [illegal] acts in the future." Standard Oil Co. v. United States, 221 U.S. 1, 77 (1911) (emphasis added). In other words, an injunction simply forbidding the specific conduct found to be illegal is ordinarily considered sufficient. Or, as the District Court recently explained, "the scope of any proposed remedy must be carefully crafted so as to ensure that the enjoin[ed] conduct falls within the . . . behavior which was found to be anticompetitive." Hearing Tr., Sept. 28, 2001, at 8.
Some commentators have nevertheless argued that the District Court is obligated to "terminate" Microsoft's dominant market position, which they characterize as an "illegal monopoly." Jennifer Bjorhus, Settlement Draws Frustration From Few Tech Giants That Spoke Out, San Jose Mercury News, Nov. 3, 2001, at 20A. But this argument rests on a misinterpretation of the pertinent case law, including the Court of Appeals' decision.
Contrary to this argument, the law does not require that a remedy attempt to recreate the world as it might have existed absent the violation or deprive a defendant of the proceeds of its business. Instead, where a violation is found, the remedy, as the Court of Appeals pointed out, should be designed to '''unfetter'" the market from the "'anticompetitive conduct.'" Microsoft, 253 F.3d at 103 (quoting Ford Motor Co. v. United States, 405 U.S. 562, 577 (1972)) (emphasis added). That, moreover, is why the Court of Appeals placed heavy focus on the requirement that, before a court can seek to "undo" an alleged monopoly, there must be a "significant causal connection" between the allegedly illegal conduct and the existence of that monopoly. The District Court recently echoed this same theme when it remarked that it intended to fashion an injunction that would "avoid a recurrence of the violation and . . . eliminate its consequences." Hearing Transcript, Sept. 28, 2001, at 9 (emphasis added).
There is a vast difference between "unfettering" or "unshackling" a market from prior anticompetitive behavior, and attempting to reconstruct the market as it might have existed absent that conduct. The former is a legitimate objective of an antitrust remedy; the latter is not. In the District Court's words, attempting to reconstruct the market as it might have been absent the conduct at issue goes well beyond simply "eliminating the consequences" of anticompetitive conduct. Antitrust law does not attempt to recreate or to maintain by detailed regulation a perfect world. Its goal is to restore competition, including legitimate competition by the dominant firm. Ford Motor Co. v. United States, 405 U.S. 562, 577-78 (1972).
Another "'traditional principle of equity jurisdiction,'" Grupo Mexicano, 527 U.S. at 319, is that any relief imposed by a court should not inflict unnecessary harm on third parties. Atchison, Topeka & Santa Fe Railway Co. v. Wichita Board of Trade, 412 U.S. 800, 824 (1973) (plurality opinion); Gilbertville Trucking Co. v. United States, 371 U.S. 115, 130 (1962). In this case, there is a real risk of harm to the entire IT industry as well as consumers. As explained in the attached affidavit of ACT's president, Jonathan Zuck, (Exh. A) both consumers and IT companies derive enormous benefits from the existing Windows platform. For IT companies in general, and ACT's members in particular, that platform is unusually valuable and important for at least three reasons.
1. Constant Improvement and Addition of New Features and Functionalities. One reason Windows is so valuable to the IT industry is that Microsoft has constantly improved it. For example, as Mr. Zuck explains, each new release of Windows contains software drivers for the major new printers and other peripheral devices that have been released since the prior version of Windows. This means that developers of applications such as money management software, graphics programs, etc., do not need to create their own drivers for these devices or, worse, choose from among several competing drivers. Affidavit ¶ 7.
Virtually everyone in the IT industry, moreover, has a strong interest in seeing this trend continue in the future. The addition to Windows of such new functionalities as voice recognition, for example, will allow software developers to add such features to their products at minimal cost. Those costs will increase dramatically and consumer benefits will be reduced if software developers are forced to develop their own voice recognition features or, worse, to "port" their programs to several competing voice-recognition programs. Id. ¶ 8.
2. Windows' Uniformity and Widespread Acceptance. Uniform standards are likewise crucial to an efficient, rapidly evolving IT sector. As Mr. Zuck explains, communications and Internet standards provide the language necessary for many different computers to "talk" or "network" with one another, enabling, for example, users of the World Wide Web to locate and retrieve the information they seek. Operating systems perform a similar function, allowing hardware devices and software applications to communicate with a computer. Indeed, it is Windows' consistency that makes it so valuable.
