Microsoft's Five Fatal Flaws MICROSOFTS FIVE FATAL FLAWS Alan Devlint Michael Jacobst I. Introduction .............................................................. 67 II. The Nature of Innovation, The Right to Exclude, and M icrosoft ........................................................... 76 A. Traditional and New Economy Industries ...... 76 B. The Economics of Innovation ............................. 79 C. Pre-Microsoft Jurisprudence ............................... 84 1. M agill ........................................................... 84 2. IM S H ealth ..................................................... 86 D. The Microsoft Decision ....................................... 88 E. Europe's Ordoliberal Tradition ......................... 92 III. Microsoft's Five Fatal Flaws ..................................... 95 A. Rewriting Relevant Precedent and Defining "Effective" Competition ..................................... 95 B. Relying on Consumer Survey Data .................... 99 C. Embracing the Discredited S-C-P Model ............ 101 D. Adm inistrative Costs ........................................... 104 E. The Final Flaw: Rejecting Ex Ante Incentives as an "Objective Justification" .................. 105 IV . C onclusion .................................................................. 106 I. INTRODUCTION The recent decision of the European Court of First Instance ("CFI") in an antitrust case brought against Microsoft has exposed a deep transatlantic divide about the appropriate level of scrutiny brought to bear on the unilateral conduct of a dominant firm.1 In its decision, the t Associate, Latham & Watkins LLP; BBL (Int'l), University College Dublin, Ireland; J.D., Stanford University; L.L.M, J.S.D., University of Chicago. CFI required Microsoft to share valuable intellectual property ("IP") with its smaller rivals, under circumstances that would create no such obligation in the United States.2 In addition, the court effectively imposed per se liability for bundling conduct-an approach explicitly rejected by the Court of Appeals for the District of Columbia in an earlier, unrelated proceeding against Microsoft.' A superficial comparison of the U.S. and European Commission ("EC" or "Commission") actions would reveal an asymmetry concerning their respective views of the presumed anticompetitive effect of bundling and refusals to deal. A deeper analysis, however, reveals that the chasm is both more extensive and fundamental than might initially appear. Although U.S. antitrust law has long approached claims of single firm misconduct with far greater skepticism than Europe 4-a departure explained in significant part by t Professor, DePaul University College of Law; B.A., Dartmouth College; J.D., Yale University. 1 See, e.g., Thomas 0. Barnett, Assistant Attorney General for Antitrust, Statement on European Microsoft Decision, Sept. 17, 2007, available at http://www.usdoj.gov/atr/public/pressreleases/2007/226070. pdf ("We are.., concerned that the standard applied to unilateral conduct by the CFI, rather than helping consumers, may have the unfortunate consequence of harming consumers by chilling innovation and discouraging competition."). 2 See WILLIAM H. PAGE & JOHN E. LOPATKA, THE MICROSOFT CASE: ANTITRUST, HIGH TECHNOLOGY, AND CONSUMER WELFARE 80 (Univ. of Chicago Press 2007). Compare Case T-201/04, Microsoft v. Comm'n 5 C.M.L.R. 11 (2007) [hereinafter "CFI Decision"], available at http://www.microsoft.com/presspass/presskits/eucase/docs/T-201-04EN.pdf with United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001). 3 Microsoft, 253 F.3d at 84. See, e.g., R. Hewitt Pate, Assistant Attorney Gen. in Antitrust Div., Dep't of Justice, Antitrust in a Transatlantic Context-from the Cicada's Perspective, Address to "Antitrust in a Transatlantic Context" Conference in Brussels (June 7, 2004), available at http://www.usdoj.gov/atr/public speeches/203973.pdf; Sara M. Biggers, Richard A. Mann & Barry S. Roberts, Intellectual Property and Antitrust: A Comparison of Evolution in the European Union and United States, 22 HAMLINE J. PUB. L. & POLY 209 (1999). COL UMBIA B USINESS LA W RE VIE W [Vol. 2009 MICROSOFT'S FIVE FATAL FLAWS the European ordoliberal tradition5 and greater incidence of dominance caused by privatization of formerly state-run monopolies 6-Microsoft has revealed a more significant disparity regarding an issue of the utmost economic importance. This issue involves the treatment afforded sellers of intellectual property, whose success leads them to economic monopoly as it casts their rivals to the fringe of the market. The question of whether the antitrust laws should require dominant owners of intellectual property to share it with their less successful rivals in order to facilitate a more "competitive" market structure is perhaps the most fundamental facing the field today. The CFI's opinion in Microsoft answered this question affirmatively,7 taking a position in marked contrast to the U.S. view.' An increasingly central tenet of competition policy concerns the treatment to be afforded copyright- and patent- holders, given the proliferation of new economy industries ' David Gerber, Constitutionalizing the Economy: German Neo- Liberalism, Competition Law, and the "New" Europe, 42 AM. J. COMP. L. 25 (1994); Valentine Korah, The Interface Between Intellectual Property and Antitrust: The European Experience, 69 ANTITRUST L.J. 801, 803 (2002). For a classic statement of the ordoliberal function of competition, see F.A. HAYEK, INDIVIDUALISM AND ECONOMIC ORDER (Univ. of Chicago Press 1948). 6 See Eleanor M. Fox, Monopolization, Abuse of Dominance, and the Indeterminacy of Economics: The U.S. 1E. U. Divide, 2006 UTAH L. REV. 725, 726-27 (2006). See CFI Decision, supra note 2. s See Miller Insituform, Inc. v. Insituform of N. Am., 830 F.2d 606, 609 (6th Cir. 1987) ("A patent owner who lawfully acquires a patent cannot be held liable under Section 2 of the Sherman Act for maintaining the monopoly power he lawfully acquired by refusing to license the patent to others."). Copyright holders enjoy similar rights. See Data General Corp. v. Gruman Systems Support Corp., 36 F.3d 1147, 1187 (1st Cir. 1994); Serv. and Training, Inc. v. Data Gen. Corp., 963 F.2d 680, 686 (4th Cir. 1992) ("Section 1 of the Sherman Act does not entitle a purchaser to buy a product that the seller does not wish to offer for sale."). In the United States, it remains true that there is "no reported case in which a court has imposed antitrust liability for a unilateral refusal to sell or license a patent or copyright." Image Technical Servs., Inc. v. Eastman Kodak Co., 125 F.3d 1195, 1216 (9th Cir. 1997). No. 1:67] founded on intellectual property and innovation.9 Because the public good characteristics of information render it vulnerable to easy appropriation, 10 the intellectual property laws bestow inventors and authors with exclusive rights akin to those long-enjoyed by possessors of physical property.1 Such exclusivity allows private markets to form for information goods and permits inventors and authors to extract much of the social value flowing from their creations. It is precisely the prospect of this monopoly return that serves ex ante to spur valuable innovation. This fundamental tenet of patent and copyright jurisprudence exists in apparent tension with the antitrust laws, which seek to enhance consumer welfare, in part by countering the misuse of monopoly. 2 The tension largely dissipates, however, when one realizes that intellectual property operates to deflect competition from its traditional ex post setting to an equally valuable ex ante form, where rivals compete to establish superior technologies and achieve future monopoly." This rivalrous process is a hallmark of new economy industries, where traditional price competition pales in importance to consumer demand for technological 9 See RICHARD A. POSNER, ANTITRUST LAw 245-56 (2d ed., Univ. of Chicago Press 2001); Korah, supra note 5. 10 See William M. Landes & Richard A. Posner, An Economic Analysis of Copyright Law, 18 J. LEGAL STUD. 325 (1989). 11 Patents grant an inventor the right to exclude others from making, using, offering to sell, selling, or importing the patented invention. 35 U.S.C. § 271 (2003). Copyright grants the owner the exclusive right to reproduce the work, prepare derivative works, distribute copies, perform the work publicly, and display the work publicly. 17 U.S.C. § 106 (2002). Trademarks impose liability for the unauthorized use of a registered mark in certain cases. 15 U.S.C. § 1114 (2005). 12 See, e.g., SCM Corp. v. Xerox Corp., 645 F.2d 1195, 1203-05 (2d Cir. 1981) (discussing the circumstances in which "the patent and antitrust laws necessarily clash"). 13 See POSNER, supra note 9, at 248; Atari Games Corp. v. Nintendo of Am., Inc., 897 F.2d 1572, 1576 (Fed. Cir. 1990); U.S. DEP'T OF JUSTICE & FED. TRADE COMM'N, ANTITRUST GUIDELINES FOR THE LICENSING OF INTELLECTUAL PROPERTY § 1.0 (1995), available at http://www.usdoj.gov atr/public/guidelines/ipguide.pdf. COL UMIA B USINESS LA W RE VIE W [Vol. 2009 primacy. 14 This marks the difference between static and dynamic competition, the respect for which currently constitutes the focal point of transatlantic divergence. 5 Microsoft marks an ominous shift in Europe's jurisprudential thinking, revealing in particular an unequivocal embrace of short-run consumer wealth at the possible expense of long-run innovation. 6 In a certain sense, the European preference is understandable. There is no question, for example, that intellectual property protection significantly reduces consumer welfare in the short term. Given the non-rivalrous nature of consuming information goods and the near-zero marginal cost of reproduction and dissemination in the digital era, once information products are invented and commercialized, they could conceivably be available for all to consume, free of charge. 7 But the patent and copyright laws foreclose this utopian possibility. There is no question, then, that relaxing the exclusivity afforded by these laws-given the demonstrable fact of innovation-will promote aggregate welfare by allowing all who value the invented information to consume it at a low price. The decision in Microsoft took just this path, by requiring the company to share sufficient information with its competitors to allow them to achieve interoperability with its operating system.' The fact that the information at issue was presumably protected by intellectual property was '4 See POSNER, supra note 9, at 248-49. " See Thomas 0. Barnett, Interoperability Between Antitrust and Intellectual Property, 14 GEO. MASON L. REV. 859 (2007) (opining that dynamic efficiency gains are more valuable than increments in static efficiency). 16 Europe's promotion of the short- over the long-run, which is in contrast to the U.S. view, has been recognized for some time, though Microsoft renders the distinction explicit. See, e.g., J. Bruce McDonald, Deputy Assistant Attorney Gen. in Antitrust Div., Dep't. of Justice, Section 2 and Article 82: Cowboys and Gentlemen, Presentation to the Modernisation of Article 82 Conference (June 16, 2005) available at http://www.usdoj.gov/atr/public/speeches/210873.htm. 17 See SUZANNE SCOTCHMER, INNOVATION AND INCENTIVES 31-32, 34-39 (MIT Press 2004). 