18 Berkeley Tech LJ 853

DATE DOWNLOADED: Sun Apr 9 13:39:11 2023
SOURCE: Content Downloaded from HeinOnline
Citations:
Please note: citations are provided as a general guideline. Users should consult their preferred
citation format's style manual for proper citation formatting.
Bluebook 21st ed.

Colleen Chien, Cheap Drugs at What Price to Innovation: Does the Compulsory Licensing
of Pharmaceuticals Hurt Innovation, 18 BERKELEY TECH. L.J. 853 (2003).
ALWD 7th ed.

Colleen Chien, Cheap Drugs at What Price to Innovation: Does the Compulsory Licensing
of Pharmaceuticals Hurt Innovation, 18 Berkeley Tech. L.J. 853 (2003).
APA 7th ed.

Chien, C. (2003). Cheap drugs at what price to innovation: does the compulsory
licensing of pharmaceuticals hurt innovation. Berkeley Technology Law Journal, 18(3),
853-908.
Chicago 17th ed.

Colleen Chien, "Cheap Drugs at What Price to Innovation: Does the Compulsory
Licensing of Pharmaceuticals Hurt Innovation," Berkeley Technology Law Journal 18,
no. 3 (2003): 853-908
McGill Guide 9th ed.

Colleen Chien, "Cheap Drugs at What Price to Innovation: Does the Compulsory
Licensing of Pharmaceuticals Hurt Innovation" (2003) 18:3 Berkeley Tech LJ 853.
AGLC 4th ed.

Colleen Chien, 'Cheap Drugs at What Price to Innovation: Does the Compulsory
Licensing of Pharmaceuticals Hurt Innovation' (2003) 18(3) Berkeley Technology Law
Journal 853
MLA 9th ed.

Chien, Colleen. "Cheap Drugs at What Price to Innovation: Does the Compulsory
Licensing of Pharmaceuticals Hurt Innovation." Berkeley Technology Law Journal, vol.
18, no. 3, 2003, pp. 853-908. HeinOnline.
OSCOLA 4th ed.

Colleen Chien, 'Cheap Drugs at What Price to Innovation: Does the Compulsory
Licensing of Pharmaceuticals Hurt Innovation' (2003) 18 Berkeley Tech LJ 853
Please note: citations are provided as a general guideline. Users should consult
their preferred citation format's style manual for proper citation formatting.
-- Your use of this HeinOnline PDF indicates your acceptance of HeinOnline's Terms and
Conditions of the license agreement available at
https://heinonline.org/HOL/License
-- The search text of this PDF is generated from uncorrected OCR text.
-- To obtain permission to use this article beyond the scope of your license, please use:
Copyright Information
CHEAP DRUGS AT WHAT PRICE TO INNOVATION:
DOES THE COMPULSORY LICENSING OF
PHARMACEUTICALS HURT INNOVATION?
By Colleen Chient
ABSTRACT
The patent system is built on the premise that patents provide an in-
centive for innovation by offering a limited monopoly to patentees. The
inverse assumption that removing patent protection will hurt innovation
has largely prevented the widespread use of compulsory licensing-the
practice of allowing third parties to use patented inventions without pat-
entee permission. In this Article, I empirically test this assumption. I
compare rates of patenting and other measures of inventive activity be-
fore and after six compulsory licenses over drug patents issued in the
1980s and 1990s. As reported below, I observe no uniform decline in in-
novation by companies affected by compulsory licenses and find very lit-
tle evidence of a negative impact, which is consistent with earlier empiri-
cal work. While anecdotal, these findings suggest that the assertion that
licensing categorically harms innovation is probably wrong. Based on the
data, I comment on the use of compulsory licensing to reduce the price of
AIDS and other drugs for developing countries. I suggest that, based on
past experience, compulsory licenses need not result in a decline in inno-
vation and that this policy option for increasing access to medicines de-
serves greater exploration.
TABLE OF CONTENTS
1. IN TRO DUCTION ....................................................................................................... 855
II. COMPULSORY LICENSING OVERVIEW ..................................................................... 857
A. General O verview ......................................................................................... 858
B. United States versus Developing Country Perspectives on Compulsory
Licen sin g ....................................................................................................... 860
C. Compulsory Licensing in the United States .................................................. 862
© 2003 Colleen Chien

t Associate, Fenwick & West; J.D., University of California at Berkeley School of
Law (Boalt Hall), 2002; B.S. & A.B., Stanford University, 1996. I would like to thank
Mark Lemley, Fred Abbott, Jenny Lanjouw, Sasha Blaug, and Dirk Calcoen for their
helpful comments and support. In addition, I would like to acknowledge Biospace Inc.,
for generously providing access to its Clinical Competitive Intelligence System clinical
trials database. An earlier version of this paper was awarded Grand Prize in the 2002
Third Annual Foley & Lardner Intellectual Property Competition.
854 BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 18:853
III. THE COMPULSORY LICENSING OF DRUGS ............................................................... 864

A. Patents and Drug Innovation ......................................................................... 864
B. Compulsory Licensing of Drugs in the United States ................................... 866
C. Compulsory Licensing of Drugs under TRIPS .............................................. 869
IV. EMPIRICAL BACKGROUND (LITERATURE REVIEW) ................................................. 872
A. Compulsory License Design ......................................................................... 873
B. Literature Review .......................................................................................... 874
1. Compulsory Licensing under US. Antitrust Consent Decrees............... 875
2. Compulsory Licensing of D rugs in Canada........................................... 876
3. Licensing of U.S. Government Inventions .............................................. 877
4. HypotheticalLicensing ofAll PharmaceuticalInventions..................... 879
C. Summ ary of Results ...................................................................................... 879
V. CASE STUDIES OF INVESTMENT IN INNOVATION AFTER Six ANTITRUST
CONSENT DECREES ................................................................................................. 880
A . The Antitrust Drug Licenses ......................................................................... 881
B. M easurement of the Impact of Drug Licenses ............................................... 883
C. Impact of Compulsory Licensing .................................................................. 885
1. Sporadic Licensing................................................................................. 885
2. PredictableLicensing............................................................................. 887
D . Results ........................................................................................................... 891
VI. IMPLICATIONS FOR DRUG LICENSING IN DEVELOPING COUNTRIES ........................ 892
VII. CONCLUSION .......................................................................................................... 896
VIII. APPENDIX: ANTITRUST LICENSE CASE STUDIES ..................................................... 897
A . Baxter/Fibrin Sealant ..................................................................................... 897
1. The Order............................................................................................... 897
2. Impact on Innovation ............................................................................. 898
B. M arion Merrell Dow/Dicyclom ine ................................................................ 899
1. The Order............................................................................................... 899
2. Subsequent D evelopments ...................................................................... 900
C. Eli Lilly/Insulin ............................................................................................. 901
1. The Order............................................................................................... 901
D. Connaught/Rabies Vaccine ........................................................................... 903
1. The Order............................................................................................... 903
E. Chiron/HSV-tk Related Therapeutics ............................................................ 905
1. The Order........................................................... 905
F. Roche/CD-4 ................................................................................................... 906
1. The Order............................................................................................... 906
20031 CHEAP DRUGS AT WHAT PRICE TO INNOVATION
I. INTRODUCTION
The international AIDS crisis has posed an acute challenge to the ro-
bustness of the patent system. Critics contend that the basic bargain be-
tween patentees and the public, namely innovation in exchange for a lim-
ited monopoly, is irreparably skewed in favor of drug companies.' De-
fenders of strong patent rights, on the other hand, insist that any weaken-
ing of existing protections would undermine the potential for future inno-
vation.2
Compulsory licensing, the practice of authorizing a third par 7 to
make, use, or sell a patented invention without the patentee's consent, has
long provided an antidote to the perceived ills of the patent system.4 In the
context of the AIDS crisis, compulsory licensing offers one way to lower
drug prices and increase access to patented medicines in developing coun-
tries in which pharmaceuticals have chosen to secure patent protection and
the markets supplied by these countries. 5 Under the Agreement on Trade-
1. See, e.g., PASCALE BOULIN ET AL., MtDECINS SANS FRONTItRES, DRUG PAT-
ENTS UNDER THE SPOTLIGHT: SHARING PRACTICAL KNOWLEDGE ABOUT PHARMACEUTI-
CAL PATENTS 2 (2003) ("Patents are not god-given rights. They are tools invented to
benefit society as a whole, not to line the pockets of a handful of multinational pharma-
ceutical companies."), available at http://www.accessmed-msf.org/documents/patents_
2003.pdf (last visited July 19, 2003); Larry Elliot, Evil Triumphs in a Sick Society,
GUARDIAN, Feb. 12, 2001 (criticizing global patent law for favoring large pharmaceutical
companies), availableat 2001 WL 11917250.
2. See, e.g., Gregory J. Glover, Statement on Behalf of Pharmaceutical Research
and Manufacturers of America Before the Federal Trade Commission and the Depart-
ment of Justice-Antitrust Division, Competition in the Pharmaceutical Marketplace 6
(Mar. 19, 2002) ("[C]ompanies would not be able to invest the huge amount of time and
money it takes to discover and develop a new medicine if they did not have a sufficient
opportunity to make a sufficient return before generic competitors copy and market the
drug at greatly reduced cost."), available at http://www.ftc.gov/opp/intellect/020319
gregoryjglover.pdf, Richard Tren, Free Industry, Not the Drugs, WALL ST. J. EUR.,
July 11, 2002, at A10.
3. F.M. SCHERER & JAYASHREE WATAL, POST-TRIPS OPTIONS FOR ACCESS TO
PATENTED MEDICINES IN DEVELOPING COUNTRIES 13 (Comm'n on Macroeconomics &
Health, Working Paper No. WG4:1, 2001), available at http://www.cmhealth.org/docs/
wg4_paperl .pdf (last visited Aug. 27, 2003).
4. Compulsory licensing was a component of a late nineteenth-century English
patent reform bill. See Fritz Machlup & Edith Penrose, The Patent Controversy in the
Nineteenth Century, 10 J. ECON. HIST. 1,4 (1950). The United States instituted a compul-
sory licensing provision as early as 1910. Act of June 25, 1910, ch. 423, 36 Stat. 851, 853
(current version at 28 U.S.C. § 1498).
5. Other options include price regulation and improved health infrastructure. See,
e.g., Amir Attaran & Lee Gillespie-White, Do Patentsfor AntiretroviralDrugs Constrain
Access to AIDS Treatment in Africa, 286 JAMA 1886, 1890 (2001) (noting that numerous
drugs are not patented or are off-patent in a number of developing countries, arguing that
BERKELEY TECHNOLOGY LAW JOURNAL [Vol. 18:853
Related Aspects of Intellectual Property Rights ("TRIPS"), 6 compulsory

