abd9.doc

IN THE UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF COLUMBIA

UNITED STATES OF AMERICA
Plaintiff,

vs.

MICROSOFT CORPORATION
Defendant.

Civil Action No. 98-1232 (TPJ)

ATTORNEY

GENERAL ELIOT SPITZER, et al.,
Plaintiffs

vs.

MICROSOFT CORPORATION,
Defendant.

Civil Action No. 98-1233 (TPJ)

MICROSOFT CORPORATION
Counterclaim-Plaintiff

vs.

ELIOIT SPITZER,
Attorney General of the State of New York,
In his official capacity, et al.,
Counterclaim-Defendants.

BRIEF OF PROFESSOR LAWRENCE LESSIG

AS AMICUS CURIAE

February 1, 2000

TABLE OF

CONTENTS

TABLE OF AUTHORITIES......................................................................................................................i

STATEMENT OF INTEREST.................................................................................................................iv

PRELIMINARY STATEMENT.................................................................................................................4

RELEVANT FINDINGS..........................................................................................................................3

TYING....................................................................................................................................................6

ASSUMING MICROSOFT II CONTROLS..............................................................................................8

ASSUMING MICROSOFT II DOES NOT CONTROL..........................................................................18

WHAT IS A SOFTWARE PRODUCT?...................................................................................................19

II.

ARE

WINDOWS

95/98 AND

IE BROWSER FUNCTIONALITY A “SINGLE PRODUCT”?..................23

APPLYING

JEFFERSON PARISHDIRECTLY

...............................................................................23

CONSIDERATIONS IN

APPLYING

JEFFERSON PARISHTO

SOFTWARE

TIES.............................27

EXTENDING

JEFFERSON PARISHTO

SOFTWARE

TIES...............................................................30

Competitive Market Practices [CMP].......................................................................................30

New Product Rationale...............................................................................................................32

Same Product Rationale..............................................................................................................41

A SUMMARY OF THE

TEST

APPLIED

......................................................................................43

CONCLUSION......................................................................................................................................44

TABLE OF

AUTHORITIES

CASES

Advanced Computer Servs.v. MAI Sys. Corp., 845 F. Supp. 356 (E.D. Va. 1994).............................................1

Anderson Foreign Motorsv. New England Toyota Distributors, 475 F. Supp. 973 (D. Mass. 1979)................31

Caldera, Inc.v. Microsoft Corp., 72 F. Supp.2d 1295 (D. Utah 1999).............................................1, 10, 24, 38

Castrignanov. E.R. Squibb & Sons, Inc., 546 A.2d 775 (R.I. 1988)...................................................................1

Data Gen. Corp.v. Grumman Sys. Support Corp., 36 F.3d 1147 (1st Cir. 1994)................................................1

Data General Corp. Antitrust Litigation, 490 F. Supp. 1089 (N.D. Cal. 1980)...............................................31

Eastman Kodak Co.v. Image Technical Servs., Inc., 504 U.S. 451 (1992)..................................................passim

Fortner Enters. v. United States Steel Corp., 394 U.S. 495 (1969).......................................................................6

ILC Peripherals Leasing Corp.v. IBM Corp., 448 F. Supp. 228 (N.D. Cal. 1978)..........................................31

Information Resources, Inc.,v. A.C. Nielsen Co., 615 F. Supp. 125 (N.D. I. 1984)...........................................31

Innovation Data Processingv. IBM Corp., 585 F. Supp. 1470 (D. N.J. 1984).................................................42

Jefferson Parish Hosp. Dist. No. 2v. Hyde, 466 U.S. 2 (1984)....................................................................passim

Kumho Tire Co., Ltd.v. Carmichael, 119 S.Ct. 1167 (1999)...............................................................................1

Maislin Industriesv. Primary Steel, 497 U.S. 116 (1990)....................................................................................7

Montgomery County Assoc. of Realtors.v. Realty Photo Master Corp., 783 F. Supp. 952 (D. Md.
1992)..............................................................................................................................................................31

Multistate Legal Studiesv. Harcourt Brace, 63 F.3d 1540 (10th Cir. 1995)......................................................31

NCAAv. Bd. of Regents, University of Oklahoma, 468 U.S. 85 (1984).............................................................43

Telex Corp.v. IBM, 367 F. Supp. 258 (N.D. Okla. 1973)...............................................................................31

