Course No. 9200 711 001
Th 6:30 - 9:30 p.m.
|Professor Jay Dratler, Jr.||
Room 231D (IP Alcove)
|Copyright © 2000, 2001, 2002, 2004, 2007 Jay Dratler, Jr. For permission, see CMI.|
The D.C. Circuit's Decision in Microsoft III:
More Questions than Answers
1. The D.C. Circuit begins its opinion in an odd way, describing how the computer industry is special and unique. Of course antitrust analysis must take account of an industry's special economic characteristics. But does that mean that the applicable law or principles of antitrust analysis differ for each industry?
The traditional answer to that question has been a resounding "no." The leading case is National Society of Professional Engineers v. United States, 435 U.S. 679, 98 S.Ct. 1355, 55 L.Ed.2d 637 (1978). There the government challenged a professional society's ethical rule against competitive bidding by engineers for work on construction projects. 435 U.S. at 682-683. The society tried to justify its rule by arguing that competitive bidding on price would de-emphasize the quality of engineering services, thereby threatening the safety and soundness of engineered structures. The Court rejected this argument in a passage that is worth quoting at length:
Since National Society, the Supreme Court has consistently taken the position that no segment of our economy (except for baseball, for unique historical reasons!) is exempt from the usual antitrust rules. See, e.g., FTC v. Superior Court Trial Lawyers Association, 493 U.S. 411, 423, 110 S.Ct. 768, 107 L.Ed.2d 851 (1990) (condemning boycott and price-fixing agreement by lawyers representing indigent criminal defendants, despite arguments that increased rates would improve quality of constitutionally-required representation); FTC v. Indiana Federation of Dentists, 476 U.S. 447, 464, 106 S.Ct. 2009, 90 L.Ed.2d 449 (1986) (condemning dentists' collective refusal to provide X-rays to insurance companies, notwithstanding alleged justification based on quality of care); Arizona v. Maricopa County Medical Society, 457 U.S. 332, 349-351, 102 S.Ct. 2466, 73 L.Ed.2d 48 (1982) (doctors: "[w]e are equally unpersuaded by the argument that we should not apply the per se rule in this case because the judiciary has little antitrust experience in the health care industry"); Goldfarb v. Virginia State Bar, 421 U.S. 773, 786-788, 95 S.Ct. 2004, 44 L.Ed.2d 572 (1975) (rejecting argument that "learned profession" of law is not "trade or commerce"). Cf. National Collegiate Athletic Association v. Board of Regents of the University of Oklahoma, 468 U.S. 85, 100-101, 104 S.Ct. 2948, 82 L.Ed.2d 70 (1984) (NCAA) (applying Sherman Act § 1 to collegiate sports but refusing to apply per se rule to horizontal price-fixing and limitations of output by league because collegiate sports comprise "an industry in which horizontal restraints on competition [such as leagues and rules] are essential if the product is to be available at all"). But see Flood v. Kuhn, 407 U.S. 258, 283-284, 92 S.Ct. 2099, 32 L.Ed.2d 728 (1972) (declining to overturn previously judge-made exemption for professional baseball because Congress "by its positive inaction" had allowed prior decisions to stand).
In hinting strongly that the software industry is somehow special, the court refers to the "network effects" of software, the need for software standards and the efficiencies of software integration (of which much more below). Are these things more important than the safety and soundness of structures in which people work and live, the quality of medical and dental services, and the danger of poorly made pharmaceuticals? In the cases cited above, the Supreme Court ruled or reasoned that none of these considerations trumps competition as our basic national rule of trade, precisely because competition is one of the best means to avoid the dangers of poor-quality products. Was the Microsoft III court bamboozled by the presumed complexity of computer software and fancy economic theory?
2. As the D.C. Circuit saw it, the personal computer industry is unique in part because of the importance of "network effects." The term "network effects" is just economic jargon for the added value and consumer benefit that a complex product or technology involving human cooperation produces as more and more people use it. As the "network" of users expands, some products (particularly those involved in communication), become more valuable to everyone.
Network effects are intimately related to the notion of standards. Perhaps the best example is the QWERTY keyboard. QWERTY has become a worldwide standard not only for English, but also for foreign languages that use the Roman alphabet. Using the QWERTY keyboard as a universal standard allows someone from London or (apart from special Portugese characters) Lisbon—let alone New York—to use the very same keyboard as a colleague in San Francisco without re-learning to type.