1" CFI Decision, supra note 2. No. 1:67] MHCROSOFT'S FIVE FATAL FLA WS insufficient to justify Microsoft's refusal to deal.19 The CFI held that the information involved was "indispensable" to maintaining "effective" competition in the market for work- server operating systems, and that Microsoft's refusal to share amounted to an abuse of its dominant position.2" The obvious purpose of the court's judgment was to allow Microsoft's rivals to offer consumers alternative products fully interoperable with Microsoft's." Despite the unquestionable fact that the CFI ruling-if properly implemented 22 -would enhance consumer well- being in the short run, the wisdom of its decision is hardly self-evident. Notwithstanding the uncertain nature of innovation and the unknowable effect that a dilution in IP protection may have on particular research activity, a general reduction in IP scope, duration, or exclusivity will necessarily have a negative effect on innovation in some settings.3 Consequently, the precedent established in 19 Id. 690. 20 Id. 1 664, 694. 21 Id. 22 As explored below, the administrative task of implementing such a ruling would be immensely problematic, yet the CFI did not address it. See infra Part III.D. Not surprisingly, the Commission initially struggled to coerce Microsoft into making the relevant proprietary protocols available to its competitors at a "desirable" price following the CFI's judgment. On February 27, 2008, the European Commission fined Microsoft $1.3 billion-the largest fine in the history of antitrust enforcement-for charging rivals "unreasonable prices" for gaining access to interface information for work group servers that would facilitate interoperability. See Commission Press Release, Antitrust: Commission imposes 6899 million penalty on Microsoft for non-compliance with March 2004 Decision, (Feb. 27, 2008), available at http://europa.eulrapid/press ReleasesAction.do?reference=IP/08/318&format=HTML&aged=0&languag e=EN&guiLanguage=en. The offending prices were initial royalty rates of 3.87% of licensees' product revenues for patented information and 2.98% for confidential communication information. 23 As a general matter, high-risk, capital-intensive research endeavors are most vulnerable to any dilution in the right to exclude underlying an IP grant. As noted below, the pharmaceutical industry provides the best example of an industry whose innovative output depends critically on ex post exclusivity. See, e.g., Richard A. Posner, Pharmaceutical Patents, BECKER-POSNER BLOG, (Dec. 12, 2004) (noting that "[tihe pharmaceutical- COL UMBIA B USINESS LA W RE VIE W [Vol. 2009 Microsoft will inevitably lead to some reduction of R&D in the future. This is especially worrisome as the duty to share proprietary information imposed by the CFI applies only to those inventions yielding the greatest value and return, which are of course the most socially desirable innovations. 4 Competitors will rarely, if ever, need or demand access to technologies for which ready substitutes exist. Only when IP-protected information is of such value that it has become a prerequisite of "effective" competition will access to it be "indispensable" under the law.2" Perversely, then, E.C. competition law seeks to dilute only the most valuable intellectual property protection, thus reducing the incentive to invent the most important technologies.26 Nevertheless, we cannot confidently assert that the CFI was mistaken. Although the duty to share arises only with respect to the most valuable inventions, the pecuniary rewards flowing to the inventors of such technologies are apt to be far larger than those accompanying more modest drug industry is the industry that can make the strongest case for needing patent protection," though expressing concern that even this industry may be over-incentivized by the contemporary patent system), available at http://www.becker-posner-blog.com/archives/2004/12/pharmaceuticall.ht ml. A notable counter-example is the fashion industry, where pervasive appropriation of design has not given rise to demonstrable reductions in creative output. See Kal Raustiala & Chris Sprigman, The Piracy Paradox: Innovation and Intellectual Property in Fashion Design, 92 VA. L. REV. 1687 (2006). The issue of whether rates of innovation will be reduced, however, is quite distinct from the question of whether a reduction in IP exclusivity is desirable. Whenever ex post return is reduced, the ex ante incentive to pursue that goal is similarly diluted. 24 This point was recognized by the CFI in Microsoft, see CFI Decision, supra note 2, 694, though the ECJ did not itself explicitly adopt it. Id. 710. 25 Id. 694. 26 One might challenge this assertion on the ground that one may not know ex ante that an invention will be especially important and hence valuable. However, this alone has little impact on incentives. The prospect of commercial success drives the innovative process-if the law provides that the most valuable innovative creations will be denied full IP protection, this fact will reduce ex ante incentives. By contrast, were the judiciary to deny holders of worthless inventions full exclusive rights, this would have little, if any, impact on innovation. No. 1:671 MECROSOFT'S FIVE FATAL FLAWS COLUMBIA BUSINESS LA W RE VIE W discoveries. If an inventor has already earned a sufficiently high return to have spurred the relevant activity ex ante, there is weaker normative ground for providing him with yet further monopoly returns. On this parsimonious approach to the appropriate rewards for invention, the CFI would be correct to facilitate interoperability if the dominant company has already earned, or will earn, a "sufficient" return in spite of the duty to share, assuming of course that a workable consensus can form around the notion of "sufficiency." Whether European law governing interoperability improves or worsens consumer welfare is an empirical matter that cannot yet, or perhaps never can, be resolved. In the absence of sufficient information, policymakers are forced to choose between either the short- or long-term. Those who emphasize maintaining ex ante incentives and ensuring high levels of innovation view the immediate consumer harm associated with strong intellectual property rights as a necessary cost of achieving a greater good over the longer term." Others, skeptical of the need for inventors to recoup a large proportion of their social contribution to spur desirable rates of research and development, focus on the short-run harm to consumers, who are denied access to a broader range of products at lower prices. Such commentators view intellectual property protection as potentially excessive, and the antitrust laws as an appropriate tool to temper such excess in appropriate circumstances." The CFI evidently elected the latter route, while the United States has chosen the former.80 27 See Dana Rohrabacher & Paul Crilly, The Case for a Strong Patent System, 8 HARv. J.L. & TECH. 265, 270-72 (1995). 2 See Michele Boldrin & David K. Levine, The Case Against Intellectual Monopoly, 45 INT'L ECON. REv. 327 (2004). 2" See, e.g., Harry First, Strong Spine, Weak Underbelly: The CFI Microsoft Decision, at 3 (2007), available at http://www.antitrustinstitute.o rg/archives/files/FirstStrong%20Spine-100220071356.pdf. so SCOTCHMER, supra note 17. [Vol. 2009 No. 1:67] MICROSOFT'S FIVE FATAL FLAWS 75 We have articulated our view on the prudence of these respective approaches elsewhere. 1 Here, we concede that a conclusive answer to the short/long-run question cannot yet be articulated. Instead, we pursue a more modest inquiry, considering the specific tenets of the CFI's judgment. In doing so, we uncover five flaws that fatally undermine the decision's legitimacy. Given the undoubted importance of the Microsoft decision to the development of competition doctrine in the United States, Europe, and elsewhere, these shortcomings are both noteworthy and unsettling. In exposing the inadequacies underlying the ruling, this Article argues that the precedential value of Microsoft should be narrowly construed for the purpose of future cases under Article 82 EC. Moreover, policymakers in the United States, and in emerging competition law regimes such as India and China, should look with consternation and caution at the implications of a decision that not only departs dramatically from the U.S. view, but does so on the basis of so suspect a foundation. Part II briefly explores the nature of innovation, the information asymmetry that frustrates efficient application of theory, Europe's ordoliberal tradition, and the precedent upon which the CFI relied in Microsoft. It concludes with a detailed examination of the facts and outcome of the Microsoft case itself. The heart of the Article lies in Part III, where five fatal flaws underlying the CFI's judgment are exposed. We argue that, irrespective of one's ideological conviction, the Microsoft judgment was a profoundly poor one. It rests on an artificial and strained reading of prior precedent, revives the thoroughly discredited Structure- Conduct-Performance ("S-C-F") paradigm of the 1950s, incongruously seeks to impose ordoliberal views on the new economy setting, fails to address the crucial challenge of actually implementing a regime of mandatory disclosure, and, perhaps worst of all, creates a nearly insurmountable " Alan Devlin, Michael Jacobs & Bruno Peixoto, Success, Dominance, and Interoperability, 84 IND. L.J. (forthcoming 2009). 32 See infra Part III. obstacle for a dominant company that seeks to justify its refusal to share on the basis of harm to innovation. II. THE NATURE OF INNOVATION, THE RIGHT TO EXCLUDE, AND MICROSOFT To appreciate the economic repercussions of the Microsoft judgment, one must understand the distinction between traditional and new economy industries, and recognize the respective roles of intellectual property and antitrust policy in spurring innovation in the latter setting. This Part briefly explores these topics and concludes with a discussion of Europe's ordoliberal tradition, which may have unfortunate repercussions if used to inform competition policy in the context of information goods. A. Traditional and New Economy Industries The traditional markets in which antitrust doctrine largely developed are characterized by multi-firm production, diseconomies of scale beyond a relatively modest level of output, significant marginal cost, and relatively stable market shares." Goods tend to be homogeneous, with few technological breakthroughs of the kind likely to revolutionize the nature of the products offered or the manner in which they are sold. 4 What technological progress does occur tends to lower the cost of production modestly. In this setting, monopoly is undesirable for two major reasons-one ex ante, the other ex post. From the ex post perspective, monopoly-whether achieved unilaterally or in concerted fashion in the form of a cartel-leads to pricing in excess of marginal cost, with the result that some consumers who value the product more than its cost of production are denied access to it." Economists refer to this misallocation as "deadweight loss," 33 See POSNER, supra note 9, at 245. 