INVENTIONS, INDUSTRY STANDARDS, AND
INTELLECTUAL PROPERTY
By Mark R. Patterson
ABSTRACT
When an industry standard incorporates a patented invention, the
demand for products that comply with the standard has two components.
Some of the demand may be for the inherent technical advantages of the
invention; the patentee is generally entitled to revenues attributable to
this demand. But some of the demand is for the benefits of standardiza-
tion, such as interoperability, and the patentee is not entitled to revenues
attributable to this demand. From this point, the article draws two con-
clusions. First, the amounts to which a patentee is entitled, either in liti-
gation or in licensing negotiations, should be calculated by determining
the portion of demand that is attributable to its invention. In some cases,
there will be evidence from which one can make this determination di-
rectly; in others, there may be no such direct evidence, but it may still be
possible to draw inferences regarding the contributions of the patentee.
Second, because the contributions of the standard itself, like interopera-
bility, are economically distinct, the "owner" of a standard--typically a
standard-setting organization--should be allowed to negotiate license
fees with the patentee of an invention incorporated in the standard.
I. INTRODUCTION
When an industry standard incorporates a patented invention, the legal
challenge is to distinguish several market effects. Some of the demand for
products that comply with the standard may be for the inherent technical
advantages of the invention. A patentee is generally entitled to revenues
 2002 Mark R. Patterson
 Professor of Law, Fordham University School of Law; Visiting Professor of Law,
Suffolk University Law School, 2002-2003. I would like to thank Marcos Vigil of the
Fordham University School of Law Class of 2002 and Jeffrey Libby of the Suffolk Uni-
versity Law School Class of 2003 for excellent research assistance. The preparation of
this essay was supported by a grant from Oracle Corporation and the American National
Standards Institute, for a Symposium on Patent and Related Issues in Standard Setting
held at The George Washington University Law School on April 19, 2002. I would also
like to thank the participants at that symposium for their comments.
BERKELEY TECHNOLOGY LAW JOURNAL
attributable to this demand.1
But some of the demand may also be created
by the adoption of the standard.2
The patentee is not entitled to revenues
attributable to this demand. Although it may be difficult to distinguish
demand for the invention and demand for the standard, it is that distinc-
tion--rather than, say, a focus on "inequity" or some other concept--that
best fosters the incentive-creating goals of patent law while minimizing
negative effects on competition.3
The distinction can be illustrated by a specific example. Rambus, Inc.
recently alleged infringement of its patents,4
which it says are required to
comply with standards for memory technology developed by JEDEC, an
industry standard-setting organization.5
The claims of the Rambus patents
are directed at "inventions designed to increase the operating speed of
memory in computers."6
The JEDEC standards, however, are intended to
provide compatibility or interoperability among the products of different
manufacturers.7
Although some of the demand for the Rambus inventions
1. See Rite-Hite Corp. v. Kelley Co., 56 F.3d 1538, 1569 (Fed. Cir. 1995) (Nies, J.,
dissenting in part) (describing the Federal Circuit's view that award of lost profits is de-
pendent on the consumer demand for patentee's invention).
2. See infra Part I.A.
3. That is not to say that inequitable conduct on the part of a patentee is irrelevant.
Such conduct might also justify placing limits on the patentee's entitlement to returns
from its invention. Indeed, most of the commentary on the issue of patents and standard-
setting addresses the issue in that context. See Janice M. Mueller, Patenting Industry
Standards, 34 J. MARSHALL L. REV. 897 (2001); Daniel I. Prywes, Patent Ambushes and
Licensing in Computer Standard-Setting Groups, ANTITRUST REP. 17 (March 2001). This
essay argues that even when the patentee has engaged in no bad conduct, it should be
limited to returns that are attributable to its invention.
4. See Rambus, Inc. v. Infineon Techs. AG, 155 F. Supp. 2d 668, 671 (E.D. Va.
2001), appeal docketed, 2002 WL 554344 (Fed. Cir.).
5. See JEDEC, at www.jedec.org (last visited July 28, 2002). JEDEC is an acro-
nym for Joint Electron Device Engineering Council.
6. Memorandum Opinion, Rambus, Inc. v. Infineon Techs. AG, No. 3:00cv524,
2001 U.S. Dist. LEXIS 10990, at *4 (E.D. Va. Mar. 15, 2001), appeal docketed, 2002
WL 554344 (Fed. Cir.). One of those claims is presented in the text accompanying note
99 infra.
7. In the specification of one standard at issue in Rambus, JEDEC describes that its
goals include "eliminating misunderstandings between manufacturers and purchasers"
and "facilitating interchangeability and improvement of products." JEDEC Solid State
Technology Association, JEDEC Standard JESD79, Double Data Rate (DDR) SDRAM
Specification, "NOTICE" page (June 2000), at http://www.jedec.org/download/search/-
JESD79R1.pdf. To that end, the standard defines the memory's "features, functionality,
AC and DC parametrics, packages and pin assignments." Id. at i.
See also Daniel I. Prywes, Amicus Curiae Brief of JEDEC Solid State Technol-
ogy Association in Support of Appellees 4, Rambus Inc. v. Infineon Techs. AG, (Feb. 8,
INVENTIONS, INDUSTRY STANDARDS, AND IP
may be due to the increased speed benefits that they provide, much of the
demand is no doubt due to the benefits of interoperability, which are made
possible by standardization, not by the invention. Therefore, although
Rambus is entitled to returns from its technical contribution, it is not enti-
tled to returns from the interoperability provided by the JEDEC stan-
dards.8
This is so even if Rambus did not mislead JEDEC regarding the
existence of its patents, although deception or other inequitable conduct
might be an independent reason to deny a patentee the right to enforce its
invention.9
This distinction between a standard's technical contributions, which
may be attributable to a patented invention, and its other contributions,
such as interoperability, has two important implications. The first implica-
tion is that the patented invention should be treated as only one contributor
to the economic value of the standard. The second implication is that the
other contributions of a standard--like interoperability--should be given
independent legal significance. Even if the creation of interoperability is
not patentable,10
it creates product demand independent of those aspects of
2002) (not filed with Federal Circuit, on file with the Berkeley Technology Law Journal)
(stating that "JEDEC develops technical standards that will permit DRAM products made
by different manufacturers to be interchangeable with each other, and to be inter-operable
on a standardized basis with other computer-system components") [hereinafter Amicus
Curiae Brief].
8. This oversimplifies somewhat. In a case where an invention contributes directly
to the purpose of a standard, the invention may be entitled to returns from the standardi-
zation. See infra Part II.C.
9. See supra note 3. The Federal Trade Commission recently filed suit against
Rambus, alleging that its conduct in the course of JEDEC's standard-setting process was
anticompetitive. See In re Rambus, Inc., No. 9302 (F.T.C. June 18, 2002), available at
http://www.ftc.gov/os/2002/06/rambuscmp.htm. In its enforcement efforts in this area,
the Commission has distinguished the contributions of the invention from the standard,
and it has also relied on whether the patentee disclosed the patent's existence:
If a company misrepresents its patent rights to a standard-setting-
organization, thereby leading the organization to adopt a particular
standard that may infringe on the company's patent rights, the com-
pany's later efforts to take advantage of market power resulting from
the standard, rather than from some inherent value of the patent, consti-
tutes a violation of [FTC Act § 5].
In re Dell Computer Corp. 60 Fed. Reg. 57,870 (F.T.C. Nov. 22, 1995) (consent agree-
ment with analysis to aid public comment), 1995 WL 688383. See also Ted Bridis, FTC
to Probe Sun, Rambus On Disclosure, WALL ST. J., Sept. 11, 2001, at A16.
10. Although standards possess the two main economic characteristics of intellec-
tual property, in that they are expensive to create but easy to copy, interoperability stan-
dards are not likely to meet the legal requirements for patentability, because they will not
often be nonobvious. In addition, standards possess other characteristics that may make it
undesirable to give them all the legal protections of intellectual property (particularly the
BERKELEY TECHNOLOGY LAW JOURNAL
the products that comply with the standard. Consequently, this essay ar-
gues that the "owner" of a standard--typically a standard-setting organiza-
tion--should be allowed to negotiate license fees with the patentee of an
invention incorporated in the standard. That is, the organization should be
permitted to negotiate on behalf of its members' collective interests, just
as a patentee can negotiate on behalf of its licensees' collective interests.11
The essay first contends that standards, like inventions, provide inde-
pendent contributions to the demand for the products that conform to
them. Part I supports this argument using the doctrinal framework of pat-
ent law, and Part II illustrates how to measure the standard's contribution
to demand. Part III then presents the argument for allowing the "owner" of
a standard, which is typically a standard-setting organization, to negotiate
with patentees on behalf of the standard. Part IV summarizes the implica-
tions of these proposals.
II. THE CURRENT LEGAL FRAMEWORK
Three bodies of rules currently govern the relationship between pat-
ented inventions and industry standards. First, patent law establishes the
rules that determine damages in the case of infringement. The rules not
only govern in the infringement context, but also influence the incentives
of patentees and potential licensees in the standard-setting process. Sec-
ond, the rules of standard-setting bodies can affect, through contract or
fraud-based theories,12
the licensing fees that their members may charge.13
right to exclude, which raises anticompetitive concerns). See, e.g., Allied Tube & Conduit
Corp. v. Indian Head, Inc., 486 U.S. 492, 500-01 (1988); Am. Soc'y of Mech. Eng'rs,
Inc. v. Hydrolevel Corp., 456 U.S. 556, 571 (1982). Nevertheless, a recognition that stan-
dards share much of the distinctive characteristics of intellectual property can help clarify
the issues at stake.
11. See infra Part III.
12. One contract-based approach is that participation in the standard-setting process
resulted in a contract between the patentee and the standard-setting organization, and that
those who use the standard are third-party beneficiaries of that contract. Alternatively, the
patentee could be viewed as having engaged in a fraud on the standard-setting organiza-
tion. See Rambus, Inc. v. Infineon Techs. AG, 155 F. Supp. 2d 668 (E.D. Va. 2001), ap-
peal docketed, 2002 WL 554344 (Fed. Cir.) (presenting both theories).
13. For example, in a recent infringement case the defendant argued that to the ex-
tent that the patents at issue were adopted as industry standards, the patentee "had an ob-
ligation to license its standards patents on reasonable, nondiscriminatory terms." Lucent
Techs., Inc. v. Newbridge Networks Corp., 168 F. Supp. 2d 181, 264 (D. Del. 2001).
The federal government also requires that when federal agencies use private
standards, the standard-setting process should comply with certain procedures, including
a requirement that licenses be made royalty-free or at a "reasonable" royalty. Federal
Participation in the Development and Use of Voluntary Consensus Standards and in Con-
INVENTIONS, INDUSTRY STANDARDS, AND IP
Third, antitrust law places restrictions on collective actions that control
whether standard-setting organizations, on behalf of their members, can
negotiate with patentees.
A. Patent Law
Patent law has not addressed the precise problem posed by the incor-
poration of inventions in standards. This problem can be characterized as
the difficulty of allocating entitlements to the benefits of an invention,
when some of those benefits stem not from the contribution of the inventor
but from the actions of others. That is, even when infringement is clear, it
may not be accurate to characterize all the patentee's lost profits as due to
the infringement, or to use the demand for the invention as a determinant
of the patentee's entitlement to a royalty. Patent law has, however, ad-
dressed similar problems in the rules for calculating infringement damages
and in the rules governing patent misuse.
1. Infringement Damages
The law applicable in calculating damages for infringement requires
consideration of the source of the demand for the patentee's product. In
the Federal Circuit, "lost profit awards have been dependent, inter alia, on
proof that consumer demand for the patentee's goods is created by the ad-
vantages of the patented invention."14
This rule is usually applied where
the allegedly infringing product includes not only the invention but also
some other desirable feature.15
The standard-setting context is different, and more difficult. If the in-
vention is incorporated in the standard, the two are not distinct, as an in-
vention and some other feature of a product can be. Indeed, in some in-
stances the invention and the standard are identical.16
Hence, in the stan-
formity Assessment Activities, 63 Fed. Reg. 8546, 8554 (Feb. 19, 1998). The federal
policy requires that the agencies "must use voluntary consensus standards in lieu of gov-
ernment-unique standards . . . except where inconsistent with law or otherwise impracti-
cal." Id. "Voluntary consensus standards" are defined by several characteristics, among
which the most relevant here is that they include "provisions requiring that owners of
relevant intellectual property have agreed to make that intellectual property available on a
nondiscriminatory, royalty-free or reasonable royalty basis to all interested parties." Id.