Often the very existence of a standard promotes efficiency, even if it is not the best possible standard. Again, the QWERTY keyboard is a good example. It became the standard for historical reasons only: the keys on old mechanical typewriters used to jam with "faster" keyboards. There are indeed a number of "faster" keyboards, such as the Dvorak keyboard, which is based upon scientific study of the human hand and the relative frequency of use of the letters in the English language. It is universally believed that the Dvorak keyboard is faster and more efficient than the old QWERTY keyboard, but the latter still predominates. The primary reasons are the "network effects" of so many people using it and the "lock-in effect" of the cost and delay that retraining all those people would entail.
Even such entrenched standards, however, are not necessarily eternal. Modern software and the sale of keyboards as separate products have made it easier to switch among different keyboard designs. As a result, a small but increasing number of younger computer users are training their hands and minds to use the more efficient but nonstandard Dvorak keyboard. (To verify this point, just type "Dvorak keyboard" into any Internet search engine.)
3. Do the facts of Microsoft III implicate the importance of standards per se or the importance of how standards are made? By the time of the appellate court's decision, Microsoft's Windows operating system was thoroughly entrenched as a near-universal standard for Intel-based personal computers. Microsoft had 95% of the market, which encompassed an estimated 100 million computers or more. Few would question the value, consumer benefit and economic efficiency of having a standard operating system for so many computers, even if, like the QWERTY keyboard, it is not the best.
However, was the issue in Microsoft III the entrenched standard of the operating system itself? Or was it whose product might become a new standard for Internet browsers? Netscape Communications Corp. (later acquired by AOL and then by Time-Warner) had invented the Internet browser as a widely used commercial consumer product. When Microsoft entered the browser field, Netscape had an 80% share of the browser market, and Microsoft had to play "catch-up ball." See: Sandberg, "Sun and Netscape Are Forming Alliance Against Microsoft on Internet Standard," Wall Street Journal, at B3 (Dec. 4, 1995) ("Netscape holds 80% of the market for Web ‘browser' software used by individuals' PCs"); Sandberg and Hill, "Microsoft Probe Spurs Subpoena Tied to Internet," Wall Street Journal, at A3 (Dec. 4, 1995) ("In the estimate of Netscape and industry observers, roughly 80% of World Wide Web users, a population pegged at more than 17 million, use Netscape's software"). Was the issue whether there would be a standard for Internet browsers, or who would set it and how?
Normally, technological standards are set through voluntary agreement among competing manufacturers, typically through trade associations. Among the many personal-computer standards set in this way are the JPEG and TIFF graphics formats, the popular MP3 format for songs, and the 802.11b and 802.11g standards for home networks and WiFi networks, respectively. The voluntary standard-setting process is not immune from antitrust scrutiny. See Allied Tube & Conduit Corp. v. Indian Head, 486 U.S. 492, 496-497, 509-511, 108 S.Ct. 1931, 100 L.Ed.2d 497 (1988) (competitor who "packed" standard-setting meeting to exclude rival's product from model electrical code was not immune).
Microsoft III, however, involves an entirely different standard-setting "process" (if it can be called that), namely, the victory of a single party in a competitive battle between differing rival products. Microsoft had certainly won the standard-setting battle for Intel-compatible PC operating systems hands down. But the battle to set the standard for Windows-compatible Internet browsers had only just begun, with Netscape and Microsoft's Internet Explorer (which the court abbreviates as "IE") as the two leading contenders.
As first, Netscape had a solid and seemingly insurmountable lead, with its 80% market share. By the time the government's lawsuit came to appeal, however, Microsoft had forced a nearly complete reversal of positions, achieving a 75% market share. See Maney, "Lords of the Net duel—but probably not to the death," U.S.A Today, at 18 (Aug. 16, 2001) (reporting browser market shares, as of June 2001, as follows: Microsoft: 75%, Netscape 18%, other 7%). The question before the Microsoft III court was whether Microsoft had achieved this dramatic reversal of fortune by lawful means.
Does the well-recognized economic value of standards generally bear on the answer to that question? Might not either Netscape or Internet Explorer have provided a useful standard for Internet browsers, as long as each was compatible with Microsoft's Windows operating system? If so, does the self-evident value of standards generally really have much to do with this case?
4. The second way in which the Microsoft III court sees the computer industry as special is its speed of development. Although ultimately rejecting Microsoft's argument for a special standard of monopoly power, the court dwells on computer software's status as a "technologically dynamic" market in discussing remedies. It suggests that the game of software is so fast that the referee cannot make timely calls of "foul."
But which way does/should this factor cut? Do fast-moving industries require more probing antitrust scrutiny for fear that anticompetitive conduct might have more devastating effects, or that monopolies might become entrenched more quickly? Is basketball any less in need of referees because it is a faster-paced, more active game than baseball?