As the Court recognized in its Findings of Fact, Windows exposes a set of "'application programming interfaces'" that lets software interact in a consistent way with any Intel-compatible PC. United States v. Microsoft Corp., 84 F. Supp. 2d 9, 12-13 (D.D.C. 1999) ("Findings of Fact"). This means that the same software will run on all Windows-based PCs and, by and large, all hardware devices can be used as well. Affidavit ¶ 10. Hence, the consumer avoids the need for time-consuming, often expensive retraining, and thus has a greater incentive to learn how to use the existing system. Also, the widespread acceptance that Windows enjoys makes it easier to ensure that computer products (both hardware and software) work the way they are supposed to, and work well with each other. Operating system consistency usually means that software will operate normally even if the type of computer changes. For example, WordPerfect will function as advertised on a Windows-based Dell computer or a Windows-based Compaq computer. Id.(2)
For these reasons, as Mr. Zuck explains, the cost per potential customer of developing a piece of software for the Windows operating system is significantly lower than the cost for the UNIX operating system. And that, of course, translates into more software and lower prices for consumers. Id. ¶ 13.
In addition, more than any other operating system, Windows has remained compatible with software written for older Windows versions. As a result, consumers have much greater confidence that the software they purchase will work when they upgrade to a new Windows release. Hardware manufacturers and developers similarly face much less risk that their research and development expenditures will be stranded if Microsoft releases a new version. Id. ¶ 14.
3. Windows' Low Cost to Consumers. The Windows operating system also allows the developer, or other providers of support services, to support end-users at minimal cost. As Mr. Zuck explains, each operating system not only has signature application interfaces and user commands, it also presents its own set of bugs and system errors. Thus, to provide software or hardware support, a developer must train personnel to identify and understand the idiosyncrasies of each operating system under which it markets its product. These increased support costs increase prices and decrease consumer demand for products and services. Id. ¶ 15.
Consumers, moreover, obtain all of these benefits inexpensively. Compared to the cost of a typical PC, and to the cost of the software typically installed on that PC, the cost of Windows (at about 5% of the PC's price) is relatively small. A low price, coupled with all the benefits stemming from Windows' widespread use, drives up demand by making computer products more affordable and attractive to consumers. Id. ¶ 16.
As Mr. Zuck explains, the widespread use of an inexpensive, constantly evolving operating system is particularly important in an industry as dynamic as the information technology industry, which constantly generates both new products and new uses for those products, and for which new developments such as the Internet can redraw the competitive landscape overnight. A popular operating system like Windows allows consumers and developers to act quickly and with confidence that software and hardware will work on most PCs today and in the future. And the fact that many consumers choose Windows adds a measure of stability to a highly dynamic industry.
For all these reasons, any "remedy" that resulted in the balkanization of Windows would have a disastrous effect on the entire IT industry. Software developers, Internet access providers, and others rely on the widely installed, constantly improving Windows platform as the groundwork for their own products. If there were no consistent platform, software developers would have to try to "port" their products to various operating systems, increasing those products' costs substantially, or else they would have to accept a much smaller market share. This, too, would drive up prices because the cost of distributing software is tiny compared to the cost of developing it.
Windows' importance as a consistent platform is illustrated by the fact that, when it appeared that Microsoft might be broken up, stock prices in the rest of the IT industry fell. Kenneth G. Elzinga, David S. Evans, Albert L. Nichols, United States v. Microsoft: Remedy or Malady?, 9 Geo. Mason L. Rev. 633 (2001). Likewise, any remedy such as those proposed by the Litigating States that would fragment Windows would be unlawful because of the harm it would impose on third parties.
Finally, any remedy should be judicially administrable and not put the courts in the position of having to "oversee product design." United States v. Microsoft Corp., 147 F.3d 935, 948 (D.C. Cir. 1998). Some have suggested that the kinds of extreme remedies proposed by the Litigating States are in some sense alternatives to regulation. But history suggests quite the opposite.
In 1982, for example, AT&T entered into a consent decree designed to remedy what the government perceived as anticompetitive practices, and to allow AT&T to compete in new markets. Then too, the provisions of that decree were touted as an alternative to regulation. But in practice, the break-up of AT&T generated pervasive judicial participation in the telecommunications industry. For example, between 1984 and 1995, the court ruled on over 250 waiver requests pursuant to the consent decree. Most of these were necessary to allow the companies spun off from AT&T to respond to market developments that had not been anticipated when the decree was entered. Although 96 % of the requests were eventually approved, the average delay prior to approval was four years. It is not surprising, then, that Congress put the court out of the telecommunications business when it passed the Telecommunications Act of 1996.
