Erasmus Law and Economics Review 1 (February 2004): 122.

CATCH ME IF YOU CAN! THE MICROSOFT SAGA AND THE SORROWS OF OLD ANTITRUST
Andrea Renda*
LE Lab, Luiss Guido Carli, Rome, Italy Abstract At the intersection between intellectual property protection and antitrust, the Microsoft case has now become a synecdoche, a part for the whole: the browser war has indeed evolved as a marginal skirmish that should be interpreted in the light of a more general IPR war. The paper describes the peculiar aspects and the questions that remain unsolved in the Microsoft case and points at the hidden places of digital capitalism, in search of an approach that reconciles the outstanding potential of e2e architecture with the need to ensure that content producers are able to control the diffusion of contents circulating on the Net. Keywords: Microsoft case, Antitrust, Digital capitalism, Software market. JEL classification: K21, L49.

While the opposed parties get prepared for the next episodes of a saga that lasted so far more than a decade, the US trustbusters seem to hardly resist the temptation to quickly leave the Microsoft case behind by botching up an ephemeral solution, rather than insisting in the quest of a satisfactory compromise between the interests at stake. And even after District Judge Colleen Kollar-Kotelly, the last in charge of solving the controversy, concluded that the consent decree signed by Microsoft, the US Department of Justice and nine of the eighteen plaintiff States is in the public interest, the Microsoft case cannot be assumed to have finally breathed its last.1 There are
*

Senior Research Fellow, LE Lab, Luiss Guido Carli, Rome. Email comments at [email protected]. The entry of the Consent Decree was conditioned to an amendment of Section VII of the agreement. Microsoft then filed a Third Revised Proposed Final Judgment that was

@ 2004 Andrea Renda. Subject to ELER Public License 1.0. URL: http://www.eler.org

Andrea Renda

indeed several issues that remain pending, such as the tying and monopolization claims initially formulated by Judge Thomas Penfield Jackson; the settlement of suits between Microsoft and its major rivals; and, finally, the uncertain outcome of the European litigation.2 After all, the solution devised by the District Judge far from being a last word appears as a quick and temporary one. This sensation is strengthened by the zeal with which the judge urged the parties to engage in negotiations seven days a week and around the clock in order to reach an agreement (consent decree) before the trial, supporting this need even by recalling the tragic events of nine-eleven3. And while the accusations that hit the Redmond-based giant, after the order issued by the Court of Appeals on 7 June 2001, seemed to head back to their original magnitude after the expansion threatened by Judge Jacksons Conclusions of Law on the other hand, the importance of the controversy at stake humongously grew and gradually transformed over time, becoming the expression of a war without frontiers, in which the famous browsers war now simply plays the role of a marginal skirmish.4 In my opinion, it is of critical importance to approach the analysis of the
approved by the Court. Judge Kollar-Kotelly modified the SRPFJ in order to preserve the District Courts jurisdiction in monitoring Microsofts compliance with the terms of the Decree. The text of the Final Judgment, issued by the District Court on November 12, 2002 is available online at http://www.dcd.uscourts.gov/98-1232cq.pdf.
2

For a disquieting illustration of the pending streams of the Microsoft litigation, see e.g. Dan Richman, From AOL to EU, More Legal Battles Loom, available at http://seattlepi.nwsource.com/business/93929_legal.102shtml (last visited: January 4, 2004). The saga is becoming more and more complex even from a chronological standpoint: consider that Judge Jacksons Findings of Fact were recently considered applicable by District Judge Frederick Motz during an oral hearing in the Netscape v. Microsoft case started on January 22, 2002. The hearing took place on October 24, 2002, more than one year after the Court of Appeals had declared Judge Jacksons conduct to be characterized by an appearance of actual bias. Microsofts lawyers and executives are reasonably panicking vis--vis such a daring attempt to exhume already vacated solutions. According to what has been reported by the Wall Street Journal and by the Seattle PostIntelligencer on 25 October 2002, Judge Motz opened the hearing by asking Do I have to sit here and listen to all the facts again? [Judge Jackson] found that Microsoft did some pretty bad things. A similar approach was reproposed by California Superior Court Judge Paul Alvarado, according to which Judge Jacksons Findings of Fact should be binding in all class actions filed in one of the dissenting States. See Peter Kaplan and Reed Stevenson, Microsoft Antitrust Findings Still Apply, Reuters English News Service, 4 November 2002, available at http://www.stern.nyu.edu/ News/news/2002/november/1104reuters.html (last visited: December 18, 2003). I refer to the words pronounced by Judge Kollar-Kotelly in the Order filed on 28 September 2001, available at http://www.dcd.uscourts.gov/98-1232gg.pdf. According to one commentator, the Microsoft case is now similar to an end game, as the final stage of a chess game after most of the pieces have been removed from the board. See Leonard Orland, The Microsoft End Game, Conn. L. Rev., Fall, 2001, 221.

Microsoft Saga and the Sorrows of Old Antitrust

Microsoft case by keeping in mind the emergence of such a broader war, denouncing at the same time the prolonged impasse of US and EU antitrust enforcement and the need to thoroughly revisit a conceptual apparatus in desperate need for modernization, especially when faced with the emergence of whole new sectors of the economy, in which the idea of market in and of itself seems to be losing meaning and appeal, as competitive dynamics exhibit totally new features. For such reason, and without pretending to run rebours through the tortuous trajectory of the litigation already documented by plenty of pages in the legal and economic literature this paper seeks to produce a humble attempt to reduce complexity and pierce the dense veil of ignorance that still covers the open wounds of a querelle that neither lawyers nor economists have so far managed to put to rest. I will then offer some answers to often echoing summae quaestiones: how important is the Microsoft case today? To what extent will it affect the evolution of digital capitalism? Furthermore, is it correct to apply the apparatus of old antitrust to the dynamics of wholly new marketplaces? There is no doubt that the solution in the Microsoft case more precisely, Microsoft IV played a featuring role for many powerful actors of digital capitalism5. And it still does. From many points of view, indeed, Microsofts judicial vicissitudes naively labelled as the case of the century during last century continues to beat all records6. Think of the number of public comments received by the District Court during the 60-day period in which as prescribed by the Tunney Act the text of the agreement was submitted to the public opinion. Not less than 32,000 comments were filed, ranging from contemptuous condemnations to ecumenical proclamations, from the disdained disapproval of Microsofts rivals to the optimistic forecasts of those who glimpsed with scarce farsightedness the imminent end of the uncertainty that dominated the high-tech industry over the whole lifespan of the litigation7.

Microsoft I coincides with the decision issued by District Court Judge Penfield Jackson to enter the consent decree signed by Microsoft and the DoJ in 1994 (U.S. v. Microsoft Corp., 159 F.R.D. 318 (D.D.C.), revd, 56 F.3d 1448 (D.C. Cir. 1995). Microsoft II concerned the first tying claim (between Windows 95 and Internet Explorer, in violation of the 1994 consent decree): Microsoft was first condemned, then partly rehabilitated by the Court of Appeals U.S. v. Microsoft Corp., 980 F. Supp. 537 (D.D.C. 1997), revd, 147 F.3d 935 (D.C. Cir. 1998). Microsoft III coincides with Judge Jacksons Findings of Fact, which included the divestiture remedy aimed at creating the baby bills, a solution later rejected by the Court of Appeals U.S. v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. June 28, 2001). See, i.a., Case of The Century?, Newsweek, 16 November 1998. Consider that the first consent decree signed between Microsoft and the DoJ received only five comments. The proposed final judgment of the famous AT&T case, which marked a new epoch in the US telecommunication sector, generated in 1982 only six hundred comments. (See U.S. v. AT&T, 552 F. Supp. 131, 135 (D.D.C. 1982).