5. For our purposes, the primary value of studying Microsoft III is that it introduces us to the application of Sherman Act § 2 to licensing behavior. As discussed above, Section 1 focuses on so-called "concerted conduct," i.e., contracts, combinations and conspiracies in restraint of trade. Because it involves two or more parties, any contractual restraint in a licensing agreement can fall afoul of Section 1 if it is unreasonably anticompetitive, either per se or under the rule of reason. Thus Section 1 is the one most commonly applied to licensing agreements generally.
Unlike Section 1, Sherman Act § 2 applies to unilateral action by a single party (except for conspiracies to monopolize, which are relatively rare). A single party, all by itself, can monopolize or attempt to monopolize a defined market, without any contract or collusion with any other party. Microsoft's acts of tying its IE browser technologically to its operating system, as well as its threats to discontinue offering its Office product for Apple's MAC computers, were unilateral acts of monopolization. The leading case explaining the difference between concerted and unilateral action is Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 770-772, 104 S.Ct. 2731, 81 L.Ed.2d 628 (1984), which held that a corporation cannot conspire with its own division or wholly-owned subsidiary for purposes of Section 1.
6. In one respect, monopolization and attempt-to-monopolize claims under Sherman Act § 2 are easier to prove than Section 1 claims: they do not require proof of concerted action. On the other hand, Section 2 claims are harder to prove in another respect: they require proof of actual or threatened monopoly power. That is, they require proof of (1) a defined product and geographic market and (2) actual or threatened monopoly power in that market. This proof in turn requires analysis of the structure and performance of the defined market, usually including the leading industry participants' market shares and relevant conditions (such as barriers to entry and excess capacity) that affect their market power.
Because Section 2 claims require such detailed market analysis, they virtually never turn on upon simple categorization of behavior, such as horizontal price-fixing or division of markets. In that respect they differ starkly from Section 1 claims. Although courts seldom make the point in haec verba, all claims under Section 2 require application of the rule of reason. There is no such thing as per se liability under Section 2. The rule of reason is appropriate for Section 2 because it covers any of the myriad variant forms that attempts to quash competition on the merits can take. The absence of per-se rules under Section 2, however, has one unfortunate result: Section 2 cases are costly, time-consuming, and expensive and require a great deal of preparation.
7. Section 2 cases have three elements: (1) definition of a market (product or service and geographic), (2) demonstration of monopoly power (or, in attempt cases, near monopoly power) in the defined market, and (3) proof of conduct by the defendant that suppresses competition on the merits and tends to achieve, maintain, entrench or extend a monopoly in the defined market (or, in attempt cases, creates a dangerous probability of success in monopolizing the defined market)..
What was the defined market at issue in Microsoft III? Do you agree with the court that Apple's MAC operating systems and such "middleware" as Java and the Internet browsers at issue in the case were not part of that defined market? What are the best factual arguments for not including them in that market, and on what legal rules for market definition are they based?
8. Under Section 2, "monopoly power" is just market power that tends toward a monopoly. "Market power" is the power to raise prices and/or reduce output for a substantial period of time without rivals stealing market share. Market share is usually a good "proxy" for (substitute for or measurement of) market power, but it is not conclusive. Depending upon other conditions in the defined market, a firm with a high market share may have low market power, or vice versa. See Jay Dratler, Jr., Licensing of Intellectual Property § 5.02[b][i][A] (Law Journal Press, 1994 & Supps.), available on LEXIS: Secondary Legal: Law Journal Press..
In discussing monopoly power, the Microsoft III court says that "monopoly power may be inferred from a firm's possession of a dominant share of a relevant market that is protected by entry barriers." 253 F.3d at 51 (emphasis added). This statement implies that entry barriers are a necessary condition of monopoly power. For it, the Microsoft III court cites a single Ninth Circuit decision, Rebel Oil Co. v. Atlantic Richfield Co., 51 F.3d 1421, 1434 (9th Cir. 1995). There are, however, hundreds of decisions discussing market power and monopoly power; few of them impose an absolute requirement for entry barriers as a condition of proving market or monopoly power. See generally, Jay Dratler, Jr., supra, § 5.02[b][i][A] (discussing generally how market power is assessed). Rather, most of them consider barriers to entry to be just one of many market-structure and market-condition factors to be considered in assessing market power and monopoly power. See id. Yet, on the basis of a single precedent, the Microsoft III court states that barriers to entry are a required element of claims not only for monopolization, but also attempts to monopolize, see 253 F.3d at 81, 82 (attempt claim requires plaintiff "to demonstrate that substantial barriers to entry protect [a] market"), and tying claims under the rule of reason, see 253 F.3d at 95.