14. Rite-Hite Corp. v. Kelley Co., 56 F.3d 1538, 1569 (Fed. Cir. 1995) (Nies, J., dis-
senting in part).
15. See Georgia-Pac. Corp. v. United States Plywood Corp., 318 F. Supp. 1116
(S.D.N.Y. 1970) (distinguishing between advantages of underlying patent and those of
infringer's additions when calculating damages award).
16. See In re Dell Computer Corp., 121 F.T.C. 616, 617 (1996) (noting that Dell's
patent covers a mechanical slot (for computer buses) which had been adopted as a stan-
dard).
BERKELEY TECHNOLOGY LAW JOURNAL
dard-setting context, there is in fact demand for the invention, even if the
demand does not exist because of the invention.
Nevertheless, the same principle is applicable in both contexts: dam-
ages should derive from the technical advance made by the invention, as
reflected in the patent claims. This seems the fairest reading of the Federal
Circuit's statement quoted above, since "the advantages of the patented
invention" are presumably those that are inherent in the invention, not
those associated with its incorporation in an industry standard.17
The rules
applicable in calculating a reasonable royalty (as distinguished from lost
profits) also reflect this interpretation, where among the factors to be con-
sidered are the invention's "usefulness and commercial value as shown by
its advantages over other things."18
Although the "advantages" of an in-
vention could include not just its inherent technical advantages but also its
advantages in appropriating value from the efforts of others,19
that inter-
pretation would be inconsistent with other inquiries in patent law.20
Most basically, "[t]o recover lost profits damages for patent infringe-
ment, the patent owner must show that it would have received the addi-
tional profits `but for' the infringement."21
An example illustrates this
principle. Take a case involving sales by a number of infringing sellers of
products that comply with an industry standard and incorporate the pat-
17. That is so unless the invention contributes to the purpose of the standard. See in-
fra Part II.C.
18. Panduit Corp. v. Stahlin Bros. Fibre Works, Inc., 575 F.2d 1152, 1159 (6th Cir.
1978) (quoting United States Frumentum Co. v. Lauhoff, 216 F. 610, 617 (6th Cir.
1914)).
19. See Georgia-Pac. Corp., 318 F. Supp. at 1120 (listing among the factors to be
used in calculating a reasonable royalty "[t]he portion of the realizable profit that should
be credited to the invention as distinguished from non-patented elements, the manufactur-
ing process, business risks, or significant features or improvements added by the in-
fringer").
20. For example, in determining patentability, one of the factors that can be consid-
ered in the obviousness inquiry is the commercial success of the invention. The rationale
for the use of the commercial success test is that an invention that meets with such suc-
cess was presumably nonobvious, else the commercial need would previously have been
met. The law recognizes, though, that this test is effective only if the commercial success
of the product is due to the invention embodied therein and not to some other factor.
Thus, the Federal Circuit has said that commercial success "must be shown to have in
some way been due to the nature of the claimed invention, as opposed to other economic
and commercial factors unrelated to the technical quality of the patented subject matter."
Cable Elec. Prods., Inc. v. Genmark, Inc., 770 F.2d 1015, 1027 (Fed. Cir. 1985), over-
ruled on other grounds by Midwest Indus., Inc. v. Karavan Trailers, Inc., 175 F.3d 1356
(Fed. Cir. 1999) (en banc).
21. King Instruments Corp. v. Perego, 65 F.3d 941, 952 (Fed. Cir. 1995).
INVENTIONS, INDUSTRY STANDARDS, AND IP
entee's invention. Suppose also that some portion of the demand for the
infringers' products is created by the interoperability made possible by the
products' compliance with the standard, and that compliance with the
standard requires the sellers to incorporate the invention. In this case, for
the patentee to show that it would have made the sales but for the in-
fringement, it would have to show that all the infringing sellers would
have sold the same products even if they were paying licensing fees, in-
stead of turning to a noninfringing alternative. If some of the sellers de-
clined to license the invention, and thus declined to comply with the stan-
dard, the advantages of interoperability would be lessened, and sales might
decrease even for those sellers that were willing to license.22
Other Federal Circuit statements, though not always very clear, also
support a focus on the technical aspects of the invention. For example, in
Slimfold Manufacturing Co. v. Kinkead Industries, Inc.,23
the Federal Cir-
cuit affirmed the district court's conclusion that there were acceptable
noninfringing alternatives24
because the patentee "failed to show that buy-
ers of bi-fold metal doors specifically want a door having the advantages
of the Ford patent"25
The reference to the "patent," rather than the inven-
tion, suggests that the relevant advantages are those made possible by the
claims of the patent, not extrinsic factors such as the invention's incorpo-
ration in a standard.
The district courts have also made statements that appear to distinguish
the patentee's contribution from other factors. For example, in Polaroid
Corp. v. Eastman Kodak Co.,26
the court addressed the proof required of
the plaintiff:
The patent holder must show that it had the marketing capability
to make the sales. Typically this requires proof of factors such as
an adequate distribution system and sales personnel. This factual
inquiry is consistent with the role of marketing in markets where
demand is relatively inelastic or dependent on variables outside
22. In theory, if the patentee had licensed some sellers, it might be able to recover
damages for lost profits based on the sales of others, because the limited licensing might
be sufficient to create the additional interoperability-based demand.
23. 932 F.2d 1453 (Fed. Cir. 1991).
24. The existence of noninfringing alternatives is part of the damages inquiry be-
cause the absence of such alternatives "tends to prove that the patentee would not have
lost the sales to a noninfringing third party rather than to the infringer." Rite-Hite Corp. v.
Kelley Co., 56 F.3d 1538, 1548 (Fed. Cir. 1995) (en banc).
25. Slimfold, 932 F.2d at 1458.
26. No. 76-1634-MA, 1990 WL 324105 (D. Mass. Oct. 12, 1990).
BERKELEY TECHNOLOGY LAW JOURNAL
the seller's control, such as a rate of new construction or popula-
tion growth.27
The reference to factors "outside the seller's control" suggests that the
patentee is not entitled to returns from such factors, including industry
standardization efforts.
One might argue that even if some of the demand for the patented in-
vention derives from its standardization, rather than its technical merits,
the functional relationship between the two entitles the patentee returns
from both. After all, the Federal Circuit has stated that the "entire market
rule" is appropriate when "the patented and unpatented components to-
gether are `analogous to components of a single assembly,' `parts of a
complete machine,' or `constitute a functional unit,' but not where the un-
patented components `have essentially no functional relationship to the
patented invention and . . . may have been sold with an infringing device
only as a matter of convenience or business advantage.'"28
This principle
should only apply, though, when the demand for the product derives from
the larger "functional relationship." In the standardization context, even
where there is a functional relationship between the invention and the
standard, the demand for the standard may not relate to the functional fac-
tors at all, but derives from the basic fact of standardization.29
2. Patent Misuse
Another way to approach this issue is to consider it as analogous to the
leveraging problem, as reflected in tying law and in the law of patent mis-
use. In a typical patent-leveraging context, the problem is distinguishing a
patentee's legitimate return on its invention in a leveraging market from
illegitimate efforts to extend its power to a related, leveraged market.30
The Federal Circuit has stated that the patentee's power is limited by the
scope of its patent claims, and that the misuse inquiry turns on whether the
27. Id. at *15 (emphasis added) (citations omitted).
28. Tec Air, Inc. v. Denso Mfg. Mich., Inc., 192 F.3d 1353, 1362 (Fed. Cir. 1999)
(quoting Rite-Hite, 56 F.3d at 1550).
29. Part II.C infra discusses those cases in which standardization is made possible
by the technical advances of the invention.
30. As the Supreme Court said, it "has held many times that power gained through
some natural and legal advantage such as a patent, copyright, or business acumen can
give rise to liability if `a seller exploits his dominant position in one market to expand his
empire into the next.'" Eastman Kodak Co. v. Image Technical Servs., Inc., 504 U.S.
451, 479 n.29 (1992) (citations omitted).
INVENTIONS, INDUSTRY STANDARDS, AND IP
patentee seeks to extend its power further.31
The issue can be clarified by
recognizing that the invention itself is generally distinct from the products
sold in both the leveraging and the leveraged markets, and that the legiti-
macy of the patentee's leveraging depends whether the invention is of
value in only one of those markets, or in both.32
In the standard-setting context, the market relationships are analogous.
Again, there are three economically distinguishable "products": the inven-
tion (and the patentees' right to exclude others from it); the invention's
technical benefits; and the interoperability benefits of the standard.33
Just
as a patentee in the leveraging context may seek to use its power over an
invention with value only in the leveraging market to reap profits in the
leveraged market, a patentee in the standards context may seek to use its
power over an invention that provides only particular technical benefits to
reap profits from interoperability. But the market effects are more difficult
to distinguish, for two reasons. First, whereas in the leveraging context
two of the products--the leveraging and leveraged products--are typically
sold in active markets, in the standard-setting context the standardized
product--incorporating both the invention's technical benefits and the
standard's interoperability benefits--may be the only product actually
sold. Second, whereas in the leveraging context the invention may not be
used in the leveraged product market, in the standard-setting context the
invention may be identical to the standard, though the demand for the
benefits provided by the two are distinct. Although these differences may
make the analysis more difficult, it still may be possible to determine
whether a patentee is extending power beyond the scope of its patent
claims.34
31. B. Braun Med., Inc. v. Abbott Labs., 124 F.3d 1419, 1426-27 (Fed. Cir. 1997);
Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 708 (Fed. Cir. 1992).
32. In a common example, a patentee claims that its patented invention entitles it to
supracompetitive profits not only in a market for parts that incorporates its invention, but
also in a market for service, which is not dependent on the invention. See Mark R. Pat-
terson, When Is Property Intellectual? The Leveraging Problem, 73 S. CAL. L. REV. 1133,
1141-42 (2000); see also Julie E. Cohen & Mark Lemley, Patent Scope and Innovation in
the Software Industry, 89 CAL. L. REV. 1, 25-26 (2001) (relying on same distinction).
33. For a recent discussion distinguishing these three in the context of expert testi-
mony regarding infringement, see Lucent Technologies., Inc. v. Newbridge Networks
Corp., 168 F. Supp. 2d 181, 223-25 (D. Del. 2001).
34. See infra Part II.
BERKELEY TECHNOLOGY LAW JOURNAL
Even if a standard were viewed as formally within the scope of a pat-
ent incorporated in it, the patentee's power should be limited.35
For exam-
ple, one might consider whether the reverse doctrine of equivalents should
bar extension of a patent covering a technical innovation to the use of that
innovation in an industry standard. As Merges and Nelson have argued,
the reverse doctrine of equivalents can be viewed as a means of avoiding
holdups that could deter innovation.36
The application of the reverse doc-
trine of equivalents in this context would not be a traditional one, in that it
would not eliminate infringement liability entirely, but would only limit
that liability to the returns on the patent's technical contribution.
The result, as Merges and Nelson also suggest, would resemble a
compulsory licensing scheme.37
Although U.S. patent law does not explic-
itly provide for compulsory licensing, a similar outcome would be
achieved by limiting damages and using the reverse doctrine of equiva-
lents to eliminate injunctive relief. Such a result does not seem unfair
since in most of the cases in which patentees have sought to enforce pat-
ents that have been incorporated in standards, the patentees have been
members of the standard-setting organizations. As a result, they would in
most cases have agreed to license on reasonable and nondiscriminatory
terms,38
and a refusal to license would not be at issue.