34 Id. 35 Id. at 12-13. [Vol. 2009COL UMBIA B USINESS LA W RE VIE W and it constitutes both the principal objection to monopoly and the raison d'Otre of the competition laws." The ex ante cost comes from wasted expenditures that might accompany the competition to obtain a monopoly position. 7 In theory, a rational company would spend up to the full amount of the anticipated monopoly surplus in order to acquire a monopoly position.3 8 Because that expenditure produces no social product, it is wasteful, and society would be better off without it.3 9 Cognizant of these potential costs, the antitrust laws attempt to facilitate a competitive market structure and thus to avoid a monopoly outcome. Monopoly itself is not forbidden, because to do so would be to outlaw the prize that drives the competitive process.4" Indeed, the prospect of success leads each firm to compete on the basis of price and quality in the hope of vanquishing its rivals and thereby obtaining the dominance it seeks. Of course, each competitor reasons the same way and, in a traditional market structure, rising long-run average cost typically means that no one entity is likely to achieve monopoly. Instead, inefficient and otherwise inferior competitors will exit, and the market itself will take on either an oligopolistic or near-competitive structure. Horizontal (and sometimes vertical) contracts are scrutinized under the competition laws to prevent rivals from bypassing the competitive process and artificially achieving monopoly outcomes.41 If monopoly is the evil against which antitrust law is directed in the traditional setting, it may be the sine qua non of the new economy." The latter context is defined by output 36 id. V7 Id. at 13-14. 38 Id. 39 Id. 40 See United States v. Aluminum Co. of Am., 148 F.2d 416, 430 (2d Cir. 1945) ("The successful competitor, having been urged to compete, must not be turned upon when he wins."). 41 15 U.S.C. § 1 (2004). 42 See Daniel L. Rubinfeld, Deputy Assistant Attorney Gen., Antitrust Div., U.S. Dep't of Justice, Competition, Innovation, and Antitrust No. 1:67] MI1CR OSOFT 'S FIVE FA TAL FLA WS COLUMBIA BUSINESS LAW REVIEW in the form of information (often in digital form), downward- sloping long-run average cost curves (caused by the high ratio of fixed-to-marginal cost), and powerful externalities in consumption (network effects)." These traits cause new economy markets to be uniquely predisposed to monopoly, a characteristic that has importance for antitrust policy." Network effect-driven markets that sell low-marginal cost products are subject to tipping effects and tend to have unstable equilibria outside the monopoly outcome.45 This phenomenon is often referred to as a "winner-take-all" structure.46 The information goods sold in these markets are invariably protected by intellectual property laws.47 Indeed, and as will shortly be explored, many information markets would not exist without patent or copyright protection that provides inventors and authors with exclusive rights to control the dissemination, marketing, and (to a degree) use of their creations. As a result of the foregoing, information markets are uniquely prone to monopoly control on the basis of patent or copyright. Despite an ostensible absence of competition, it is a mistake to think that monopolized new economy markets reflect a failed competitive process.48 Such a myopic view would fail to account for the fact that ex post dominance may mask the reality of fierce ex ante competition.49 Monopolists in information markets do not enjoy particularly secure positions, given both the overwhelming primacy consumers Enforcement in Dynamic Network Industries, Address to the Software Publishers Assocation (Mar. 24, 1998), at 14, available at http://www.usdo j.gov/atr/public/speeches/161 1.htm. POSNER, supra note 9, at 245-46. " See Michael L. Katz & Carl Shapiro, Network Externalities, Competition, and Compatibility, 75 AM. ECON. REV. 424 (1985). 41 See WILLIAM H. PAGE & JOHN E. LOPATKA, THE MICROSOFT CASE: ANTITRUST, HIGH TECHNOLOGY, AND CONSUMER WELFARE 92-96 (Univ. of Chicago Press 2007). See generally ROBERT H. FRANK & PHILIP J. COOK, THE WINNER- TAKE-ALL SOCIETY (The Free Press 1995). 47 POSNER, supra note 9, at 246. 4 Id. at 248-49. 49 Id. [Vol. 2009 place on quality over price in the technological sector and the low marginal cost of expanding production, which allows ready displacement of inferior incumbents." New economy settings thus bear witness to Schumpeterian waves of "creative destruction," as a series of ephemeral monopoly positions rise, only to fall in the face of superior, rival technology.51 Given the characteristics of information markets, the most valuable form of competition occurs (perhaps counter-intuitively) not over price, but quality. Innovation is therefore the driving force of the new economy, and a factor that must enjoy central importance in the construction of antitrust rules. Though the distinction between traditional and new economy settings is pronounced, a failure to appreciate it may frustrate, or even defeat, the formulation of prudent competition policy. In particular, applying traditional antitrust principles to attack ex post monopoly may be calamitous in information markets, by eviscerating the ex ante incentive to engage in the innovative process. This danger looms especially large with respect to intellectual property rights that afford their owner with a dominant position. Depriving a patent- or copyright-holder of the right to exclude will allow its rivals to compete on equal footing, but may result in a Pyrrhic victory for consumers. By devaluing the prize that drives inventive efforts, antitrust enforcers may unwittingly sacrifice long-term innovative activity for short-run gain. B. The Economics of Innovation We have emphasized that a critical distinction must be drawn between traditional and new economy settings, particularly when analyzing the relationship between the antitrust and intellectual property laws. For information markets, where dominance is not an unusual outcome, the key question is: to what extent, if any, should antitrust dilute exclusivity to maximize social welfare in the short- "o Id. at 249-50. 61 Id. No. 1:671 MIfCROSOFT'S FIVE FA TAL FLA WS COLUMBIA BUSINESS LA W RE VIEW term? Innovation lies at the heart of the new economy, and the right to exclude inherent in every patent and copyright bestows inventors and authors with commercial incentives to engage in this process. On the other hand, unqualified exclusivity imposes an undeniable social cost in the form of higher prices and deadweight loss. Crafting optimal rules therefore requires some appreciation for how and why advances in the sciences and arts take place. Unfortunately, the process of innovation is not well understood.52 In a general sense, though, innovation will occur when the prospect of ex post reward discounted both to present value and by the risk of failure exceeds the expected cost of development and commercialization. The need to incur these costs and risks, and to be compensated for them ex post, provides the economic rationale for the intellectual property regime.53 In practice, however, the incentives required to spur optimal levels of innovative activity seem heavily context- and industry-specific.54 In some circumstances, little or no patent or copyright protection is needed to induce such activity, such as in the instance of a student writing a doctoral dissertation for graduation requirements, an artist engaging in creative expression for the joy inherent in the activity, a company developing superior methods of doing business, or simple altruism.5 In other settings, with the pharmaceutical industry serving as a paradigmatic example, the level of expected ex post financial return is the sine qua non of inventive 52 See SCOTCHMER, supra note 17, at 118, passim. 5 See Landes & Posner, supra note 10, atpassim. See Mark A. Lemley, Intellectual Property Rights and Standard Setting Organizations, 90 CAL. L. REV. 1889, 1892 (2002) ("People innovate for many reasons, and in many industries the existence of IP rights doesn't appear to be chief among them. .. "). " See, e.g., Alfred C. Yen, Restoring the Natural Law: Copyright as Labor and Possession, 51 OHIO ST. L.J. 517, 537 (1990). [Vol. 2009 activity.5" In these circumstances, an absence or even significant dilution of copyright or patent protection would have profound and disproportionate effects on research and commercialization activity. Intellectual property laws, particularly patent doctrine, have largely adopted a "one- size-fits-all" approach 57-an evidently imperfect path that may nevertheless be justified on heuristic grounds. Nevertheless, as a consequence of patent and copyright laws' crude reflection of specific incentive requirements, a reduction in a patent- or copyright-holder's right to exclude may promote social welfare in one context and reduce it another. Antitrust enforcers could conceivably improve this incentive structure, by reducing the scope and exclusive force of intellectual property in some cases, and treating it as inviolable in others.5 One way to accomplish this would be to employ liability rules in the former context and apply property rules in the latter. 9 Unfortunately, policymakers and courts lack the empirical data required to perform this calculus reliably. Of necessity, they are left to form heuristic rules in an 56 See RICHARD EPSTEIN, OVERDOSE: How EXCESSIVE GOVERNMENT REGULATION STIFLES PHARMACEUTICAL INNOVATION (Yale Univ. Press 2006). " Andrew F. Christie & Fiona Rotstein, Duration of Patent Protection: Does One Size Fit All?, 3 J. INTELL. PROP. L. & PRAc. 402 (2008). ' This position need not be entirely far-fetched-indeed, in the copyright realm, the courts have (rather successfully) performed just such a case-by-case analysis in the context of fair use. See generally Pierre N. Leval, Toward a Fair Use Standard, 103 HARV. L. REV. 1105 (1990). " For a discussion of the difference between liability and property rules, see Richard Epstein, A Clear View of The Cathedral: The Dominance of Property Rules, 106 YALE L.J. 2091, 2091 (1997) ("A property right gives an individual the right to keep an entitlement unless and until he chooses to part with it voluntarily. Property rights are, in this sense, made absolute because the ownership of some asset confers sole and exclusive power on a given individual to determine whether to retain or part with an asset on whatever terms he sees fit. In contrast, a liability rule denies the holder of the asset the power to exclude others or, indeed, to keep the asset for himself. Rather, under the standard definition he is helpless to resist the efforts by some other individual to take that thing upon payment of its fair value, as objectively determined by some neutral party."). No. 1:67] MICR OSOFT'S FIVE FA TAL FLA WS environment of incomplete and asymmetric information. Unsurprisingly, the formulation of such rules closely tracks the intuitions of the people formulating them. Maximalists advocate obeisant treatment of intellectual property and support the near-Blackstonian position" that a property owner's ability to exclude others should be essentially inviolate. 