Times-Picayune Publishing Co.v. United States, 345 U.S. 594 (1953)..........................................................5, 19

United Shoe Mach. Corp.v. United States, 258 U.S. 451 (1922).........................................................................5

United Statesv. Jerrold Electronics Corp., 187 F. Supp. 545 (E.D. Pa. 1960), aff’d, 365 U.S. 567
(1961)..........................................................................................................................................23, 24, 31, 37

United Statesv. Microsoft Corp., 1998 WL 614485 (D.D.C. 1998)...........................................................11, 40

United States v. Microsoft Corporation,Civ. No. 94-1564, Response of the United States to Public
Comments, 59 FED REG59426 (1994)...................................................................................................9, 24

United Statesv. Microsoft, 147 F.3d 935 (D.C. Cir. 1998)........................................................................passim

United Statesv. Microsoft, 56 F.3d 1448 (D.C. Cir. 1995)...............................................................................10

OTHER AUTHORITIES

Alan Meese, Antitrust Balancing in a (Near) Coasean World, 95 MICH. L. REV. 111 (1996)........................27

Alan Meese, Monopoly Bundling in Cyberspace: How Many Products Does Microsoft Sell,
ANTITRUST BULLETIN(Spring 1999)....................................................................................................1, 15

Alan Meese, Tying Meets the New Institutional Economics, 146 U. PENN. L. REV. 1 (1997)..........................6

Benjamin Klein and Lester Saft, The Law and Economics of Franchise Tying Contracts, 28 J. L. &
ECON. 345 (1985)...........................................................................................................................................6

CARLISS Y. BALDWIN, KIM B. CLARK, DESIGN RULES: THE POWER OF MODULARITY
(1999).............................................................................................................................................................25

Einer Elhauge, The Court Failed My Test, Washington Times, A19, July 10, 1998......................................14

Frank H. Easterbrook, Allocating Antitrust Decisionmaking Tasks, 76 GEO. L.J. 305 (1987)...........................6

iii

Frank H. Easterbrook, Vertical Arrangements and The Rule of Reason, 53 ANTITRUST L. J. 135
(1984)...............................................................................................................................................................6

HERBERT

HOVENKAMP, FEDERAL

ANTITRUST

POLICY

(1999)..................................................................6

John E. Lopatka & William H. Page, Antitrust on Internet Time, 7 SUP. CT. ECON. REV. 157
(1999).............................................................................................................................................................10

Keith Wollenberg, Note, An Economic Analysis of Tie-in Sales: Re-examining the Leverage Theory,
39 STAN. L. REV. 737 (1987)........................................................................................................................6

Letter of January 21, 1998, from Richard Urowsky to Lawrence Lessig..........................................................9

Linus Torvalds, The Linux Edge, in OPEN SOURCES—VOICES FROM THE OPEN SOURCE
REVOLUTION(Chris DiBona et al. eds., 1999)..........................................................................................25

Michael Woodrow De Vries, United Statesv. Microsoft, 14 BERKELEY

TECH. L.J. 303 (1999)..............9, 29

Microsoft “Halloween Document,” < http://users.andara.com/~sdinn/halloween.html>...............................25

Norman W. Hawker, Consistently Wrong: The Single Product Issue and the Tying Claims Against
Microsoft, 35 CA. W. L. REV. 1 (1998).........................................................................................................9

Rachel V. Leiterman, Comment: Smart Companies, Foolish Choices? Product Designs that Harm
Competitors, 15 SANTA CLARACOMPUTER & HIGH TECH. L.J. 159 (1999)........................................29

RICHARD

A. POSNER

& FRANK

H. EASTERBROOK, ANTITRUST

(2d ed. 1981)........................................6

RICHARD

A. POSNER, ANTITRUST

LAW

(1976).............................................................................................6

ROBERT

BORK, THE

ANTITRUST

PARADOX(1978)......................................................................................6

Sandra Loosemore, with Richard M. Stallman, Roland McGrath, Andrew Oram, and Ulrich
Drepper, The GNU C Library Reference Manual, < http://www.gnu.org/manual/glibc-
2.0.6/html_chapter/libc_1.html>..................................................................................................................25

Thomas Krattenmaker & Steven Salop, Anticompetitive Exclusion: Raising Rivals’ Costs to Achieve
Power Over Price, 96 YALE

L.J. 209 (1986).................................................................................................42

Transcript of Teleconference, United States v. Microsoft, No. 94-1564, January 16, 1998..........................21