B. Standard-Setting Organizations' Rules
Some standard-setting bodies are moving away from policies requiring
their members to agree to royalty-free licensing and are instead adopting
policies requiring "reasonable and nondiscriminatory" ("RAND") licens-
ing.39
Although the "nondiscriminatory" element of these policies is
straightforward, since it requires that patentees license to all on the same
terms, the definition of "reasonable" is not so clear. Moreover, the stan-
35. See generally Robert P. Merges & Richard R. Nelson, On the Complex Econom-
ics of Patent Scope, 90 COLUM. L. REV. 839 (1990).
36. Id. at 865-67. Merges and Nelson argue that reverse doctrine of equivalents can
be used to free a subservient patent holder from being blocked by the original patentee.
37. Id. at 866 n.118.
38. See infra Part I.B.
39. This shift is not universal, however. The World Wide Web consortium recently
abandoned a proposal for a RAND policy, in favor of a royalty-free one. See Margaret
Kane, W3C retreats from royalty policy, NEWS.COM (Feb. 26, 2002), at http://news.-
com.com/2100-1023-845023.html; W3C, Patent Policy Working Group, Royalty-Free
Patent Policy (Feb. 26, 2002), at http://www.w3.org/TR/2002/WD-patent-policy-
20020226/.
INVENTIONS, INDUSTRY STANDARDS, AND IP
dard-setting bodies themselves make little effort to define the term.40
In-
deed, the American National Standards Institute ("ANSI") says that de-
termination of reasonableness is not a proper subject for the standard-
setting process.41
As a result, it is unclear whether the standard-setting
40. See, e.g., INTERNET ENGINEERING TASK FORCE, REQUEST FOR COMMENTS
(RFC) 2026 (INTERNET STANDARDS PROCESS) § 10.3.3, at http://www.ietf.org/IESG/-
Section10.txt ("The IESG will not make any explicit determination that the assurance of
reasonable and nondiscriminatory terms for the use of a technology has been fulfilled in
practice. It will instead use the normal requirements for the advancement of Internet Stan-
dards to verify that the terms for use are reasonable.") (last visited July 28, 2002);
INTERNATIONAL TELECOMMUNICATION UNION, TELECOMMUNICATION
STANDARDIZATION BUREAU, STATEMENT ON TSB PATENT POLICY, PATENT DEC-
LARATION FORMS AND GUIDELINES FOR THE IMPLEMENTATION OF THE TSB PATENT POL-
ICY Annex 1 § 2.2 (Feb. 2, 2000) (Statement on TSB Patent Policy), at http://www.-
itu.int/itudoc/itu-t/circ/circ5/245_ww9.doc ("The patent holder is not prepared to waive
his rights but would be willing to negotiate licenses with other parties on a nondiscrimi-
natory basis on reasonable terms and conditions. Such negotiations are left to the parties
concerned and are performed outside the ITU-T.") [hereinafter ITU TSB PATENT
POLICY]; AMERICAN NATIONAL STANDARDS INSTITUTE, PROCEDURES FOR THE
DEVELOPMENT AND COORDINATION OF AMERICAN NATIONAL STANDARDS § 1.2.11.1
(Jan. 2002), at http://www.ansi.org/public/library/std_proc/anspro/due_proc1.html (last
modified Mar. 2002). The same is true of some commentary on this issue. See Mueller,
supra note 3, at 933-34 (stating that "some competent authority must set a licensing fee
structure that will determine the patentee's remuneration," but not proposing any criteria
for that determination other than recommending that industry experts instead of the gov-
ernment should create the licensing fee schedules).
41. ANSI apparently believes that reasonableness is a subject only for the parties to
a license:
It should be reiterated, however, that the determination of specific li-
cense terms and conditions, and the evaluation of whether such license
terms and conditions are reasonable and demonstrably free of unfair
discrimination, are not matters that are properly the subject of discus-
sion or debate at a [standards] development meeting. Such matters
should be determined only by the prospective parties to each license or,
if necessary, by an appeal challenging whether compliance with the
Patent Policy has been achieved.
American National Standards Institute, Guidelines for Implementation of the ANSI Pat-
ent Policy: An Aid to More Efficient and Effective Standards Development In Fields
That May Involve Patented Technology § III.B, at http://www.ansi.org/public/-
library/guides/ppguide.html [hereinafter ANSI Guidelines].
Although the ANSI GUIDELINES do not provide further rationale for this approach,
the organization may be concerned about antitrust issues. Cf. infra Part I.C; Sony Elecs.,
Inc. v. Soundview Techs., Inc., 157 F. Supp. 2d 190, 193 (D. Conn. 2001) (noting pat-
entee's allegations that a standards organization and its members conspired "to `avoid
unreasonable royalty demands'" on the members). On the other hand, ANSI will appar-
ently consider the reasonableness of terms and conditions outside the initial standard-
BERKELEY TECHNOLOGY LAW JOURNAL
bodies would approve (or mandate) an approach like that described in the
previous section.
Where an attempt is made to define "reasonable," however, the focus
is on factors that contribute to the patentee's technical contribution, not to
those related to standardization.42
For example, the International Tele-
communication Union, which has a RAND policy, states that "in order to
define what is fair and `reasonable' in a given case, one needs to know
development and manufacturing costs, profits, etc."43
The "development
and manufacturing costs" of an invention would presumably be unrelated
to its incorporation in an industry standard. Furthermore, although the
"profits" from an invention could be greater if the patentee were viewed as
entitled to returns from standardization, the mention of profits in the con-
text of costs suggests that it refers to pre-standardization profits, which
would reduce the post-standardization royalty required to provide the pat-
entee with a reasonable rate of return.
Thus, both patent law and the policies of standard-setting organiza-
tions can be seen as limiting patentees' entitlements to their specific tech-
nical contributions to products that incorporate their inventions. Neither
set of principles supports allowing patentees to reap profits from the con-
tributions of others to the standard and to the distinct benefits, such as in-
teroperability, that standardization provides. But neither the law nor the
organizational policies speaks clearly on this issue, so it is not surprising
that licensing negotiations are fraught with uncertainty.
C. Blocking Patents and Antitrust Law
The combination of a patented invention and a standard incorporating
the invention resembles two "blocking" patents.44
The typical blocking
situation arises when a new patent offers a narrow improvement on an ex-
setting process, stating that a decision on reasonableness "is the exclusive province of the
Board of Standards Review (or, on appeal, the ANSI Appeals Board)." ANSI GUIDE-
LINES § II.
42. A recent article made a somewhat similar point. See Prywes, supra note 3, at 26
("The determination of a reasonable royalty must take into account the benefits of stan-
dardization to the patent-holder at the time a standard is being developed.") The focus of
Prywes, however, is on supply, rather than demand. "A patent-holder . . . usually stands
to benefit from the adoption of its design as an industry standard, because that status will
promote greater production and, in turn, lower costs for items needed for the patent-
holder's own products." Id. A footnote that Prywes adds to this statement refers to net-
work effects, which are a demand-side phenomenon, but Prywes discusses only costs. Id.
at 28 n.31.
43. ITU TSB PATENT POLICY, supra note 40, at app. 1, § 2.3.
44. See Merges & Nelson, supra note 36, at 860.
INVENTIONS, INDUSTRY STANDARDS, AND IP
isting, broader patent.45
In that situation, licenses under both patents are
needed to produce the improved product.46
Thus, the owner of either pat-
ent can block the production of the product. However, the patentees will
often negotiate a cross-licensing or pooling agreement, since it is in both
their interests to profit from their innovative contributions.47
The patent-standard combination is similar. To produce a product that
embodies both a patented invention and standardization, a seller must both
obtain a license under the patent and comply with an agreed-upon stan-
dard. Just as with the blocking patents, two independent factors contribute
to the value of the ultimate product.48
The patentee, of course, provides the
innovative contribution for the invention, and the standard-setting organi-
zation and its members invest considerable effort in creating the standard.
(Usually, the standard itself is not patented,49
but if it were, cross-licensing
could be used in these circumstances as well.)
As a result, the negotiation process between the patentee and the stan-
dard setting organization can be difficult. On one side is the patentee, and
on the other are the sellers who seek to offer products that comply with the
standard. One way to solve the problem would be to allow the standard-
setting organization to negotiate a licensing arrangement with the patent
holder, just as the holder of one blocking patent generally negotiates with
the other. There are antitrust risks to that approach, though, because the
standard-setting organization could be seen as a vehicle for price-fixing
collusion by its members, as at least one case has suggested.50
Antitrust
liability for such negotiations would be unfortunate, because those nego-
tiations could provide all of the pro-competitive benefits cited by the anti-
trust agencies in their Antitrust Guidelines for the Licensing of Intellectual
Property, "by integrating complementary technologies, reducing transac-
tion costs, clearing blocking positions, and avoiding costly infringement
litigation."51
45. See U.S. DEPARTMENT OF JUSTICE AND FEDERAL TRADE COMMISSION, ANTI-
TRUST GUIDELINES FOR THE LICENSING OF INTELLECTUAL PROPERTY § 2.3 (1995), avail-
able at http://www.usdoj.gov/atr/public/guidelines/ipguide.htm [hereinafter DOJ-FTC IP
GUIDELINES].
46. Id.
47. See generally id. § 5.5.
48. That is, they potentially do so, if there are independent benefits from the inven-
tion and the standard.
49. See supra note 10 and accompanying text (on standard that is not patentable).
50. See Sony Elecs., Inc. v. Soundview Techs., Inc., 157 F. Supp. 2d 180, 181 (D.
Conn. 2001).
51. DOJ-FTC IP GUIDELINES, supra note 46, § 5.5.
BERKELEY TECHNOLOGY LAW JOURNAL
Antitrust law can and should distinguish, however, between collective
action that facilitates negotiation in the patent-standard context and anti-
competitive collusion among potential licensees. As the Guidelines make
clear, it is important to distinguish the market for technology (the stan-
dard) from the market for goods (the products that comply with the stan-
dard).52
The Guidelines state that "[t]he Agencies will not require the
owner of intellectual property to create competition in its own technol-
ogy,"53
and the same rule should apply to standards, even if the standards
themselves are not patented, as discussed below.
Ultimately, then, the relevant background rules support treating inven-
tions and standards separately for the purposes both of calculating in-
fringement damages and of negotiating between patentees and standard-
setters. This basic principle does not, however, establish how one should
determine specific entitlements in particular instances. The following sec-
tion addresses that question.
III. DETERMINING LICENSE FEES FOR PATENTED
INVENTIONS IN STANDARDS
The analysis presented in this section provides an approach to deter-
mine license fees. That approach considers whether demand is due to the
technical advance of the patentee's invention or is instead the product of
the adoption of a standard that happens to incorporate the invention.54
This
section first considers this issue from a general perspective and then con-
siders three special cases: inventions that contribute directly to the goal of
the standard that incorporates them; de facto standards; and "standards"
created by "lock-in" (i.e., the existence of investments that would make it
costly to switch to an alternative).
As an aid in illustrating the first part of the analysis, consider the fol-
lowing table, the cells of which represent products that might be sold.
Each product may incorporate a patented invention, comply with a stan-
dard, or neither, or both.
52. Id. § 3.2.
53. Id. § 3.1.
54. In the ongoing Rambus litigation, the standard-setting organization, JEDEC, said
in an amicus brief that a previous court had stated that "JEDEC's adoption of an item as
an industry standard can vastly increase demand for that item, in some cases into a
`multi-billion dollar market.'" Amicus Curiae Brief , supra note 7, at 11 (quoting Wang
Labs., Inc. v. Mitsubishi Elecs. Am., Inc., 103 F.3d 1571, 1579) (Fed. Cir. 1997)). In fact,
Wang did not say that the size of the market in that case was a result of the adoption of a
standard, but the basic point is no doubt a valid one.
INVENTIONS, INDUSTRY STANDARDS, AND IP
Does not incor-
porate a patented
invention
Incorporates
patented inven-
tion P
Incorporates al-
ternative patented
invention Pa
Does not comply with
standard
A: demand: $6
cost: $5
B: demand: $7
cost: $5
C: demand: $9
cost: $6
Complies with chosen
standard S
D: demand: $9
cost: $8
E: demand: $10
cost: $6
F: demand: $9
cost: $6
Complies with alternative,
unchosen standard Sa
G: demand: $8
cost: $7
For each cell's hypothetical product, the table provides a hypothetical
cost of manufacture and a hypothetical value for buyer demand (or will-
ingness to pay).55
Although the numbers chosen are hypothetical, the rela-
tionships among them are intended to be plausible.