6 7

4 1.

Andrea Renda

But let us proceed step by step. After the Final Order issued by the Court of Appeals8, which partly endorsed and partly rejected Judge Jacksons Conclusions of Law in Microsoft III, averting the perspective of a frankly inappropriate divestiture, two issues remained to be ascertained: a) whether Microsoft violated the antitrust laws in tying its browser Internet Explorer to the dominant Operating System (OS), Microsoft Windows; and b) whether the Redmond-based firm had successfully managed to preserve its hegemony in the sale of OS for Intel-Compatible PCs through aggressive contractual arrangements vis--vis its commercial partners, aimed at deterring them from promoting or distributing rival software, which could allegedly replace Windows as the dominant platform for application developers and final users. The approach adopted by the District Judge for solving such questions, though, relied on negotiations between the parties rather than trial litigation. Indeed, the US Department of Justice (DoJ) and Microsoft eventually and laboriously managed to reach an agreement, thanks also to the effort of a mediator that succeeded where Richard Posner, two years before, had failed9. The agreement was then signed by nine of the eighteen States that had originally filed suit against Microsoft, and was subject to the 60-day public consultation procedure, reaching under the name of Second Revised Proposed Final Judgment (SRPFJ) the benches of the District Court: Judge Kollar-Kotelly was simply in charge of judging its compatibility with the public interest10. In case of a negative judgment, the judge should turn to the alternative remedial proposal formulated by the nine non-settling States, which as I explain below contained far harsher conditions for Gates & Co., even jeopardizing their

For a comment on Jacksons Findings of Fact and Conclusions of Law, see e.g. Kenneth G. Elzinga, David S. Evans, and Albert L. Nichols. United States v. Microsoft: Remedy or Malady?, George

Mason L. Rev., Spring 2001, 633. For a comment in Italian language, see Andrea Renda, Microsoft: Cronaca di una Condanna Annunciata, Il Foro Italiano, 2000, IV, 229.
9

The mediation failed, according to the Judge Posner, because of the obstructionism of the nineteen Plaintiff Statess Attorney Generals. The new mediator, Eric Green (Professor of Mediation at the Boston University School of Law), intervened after Microsoft and the DoJ filed with Judge Kollar-Kotelly a Joint Status Report that was considered as totally insufficient. (The reading is quite amusing. Check it out at www.usdoj.gov/atr/cases/f9000/9085.pdf). The SRPFJ is available at http://cyber.law.harvard.edu/openlaw/msdoj/022702/secondrevised-pfj.pdf (last visited on December 18, 2003). The nine states that accepted the agreement are New York, Ohio, Illinois, Kentucky, Louisiana, Maryland, Michigan, North Carolina and Wisconsin. The Tunney Act (5 of the Clayton Act 15 U.S.C. 16) prescribes at Section (e) the Public Interest Determination, i.e. an assessment of the impact that the proposed settlement will exert on society as a whole and secondarily on the economic interest of the opposed parties. In order to access such procedure, the DoJ had to file a Competitive Impact Statement (CIS) containing such preliminary evaluation. The CIS is available at www.usdoj.gov/atr/cases/f9500/9549.pdf (last visited on December 18, 2003).

10

Microsoft Saga and the Sorrows of Old Antitrust survival in the marketplace.11

The proposed final judgment signed by Microsoft and the DoJ contains far reaching terms, mostly related to contractual arrangements between Microsoft and hardware, software and content providers. Put simply, no structural remedies were contemplated, in line with the approach suggested by the Court of Appeals in June, 2001. According to Attorney General Charles James, the remedies contained in the SRPFJ are prompt, certain and effective, and are apt to restoring effective competition in the relevant market.12 The SRPFJ contains a sort of charter of rights for Microsofts commercial partners, and in particular for original equipment manufacturers (OEMs), the most important distribution channel for OS and middleware products13. Under the wording of the SRPFJ, Microsoft will be forced to leave OEMs and final users the freedom to configure their PCs at will, by installing or removing any software product (hence, including Windows or Internet Explorer)14. OEMs shall then be left completely free to modify the BIOS by inserting in the boot procedure non-Microsoft OS products and by placing on their desktops any icon of browsers, application software or Internet Access Providers (IAP), even when such products are not welcomed by Microsoft. The issue of Microsofts overwhelming bargaining power in dealing with OEMs had already emerged during Microsoft III, in which Microsoft was considered able to limit similar initiatives by OEMs by threatening to modify the license terms or even withdraw the Windows license at once. The SRPFJ brilliantly solves this problem: it is now possible to imagine PCs that automatically run non-Microsoft OS products and carry the icons of competing middleware and application software products right on the desktop: furthermore, the Redmond-based firm now does not have any possibility to discourage the promotion of rival or unwelcome software through abuses of partners economic dependency and manipulation of contractual terms.15 In order to achieve such result, the SRPFJ winks at the essential facility
11

The dissenting States are California, Connecticut, Florida, Iowa, Kansas, Massachusetts, Minnesota, Utah, Virginia and D.C. The text of the proposed final judgment is available at http://cyber.law.harvard.edu/msdoj/states-fj.pdf (last visited on December 18, 2003). See the Competitive Impact Statement, supra note 10, at 2. I refer to the options and alternatives listed at Section III.C.1 of the SRPFJ, supra note 10. As a result, the so-called dual boot procedure will become possible for OEMs and final users. See the SRPFJ, supra note 10, at Section III.H.1. The US law does not include any rule specifically aimed at sanctioning the abuse of economic dependency (or, as is sometimes called, abuse of relative dominant position). However, the economic theory shows that, in many cases, the absence of commercially praticable contractual alternatives and the existence of specific investments leave one of the parties in the abyss of path-dependency and lock-in, providing the counterparty independently of its willingness to exploit such situation the chance to renegotiate the

12 13 14

15

Andrea Renda

doctrine, imposing on Microsoft an obligation to deal with its partners at FRAND (Fair, Reasonable And Non Discriminatory) conditions16. First, Microsoft shall apply the same license conditions (including the royalty level) to all major OEMs (the so called Covered OEMs), and shall publish such license terms on its Web site in order to allow all parties to monitor compliance with this commitment. Microsoft will still be free to grant market development allowances or other discounts, provided that such discounts are based on objective and verifiable criteria that shall be applied on a uniform basis for all Covered OEMs17. Moreover, Microsoft will be bound to disclose the code lines known as Application Programming Interfaces (API) to all hardware, software and content producers as well as Internet access providers that will request them18; this, according to the SRPFJ, for the sole purpose of interoperating with a Windows Operating System Product, hence only to ensure interoperability with Microsofts software platform19. The same approach is adopted also for the communications protocols that optimize communication between a client computer and a central server: Microsoft shall disclose such protocols to software developers, again in order to ensure interoperability, allowing Windows to communicate with both Microsoft and nonMicrosoft servers.20
contract terms imposing patently unbalanced conditions. See, for the economics of the Italian rule on the abuse of economic dependency (article 9 of Law 192/1998), Andrea Renda, Esito di Contrattazione e Abuso di Dipendenza Economica: un orizzonte pi Sereno o la Consueta Pie in the Sky?, Rivista di Diritto dellImpresa, 2000. Note that the SRPFJ significantly limits Microsofts ability to threaten to withdraw from the contractual relationship, by prescribing that Microsoft shall not terminate a Covered OEMs license for a Windows Operating System Product without having first given the Covered OEM written notice of the reasons for the proposed termination and not less than thirty days opportunity to cure. See the SRPFJ, supra note 10, at III.A.
16