licensing is authorized under certain circumstances, such as public health
emergencies. However, until recently, few compulsory licenses had been
actually issued under TRIPS. 7 One of the most important reasons for this,
and the one this Article focuses on, is the perception that compulsory li-
censes harm the incentive for innovation. In the words of one pharmaceu-
tical executive: "[T]hreatening compulsory licensing ...will only act as
[a] disincentive[] to the development and marketing of new drugs."8 The
twin goals of increasing access to existing medicines and promoting re-
search and development of new medicines have been portrayed as compet-
ing with each other.
This Article questions this fundamental assumption. It explores
whether past compulsory licenses over drugs have been accompanied by a
reduction in innovation, drawing upon past research efforts and the results
of an empirical analysis that I performed on six cases of compulsory drug
licenses issued in the United States by the Department of Justice ("DOJ")
in the 1980s and 1990s. The analysis compares rates of innovation within
a therapeutic area, measured by patent counts and other indicia, before and
after compulsory licenses were issued.
the absence of patent protection neither guarantees nor increases access to drugs, and
suggesting that factors such as political will and poverty levels restrict access more than
patent protection does); Tobias Buck, EU Acts to Speed up Flow of Cheap AIDS Drugs,
FIN. TIMES, May 27, 2003, at 6 (describing a proposal by the European Union to cap the
price of AIDS, malaria, and tuberculosis drugs sold to developing countries).
6. Agreement on Trade-Related Aspects of Intellectual Property Rights, Apr. 15,
1994, Marrakesh Agreement Establishing the World Trade Organization [hereinafter
WTO Agreement], Annex IC, LEGAL INSTRUMENTS-RESULTS OF THE URUGUAY
ROUND, vol. 31, 33 I.L.M. 81 (1994) [hereinafter TRIPS Agreement].
7. Nonetheless, the threat of compulsory licensing under TRIPS arguably encour-
ages pharmaceutical companies to voluntarily reduce prices. See JEROME H. REICHMAN &
CATHERINE HASENZAHL, NON-VOLUNTARY LICENSING OF PATENTED INVENTIONS: HIs-
TORICAL PERSPECTIVE, LEGAL FRAMEWORK UNDER TRIPS, AND AN OVERVIEW OF THE
PRACTICE IN CANADA AND THE UNITED STATES OF AMERICA 13 nn.92-93 (2002) (de-
scribing the experiences of Brazil and the United States in using the threat of compulsory
licensing to drive down the prices of AIDS drugs), available at http://www.ictsd.org/
iprsonline/unctadictsd/docs/reichmanhasenzahl.pdf (last visited July 19, 2003); Patent
Remedies, ECONOMIST, Oct. 25, 2001 (stating that the U.S. Department of Health used
the specter of domestic compulsory licensing to obtain a half-price discount on Cipro
from Bayer), 2001 WL 7320684; Tina Rosenberg, Look at Brazil, N.Y. TIMES, Jan. 28,
2001 (detailing how Brazil effectively used the threat of compulsory licensing to leverage
discounted prices on AIDS drugs), http://www.nytimes.com/library/magazine/home/200
10128mag-aids.html (last visited July 19, 2003).
8. See Tren, supra note 2, at A10.
In five of the six cases I studied, I observed no measurable decline in

innovation. This finding is consistent with earlier work. By available
measures, the companies affected by licenses continued to perform re-
search and development ("R&D") in the therapeutic areas targeted by the
license. Even in the case of forward-looking compulsory licenses that
spanned several years, the decline in R&D that advocates for strong patent
rights might predict was not observed. While limited and anecdotal, this
and past work suggest that concerns about compulsory licensing are over-
stated and that the blanket assertion that licensing categorically harms in-
novation is probably wrong.
This Article also discusses how the structure and implementation of
compulsory licenses affects R&D. Based on past research and common
sense, I postulated that two factors are extremely important-the degree to
which a company can predict that a compulsory license will be taken on a
patent ("predictability") and the relative importance of the markets af-
fected by the license ("importance"). In the six cases analyzed, licenses
that were either unpredictable or did not affect important markets had no
discernable impact on R&D, all other things being equal. In all cases but
one, the license was either unpredictable or did not impact a developed,
existing product market. I observed no reduction in R&D activity in these
cases. However, in the one case where licensing was both predictable and
impacted a developed market for a drug, there was some evidence of a de-
cline in R&D. Although too few in number to be conclusive, these cases
and earlier work provide hope that compulsory licensing need not discour-
age innovation. They also underscore that the manner in which compul-
sory licenses are structured and implemented matters, and suggest that the
factors of predictability and market impact deserve special attention.
Part II of this Article provides an overview of compulsory licensing.
Part III explores the role of patents in pharmaceutical innovation and dis-
cusses the compulsory licensing of drugs. Part IV discusses the existing
literature on the impact of compulsory licensing on innovation. Part V re-
ports the results of empirical analyses performed on six case studies of
pharmaceutical compulsory licensing. Part VI discusses the implications
of these results for policymaking.
II. COMPULSORY LICENSING OVERVIEW

Compulsory licenses are generally defined as "authorizations permit-
ting a third party to make, use, or sell a patented invention without the pat-
ent owner's consent." 9 Because they limit the power conferred by patents,
compulsory licenses have long been controversial.' 0 This part briefly re-
views the origins of compulsory licenses, the arguments for and against
them in both the United States and developing countries, and the record of
their implementation in the United States.
A. General Overview
The current debate over compulsory licensing is nothing new. In the
United States Senate in 1790, in the House of Lords in Britain in 1851,
and in Germany in 1853,11 policy makers discussed compulsory licensing
as a way to preserve the benefits of the patent system while minimizing its
evils. On the one hand, patents created positive incentives for innovation
and the disclosure of inventions, granted "just rewards" to inventors, dem-
onstrated society's recognition of the "natural" property rights
of inventors, and generally addressed the public goods problems associ-
ated with creation of knowledge.' 2 On the other hand, these benefits came
at a cost, including the potential abuse of monopoly power by patentees,
the use of patents to block inventive activity by third parties, the diversion
of productive activity disproportionately towards patentable activity, and
the substantial administrative costs of operating a patent system. 13
With these benefits and costs in mind, patent critics and advocates ac-
cepted compulsory licensing as a "strategic compromise" in 1873 at the
Patent Congress in Vienna.' 4 In order to preserve the incentive for innova-
9. See SCHERER & WATAL, supra note 3, at 12.