One who knew all of this information could draw conclusions about
the value of patented invention P. For example, the manufacturing cost
advantage provided by invention P in complying with the standard (i.e.,
E's cost, without license, of $6, as compared to D's cost of $8) shows that
the invention provides value in itself. The greater demand for a standard-
ized product that incorporates invention P, as distinguished from a product
that does not incorporate any invention or incorporates alternative inven-
tion Pa (E's demand of $10, as compared to D's and F's demands of $9),
also shows that invention P has independent value.
Unfortunately, this information is rarely available. For example, when
the invention and the standard are coextensive, the possibility of comply-
ing with the standard without incorporating the invention does not exist.56
Nevertheless, these hypothetical products must be considered to distin-
guish desirable licensing of technical advances from opportunistic exploi-
tation of standardization. In many instances, one can infer approximations
55. In a real market, of course, one could rarely characterize demand as a single
value, because demand often differs for different consumers. This approach is a helpful
simplification, though, and it may even be plausible for some markets. If, for example,
the product at issue is incorporated as part of a larger product, and if it is a small part of
that larger product, the value of the incorporated product may be consistent across buy-
ers. This might be the case, for example, of a product like the VL Bus, which was at issue
in the Dell case, in that the bus was a small part of a larger computer product. In re Dell
Computer Corp., 121 F.T.C. 616, 617 (1996).
56. Even in some such cases, though, a variation on this analysis can be helpful, as
discussed later in this section.
BERKELEY TECHNOLOGY LAW JOURNAL
of the values in the table, as is discussed below. Even more importantly,
the principles of the analysis provide an appropriate background for li-
censing negotiations, even where the data for precise application of the
principles are unavailable.57
A few more points should be made before moving to the analysis. As
the discussion above indicates, an invention can provide value to a stan-
dard that incorporates it either by reducing the cost of compliance with the
standard or by increasing the attractiveness of the standardized product.
Either possibility causes sellers of the standardized product to demand the
invention. In the former case, though, the standardized product is assumed
to be equally desirable to an end consumer whether or not it incorporates
the invention, so there is no independent demand for the invention by
downstream buyers of the standardized product. Only when the invention
has independent technical value will there be additional buyers who pro-
vide a demand for the invention above and beyond the demand for the
standard.
These two possibilities present different problems when inferring the
importance of the invention. When the demand for the invention is due
purely to cost savings in complying with the standard, the invention and
the standard may be coextensive. In that case, the benefit of the invention
is inseparable from the benefit of the standard, and one can only measure
it by considering alternative standards. Where the invention provides tech-
nical benefits beyond those provided by the standard, on the other hand,
one can measure its benefits by considering alternative means of comply-
ing with the standard.
A. License Fees Attributable to Cost Savings
In principle, an invention may reduce the cost of complying with a
standard by a measurable amount. One could measure this reduction in
cost to determine the value of the patentee's contribution and thus the li-
cense fees to which it is entitled. One might question whether measuring
the demand for an invention, as courts do when calculating patent dam-
ages,58
overlooks the benefit of this cost reduction. But an invention's ca-
pacity for reducing costs for product manufacturers will in fact create de-
mand for the invention. The demand is defined, though, by the cost reduc-
tion that the invention makes possible. This possibility was acknowledged
by the Federal Circuit in Slimfold, where it said that "the advantage of the
57. See infra Part III.
58. See supra Part I.A.1.
INVENTIONS, INDUSTRY STANDARDS, AND IP
Ford invention was primarily a manufacturing advantage . . . and did not
greatly increase the value of the entire door."59
Does not incorporate a
patented invention
Incorporates
patented invention P
Complies with chosen
standard S
D: demand: $9
cost: $8
E: demand: $10
cost: $6
The most straightforward circumstances in which cost savings can be
measured are those in which there are alternative means of attaining com-
pliance. For example, in the table above, where it is possible to comply
with the standard without using any patented invention, at a manufacturing
cost of $8 (cell D), but use of invention P allows compliance at a cost of
$6 (cell E), the invention provides a cost savings of $2. The patentee is
certainly entitled to a license fee of $2 for the use of its invention in these
circumstances, at least if the standard-setter knew of the patent when it
adopted the standard, or did not know of the invention at all (in which case
it would not have relied on the existence of the invention in selecting the
standard).
If the standard-setting organization did not know of the patent, and the
organization relied on the ability to comply with the standard by using the
invention, the situation is more complicated.60
In this case, the problem is
that the invention is unexpectedly expensive,61
and therefore compliance
with the standard may also be unexpectedly expensive. In some cases, this
will occur because the cost difference between compliance with the inven-
tion and by alternative means is great; in others--perhaps the majority--it
59. Slimfold Mfg. Co. v. Kinkead Indus., Inc., 932 F.2d 1453, 1459 (Fed. Cir.
1991).
60. This has been the circumstance in most of the litigated cases. See Rambus, Inc.
v. Infineon Techs. AG, 155 F. Supp. 2d 668, 672 (E.D. Va. 2001), appeal docketed, 2002
WL 554344 (Fed. Cir.) (describing Rambus's failure to disclose existence of pending
patent to JEDEC); see also Dell, 121 F.T.C. at 617 (noting that Dell never disclosed exis-
tence of its patent to a standard-setting organization).
61. The ongoing Rambus litigation apparently presents an example of this situation.
See Tony Smith, Rambus's 'Very High' DDR Royalty Revealed, THE REGISTER (Mar. 5,
2001), at http://www.theregister.co.uk/content/3/18706.html (reporting that Rambus was
charging a royalty of 3.5% of sales for rights to patents that had been incorporated in a
standard, as compared with a 0.75% rate for some of its other patents).
BERKELEY TECHNOLOGY LAW JOURNAL
will occur because the invention is the standard.62
In either case, the prob-
lem is evaluating the possibility of alternative standards.63
Does not incorporate a
patented invention
Incorporates
patented invention P
Complies with chosen
standard S
D: demand: $9
cost: $8
E: demand: $10
cost: $6
Complies with alternative,
unchosen standard Sa
G: demand: $8
cost: $7
For example, suppose that the standard-setting organization relied on
the ability to comply with standard S at a cost of $6 in selecting that stan-
dard (cell E in the table). That is, suppose that the organization knew of
and could have selected an alternative standard Sa that cost $7 (in cell
G).64
So long as compliance with S at a cost of $6 is possible, there may
be no reason to consider Sa, but knowledge that compliance with S would
cost $8 (cell D, or cell E with the $2 license that the patentee would likely
charge) could provide such a reason. If the $1 cost difference seems insuf-
ficient, one can imagine a much higher difference. When S is coextensive
with P, and the patentee refuses to license, the cost is infinite. If the or-
ganization considered an alternative standard, and rejected it only on the
basis of cost, limiting the patentee to a license fee of only $1 (the cost dif-
ference between E, with the chosen standard, and G, the alternative) seems
straightforward.
The circumstances are more difficult when the organization did not
consider any such alternative standards. The question then is one of hypo-
thetical noninfringing alternatives: if there had been disclosure of the pat-
ent, and therefore disclosure of the higher-than-expected cost, would the
standard-setting organization have chosen a different standard that would
have served as a noninfringing alternative? Assuming that the patentee
was not part of the standard-setting organization, the purpose of asking
62. See infra note 104.
63. That is so, at least, if any alternative means of compliance with the standard has
an unacceptably high cost.
64. At least in some cases evidence of such noninfringing alternatives may be avail-
able. See Dell Computer Corp. Consent Agreement Statement, Federal Trade Commis-
sion, (June 17, 1996) at http://www.ftc.gov/opa/1996/9606/dell2.htm (describing the Dell
case as one in which "there is evidence that the association would have implemented a
different non-proprietary design had it been informed of the patent conflict during the
certification process").
INVENTIONS, INDUSTRY STANDARDS, AND IP
this question is not to determine whether the patentee behaved unethi-
cally--it had no obligation to disclose--but to determine just how much
its invention contributed to the demand for the standard. If, with knowl-
edge of the patent, the standard-setting organization would have found a
noninfringing standard at the same cost, it is fair to infer that the demand
for the standard is not due to the incorporation in it of the patentee's in-
vention.
The patentee should be free to contest such an inference, though. For
example, the patentee might argue that the circumstances are as shown in
the table above. As shown there, the selection of Sa would have provided a
noninfringing alternative with $1 of the $2 cost savings provided by in-
vention P in standard S; if so, the patentee of P is entitled only to the addi-
tional $1 in cost savings that its invention provides. But the demand for
the standard Sa in cell G is less than the demand for the standard S (incor-
porating invention P) in cell E. If the greater demand for standard S is due
to the technical contribution of invention P, the patentee is entitled to the
revenue from that greater demand; this issue is taken up in more detail in
the next section.
Although the principles described above apply regardless of whether
the patentee is a member of the standard-setting organization, organization
membership is not irrelevant. The arguments above depend on the exis-
tence of an alternative standard Sa. Those arguments will always be
somewhat difficult for the parties to make, given their hypothetical nature,
so allocating the burden of proof will be important. If the patentee is not
part of the standard-setting organization, it seems appropriate to allocate
the initial burden of showing that an alternative standard could have been
chosen to the defendant infringer (or to the standard-setting organization
itself). But if the patentee is a member of the standard-setting organiza-
tion, it is reasonable to charge it with acting so as to facilitate the organi-
zation's process.65
Consequently, if by its nondisclosure it fails to do that,
it is reasonable to require, if it later brings an infringement suit, that it bear
the burden of showing that no alternative standard would have been cho-
sen had it disclosed its patent.66
65. See Rambus, Inc. v. Infineon Techs. AG, 164 F. Supp. 2d 743, 751 (E.D. Va.
2001) (discussing JEDEC committee chairman's testimony that committee's successful
operation depended on disclosure requirement).
66. These allocations of the burden of proof are generally consistent, or at least not
inconsistent, with current law. As the Federal Circuit has said, the patentee generally has
the burden of showing that the four Panduit factors (one of which is the absence of a non-
infringing alternative) are met. Rite-Hite Corp. v. Kelley Co., Inc., 56 F.3d 1538, 1545
BERKELEY TECHNOLOGY LAW JOURNAL
This burden is nevertheless less severe than some would impose. For
example, JEDEC, the relevant standard-setting organization in the Rambus
litigation, suggests in its amicus brief67
that nondisclosure should make a
patent unenforceable.68
JEDEC argues, as suggested above, that with
knowledge of the existence of patent protection for technology it is con-
sidering, a standard-setting organization might "seek to use alternative
technology that is not saddled with actual or potential patent rights."69
Does not incorporate a
patented invention
Incorporates
patented invention P
Does not comply with
standard
A: demand: $6
cost: $5
B: demand: $7
cost: $5
Complies with chosen
standard S
D: demand: $9
cost: $8
E: demand: $10
cost: $6
JEDEC says that it will use patented technology, but only if the pat-
entee agrees to a royalty-free or reasonable-and-nondiscriminatory licens-
ing policy.70
Given the vague meaning of "reasonable," as noted earlier, it
is difficult to assess JEDEC's policy. But to the extent that the require-
(Fed. Cir. 1995) (en banc). "The burden then shifts to the infringer to show that the infer-
ence is unreasonable for some or all of the lost sales." Id. (citation omitted).
Here, where the noninfringing alternative is a hypothetical one, the patentee
would generally meet its burden by showing the absence of any actual noninfringing al-
ternative. Therefore, the burden would be on the infringer to show that the inference of
damages was an unreasonable one, as by showing that the standard-setting organization
would have chosen a noninfringing alternative if it had had information about the patent.
But if the patentee was a member of the organization, the approach proposed here would
provide for a re-shifting of the burden to the patentee. One could view this approach as
conforming to the Rite-Hite analysis simply by treating a showing of the patentee-
member's nondisclosure as a showing that the inference of damages was unreasonable.
The patentee could then show once again that the inference was reasonable by showing
that no noninfringing alternative would have been chosen.
