FALL 2007

Computer Law

Course No.  9200 711 001
Th 6:30 - 9:30 p.m.
Room W-214
Professor Jay Dratler, Jr.
Room 231D (IP Alcove)
(330) 972-7972
dratler@uakron.edu, dratler@neo.rr.com
Copyright © 2000, 2001, 2002, 2004, 2007   Jay Dratler, Jr.   For permission, see CMI.

Microsoft III: Notes and Questions on Tying Claim

1.  Tying is a Sherman Act § 1 offense.  It involves forcing or requiring a customer to buy A in order to get B.  In short, tying involves the forced "bundling" of two products or services.  (Although the terms "tying" and "bundling" are often used interchangeably, "bundling" is a softer, nonlegal term.  It often connotes offering two products or services together without coercion, for convenience or at a discounted price.)

The legal essence of tying is coercion.  The seller says to the buyer, in effect, "you can't have B (the ‘tying' product) unless you also take A (the ‘tied' product)."  In antitrust analysis, market power is generally a proxy for coercion.   If the seller has market power in tying product B, then the tie can coerce the buyer to take the unwanted (tied) product A just to get the desired product B.  If the seller has no market power in B, then the buyer can get B (or an equivalent substitute) from others at comparable price and comparable terms, so the tie does no harm.

When there is market power in the tying product B, the tie causes two types of harm.  First, it coerces the buyer into buying A (instead of another seller's similar product, a different product with similar function, or no such product at all).  Second, it forecloses other sellers of products like A from the market, based not on competition on the merits of product A, but on the tie.  See generally, Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 2, 11-19, 104 S.Ct. 1551, 80 L. Ed. 2d 2 (1984) (discussing history and economics of Section 1 tying offense).

When tying arrangement have the requisite coercive power, they are illegal per se under Sherman Act Section 1, as well as Clayton Act § 3, 15 U.S.C. § 14, which more directly addresses tying.  As the Supreme Court has said, "It is far too late in the history of our antitrust jurisprudence to question the proposition that certain tying arrangements pose an unacceptable risk of stifling competition and therefore are unreasonable ‘per se.'"  Id., 466 U.S. at 9.  The elements of a per se tying offense are: (1) two separate products, (2) a contractual or other tie that bundles them, (3) appreciable economic power (or market power) in the tying product or service, and (4) an effect upon a substantial volume of commerce in the tied product.  See id., 466 U.S. at 20-21, 26; Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451, 461-462, 112 S.Ct. 2072, 119 L.Ed.2d 265 (1992).  When a tie does not satisfy all these requirements, its analysis falls under the rule of reason.  See Jefferson Parish, supra, 466 U.S. at 29-30.

2.   Did Microsoft's tie between its browser (IE) and its Windows operating system satisfy these four requirements of the per se rule?  After the Microsoft III court's analysis of monopoly power in connection with the monopolization claim, is there any doubt that Microsoft's Windows monopoly had coercive power?  Microsoft's Windows operating systems enjoyed a 95% market share, reinforced by a substantial "applications barrier to entry."  Can you think of any more coercive combination?  If a consumer didn't want IE, she could (1) buy an Apple computer at a much higher price and lose easy compatibility with Microsoft's 100-million-user network, or (2) write her own operating system and port to it the 70,000 applications programs (or as many as she wanted to use).  Might those choices be enough to convince the average PC user to say "All right, I guess maybe I don't really like Netscape that much more than IE after all"?  Note that tying claims require only market power, i.e., some substantial power to coerce, not power that rises to the level of a monopoly.

Was there a tie?  The district court had refused to consider the technological bundling (commingling IE code with the operating system's code) as part of the tie.  See 253 F.3d at 85.  Did the court in so doing hope to avoid the difficult question whether the technological tie had technological advantages?  In any event, is there any doubt, given the other listed facts, as well as all of Microsoft's illegal monopolization conduct—most of which was designed to force Internet Explorer down the throats of OEMs and computer users, among others—that there was a tie?  As for an effect on a substantial volume of commerce, don't the facts suggest that?  IE went from a less than 20% market share to 75% in less than six years.  Doesn't that sound substantial?