61 They argue that strong property rights not only allow valuable information to be allocated efficiently through contract, but facilitate post-grant commercialization 62 and create optimal incentives by allowing patentees and copyright-holders to extract the full social value of their inventions. 63 Others reject a property- based view, favoring instead a liability approach that would grant third parties access to intellectual property-protected information for a fee. Such commentators typically adopt a parsimonious perspective, accepting that patent or copyright is required to overcome the public good nature of information, but advocating a regime that provides the minimum level of pecuniary compensation necessary to induce the production of the relevant innovation.64 Still others consider the allocative inefficiency and blunt incentive characteristics of intellectual property as good reasons for 60 WILLIAM BLACKSTONE, COMMENTARIES ON THE LAW OF ENGLAND Book II (describing property as "that sole and despotic dominion which one man claims and exercises over the external things of the world, in total exclusion of the right of any other individual in the universe"). 61 See, e.g., Richard Epstein, Intellectual Property: Old Boundaries and New Frontiers, 76 IND. L.J. 803, 805, 816-21 (2001) (supporting an analogy between intellectual and tangible property, but rejecting Blackstone's conception of the latter as excessive and inaccurate). 62 See F. Scott Kieff, Property Rights and Property Rules for Commercializing Inventions, 85 MINN. L. REV. 697, 719-27 (2001). ' See, e.g., Wendy J. Gordon, On Owning Information: Intellectual Property and the Restitutionary Impulse, 78 VA. L. REV. 149, 156-67 (1992). ' See, e.g., Stewart E. Sterk, Rhetoric and Reality in Copyright Law, 94 MICH. L. REV. 1197, 1205, 1209, 1213-15 (1996); Stephen Breyer, The Uneasy Case for Copyright: A Study of Copyright in Books, Photocopies and Computer Programs, 84 HARv. L. REV. 281, 322-23 (1970). [Vol. 2009COL UMIA B USINESS LA W RE VIE W replacing the patent system with a reward structure. 5 Finally, some doubt that intellectual property protection is even necessary to induce desirable innovation," chanting the mantra that "information wants to be free."67 The normative justification for this agnostic position emphasizes above all the altruistic nature of much innovation (particularly in the copyright realm).6 In fact, though, society lacks sufficient information to conclusively judge the merits of the preceding arguments. This is not to say that one should be indifferent to dilutions imposed by antitrust upon intellectual property or that prudent policy does not dictate a single approach. Indeed, we strongly believe that the optimal rule for now is the U.S. one, which we have explained in detail elsewhere.6 9 For present purposes, however, we concede that whether antitrust law should sometimes impose mandatory sharing obligations on the holders of intellectual property remains an empirically indeterminate inquiry. 65 See Kieff, supra note 62, at 705-07 (citing Steven Shavell & Tanguy van Ypersele, Rewards Versus Intellectual Property Rights (Nat'l Bureau of Econ. Research, Working Paper No. 6956, 1999) and Michael Kremer, Patent Buy-Outs: A Mechanism for Encouraging Innovation (Nat'l Bureau of Econ. Research, Working Paper No. 6304, 1997)). 6 See, e.g., MICHELE BOLDRIN & DAVID K. LEVINE, AGAINST INTELLECTUAL MONOPOLY Ch. I, 11 (2007) ("Since there is no evidence that intellectual monopoly achieves the desired purpose of increasing innovation and creation, it has no benefits. So there is no need for society to balance the benefits against the costs. This leads us to our final conclusion: intellectual property is an unnecessary evil."); see also Michele Boldrin & David K. Levine, Rent-Seeking and Innovation, 51 J. MONETARY ECON. 127 (2004). 6 Thomas 0. Barnett, Interoperability Between Antitrust and Intellectual Property, 14 GEO. MASON L. REV. 859, 865 (2007). This position is not universally improper-in many circumstances, intellectual property is not needed to spur innovation. See, e.g., Lemley, supra note 54. 6" Devlin et al., supra note 31. No. 1:67] MHCROSOFT'S FIVE FA TAL FLA WS COLUMBIA BUSINESS LAW REVIEW C. Pre-Microsoft Jurisprudence As noted, the ideology underlying the CFI's decision in Microsoft is not susceptible to conclusive disproof, given existing data. Notwithstanding this fact, however, numerous inadequacies undermine the decision's legitimacy, which-we submit-are sufficiently severe to justify limiting its precedential effect. This section reviews the prior case law upon which Microsoft relied. It bears noting that the following cases were, and are, considerably controversial, given their imposition of a duty to share on an intellectual property holder. 0 1. Magill Magill was the first European case to impose a duty to share intellectual property on a dominant undertaking.7 The impetus for the case came from an entrepreneur, who wanted to create a new product-a weekly program listing for the three television stations broadcasting in the United Kingdom and Ireland.12 At the time, the pertinent information was available only on the same day in newspapers.73 The three television companies had copyright protection for their TV listings, with the source of that protection coming from state law.74 Mr. Magill required a license to offer his prospective product, which was denied by each of the three companies. The Commission controversially held that each company had abused its respective dominant position, thereby violating Article 82 EC.75 Accordingly, the Commission required the companies to provide the pertinent information to Mr. Magill in order to facilitate the introduction of the 70 See RICHARD WHISH, COMPETITION LAw 663 (Oxford 5th ed. 2005). 71 Magill TV Guide/ITP, BBC & RTE, 1989 O.J. (L 78) 43. 72 Id. T 10. 73 Id. 74 Id. 75 Id. [Vol. 2009 No. 1:67] new product. The Commission's decision was upheld by both the CFI and the European Court of Justice ("ECJ").7 6 The ECJ held that the violation consisted of the refusal to release basic information by relying on copyright, thereby frustrating the emergence of a new product for which there was potential consumer demand.77 In addition, the court noted that there was no objective justification (i.e., a valid business reason) for the refusal, the effect of which was to reserve to the television companies the downstream market for television guides. 8 The ensuing test (the "Magill test") for establishing an abuse of a dominant position had four parts: (1) the refusal prevents the emergence of a new product; (2) for which there is significant consumer demand; (3) no objective justification exists; and (4) the result of the refusal is to reserve the downstream market for the dominant undertaking.79 Perhaps the most noteworthy aspect of the case was the ECJ's holding that intellectual property protection did not constitute an "objective justification" for excluding others from the information there at issue."' The enormity of this holding is magnified when one considers that the very raison d'6tre of intellectual property is the right to exclude.81 As explored above, the issue of whether to promote long- run innovation at the expense of short-run consumer wealth involves an empirically indeterminate inquiry, though it is surely true that the case for diluting the exclusive force of an intellectual property grant for the benefit of consumers may 76 Case T-69/89, RTE v. Comm'n, 4 C.M.L.R. 586 (1991); Case C- 241/91, RTE v. Comm'n, 1995 E.C.R. 743. " Case C-241/91, RTE v. Comm'n, 1995 E.C.R. 743, 54. 78 Id. 55. 79 Id. 54-56. 80 Id. 55. 81 Patents grant an inventor the right to exclude others from making, using, offering to sell, selling, or importing the patented invention. 35 U.S.C. § 271. Copyright grants the owner the exclusive right to reproduce the work, prepare derivative works, distribute copies, perform the work publicly, and display the work publicly. 17 U.S.C. § 106. Trademarks impose liability for the unauthorized use of a registered mark in certain cases 15 U.S.C. § 1114. MICROSOFT'S FIVE FATAL FLA WS COLUMBIA BUSINESS LA W REVIEW be stronger in some instances than others.82 Magill serves as an interesting example of a case where imposing a duty to share may have successfully enhanced short-run consumer welfare without having a larger negative effect on future innovation. Although the court did not say so explicitly, the relevant intellectual property was weak, as few European countries would have recognized a copyright in television listings information.83 More importantly, the third party recipient of the copyrighted information did not stand in horizontal competition with the copyright holders. Thus, the television companies were not forced to subsidize horizontal rivals. As a result, their refusal to license the copyrighted information harmed consumer welfare by foreclosing a market that would otherwise exist. At the same time, enforced sharing of that information may not have appreciably harmed ex ante incentives by reducing the dominant companies' ex post return. 2. IMS Health As noted, Magill involved circumstances where adopting a liability rule would unequivocally benefit consumers in the short-run without necessarily harming ex ante incentives. The precedential value of that decision is itself a source of concern, however, as became evident in the subsequent Commission decision in IMS Health.4 There, IMS-the world's largest leader in data collection on pharmaceutical sales and prescriptions-refused to grant a license to competitors that would have given them access to its copyrighted format for processing regional sales data.85 The Commission determined that access was necessary because the IMS brick structure had become the de facto industry standard. 6 The ECJ subsequently affirmed, and held that intellectual property does not provide an objective 82 See supra Part I. See WHISH, supra note 70, at 760; Korah, supra note 5, at 811. Case T-184/01, IMS Health, Inc. v. Comm'n, 4 C.M.L.R. 1 (2002). 8 This data was known as the "1860 brick structure." See id. at 47. 