67. Amicus Curiae Brief, supra note 7, at 11-13.
68. The brief says only that a court should impose "suitable remedies," id. at 14, but
it cites with approval a number of cases in which courts have not enforced patents in such
circumstances, id. at 11-13.
69. Id. at 6. The brief also acknowledges, though, that "[i]n some cases . . . the tech-
nology that is the subject of a patent or patent application may be technically superior to
alternatives." Id. This presents the standard-setting organization's dilemma: the organiza-
tion would prefer to avoid patented inventions, but to do so might result in a standard
with significantly less technical merit.
70. Id.
INVENTIONS, INDUSTRY STANDARDS, AND IP
ment contemplates any meaningful content for the term "reasonable,"71
it
seems an undesirably strict policy. Even if the patentee demands licensing
revenues that are greater than those to which its contributions in cost re-
ductions (and technical advantages, as discussed in the next section) enti-
tle it, the adoption of the standard may be beneficial. That is, referring to
the table, cell E is an improvement over cell A or B, even if the patentee
demands a licensing fee greater than the $2 to which it is entitled based on
the manufacturing cost savings that its invention provides. This is so even
if the invention provides no independent value, so that the demand values
in cells A and B and those in cells D and E are the same, when the benefit
of standardization outweighs the cost of the license. If so, an organization
that refused to use the invention would be shooting itself in the foot if
there were no acceptable alternative standard Sa. The approach proposed
here--to require the patentee to prove that in fact there was no alternative
standard--avoids that danger, while still ensuring that the standard-setting
organization does not unnecessarily suffer from exploitation by patentees.
Moreover, this analysis justifies the policy of those standard-setting
organizations that require members to disclose any inventions that may
affect particular standard-setting activities, even if the patentee is not itself
a part of those particular activities.72
One might argue that if a member did
not participate in a particular standard-setting effort, it could not decep-
tively promote standards that would incorporate its patents. But the ap-
proach described above places the burden on the patentee not as a penalty
for any particular act of deception, but in order to facilitate the standard-
71. If the policy merely requires that the patentee will license its patent on some
terms, rather than refuse to license it entirely, the term "reasonable" should be said to
have no meaning. But the policy might in fact not contemplate any such meaning, if its
focus is on ensuring that patentees act nondiscriminatorily.
72. For example, the W3C Patent Policy Framework requires disclosure from all
W3C members:
W3C Members agree to use good faith efforts to disclose all patents
known to them which may contain Essential Claims. Disclosure obliga-
tions stated here cover:
1. W3C Members: whether or not they are part of a given Working
Group
2. Working Group participants: individuals, whether or not they
are in good standing, who have joined a Working Group, and their al-
ternates
3. Working Group contributors: Working Group members who
make formal contributions to a Working Group W3C Patent Policy
Framework § 7 (Aug. 16, 2001 working draft), at http://www.-
w3.org/TR/2001/WD-patent-policy-20010816.
BERKELEY TECHNOLOGY LAW JOURNAL
setting process. By joining a standard-setting organization, an organization
member presumptively commits itself to the goals of the organization, and
likely benefits in its industry relations from that commitment.73
It there-
fore should not be permitted to disavow its commitment later, when it sees
an opportunity for profit. The adoption by standard-setting organizations
of general--i.e., nonstandard-specific--disclosure policies is consistent
with this understanding.
B. License Fees Attributable to Technical Advances
Demand for a patented invention may also arise not from any cost sav-
ings that it provides, but from its contribution to the desirability of the
standardized product. Generally speaking, it is more difficult to derive ob-
jective measures of the demand for an invention than it is to derive meas-
ures of the cost savings it provides.74
Nevertheless, one can sometimes
distinguish demand for an invention and demand for a standard, even
where the invention and the standard coincide, by drawing inferences from
market conditions before and after the invention is incorporated into the
standard.
Does not incorporate a
patented invention
Incorporates
patented inven-
tion P
Incorporates alter-
native patented
invention Pa
Does not comply
with standard
A: demand: $6
cost: $5
B: demand: $7
cost: $5
C: demand: $9
cost: $6
Complies with
chosen standard S
D: demand: $9
cost: $8
E: demand: $10
cost: $6
F: demand: $9
cost: $6
Suppose, referring to the table above, that three approaches, A, C, and
E, each accomplish a particular goal, and that C and E are patented, but
that the patent protection for E is secret.75
Suppose further that the indus-
try has generally adopted either A or C, but that an industry standardiza-
tion effort selects a standard based on E, perhaps to avoid giving an advan-
73. See Marvin Lumber & Cedar Co. v. PPG Indus., Inc., 223 F.3d 873, 883 (8th
Cir. 2000) (accepting district court's view that activity in standard-setting organization
indicates possession of specialized knowledge in field).
74. See generally William M. Landes & Richard A. Posner, Market Power in Anti-
trust Cases, 94 Harv. L. Rev. 937 (1981) (explaining the difficulty of measuring de-
mand).
75. The significance of this latter assumption is explained in the text following note
76 infra.
INVENTIONS, INDUSTRY STANDARDS, AND IP
tage to the users of either A or C. Consumers demand the standardized
product, and the industry switches to E. Under these circumstances, the
fact that users had not selected E before it was chosen as a standard, but
did so after it became the standard, suggests that their later choice of E
was not due to its intrinsic value but to the standardization.76
Moreover,
the initial secrecy of the patent supports this conclusion, because it indi-
cates that the industry did not avoid E merely to avoid paying licensing
fees.
If E was widely used even before its incorporation into a standard, the
situation is more complicated. One might then conclude that the standard-
setting organization adopted E for its intrinsic value. But since the patent
on E was secret before its incorporation into the standard, E might only
have been widely used because it had no associated licensing fees. That is,
one cannot conclude from users' adoption of an invention when the use of
it is free that they would also be willing to pay for it. For that reason, the
use of a patented invention prior to its standardization does not justify the
patentee's post-standardization imposition of more onerous licensing
terms.
The patentee would, however, be justified in continuing to impose
whatever licensing terms it imposed before standardization.77
That is, sup-
pose P had been adopted by some, but not all, users before being incorpo-
rated into a standard. Suppose that after P's incorporation into a standard,
other users also sought to license it. If those users had previously used
some technology other than P, and in fact had declined to license P on its
pre-standardization terms, they might contend that, for them, it was P's
value as a standard, not as technology, that they sought. But even if they
did not prefer the technology of P to other alternatives, given their relative
pre-standardization prices, the pre-standardization license terms of P are
still the best estimate of its value. That is particularly so in that P might
have technological benefits in a standardized context that it did not have
when there was no standard. This is reflected in the table in the greater
76. The possibility that the standard might have been made possible only by the ex-
istence of E, so that E should be entitled even to the returns from standardization, is dis-
cussed below. See infra Part II.C.
77. IBM, for example, takes this approach, at least in some instances. Its statement
regarding licensing of a patent that it believes is relevant to the standard-setting activities
of the Internet Engineering Task Force states that "IBM is willing upon written request to
grant a nonexclusive license under such patents on a nondiscriminatory basis and on rea-
sonable terms and conditions, including its then-current terms and royalty rates." See
Posting of Chuck Adams, Jr., wcadams@us.ibm.com, at http://www.ietf.org/ietf/-
IPR/IBM-SNMP (May 14, 2001).
BERKELEY TECHNOLOGY LAW JOURNAL
demand for E over F or D, though the demand for B was less than that for
C.
It must be clear, though, that the pre-standardization terms represent an
objective estimate of value. For example, the court in Townshend v. Rock-
well International Corp.78
might too readily have accepted an argument
along this line, where the alleged infringer argued that the patentee had
sought unfair licensing terms after adoption of a standard:
Even if the [c]ourt were to consider the unfair terms alleged by
[the alleged infringer], the [c]ourt finds that these terms do not
state an injury to competition. First, with respect to the proposed
royalty rates, the [c]ourt notes the initial licensing proposal dated
September 1997 sought a maximum $1.25 per-unit royalty for
client-end products and a maximum $9.00 per-port royalty for
server-end products. In September 1998, after the V.90 standard
had been adopted, [the patentee] submitted a revised licensing
proposal which sought a maximum $1.25 per-unit royalty for cli-
ent-end products and a maximum $2.50 per-port royalty for
server-end products.79
If the court meant to suggest here that the absence of any royalty in-
crease after adoption of the standard was evidence of the reasonableness
of the terms, that was incorrect. The initial licensing proposal was submit-
ted to the standard-setting organization,80
and thus was made in anticipa-
tion of standardization. For pre-standardization licensing terms to be sig-
nificant, they must have been determined prior to any effect, actual or an-
ticipated, of standardization.
That is not to say, however, that if the patentee did not seek license
fees prior to standardization, it would never be justified in demanding such
fees after standardization. In such circumstances, the analysis could rely
on inferences from the post-standardization market, but this is only possi-
ble in certain circumstances. When saving costs is the issue, as discussed
in the previous section, one can hypothesize alternative standards, because
one can objectively evaluate the cost of compliance with hypothetical
standards.81
In contrast, it is more difficult to estimate the demand for hy-
78. No. C 99-0400 SBA, 2000 U.S. Dist. LEXIS 5070 (N.D. Cal. Mar. 28, 2000).
79. Id. at *23-*24.
80. Id. at *21.
81. It is also possible to infer, with at least some confidence, whether a standard-
setting organization would have adopted an alternative standard.
INVENTIONS, INDUSTRY STANDARDS, AND IP
pothetical inventions or standards.82
Therefore, only where the post-
standardization market presents actual alternatives can one reliably esti-
mate the relative significance of the invention and the standard.
Does not incorporate a
patented invention
Incorporates
patented invention P
Complies with chosen
standard S
D: demand: $9
cost: $8
E: demand: $10
cost: $6
As an example of an instance where such an alternative is available,
suppose, referring again to the table, that it is possible to comply with the
standard either by using invention P or without using any invention. Sup-
pose also, as shown in the table, that the alternatives for compliance with
the standard are D and E, that the cost of compliance with D is $8, and that
the cost of compliance with E is $6. Under these circumstances, the pat-
entee would presumably demand, and users would pay, a license fee of at
least $2 for E. If the license fee paid exceeded $2, one could assume that
users license P for its technical contribution, as well as (or rather than) for
the cost savings that it provides. Therefore, the patentee's entitlement to
the $2 from cost savings would be determined as described in the previous
section, but the patentee would be entitled to any license fees beyond the
fees for cost savings, which must be due to demand for the invention's
technical contribution.83
Incorporates
patented invention P
Incorporates alternative pat-
ented invention Pa
Complies with chosen
standard S
E: demand: $10
cost: $6
F: demand: $9
cost: $6
One can draw similar inferences from alternatives that appear subse-
quent to standardization, though the process is somewhat more compli-
cated. For example, suppose that there is no alternative D, but that after
standardization, an alternative invention Pa is created that allows compli-
ance with the standard, with cost and demand of F in the table. Suppose
82. Cf. Micro Chem., Inc. v. Lextron, Inc., 161 F. Supp. 2d 1187 (D. Colo. 2001)
(illustrating difficulty of proving market demand for specific claimed features of patent).
83. That is, it might be appropriate to reduce the patentee's fee entitlement for cost
savings, as described in the previous section, but its total fee for both demand and cost
contributions could still be greater than $2.
BERKELEY TECHNOLOGY LAW JOURNAL
that F is licensed for $2, and that E had been licensed for $3. If buyers
choose F over E under these conditions, one could infer that the combina-
tion of E's cost savings and technical advantages is no more than $1
greater than the same combination for F. Therefore, one could use the rela-
tive costs of E and F to determine how much buyers would be willing to
pay for E's technical advantages. For example, suppose that the cost to
comply with the standard were the same for E and F (absent license fees).