This kind of intrusive, time-consuming regulation is particularly ill-suited to a rapidly-changing industry such as IT. For example, many settlement opponents have made proposals resting on a distinction between "middleware" and the "operating system." But this distinction is dubious even now, and is rapidly being eroded. The federal courts are not equipped to draw lines in the shifting sands of information technology.
Notwithstanding this reality, some settlement opponents have proposed "ongoing regulation of Microsoft's conduct," or detailed enforcement provisions envisioning ongoing judicial involvement in Microsoft's management. Some have even proposed egregious private attorney general provisions that would simply foment litigation and enrich plaintiff's lawyers. All of these proposals would create the kinds of problems that arose in abundance in the wake of the AT&T consent decree.
Other cases demonstrate the dire consequences that can arise when courts attempt to regulate an industry under the guise of an antitrust decree. For example, in United States v. United Shoe Machinery Corp., 110 F. Supp. 295 (D. Mass. 1953), aff'd, 347 U.S. 521 (1954), the district court imposed extensive regulation on the shoe machinery industry over a ten-year period. The remedies were meant to end United's practice of distributing shoe machinery through long-term leases and to make shoe machinery available from a variety of sellers. To this end, the court restricted lease terms, required United to offer its machines for sale in addition to leasing them, and required United to charge separately for services such as repairs. Id. at 352-53. However, a 1993 study concluded that the court order destroyed many efficiencies arising out of the technical realities of the shoe manufacturing industry, impaired the quality of United's performance, and likely contributed to the dramatic decline of the domestic shoe industry in the 1960s and beyond. Scott E. Masten & Edward A. Snyder, United States v. United Shoe Machinery Corp.: On the Merits, 36 J.L. & Econ. 33 (1993); see also Lino A. Graglia, Is Antitrust Obsolete?, 23 Harv. J.L. & Pub. Pol'y 11, 17 (1999).
For all these reasons, judicial regulation of the IT industry, or any portion of that industry, is to be avoided at all costs.
Indeed, that appears to be the main message of the D.C. Circuit's earlier decision rejecting the preliminary injunction that the Government sought. Microsoft, 147 F.3d at 948 ("Antitrust scholars have long recognized the undesirability of having courts oversee product design, and any dampening of technological innovation would be at cross-purposes with antitrust law."). And the Court of Appeals' most recent decision is entirely consistent with that message. Microsoft, 253 F.3d at 101-07. Indeed, even Judge Jackson has acknowledged that in this case, as in others: "'The less supervision by this court, the better.'" John R. Wilke, For Antitrust Judge, Trust, or Lack of It, Really Was the Issue, Wall St. J., June 8, 2000, at A1.
On balance, the RPFJ complies with these four principles and is therefore in the public interest. Like most settlements, it is less than perfect. However, the purpose of this proceeding is not to produce a perfect order. The court must review the settlement that the parties have agreed to, and enter it "so long as the proposal falls 'within the reaches of the public interest.'" Microsoft, 56 F.3d at 1458 (D.C. Cir. 1995) (emphasis in original; citations omitted); see 15 U.S.C. § 16(e) ("Before entering any consent judgment proposed by the United States under this section, the court shall determine that the entry of such judgment is in the public interest.").
It is clear that entry of the RPFJ is in the public interest. The federal government has explained at length in its Competitive Impact Statement that the RPFJ will provide a "prompt, certain and effective remedy for consumers" by enjoining the conduct that the Court of Appeals found to be illegal, and by restoring competitive market conditions. Competitive Impact Statement at 2, United States v. Microsoft, No. 98-1232 (D.D.C. Nov. 15, 2001) ("CIS"). Each of the Court of Appeals' findings of anticompetitive conduct is addressed by at least one provision of the proposed final judgment. See Exh. B (table showing which provisions address each finding of illegality). Indeed, the RPFJ's provisions regarding "server protocols," and its enforcement provisions, extend beyond the anticompetitive conduct found by the Court of Appeals. Accordingly, any notion that the RPFJ only tells Microsoft "to go forth and sin no more," United States v. Microsoft Corp., 159 F.R.D. 318, 334 (D.D.C. 1995), rev'd, 56 F.3d 1448 (D.C. Cir. 1995), is ludicrous.