On the implicit applicaton of the Aspen/Kodak rule in Microsoft, see Renda, supra note 8, at 223-224. See the SRPFJ, supra note 10, Section III.B.3.b. For the purposes of Section III.D of the SRPFJ, APIs are defined as the interfaces, including any associated callback interfaces, that Microsoft Middleware running on a Windows Operating System Product uses to call upon that Windows Operating System Product in order to obtain any services from that Windows Operating System Product. See the SRPFJ, supra note 10, at Section III.D. These obligations arise after the relase of Service Pack 1 for Windows XP (or, if earlier, three months after entry of the SRPFJ). For middleware and OS products, the prescription is even more stringent: according to the text of the SRPFJ, [i]n the case of a new major version of Microsoft Middleware, the disclosures required by this Section III.D shall occur no later than the last major beta test release of that Microsoft Middleware. In the case of a new version of a Windows Operating System Product, the obligations imposed by this Section III.D shall occur in a Timely Manner. This prescription is way more understandable than the corresponding term included in the

17 18

19

20

Microsoft Saga and the Sorrows of Old Antitrust

The two issues that remained pending technological tying and monopolization are tackled with an acceptable set of remedies. On the one hand, the SRPFJ avoids the imposition of mandatory (technological) unbundling between the OS and the browser, a solution which was particularly undesirable, since the supply of an integrated product as I explain in more detail below is welfare-enhancing for nonprofessional users, and as such would deserve to come out immune from the Caudin Forks of the rule of reason21. On the other hand, as far as the monopolization claim is concerned, many authors (including myself) already highlighted the virtues of a contractual remedy a few years ago, during the divestiture era22. For the purposes of this paper, it is enough to recall that evoking the behavioralist vein of antitrust in the case at hand presents the further advantage of preserving consistency and continuity with the wording used by the Court of Appeals in June 2001: after all, the SRPFJ was drafted as a germination of that Order. The SRPFJ can be usefully read from the viewpoint of the obligations to which Microsoft will not be bound. First, the wording of the agreement debases but only partially the possibilities for Microsoft to exploit its enormous intellectual property potential: although considered as an essential facility holder, the Redmond-based firm will indeed be able to invoke copyright and (when applicable) patent protection in order to stop third-party appropriation of its code lines, provided that such activity does not contrast with the prescriptions contained in the SRPFJ.23 As I recall later in this paper, this clause might significantly affect the overall ongoing debate on the limits and boundaries of intellectual property protection in high-tech markets. Moreover, Microsoft will have the possibility to avoid disclosing or licensing its APIs and protocols whenever such disclosure or licensing is likely to compromise the security of its OS, in particular for anti-piracy and anti-virus systems installed therein. Microsoft shall also be able to refuse licensing APIs and protocols whenever the would-be licensee has no reasonable business need for requesting the license, or has a history of software counterfeiting or piracy or willful violation of intellectual property rights24.
previous proposed final judgments. See the United States Memorandum Regarding Modifications Contained in the Second Revised Proposed Final Judgment, available at www.usdoj.gov/atr/cases/f10100/10144.pdf (last visited: 18 december 2003), at 6.
21

On technological tying see, i.a., G. Sidak, An Antitrust Rule for Software Integration, 18 Yale J. Reg. (2001). See Andrea Renda, Microsoft: Cronaca di una Condanna Annunciata, supra note 8. See the SRPFJ, supra note 10, Section III.A and III.F.3. See the SRPFJ, supra note 10, Section III.J.2. According to the Free Software Foundation, this term of the agreement was formulated vaguely in order to leave Microsoft with stronger contractual freedom. See Eben Moglen, FSF Statement in Response to PRFJ in Microsoft v.

22 23 24

Andrea Renda

Finally, the SRPFJ introduces a complex control mechanism for monitoring Microsofts compliance with the agreement, which includes the appointment of a Compliance Committee composed by at least three independent experts in the field of software and programming, which will be in charge of assisting the Court in monitoring compliance. The Compliance Committee will then be supported by an Internal Compliance Officer, in charge of facilitating the Committee in the course of its activity and periodically reporting to the Committee on Microsofts compliance with the SRPFJ.25 Furthermore, the SRPFJ allows plaintiff States to access during normal office hours to inspect any and all source code, books, ledgers, accounts, correspondence, memoranda and other documents and records in the possession, custody, or control of Microsoft, which may have counsel present, regarding any matters contained in this Final Judgment26. 2. The text of the SRPFJ, no surprise, was subject to a fierce debate.27 Gazillion stakes were involved, as the antagonists were in quest of long-run effects in a market in which such a time horizon ends up being inevitably nebulous. Competitors, no doubt, would have preferred that the District Judge issue a remedial judgment imposing both structural and behavioral remedies, with the effect of profoundly devitalizing the competitive potential of the Redmond-based giant. The alternative proposal filed by the nine dissenting States patently moved in this direction, stirring up the unconditioned approval of Open Source advocates and of those rivals that though overtly determined to protect their intellectual property rights would have benefited from such a solution by finding new competitive thrust and trust28.
US, available at www.gnu.org/philosophy/ms-doj-tunney.html (last visited on December 29, 2003).
25 26 27

See the SRPFJ, supra note 10, at Section IV. Id., Section IV.A.2. The last assault to the text of the SRPFJ occurred on November 6, 2002, when Judge Kollar-Kotelly rejected the request filed by a number of academicians close to the open source community and by the Consumers for Computing Choice and Open Platform Working Group to act as amici curiae in the trial. See the order issued by the District Judge at www.dcd.uscourts.gov/98-1232cn.pdf. The Microsoft saga elicited a heated debate on the need to avoid that single plaintiff states play the role of parentes patriae by representing what they deem to be the interest of their citizens. As was largely expectable, Microsoft contested the active legitimation of plaintiff states in this respect. What was less expectable is that the Federal Government endorsed Microsofts position by filing a document in which it evoked its exclusive legitimation to act as representative of the public interest. See http://www.usdoj.gov/atr/cases/f10900/10980.pdf. For what concerns Microsofts rivals, it is hardly the case to recall that AOL/Netscape joined the array of plaintiffs in January 2002