88 Id. [Vol. 2009 justification for a refusal to license and that an abuse of a dominant position follows if the refusal both prevents the emergence of a new product for which there is potential demand and is capable of eliminating all competition on the relevant market.87 Importantly, in IMS Health the ECJ expanded the scope of the already-controversial Magill decision. In particular, it chose to adopt a broad definition of the "new product" requirement set forth in Magill. The Court could have interpreted that requirement to mean that a dominant entity would be compelled to license its intellectual property only when a refusal would prevent the emergence of a new market. This view would have fit with the facts of Magill, where the third party sought copyrighted information that would enable it to offer a comprehensive TV listings guide-a product for which no substitutes then existed. The ECJ in the subsequent case of Bronner seemed to adopt this perspective, holding that for Magill to apply, the complainant must show that there is no realistic actual or potential substitute for the "new product" that would be foreclosed by the refusal to supply.8" In other words, the refusal to supply would have to eliminate all competition in the downstream market."9 In IMS Health, however, the ECJ made clear that the term "new product" includes one that will-or even might- be offered in competition with the intellectual property holder's existing product. In other words, no new market need be foreclosed by the refusal to supply. This distinction is fundamental and may have serious ramifications for future innovation incentives, given that a refusal to license to one's horizontal competitors constitutes the heart of a patent or copyright's commercial value. 87 id. ' Case C-7/97, Oscar Bronner GmbH v. Mediaprint Zeitungs-und Zeitschriftenverlag GmbH, 1998 E.C.R. 1-7791; Case COMP/C-3/37.792 Microsoft, Commission Decision of 24 Mar. 2004, 585 available at http://ec.europa.eu/competition/antitrust/cases/decisions/37792/en.pdf [hereinafter "Commission Decision"] (citing Bronner for this proposition). " Oscar Bronner, 1998 E.C.R. 1-7791, $ 41. No. 1:671 MCROSOFT'S FIVE FA TAL FLA WS D. The Microsoft Decision The European action against Microsoft resulted from the latter's refusal to grant Sun Microsystems the specifications necessary to render Sun's Solaris server operating system interoperable with Windows client PCs and servers.9" The Commission (the regulatory agency) found that Microsoft had a dominant position in the market for work group server operating systems, 91 basing its determination in part on network effects and in part on the finding that Microsoft's withholding interoperability information created an additional entry barrier.92 In analyzing the propriety of Microsoft's actions, the Commission began by noting that a simple refusal to license intellectual property rights can amount to an abuse of a dominant position.93 Observing that "interoperability with the client operating system is of significant competitive importance in the market for work group server operating systems"9-an observation that would seem to fall far short of the ECJ's determination in Bronner that the thing to be disclosed be "indispensable" for "any" competition 9 -the Commission concluded that Microsoft had abused its position.96 In making this determination, the Commission noted that "Microsoft's refusal puts Microsoft's competitors at a strong competitive disadvantage."97 Placing the focus on "disadvantage," though, departs significantly from previous case law's requirement that a refusal eliminate all competition in the downstream market.9 Far from recognizing this deficiency, the Commission compounded its " Commission Decision, supra note 88, 277. 9' Id. If 514, 541. 92 Id. IT 515-25. 93 Id. 547. 94 Id. 547 (emphasis added). 9' Case C-7/97, Oscar Bronner GmbH v. Mediaprint Zeitungs-und Zeitschriftenverlag GmbH, 1998 E.C.R. 1-7791, 41. 96 Commission Decision, supra note 88, % 692, 791. 9' Id. 589 (emphasis added). " Oscar Bronner, 1998 E.C.R. 1-7791, 41. COLUM1BIA B USINESS LA W RE VIE W [Vol. 2009 error. It did so by finding that open industry standards, the distribution of client-side software on the client PC, and reverse engineering were insufficient to alleviate the "disadvantage," as they fell "short of enabling competitors to achieve the same degree of interoperability with the Windows domain architecture as Windows work group server operating systems do."99 Thus, "it can be concluded that this degree of interoperability proves to be insufficient for them to viably compete in the market."100 Moreover, the Commission highlighted the importance of both an "effective competitive structure"1 ' and unfettered "customer choice," as critical factors in determining whether viable competition existed. 02 Microsoft's strongest argument may have been that its refusal to supply intellectual property-protected information was justified by the need to protect its incentives to innovate. 3 But the Commission's reaction to this argument was dismissive and unsettling. It concluded that "[t]he central function of intellectual property rights is to protect the moral rights in a right-holder's work and ensure a reward for the creative effort. But ... [a] refusal by an undertaking to grant a license may... be contrary to the general public good."10 4 Such deontological conceptions of intellectual property have no place in U.S. law° 5-- and for good reason. In Commission Decision, supra note 88, T 669 (emphasis added). Id. (emphasis added). See also id. 91% 687, 700. 101 Id. 704 (quoting Hoffman La Roche). 102 Id. 706. 103 Id. 709. 'O4 Id. 711. 10' See Peter Lee, The Evolution of Intellectual Infrastructure, 83 WASH. L. REV. 39, 53-54 (2008) ("The Supreme Court has recognized that the 'ultimate goal of the patent system is to bring new designs and technologies into the public domain through disclosure.'"). While acknowledging the valuable contributions of inventors, the Court has accordingly rejected any moral rights or Lockean labor theory justifications for granting patents. See Yochai Benkler, Siren Songs and Amish Children: Autonomy, Information, and Law, 76 N.Y.U. L. REV. 23, 59 (2001) (noting that "the basic ideological commitment of American No. 1:671 MI7CROSOFT'S FIVE FA TAL FLA WS particular, moral rights theories not only fail to yield reliable policy predictions, they fail to even provide a guiding, normative framework. °6 Although a teleological economic approach can be legitimately criticized for sometimes failing to yield black-letter policy conclusions," 7 its underlying conceptual framework is at least coherent and susceptible to unanimous agreement. Moreover, the moral rights tradition faces a serious reconciliatory problem with the unquestionable social costs that would follow a legal regime that departs from the utilitarian framework. It is in any event difficult to reconcile the utilitarian framework of competition law with a natural rights theory of patent and copyright. Yet, such reconciliation becomes necessary when the two fields coincide, as they do so explicitly when rivals seek access to another's intellectual property under the antitrust laws. The concepts of "viability" and achieving the "same degree of interoperability," each crucially important to the Commission's decision, have serious and unfortunate implications for competition policy and enforcement. Their importance is signaled, among other things, by the Commission's conclusion that the fact Microsoft would have to divulge some intellectual property-protected information was an inadequate justification for its otherwise "anti- competitive" conduct.08 Adopting "viability" as the standard by which to judge harm to the competitive process allegedly intellectual property is actually heavily utilitarian, not Lockean or Hegelian"); Sinclair & Carroll Co. v. Interchemical Corp., 325 U.S. 327, 330-31 (1945) ("The primary purpose of our patent system is not reward of the individual but the advancement of the arts and sciences."); id. at 331 n.1 (observing that the role of the patent system is "much deeper and the effect much wider than individual gain"). .08 Notably, even avid proponents of the natural rights view often acknowledge the indeterminacy underlying that perspective. See, e.g., Yen, supra note 55, at 517 n.1 (conceding in the first sentence that "[d]efining the natural law is a difficult task, and any definition will be incomplete"). 107 See id. at 520 (doubting "whether economics is in fact capable of yielding reliable policy recommendations"). 108 Commission Decision, supra note 88, 1005-09 COL UMIA B USINESS LA W RE VIE W [Vol. 2009 effected by unilateral behavior is to discard widely accepted economic theory in favor of either an ordoliberal view, or one clearly influenced by the discredited S-C-P paradigm of the 1950s. 1°9 Despite the possibility that the CFI would temper the Commission's overbroad standards, on September 17, 2007, the CFI essentially approved the Commission Decision in all material aspects.11 ° In particular, the court rejected the argument that Microsoft had an unqualified right to the exclusive use of its intellectual property-protected information.' In doing so, it purported to rely on the Magill and IMS decisions explored above, but construed them so broadly as to re-write them. 12 Most worryingly, the court held that not all competition would have to be eliminated in secondary markets to trigger a duty to share and that merely hindering the technical development of new products might also suffice."' Ultimately, as noted above, the decision rests on the implicit assumption that any harm to incentives that might flow from subjecting Microsoft to a duty to share is outweighed by the short-run boon to consumers."' But, this assumption is both unproven and unprovable. It represents either a guess about relative benefits, a political bias in favor of the short term, or skepticism about the relationship between incentives and innovation. Acknowledging the critical indeterminacy underlying the CFI's holding, we explore in the following Part five deficiencies in the judgment that are of such magnitude as to warrant restricting the precedential value of the decision. These will 109 Harold Demsetz, Two Systems of Belief about Monopoly, in INDUSTRIAL CONCENTRATION: THE NEW LEARNING (Harvey J. Goldschmid et al. eds., 1974) (debunking the S-C-P paradigm). ... CFI Decision, supra note 2. 111 Id. 690. 112 Id. 113 Id. 55 709-711. 11 This approach was explicitly adopted by the Commission. See id. 669 (citing Commission Decision, supra note 88, 783). No. 1:67] MICR OSOFT'S FIVE FATAL FLA WS COLUMBIA BUSINESS LAW REVIEW be explored one at a time in Part III, once the European focus on ordoliberal principles is briefly explained. E. Europe's Ordoliberal Tradition A relatively obscure philosophic approach to competition known as "ordoliberalism" has played a significant role in the development and evolution of EC antitrust jurisprudence. 115 Europe's ordoliberal tradition has significant-and perhaps dire-consequences for the formulation of competition policy for the new economy. It also has appreciable explanatory power in analyzing the outcome in Microsoft. The so-called Freiburg School of ordoliberalism arose in the 1930s following the collapse of the Weimar Republic. The contributors to this school of thought attributed the socio-economic and political breakdown of the Weimar Republic to the flawed legal institutions of the time and the resulting failure to check the abusive exercise of economic power by cartels and monopolists. The central tenets of ordoliberal thought are succinctly explained by David Gerber: Against the backdrop of chaos, devastation and amassed power, the ordoliberals set out to create a tolerant and humane society that would protect human dignity and personal freedom. Drawing on the central values of classical liberalism, they envisioned a society in which individuals were as free as possible .... The Weimar experience led ordoliberals to demand the dispersion of not only political power, but economic power as well. For most, this meant the elimination of monopolies. For others, such as Wilhelm Ropke, the concentration of economic resources was an evil unto itself; they sought an 115 See Gerber, supra note 5. [Vol. 2009 MICROSOFT'S FIVE FATAL FLAWS economy composed, to the extent possible, of small and medium-sized firms." 