10. Although this Article focuses on compulsory licenses in the patent context, these
licenses also arise in the context of other intellectual property, such as copyrights. See
ROBERT A. GORMAN & JANE C. GINSBURG, COPYRIGHT 498-505 (6th ed. 2002) (describ-
ing the introduction of compulsory licensing into U.S. copyright law in 1909 and discuss-
ing 17 U.S.C. § 115, which permits the taking of licenses to publicly distributed phonore-
cords without the permission of the copyright holder).
11. See FRITZ MACHLUP, STAFF OF SUBCOMM. ON PATENTS, TRADEMARKS, AND
COPYRIGHTS OF THE SENATE COMM. ON THE JUDICIARY, 85TH CONG., AN ECONOMIC RE-
VIEW OF THE PATENT SYSTEM 5 (Comm. Print 1958) [hereinafter ECONOMIC REVIEW OF
THE PATENT SYSTEM].
12. See, e.g., Machlup & Penrose, supra note 4, at 10-11.
13. See, e.g., Michael A. Heller & Rebecca S. Eisenberg, Can Patents Deter Innova-
tion? The Anticommons in Biomedical Research, 280 SCI. 698 (1998) (describing the
problem of "blocking" patents in biomedical research); Machlup & Penrose, supra note
4, at 23-28 (discussing general critiques of the patent system); James Love, Paying for
Health CareR&D: Carrots and Sticks, Consumer Project on Technology, Oct. 19, 2000,
(articulating perceived abuses of the patent system), at http://www.cptech.org/ip/health/
md/carrotsnsticks.html (last visited Aug. 27, 2003).
14. A battle occurred between the anti-patent movement of the 1850s through 1870s
and the patent advocates of the 1870s through 1910s: "The strategic compromise was the
tion while increasing access to innovations themselves, the Congress

adopted a requirement that licensees pay patent holders reasonable com-
pensation for their licenses.' 5 With the subsequent adoption of compulsory
licensing by the 1883 Paris Convention,' 6 the world's foremost interna-
tional patent agreement, compulsory licensing became a fixture in almost
all patent systems.17
While specific provisions vary, compulsory licenses are generally au-
thorized in the event of undesirable behavior by the patentee, such as anti-
competitive, non-working, or blocking behavior; in the event of "public
need," such as government infringement or national emergency; or in the
context of food and drugs.' 8 Licensees are commonly required to pay ade-
quate compensation to a patentee in exchange for use of a patent. The re-
quired amount is generally more than a "reasonable royalty," the floor for
infringement compensation in the United States,' 9 but less than "lost prof-
its," another basis for calculating infringement damages. 20 The amount of
compensation varies among countries; commentators have observed that
"the United Kingdom has provided the most generous compensation in its
drug patent licensing decisions; the United
21
States the least generous com-
pensation in key antitrust decisions.",
acceptance of the principle of compulsory licensing--of compelling all patentees to li-

cense others to use the invention at reasonable compensation . ... The patent advocates
and the free traders compromised on this general limitation on the patentees' monopoly
power." See ECONOMIC REVIEW OF THE PATENT SYSTEM, supra note 11, at 5.
15. Id.
16. See Paris Convention for Protection of Industrial Property, Mar. 20, 1883, art.
5A, 21 U.S.T. 1583 (last revised at Stockholm, July 14, 1967).
17. As of Feb. 13, 2002, 163 states, including most industrialized countries, had
ratified the Convention. See World Intellectual Property Organization [hereinafter
WIPO], ParisConvention for the Protectionof IndustrialProperty,at http://www.wipo.
intltreaties/documents/english/word/d-paris.doc (last visited July 18, 2003).
18. See, e.g., FREDERICK M. ABBOTT, COMPULSORY LICENSING FOR PUBLIC HEALTH
NEEDS: THE TRIPS AGENDA AT THE WTO AFTER THE DOHA DECLARATION ON PUBLIC
HEALTH (Quaker United Nations Office, Occasional Paper No. 9, 2002); Gianna Julian-
Arnold, InternationalCompulsory Licensing: The Rationales and the Reality, 33 IDEA
349, 349-55 (1993).
19. 35 U.S.C. § 284 (2000) (stating that the damages for patent infringement "shall
[be] ... adequate to compensate for the infringement, but in no event less than a reason-
able royalty for the use made of the invention by the infringer").
20. ROBERT P. MERGES, PATENT LAW AND POLICY 1038-84 (2d ed. 1997).
21. SCHERER & WATAL, supra note 3, at 28.
B. United States versus Developing Country Perspectives on

Compulsory Licensing
Within the general framework of compulsory licensing, however, there
has been little consensus on how best to implement it. In modem times,
nowhere has the divergence in views been more pronounced than in the
context of the compulsory licensing provisions of TRIPS. This was par-
ticularly evident during the negotiations behind these provisions. While
the United States viewed these provisions with distrust and suspicion, de-
veloping countries claimed them to be an essential part of a workable pat-
ent system.2 2 Commentators have noted that the resulting provisions, dis-
cussed below, were left intentionally vague, reflecting the parties' inability
to come to an agreement.23
The contrast in views on patents between the United States and devel-
oping countries is driven in part by differences in economic status. In de-
veloping countries, foreigners file most of the patents.24 As a result, the
22. During TRIPS negotiations in 1989, the U.S. representative characterized com-
pulsory licensing as prone to "mischievous use," and favored a more restrictive, excep-
tional regime in which licensing would be permitted only for "legitimate purposes." Note
by the Secretariat, Meeting of Negotiating Group of 12-14 July 1989, 14.doc, 83.2,
available at WTO, http://www.wto.org/english/tratope/trips e/tripse.htm (download
derestricted official document archive under heading History: Derestricted Uruguay
Round NegotiatingDocuments on TRIPS) (last visited Aug. 20, 2003). In contrast, India's
representative stated that compulsory licensing should be viewed as a means for balanc-
ing the rights and obligations of patent holders; compulsory licenses should not be nar-
rowly circumscribed, particularly since they are vital to the transfer of technology. Id.
83.3. This difference in views led to competing draft legislation in 1990. The version
supported by the United States and other developed nations narrowly defined the basis
for licenses, whereas the version supported by developing countries was much more
open-ended. See generally ABBOTT, supra note 18.
23. See JEAN 0. LANJOUw, INTELLECTUAL PROPERTY AND THE AVAILABILITY OF
PHARMACEUTICALS IN POOR COUNTRIEs 25 (Ctr. for Global Dev., Working Paper No. 5,
2002). But see REICHMAN & HASENZAHL, supra note 7, at 12-13 (stating that the result-
ing language ultimately vindicated the stance of developing countries over that of the
United States).
24. Consider, for example, Brazil and South Africa. These are two developing coun-
tries against which U.S. government and industry have initiated significant patent dis-
putes over compulsory licensing. Brazil held less than 0.1% of the U.S. patents issued in
1998, while the United States captured nearly 40% of the patents issued in Brazil that
same year. In South Africa, foreigners applied for over 99% of the patents in 1999 (is-
sued patent data is not available), and 40% of those applications were from the United
States. In contrast, South African inventors captured less than 0.1% of U.S. patents issued
in 1998. See 1 NAT'L Sci. BD., SCIENCE & ENGINEERING INDICATORS-2002, source data
for 6-21 fig.6-23, source data for 6-25 fig.6-27 [hereinafter NAT'L SCI. BD. SOURCE
DATA], at http://www.nsf.gov/sbe/srs/seind02/pdf/volumel.pdf (last visited Aug. 24,
2003) (source data for fig.6-23, at http://www.nsf.gov/sbe/srs/seindO2/c6/figO6-23.xls;
2003] CHEAP DRUGS AT WHAT PRICE TO INNOVATION
patent system facilitates the transfer of monopoly rents to foreigners out-