28

Microsoft Saga and the Sorrows of Old Antitrust

A careful reading of the alternative remedial proposal filed by the nine dissenting States on December 2002 suggests that it would have been quite hard for the defendant and such aggressive plaintiffs to reach a mutually satisfactory agreement. The obligations and sanctions imposed by this remedial proposal would pillory Microsoft for as long as ten years (renewable).29 First, the Redmond-based firm would have been forced to develop a version of Windows without browsing functionality i.e., without Internet Explorer and distribute it at a discounted price. According to the dissenting States, such a remedy was meant to avoid the code commingling that represented the fundamental matter in Microsoft III and still is one of the most debated issues within the tying claim. However, such a prescription was not based on any acceptable theoretical justification. As a matter of fact, it is hard to imagine why Microsoft should be forced to deprive Windows of its browsing functionality, in a moment in which the need to reduce complexity and the lessons drawn from the economics of attention call for the creation of multi-functional environments in which non-professional users easily find solutions to their operative needs.30 In other words, such a mandatory reshaping of Windows would be contrary to sound economic thinking, and would force Microsoft to abandon a technological tying that certainly benefits final users; not to mention the costs that Microsoft would face in order to modify the code lines of a software designed as a single product, hence characterized by the integration of OS and browser code lines.31 Furthermore, from a legal standpoint, such a remedy would
by filing its own complaint for an injunctive relief and treble (punitive) damages aimed at restoring the competitive equilibrium and avoiding Microsofts future abusive behaviour. See http://eon.law.harvard.edu/openlaw/msdoj/aol-complaint.html. Microsoft and AOL recently reached an agreement to settle their litigation: according to the agreement, Microsoft will pay AOL $750 million. See infra, note 69.
29

See Section 21.b of the final judgment proposed by the dissenting states, supra note 11. It is interesting to read also the following Section (21.c), which mandates that any term of the agreement must be interpreted extensively. According to Bill Gates, if the Judge approved the alternative remedial proposal, Microsofts market value would have decreased by roughly $300 million. See Gates testimony before the District Judge, 22 April 2002, available at http://www.microsoft.com/presspass/trial/mswitness/2002/billgates/billgates.asp, 185206 (last visited: January 3, 2004). The so called information overload the need to process a remarkable amount of complex information leads non professional users to demand more and more integration in their use of computer products. PC producers are no doubt more equipped than final users in assembling and integrating OS and browser products, for reasons deeply related to the specialization of labour and to the existence of economies of scale in learning and knowhow. In his testimony before the District Court of 22 April 2002, Microsofts co-founder Bill Gates stated that it was economically impracticable for Microsoft to comply with the obligation to separate the source codes of Windows and Explorer as provided by the

30

31

10

Andrea Renda

certainly be at odds with the wording of the Court of Appeals in Microsoft III, which glamorously rejected the approach previously endorsed by District Judge Penfield Jackson, hinged on the application of the (modified) per se rule devised in Jefferson Parish. The Court of Appeals clearly advocated the adoption of a rule of reason approach, whose outcome depends on the assessment of costs and benefits produced by the tying arrangement for final users.32 And this is just the beginning. Scrolling down the proposal invoked by the nine dissenting States, many other strangling and legally flawed remedies surface. Think about the obligation, for Microsoft, to freely distribute and promote the Java programming language for as long as ten years, by developing a version of the Java Virtual Machine that ensured compatibility with the technical specifications provided by Sun Microsystems, in order to allow the diffusion and success that this crossplatform programming language so far lacked according to the States because of Microsofts fierce obstructionism.33 Or, consider the obligation to ensure the portability of Microsofts application software on alternative platforms in primis Apples even running an auction in order to license the right to develop and market a version of MS Office perfectly compatible with the MacOS and other operating systems.34 As a result, Microsoft would be forced to open its arms to rivals note, even to Apple, whose OS was excluded from the relevant market from the very beginning of the litigation playing the role of a modern Saint Francis and giving its
dissenting states remedial proposal, and that such a code separation would make it impossible for Microsoft to comply to another prescription of the remedial proposal (contained in Section 21), i.e. the obligation to adhere to commonly adopted standards and non-proprietary standards. See Gates testimony, supra note 29, at 185-206.
32

According to Judge Jackson, the essential characteristic of tying arrangements is that market power is used to force the buyer into a purchase of a tied product that the buyer either did not want at all, or might have preferred to purchase elsewhere on different terms. See Jefferson Parish Hospital District No. 2 Vs. Hyde, 466 U.S. 2, 12 (1984). The reply by the Court of Appeals was crystal clear: not all ties are bad ... if integration has efficiency benefits, these may be ignored by the Jefferson Parish proxies ... simple integration of the rules separate product test may make consumers worse-off. See U.S. vs. Microsoft, 253 F. 3d at 73 and 77. For an economic analysis of the evolution of the US and EU approach towards tying, see Christian Ahlborn, David Evans e A. Jorge Padilla, The Antitrust Economics of Tying: a Farewell to Per Se Illegality?, forthcoming on Antitrust Bulletin, Winter 2003, available at http://www.nera.com/wwt/publications/6301.pdf (last visited on December 29, 2003). See the dissenting States remedial proposal, supra note 11, at 17-18 (Section I): Microsoft continues to enjoy the benefits of its unlawful conduct, as Suns Java technology does not provide the competitive threat today that it posed prior to Microsofts campaign of anticompetitive conduct ... Microsoft must be required to distribute Java with its platform
software (i.e., its operating systems and Internet Explorer browser), thereby ensuring that Java receives the widespread distribution that it could have had absent Microsofts unlawful behavior.

33

34

Id., at Section 14.

Microsoft Saga and the Sorrows of Old Antitrust mantle to competitors in hardship.

11

If the dissenting States remedial proposal stopped here, this would already be enough to transform the initial goal (restoring the level playing field in the relevant market) into an unprecedented punishment in other words, the prescription would transform into a proscription. Yet, the proposal did not end here. Between the multitude of blatant remedies contained therein, the most important are also the most tricky and hidden. In fact, in forcing Microsoft to share all relevant code lines to enable interoperability of software applications with Windows and communication between client computers and non-Microsoft servers, the dissenting States remedial proposal did not include the expression for the sole purpose of interoperating that in the SRPFJ acted as functional ceiling to Microsofts duty to disclose its private information. Such a circumstance granted de facto open access to Windows APIs to all rivals hence, not just interoperability, but also and above all appropriability.35 Not surprisingly, the breadth of such a proposed remedy was noticed by the advocates of open source software represented in the public hearing by Eben Moglen of the Free Software Foundation which invoked the elimination of the sole purpose limitation from the SRPFJ, supporting this claim with the need to allow competitors the chance to successfully market alternative OS: amongst the latter, Moglen quoted with unaware impudence the so-called WINE (or Windows Emulator), which faithfully clones Windows graphical user interface and functionalities, replacing it as platform for third-party application software.36 The reason why such a prescription would benefit the population of final users remains rather obscure. Endorsing such solution would be as hazardous as asking record companies to invest in the free distribution of their records for sale on street stalls, ignoring the understandable effects that this would generate in terms of incentives to invest. One last thing: dulcis in fundo, scraping the surface of the remedial proposal, one could clearly read that Microsoft would be forced to publish the whole code of Internet Explorer, thence putting it in the public domain. Put differently, Microsoft
35

The remedial proposal is clear on this issue. Interoperability was not also related to middleware producers, but to any other Microsofts potential competitor. See the remedial proposal at 10-11 and the proposed judgment, supra note 11, at Section 4. See Eben Moglen, supra note 24: the sole purpose requirement means that Defendant does not have to make any such API information available to developers of software whose purpose is to make competing Intel-compatible PC operating systems. Moglen defines the wording of SRPF as an artful technical loophole inserted by Microsoft to escape the consequences of the SRPFJ. According to Moglen, [i]f Defendant were required to release information concerning its APIs to the developers of free software, GNU, Linux, the X windowing system, the WINE Windows Emulator, and other relevant free software could interoperate directly with all applications that have been developed for Windows.