6 The Freiburg School thus takes a hostile view of the accumulation of economic power by a single entity, regarding such concentration as a potential danger to consumers and society at large. Importantly, this does not mean that ordoliberals oppose competition. To the contrary, they embrace it, viewing an effective competitive process as the heart of a socioeconomic structure that maximizes consumer well-being and political stability. However, the role of "competition" under this view clearly conflicts with new economy economics, which teach that exclusivity and monopoly may drive the competitive process in innovation." 7 By contrast, the ordoliberal view is deeply suspicious of monopolies, seeks to guarantee open access to markets, to facilitate the free movement of goods and information, and to generate cooperation, coherence, and political stability."' The potential for conflict becomes apparent when one recognizes that "ordoliberal thought set the tone for thinking about competition law within the Communities."" 9 Several aspects of the CFI's decision in Microsoft echo the ordoliberal view of competition. In particular, the court 116 Id. at 36-37. 117 See supra Part II.B. 118 For a particularly helpful articulation of this view, see Calixto Salomdo Filho, Legal Theory of Economic Knowledge, available at http:/! islandia.law.yale.edu/sela/esalamao.pdf ("For the Ordo-liberals, the great advantage of the competition system is the fact that through the transmission of information and freedom of choice, it allows the discovery of the better available options and the most rational behaviors. This is Hayek's classic definition of the competition system as a Entdeckungsverfahren (discovery process). The existence of a free choice process and discovery of the best options in the market depends on not only the existence of a price that is not altered by artificial offer and demand conditions, which would represent the marginal utility of a product, as the Neoclassics wish, but also on the existence of actual or potential possibilities of choice among products, based on price, quality, regional preferences, etc. The possibility of choice has its own value."). 119 Gerber, supra note 5, at 73. No. 1:671 lamented the "limitation thus placed on consumer choice" caused by Microsoft's actions, effectively equating the fact of choice with consumer welfare. 2 ° While an economic construction of consumer welfare might incorporate this concern, it would place much more importance on considerations of quality or price. The Freiburg view, however, considers the presence of choice to be critical, even though a monopolist facing no competition at all might conceivably offer a single, objectively superior product. Likewise, the CFI spoke of an "effective competitive structure," necessarily implying that a highly concentrated or dominated market cannot be competitive'l-a position that was exposed as clearly erroneous by the Chicago School in the 1970s, 2 but that nevertheless remains central to ordoliberalism. Similarly, the CFI viewed the fact that Microsoft's "competitors are placed at a disadvantage" in the competitive process as highly significant.'23 While an economics-based approach would consider the fact of harm to competitors as irrelevant, without some indication of harm to allocative efficiency,'24 an ordoliberal policy seeks to maintain a competitive market by ensuring that all competitors are free to compete on equal footing, regardless of preexisting allocative harm. One fundamental problem with the ordoliberal approach lies in its stark inconsistency with new economy economics. Monopoly needs to be recognized as a legitimate, even an inevitable, outcome in information markets, in order for innovation-based competition to flourish. The Freiburg approach threatens to impose obsolete and incongruous principles on a market setting that simply cannot 120 CFI Decision, supra note 2, J 652. 121 Id. 664. 122 See, e.g., Frank H. Easterbrook, The Chicago School and Exclusionary Conduct, 31 HARv. J.L. & PUB. POLY 439, 440 (2008); Timothy J. Muris, Economics and Antitrust, 5 GEO. MASON L. REv. 303, 303-06 (1997); Timothy F. Bresnahan & Peter C Reiss, Entry and Competition in Concentrated Markets, 99 J. POL. ECON. 977 (1991) 123 CFI Decision, supra note 2, 653. 124 Easterbrook, supra note 122, at 439-40. COL UMBIA B USINESS LA W RE VIE W [Vol. 2009 accommodate them. As should be clear throughout the following Part, the ordoliberal influence underscores many of the significant flaws plaguing the Microsoft decision. III. MICROSOFTS FIVE FATAL FLAWS Five fatal flaws plague the CFI's decision in Microsoft. While any one of them would suffice to cast serious doubt on the wisdom of the opinion, collectively they demonstrate its utter weakness. This Part addresses each shortcoming in turn, explaining why it is unequivocally erroneous. A. Rewriting Relevant Precedent and Defining "Effective" Competition Despite ostensibly tailoring its decision to relevant precedent governing refusals to supply, the CFI significantly transformed prior case law to facilitate its desired holding. In particular, the court subtly re-wrote the Magill test to inject a requirement that the refusal exclude any "effective" competition in a neighboring market.125 This seemingly innocuous alteration will likely create a host of problems. More specifically, summarizing Magill, the CFI observed that the refusal to license there, in the first place, "concerned a product.., the supply of which was indispensable to the exercise of the activity in question"; in the second place, "prevented the appearance of a new product"; in the third place, "was not justified"; and in the fourth place, "reserved to [the dominant firms] themselves a secondary market."2 ' The CFI then referred to Bronner and its affirmation of the "exceptional circumstances" test described above,'27 and then to IMS Health, which it characterized as holding that proof of "three cumulative conditions" was sufficient to satisfy the test: "that the refusal prevents the emergence of a new product for which there is a potential consumer demand, 125 CFI Decision, supra note 2, 1% 324, 332. 126 Id. 324. 127 Id. 1 325, 328. No. 1:67] MIfCROSOFT'S FIVE FATAL FLA WS COLUMBIA BUSINESS LAW REVIEW that it is unjustified and that it is such as to exclude any competition on a secondary market."1"8 At paragraph 332, however, the CFI restated that test in its own words, slightly but significantly modifying it. Thus: [T]he following circumstances, in particular, must be considered to be exceptional: in the first place, the refusal relates to a product or service indispensable to the exercise of a particular activity on a neighbouring market; in the second place, the refusal is of such a kind as to exclude any effective competition on that neighbouring market; [and] in the third place, the refusal prevents the appearance of a new product for which there is potential consumer demand. 129 This is a remarkable misconstruction. The requirement under Magill and IMS Health that the refusal to deal foreclose "any" competition, and Microsoft's insistence that "effective" competition be precluded, are subtly but profoundly different. The small change in language creates and conceals a far more expansive, and potentially intrusive, test-one that is susceptible to wide and imperfect application. Indeed, prior to Microsoft, many in Europe had viewed Magill and IMS as essential facilities cases because in each the recalcitrant rights holder had arguably used its power to interpose a bottleneck that prevented would-be rivals from offering any alternative to its product.13 ° A standard focused on "any" competition would permit a broad range of exclusive behavior, even if the refusal at issue severely disadvantaged existing rivals, raised their costs, or denied them access to quality-enhancing technology. As long as the refusal was not so severe as to entirely foreclose rivals from access to the market-thus precluding "any" competition-the dominant firm's exclusive behavior would 12 Id. IT 329-330 (emphasis added). 129 Id. T 332 (emphasis added). 130 See Robert Pitofsky, Donna Patterson & Jonathan Hooks, The Essential Facilities Doctrine Under U.S. Antitrust Law, 70 ANTITRUST L.J. 443, 444-45 (2002). (Vol. 2009 MICROSOFT'S FIVE FATAL FLAWS be lawful. Such a rule would conceivably allow a dominant undertaking to maintain superiority over its inferior rivals, who would occupy a fringe position in the market. It would also have the enormous benefit of ease of application: when "all" competition is prevented, the rule may be invoked; when "some" exists-whether it is "effective" or not-the rule would not apply. In contrast, depending on the definition of "effective"- described in more detail momentarily-the CFI's standard could condemn refusals to supply that do no more than hinder rivals' ability to compete on equal ground. In fact, the more valuable the intellectual property right at issue, the greater the case for mandatory disclosure under a rule founded on maintaining "effective" competition. At its logical limit, a duty to disclose proprietary information necessary to facilitate such competition amounts to a right of open access to valuable intellectual property. Under such a standard, any nonsubstitutable technology protected by copyright or patent would be subject to a liability rule and mandatory access. As applied to Microsoft, the CFI's standard could conceivably be employed to justify rivals' access to the source code for its dominant software. Although the Commission went to great lengths to emphasize that it sought to compel no such right of access, 1 3 this assertion is quite distinct from the creation of a precedent that would by its terms facilitate such access in the future. On this basis alone, one can appreciate the serious ramifications of the CFI's decision to rewrite precedent. The fact that the CFI misconstrued vital precedent is not the end of the problem. By moving to a criterion based upon the prevention of "effective" competition, the court necessarily took on the task of describing in coherent terms, if not defining, the meaning of the term "effective." We have already seen that various definitions of the term could give rise to extraordinarily broad incursions into a dominant company's intellectual property. Avoiding this outcome 131 Commission Decision, supra note 88, 569-72, 714; CFI Decision, supra note 2, 40. No. 1:67] COLUMBIA BUSINESS LAW REVIEW would require careful extrapolation of what is-and, just as important-what is not encompassed by this standard. Unfortunately, the CFI addressed this issue in a profoundly unsatisfactory way, by equating effectiveness with "viability,"'32 an equation that bears the potential for highly perverse application to unilateral behavior in the new economy setting. Presumably, "viability" refers to some characteristic of the dominant firm's rivals, and not necessarily to the state of competition or to an economically defined test of consumer welfare. But which characteristic? The court did not say. Although the court did not unequivocally embrace the "viability" test, it did so by implication-given its frequent reference to the concept. 