side the country, although it is also true that companies may choose not to25
patent inventions in markets they regard as too small to be significant.
Furthermore, the high price of products covered by patents can put needed
technology out of the reach of developing country consumers, who are
generally required to pay for drugs out of pocket due to the lack of health-
care infrastructure. 26 To compensate for these patent system costs, permis-
sive compulsory licenses are used to widen distribution of and increase
access to patented technologies. The situation is different in the United
States since U.S. inventors capture a large share of patents both domesti-
27
cally and abroad. Patent profits from both domestic and international
markets reward and support research performed locally by U.S. inventors.
Another basic reason for the difference in perspectives derives from
the rationales behind each country's patent system. Generally, countries
with relatively few patents view the patent system as a means to promote
the transfer of technology from other countries. 2 8 Compulsory licensing
provides an important safeguard to ensure that technology transfer hap-
pens in the event of non-working or high prices. In contrast, countries such
source data for fig.6-27, at http://www.nsf.gov/sbe/srs/seind02/c6/figO6-27.xls); U.S.

PAT. & TRADEMARK OFF., PATENTING TRENDS CALENDAR YEAR 1999, at http://www.
uspto.gov/web/offices/ac/ido/oeip/taf/pattr99.htm (last visited Aug. 24, 2003); WIPO,
WIPO INDUSTRIAL PROPERTY STATISTICS 1999, WIPO Doc. IP/STAT/1999/B (1999), at
http://www.wipo.org/ipstats/enlpublications/b/1999/i/pattabl.pdf (last visited Aug. 24,
2003).
25. See Attaran & Gillespie-White, supra note 5, at 1890.
26. See WORLD HEALTH ORG. & WORLD TRADE ORG. SECRETARIATS, REPORT OF
THE WORKSHOP ON DIFFERENTIAL PRICING AND FINANCING OF ESSENTIAL DRUG [sic] 6
(2001) (reporting that 90% of the population in developing countries buys medicines out-
of-pocket, whereas only 20% of the population does so in high income countries), at
http://www.who.int/medicines/library/edm-general/who-wto-hosbjor/wholereporthosbjo
rworkshop-fin-eng.pdf (last visited Aug. 24, 2003).
27. In the United States, a thin majority (54%) of patents were granted to U.S. resi-
dents in 2001. U.S. PAT. & TRADEMARK OFF., 2001 PERFORMANCE AND ACCOUNTABIL-
ITY REPORT 115, 118 (2001), at http://www.uspto.gov/web/offices/com/annual/2001/01
performreport.pdf. Most of the remaining patents were granted to inventors from devel-
oped countries. Id. Because U.S. inventors capture an extensive share of the patents in
these developed countries, the costs of foreign patenting are counterbalanced by the bene-
fits to U.S. inventors of obtaining patents abroad. In 1998, for instance, the United States
captured 45%, 28%, 28%, and 30.4% of the patents awarded to foreigners in Japan, Ger-
many, France, and the United Kingdom, respectively, while the same countries each cap-
tured 46%, 14%, 5%, and 5%, respectively, of the patents awarded by the United States
to residents of foreign countries. See NAT'L SCI. BD. SOURCE DATA, supra note 24.
28. See Edith Penrose, InternationalPatenting and the Less-Developed Countries,
83 ECON. J. 768, 771 (1973).
as the United States claim a relatively large share of the world's patents
and look to the patent system primarily as an incentive to innovate and a
means to stimulate technology creation. 29 This innovation-based focus
leads to the selective application of compulsory licensing to cases where
patents hinder rather than advance innovation.
C. Compulsory Licensing in the United States
Consistent with a focus on innovation, the U.S. government has used
compulsory licenses to curb anti-competitive behavior. 30 By 1977, the
Federal Trade Commission ("FTC") and DOJ had issued approximately 31
125 decrees over thousands of patents and a wide range of technology.
Recently, such decrees have been ordered in the context of mergers, price-
fixing, and the abuse of monopoly or market power. 32 Compulsory licens-
ing has also been proposed as a solution to the problem of patent thickets,
wherein broad or multiple patents over technology areas prevent follow-on
research. Voluntary or compulsory patent pools, in which the rights to use
multiple patents are exchanged among patentees have been proposed as a
way to overcome the refusal of patentees to license an invention and the
administrative burden associated with licensing.33
However, compulsory licensing has also been used to further public
interests, primarily by enabling the U.S. government to use patented in-
ventions without permission. Although courts have emphatically resisted
issuing compulsory licenses merely because a patentee chooses not to use
29. See Clarisa Long, Patents and Cumulative Innovation, 2 WASH. U. J.L. & POL'Y
229, 231 (2000).
30. See, e.g., United States v. Nat'l Lead Co., 332 U.S. 319 (1947); Hartford-Empire
Co. v. United States, 323 U.S. 386, 417 (1945); see also Arti K. Rai & Rebecca S.
Eisenberg, Bayh-Dole Reform and the Progress of Biomedicine, 66 LAW & CONTEMP.
PROBS. 289, 297 n.46 (2003) (noting the role of the U.S. military in ensuring cross-
licenses between the Wright Brothers and follow-on innovators).
31. See F.M. SCHERER, THE ECONOMIC EFFECTS OF COMPULSORY PATENT LICENS-
ING 47-48 (1977).
32. See, e.g., Compulsory Licensing as Remedy to Anticompetitive Practices,Con-
sumer Project on Technology, at http://www.cptech.org/ip/health/cl/us-at.html (last vis-
ited July 18, 2003) (reporting that of twenty-five compulsory licenses issued since the
mid-90s, roughly half resulted from mergers and acquisitions, while the remainder re-
sulted from other forms of anticompetitive behavior).
33. See generally JEANNE CLARK ET AL., U.S. PAT. & TRADEMARK OFF., PATENT
POOLS: A SOLUTION TO THE PROBLEM OF ACCESS IN BIOTECHNOLOGY PATENTS? 8-12
(2000) (discussing the use of patent pools as a solution to the problems associated with
biotechnology patents), available at http://www.uspto.gov/web/offices/pac/dapp/opla/
patentpool.pdf (last visited July 19, 2003).
her invention,34 the U.S. government routinely relies on 28 U.S.C. § 1498