Is this judicial innovation in antitrust law holding or dictum? Is the "applications barrier to entry" which the court discusses at length (see also below) a sufficient barrier to entry? If so, does this case suggest that such a barrier is required in all other cases, or just that one is present and sufficient in this case?
9. Now consider the barrier to entry found in this case. The district court found, and the appellate court affirmed, that Microsoft had a 95% share of the defined market. In addition, both courts found a significant barrier to entry—the so-called "applications barrier to entry"—that further reinforced Microsoft's predominant market share.
To understand the applications barrier to entry, note that independent software vendors had developed 70,000 application programs to run on Microsoft's Windows operating systems. See 253 F.3d at 55. Suppose the entire personal-computer industry wanted to abandon Windows and switch to another operating system, called New OS, without losing a thing. All of the independent software vendors that produced the 70,000 application programs would have to rewrite their programs to run on New OS. This job alone would require thousands or tens of thousands of person-years of effort. But that wouldn't be all. In addition, all of the millions of users of the those 70,000 programs would have to learn or re-learn how to use them on New OS, requiring an additional substantial efforts on their part. The effort would make the job of converting the entire Western world from QWERTY to Dvorak keyboards seem like child's play. These "network effects" of Microsoft's monopoly make it much stronger than even the high market share would suggest.
Can you think of any other firm, either in your own experience or in the history of American industry, that enjoyed or enjoys such a monopoly? Monopolies on oil, coal, steel and aluminum rely on such things as the high capital investment required to build a new plant, patent protection, and perhaps the need to find new natural resources. Patents may be difficult to "invent around," but banks have lots of money, and scarcity of natural resources in America (oil, for example), is a relatively recent phenomenon. Did circumventing earlier monopolies ever involve the need to restructure an entire industry, requiring massive effort on the part of tens of thousands of independent vendors and millions of users? Does this analysis suggest that the danger of monopolization is stronger or weaker in the "new" economy than in the "old" economy? Is the supposed uniqueness of the "new economy" reason for prudent hesitation of for more vigorous enforcement of antitrust principles?
10. The "conduct" element of a Section 2 claim is difficult to define or delimit in general. Firms and their employees are often ingenious and creative in devising ways to crush competition on the merits yet to make what they are doing seem innocent or even beneficial. Therefore the law has no specific definition of prohibited conduct, which in any event would quickly be circumvented.
It is black-letter law that conduct unlawful under Section 2 need not be illegal in itself, either under Section 1 or under any other law. See Dratler, supra, § 5.02[b][ii]. If a person with monopoly or near-monopoly power engages in conduct having the purpose or effect of suppressing or avoiding competition on the merits, that is enough to invoke Section 2.
The conduct in which Microsoft engaged to maintain and extend its monopoly was comprehensive indeed. A number of elements are omitted from the edited opinion in Microsoft III above, in order to reduce its length. Here is a more complete list of what the court found unlawful and lawful:
11. As the foregoing list of disapproved conduct suggests, Microsoft's effort to replace Netscape's pioneering Internet browser with its own, using all the considerable power of its operation-system monopoly, was broad and comprehensive. Ultimately, that attempt was successful, causing Microsoft's Internet Explorer nearly to change places in market share with Netscape's pioneering product in less than six years. All this occurred despite the fact that, as the district court's unaltered finding stated, Microsoft's product was "not demonstrably the current ‘best of breed' Web browser, nor [was] it likely to be so at any time in the immediate future." United States v. Microsoft Corp., 87 F. Supp.2d 30, 40, 54 U.S.P.Q.2d (BNA) 1365 (D.D.C. 2000), aff'd in part and rev'd and vacated on other grounds 253 F.3d 34 (D.C. Cir. 2001) (Microsoft III) (per curiam).
Did all this happen by accident, or did Microsoft have a carefully elaborated and assiduously implemented plan? In addressing the substance of the monopolization claims, the Microsoft III court de-emphasized Microsoft's intent. 253 F.3d at 59. This approach is undoubtedly correct as a matter of antitrust liability. Modern antitrust law de-emphasizes intent, because an intent that appears anticompetitive may be nothing more than an intent to compete vigorously but not unlawfully. In a seminal decision involving alleged predatory pricing, the Seventh Circuit put it this way:
This approach makes eminent sense in determining whether conduct is unlawful under the antitrust laws. But what about the remedy? Once unlawfully anticompetitive conduct has been proven, should courts take the violator's intent into account in determining the appropriate remedy? Does intent at least bear on what a crusher of competition may be likely to do next? Should the remedy be the same, for example, if Microsoft had only been liable for one series of exclusive agreements, rather than for the whole laundry list of anticompetitive conduct set forth above?