36

12

Andrea Renda

would be forced to abjuration and to spontaneously embrace the open source philosophy.37 In conclusion, if the dissenting States meant to show the muscles of tough negotiators in drafting their alternative proposed final judgment, they certainly succeeded. Yet, such an attempt went so much beyond the line of reasonableness that the District Judge could never endorse such an imaginative option. Adopting this approach would strengthen the suspicion that the District Court intended, besides imposing an exemplary punishment on Microsoft, to favor the decline of proprietary software as a whole, confirming the vaticination of John Perry Barlow, which announced the death of copyright already a few years ago.38 The shortcomings would be, to say the least, tragic. Such claims raised many other concerns both from an economic and a legal standpoint. First, where the Court of Appeals had averted a structural solution (by rejecting the divestiture remedy devised by Judge Jackson, that would have led to the creation of two baby bills), the remedial proposal filed by the dissenting States reproposed the structuralist soul of antitrust, ignoring the developments of latest trial stages. Even if the word divestiture was never mentioned in the text, the remedies proposed directly affected the variable that mostly represents the structural element in the new sectors of digital capitalism: the so called architecture, or code.39 A solution that mandates Microsoft to develop a browserless version of Windows and to wholly disclose its browsers source code is far worse than a traditional divestiture remedy. Hence, there is no ground for endorsing a similar conceptual golpe, by exhuming anachronistic and socially harmful solutions, emerged also as the result of a malice that of Judge Jackson glamorously censured by the Court of Appeals as appearance of actual bias.40 An even more malicious interpretation would hinge on the insistent voces populi that relate the dissenting States behavior to the heavy lobbying efforts of major Microsofts rivals.41 True or false, the remedial proposal
37

See Section 12 of the remedial proposal (at 16-17), whose title is Internet Browser OpenSource License. I refer to John Perry Barlow, The Economy of Ideas: a Framework for Patents and Copyrights in the Digital Age( Everything You Know About Intellectual Property is Wrong), Wired, March 1994. See also Eben Moglen, Anarchism Triumphant: Free Software and the Death of Copyright, available at http://emoglen.law.columbia.edu/ publications/anarchism.html; and A. Minassian, The Death of Copyright: Enforceability of Shrinkwrap Licensing Agreements, 45 UCLA L. Rev. 569 (1997). Amongst many others, see Lawrence Lessig, Code and Other Laws of Cyberspace. Basic Books, New York, 1999. The Court only identified the appearance, not the existence of an actual bias. In particular, the press reports rumours regarding the State of California, where both Sun Microsystems and Oracle two of Microsofts most powerful rivals are based. See e.g.

38

39

40 41

Microsoft Saga and the Sorrows of Old Antitrust

13

was clearly incompatible with the most classical of antitrust tests: protecting competition, not competitors. Finally, that same solution would produce the effect of permeating the whole software sector with a diffused aura of uncertainty that could paralyze all those firms that compete for the market by relying on the possibility to recover their R&D investments through the licensing of their successful products.42 As is easily observed, the Microsoft case had transformed into a synecdoche, a part for the whole, a test-bed whose consequences could shake the whole functioning of the digital economy, producing an uncontrolled and unpredictable drift. 3. Faced with such a quagmire, the US antitrust was forced to react. Alas, the observable trend does not leave much room for optimism: whereas Microsoft III was defined by an authoritative commentator as an example of old rather than new antitrust, its subsequent developments although technically more appreciable have shown no evidence of a modernized conceptual apparatus.43 After all, it is fairly straightforward that what is sometimes termed protection of competition and the market faces some problems when applied to an economic sector in which the features of competition change and the boundaries of the relevant market become more and more evanescent. In my opinion, most economists are late in realizing that the peculiarities of digital capitalism are far from compatible with literal applications of the modus operandi of traditional antitrust. And in some cases the best solution might even be an abdication in favor of other legal remedies, more apt to governing the unremitting dynamics of knowledge-based industries by putting individual incentives back on the efficiency path.
States Splintered on Microsoft Settlement; Antitrust: California Leads Holdouts that want to Continue the Landmark Litigation, Los Angeles Times, November 7, 2001. Even Richard Posner, after the failure of his attempt to mediate between the interests of the opposed parties, observed that [States] are too subject to influence by interest groups that may represent a potential antitrust defendants competitors. This is a particular concern when the defendant is located in one state and one of its competitors in another, and the competitor, who is pressing his states attorney general to bring suit, is a major political force in that state. (R.A. Posner, Antitrust in the New Economy, 68 Antitrust L. J. 925, 940-941 (2001), also in Antitrust Law, 2nd Ed., Chicago 2001.
42

Consider the sentence with which the Court of Appeals reversed the approach adopted in Lotus v. Borland (49 F.3d 807 (1st Cir. 1995), by stating that the menu tree introduced by Lotus 1-2-3 was not copyrightable because it was essential, hence more an idea than an expression. See Herbert Hovenkamp, A Reckoning of Post-Chicago Antitrust, in Cucinotta, Pardolesi and Van Den Bergh (Eds), Post-Chicago Developments in Antitrust Law, Edward Elgar Publishing, February 1, 2003.

43

14

Andrea Renda

The most crystal clear signs of impasse in the application of antitrust rules are observed in market definition exercises although this is just the tip of an iceberg. In 1999, while the Federal Trade Commission was defining Intel as holding a monopoly position strangely enough in the market for Intel processors, eliciting at once hilarious and concerned comments, in Microsoft III Judge Penfield Jackson labelled Microsoft as both a quasi-monopolist and a fierce competitor.44 It was way too clear that Microsofts most powerful rival was not operating in the same relevant market as the Redmond-based softwarehouse: in digital capitalism and particularly in the field of software platforms firms compete for the market rather than in the market, and in the competitive game that preludes to the election of the de facto standard, economists use to say, the winner takes it most; hence, the market is identified at any given time with a single product, and competitive pressure inevitably comes from the future.45 As a result, there is a neat juxtaposition between competition and quasimonopoly whose virtue lies in the unceasing dynamics with which product generations overlap. And this same coexistence creates more than a few headaches to the traditional apparatus of antitrust. How can we assess the degree of market power of a firm enjoying a temporary monopoly position? Should we refrain from considering as relevant firms exerting competitive pressure on the monopolist, though belonging to another relevant product market? And most importantly, if we accept as true and in the US, such a conclusion seems now self-evident that the ultimate goal of antitrust enforcement is the protection of consumer welfare, what happens when the market, and hence the competitive mechanism loses the crown of supreme resource allocator, of reserved lane towards the efficiency frontier? How can we imagine to pursue an optimal market equilibrium in a sector in which the idea of market equilibrium is no more than a fleeting glimpse? Well, traditional antitrust policy is not well-equipped to catch such dynamics. As a result, those who believe in the application of old antitrust tools to new sectors advocate for necessarily short-sighted solutions. In Microsoft, for instance, Judge Penfield Jackson stubbornly insisted on traditional market definition exercises,