to immunize its use of inventions without the patentee's permission. The
statute limits a patentee's remedy for infringement by the government or a
government contractor to "reasonable and entire compensation." 35 By not
allowing for injunctive relief, the statute effectively strips patentees of the
right to prevent others from using their inventions.
Although the statute was originally conceived with wartime urgency in
mind, 36 the government has used it in a wide range of circumstances.
Since 1948, the year of the statute's enactment in its current form,37 the
Court of Federal Claims and its predecessors have decided almost 300
cases, involving a wide variety of technologies, under § 1498.38 Although
this figure is surprisingly large, it arguably understates the use of compul-
sory licenses by the government because it excludes cases resolved with-
out litigation and infringement that goes unnoticed by the patentee. In in-
fringement suits against the government that have been decided on the
34. See, e.g., Cont'l Paper Bag Co. v. E. Paper Bag Co., 210 U.S. 405, 429 (1908)
(holding that "it is the privilege of any owner of property to use or not to use it, without
question of motive"); see also 35 U.S.C. § 271(d)(4) (2000) (confirming by amendment
under the Patent Misuse Reform Act of 1988 that the refusal to license or use one's pat-
ents rights does not by itself constitute misuse for which compulsory licensing would be a
remedy).
35. 28 U.S.C. § 1498(a). The subsection states:
Whenever an invention described in and covered by a patent of the
United States is used or manufactured by or for the United States with-
out license of the owner thereof or lawful right to use or manufacture
the same, the owner's remedy shall be by action against the United
States in the United States Court of Federal Claims for the recovery of
his reasonable and entire compensation for such use and manufacture
Id.(emphasis added).
36. In Richmond Screw v. United States, the Supreme Court commented about the
statute:
The intention and purpose of Congress ...was to stimulate contractors
to furnish what was needed for the war, without fear of becoming liable
themselves for infringements to inventors or the owners or assignees of
patents ....To accomplish this governmental purpose, Congress exer-
cised the power to take away the right of the owner of the patent to re-
cover from the contractor for infringements.
275 U.S. 331, 345 (1928).
37. See Lionel Marks Lavenue, Patent Infringement Against the United States and
Government Contractors Under 28 U.S. C. § 1498 in the United States Court of Federal
Claims, 2 J. INTELL. PROP. L. 389, 415 (1995).
38. Id. at 496 n.563 (noting that there have been 240 cases from 1949 to Apr. 1,
1994); LEXIS search, Genfed Library, FED File (Apr. 2, 1994 through Aug. 10, 2002)
using search terms "28 U.S.C. § 498", "government", and "patent."
merits, plaintiff patentees have won just over one-third of the time,39 as
compared to a 58% success rate of patentees against accused infringers in
general. 4 ° Outside the context of section 1498, compulsory licenses have
been authorized for public policy reasons, but on a more limited scale.4 '
III. THE COMPULSORY LICENSING OF DRUGS

Against the backdrop of compulsory licensing, this part discusses the
role of patents in drug innovation and the compulsory licensing of drugs
both in the United States and under TRIPS.
A. Patents and Drug Innovation
Drugs have been singled out for special treatment both in terms of pat-
enting and compulsory licensing primarily because of their role in promot-
ing public health. For many years product patents were not awarded over
pharmaceuticals. In the developed world, Japan did not introduce product
patents for drugs until 1976, and pharmaceutical powerhouse Switzerland 42
waited until 1977 to introduce patents covering pharmaceutical products.
Spain, Portugal, Greece, and Norway introduced product patents over
drugs as recently as 1992. 4 3 At the end of the 1980s, at least forty develop-
ing countries, including the most populous, provided no patent protection
for pharmaceuticals. 44 The rationale behind this policy of non-protection
39. This figure is based on an analysis of cases from 1982 to 1993. See Lavenue,
supra note 37, at 502.
40. Kimberly A. Moore, Judges, Juries, and Patent Cases-An Empirical Peek In-
side the Black Box, 99 MICH. L. REv. 365, 385 (2000).
41. See 42 U.S.C. § 2183 (2000) (allowing the Atomic Energy Commission to com-
pel licensing of certain "public interest" patents); id. § 7608 (allowing compulsory li-
censes if use of the patented invention is required to meet emission requirements, no rea-
sonable alternative is available to meet the requirements, and the lack of availability of
the patentee would tend to lessen competition). In several cases, courts have defacto au-
thorized compulsory licensing by rewarding damages but refusing to enjoin infringement
for public interest reasons. See Vitamin Technologists v. Wis. Alumni Research Found.,
146 F.2d 941 (9th Cir. 1945) (stating that the partial refusal to license production of vi-
tamin D in oleomargarine amounted to patent misuse and suggesting that an injunction
could be denied if the refusal to license was against public interest); Milwaukee v. Acti-
vated Sludge, Inc., 69 F.2d 577 (7th Cir. 1934) (declining to issue an injunction against a
patent infringing sewage plant because it would cause lake pollution), cert. denied, 293
U.S. 576 (1934).
42. JEAN 0. LANJOUW & IAIN COCKBURN, Do PATENTS MATTER? EMPIRICAL Evi-
DENCE AFTER GATT 1 n.2 (Nat'l Bureau of Econ. Research, Working Paper No. 7495,
2000).
43. Id.
44. Id. at 1. However, all WTO members are obligated to offer pharmaceutical pat-
ent protection by 2016. See ABBOTT, supra note 18, at 11.
was that drugs are too important to patent and leave vulnerable to monop-
oly abuses.
However, a competing rationale has stimulated the recent trend toward
granting patent protection for drugs. Drug development is enormously
time-consuming, risky, and expensive,45 intensifying the importance of the
patent incentive. In addition, drug patents tend to be more effective in se-
curing commercial advantage because, once invented, drugs are relatively
easy to copy, and because a few key patents usually cover a single drug
product.46 Accordingly, surveys published in 1986 and 2000 all concluded
that the pharmaceutical, biotechnology, and chemical industries rely more
heavily on patents than other industries.47 Pointing to these facts, critics of
compulsory licensing have concluded that drugs are too crucial not to be
protected by patents.
The U.S. system reflects this inherent tension, extensively regulating
drug development on one hand and providing special incentives for drug
innovation on the other. In terms of regulation, pharmaceutical companies
must undergo a lengthy drug approval process administered by the Food
45. Precisely how expensive is highly contested. Researchers at Tufts estimate the
cost of developing a new drug to be $802 million. Joseph A. DiMasi et al., The Price of
Innovation: New Estimates of Drug Development Costs, 22 J. HEALTH ECON. 151, 166
(2003). However, roughly half of this figure reflects opportunity costs within the indus-
try. Id. Using data from PhRMA, Public Citizen estimates the cost of development to be
between $114 million and $150 million. PUBLIC CITIZEN, Rx R&D MYTHS: THE CASE
AGAINST THE DRUG INDUSTRY'S R&D "SCARE CARD" 7 (2001), available at http://www
.citizen.org./documents/acfdc.pdf (last visited Aug. 27, 2003). The Boston Consulting
Group, who estimates a development cost of $880 million, suggests that $165 million is
spent in target identification, $205 million is spent on target validation, $40 million is
spent on screening, $120 million is spent on optimization, $90 million is spent on pre-
clinical development, and $260 million on clinical development. The time expended in
each of these phases is estimated at 1, 2, 0.4, 2.7, 1.6, and 7 years, respectively. See PE-
TER TOLLMAN ET AL., THE BOSTON CONSULTING GROUP, A REVOLUTION IN R&D: How
GENOMICS AND GENETICS ARE TRANSFORMING THE BIOPHARMACEUTICAL INDUSTRY 12
(2001), available at http://www.bcg.com/publications/files/enggenomicsgenetics-rep_
1l01.pdf.
46. See F.M. Scherer, The PharmaceuticalIndustry and World Intellectual Property
Standards, 53 VAND. L. REV. 2245, 2247 (2000) (estimating that it costs only $1 million
to copy a drug); cf DiMasi, supra note 45 (estimating the development cost of a new
pharmaceutical to be $802 million).
47. This difference in reliance on patents is decreasing. See WESLEY M. COHEN ET
AL., PROTECTING THEIR INTELLECTUAL ASSETS: APPROPRIABILITY CONDITIONS AND
WHY U.S. MANUFACTURING FIRMS PATENT (OR NOT) 11-14 (Nat'l Bureau of Econ. Re-
search, Working Paper No. 7552, 2000), available at http://papers.nber.org/papers/
w7552.pdf (last visited Aug. 27, 2003); Edwin Mansfield, Patents and Innovation: An
EmpiricalStudy, 32 MGMT. SCI. 173, 175 (1986).
and Drug Administration ("FDA") prior to selling a new drug to the pub-
lic. Companies must prove the efficacy and safety of the new drug. Direct-
to-consumer drug advertising, liberalized in 1997, remains heavily regu-
lated.48 The government has occasionally authorized or ordered the com-
pulsory licensing of patented drugs as well, as discussed below.
In terms of incentives, the Orphan Drug Act of 1983 provides market-
ing exclusivity, tax incentives, and research grants for companies engag-
ing in research on rare "orphan" diseases that affect a small share of the
population. 49 Similarly, the Hatch-Waxman Act of 1984 extends the pe-
riod of exclusivity granted by drug patents in order to compensate for time
lost in FDA approvals. 50 These extensions are meant to encourage not only
the initial R&D that leads to the discovery of patentable drug inventions,
but the expensive and time-consuming testing and commercialization of
inventions after their discovery. In fact, according to one estimate, close to
50% of expenditures take place post-patenting. 51 Although post-patenting
development activities are highly worthwhile and for all practical purposes
required in order for the public to benefit from the patented innovation,
they are not necessarily "innovative" in the sense typically thought of, es-
pecially given that they are carried out downstream from the patentable
52
invention, often by parties other than the inventor.
B. Compulsory Licensing of Drugs in the United States
To date, Congress has resisted enacting specific provisions authorizing
the compulsory licensing of drugs, although pharmaceutical-specific price
48. See Tamar V. Terzian, Direct-to-ConsumerPrescriptionDrug Advertising, 25