In contrast, the Litigating States and other critics of the RPFJ have proposed a variety of radical "remedies" that they claim would be more effective than the RPFJ in restoring competition. However, these proposals violate the four principles described above, and are in fact designed to benefit Microsoft's competitors. Indeed, these proposals would advantage Microsoft's competitors in areas other than PC operating systems, which is the only market at issue in this case. Moreover, rather than seeking to restore competition, these proposals and others like them seek to impose a court-designed, court-regulated regime that is especially inappropriate for a rapidly changing area such as IT. A principle-by-principle analysis highlights the flaws in these proposals.
As noted above, the most vital principle in designing an antitrust remedy is that it must be designed to benefit consumers rather than competitors. Unlike the Litigating States' proposals, the RPFJ easily complies. Consumers will benefit from the guaranteed flexibility and choice provisions in the RPFJ. All new Microsoft operating systems, including Windows XP, will have to allow end users to readily remove or re-enable Microsoft's middleware products such as its Internet browser, instant messaging tools, media player, and email utilities. While end users can already remove Microsoft middleware from Windows XP, the RPFJ will make it easier for users to switch and compare among competing middleware products, including those installed by computer manufacturers and those readily accessible over the Internet.
Most importantly, the RPFJ preserves the integrity of the Windows standard while making it easier for other platforms to compete with Windows. As discussed above, the "network effects" that characterize the operating system market mean that consumers and the IT industry both benefit when they know that the platform they rely on is widely used, and will continue to be widely used in the future. Findings of Fact at 19-23; see also Affidavit ¶¶ 9-14. By and large, the RPFJ avoids requirements that would encourage the emergence and sale of multiple, incompatible operating systems under the Windows brand name.
At the same time, the RPFJ protects Microsoft's competitors in several ways. Most importantly, it forbids retaliation against OEMs, § III.A, requires uniform license terms for the twenty largest OEMs, § III.B, and prevents Microsoft from including various restrictive provisions in OEM licenses, § III.C. Thus the RPFJ opens up the valuable OEM distribution channel to competitors, addressing the Court of Appeals' most substantial concerns. By increasing competitors' access to OEMs and by preventing Microsoft from negotiating quotas with IAPs, the RPFJ reasonably ensures that consumers will have access to whatever products they want.
By contrast, a central thrust of the Litigating States' proposals is to break Microsoft's control over the "Windows" brand. Forcing Microsoft to break up Windows into what a court conceives of as its component parts both destroys the utility of the standard Windows platform and entangles judges in a maze of technical regulation that they are poorly equipped to solve. If implemented, the LSPFJ would result in the creation of as many as 4,000 different versions of Windows, each requiring support not only by Microsoft but also by OEMs, software developers, and other IT professionals. This outcome would worsen, not improve, the lot of consumers. It would only serve to weaken Microsoft's product offerings, confuse users, drive up prices, and limit software choices.
Such remedies would also create concerns about privacy and security. Consumers are concerned and rightly so about on-line privacy and the security of their electronic information. E.g. David Ho, Identity Theft Tops Fraud Complaints, Wash. Post, Jan. 24, 2002 at E4. Because Microsoft would have almost no control over access to its code and to its technical information under the states' plan, hackers and other unsavory characters would find it much easier to penetrate the most common privacy and security protections. It would also be harder for Microsoft to control computer piracy, which in the end drives up prices to consumers.
By making the fruits of Microsoft's innovations readily available to competitors, the Litigating States' proposals would also harm consumers by reducing Microsoft's incentive to innovate in the future. Indeed, it is likely that Microsoft's research and development budget, which has historically been the largest in the industry, would be substantially reduced to the detriment of consumers. Property ownership is the cornerstone of a free market system; as property rights are eroded, so is the incentive to put that property to its most valuable use.
Beyond these problems, the Litigating States' proposals are patently designed to provide specific benefits to Microsoft's principal competitors, and to reinforce their dominant positions in markets that are irrelevant to this litigation. This approach to remedies is contrary to the interests of consumers and the rest of the IT industry, and contrary to antitrust law.
Benefits to AOL Time Warner. Some of the Litigating States' proposals will directly benefit AOL Time Warner. For example, the Litigating States' proposal to break Microsoft's control over the "Windows" brand, and the proposed prohibition on making Microsoft middleware the default for any functionality, LSPFJ § 10, unless the OEM or other licensee can override the setting and designate a different default or give the end-user a "neutrally presented choice" means that consumers who think they are buying a coherent, integrated operating system designed by Microsoft will get something quite different.