Victor H. Kramer, The Supreme Court and Tying Arrangements: Antitrust as History, 69 MINN. L.
REV. 1013 (1985)............................................................................................................................................6

William H. Page & John E. Lopatka, The Dubious Search for “Integration” in the MicrosoftTrial,
31 CONN. L. REV. 1251 (1999)...................................................................................................................10

TREATISES

HERBERT H. HOVENKAMP, 1998 SUPPLEMENT TO AREEDA, ET AL., ANTITRUST LAW
(1998).............................................................................................................................................................32

HERBERT H. HOVENKAMP, 1999 SUPPLEMENT TO AREEDA, ET AL., ANTITRUST LAW
(1999)......................................................................................................................................................passim

IX PHILLIP

E. AREEDA, ANTITRUST

LAW(1991).........................................................................................6

X PHILLIP E. AREEDA, EINER ELHAUGE & HERBERT HOVENKAMP, ANTITRUST LAW
(1996)......................................................................................................................................................passim

STATEMENT OF

INTEREST

At the Court’s request, United Statesv. Microsoft, Nos. 98-1232, 1233, Order, November 19,

1999, I submit this brief addressing the question of how tying law under Section 1 of the Sherman Act

should apply to software products.

I am the Jack N. and Lillian R. Berkman Professor for Entrepreneurial Legal Studies at Harvard

Law School. I have been a law professor since 1991, initially at the University of Chicago Law School. I

have taught antitrust law at the Yale Law School, and courses related to the law of cyberspace at Harvard,

including a seminar devoted to this case. I am presently a Fellow at the Wissenschaftskolleg zu Berlin, in

Berlin, Germany.

I have no financial interest in the Defendant in this case; nor do I have any financial interest in

any competitor of Microsoft. I own shares in a number of mutual funds. The only stock I own is issued by

Marimba Corporation.

The views expressed in this brief are my own. I have drawn upon the advice of a number of col-

leagues and friends in forming these views. These include Harvard Professor Einer Elhauge, MIT Pro-

fessors Harold Abelson and Jerome Saltzer, William & Mary Professor Alan Meese, as well as Ben

Edelman, Renato Mariotti, Bettina Neuefeind, and Chris Parry.

PRELIMINARY STATEMENT

The government alleges that Microsoft has violated Section 1 of the Sherman Act, by “tying” its

“browser product” to various versions of its Windows operating system. Both alleged “products” are soft-

ware products — products that provide computer functionality through software code. The Court has

asked me to address the question of how the law of “tying” applies to an alleged tie of software products.

There is reason to believe that the law in this area is unsettled. While the Supreme Court’s most

extensive examination of the test for “tying” two products, Jefferson Parish Hosp. Dist. No. 2v. Hyde, 466

U.S. 2 (1984), has been reaffirmed by the Court, Eastman Kodak Co.v. Image Technical Servs., Inc., 504

U.S. 451 (1992), Jefferson Parishdealt with a tie between two service products, and Eastman Kodakdealt

with a tie between a physical product and a service product. The Court has not expressly considered the

question of a tie between two software products. Lower courts have, but not without some difficulty.

While courts post-Jefferson Parishhave applied the test to cases raising technological questions,1courts

considering claims that software products are tied have hesitated before doing so directly. See, e.g., Cal-

dera, Inc.v. Microsoft Corp., 72 F. Supp.2d 1295 (D. Utah 1999). The Court of Appeals decision in United

Statesv. Microsoft, 147 F.3d 935 (D.C. Cir. 1998) [Microsoft II], if it can be read to define the law of tying

for the Circuit, is just one, if prominent, manifestation of this uncertainty.

This hesitation comes from a reluctance by courts to investigate the intricacies of software design.

Despite the fact that in other contexts, courts routinely review design decisions, see, e.g., Kumho Tire Co.,

Ltd.v. Carmichael, 119 S.Ct. 1167 (1999) (tire design) and Castrignanov. E.R. Squibb & Sons, Inc., 546

A.2d 775, 781-83 (R.I. 1988) (drug design); see also Alan Meese, Monopoly Bundling in Cyberspace: How

Many Products Does Microsoft Sell, ANTITRUST

BULLETIN106 (Spring 1999), the apparent feeling among

a number of courts and commentators is that code is different: that the task of evaluating design decisions

involved in technological products is uniquely beyond the ken of federal courts.