AM. J.L. & MED. 149 (1999) (describing the FDA's in-depth regulations of prescription
drug broadcast advertisements).
49. See, e.g., F.M. Scherer, Pricing, Profits, and Technological Progress in the
PharmaceuticalIndustry, 7 J.ECON. PERSP. 97 (1993) [hereinafter Scherer, Pricing].
50. See THE BOSTON CONSULTING GROUP, SUSTAINING INNOVATION IN U.S. PHAR-
MACEUTICALS: INTELLECTUAL PROPERTY PROTECTION AND THE ROLE OF PATENTS 16-18
(1996).
51. TOLLMAN ET AL., supra note 45 and accompanying text. In the Boston Consult-
ing Group model, it is assumed that there are eleven years of patent protection after clini-
cal development. Id. at 59-60. Based on a patent life of twenty to twenty-three years, after
Hatch-Waxman extensions, this means that patents are issued nine to twelve years before
FDA approval, before the time consuming clinical and development phases, which con-
sume 45% of total expenditures. See The Boston Consulting Group, supra note 50, at 35
(stating that average extensions are two to three years in length).
52. See generally DATAMONITOR, REP. NO. DMHC1554, CREATING WIN-WIN BIO-
TECHNOLOGY AND PHARMACEUTICAL DEALS 22 (Oct. 2000) (describing the various ways
in which biotechnology firms may license their inventions to pharmaceuticals in the de-
velopment phase and estimating that 30% of pharmaceuticals use portfolio management,
a strategic tool that specifically contemplates drug development partnerships).
regulation has been contemplated intermittently since the 1950s. In the

late 1950s and early 1960s, the Kefauver hearings turned public scrutiny
on the industry's above-average profit levels, price markups, false and
misleading advertising, and general lack of price competition.53 In 1962,
Congress enacted the Kefauver-Harris amendments, which increased the
FDA's involvement in the development and advertising of drugs. 54 U.S.
lawmakers again addressed price control mechanisms in 1972 with the
proposal of the Public Health Price Protection Act, which was ultimately
unsuccessful.55
During the 1990s, several trends came together to focus attention on
drug pricing, the most prominent being the "relentless escalation" of
health care costs. 56 By 1992, the United States devoted 14% of its Gross
National Product to healthcare costs, more than any other industrialized
country.57 Prices rose much faster on drugs than on other goods, and
pharmaceutical profitability levels topped those of all other industries. 58
The unsuccessful Hart Bill of 1993 and Affordable Prescription Drugs Act
of 1999 proposed compulsory licensing of health related patents in various
circumstances, such as unreasonable pricing. 59 In 2000 and 2002, Presi-
dent Clinton and President Bush, respectively, blocked the implementation
of bills that would have enabled prescription60 drug wholesalers to import
drugs from countries where they are cheaper.
As another form of price regulation, compulsory licenses over drug
patents have been granted in two contexts-under 28 U.S.C. § 1498 and
under antitrust consent decrees. Although few in number, drug licenses
taken pursuant to the statute have involved deliberate infringement by the
53. See Mary T. Griffin, AIDS Drugs & the PharmaceuticalIndustry: A Need for
Reform, 17 AM. J.L. & MED. 363, 377 (1991).
54. Id.
55. Id. at 404.
56. Scherer, Pricing,supra note 49, at 97.
57. Id.
58. Id. at 98.
59. See Joseph A. Yosick, Compulsory PatentLicensing for Efficient Use of Inven-
tions, 2001 U. ILL. L. REV. 1275, 1278 (2001).
60. See Mason Essif, Prescription Drugs are Crossing Borders to Buyers,
CNN.cOM, Mar. 12, 2001, at http://www.cnn.com/2001/HEALTHI/03/12/prescription
.drugs (last visited Aug. 1, 2003); Robert Pear, Plan to Import Drugs From Canada
Passes In Senate, but Bush Declines to Carry It Out, N.Y. TIMES, July 18, 2002, avail-
able at 2002 WL 24463223. But see Import Drug Bill Clears House, CBSNEWS.coM,
July 25, 2003, at http://www.cbsnews.com/stories/2003/07/25/politics/main565066.shtml
(last visited Aug. 11, 2003) (describing House passage of a bill that allows for importa-
tion of drugs from Canada and the European Union and that, unlike past bills, specifically
avoids presidential oversight of the drug approval process).
government to produce drugs for public health purposes. In the 1960s and
1970s, the U.S. government made and used tetracycline and mepro-
bamate 62 for the military without permission from patent holders. Simi-
larly, in the fall of 2001, the threat of a compulsory license was used to
drive down the price of the patented drug Cipro by almost 50%.63
Antitrust orders have generated many more compulsory licenses, and
have been issued to remedy patent misuse and the use of patents in prce-
fixing, entry-restricting cartels, and market concentration schemes.6 One
of the most notable early cases involved the licensing of tetracycline, am-
picillin, and related products as part of a judgment against Pfizer, Ameri-
can Cyanimid, and other pharmaceutical
65
companies, in response to an an-
tibiotic price-fixing scheme.
In the 1970s, the FTC created a division, staffed with thirty-five law-
yers and investigators within the Bureau of Competition, to work exclu-
sively on health care antitrust issues. 66 In the second half of the 1980s and
early 1991, in response to the rising number of pharmaceutical mergers,
the division issued twelve consent decrees. Five decrees involved horizon-
tal mergers between direct competitors, three involved mergers between
potential competitors, and four involved the proposed combination of
R&D "innovation" markets. 67 Six of the twelve decrees ordered the com-
pulsory licensing of patented drugs; these form the basis of the analysis in
Part IV.
Although all antitrust licensing orders seek to address antitrust con-
cerns, their provisions have varied, depending on whether their objective
was to increase access to existing competitors, facilitate entry of new
competitors, or redress past wrongs by the patentee. Under a decree, re-
muneration for licenses may be negotiated by the parties, set by the court,
61. See SCHERER& WATAL, supra note 3, at 26.