33 In any event, the CFI did not expressly disapprove the test, even though the Commission directly relied upon it. 3 1 So construed, the "viability" standard facilitates the worst examples of mandatory access referenced above. The more valuable the intellectual property right, and thus the greater the need for competitors to access the protected technology, the more plausible is the claim that a refusal to supply denies rivals the ability to compete viably. It is not difficult to apply this line of reasoning to the pursuit of source code, amongst other nefarious scenarios. The final flaw under this first rubric lies in an ordoliberal conception of welfare that would conflict with an approach founded on economics. In attempting to link consumer welfare more directly to the "efficacy" test, the court seemed roughly to view improved consumer welfare solely in terms of additional consumer choice,1 3 ' a view whose implications would open up many pieces of intellectual property to the enforced sharing obligation imposed in Microsoft. The main constituent elements of consumer welfare- economically defined-are output, price, and quality. 132 CFI Decision, supra note 2, atpassim. 133 Id. 131 Commission Decision, supra note 88, IT 421, 455, 457, 523, 565, 606, 666, 679, 685, 687, 700, 722, 817. 135 CFI Decision, supra note 2, I 660-63. [Vol. 2009 MICROSOFT'S FIVE FA TAL FLAWS Although some consumers may value the easy availability of an eclectic range of goods or services, choice itself is no substitute for access to an objectively superior product. In the world of the new economy, where intellectual property drives innovation, a strong tension may exist between the provision of choice and the promotion of improved technology. Only by granting the inventor of a qualitatively superior standard an exclusive right to reproduce it, and thereby denying consumers a choice between multiple purveyors of the technology, can desirable incentive structures be maintained. By focusing on the fact of choice itself as an integral constituent of consumer welfare-a position no doubt informed by ordoliberalism-the CFI ignored this crucial trade-off. In doing so, it has sown the seeds for extraordinarily broad rights of access to intellectual property, the very purpose of which is to deny choice.136 B. Relying on Consumer Survey Data By adopting an ordoliberal construction of the competitive process, which accords primacy to consumer choice and markets comprised of "viable" rivals, the CFI eschewed Chicago-derived conceptions of consumer welfare. Unlike Chicago and post-Chicago approaches to competition policy, which never inquire into the status of a single competitor, 137 the CFI adopted a test that sedulously tracks the well-being of individual rivals. This aspect of the Microsoft decision is not surprising-a frequent accusation leveled at EC 136 The intellectual property laws are specifically (and in the United States exclusively) designed to spur the creation of easily appropriated information. Only by bestowing information goods with the trait of physical property-excludability-can an inventor overcome the public good nature of information and engage in the costly process of innovation with an expectation of recouping its cost of investment and earning a sufficient premium. The very fact of excludability demands exclusivity, which is the antithesis of choice. This aspect of the CFI's decision amounts to little more than a direct affront to the purpose of intellectual property. .37 See generally Herbert Hovenkamp, Post-Chicago Antitrust: A Review and Critique, 2001 COLUM. Bus. L. REv. 257 (2001). No. 1:67] COLUMBIA BUSINESS LAW REVIEW competition doctrine has been its tendency to protect individual competitors 3 ' in place of promoting the U.S. model of Darwinian competition where "[a]ntitrust and bankruptcy go hand in hand."'3 9 The CFI's focus on "viable" competition gives primacy of place to the status of competitors, naturally raising the question whether the decision considered the ultimate benefit of consumers at all. Perhaps surprisingly, it is clear from the judgment that the CFI was indeed concerned with the well-being of consumers. Notably, an important component of the court's three-part test requires a showing of "potential consumer demand" for the product foreclosed by the relevant refusal to supply. 4 ° From a Chicago-oriented perspective, the idea that a standard that looks first to the status of competitors may somehow be consistent with the promotion of consumer welfare would be nonsensical, but, as noted, the Freiburg definition of welfare is distinct.' From an ordoliberal perspective, a legal regime that maintains "effective" competition, ensures free choice, and thus avoids market concentration, is considered apodictically to promote consumer welfare. The court's emphases were consistent with this latter point, noting as it did that consumer harm need not take place directly for a violation of Article 82 EC to be found-indirect injury will suffice. 14 2 An important question then is how such indirect harm to consumers should be demonstrated. One would imagine that the answer is straightforward, and that satisfying Microsoft's ' See Eleanor M. Fox, We Protect Competition, You Protect Competitors, 26 WORLD COMP. L. & ECON. REV. 149, 149 (2003); John R. Wilke, U.S. Antitrust Chief Criticizes EU Decision to Reject Merger of GE and Honeywell, WALL ST. J., July 5, 2001, at A3 (quoting Assistant Attorney General Charles James: "Clear and longstanding U.S. antitrust policy holds that the antitrust laws protect competition, not competitors... [the European Commission's decision] reflects a significant point of divergence."). ... See Easterbrook, supra note 122, at 440. 140 CFI Decision, supra note 2, 332. 141 See supra Part II.E. 142 CFI Decision, supra note 2, 664. [Vol. 2009 three-part test would suffice. 14 3 Indeed, the court seemed to make clear that a dominant company's refusal to supply something indispensable to a neighboring market, which excludes any effective competition and prevents the appearance of a new product for which there is potential consumer demand, constitutes an abuse of its position from which indirect harm to consumers can be presumed. 44 If such a presumption exists, however, there should be no need for a court to take the superfluous step of ascertaining actual consumer preferences. Yet, this is precisely what the CFI did in Microsoft, concluding on the basis of documentary evidence alone (the Mercer survey) that consumers preferred non-Microsoft work group server operating systems "on a number of features to which they attach great importance."145 This stark incongruity is a considerable weakness in the decision. As a preliminary matter, it is not clear whether courts are competent to make such a determination, even assuming that consumer preference surveys are sufficiently reliable to inform a judicial finding of abuse.'46 More fundamentally, however, and setting aside the issue of judicial competency, the conclusion seems to contradict in greater or lesser degree all three limbs of the "exceptional circumstances" test. C. Embracing the Discredited S-C-P Model Apparently uncomfortable with its previous statements about the meaning of "effective competition," the court announced that Article 82 concerns itself not just with "practices which may prejudice consumers directly" but also those which indirectly prejudice them by "impairing an 143 Id. 332. 144 Id. 145 Id. 661. 141 Surveys may be vulnerable to bias on the part of the surveyor, consumer myopia (which may be exacerbated by the cascade effect) and, worst of all, the sheer indeterminacy of a questionnaire that asks participants to make determinations concerning a hypothetical market environment that does not exist. No. 1:67] MCR OSOFT'S FIVE FA TAL FLA WS COLUMBIA BUSINESS LA W REVIEW effective competitive structure."'47 The CFI went further by concluding that "Microsoft impaired the effective competitive structure on the work group server operating systems market by acquiring a significant market share on that market."148 These statements do several things: first, they suggest that perhaps consumers in these markets were not prejudiced directly;'49 second, they assume that the term "an effective competitive structure" needs no description or further elaboration; and third, they implicitly equate "effective competition" with an "effective competitive structure," and assume that a concentrated market is an ineffective one. This limb of the analysis is particularly weak. Not only does it echo the discarded "structure-conduct-performance" paradigm of the 1950s-a paradigm that was rejected because it failed to reflect economic and commercial reality- but it also requires someone, be it the courts, Commission, or some other party, to hypothesize what the relevant structure might be for each market, taking into account economies of scale, first-mover advantage, risk-taking, minimum efficient scale, dominance properly achieved, path dependency, and so on. The opinion says nothing about who is to do this, or how. All we know is the court's summary conclusion that Microsoft "impaired the effective competitive structure ... by "9150acquiring a significant market share .... The CFI's focus on market structure hearkens back to Bain's well-known 1951 work, where he found that rates of return in relatively concentrated industries significantly surpassed those in relatively unconcentrated markets.'5 ' He concluded that the results supported the theory of the SCP paradigm, which predicts that higher levels of concentration in an industry (structure), will facilitate collusive and "4 CFI Decision, supra note 2, 664. 