3.  Thus, the crux of the matter seemed to be whether there were two separate products.  Netscape certainly thought there were.   It had invented the product category of browsers (at least for broad consumer applications), and its separate product, marketed by a separate firm, had been widely used.  Consumers, most likely, thought the same.  Until Microsoft began offering IE, 80% of them installed or saw a separate program from a separate vendor, both called "Netscape," on their desktop.

As the Microsoft III court discusses at some length, the legal test for separate products is separate consumer demand, not their functional relationship.  See Jefferson Parish, supra, 466 U.S. at 19-20 & n.30.   If anesthesia is a service separate from surgery because it is offered by different doctors and billed separately, see id. at 22-23, isn't a browser separate from an operating system on the same rationale?

4.  In refusing to find separate products for the purpose of the per se rule, the D.C. Circuit focuses on the question of the (presumed) efficiencies of software "integration," i.e., in this case offering an operating system and a Web browser as a single package.  What are these efficiencies?  Do they require that the two programs (operating system and browser) be offered by a single vendor, here Microsoft?  If Netscape had maintained its originally dominant 80% market share, couldn't it also have offered the same efficiencies of software "integration"?  In assuming that only Microsoft had the legal right or technical ability to offer these benefits of integration, did the Microsoft III court in effect assume its conclusion?

In speaking of the benefits of software integration, the court seems to focus on so-called "Application Programming Interfaces" or "APIs."   They include function and subroutine calls and other technological "hooks" that independent software programmers must use to develop new programs that work with an existing program.  Both the operating system (in this case Windows) and the browser (in this case either Netscape or Internet Explorer) have APIs.  The court seems to think that having a standard set of APIs would provide benefits both to independent software developers and computer users.  Yet, when Microsoft began its comprehensive scheme to crush Netscape, Netscape had had an 80% market share.  Didn't Netscape then provide a standard set of APIs for independent software vendors and consumers to use in making their programs compatible with the browser?  Does the court note any reason for assuming that Microsoft's own APIs would provide any better standard?  Isn't the underlying issue not whether standards are good, but who will set them?

More fundamentally, the court seems to think that the well-established "separate demand" test for separate products is just a poor reflection of the underlying goal: economic and technological efficiency.  It asserts that "[t]he per se rule's direct consumer demand and indirect industry custom inquiries are, as a general matter, backward-looking and therefore systematically poor proxies for overall efficiency in the presence of new and innovative integration."  253 F.3d at 89.  But who is ultimately supposed to determine what is better for "overall efficiency," judges or consumers?  Isn't the whole premise of antitrust law that consumers, through competition, make this decision?  Isn't that what the cases cited in Note 1 of the frist set of Notes above are all about?

Do/should antitrust judges sit in judgment on the economic and technological efficiencies of various combinations of products and services?  Are they competent to do so?  Or do the antitrust laws simply require them to make sure that consumers get a fair choice and producers a fair chance to compete?  Is using the awesome power of Microsoft's operating-system monopoly to coerce OEMs, independent software vendors and consumers to select Microsoft's browser rather than Netscape's giving them a fair choice?

5.  From a policy perspective, the Microsoft III court saw itself as having to make a choice.  Applying the per se rule in this case, it reasoned, would "chill," if not legally preclude, Microsoft's ability to provide integrated products and expand the capability of its highly popular operating system.

But is that so?  If Microsoft had developed the Web browser first and had made it an integral part of its Windows operating system, would there be any question of its being a "separate product" for tying purposes?  Didn't the very existence of the issue of "separate products" in this case arise out of the fact that Netscape had developed the product first and had marketed it as a separate add-on?  Was the real issue whether tying jurisprudence would ever allow Microsoft to expand its operating systems and make them better by adding new programs with new functionality?  Or was it whether it the law should allow Microsoft to copy the functionality of others' pioneering products and, by exerting the tremendous leverage of its operating-system monopoly, exclude them from the very market that they had pioneered and developed?