To see how this benefits AOL, consider the following scenario: AOL's "Magic Carpet" service will compete with Microsoft's .Net services. If Microsoft designates .Net as a default service in Windows, AOL can ask computer sellers to re-direct the default to Magic Carpet. Indeed, AOL's strategy is to do just that. Alec Klein, AOL to Offer Bounty for Space on New PCs, Wash. Post, July 26, 2001, at A1 ("In internal AOL documents, the media giant lays out a strategy that calls on manufacturers to build into their new personal computers icons, pop-up notices and other consumer messages aimed at pushing aside Microsoft by giving AOL's own products prominent placement on PCs. It's the latest foray in an intensifying feud between the two technology titans over consumers and supremacy on the Internet.") Yet this hybrid product will still be marketed as a Windows system, making Microsoft responsible in consumers' eyes for programs it has no control over, and giving AOL a free ride on Microsoft's reputation and marketing.
Other users will be provided with a bewildering array of choices, all presented in a "neutral" manner, i.e. without guidance as to what product best suits their needs. Yet sophisticated users who have information about middleware alternatives do not need "neutrally presented choices" to help them make their decisions. Less sophisticated consumers are entitled to get the brand they paid for, or at least to be told how to get that brand. The RPFJ's Section III, by contrast, puts Microsoft and its competitors on a level playing field, with minimal judicial intervention.
Benefits to Sun Microsystems. Another Microsoft rival, Sun, would also benefit directly from the Litigating States' proposals. Sun would benefit most obviously from the proposal that Microsoft include Sun's Java with every copy of Windows. LSPFJ § 13. Apparently Sun sees no conflict between that proposal and the proposal that Microsoft make available middleware-free versions of Windows at reduced prices. It is hard to argue that this requirement would benefit consumers, who can already get Sun's Java free from those web sites that use it. The federal government's settlement with Microsoft will make Sun's Java even easier for consumers to obtain by allowing OEMs, IAPs, and ISVs to provide it to their customers without fear of retaliation. But under the Litigating States' proposal, all consumers would have Sun's Java forced on them.
Benefits to IBM and Apple. The Litigating States' proposals also benefit IBM and Apple, giving them each an Office suite. IBM wants Office for Linux, and under the Litigating States' proposal it will get its wish by snatching Microsoft Office source code at the auction price. Under that proposal, Microsoft must maintain and support Office for the Macintosh even if it is a money-losing proposition. And if Apple is unhappy with the Office support Microsoft has to provide, it can snatch the source code at auction, and have an Office all its own. LSPFJ § 14. These "porting" proposals go far beyond the scope of this case, which is the Windows operating system market.
Conversely, the federal government's settlement with Microsoft addresses the Court of Appeals' only holding of anticompetitive behavior involving Apple, namely the agreement that Apple would distribute Internet Explorer exclusively. Under the RPFJ, Apple, like all ISVs, is free to distribute and promote non-Microsoft platform software without fear of retaliation. The states' proposal would give a free ride to a handful of companies and would impose an unnecessary burden on Microsoft but would not benefit consumers.
The states' proposals also provide free source code for Microsoft's Internet Explorer, LSPFJ § 12, giving IBM a good browser for the entire line of IBM computers and Apple a leg up on its software design. But once again, the problem with all this generosity is that its sole purpose is to benefit competitors and harm Microsoft, not to benefit consumers.
Another key flaw in the Litigating States' proposals is that they go well beyond the Court of Appeals' ruling. Indeed, the sweeping scope of the Litigating States' proposals suggests that they mistakenly read the Court of Appeals' decision on liability as a broad affirmance, rather than as it was in fact a reversal in part containing very precise, narrow holdings on liability. Indeed, the D.C. Circuit reversed the District Court's findings that Microsoft had committed attempted monopolization and illegal tying.
As to the remaining findings, the Court of Appeals affirmed only some of the District Court's findings that Microsoft had illegally maintained its monopoly. Microsoft, 253 F.3d 34. The Court of Appeals held that some exclusionary contracts and negotiating tactics were unlawful; that Microsoft had acted illegally in deceiving developers about its own Java language; and that Microsoft had illegally excluded Internet Explorer from its Add/ Remove facility and intermingled its Internet Explorer and operating system code. The Court also emphasized that, on remand, the District Court "must base its relief on some clear 'indication of a significant causal connection between the conduct enjoined or mandated and the violation found directed toward the remedial goal intended.'" Id. at 105 (quoting 3 Philip E. Areeda & Herbert Hovenkamp, Antitrust Law ¶ 653(b), at 91-92 (1996)).