148 Id. 149 The repercussions of this point have been explored above. See supra Part III.B. 150 CFI Decision, supra note 2, 664. ..' Joe S. Bain, Relation of Profit Rate to Industry Concentration: American Manufacturing, 1936-40, 65 Q. J. ECON. 293 (1951). [Vol. 2009 monopolistic pricing (conduct), which will in turn lead to abnormally high rates of return (performance). Consequently, the SCP model regards higher levels of concentration with grave suspicion. As applied to Microsoft, the CFI was concerned that in the concentrated market for work server operating systems,'52 Microsoft's refusal to supply had prevented its rivals from obtaining a "viable" position-greater market share presumably-and had thus entrenched Microsoft's dominance. The far-from-subtle implications, of course, are that reducing Microsoft's market share is a necessarily desirable outcome; that the reduction is valuable for its own sake; and that there is a determinable level of reduction that would restore, or create, "effective" competition. The economics of new economy industries directly contradict this assumption, as has been explored above.'53 The CFI's reliance on the SCP model is flawed for another reason, however; namely, that the model has been conclusively debunked on numerous grounds. In particular, the Chicago School successfully challenged the assumption that higher rates of return suggest the presence of market power, arguing instead that superior efficiency explains enhanced profitability.' More specifically, the acquisition of scale economies with increasing market share will enhance profitability as long as the company is on the downward- sloping portion of its long-run average cost curve. If efficiency leads to success, and success leads to dominance, attacking concentrated markets will punish efficiency and serve only to stunt competition and harm consumers. In short, the application of the SCP paradigm leads to perverse results. This conclusion holds equally true with respect to the CFI's holding that a refusal to deal will constitute an abuse of a dominant position when it "impair[s] an effective 152 Commission Decision supra note 88, 473-99. 153 See supra Part II.B. " Harold Demsetz, Industry Structure, Market Rivalry and Public Policy, 16 J.L. ECON. 1 (1973). No. 1:67] MCR OSOFT'S FIVE FATAL FLA WS competitive structure."155 One of the implications of this test would seem to be that if, in some subsequent round of communication, Sun or another rival should request that Microsoft provide it with the source code for its operating system-claiming plausibly, one assumes, that the provision of the interoperability information failed to bring about an "effective competitive structure" in the relevant market-the logic of the newly minted "exceptional circumstances" test might well suggest that the source code should be shared as well. It involves little in the way of hyperbole to suggest that the CFI's test could promote an anticompetitive antitrust policy. D. Administrative Costs The decision of whether a given situation is so exceptional as to warrant interoperability does not mark the end of a court's proper inquiry-serious problems remain with respect to application. Even if enforcers had reliable means by which to ensure that ex post rewards would never fall below desirable levels, which would of course enable them to demarcate whether and when compulsory licensing would be proper, the formal decision that requires interoperability precedes the major problem of access. More specifically, if one decides that a dominant intellectual property-holder should make its protected code available to rivals, at what price should it be made available, and who should monitor and approve changes in price? The same problems that have been so well catalogued in the literature criticizing the administration of an essential facilities test apply in this area, and with no less vigor. How should access be priced? At the monopoly level? At "cost"? Would "cost" include the risk that the investment sunk in the innovative process might have been lost? For how long would an administered price remain in effect, and what circumstances might require or allow it to change? Should the price rise as rivals' market shares increase? Who would resolve disputes about pricing and access, and what sort of 155 CFI Decision, supra note 2, 664. [Vol. 2009COL UMBIA B USINESS LA W RE VIE W MICROSOFT'S FIVE FATAL FLAWS process would exist for dispute resolution? This is the short list of difficult questions; the long one would fill many pages. At the very least, it was incumbent upon the CFI to articulate some sound basis for believing that these challenges are surmountable and, indeed, to provide some minimum guidance on how mandatory access should be managed. The court conspicuously failed to address either issue, which constitutes a rather fundamental shortcoming, given the very real difficulties inherent in administering interoperable remedies. E. The Final Flaw: Rejecting Ex Ante Incentives as an "Objective Justification" Part II explored the crucial distinction between traditional and new economy industries, and how monopoly inventors may serve as particularly important catalysts to economic growth in the latter context. At a minimum, one would expect a prudent rule to display some sensitivity to the critical tradeoff between the long- and short-run. If one generally places primacy on the immediate welfare of consumers, and regards the danger of long-term harm to innovation with some skepticism, at the very least some mechanism must exist to highlight that danger in particularly egregious circumstances. That is, there must be some short-term benefits so small or ephemeral, and some long-term incentive effects so large and self-evident, that even the skeptic would insist on a means of weighing them off empirically, rather than by a priori assumption. For instance, in especially high-risk, capital intensive technology markets, the risk of diluting valuable intellectual property is considerably greater than in other contexts. 5 ' What is perhaps most striking-and damning-about the Microsoft decision is its effective disavowal of any defense based on maintaining proper incentives. Microsoft tried to offer an "objective justification" for its refusal, claiming that the compulsory sharing of the kind ordered by the court would adversely affect incentives to 15 See supra note 23. No. 1:67] innovate-its own, its current and future rivals, and those of other firms not involved in this controversy.'57 According to the CFI, Microsoft failed to establish its "burden of proof' on this point, and simply put forward "vague, general, and theoretical" arguments.15 But how else would anyone "prove" the disincentive effects of such a legal rule? As a general social matter, we believe that law can alter incentives to innovate but we- society in general, that is-accept that proposition without having any real "proof' of it. The CFI's treatment of the issue means that adverse incentive effects are effectively eliminated as an objective justification. Given the overriding importance of maintaining proper incentive structures in the new economy, Europe's decision to discard any consideration of them is highly ill-advised. IV. CONCLUSION The CFI's decision in Microsoft is of considerable importance to the future of antitrust policy in high- technology markets. Among other things, the CFI re-wrote binding precedent to create a potentially open-ended standard that is difficult both to define and to administer, and that could conceivably justify incursions into source code and other centrally important intellectual property-protected technology. Despite the Commission's repeated emphasis that it never sought to compel the release of such information,'59 the fact of facilitative precedent itself creates an appreciable risk that a suitably minded regulator may attempt precisely such a feat in the future. One need merely look at Magill to see how a case premised on narrow facts and refusals to supply intellectual property of dubious legitimacy can be employed to justify far greater incursions in the future. C OFI Decision, supra note 2, 627, 659, 668. 15 Id. 697-98. '5 See Commission Decision, supra note 88, 569-72, 714; CFI Decision, supra note 2, 140. [Vol. 2009COL UMBIA B USINESS LA W RE VIE W Moreover, and despite European officials' repeated emphasis that EC competition law is premised on notions of efficiency and consumer welfare, 16 1 Microsoft is remarkable for its embrace of the discredited SCP paradigm. The CFI's adoption of SCP-inspired jurisprudence is further exacerbated by its ordoliberal-influenced perspective on effective competition. The economic literature now clearly demonstrates that heavily concentrated, or even ostensibly monopolized, information markets may be highly competitive.'61 To assume, as the CFI did, that an absence of ex post choice to consumers implies a market failure is to disregard entirely the fundamental teachings of economics. Although Freiburg-driven principles of ordoliberalism could conceivably be applied reasonably to traditional settings, they are anathema to new economy markets. Innovation-based markets may well operate at their best when they take the form of sequential, short-lived monopoly.'62 To impose ordoliberal principles of free access to information markets-in other words, to dilute the exclusivity of intellectual property-may be to eviscerate incentives and ultimately to prevent competition itself. The CFI's final transgression is the most serious. As noted, there may conceivably be cases where tempering the exclusive force of intellectual property through a liability rule may benefit consumers in the short-run by more than it harms long-run innovation. To facilitate such a calculus, however, at the very minimum the applicable test must recognize the fact of possible harm to ex ante incentives in some settings. The CFI discards this crucial prerequisite by 160 See, e.g., Neelie Kroes, European Commissioner for Competition, Preliminary Thoughts on Policy Review of Article 82, (Sept. 23, 2005), available at http://www.abanet.org/intlaw/calendar/fall2005/policyreview- article82.doc ("I am aware that it is often suggested that-unlike Section 2 of the Sherman Act-Article 82 is intrinsically concerned with "fairness" and therefore not focused primarily on consumer welfare .... My own philosophy on this is fairly simple. First, it is competition, and not competitors, that is to be protected. Second, ultimately the aim is to avoid consumers [sic] harm."). 161 See Demsetz, supra note 154. 162 See POSNER, supra note 9, at 248-50. MCROSOFT'S FIVE FATAL FLAWSNo. 1:671 COLUMBIA BUSINESS LAW REVIEW requiring a defendant to provide factual evidence of diminished incentives, and ignoring powerful theoretical evidence." 3 Yet, reliance on theory may be the only way to make the case for harm to the innovative process. The court's test therefore amounts to a rejection of ex ante considerations altogether. Combined, these flaws serve to condemn the case on several important grounds. At a minimum, the CFI's judgment in Microsoft should be construed narrowly for the purpose of future cases where rivals seek access to a successful company's intellectual property. With respect to the decision's implications for U.S. domestic antitrust policy, the Justice Department's initial reaction of skepticism and concern was the correct one." Newer antitrust regimes should recognize that the decision departs radically from both prior EC law and the U.S. view. Unfortunately, this latest chapter in the Microsoft saga has further revealed a considerable transatlantic divide in the treatment to be afforded unilateral behavior that triggers competitor complaint. To borrow from William Kolasky, it is indeed a long way from Chicago to Brussels."5 163 CFI Decision, supra note 2, 698. 16 See supra note 1. 165 William Kolasky, Deputy Assistant Attorney Gen. in Antitrust Div., Dep't. of Justice, Conglomerate Mergers and Range Effect: It's a Long Way from Chicago to Brussels, address to the George Mason University Symposium, Nov. 9, 2001, available at http://www.usdoj.gov/atr/public/ speeches/9536.htm. [Vol. 2009