As a matter of judicial policy, I believe it is a mistake to fetishize code in this way. While I agree

that an overly invasive antitrust policy can stifle innovation, I am not a skeptic of courts’ ability to under-

stand how software functions; nor do I believe that software technology is so benign that it is advisable for

courts to ignore the competitive impact of code-based restraints. Nevertheless, my aim in this brief is to

evaluate how the law of the D.C. Circuit applies to ties of software products, and ultimately, how it might

apply to the government’s Section 1 tying claim in this case.

The answer, in my view, depends upon whether the test articulated by the Court of Appeals in

Microsoft IIis the law applicable to this case. If it is, then it is my view that given the findings of this

Court, the government has not made out a claim of tying under Section 1 of the Sherman Act. If Micro-

soft II does not apply, then it is my view that under the best reading of the Supreme Court’s tying juris-

prudence, the government has made out a claim of tying under Section 1 of the Sherman Act. Whether

or not Microsoft IIapplies to this case is a question I consider but offer no opinion about.2

RELEVANT

FINDINGS

While the first version of Windows 95 offered in the retail channel did not include any browser

technology, every version of Windows 95 and Windows 98 offered in the Original Equipment Manufac-

turer [OEM] channel has had browser technology included in the operating system package. Findings at

¶202. This browser technology has been referred to by Microsoft as “Internet Explorer” [IE], and it has

undergone a number of revisions. Internet Explorer 1.0 was included in the first version of Windows 95

supplied to OEMs in July 1995. IE 2.0 was released with OEM Service Release (“OSR”) 1 of Windows

95 in November 1995. OSR 2, released in August 1996, included IE 3.0. OSR 2.5, released in September

1997, included IE 4.0. And Windows 98 (second edition), released in March 1999, includes IE 5.0. MS

Proposed Findings ¶44 [MS Findings]. At roughly equivalent times, Microsoft also released versions of

its IE technology independently of operating system service releases, both to permit earlier versions of its

operating system to be upgraded to the OSR release level, and to give users of other operating systems

similar IE functionality.

1 See cases cited in United States v. Microsoft, 147 F.3d 935, 960 (D.C. Cir. 1998) (Wald, J, concurring). See also
Data Gen. Corp. v. Grumman Sys. Support Corp., 36 F.3d 1147 (1st Cir. 1994); Advanced Computer Servs. v. MAI Sys.
Corp., 845 F. Supp. 356, 367-70 (E.D. Va. 1994).

2I also do not discuss (1) whether Microsoft has “appreciable economic power” in the market for operating systems;
(2) whether a substantial volume of commerce is affected; (3) whether, assuming the Court finds “two products,”
they are properly considered tied together; (4) whether Microsoft’s alleged tying behavior violates Section 2 of the
Sherman Act; (5) whether the Court should adopt a “rule of reason” approach to this case rather than apply the “per
se” rule.

Though the name of the browser technology has remained constant, the underlying architecture

of IE on the Windows operating system has changed substantially. MS Findings ¶537; ¶161. Versions

1.0 and 2.0 of IE were application programs that were shipped in the package of OEM software denomi-

nated “Windows 95.” They were not architected to “share code” and could be removed from Windows 95

without affecting its functionality. ¶175. Beginning with IE 3.0, that architecture changed. ¶161, 164.

Though the findings do not precisely map the code, it is clear that Microsoft designed these later versions

of its browser technology to be more modular. Code providing particular functionality was isolated into

modules; these modules could then be “called” by other programs, including the operating system. ¶¶162,

163. The result was that more of the browser functionality was made available to programs on the Win-

dows platform, without those programs having to launch a stand-alone browser running in its own appli-

cation window. Cf. ¶281.

This architecture has obvious benefits. In particular, it makes it easier for other programs to use

or rely upon browser related technologies. Other applications’ program developers need not write code to

support those technologies. They can instead simply rely upon Microsoft’s code. ¶44.

At some point, however, the strategy of this modularized design took on a particular form. Rather

than a design that simply isolated browser functionality into one set of modules and operating system

functionality in another, the Court has found that Microsoft began building modules that bundled code

that provided browsing functionality and operating system functionality in the same file. ¶¶160, 161, 164.