44

For an analysis of the Intel case (Intergraph Corp. V. Intel Corp., 3 F. Supp. 2d 1255, 1998), see Randal C. Picker, Regulating Network Industries: A Look At Intel, John M. Olin Law and Economics Working Papers Series n. 84, 1999. For what concerns Microsoft III, see R.B. McKenzie, Trust on Trial, Perseus Books, Cambridge, MA, 2000 (where the inherent contradiction in Judge Jacksons analysis of Microsofts market position is highlighted). See Roberto Pardolesi and Andrea Renda, How Safe is the Kings Throne? Network Externalities on Trial, in Cucinotta, Pardolesi and Van Den Bergh (Eds), Post-Chicago Developments in Antitrust Law, Edward Elgar Publishing, February 1, 2003. in Italian Language, see Roberto Pardolesi and Andrea Renda, Appunti di un Viaggio nel Capitalismo Digitale: Reti e Retaggi Culturali nel Diritto Antitrust, in Lipari and Musu (Eds.), La Concorrenza tra Economia e Diritto, Bari, 2000.

45

Microsoft Saga and the Sorrows of Old Antitrust

15

excluding from the relevant market Microsofts fiercest and most effective rivals, such as Apple, Sun and Netscape. 4. These and many others are the sorrows of old antitrust. But the story does not end here. In fact, for many reasons including its extraordinary longevity the Microsoft case gradually became a battle to set out the boundary markers on the Net. As I already mentioned, the decision issued by Judge Kollar-Kotelly may exert an outstanding impact on the future of innovation and on the outcome of the so-called IPR war. The latter is fought by two opposed and easily identified coalitions. On one side, those who not holding any valuable content see cyberspace as a land of conquer, in which the appropriation, exchange and costless distribution of copyrighted contents become successful strategies. I refer to these as the advocates of architecture, since they hinge on the need to fully exploit the open, end-to-end architecture that the pioneers of the Web mostly, Tim Berners Lee designed for cyberspace.46 On the other side, the large content producers, concerned by the ease with which information goods can be duplicated and distributed on the Net, which thwarts any attempt to control their diffusion. I refer to these as the advocates of control.47 In cyberspace, architecture defines whats possible, and the degree of control identifies the efficiency path.48 The growing tension between these two pillars may be explained by means of a few examples, and of course produced its effect also in the Microsoft case. Consider Napster, a champion in exploiting the potential of

46

For an explanation of the end-to-end (e2e) design principle, see Mark Lemley and Lawrence Lessig, The End od End-to-End: Preserving the Architecture of the Internet in the Broadband Era, UC Berkeley Law & Econ Research Paper No. 2000-19 (2000). The first to illustrate the e2e principle were Jerome Saltzer, David Reed and David Clarke, Endto-End Arguments in System Design, available at http://web.mit.edu/Saltzer/www/publications/endtoend/ endtoend.pdf. Control includes whichever extension of intellectual property rights on the Net. In particular, many commentators expressed concern that the extension of the patentable subject matter which is spreading like wildfire in cyberspace, and today covers also hardly creative business methods might jeopardize future incentives to invest in R&D. On this issue, it is important to notice that IPRs should be seen more as bargaining chips than as per se property rights carrying a rigid jus excludendi: a careful and clever management of IP portfolios has become a key driver in digital capitalism. For an amusing and brilliant illustration, see K.G. Rivette and D. Kline, Rembrandts in the Attic: Unlocking the Hidden Value of Patents, Harvard Business School Press, 2000. The idea that in cyberspace architecture (i.e., code) defines whats possible is due to Larry Lessig, Code, supra note 39.

47

48

16

Andrea Renda

architecture, which led content producers to react and invoke stronger control.49 Or think about the community of open source programmers, which fully exploits the Nets potential, but seems to require proprietary software as lymph needed to gain momentum.50 These are all phenomena which seem to substantially limit the potential for controlling the diffusion of information on the Net. But countervailing examples of exaggerated attempts to enhance control also exist. A good example is the current evolution of Internet broadband services, where dominant operators AT&T and Time Warner (better, the recently declined giant conglomerate Time Warner/ Netscape/AOL) recently introduced operating systems and license contracts whose ultimate effect is to jeopardize end-to-end architecture and transform the Net into a sort of broad- or narrowcasting network similar to cable TV, hence destroying the potential for user-to-user communication that always characterized the Internet.51 But there are many other examples. Consider recently adopted legal initiatives such as the 1998 Sonny Bono Copyright Term Extension Act, which extended the duration of copyright by yet another twenty years after the authors death and for this reason ironically labelled as the Disney Act, since it was passed just as the copyright on Mickey Mouse was close to expire now subject to a heated trial suit upon initiative by publisher Eric Eldred under the supervision of Larry Lessig and Jonathan Zittrain.52 Finally, consider the growing success of firms specialized in supporting
49

On the Napster litigation, see the December 2000 issue of the American University Law Review, which hosted a symposium titled Beyond Napster. Debating the Future of Copyright on the Internet. See also R. Shih Ray Ku, The Creative Destruction of Copyright: Napster and the New Economics of Digital Technology, 69 U. Chi. L. Rev. 263 (2002). A lot has been written on the potential and limits of open source software. I will just mention what I consider to be the most innovative and seminal contributions: David McGowan, Legal Implications of Open Source Software, U. Ill. L. Rev. 241, 2001; Robert Gomulkiewicz, How Copyleft Uses License Rights To Succeed in the Open Source Revolution and the Implications for Article 2B, 36 Houston L. Rev. 179, 1999; Josh Lerner and Jean Tirole, The Simple Economics of Open Source, available at http://www.hbs.edu/research/facpubs/workingpapers/ papers2/9900/ 00-059.pdf. As far as the viral nature of open source software is concerned, the topic is of course fiercely debated. The GNU project, born in 1984 and ancestor of all the GNU/Linux operating systems, even drew its name from a family of operating systems (developed by AT&T) that later became proprietary. GNU is a recursive acronym: Gnus Not Unix. See Mark Lemley and Lawrence Lessig, supra note 46, and Lawrence Lessig, The Future of Ideas: the Fate of the Commons in a Connected World, Random House, New York, 2001. According to Metcalfes Law, the value of an e2e network grows exponentially with the number of its users. To the contrary, when a network does not have an e2e architecture, such a relation is not satisfied. See Roberto Pardolesi and Andrea Renda, supra note 45. On February 2002, the Supreme Court granted the writ of certiorari in Eldred v. Reno, in which the constitutionality of the Sonny Bono Copyright Term Extension Act is challenged. Eldred, a publisher, and Harvards most libertarian fringe (the Berkman Center for Internet & Society) launched a campaign called Free the Mouse.

