European Union: The European Commission’s Microsoft Decision

In its Decision of 24 March 2004, the European Commission found Microsoft guilty of abusing Article 82 of the EC Treaty by withholding from competitors information required to permit interoperability between the Windows PC and server operating systems and other operating systems, and by bundling Media Player with the PC operating system. By way of remedy, it ordered Microsoft to supply specified interface information to competitors, to offer a version of Windows without a media player as well as the version with it, and to pay a fine of EUR 497.2 million (approximately USD 610 million). This memorandum provides a brief analysis of the decision (the "Decision") and its implications.

In summary:

• In one sense, the Decision can be seen as sui generis. The Commission has never before been confronted with such a position of "super-dominance" in a market occupying such a central position in the economy. The special characteristics of software add an additional layer of difficulty and challenge.

• The Decision is clearly intended to establish a framework for the Commission’s handling of a number of important and contentious issues, already the subject of pending complaints and proceedings, regarding (1) the role which Microsoft, will play in the information technology sector and (2) the standards for determining an abuse of a dominant position in the high-tech sector and generally, especially for "refusals to supply" intellectual property and bundling complementary technologies into an existing product. In effect, the Decision sets up a situation which, if left undisturbed by the appeals process, allots to the Commission a supervisory and regulatory role in relation to Microsoft’s continuing operation in the various areas where it can bring to bear its existing market power – areas which include the internet, mobile telephony, digital rights management, etc.

• The Commission must now run the gauntlet of the appeals process in the European courts – starting with the request which Microsoft will make to the President of the Court of First Instance ("CFI") for interim measures suspending the decision (a proceeding which could yield a decision in some number of months), and proceeding to the full decision on the merits by the CFI and, if necessary, the European Court of Justice ("ECJ") (a process which could last several years). Given the nature of the subject matter, and the speed of evolution in the IT sector, the outcome of the interim measures proceeding will be of primary importance. Not surprisingly given the novelty and complexity of the case, there are a large number of issues where the Commission may be at risk on appeal. Two which stand out in particular are (i) Microsoft’s accusation that the Commission’s infringement finding and remedy order with respect to interoperability are not consistent with the case law flowing from the Magill judgment of 1995¹ concerning the circumstances in which failure to license IPRs may be an abuse under Article 82 – a challenge which may have been strengthened by the judgment handed down by the ECJ on April 29 in the IMS case² – and (ii) the Commission’s assimilation of software bundling to tying under Article 82(d), the standard of harm to competition applied with respect to bundling, the novel unbundling remedy, and the implications for software producers when they incorporate new features in their products.
• Haunted by the memory of a series of stinging defeats in 2002 in merger regulation cases, where the CFI castigated the Commission for sloppy fact finding as well as bad economics, the Commission has gone to exceptional lengths to try to put together a strong factual and legal case. In doing so, it has been aided by the peculiar character of the existing case law under Article 82. The new wave of economic analysis which has been so prominent in relation to Article 81 and to the Merger Regulation has so far only made limited and uncertain strides in relation to Article 82. The case law as it stands appears to afford the Commission a lot of latitude in unilateral infringement cases in crafting theories of infringement and remedies. The Decision exploits this latitude to the full, adducing a large number of different theories. The Commission seems to have used the Decision to attempt to articulate a more systematic analytical framework for dealing with cases involving interoperability and bundling, and more generally to unilateral infringement cases in the high technology sector. One major question is how this framework (with any changes to it which may emerge from the appeal process) will apply in the software sector to situations which may involve dominance but lack the "super-dominant" character of Microsoft’s market position.

• Finally, the Decision shines a spotlight on some real and deep divergences between the law and practice on the two sides of the Atlantic in relation to unilateral infringing behavior, which will presumably have to be addressed in the future, especially in relation to cases like this one, where the outcome of a proceeding by either the U.S. or the EU is intrinsically global in scope. A comparison of the Decision with the recent decision by the U.S. Supreme Court in Verizon Communications v. Trinko (Trinko) reveal that U.S. and EU law in this area have a substantially different structure and analytical framework. For example:

• Under U.S. antitrust law, Section 2 of the Sherman Act prohibits unlawful "monopolization" and "attempts to monopolize." To establish a "monopolization" claim, proof of monopoly power alone is not sufficient. As the Court emphasized in Trinko, "[t]o safeguard the incentive to innovate, the possession of monopoly power will not be found unlawful unless it is accompanied by an element of anticompetitive conduct³ For the conduct to be anticompetitive, it must serve no legitimate purpose or be unprofitable and undertaken solely to exclude or weaken competitors in anticipation of later recouping supra-competitive profits.⁴