Section III of the RPFJ addresses each of these holdings. As to exclusionary contracts and high-pressure negotiations, the RPFJ forbids Microsoft to retaliate against OEMs, § III.A; requires Microsoft to sell Windows to the twenty largest OEMs under uniform license terms, § III.B; and forbids retaliation against, or exclusionary agreements with, ISVs or IHVs, § III.G, § III.F. As to Java, the RPFJ requires disclosure of information needed to design other software to be fully compatible with Windows, § III.D, and requires Microsoft to license its intellectual property to rivals, § III.I. As to Internet Explorer, the RFPJ forbids Microsoft to restrict any OEM from modifying their computer interfaces in various ways, such as removing the Internet Explorer icon, § III.C, and requires Microsoft to allow end-users to remove access to Microsoft Middleware or to designate a non-Microsoft middleware product as the default instead of the Microsoft product, § III.H.
The Court of Appeals was also quick to note that much of the conduct that Microsoft was accused of and even conduct that was found to be anticompetitive in particular settings is common in business, and is usually not anticompetitive. But the states' proposed categorical bans sweep in a host of pro-competitive conduct, in disregard of the Court of Appeals' instruction that any remedy be "narrowly tailored" to specific holdings of illegality.
For example, the states would ban exclusive dealing across the board. Yet the Court of Appeals explained that:
Microsoft, 253 F.3d at 70.
Similarly, the proposed judgment reflects an implacable hostility to integrating an internet browser or any additional functionality with the basic Windows operating system. Yet, as the Court of Appeals observed, "[a]s a general rule, courts are properly very skeptical about claims that competition has been harmed by a dominant firm's product design changes." Id. at 65.
In perhaps the Litigating States' most egregious proposal, Sun CEO Scott McNealy got a special gift he has always wanted, see Peter Burrows, Face-Off, Bus. Wk., Nov. 19, 2001, at 104, -- the ability to stop Microsoft from buying anything that could help it compete with Sun. If Microsoft wants to make an acquisition, an investment, or an exclusive license, it must notify the plaintiff states' attorneys two months in advance. LSPFJ § 20. The states make this proposal despite the total absence of any takeover-related findings anywhere in this case.
It was precisely this type of overreaching that the Court of Appeals soundly rejected in 1995, when it reversed Judge Sporkin's refusal to approve the federal government's settlement with Microsoft and reassigned the case to a different district judge. Microsoft, 56 F.3d 1448. Judge Sporkin had gone beyond the complaint to try to force the parties to address his own concerns about "vaporware." The Court of Appeals found that effort inappropriate. And it is no more appropriate for the Litigating States, at this late date, to try to drag in new issues and punish Microsoft for conduct that it never had a chance to defend. If a "claim is not made, a remedy directed to that claim is hardly appropriate." Id. at 1460.
Another example of overreaching is buried in the Litigating States' proposals regarding orders and sanctions, and which singles out for punishment any "groundless" claim Microsoft makes of intellectual property infringement. Again, Microsoft's conduct in intellectual property litigation is no part of this case.
Finally, the Litigating States' proposed ban on retaliation against those who participated in the litigation is not grounded in any finding of illegality, even though Microsoft has been enmeshed in antitrust cases for years and has presumably had ample opportunity to retaliate unhindered. The RPFJ retaliation ban, in contrast, is clearly aimed at the possibility that Microsoft might try to punish companies that do not cooperate with Microsoft's business goals. The Court of Appeals envisioned that Microsoft would continue its normal business relations, albeit with injunctions in place against specific conduct found to be anticompetitive. The RPFJ provision implements that vision, while the states' proposal would open the door to unfounded claims of retaliation by any disgruntled participant in the litigation.
Of course, the RPFJ itself is overbroad in some respects.(3)Yet despite these problems with its scope, it is clear that as a whole, the RPFJ falls "within the reaches of the public interest." It addresses the Court of Appeals' findings of illegality, remedies them all, and ensures competitive conditions in the market for Intel-compatible PC operating systems.
The RPFJ also offers significant advantages to the IT industry. Most importantly, of course, it preserves the integrity of Windows. But it also serves the IT industry by achieving a relatively quick resolution of this dispute. Litigation over remedies, possibly followed by appeal and remand or further appeal, could take years. The Supreme Court has recognized that a government antitrust consent decree is a contract between the parties to settle their disputes and differences, United States v. ITT Continental Baking Co., 420 U.S. 223, 235-38 (1975); United States v. Armour & Co., 402 U.S. 673, 681-82 (1971), and "normally embodies a compromise; in exchange for the saving of cost and elimination of risk, the parties each give up something they might have won had they proceeded with the litigation." Armour, 402 U.S. at 681. The RPFJ has the virtue of bringing the IT industry certain benefits and protections without the uncertainty and expense of protracted litigation, Armour, 402 U.S. at 681; Microsoft, 56 F.3d at 1459; it will provide "prompt, certain and effective remedies," CIS at 3.