50

51

52

Microsoft Saga and the Sorrows of Old Antitrust

17

multimedia content producers by using all sort of methods to deter final users from sharing copyrighted files through peer-to-peer exchange networks. Firms such as Media Defender or Overpeer even follow the Maoist saying punish one, educate a hundred in the attempt to stop what MPAA Chairman Jack Valenti a powerful advocate of control considers to be just theft.53 Faced with all these developments, the Microsoft case became a battlefield of utmost importance. And the signs of this new relevance are already evident. Think about the reactions elicited by the uniform software licenses formulated by the Redmond-based company for communication between client PCs and servers in (anticipated) compliance with the SRPFJ. In particular, consider the license contract for the Common Internet File System Protocol (CIFS), in which Microsoft expressly requires licensees not to sublicense code information under the General Public License (GPL) or Library GPL IPR schemes used by the open source community.54 Needless to say, such contract term excited some anger at the Free Software Foundation, where open source software developers were eager to take a free ride on the information needed to emulate and replace Microsoft products on the server market.55 From this viewpoint, it is far from surprising that Microsoft seeks to preserve some degree of control on its IPRs. After all, the SRPFJ is compatible with this conduct.56 However, in my opinion, relying on an antitrust case to set the efficient degree of intellectual property protection in the trade-off between architecture and
53

According to the New York Times, Overpeer is one of a dozen companies developing software that would sabotage the computers and Internet connections of people who download pirated music. These companies develop programs for record companies, such as silence (which searches for and deletes music files from hard drives); freeze (which locks up a computer for minutes or hours); Trojan Horses and viruses. See Andrew Ross Sorkin, Software Bullet is Sought to Kill Music Piracy, The New York Times, May 4, 2003. GPL and LGPL are licenses that allow users free access to the source code and the possibility to modify it at their will, provided that any modification is made available to the community of programmers, i.e. put in the public domain and subject to free licensing. See http://www.gnu.org/copyleft/gpl.html.

54

55

The terms of the license agreement are available in the Windows Development section of the MSDN-Microsoft Development Network (www.msdn.microsoft.com) and is named Royalty Free CIFS Technical Reference License Agreement. Microsoft, after defining the GPL and the LGPL as IPR Impairing Licenses, states that the licensee shall not distribute its implementation of the CIFS destined to non-Microsoft platforms under an IPR impairing licensing agreement. The complaint moved by the open source movement in this respect is found at http://www.gnu.org/press/2002-04-11-ms.patent.html, where the communitys spokesmen claim that Microsoft has veiled its attack in the wrappings of a gift. In other words, Microsoft seems overly generous but, scraping the surface, it is just trying to illegally preserve its monopoly to say it with Virgil, timeo [Microsoft] et dona ferentes. 56 The SRPFJ, supra note 10, at Section III.l.3, explicitly allows Microsoft to require its licensees not to sublicense its copyrighted products.

18

Andrea Renda

control would be foolish. It would be foolish even if antitrust enforcement were not characterized by those ongoing problems that challenge its ability to remain updated with the markets new developments and it certainly is. It suffices to observe that while Microsofts behavior was subject to fierce scrutiny at the antitrust authorities for more than a decade, in more hidden gorges of cyberspace, AOL, Netscape and Time Warner were creating a giant conglomerate, whose power could lead to the elimination of the whole Net potential; and RealNetworks was signing an agreement with Intel according to which multimedia software RealPlayer will be shipped together with Intels motherboards in order to encourage OEMs to preinstall both products on their PCs. I just mean to send a caveat. An invitation not to empower the Microsoft case with a role that only a serious information policy could effectively and satisfactorily play. A thorough regulation of cyberspace is needed, based on two fundamental goals: (a) maximizing the value of the Net, by preserving its architecture; and (b) maximizing the value of content exchanged on the Net, by ensuring that content owners can control the diffusion and reproduction of their products. 5. The outcome of the US Microsoft case is certainly going to exert a significant impact on the solution in the EU proceeding. As widely known, a few years ago the European Commission started an investigation after receiving a complaint from Sun Microsystems in order to ascertain whether Microsoft unlawfully attempted to monopolize the market for server OS and the market for Web services by leveraging its dominant position in OS for Intel-compatible PCs and PC application software (so called Personal Productivity Applications or PPAs, such as the MS Office Suite).57 The two proceedings are similar in the object and in the likely effects. On both sides of the Atlantic, antitrust authorities are regulating the limits of mandatory disclosure of information, in particular for what concerns APIs and client-server communication protocols.58 From this standpoint, the EU case seems to be moving even more clearly
57

In Europe, Microsoft must also face a bundling claim relative to its audiovisual software, Windows Media Player. EC Commissioner Mario Monti integrated the two proceedings, in order to find out whether Microsoft reiteratedly abused its market power by attempting to monopolize more than one market. Funnily enough, Monti seems to have realized himself the benefits of integration ante litteram by merging the two proceedings. According to Sun Microsystems, Microsoft does not completely disclose the information needed to achieve a sufficient degree of interoperability between Windows and the Solaris servers implemented by Sun. See the Commissions Statement of Objections Against Microsoft Corporation, IV-C-e/37.545, 1 August 2000, at 1.5.1.2.

58

Microsoft Saga and the Sorrows of Old Antitrust

19

towards the application of an essential facility approach59, by imposing Microsoft mandatory licensing of APIs and communication protocols to downstream operators, with no obligation to share its code with rivals in the relevant market.60 In some other respects, the outcome of the EU proceeding might move exactly in the same direction as the SRPFJ; most practitioners exclude structural remedies if not for the likely mandatory unbundling of Windows and its Media Player and expect the Commission to charge Microsoft with mandatory free licensing of communication protocols, in order to restore the level playing field in the server market. On this issue, Microsoft already announced that it will not agree with the Commission on any further concession than those already prefigured in the SRPFJ.61 A detailed illustration of the law and economics of the EU Microsoft case would fall outside the scope of this article. One question, anyway, would be worth asking: what would happen if Microsofts behaviour is judged differently on the two sides of the Atlantic? As the remedies considered typically include mandatory information sharing, if one of the two proceedings end with the imposition of mandatory disclosure, the same result would automatically emerge on the other side of the ocean. Put differently, it is quite evident that, after Judge Kollar-Kotellys decision, the European Commission is called to adopt a similarly cautious approach, aimed at avoiding that the goal of ensuring reasonable interoperability between server and client computers transforms into a threat to all those firms that invest in R&D relying on proprietary architectures.62
59

The Commission explicitly refers to the decision of the ECJ in Magill at 298 of the Statement of Objections against Microsoft, supra note 57. The Commissions approach to the essential facility doctrine differs substantially from the one expressed by the ECJ. A famous and recent application of the Magill rule is found in NDC Health/IMS Health (COMP/38.044, 3 July 2001), a case in which the Commission showed a marked tendency towards compulsory licensing of de facto industry standards. As far as open source software is concerned, the 2000 Statement of Objections against Microsoft does not even tackle the issue of source code disclosure. The issue became even more complex given the impasse in which the EC currently finds itself, after its decision to block the merger between General Electric and Honeywell and the recent contrasts with the CFI in Airtours, Schneider/Legrand and Tetra/Laval. Many advocates of stricter remedies hinged on Montis presumed tendency to be in the limelight: for example, Thomas Vinje, from the Brussels office of Morrison & Foerster, which represents a U.S.-based anti-Microsoft industry group, stressed that It would be an abdication of Mr. Monti's responsibilities to go down the same settlement path as the U.S. government. See, e.g., EU Continues Own Microsoft Probe, CBS News, November 4, 2002, available at http://www.cbsnews.com/stories/2002/11/04/ tech/main528026.shtml. In the Statement of Objections issued on August 2001, the European Commission informed Microsoft that it is close to consider the Redmond-based company as having unlawfully monopolized the server market and that it considers the integration of Windows Media Player and Windows to be unlawful. If one also considers Montis tendency to deviate from the approach normally adopted on the other side of the Atlantic see, inter alia, the