62. Carter-Wallace, Inc. v. United States, 496 F.2d 535 (Ct. Cl. 1974).
63. See Kavaljit Singh, Anthrax, Drug Transnationals,and TRIPs, FOREIGN POLICY
IN Focus NEWSLETTER, Apr. 29, 2002, at 1-3, available at http://www.fpif.org/pdf/gac/
OUS0204trips.pdf (last visited Aug. 8, 2003); Press Release, U.S. Dep't Health & Human
Servs., HHS, Bayer Agree To Cipro Purchase (Oct. 24, 2001), available at http://www.
hhs.gov/news/press/200l pres/20011024.html (last visited Aug. 8, 2003).
64. SCHERER & WATAL, supra note 3, at 17.
65. See In re Am. Cyanamid Co., 63 F.T.C. 1747 (1963); Peter Temin, Technology,
Regulation, and Market Structure in the Modern PharmaceuticalIndustry, 10 BELL J.
ECON. 429, 435-41 (1979).
66. See HEALTHCARE SERVS. & PRODS. Div., FED. TRADE COMM'N, FTC
ANTITRUST ACTIONS IN HEALTHCARE SERVICES AND PRODUCTS 1-2 (Apr. 2003),
availableat http://www.ftc.gov./bc/hcupdate03040l.pdf.
67. Id.
or set at zero (royalty-free). Most often, orders call for reasonable royalties
and let the parties decide on the price. The court only intervenes if the par-
ties cannot agree. Some orders indicate specific monetary licensing terms,
while others authorize cross-licenses as an alternative. 68 Royalty-free
69
li-
censes are issued more rarely-usually in cases of misconduct.
Additionally, to ensure that the license issues to a viable or prospective
competitor, the DOJ or FTC approval of the licensee and additional li-
cense terms is sometimes required. 70 Also, to increase the likelihood that
patents will be used efficiently, the license may cover know-how, manu-
facturing capability, or other tangible or intangible assets in addition to the
patents. Special precautions are often taken in the case of pharmaceutical
licenses because of the special challenges posed by the time-consuming
and expensive drug development process. 7 1 This has led to the creation of
additional obligations for the patentee, such as providing ongoing support
until the licensee's product is approved, and the possibility of a continuing
relationship with the licensee.
C. Compulsory Licensing of Drugs under TRIPS
TRIPS contains a comprehensive framework for the compulsory li-
censing of patented inventions. The agreement also makes clear that, for
public health reasons, countries may suspend patent protection over drugs.
The primary provision for compulsory licensing is Article 31, which is
entitled "Other Use without Authorization of the Right Holder." This pro-
68. See, e.g., Hartford-Empire Co. v. United States, 323 U.S. 386, 414-17 (1945).
69. See, e.g., Lawrence Schlam, Compulsory Royalty-Free Licensing as an Antitrust
Remedy for Patent Fraud: Law, Policy and the Patent-AntitrustInterface Revisited, 7
CORNELL J.L. & PUB. POL'Y 467 (1998).
70. Seeln re Institut Merieux S.A., 113 F.T.C. 742 (1990).
71. See BUREAU OF COMPETITION, FED. TRADE COMM'N, A STUDY OF THE COMMIS-
SION'S DIVESTITURE PROCESS 40-41 (1999), available at http://www.ftc.gov/os/1999/08/
divestiture.pdf (last visited Aug. 28, 2003). The study stated that:
The pharmaceutical orders played an important role in the development
of the divestiture remedies because they posed, in a more obvious form,
some of the difficulties found in the Study. ... Foremost among them
is the fact that divestiture is not possible unless the Food and Drug
Administration authorizes the buyer to produce the drug or health
product. Until approval is obtained, the most that the buyer could ex-
pect to do under FDA rules is to market and distribute the products
made by the respondent. In the meantime, the buyer would be required
to build and replicate exactly the respondent's production facilities. The
orders had to reflect these realities through provisions requiring interim
supply agreements and technical assistance for a substantial period of
time.
vision permits WTO member countries to authorize compulsory licenses

for use by the government or third parties subject to certain restrictions.
Under all circumstances, patentees are to receive "adequate remuneration
...taking into account the economic value of the authorization." 72 Before
licenses are granted, the proposed user must try unsuccessfully for a rea-
sonable amount of time to secure a license on reasonable terms. 73 How-
ever, this requirement is waived if there is "a national emergency" or a
"circumstance[] of extreme urgency," or if the patented invention is used
for "public noncommercial use.",74 Such use must be non-exclusive and
non-assignable. 75 Additionally, unless the patentee has engaged in anti-
competitive behavior, the use must predominately supply the domestic
market. 76 Finally, the scope and duration of use is limited to the purpose
authorized with a license subject to termination "if and when the circum-
stances which led to it cease to exist and are unlikely to recur." 77 Article
30 authorizes general exceptions to patent protection, presumably includ-
ing compulsory licensing, but states that these exceptions must neither
"unreasonably conflict with a normal exploitation of the patent" nor "un-
reasonably prejudice the legitimate interests of the patent owner., 7 8
While Articles 30 and 31 apply to patents in all fields, Articles 8 and
27, as well as the Doha Declaration on the TRIPS Agreement and Public
Health ("Doha Declaration"), explicitly address the relationship between
TRIPS and public health. Article 8 states that "[m]embers may ...adopt
measures necessary to protect public health," but adds the requirement that
"such measures are consistent with the provisions of this Agreement., 79
Article 27 allows member countries to exclude from patentability inven-
tions needed to protect public health. 80 The Doha Declaration on TRIPS,
adopted in October 2001 by the WTO Ministerial Conference, affirms that
countries may undertake compulsory licensing for public health reasons.
Heralded as a major step forward in paving the way for cheap drugs for
the poor, 8 1 it states in part:
72. TRIPS Agreement art. 31 (h).

73. Id.art. 31(b).
74. Id.
75. Id.art. 31(d)-(e).
76. Id.art. 31 (f), (k).
77. Id. art. 3 1(g).
78. Id.art. 30.
79. Id.art. 8(1).
80. Id.art. 27(2).
81 . See, e.g., Ellen 't Hoen, TRIPS, PharmaceuticalPatents, andAccess to Essential
Medicines: A Long Way from Seattle to Doha, 3 CHI. J. INT'L L. 27, 28 (2002) (describing
We stress the importance we attach to implementation and inter-

pretation of the Agreement on Trade-Related Aspects of Intellec-
tual Property Rights (TRIPS Agreement) in a manner supportive
of public health, by promoting both access to existing medicines
and research and development into new medicines and, in this
connection, are adopting a separate declaration.82
In addition, the Doha Declaration clarifies that member countries may

define for themselves "what constitutes a national emergency or other cir-
cumstances of extreme urgency." 83 When a country declares an emergency
in good faith, this waives the obligation to negotiate under Article 3 1(b)
before issuing a compulsory license. Finally, the Declaration acknowl-
edges the problems imposed by Article 31 's requirement that manufactur-
ing be done primarily to service the domestic market, which prevents
countries without generic
84
drug manufacturing capabilities from making
use of the provision.
While developing countries have pressed for a broad interpretation of
the Doha Declaration, and thus a large list of diseases for which patent
rules will be relaxed, drug companies and their respective governments
have advocated for a narrow interpretation of the Declaration. 85 Although
the Declaration required that the TRIPS Council find an "expeditious solu-
the Declaration as an "important achievement" that "broke new ground in guaranteeing

Members' access to medical products").
82. WTO, Ministerial Declaration, Fourth Ministerial Conference in Doha, Qatar,
117 (adopted Nov. 14, 2001).
83. See WTO, Declaration on the TRIPS Agreement on Public Health, Fourth Min-
isterial Conference in Doha, Qatar, 5(c) (adopted Nov. 14, 2001) [hereinafter Doha
Declaration].
84. Id. 6. The Doha Declaration states
[w]e recognize that WTO members with insufficient or no manufactur-
ing capacities in the pharmaceutical sector could face difficulties in
making effective use of compulsory licensing under the TRIPS Agree-
ment. We instruct the Council for TRIPS to find an expeditious solu-
tion to this problem and to report to the General Council before the end
of 2002.
Id.
85. See, e.g., Sarah Boseley & Charlotte Denny, Prescriptionfor World's Poorest
Stays Unwritten: WTO ConferenceDeadlock as US Shows no Sign of Loosening Veto on
PharmaceuticalPatent Rights, GUARDIAN, Feb. 20, 2003, available at http://www.
economist.com/agenda/displayStory.cfi?storyid=1589657 (last visited July 19, 2003);
Negotiators Meet Again; Minds Don't, ECONOMIST, Feb. 19, 2003, available at http://
www.economist.com/agenda/displayStory.cftn?story_id=1589657 (last visited July 19,
2003).
tion" to these issues by the end of 2002,86 it was not until late August 2003
that an accord was reached.87
IV. EMPIRICAL BACKGROUND (LITERATURE REVIEW)

One major obstacle to the widespread use of compulsory licenses has
been the perception that licenses reduce the incentive for innovation of-
fered by the patent system. 88 Insofar as patents are needed to induce inno-
vation, the argument goes, weakening patents through compulsory licenses
will reduce innovation. This notion has special import for the drug indus-
try.
First, it is often repeated that drugs, due to the costs and risks associ-
ated with drug development, are different than other inventions, and that
the drug industry relies on patents more than other industries. 89 Because of
this unique dependence on patents, more is at stake for the drug industry
than for other industries when measures that reduce patent protection such
as compulsory licensing are contemplated.
Second, in light of the current public health crisis, relaxation of patent
rules will likely take place to some degree, regardless of any potential ef-
fect on innovation. 90 In the context of the AIDS crisis and public health
generally in developing countries, at least two kinds of incentives are rele-
vant-those that prompt research in diseases of common interest to devel-
oped and developing countries (e.g., AIDS), and those that encourage re-
search in areas specifically relevant to developing countries (e.g., ma-
laria). 91 Particularly problematic could be a negative impact on R&D spe-
cific to the developing countries, the growth of which is anxiously antici-
pated with the introduction of stronger patent protection. 92 In light of these
86. Doha Declaration, supra note 83, 6.