The RPFJ also directly helps OEMs and other IT firms. Many of the options that will benefit consumers will also benefit the companies they buy from. As discussed above, OEMs that equip their products with any Microsoft operating system will benefit from guaranteed flexibility under the RPFJ. The twenty largest OEMs will also be entitled to uniform licensing terms, with some flexibility for volume discounts and marketing allowances. OEMs will have the ability to "lease" desktop space as well as "space" in the boot sequence on their computers by installing or promoting non-Microsoft products and services; IT companies will thus have the option to negotiate with the OEM(s) of their choice for that "space".
By contrast, the states' proposal to give the OEMs the choice of which parts of Windows to include on their computers and forcing Microsoft to accommodate those choices would fragment the Windows standard. As explained above and in Mr. Zuck's affidavit, such fragmentation would have disastrous effects. Creating multiple versions of Windows would slow the release of new versions of Windows and would make it impossible for software developers to program with confidence. Either they would write only to the leanest version available, depriving consumers of the benefits of most of Windows' functionality, or they would have to write multiple versions of each program, substantially increasing development costs and customer confusion. A stagnant, fragmented Windows would hurt the entire industry.
On another front, the RPFJ benefits all IT providers, including Microsoft's competitors, by guaranteeing access to technical specifications. Microsoft would have to promptly disclose technical information that enables any Windows operating system to communicate with Microsoft servers and with all Microsoft middleware products. §§ III.D, III.E. To encourage more non-Microsoft middleware, the settlement forces Microsoft to license any intellectual property rights that others might need to compete with Microsoft. § III.I. And as with OEMs, Microsoft could not penalize any software developer, service provider, or hardware vendor that develops or sells products that compete with Windows and Microsoft middleware. §§ III.A, III.F.
By contrast, as discussed above, the Litigating States' proposals would stifle innovation further by weakening or entirely eliminating Microsoft's intellectual property rights, thereby reducing its incentive to innovate. E.g. LSPFJ §§ 1 (stripping down Windows), 2(a) (mandatory licenses), 3 (mandatory licensing of predecessor versions), 4 (disclosure of APIs and technical information), 12 (giving away browser), 14 (mandatory porting), 15 (intellectual property licenses), 19(f) (intellectual property claims). These provisions would not only hurt Windows, but also would instill in any sensible IT executive the fear that success will lead to confiscation.
Even if these proposals did not end Microsoft's improvements to Windows, another provision would likely do so. That is the Litigating States' proposal to require Microsoft to notify any ISV of non-Microsoft middleware of any planned action, sixty days in advance, if the action will interfere with the middleware's performance or compatibility with Windows, unless the action is taken for good cause. LSPFJ § 5. After the notification, the ISV could complain to Microsoft's court-installed regulators to try to block the change.
The states' broad prohibitions on exclusive dealing and on agreements limiting competition also would prohibit Microsoft from entering into joint ventures with any other members of the IT industry. Because IT products are so interdependent, both consumers and companies would suffer if the only option is to design around Microsoft products, and the option of collaborating with Microsoft on entirely new projects is excluded.
Finally, the approach of the RPFJ is not unduly regulatory. To be sure, the enforcement mechanism is too intrusive and could be substantially improved. However, the substantive provisions of the RPFJ focus on improving competition rather than micromanaging markets or product design. Thus, most of the injunctions tell Microsoft what not to do, rather than imagining what a perfect competitor might do and then attempting to enforce that vision.
Not so the proposal by the Litigating States. They have proposed ongoing regulation of Microsoft's conduct, including ongoing judicial involvement in Microsoft's management, by a "special master" who would serve as an investigator, prosecutor, judge, and potentially even witness against Microsoft. LSPFJ § 18. The special master would be free to receive and act on even anonymous complaints, again a procedure that the Court of Appeals harshly criticized when Judge Sporkin used it. Microsoft, 56 F.3d at 1464. These proposals are most likely unlawful, if not unconstitutional. Id.; Microsoft, 147 F.3d. at 954 (granting mandamus to vacate non-consensual reference to a special master where "[t]he issue here is interpretation, not compliance; the parties' rights must be determined, not merely enforced"). And in all events, they would allow Microsoft's rivals to thwart competition at every turn.