60

61

62

20

Andrea Renda

The mutual dependency of the two proceedings may be observed also in practice. For instance, Microsoft recently announced its availability to provide Sun with relevant code information about the Kerberos security system and about the CIFS. The former is an open standard developed by the Massachusetts Institute of Technology for improving network security, later implemented by Microsoft with a number of proprietary extensions, generally named Security Service Provider Interface (SSPI).63 The latter as I already mentioned is a protocol that allows client PCs to locate and load files saved on other computers in a local network (LAN) by communicating with a central server. CIFS is the open version of a protocol developed and used by Microsoft and named Server Message Block.64 For both products, Microsoft expressed its availability to grant concessions to its rivals, moving towards more sharing of information for interoperability purposes. The stark similarity between the two cases the US and the EU investigations on Microsoft highlights once again the need for more cooperation and coordination between antitrust authorities on the two sides of the Atlantic. In the case at hand, it would be highly recommended that both cases do not move beyond imposing mandatory disclosure for interoperability purposes. Beyond that threshold, any decision would threaten to produce undesirable and indeed uncontrollable distortions in the competitive dynamics of the industry.65 6. In conclusion, this paper aimed at illustrating the problems that remain unsolved in
decision in GE/Honeywell and his reputation as a regulatory pitbull, then it becomes clear why Microsoft considers the Commission as a tough rival, hardly available to negotiations of any kind. See, e.g., Microsoft: Will the EU Crack Down? Gates & Co. Face a Regulatory Pitbull, Business Week, 5 November 2001.
63 64

See the EC Statement of Objections against Microsoft, supra note 57, at 71. The CIFS is the standard proposed by the Open Group to the IETF, Internet Engineering Task Force. See the Statement of Objections, supra note 57, at 57. The Open Group is a consortium that aims at improving interoperability between computer systems. Microsoft is a member of the Group as well as Sun Microsystems, IBM, Motorola and many others. See www.opengroup.org. The are many other factors that might exert an impact on Microsofts position in Europe in the near future. The balance painfully struck by the SRPFJ, which allows Microsoft to preserve at least, partially the value of its intellectual property portfolio, might turn out being more disadvantageous for Microsoft in the EU, but a lot will depend on the wording of the oncoming EC Directive on the patentability of computer-implemented inventions, which aims, i.a., at harmonizing the EU treatment of software patentability with the US approach and the wording of article 17 of the TRIPS agreement. See the EC proposed Directive (COM(2002) 92) at http://europa.eu.int/comm/internal_market/en/indprop/comp/com02-92en.pdf (last visited on December 26, 2003).

65

Microsoft Saga and the Sorrows of Old Antitrust

21

the Microsoft case and in the legal and economic approach to it. No doubt, the SRPFJ moves in a satisfactory direction, which may produce the degree of stability that the whole sector has been invoking for ages. But the major virtue of the agreement reached between Microsoft and the DoJ lies in avoiding structural remedies: following the path proposed by the non-settling States would lead to disavowing the role of intellectual property, qualifying the de facto standard not just as an essential facility, but as a public good, available for use and free ride on the market. Instead of a game in which the winner takes most, such a solution would create a paradoxical competitive arena in which losers take it all. I am not suggesting that a future in which open standards dominate the scene is not attainable at all in cyberspace. But there is no doubt that a similar trajectory should not be drawn by courts and judges, especially when the industry itself seems well-equipped to move towards a new balance between closed and open (but still mostly proprietary) architectures. Microsoft itself seems to be adopting a different commercial strategy, more apt to exploiting the huge potential of cyberspace through the marketing of interoperable platforms.66 For what concerns the Microsoft litigation, it was certainly important to ensure that Microsoft could not exploit its superior bargaining power vis--vis upstream and downstream operators and hence to inhibit the natural overlapping of product generations in such a dynamically competitive sector. The SRPFJ no doubt reaches this goal and seems perfectly crafted to effectively restore the level playing field in the OS sector. Lawrence Lessig ipse, one of Microsofts toughest defamers and brilliant guru of cyberspace regulation stated that the remedial proposal filed by the nine non-settling States appeared frankly excessive if compared with the wording of the decision adopted by the Court of Appeals in June 2001. According to Lessig, the SRPFJ certainly moves in the right direction, and its limits are to be found only in the setting of compliance control mechanisms that, according to Judge Penfield Jacksons former special master, might prove too weak.67 But the true conclusion is yet another: an invitation to update the conceptual apparatus of antitrust enforcement, freeing it from way too obsolete schemes. In order

66

The new Microsoft platform for Web Services, .Net, uses an open language, the XML, developed by the World Wide Web Consortium as the successor of the glorious HTML. This may ensure full interoperability between Microsofts platform and major server OS (such as Windows 2000, Solaris, UNIX) and with OS conceived for handhelds, cell phones etc. See Lessigs testimony at http://lessig.org/content/testimony/msft/msft.pdf (last visited: January 3, 2004). The approach adopted by Lawrence Lessig might have exerted an influence on the District Judges Final Order, which amended the SRPFJ only in the Section related to compliance procedures (i.e., Section VII).

67

22

Andrea Renda

to successfully carry antitrust to the shore of high-tech market, one might evoke the lighter vessel needs in Dantes Inferno.68 And above all, in order to avoid futile attempts to catch the market optimum, which here is certainly fleeting, it is necessary to integrate antitrust and intellectual property protection in a wider context of information policy, by regulating the whole sector in terms of architecture and control, in order to maximize both the value of the Net and the value of the content circulating therein. Those should be, in my opinion, the new milestones of antitrust in the new economy69. Thence, there is one, none and a hundred thousand Microsoft cases. And yet one certainty. Any attempt to solve such transcendent issues in a courtroom would lead to an overly dangerous path. The suggestion is to avoid stubbornly clinging to the solution of an issue that though extraordinarily important is far from being confined to the individual interest of the opposed parties. This would mean behaving like those soldiers that continue to remain secured in their blockhouses without knowing that the war is over. And in the case at hand the war is certainly over, and paved the way to another, far more crucial conflict. It is undoubtedly necessary to find a definitive solution for the Microsoft case, but only if policymakers contextually manage to set the rules that will govern information flows in the future of the Net. On this side, no doubt, a lot still needs to be done.

68 69

Divine Comedy, Inferno, Canto III. While I was completing this article, Microsoft and AOL reached an agreement to settle the pending lawsuit initiated by AOL on January 2002. According to the terms of the agreement, Microsoft will pay AOL 750 million USD, and commits to license its browsing technology to AOL for seven years. In exchange, AOL will grant Microsoft access to its powerful distribution channels: this will facilitate Microsofts entry in a sector that has become more than emerging decisive in cyberspace. The settlement reached by the two giant operators will likely sentence Netscape Navigator to death. Needless to say, the constant overlap of coups de theatre makes it almost impossible to provide a real time snapshot of what is happening in high-tech markets.