One could then infer that buyers value E's technical contributions at no
more than $1 more than F's. More significantly, if the cost to comply with
the standard were $2 less for E than for F (as if the cost in cell E of the
table were $4), one could infer that E's technical contributions were val-
ued by buyers at (no more than) $1 less than F's (in which case demand
for E in the table would be $8 or less).84
The scenario in the preceding paragraphs somewhat resembles the GIF
controversy.85
When Unisys asserted its patent on the algorithm for gener-
ating compressed GIF files, competitors initiated an effort to develop a
method of creating the files without infringing the patent.86
Had this effort
been entirely successful, one could perhaps make the calculations de-
scribed above. But the alternative method that avoided the patent was not
entirely successful,87
preventing any clear comparison. Nevertheless, be-
cause the alternative method apparently worked in some applications but
not in others,88
it might have been possible to determine the relative value
of the two implementations.
84. That is, E's total contribution, relative to F, is $1, and its relative cost savings
contribution is $2. Therefore, its relative technical contribution is--$1. Because F's (ab-
solute) total contribution is $2, its technical contribution can be no more than $2, and E's
therefore can be no more than $1.
85. See Michael C. Battilana, The GIF Controversy: A Software Developer's Per-
spective, CLOANTO, at http://cloanto.com/users/mcb/19950127giflzw.html (last modified
Feb. 2, 2002).
86. Id.
87. Id.
88. Specifically, in applications where compression of the data was not critical, al-
ternatives might have worked:
Some of the most active developers decided to collaborate on the de-
sign of a patent-free evolution of GIF (and TIFF's LZW compression
mode [which Unisys had patented]). A method was quickly found to
create uncompressed GIF files without using LZW code, while remain-
ing compatible with existing GIF loaders. Also, a variety of different
procedures and data structures (such as Shannon-Fano and AVL trees)
have been used to compress data in ways similar, if not always equiva-
lent, to LZW. But a diversity in procedures and data structures alone
INVENTIONS, INDUSTRY STANDARDS, AND IP
Knowing this sort of information about relative values, one could de-
termine P's owner's total entitlement to reasonable royalties (or to dam-
ages based on reasonable royalties) based on cost savings and technical
advantages. As described in the previous section, a patentee's entitlement
to returns on cost savings that its invention makes possible may be limited
depending on whether it disclosed the existence of its invention during the
standard-setting process. However, the patentee's entitlement to returns on
the independent technical contribution made by its invention is not so lim-
ited; it is therefore important to allocate the returns sought by the patentee
to their correct sources.
The principles discussed above apply even in cases in which the pat-
entee, during patent prosecution, amends its claims to conform to a stan-
dard under consideration, as has happened in some cases.89
Because the
key issue under the approach proposed here is whether the demand for the
invention arises from its technical contribution or from its adoption as a
standard, that the claims are identical to the standard is irrelevant. Also
irrelevant is the question that some have suggested for dealing with such
cases: should the standard-setting organization, rather than the patentee,
be credited with "inventing" the patented invention?90
Under the approach
proposed here, the question is the more fundamental one of whether the
standard-setting organization or the patentee created the demand for the
invention.
C. License Fees Attributable to Direct Impact on
Performance or Interoperability
The most conceptually difficult cases under the approach proposed
here arise when the invention that is incorporated in the standard provides
some advantages over alternative approaches in achieving the goal of the
apparently does not escape the patent. As one expert said, "If the output
data is [compressed] GIF, the compressor infringes the Unisys patent
regardless of the algorithm."
Id.
89. See, e.g., Mueller, supra note 3, at 913-14 ("After the [California Air Resources
Board (CARB)] issued its regulations, the refiners contended, Unocal cancelled its origi-
nal patent claims and intentionally substituted amended claims to `resemble' the CARB
regulations.") (footnote omitted).
90. See id. This argument was made in the Unocal case, but did not prevail. See Un-
ion Oil Co. of Cal. v. Chevron U.S.A., Inc., 34 F. Supp. 2d 1222, 1224 (C.D. Cal. 1998),
aff'd sub nom Union Oil Co. of Cal. v. Atl. Richfield Co., 208 F.3d 989 (Fed. Cir. 2000),
cert. denied, 531 U.S. 1183 (2001) (stating that the defendants made a "`derivation' ar-
gument, the gist of which was that Unocal had copied the invention from CARB," but
that "[n]o competent evidence was introduced in support of that argument and the jury
did not find the patent invalid on that basis").
BERKELEY TECHNOLOGY LAW JOURNAL
standard. Here the distinction between standards directed at interoperabil-
ity and those directed at improved performance becomes important. Many
inventions embody technical advances that are not directed specifically at
interoperability.91
For those inventions, it is not difficult, at least concep-
tually, to distinguish the technical benefits of the invention and the inter-
operability benefits of the standard. Where both the invention and the stan-
dard are directed at improved performance, that distinction is more
difficult to maintain.
The Unocal case92
presents an example of an invention that is arguably
essential to a performance-directed standard. Unocal received a patent on
gasoline formulations that produce fewer emissions than previous formu-
lations.93
When the California Air Resources Board ("CARB")94
enacted a
new, more demanding emissions standard,95
Unocal notified its competi-
tors that they would be required to license its invention.96
In such a case,
where the standard is directed at improved performance, the patentee is
entitled to whatever returns it can achieve. If there is more than one way
of complying with the performance standard, the patentee will face com-
petition, which will constrain its licensing terms. Also, if the patentee's
invention is the only means of complying with the standard, or if alterna-
tives are significantly less desirable, the invention can be said to make the
standard possible, and thus the patentee is entitled to the returns derived
from demand for the standardized product.97
91. See supra Introduction (Rambus discussion).
92. Union Oil Co. of Cal. v. Atl. Richfield Co., 208 F.3d 989 (Fed. Cir. 2000), cert.
denied, 531 U.S. 1183 (2001).
93. Id. at 991.
94. Because the CARB was a state agency, Unocal presents some issues that are not
present in cases that involve an industry standard-setting body. For example, the CARB
standard was coercive in a more direct way than are "voluntary" industry standards.
These differences are not important in the present context, however.
95. See Aguilar v. Atl. Richfield Corp., No. 7008101997, WL 824017 (Cal. Super.
Ct. Oct. 17, 1997), superceded by Aguilar v. Atlantic Richfield Corp., 999 P.2d 666 (Cal.
2000) (describing adoption of new CARB standard).
96. Unocal Corp., Unocal Awarded Patent for Reformulated Gasolines; Plans to Li-
cense Patent, (Jan. 31, 1995), at http://www.unocal.com/rfgpatent/rfgnr1.htm. Some of
Unocal's competitors contended that it had behaved inequitably in keeping the existence
of its patent application secret while CARB deliberated on its standard. Such conduct, if
proven, might be an independent reason for denying a patentee license revenue, but it is
not directly relevant to the approach to these issues that is proposed here.
97. This is so under the basic approach described here, which turns on determining
the sources of market demand. If the patentee engaged in deception in connection with
INVENTIONS, INDUSTRY STANDARDS, AND IP
Where a standard is directed at interoperability, rather than perform-
ance, a particular invention generally will not be necessary to make the
standard possible. In most such instances, interoperability could be
achieved in any of a variety of ways, just as a particular level of perform-
ance can in principle be achieved in many ways. The difference in the in-
teroperability context is that the standard achieves its goal by specifying a
particular approach, thus eliminating some alternatives that might other-
wise have provided competition. As a result, even if the particular stan-
dard chosen is technically better than alternatives, and is better because it
incorporates a patented invention, the patentee may also benefit from the
standard-setting organization's exclusion of possible competition. It may
therefore be possible to distinguish two distinct sources of demand: tech-
nical performance and interoperability. When that is the case, one can ap-
ply the principles from the previous section to determine the patentee's
contribution to demand, and thus its entitlement to licensing revenue.
For example, Rambus bases its ongoing patent infringement action98
on its claims pertaining to memory devices, as in the following example:
14. A synchronous semiconductor memory device having at least
one memory section which includes a plurality of memory cells,
the memory device comprising:
a programmable register to store a value which is repre-
sentative of a number of clock cycles of an external
clock to transpire before data is output onto an external
bus in response to a read request; and
a plurality of output drivers, coupled to the bus, to output
data in response to the read request, wherein the output
drivers output data on the bus after the number of clock
cycles of the external clock transpire.99
The goal of this invention is to make "response time more predictable
. . . , thereby allowing the system to plan for transfers and improving over-
all traffic flow over the bus."100
Rambus alleges that products that comply
the standard-setting process, it might still be appropriate to limit its returns for that rea-
son. But see supra note 68.
98. Rambus, Inc. v. Infineon Techs. AG, No. 3:00cv524 (E.D. Va. filed Aug. 8,
2000).
99. U.S. Patent No. 5,953,263 col. 25, ll. 32-42 (issued Sept. 14, 1999).
100. Memorandum opinion, Rambus, Inc. v. Infineon Techs. AG, No. 3:00cv524,
2001 U.S. Dist. LEXIS 10990, at *6-*7 (E.D. Va. Mar. 15, 2001), appeal docketed, 2002
WL 554344 (Fed. Cir.).
BERKELEY TECHNOLOGY LAW JOURNAL
with a JEDEC101
standard for dynamic random access memory devices
infringe its claims. Although JEDEC's standards exist to ensure compati-
bility among different manufacturers' devices,102
the Rambus invention
improves memory speed103
and does not obviously further the goal of
compatibility.
There is, however, the possibility that an invention could be directed
specifically to, or at least could contribute to, improvements in interopera-
bility. For example, a form of computer bus connection that works with a
wide variety of circuit board configurations might contribute significantly
to interoperability. If so, its inventor could be entitled to the profits made
possible by the demand created by adoption of that invention as an inter-
operability standard.
This might in fact have been true of the invention in the Dell case.104
The patent at issue in that case, U.S. Patent No. 5,036,481, includes the
following independent claim:
1. A personal computer system having an I/O channel and a memory
channel, and having a dual purpose expansion slot, comprising:
(a) a chassis;
(b) a main logic board mounted on the chassis and incorpor-
ating the I/O channel and the memory channel;
(c) a fixed number of expansion slots, including the dual pur-
pose expansion slot, each occupying a fixed volume, posi-
tioned over the main logic board for providing space for se-
lective connections of I/O devices, implemented on full
length and short logic cards, to the I/O channel;
(d) a high speed memory system mounted on the main logic
board and connected to the memory channel, occupying a
fixed amount of space; and
(e) expansion high speed memory, mounted on the main logic
board within the fixed amount of space, connected to the
memory channel, and occupying a portion of the dual pur-
pose expansion slot, the remaining portion being occupied
101. See JEDEC, at www.jedec.org (last visited July 28, 2002).
102. Rambus, Inc. v. Infineon Techs. AG, 155 F. Supp. 2d 668. 672 (E.D. Va. 2001),
appeal docketed, 2002 WL 554344 (Fed. Cir.).
103.See supra note 6.
104. In re Dell Computer Corp., File No. 931-0097, 1995 FTC LEXIS 466 (F.T.C.
Oct. 20, 1996).
INVENTIONS, INDUSTRY STANDARDS, AND IP
by a short logic card, thereby enabling increased high speed
memory capacity without eliminating I/O capability.105
The flexibility provided by the "dual purpose" expansion slot might
well make standardization more attractive. As the patent notes, the inven-
tion "may be practiced in other personal computers with more or less
memory, [and with] more or less [sic] expansion slots with different im-
plementations of memory and connectors."106
It is plausible that it was ex-
actly this flexibility that made the invention a desirable standard.
Of course, other means might be used to achieve similar flexibility.
Thus, the principles discussed above for analyzing whether an alternative
standard might have been chosen apply in this context also, as do other
principles applicable to a patentee's nondisclosure to a standard-setting
organization. The FTC's enforcement proceeding against Dell was there-
fore appropriate, particularly under its view that if Dell's patent had been
disclosed, the standard-setting organization might have adopted a nonpro-
prietary standard.107
Indeed, where the invention at issue contributes di-
rectly to the goal of the standard-setting organization, it seems particularly
appropriate to impose on the patentee a duty to disclose.
D. License Fees for Patented Elements in De Facto
Standards
Patented de facto standards present a special case in the broader range
of patented inventions that enable standardization. A de facto standard is
one that achieves industry acceptance without the imprimatur of any offi-
cial or quasi-official standard-setting body.108
Initially, it might seem that
the incorporation of an invention in a de facto standard would indicate that
the invention contributes to making the standard possible, in the sense dis-
cussed in the previous section. That is, it might seem that when the market
chooses a standard that incorporates an invention, it will do so because the
invention best serves the purposes of standardization.