The Litigating States also err in proposing an unduly long duration period. Any remedy in this case must be sensitive to the rapid pace of technological change in the operating system market. An injunction that is appropriate today may be completely unsuited to tomorrow's market. If, as The Economist has written, operating systems are "no longer central," then there is little point in regulating that market. Microsoft: Extending its Tentacles, The Economist, Oct. 20-26, 2001, at 59. The RPFJ recognizes this reality by limiting its term to five years, with the possibility of a two-year extension. § V. Not so the Litigating States, who in their rush to ask for the most punitive "remedies" available seek a ten-year term for the judgment.
In an effort to cover unforeseeable eventualities, the States also define key terms such as "middleware," "browser," and "technical information" so broadly that the proposed judgment is in some ways absurd. For example, it appears that the "middleware" definition would include parts of Windows 3.0, which was developed before anyone thought of Java or Internet Explorer. Because they are unworkable, many of the Litigating States' proposals invite additional judicial involvement through complaints by competitors or others; indeed, the provisions for anonymous complaints invite not only involvement, but abuse.
In short, the Litigating States' proposals pose an enormous risk of ongoing judicial regulation. Not only would they require substantial modification of Microsoft's internal management structure, but they would require the District Court to set up its own regulatory agency, headed by the special master and potentially including a substantial staff, all paid by Microsoft. Courts are simply not designed for this sort of ongoing regulatory role, particularly in a field as far removed from their expertise as IT. At best, the Litigating States' proposals would create a contentious, judicially-regulated regime in place of a market. At worst, they would seriously impair IT innovation, at everyone's expense.
For all these reasons, the RPFJ should be adopted, and the Litigating States' proposals should be rejected.
1. Normally, of course, a settlement is reached before a trial on the merits. In that situation, it is clear that a reviewing court cannot expand an antitrust decree to remedy perceived problems that lie outside the scope of the complaint. That was the thrust of the Court of Appeals' 1995 Microsoft decision, 56 F.3d 1448. Furthermore, any such action by a reviewing court would most likely be unconstitutional. Id. at 1459; see also Maryland v. United States, 460 U.S. 1001, 1006 (1983) (Rehnquist, J., dissenting). Here, of course, the Court of Appeals has affirmed some of the district court's findings of liability. Expanding the remedy to address issues as to which liability has not been proven let alone issues as to which liability has never been alleged would clearly exceed the District Court's power.
2. In its consistency from one computer and software program to another, Windows is markedly different from the UNIX operating system. That "system" is in reality a collection of similar operating systems, including Sun's Solaris, Digital's UNIX, HP's HP-UX, IBM's AIX and SCO's UnixWare. See http://www.techweb.com/encyclopedia/defineterm?term=unix. Although different versions may be desirable with respect to many products, for most computer users such a proliferation promises nothing but confusion, lost time, fewer applications, and higher prices.
For example, a consumer who shifts from one UNIX-based computer to another UNIX-based computer may find that the two computers use different UNIX versions with different features, functions, and idiosyncrasies. Consequently, the consumer may have to devote considerable time and expense learning how to perform the same tasks on the second UNIX-based computer that she already knew how to perform on the first platform. Worse still, the software applications or hardware equipment she purchased for and used on the first computer may be incompatible with the version of UNIX installed on the second computer. And a UNIX user obviously has less incentive to develop skills tailored to her particular system if it is likely that she will use a different UNIX operating system in the future. Affidavit ¶¶ 11-12.
3. For example, the Proposed Final Judgment defines "Microsoft middleware" as including Outlook Express, photo and video editing software, and other products that cannot serve as competitive threats to Microsoft. RPFJ § VI.K.1. This definition clearly overreaches. This case is about Microsoft's response to the emergence of "middleware" as a competitive threat a possible alternative platform for software developers that could run on a variety of operating systems and thus would make software independent of Windows. Only "middleware" that can serve as an independent basis for software development across different operating systems poses a competitive threat to Windows. Microsoft, 253 F.3d at 53.
Similarly, the RPFJ overreaches when it requires that Microsoft disclose communications protocols used to interoperate with Windows 2000 servers and their successors. The Court of Appeals' definition of the relevant market made it clear that servers are not a part of that market and therefore, that they are not a part of this case. Microsoft, 253 F.3d at 52-53. As explained above, the only connection between servers and this case is that Microsoft's competitors in the server market have been highly influential with the Attorneys General who continue to litigate this case. The server protocols themselves are irrelevant and thus compelling disclosure is both overbroad and designed to benefit competitors rather than consumers.