• In Trinko, the Court determined that Verizon’s failure to provide access to its local phone network did not constitute unlawful monopolization or an attempt to monopolize. In addition, the Court declined to adopt the "essential facilities" doctrine developed by some lower U.S. courts to impose a duty for a monopolist controlling such a facility to deal with its rivals. In rejecting the monopolization claim, the Court reemphasized that a refusal to deal or cooperate with one’s rivals, even by a monopolist, is generally not considered to be the type of anticompetitive or exclusionary conduct prohibited by U.S. antitrust laws. While there have been a few exceptions to this rule under U.S. law, the Court indicated it has "been very cautious in recognizing such exceptions, because of the uncertain virtue of forced sharing," including (1) its potential to lessen the incentive of firms to innovate and possibly facilitate collusion and (2) the need for courts to act as "central planners, identifying the proper price, quantity and other terms of dealing - a role for which they are ill-suited."⁵ Finally, the Court rejected the "monopoly leveraging" theory accepted by the lower court and indicated that a monopolist’s efforts to extend or "leverage" its monopoly power must do more than merely result in a competitive advantage in another market⁶ To prove an unlawful attempt to monopolize, a firm’s conduct must be "predatory" or "exclusionary" with a "specific intent" to monopolize and have a "dangerous probability of achieving monopoly power." ⁷
• The corresponding concept of "dominance" under Article 82 involves a substantially lesser degree of market power than what the Sherman Act requires to prove monopoly power. Moreover, the conduct by a "dominant" company sanctioned by Article 82 is an open-ended collection of behaviors, some exploitative and some having foreclosure effects. And, the required probability and magnitude of harm is relatively easier for an enforcement agency to establish.⁸ For example, charging excessive prices or impeding or creating a risk of impeding competition, sometimes even by only one competitor, can be considered abusive conduct. The flexibility for "abuse of dominance" under Article 82 is also shown by the fact that, once dominance on a given market is found, the infringing conduct and the anti-competitive effects may be found either on that market or on other "adjacent" markets. Thus, in the Microsoft Decision, the Commission determined that Microsoft’s refusal to supply interface information was an abuse of its dominance of PC operating systems because (1) the interface information was an "indispensable" input that competing suppliers of another product (work group server operating systems) needed to remain a viable competitor and (2) Microsoft’s refusal to supply this information put competitors at a "strong competitive disadvantage" to an extent where there is a "risk of elimination of competition." Unlike a U.S. complainant under the Sherman Act, the Commission did not need to establish that Microsoft had no legitimate purpose in refusing the information (e.g., it "was willing to forsake short-term profits to achieve an anticompetitive end"⁹) or the "specific intent" to monopolize (or "dangerous probability" of monopolizing) the work group server OS market. Rather, the Decision indicates that a dominant firm should be required to deal if there is "a risk of elimination of competition, and "on balance", the possible negative impact is outweighed by the positive impact that would result from a compulsory license.

Following Sun’s initial complaint in December 1998, which was concerned only with the denial of interface information between the PC operating system and low end server operating systems, the investigation snow-balled, running through three successive statements of objections, taking in more and more interested parties, and expanding the theories of abuse to include interoperability between different server operating systems (not just between server and client operating systems), and the bundling of Media Player with Windows. The investigation team expanded correspondingly, and this has probably been the most resource-intensive proceeding concerning a single company in the history of EU competition law enforcement. The case is set to continue following the opening of new investigations concerning Microsoft’s licensing practices to OEM PC manufacturers, and further accusations of leveraging relating to other products and sectors, including mobile telephony, browsers, digital rights management, etc.

The cornerstone of the Commission’s analysis is Microsoft’s extraordinary market power on the market for PC operating systems: "Microsoft, with its market share of over 90%, occupies almost the whole market – it therefore approaches a position of complete monopoly, and can be said to hold and overwhelmingly dominant position" (para 435). The Commission emphasizes the durability of that position, stretching back at least to the mid 1990’s, and entrenched in the future by network effects and the applications barrier to entry (paras 471-72); and the reinforcement of these effects in the specific context of software due to the preference of developers and content providers to give preference to designing products to run with a "ubiquitous" operating system (para 833). It sweeps aside the argument raised by Microsoft early in the proceeding to the effect that, in assessing market power in the IT sector "it would be necessary to take into account not only demand-side and supply-side substitutability, barriers to entry, etc., but also the permanent, although unforeseeable and unspecified, threat of a possible ‘technological revolution’" (para 468). Instead, it suggests that Microsoft’s market power enables it to stifle innovation by competitors (paras 694, 711).