105.U.S. Patent No. 5,036,481, at col. 3, ll. 41-65 (issued July 30, 1991).
106. Id. at col. 3, ll. 34-37.
107. See supra note 64.
108.See Mueller, supra note 3, at 905 (noting that de facto standards arise not from
involvement of particular organizations but from market activity); see also Lauren Johns-
ton Stiroh & Richard T. Rapp, Market Power in Technology Markets, 1999 ALI-ABA 61,
70 (Apr. 22, 1999) (predicating development of de facto standard on amount of technol-
ogy dispersed to market).
BERKELEY TECHNOLOGY LAW JOURNAL
In fact, though, the market is likely to choose a standard just as a stan-
dard-setting body does, in order to maximize the combination of inherent
technical benefits and suitability to the goals of the standard (such as, for
example, interoperability), and in order to minimize the costs of searching
for a standard. Consequently, an invention may become part of a standard
not because it makes any particular contribution to the goals of the stan-
dard, but only because it provides greater (or at least no less) technical
benefit or is more widely available than alternative possibilities. Thus, a
patentee is generally entitled to revenues from the adoption of its inven-
tion in a de facto standard under the same principles described in the pre-
ceding sections for de jure standards.
There is one important difference, though. Because a de facto standard
comes into existence without a formal standard-setting process, it is more
difficult to define a point at which a duty of disclosure would arise for the
patentee. If no such duty exists, the burden of showing that if information
about the patent had been available, an alternative standard would have
been adopted will fall, as described above, on the infringer. Nevertheless,
in some cases it may be reasonable to impose upon the patentee of an in-
vention incorporated in a de facto standard a duty to disclose.
For example, market participants sometimes promote the adoption of
their approach to a particular problem, and this promotion can contribute
to the development of a de facto standard.109
Where such promotion oc-
curs, the patentee should be charged with a duty to disclose the existence
of any relevant patents. The rationale would be that the patentee, by pro-
moting its invention, is participating in the standard-setting process, even
if the process is an informal one. Moreover, the disclosure required, in this
context as with formal standard-setting processes, should be sufficient to
put potential adopters on notice that they may be obligated to pay licens-
ing fees. For example, Microsoft has promoted its "HailStorm" product
(now called "Microsoft .NET My Services"), which it describes as "a
user-centric architecture and set of XML Web services," as an industry
109. A recent example may be Geoffrey Moore's book Crossing the Chasm. See
GEOFFREY MOORE, CROSSING THE CHASM 69-86 (paperback ed. 1995). Moore formed
The Chasm Group, LLC to provide advice on putting his market strategies to work, and
the success of his books and speaking tours seem to indicate the possibility that promot-
ing his approach might lead to creation of a de facto standard for high technology market
analysis. See generally Chasm Group webpage, available at http://www.chasmgroup.com
(last visited May 24, 2002).
INVENTIONS, INDUSTRY STANDARDS, AND IP
standard.110
In its literature promoting HailStorm, Microsoft initially men-
tioned the possibility that it had applicable intellectual property rights:
Microsoft may have patents, patent applications, trademarks,
copyrights, or other intellectual property rights covering subject
matter in this document. Except as expressly provided in any
written license agreement from Microsoft, the furnishing of this
document does not give you any license to these patents, trade-
marks, copyrights, or other intellectual property.111
Although Microsoft is certainly correct that its marketing of HailStorm
should not be construed as granting a license to its intellectual property, by
promoting HailStorm as a standard, Microsoft should have the burden of
disclosing any intellectual property that would influence adoption of Hail-
Storm as a de facto standard. A general disclosure of the existence of some
intellectual property rights, as in the passage quoted above, is not suffi-
cient because it does not allow users to evaluate the cost of using Micro-
soft's product or the benefits of seeking alternatives. In the absence of
more specific disclosure, describing the nature of the intellectual property
and its applicability, Microsoft should be required, if it seeks returns from
the adoption of its products as standards, to show that even with disclosure
its products would have been chosen as standards.
110. Microsoft's marketing literature is fairly explicit:
HailStorm is the user-centric architecture and set of services for .NET
that deliver personally relevant information through the Internet to a
user, to software running on the user's behalf, or to devices working for
the user. HailStorm services are accessed through SOAP (Simple Ob-
ject Access Protocol) and XML (eXtensible Markup Language), which
are open access technologies: they can be called from any network-
connected device that supports SOAP, regardless of operating system
or service provider. SOAP and XML are the open Internet standards
Microsoft has helped champion throughout the first phase of the .NET
rollout. HailStorm is the next logical step: Microsoft began by encour-
aging the general standards and introducing the first Web services tools
and infrastructure. Now we're leading the way to the first set of com-
pelling Web services.
Microsoft Corp., Building User-Centric Experiences: An Introduction to Microsoft Hail-
storm 4-5 (Mar. 2001).
111. Id. at 13. Interestingly, the latest version of this paper, though in similar in
many ways to the older one, omits this notice regarding intellectual property. See Micro-
soft Corp., Building User-Centric Experiences with .NET Services, at http://-
www.microsoft.com/myservices/services/userexperiences.asp (visited Apr. 2, 2002).
BERKELEY TECHNOLOGY LAW JOURNAL
E. Lock-In and IMS Health
The European Commission's current action against IMS Health Inc.
presents interesting variations on these issues.112
It then, with the aid of
some pharmaceutical companies,113
developed a "brick structure" that di-
vided Germany into small geographical areas. IMS Health provided phar-
maceutical sales data to pharmaceutical companies organized according to
this structure. IMS Health received a copyright in Germany for its "1860"
brick structure,114
which became the industry standard, despite competi-
tors' efforts to introduce other structures. It refused to license the 1860
brick structure to competing data providers, and the European Commis-
sion challenged the refusal to license as an abuse of IMS Health's domi-
nant position under Article 82 of the EC Treaty.
The pharmaceutical companies' adoption of the 1860 brick structure
without the intervention of a standard-setting body appears to make it a de
facto standard. On the other hand, because the pharmaceutical companies,
through their trade group, were involved in its creation, the adoption has
some de jure character as well. Regardless, the critical point is that users
do not demand the specific geographical divisions of the 1860 brick struc-
ture, but simply the existence of some acceptable brick structure. There-
fore, because IMS Health does not have intellectual property protection
for the invention of brick structures in general, but only for its particular
brick structure, it is not entitled to all the revenues from the standardiza-
112. Initially, the Commission imposed interim measures that required IMS Health
to license its "1860 brick structure," which is described in the text. NDC Health/IMS
HEALTH: Interim Measures, 2002 O.J. (L 59) 18, available at http://europa.eu.int/-
comm/competition/antitrust/cases/decisions/38044/en.pdf [hereinafter Commission Deci-
sion]. Subsequently, the president of the Court of First Instance (C.F.I.) issued an ex
parte interim order suspending the operation of the Commission's decision. Case T-
184/01, IMS Health Inc. v. Commission, 2001 E.C.R. II-2349, available at http://curia.-
eu.int/en/jurisp/index.htm. That order was confirmed after submission of arguments by
the parties and a hearing, Case T-184/01, IMS Health Inc. v. Commission, 2001 E.C.R.
II-3193, available at http://curia.eu.int/en/jurisp/index.htm [hereinafter C.F.I. Decision],
and a challenge to it was rejected by the president of the European Court of Justice, Case
C-481/01, NDC Health Corp. v. IMS Health Inc., Apr.11, 2002, available at http://curia.-
eu.int/en/jurisp/index.htm.
113. See C.F.I. Decision, supra note 112, at  128 ("many of [IMS Health's pharma-
ceutical company] clients appear, as alleged in the contested decision . . . to have played a
significant role . . . in the development of 1 860 brick structure").
114. The ultimate validity of the copyright is as yet undetermined. See C.F.I. Deci-
sion, supra note 112, at  8-17.
INVENTIONS, INDUSTRY STANDARDS, AND IP
tion.115
The Commission seems to have been correct, then, in imposing a
licensing obligation on IMS Health,116
despite some significant differences
from the usual standards context.
First, that the intellectual property at issue is a copyright rather than a
patent does not alter the issues significantly. Although copyright law,
unlike patent law, does not prohibit independent creation (as distinguished
from copying) of the original work, the adoption of the copyrighted work
as a standard eliminates that possibility. That is, although IMS Health's
competitors did not seek to copy the 1860 brick structure per se, they did
seek to comply with the standard, which required copying the 1860 brick
structure. As a result, the copyright provided a patent-like monopoly. That
is likely to be the case in most standards contexts, where it is conformity,
not creativity, that matters after the standard is accepted.117
Second, the 1860 brick structure apparently became a standard not so
much because of a need for interoperability among firms using it, but as a
way of preserving the value of individual firms' investments in it.118
The
various pharmaceutical companies were not exchanging their data, but had
invested in adapting their operations to the 1860 brick structure. Neverthe-
less, the critical point continues to be that the demand for the standard
115. One might argue that in a sense similar to that discussed above, IMS Health's
creation of the 1860 brick structure enabled the creation of the industry standard, because
before its creation, there was no structure available for standardization. It appears not,
however, to have been the specific 1860 structure that enabled the standardization, but
merely the existence of some acceptable brick structure. Therefore, by the same reason-
ing, IMS Health is not entitled to all the revenues from the standardization.
116. Commission Decision, supra note 112, at  185. The Commission did not rely
on this rationale, however. Instead, it appears to have relied on the involvement of the
pharmaceutical companies in creating the standard:
The input which the pharmaceutical companies have made to the struc-
ture has contributed greatly to its status as a de facto industry standard
and to their current dependence on this structure as a format for the re-
ceipt of regional sales data services. It is therefore the case that refusing
access to this structure to competitors on the relevant market would ex-
clude all competition from this market, and that therefore IMS' refusal
to license the 1860-brick structure involves abusive conduct.
Id.
117. In that respect, copyrighted material incorporated into or adopted as a standard
is functional. Cf. Lotus Dev. Corp. v. Borland Int'l, Inc., 49 F.3d 807 (1st Cir. 1995),
aff'd 519 U.S. 233 (1996) (holding that functional menu trees are not copyrightable).
118.There were some interoperability issues, in that related firms used the same
1860 brick structure, but the pharmaceutical companies' primary concern seemed to be
their own investments. Commission Decision, supra note 112, at  120.
BERKELEY TECHNOLOGY LAW JOURNAL
does not arise from any inherent advantages in the intellectual property
owner's product, but from the contributions of its users.
In this context, the parallel to tying law is clear.119
IMS Health was
really offering two products: the 1860 brick structure, and pharmaceutical
data based on it. By denying access to the brick structure, IMS Health ef-
fectively extended its copyright to the pharmaceutical data market, which
could otherwise have been competitive. Moreover, the investments of the
pharmaceutical companies created switching costs of the sort that were
held to contribute to market power in the U.S. Supreme Court's tying de-
cision in Kodak.120
The pharmaceutical companies might have faced in-
formation costs as well, depending on whether they should have antici-
pated IMS Health's hold-up possibilities before choosing to invest in the
1860 brick structure. As discussed above, though, because the law in the
United States and the EC does not clearly cover those hold-up possibili-
ties, it is difficult to charge the companies with anticipating them.
A related point concerns the risk borne by IMS Health in creating the
1860 brick structure. Commentary on the case has argued that if IMS
Health is forced to license the 1860 structure, the royalty should include
compensation for the risk taken by IMS Health.121
But the involvement of
the pharmaceutical companies in the development of the 1860 structure
makes it unclear just how much risk IMS Health bore. This point, too, is a
more general one: where an intellectual property owner develops its in-
vention in parallel with a standard-setting effort, the risk of its efforts is
considerably reduced, so that the justifications for intellectual property
protections are lessened.
IV. NEGOTIATIONS BETWEEN PATENTEES AND
STANDARD-SETTERS
As described above, the relationship between a standard-setting or-
ganization and the patentee of an invention incorporated in a standard is
analogous to the relationship between two blocking patent holders.122
In
contrast to the situation of blocking patents, though, a standard-setting or-
ganization may risk antitrust liability by negotiating with a patent holder.