The consequence ofand, the Court has found, “primary motivation” for this design was that the removal

of some modules that provided browser functionality meant the removal of operating system functionality

as well. ¶164. This design, while assuring newly exposed Application Program Interfaces [APIs] would

be within the Windows OS, prevented users from disabling code that provided browser functionality

without also disabling the underlying operating system. At that stage, to remove the code providing

browser functionality would be to “break” the operating system.

The evolution of Windows 95/98 can thus be divided analytically into three phases (though the

findings are not clear about whether there was ever a pure stage two). In the first, the functionality associ-

ated with the browser technology is provided by a single application program. In the second, that func-

tionality is provided by a program that has been modularized, such that other programs can call upon

some of the functionality typically associated with browsing technology. And in the third, some of the

modules providing browsing technology have been mixed with modules providing non-browsing func-

tionality. The effect, then, of moving from phase one to phase two is that browsing technology is more

easily available to other applications. The effect of moving from phase two to phase three is that browsing

technology cannot be removed without disabling non-browsing functionality. (In all phases Microsoft

required by contract that both products be present.)

Thus while Microsoft’s consistent objective since July 1995 has been to assure that its Windows

95 and 98 operating systems include browser functionality, its techniques have changed. At first Micro-

soft relied upon contract alone to assure that browser functionality was included with Windows; later Mi-

crosoft also used the design of its code to assure that browser functionality was included with Windows.

With IE 1.0 and 2.0, OEMs were forbidden from removing the IE application from the suite of

programs packaged with the Windows operating system. ¶158. The ultimate consumer was free to re-

move IE 1.0 and 2.0, as well as other aspects of the OS package. But because these versions of IE were

stand-alone applications, removing them removed all the code supporting browser functionality. ¶175.

OEMs were likewise restricted with IE 3.0 and IE 4.0 browser technology. While Windows 95

did provide a technique (the “Add/Remove” utility) for disabling easy access to browser functionality,

¶165, OEMs, this Court has found, were contractually not permitted to use that technique to disable ac-

cess to IE. ¶176. Executing the “Add/Remove” utility would remove some of the files supporting brows-

ing functionality, but it did not remove all such files. ¶165. In particular, it left files that could be used by

other programs to provide browsing functionality. Id.

Beginning with Windows 98, the restriction on OEMs was not merely contractual. Microsoft

eliminated the technique for disabling browsing technology by not including IE technologies within the

list of functionalities that the “Add/Remove” utility could disable. ¶170. And because the code supporting

the OS and browser functionality in Windows 98 was intermixed, it was no longer technologically feasible

to disable IE functionality. ¶164. OEMs were therefore constrained both contractually and technologi-

cally from disabling IE functionality.

The Court has found three different efficiency costs associated with the move to phase three.

First, for consumers who desired no browser functionality, the Court has found that this design will in-

duce “performance degradation, increased risk of incompatibilities, and the introduction of bugs.” ¶173.

Second, for consumers who want the IE browser functionality, the Court has found the design “unjusti-

fiably jeopardized the stability and security of the operating system.” ¶174. And finally, though this cost

was not a necessary consequence of moving to phase three, the Court has found that Microsoft hard

coded its preference for IE browsing functionality in additional ways: While it has become standard on

competing platforms to permit users to select the default browsing technology — whether IE or Netscape

Navigator, or another browser technology — the Court has found that Windows 98 does not fully respect

that choice. At times, despite a user’s choice of another browser as the default, Windows 98 will launch

the IE browsing technology to view certain html documents. ¶171. Microsoft did not offer in its pro-

posed findings of fact a justification for this inconsistency, though a justification is conceivable. Nonethe-

less, this Court has not found any technological justification for this design.

TYING

The Supreme Court has interpreted antitrust law to proscribe certain “tying” contracts since

United Shoe Mach. Corp.v. United States, 258 U.S. 451, 457 (1922). The idea behind this proscription has

been that tying contracts are in their nature anticompetitive. If the consumer wanted the tied product,

there was no reason to tie it; if the consumer did not want the tied product, there was no reason to permit

the seller to “force” the tied product. See Times-Picayune Publishing Co.v. United States, 345 U.S. 594

(1953).

Early tying jurisprudence was quite undiscriminating in its condemnation of tying contracts.3

This was a mistake. As courts and scholars have argued, the instances when tying can be said to cause

3For an account of this early history, see Victor H. Kramer, The Supreme Court and Tying Arrangements: Antitrust as
History, 69 MINN. L. REV. 1013 (1985).