From this base, the Commission proceeds to examine the two specific and quite narrowly drawn cases of abuse – denial of interface information to suppliers of work group server operating systems, and bundling of Media Player – in such a way as to establish guidelines having much broader consequences for Microsoft (and possibly others) in the future. Commissioner Monti’s comments to the press at the time of the adoption of the decision (which immediately followed the failure of settlement negotiations with Microsoft) confirm the Commission’s approach of using the Decision to establish a legal framework for its continuing scrutiny of Microsoft’s practices.

A central element in this part of the case is the Commission’s identification of a separate market for work group server operating systems, defined by reference to the performance of four "core" tasks: file services, print services, and group and user administration services. Work group servers "are designed, marketed, purchased and used essentially to carry out this set of interrelated tasks" (paras 356, 379). This definition, which Microsoft has vigorously contested, yields a number of important results. The Commission finds a present market share of at least 50%, perhaps 60-75%, sufficient to support a presumption of dominance. In the market thus defined, competing operating systems such as UNIX and Linux have only a weak presence. The Commission reinforces the finding of dominance by invoking the "associative links" theory first enunciated in Tetra Pak II:¹⁰ dominance on one market may be found or supplemented if that market has close commercial and technical links to another market where the same supplier has a strongly dominant position (paras 526-540).

The Commission finds that Microsoft is abusing this dominant position "by refusing to supply Sun and other undertakings with the specifications for the protocols used by Windows work group servers in order to provide file, print and group and user administration services to Windows work group networks and allow these undertakings to implement such specifications for the purpose of developing and distributing interoperable work group server operating system products" (para 546). The Commission analyzes this abuse under the rubric of "refusal to supply," drawing on a range of cases involving refusal to supply products or services and refusal to license intellectual property rights. It is at pains to set out an approach which examines "the entirety of the circumstances surrounding a specific instance of a refusal to supply" and to refute "an approach that would advocate the existence of an exhaustive checklist of exceptional circumstances." In doing this the Commission is deliberately seeking to establish a flexible approach to the "exceptional circumstances" test relating to the duty of dominant companies to license IPRs enunciated by the ECJ in Magill; Microsoft is arguing for a narrow interpretation of that test, limited to the specific circumstances in Magill (a refusal to supply is abusive only when it prevents the development of a new product on a market separate from that on which the IPR holder is active). The Commission further considers the test of the ECJ’s key judgment on essential facilities cases in Bronner – that it is "necessary to show that supply is indispensable to carry on business in the market, which means that there is no realistic actual or potential substitute to it" (para 585). It is clearly concerned to find an interpretation of this standard which is not excessively demanding. It finds that, in light of Microsoft’s extraordinary market share on the PC operating system market and the importance for work group server operating systems of interoperability with PCs, the Bronner test is met because "Microsoft’s refusal puts Microsoft’s competitors at a strong competitive disadvantage in the work group server operating system market, to an extent where there is a risk of elimination of competition" (para 589)¹¹. In a footnote (footnote 712) the Commission adds that, while some degree of interoperability could be achieved without further disclosures from Microsoft, it would be "insufficient to enable competitors to viably stay in the market." The Commission relies heavily on customer surveys, some submitted by Microsoft and some undertaken by the Commission itself, as well as IDC market studies. It emphasizes Microsoft’s ability to prevent innovation by competing work group server operating system providers; this outweighs any reduction to Microsoft’s own incentive to innovate that might result from disclosing interface information to competitors (paras 712 – 729); in this regard, it appears to suggest that a balancing test should be applied, weighing the possible negative impact of a compulsory license on the right-holder against the positive impact on the industry as a whole (para 783). It suggests that strengthening Microsoft’s dominant position on the work group server market reinforces its position on the PC market (para 769).

The theory of the Decision is based on Article 82(d), which provides, as one of the specified (but not limitative) cases of abuses "making the conclusion of contracts subject to acceptance by the other parties of supplementary obligations which, by their nature or according to commercial usage, have no connection with the subject of such contracts." Looking only to the language of this paragraph, its application to bundling (never before the subject of a condemnation under Article 82), as opposed to contractual tying, is not obvious or uncontestable. The Commission could have used some more general theory of foreclosure of competition. Instead, it chose to rely on Article 82(d),¹² but at the same time to enunciate in a new way the elements necessary to establish an abuse under that paragraph, not only for bundling but also for conventional contractual tying:

Additional components of this test are (v) a balancing test, weighing efficiencies from tying against harm to competition, and (vi) specification of the amount and probability of harm to competition. In elaborating this test, the Commission seeks to provide a new synthesis of the existing (not highly developed) EU case law on tying under Article 82, also taking into account U.S. law on tying as reflected in the judgment of the DC Circuit in United States v. Microsoft.¹³

Applying these tests, the Decision finds that there is a separate market for media players, rejecting Microsoft’s "integrative" argument that the media player has become a natural part of an operating system. It further rejects Microsoft’s argument that to prove a violation of Article 82(d), there must be proof that that the customer is forced to pay for and use the tied product: ". . . inasmuch as tying risks foreclosing competitors, it is immaterial that consumers are not forced to ‘purchase’ or ‘use’ WMP. As long as consumers ‘automatically’ obtain WMP – even if for free – alternative suppliers are at a competitive disadvantage. This is because no other media player vendor can guarantee content and software developers similar platform ubiquity. Content providers and software developers who know that WMP is present on all Windows client PCs (more than 90% of the market) will provide Microsoft with a competitive advantage by developing content and applications primarily to WMP" (para 833).

Before providing its detailed evidence for these propositions, the Decision sets out the standard as regards the magnitude and probability of harm to competition. It draws heavily on the recent judgments of the Court of First Instance in Michelin¹⁴ and British Airways,¹⁵ which affirmed Commission decisions condemning fidelity rebates, and which held that "it is sufficient to show that the abusive conduct of the undertaking in a dominant position tends to restrict competition or, on other words, that the conduct is capable of having that effect" (para 838 and footnote 979). This enables it to reject Microsoft’s argument that other means of distribution are available for competing media players, and that other competitors are still operating with some degree of success: "in view of the indirect network effects obtaining in the media player market, the ubiquitous presence of the WMP code provides it with a significant competitive advantage, which is liable to have a harmful effect on the structure of competition in that market" (para 878). A similar approach is taken in arguing that software developers and content providers will tend to favour WMP over other media players. The Commission rejects "Microsoft’s ‘last man standing’ foreclosure rule," arguing that the evidence of harm must not be "concomitant with the ‘tipping’ of the market" (para 946). It finds that the harm to competition will also reinforce Microsoft’s operating system monopoly (para 980). The decision also finds "spill-over" effects harming competition relating to server software, formats for content and software developers, media players for wireless information devices, set-top boxes, digital rights management and on-line music delivery (para 897).

As regards the issue of efficiencies, which the Decision equates to the issue of "objective justification," it finds that Microsoft has the burden of proof and has failed to meet it, that the benefits which Microsoft claims could be achieved without tying, that they "primarily relate to Microsoft’s own profitability" and that they are in any event outweighed by the anti-competitive effects of the tying.

The full appeals process – judgment on the merits by the CFI, followed by judgment on the merits by the ECJ – will run for several years. Microsoft will accompany its appeal on the merits with a request, addressed to the President of the CFI, for a preliminary judicial order suspending the implementation of the Decision. In practice, the outcome of this preliminary proceeding will have a major and perhaps decisive effect on the continuing Commission proceedings against Microsoft.

The ECJ’s judgment of April 29, 2004 in the IMS case raises new questions about the interface information disclosure portion of the Decision, although both sides can claim support from it. In holding that refusal to license "may be regarded as abusive only where the undertaking which requested the license does not intend to limit itself essentially to duplicating the goods or services already offered on the secondary market by the owner of the copyright, but intends to produce new goods or services not offered by the owner of the right and for which there is a potential consumer demand," the ECJ appears to reject the broad, case-by-case approach to interpreting the Magill doctrine advocated by the Decision, and to make the survival of the Decision’s conclusion depend on whether it can be shown that potential licensees would produce a "new" product. Microsoft will argue that its competitors are seeking interface information essentially to duplicate its working group server product. The Commission will argue that Microsoft’s refusal to supply is abusive precisely because it prevents innovation, as clearly set forth in the Decision. The parties will also take opposing positions on whether the Decision provides proof sufficient to meet the requirement that operation by the competitor is 'impossible or at least unreasonably difficult' unless the competitor receives a license; the IMS judgment contains statements on this issue which are capable of being used by both parties.

The portion of the Decision relating to bundling also raises significant issues. One concerns the assimilation of bundling to tying under Article 82(d), which is a matter on which there is no precedent. Another concerns the standard applied concerning the degree and probability of harm to competition, where the Commission’s approach is not without support in the case law, but where some judicial clarification or reformulation is not to be excluded.

Because of the generality of this update, the information provided herein may not be applicable in all situations and should not be acted upon without specific legal advice based on particular situations.

© Morrison & Foerster LLP. All rights reserved