Generally, it is an antitrust violation for individual competitors to combine
119.See supra notes 30-38 and accompanying text.
120.See Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992).
121. Valentine Korah, The Interface Between Intellectual Property and Antitrust:
The European Experience, 69 ANTITRUST L.J. 801, 826 (2002).
122. See supra Part I.
INVENTIONS, INDUSTRY STANDARDS, AND IP
their negotiating power. In the patent-standard context, though, the under-
lying rationale for that general rule is not applicable.
The members of a standard-setting organization, or at least the organi-
zation itself, should be treated as a single entity when involved in negotia-
tions related to the standard. More specifically, when the goal of the nego-
tiation is to procure a patent license that will enable the practice of the
standard, and when the license will only be valid when it is used with the
standard, the members can be thought of as negotiating for the standard
itself. In such circumstances, the individual members are not pooling their
market shares to gain greater power, but are using the power of the stan-
dard. In that respect, they are acting just as would the owner of an im-
provement patent in a situation in which its use of the patent was blocked
by another patent.
As described above, the standard, though not technically an intellec-
tual property right, functions much like one, in that it requires a consider-
able investment to develop and provides easily duplicable benefits to those
with access to it. Although it may not meet the technical requirements of
patentability,123
it is nevertheless an additional source of market power for
the sellers that produce products in compliance with it.124
For that reason,
it should be treated as a unitary interest, and members of a standard-setting
organization negotiating for it should not be treated as parties to an im-
permissible agreement.125
Viewed from this perspective, the members are
more akin to the licensees of a patent than to competitors.
This view calls into question the allegation in Sony Electronics, Inc. v.
Soundview Technologies, Inc.,126
that the members of a standard-setting
organization had conspired to refuse to purchase a license for a patent that
was needed to comply with a standard. The challenged actions in that case
were coordinated through the standard-setting organization, and they ap-
peared to be directed solely toward enabling compliance with the stan-
dard.127
Hence, although the court refused to dismiss the antitrust claims,
123. Although it might be possible to patent, for example, the use of a particular
technical approach to achieve interoperability, it seems likely that in most circumstances
such a use would be obvious (given the availability of the technical approach itself).
124. See supra note 10.
125. In that respect, they could be treated as having the sort of "unity of interest" that
can prevent formally separate entities from forming an illegal conspiracy for the purposes
of Sherman Anti-Trust Act § 1, 15 U.S.C. § 1 (1994). See Copperweld Corp. v. Inde-
pendence Tube Corp., 467 U.S. 752, 768 (1984).
126. 157 F. Supp. 2d 180 (D. Conn. 2001).
127. See id. at 181-82.
BERKELEY TECHNOLOGY LAW JOURNAL
holding that the patentee had properly alleged a conspiracy to drive down
license fees, the members of the organization might better be viewed as
vindicating the interests of the standard itself, rather than their own inter-
ests independent of the standard.
It is true, though, that if this sort of freedom from normal antitrust
standards were provided,128
it could perhaps be used anticompetitively.
For example, the members of a standard-setting organization, if unsatis-
fied with the terms offered by a patentee, might develop a sham standard
in order to be permitted to negotiate collectively.129
However, this possi-
bility could be addressed by the same method used for determining
whether a patent provides an independent technical contribution to a stan-
dardized product.130
That is, if the purported standard provided no inde-
pendent contribution to demand, then any benefit of adopting the standard
would come from the patented invention. Thus, the standard-setting body
would not merit any antitrust accommodations in the negotiating process.
Although a patentee might choose to test this question through antitrust
litigation, a standard-setting organization could provide itself with some
security by seeking clearance of its negotiating activities from the antitrust
agencies.
V. SUMMARY OF IMPLICATIONS
To summarize, this essay proposes two changes in, or clarifications of,
existing law: an approach to determine licensing fees for patented inven-
tions incorporated in standards and a recognition of the freedom of stan-
dard-setting organizations to negotiate with patentees. Both would likely
reduce the licensing fees to which patentees would be entitled, either in
litigation or in licensing negotiations. Although these proposals might be
thought to reduce undesirably the incentives that patent law is intended to
provide, they probably would not do so in fact. The inventor will still re-
128. This assumes that a deviation from normal antitrust standards would be re-
quired. See supra note 125.
129. Alternatively, they might use licensing negotiations for a standard as an oppor-
tunity to exchange information for anticompetitive purposes. For example, some of the
information gathered by the standard-setting organization in the Soundview litigation
seems at least as useful for anticompetitive collusion as it would be for negotiating a li-
cense. See Sony Elecs., 157 F. Supp. 2d at 193 (noting that the organization's "meeting
minutes also make statements concerning the `revenue streams' of member companies").
However, because this sort of information exchange would not necessarily harm the pat-
entee, it might have no standing to challenge it.
130.See supra Part II.
INVENTIONS, INDUSTRY STANDARDS, AND IP
ceive the return on the technical advances of her invention. Furthermore,
giving the standard-setting organization the right to bargain on behalf of
its members would reduce the transaction costs of the negotiation process.
A. Effects of Reduced License Fees on Incentives of
Inventors
The limits on licensing fees proposed above would deny patentees
some returns to which they might otherwise be entitled.131
This could be
viewed as undesirable. An inventor, when considering whether to engage
in a particular research project, might include in her decision-making cal-
culus the possibility that any invention she creates might be adopted as an
industry standard, in which case she would reap returns from that adop-
tion. Therefore, so the argument would go, to deny the inventor that return
would lessen her incentive to invent. However, this should not cause con-
cern, for a number of reasons.
First, under the approach proposed here, the inventor would not be de-
nied returns derived from her technological contribution, even if her in-
vention was incorporated in a standard. The inventor would only be denied
those returns that derive from standardization, and that do not derive from
her technical innovation. Consequently, to the extent that the inventor re-
lied on the latter returns, there is no reason to think that those returns
would be correlated in any way with the desired incentive for technical
innovation. The law provides patent protection for inventions in the expec-
tation that the costs imposed by the elimination of competition in the sale
of those inventions are balanced by the benefits of increased innovation.
This balance is struck, at least implicitly, by granting the patentee the right
to exclude others from its invention.132
Permitting the patentee to exclude
others not just from its invention but also from others' standardization ef-
forts disturbs this balance.133
Over-investment, always a concern in patent
law,134
becomes a very real possibility, in that patent law would then cre-
ate an incentive for inefficient rent-seeking.
131. See supra Part I.
132. See supra note 32 and accompanying text.
133. See Patterson, supra note 32, at 1138-39.
134. See, e.g., Jennifer F. Reinganum, The Timing of Innovation: Research, Devel-
opment, and Diffusion, in 1 HANDBOOK OF INDUSTRIAL ORGANIZATION 849, 850 (Rich-
ard Schmalensee & Robert D. Willig, eds., 1989) ("The typical outcome of these com-
parisons [between models that compare noncooperative investment in research and de-
velopment with those with cooperative investment or surplus-maximizing results] is that
aggregate expenditure on R&D is too high relative to the cooperative optimum; there are
too many firms and each invests too much").
BERKELEY TECHNOLOGY LAW JOURNAL
Second, those cases that provide guidance in this area suggest that it is
often exactly this sort of rent-seeking that motivates patentees to seek re-
turns from the standardization of their inventions. Inventors seeking adop-
tion of their inventions as industry standards have distorted the standard-
setting process in a variety of ways.135
The effect of this, then, is more
than a distortion of higher prices for the selected standard--the distinction
contemplated by patent law--but also a distortion of the standard-setting
process itself. This effect must be added to the cost side of the balance,
and it strengthens the conclusion that patent protection should not be ex-
tended to the effects of standardization.
Regardless of the theoretical merits of the approach proposed here, it
may appear that the proposal would be difficult to apply in practice. But
this proposal's approach is not appreciably more difficult than other in-
quiries in patent law, such as determining the source of demand for an in-
fringing product when deciding the damages that will be awarded. When-
ever damages must be calculated in a patent case, such factors as the exis-
tence and significance of noninfringing alternatives136
must be considered.
Accounting for these factors presents difficulties that are similar both
qualitatively and quantitatively to the proposal here.
Moreover, adopting this proposal might itself reduce the need for its
application. Since patentees will be unable to rely on extracting licensing
revenue deriving from standardization, they have less incentive to engage
in the rent-seeking conduct referred to above, and they might negotiate
licensing arrangements more readily. Because no clear rule currently es-
tablishes the allocation of entitlements to profits deriving from standardi-
zation, parties can take broadly different positions in negotiations. By es-
tablishing the principle that a patentee is entitled only to revenues from its
135. As discussed in note 16 supra, the F.T.C. brought an enforcement action against
Dell Computer Corp., alleging that Dell kept the existence of a patent secret while a stan-
dard-setting organization considered a standard based on the invention claimed in Dell's
patent. Rambus, Inc., as noted earlier, and Sun Microsystems, Inc., may also have en-
gaged in similar behavior. See supra note 10 and accompanying text; see also Bridis,
supra note 9 (reporting that the F.T.C. has commenced an investigation into whether
Rambus and Sun encouraged standard-setting organizations to adopt standards covered
by their patents, which they did not disclose). Another example of questionable behavior
in this context is the use of insiders at standard-setting meetings to gather information
about the progress of the standard-setting process. See Tony Smith, Rambus Received
Leaked JEDEC SDRAM Data, THE REGISTER (June 4, 2001), at http://www.theregister.-
co.uk/content/archive/18148.html.
136. See, e.g., Polaroid Corp. v. Eastman Kodak Co., No. 76-1634-MA 1990 WL
324105 at *3 (D. Mass. Oct. 12, 1990).
INVENTIONS, INDUSTRY STANDARDS, AND IP
technical contribution, the range of disagreement is narrowed, and a nego-
tiated settlement made more likely.
B. Effects of Granting License-Negotiating Power
to Standard-Setting Organizations
As discussed above, giving standard-setting organizations the right to
negotiate on behalf of their members could improve the negotiation proc-
ess. However, this approach might provide standard-setting organizations
with so much power that they would negotiate patent licensing fees that
were too low to provide sufficient incentive for innovation. It is theoreti-
cally possible that a standard-setting organization could force a patent li-
censing fee down not just to the value of the invention's technical contri-
bution--which would be desirable137
--but below its value. However, it is
equally possible that the patentee, who is after all in a similar monopoly
position, could demand fees higher than those justified by its contribu-
tion.138
In fact, Merges and Nelson argue that a likely outcome in the simi-
lar context of original and improvement patents (analogous to a patent and
a standard)139
is that the original patentee will extract half the value of the
improvement (standard).140
Ultimately, the result is indeterminate, as
would be expected with a bilateral monopoly. It is clear, though, that cur-
rent rules allow patentees to threaten discriminatorily large fees from some
standards users.141
By eliminating that market failure, this proposal will
likely to provide better correspondence between financial returns and in-
novative contributions.
VI. CONCLUSION
The approach presented in this essay provides a means of efficiently
and fairly determining what licensing revenue is due a patentee when an
industry standard incorporates its patent. The proposed approach puts an
emphasis on distinguishing between demand for the invention and demand
for the standard, and it would allocate to the patentee only those revenues
derived from its own innovative contribution. In some cases there may be
evidence available to determine what portion of the demand for a stan-
dardized product is due to that contribution. In other cases, there may be
137. See supra Part I.A.
138. See, e.g., Rambus, Inc. v. Infineon Techs. AG, 155 F. Supp. 2d 668 (E.D. Va.
2001).
139. See supra Part I.C.
140. See Merges & Nelson, supra note 35, at 866 n.117.
141. See Smith, supra note 135.
BERKELEY TECHNOLOGY LAW JOURNAL
no such direct evidence, but it may still be possible to use indirect evi-
dence to draw inferences regarding the contributions of the patentee. Al-
though in some instances this analysis will be inconclusive, even in those
instances it clarifies the issues to be resolved. Furthermore, by giving
standard-setting organizations the power to negotiate on behalf of their
members, the equilibrium in the patent system is restored.