12. The Microsoft III court also reversed the district court's decision to impose liability based upon Microsoft's entire "course of conduct." 253 F.3d at 78-80. Does that make sense as a matter of liability? The whole cannot be greater than the sum of the parts, can it?
But, again, what about the remedy? Does a massive corporation like Microsoft present a greater danger to competition when proven to have engaged in a comprehensive scheme to crush competition like the one outlined above? Does/should it matter that Microsoft was reported to have $38 billion in cash to spend? See Steve Hamm, Faith Keenan, and Andy Reinhardt, "Making the Tech Slump Pay Off," Business Week at 84 (June 24, 2002). Should courts take such matters into account in determining antitrust remedies?
Should courts' goal be to restore a competitive market, to protect the market from further anticompetitive activity (of the same or a different sort), or just to prohibit what the violator already has done? Cf. Zenith Radio Corp. v. Hazeltine Research, 395 U.S. 100, 132, 89 S.Ct. 1562, 23 L.Ed.2d 129 (1969) ("In exercising its equitable jurisdiction, a federal court has broad power to restrain acts which are of the same type or class as unlawful acts which the court has found to have been committed or whose commission in the future, unless enjoined, may fairly be anticipated from the defendant's conduct in the past.") (citation and internal punctuation omitted).
13. The Microsoft III court did not attempt to prescribe the appropriate remedy. In clear dicta, however, it disfavored the strong medicine—splitting Microsoft up into separate operating-system and applications firms—prescribed by the district court. 253 F.3d at 105-107. It also suggested that the remedy on remand be "tailored to fit the wrong." See id. at 107.
The remedy that ultimately resulted was toothless. In a portion of the opinion not reproduced above, the Microsoft III court disqualified the trial judge for expressing strong opinions on Microsoft's conduct to reporters and others, thereby creating an appearance of bias; it therefore vacated his remedial findings and orders (while retaining his findings and conclusions on other matters) and ordered the case transferred on remand to a different judge. See id. 166-117, 199. (For more on this fascinating aspect of the case, see the second Note for next class session below.)
Having had no previous experience with this mammoth case, the new trial judge pressed the parties to settle, which they ultimately did. The settlement involved a "sin no more" consent decree requiring no divestiture and no mandatory disclosure of Microsoft's source code or other intellectual property. Attorneys general from a number of states that were parties plaintiff objected to the proposed settlement on the ground that Microsoft had already ceased most of the prohibited conduct and that the settlement did little to re-establish competitive conditions in the industry. After months of haggling, however, they declared victory and capitulated, achieving only a few changes, which most industry observers felt would have little effect on real-world competition in the industry.
14. A few years later, a similar factual situation arose. Microsoft integrated multimedia software, called Windows Media Player, into its operating systems in an attempt to compete with independently developed multimedia software. Antitrust authorities in the United States did nothing, but European antitrust authorities included this "bundling" issue in a larger antitrust case against Microsoft, which also addressed the server market. In March 2004, the European Commission fined Microsoft 497 million Euros ($ 612 million) for antitrust violations and ordered it to offer a version of Windows without Windows Media Player. It also ordered Microsoft to disclose interface information on its server software to head off a monopoly in that separate market. See Paul Andrews, "EC ruling's big impact is on bundling," Seattle Times, Business at C1 (March 29, 2004). For a summary of the economic justification for the tougher European approach, see William Bishop and Robert Stillman, "Microsoft got what it deserved in Europe," Financial Times (London), Comment at 21 (March 29, 2004).
Numerically, the fine will hardly sting Microsoft. "The European Union's biggest fine ever against one company, about $612 million, is just slightly more than 1 percent of Microsoft's cash on hand, and is actually less than the $750 million Microsoft voluntarily agreed to pay Time Warner to settle an antitrust action over America Online." Paul Andrews, Seattle Times, supra. In any event, Microsoft appealed the decision to the European Court of Justice, and the appeal is likely to take until the end of the new millennium's first decade. The European remedy, however, might be enforced pending appeal. See Patrick Thibodeau and Carol Sliwa, "Europe Demands Server Jewels From Microsoft; ECs interoperability decree calls for release of interface documentation; Microsoft will appeal," Computerworld, March 29, 2004 at 1.