6.  The court makes much of the fact that rivals to monopolists like Microsoft will still have the chance to prove ties illegal under the rule of reason, or to prove that a tie was an act or attempt of monopolization.   Proving a case under the rule of reason, however, is far more difficult, time consuming, and expensive than applying a so-called per-se rule, isn't it?

In any event, does the four-part test of Jefferson Parish really establish a "per-se" rule?  Most per-se rules, like the rule against horizontal division of markets among competitors, involve simple categorization of behavior.  In contrast, the so-called per-se rule for tying requires an assessment of the tie's actual ability to coerce (through market power) and its effect on a not insubstantial volume of commerce.  Is that sort of rule based upon simple categorization of conduct?  Or is it a shorthand application of market-based antitrust analysis, suggesting a clear and present danger to competition?

7.  The Microsoft III did not limit its holding to the case before it.   It purported broadly to hold that the rule of reason applies to any tie in which "the tying product is software whose major purpose is to serve as a platform for third-party applications and the tied product is complementary software functionality."  253 F.3d at 95.

Is this statement holding or dictum?  Is it a good rule?  Will it encourage independent firms to create those third-party applications in the form of add-ons to an economically dominant platform?  Will it encourage innovation in the software industry?  Or will it encourage purveyors of dominant platforms to copy the functionality of others' innovative add-ons, thereby discouraging independent innovation?

Would a contrary rule, which prohibited dominant platform purveyors from copying others' new functionality without unbundling it, substantially impair their ability to make truly independent innovations?  Would it hurt consumers?

8.  The very software industry in which Microsoft now weighs so heavily began by unbundling a tie.  In the late 1960s, IBM Corp. had nearly as dominant a share of the market for computer hardware as Microsoft now does of the PC operating-system market.  See Jay Dratler, Jr., "Microsoft as an Antitrust Target: IBM in Software?," 25 Sw. U. L. Rev. 671, 674 (1996) ("Although estimates of [IBM's] market share varied, they generally exceeded seventy percent, depending on the precise market at issue") (footnote omitted).  At that time, IBM also was nearly the sole purveyor of commercially useful software, including operating systems.  The reason was that IBM had developed both the machines and the software to run them.  IBM tied its software to its hardware, refusing to sell the machines separately, thereby creating a huge barrier to entry in the nascent business of writing software.

In 1969, under the threat of antitrust litigation (including tying claims) from the United States government and private parties, IBM voluntarily unbundled its software from its hardware, giving birth to the software industry.  See id., at 730-731.  An important computer industry executive later described this event as "the second milestone in the history of the computer industry, the first being the change from vacuum tube to transistor technology."   Martha Rounds, "IBM Saw ‘Limited' Software Industry; 1969 Prediction: ‘Limited But Increasing Number of Enterprises Engaged in the Development of Computer Programs for Sale,' 9 Software Mag. 37, Mar. 15, 1989, quoting William Norris, president of Control Data Corporation, an IBM competitor.

The software that IBM unbundled voluntarily was entirely of its own invention and design.  IBM had not copied the functionality of anyone else's pioneering innovation, as Microsoft had done in the case of Netscape.  Rather, IBM's employees had written the software that it unbundled in order to permit its machines to do useful work for customers.  Yet by voluntarily unbundling (albeit under the pressure of antitrust litigation), IBM created the software industry, part of which Microsoft dominates today.

Is this tale just a bit of esoteric history, or does it suggest a better way to build innovative industries and avoid undue industrial concentration?  Where would the software industry be today if the courts in the 1960s and 1970s, like the D.C. Circuit, had applied the rule of reason to tying software add-ons to IBM's hardware "platforms"?  Would IBM have voluntarily unbundled software and created a new industry?

Back to Top