87. See, e.g., Elizabeth Becker, Poor Nations Can Purchase Cheap Drugs Under
Accord, § 1, at 14.
88. See Tren, supra note 2.
89. See Richard C. Levin et al., Appropriatingthe Returns from IndustrialResearch
and Development, 1987 BROOKINGS PAPERS ON ECON. DEV. 783, 796-98 (1987); supra
note 47.
90. See infra note 80 and accompanying text.
91. See Jean 0. Lanjouw, A Patent Policy Proposal for Global Diseases (Apr. 2001),
at http://econ.worldbank.org/files/1733_lanjouw.pdf (last visited July 14, 2003).
92. See generally CARSTEN FINK, How STRONGER PATENT PROTECTION IN INDIA
MIGHT AFFECT TI-E BEHAVIOR OF TRANSNATIONAL PHARMACEUTICAL INDUSTRIES
(World Bank Group, Policy Research Working Paper No. 2352, 2000) (modeling the im-
pact of stronger patent rights on the pharmaceutical industry in India), at http://www-
wds.worldbank.org/servlet/WDSContentServer/WDSP/IB/2000/06/27/000094946_00060
905463269/Rendered/PDF/multi_page.pdf (last visited July 28, 2003); LANJOUW &
incentives, the challenge for policy makers will be to implement patent-

weakening schemes that increase access but cause minimal harm to the
patent innovation incentive.
A. Compulsory License Design
The impact of a license on the licensor's innovation depends on a vari-
ety of factors. The following paragraphs identify possible factors that
might determine how much a compulsory license impacts innovation.
It is clear that the price at which a compulsory license is set will de-
termine whether and how much innovation is affected. If a compulsory
license is priced essentially at what a patentee demands, there is no real
reason to anticipate that innovation will be substantially harmed. On the
other hand, a compulsory license whose price is set at a level far below
market could operate to effectively strip the patentee of its right to any
monopoly profits. 93 Besides price, two factors that deserve special atten-
tion are "market significance," or the extent to which a licensee actually
threatens the patentee's markets, and "predictability," or the extent to
which a licensor anticipates a compulsory license.
As to market significance, compulsory licenses can vary in degree as
to the competitive threat they pose to licensors. If a compulsory license
covers a known product in a licensor's target market, the licensor and the
licensee will have to share the same market. Under the above definition,
the market significance of this license is high because the licensor's mar-
ket is directly threatened. Conversely, if the license covers a market that is
unimportant to the licensor, or it covers a product that has yet to be proven
or for which the market is immature or untested, there is a good chance
that the licensee and licensor will not compete head to head. The signifi-
cance of this license may be relatively low.
Whether a license is predictable is also an important characteristic.
Unpredictable licenses that cover only existing technologies are more lim-
ited in scope than those that are predictable and cover future inventions.
Although the unanticipated loss of exclusivity that accompanies an unpre-
dictable compulsory license may influence a company's decisions about
investing in follow-on innovation, development, and commercialization,
COCKBURN, supra note 42, at 3-4 (establishing for future reference the current baseline of
research efforts devoted to those diseases specific to developing countries).
93. For a deeper analysis of the ethical and economic issue of what developing
countries should contribute to innovation, see WILLIAM JACK & JEAN 0. LANJOUW, FI-
NANCING PHARMACEUTICAL INNOVATION: How MUCH SHOULD POOR COUNTRIES CON-
TRIBUTE? (Ctr. for Global Dev., Working Paper No. 28, 2003), at http://www.cgdev.org/
wp/cgd wp028.pdf (last visited Aug. 8, 2003).
the licensing event may come at a point that is too late for the company to
change course. This is not the case with an order that requires licensing of
future patents. The licensor may choose to redirect R&D investment, put
off inventive activity until the license has expired, or choose trade secret
over patent protection.
B. Literature Review
For some time, researchers have not focused on compulsory licensing
and the more general phenomenon of weakening patent protection, pre-
sumably because changes to the patent system over the last several dec-
ades have been in the direction of strengthening patent protection.94 Nev-
ertheless, major studies conducted in the 1960s and 1970s on compulsory
licensing regimes concluded that, as implemented, licensing had no long-
term negative impact on licensor innovation. 95 The most thorough study to
date, which focused on U.S. antitrust consent decrees issued during the
1950s and 1960s, found that licensing had no measurable impact on future
innovation in any of the industry segments studied, including pharmaceu-
ticals.96 Another major study that focused on Canada's extensive general
compulsory licensing program similarly concluded that 97
Canada's program
had no negative impact on pharmaceutical innovation.
However, research on related questions suggests that some forms of
compulsory licensing could be detrimental to innovation. From 1967 to
1968, the Harbridge House conducted a study of civilian utilization of in-
ventions created for the government. The study demonstrated that the loss
of exclusivity due to the compulsory licensing of some of the inventions
negatively affected utilization rates of those inventions. 98 In addition, there
is a perception that compulsory licensing can discourage R&D. A survey
of British pharmaceutical executives suggested that they 99
believed that, in
some extreme forms, licensing could harm innovation.
Like the study in this Article, these studies focused exclusively on li-
censor innovation, and largely ignored the impact of compulsory licensing
on the licensee. The licensee often benefits from the "spillover" effects of
94. See LANJOUW & COCKBURN, supra note 42.

95. See SCHERER, supra note 31; Donald G. McFetridge, Intellectual Property,
Technology Diffusion, and Growth in the CanadianEconomy, in COMPETITION POLICY
AND INTELLECTUAL PROPERTY RIGHTS IN THE KNOWLEDGE BASED ECONOMY 65 (Robert
D. Anderson & Nancy T. Gallini eds., 1998).
96. See SCHERER, supra note 31, at 67-75.
97. See McFetridge, supra note 95.
99. C.T. TAYLOR & Z.A. SILBERSTON, THE ECONOMIC IMPACT OF THE PATENT SYS-
TEM 198-99 (1973) (study results further described in SCHERER, supra note 31, at 60-62).
the original innovation. 100 Indeed, follow-on innovation by competing li-

censees or by potential entrants is often the very aim of licensing orders in
the antitrust context.' 01 The question of whether a potential tradeoff be-
tween increased licensee innovation and decreased licensor innovation ex-
ists, however, is beyond the scope of this Article.
1. Compulsory Licensing under U.S. Antitrust Consent Decrees
In 1977, F.M. Scherer conducted a major study of antitrust, consent-
related compulsory licenses. His study focused on nearly seven hundred 0 2
companies, forty-two of which had been subject to compulsory licenses.'
Scherer calculated the ratio of each company's R&D expenditures to its
sales for the year 1975, and compared ratios between companies that had
been subject to significant compulsory licensing decrees and those that
had not. Scherer further modeled the relationship between compulsory li-
censing and R&D, and found a slight positive correlation between licens-
ing and high R&D-to-sales ratios. On average, companies subjected to
compulsory licensing actually spent more on R&D than similar firms in
their industry that had not been subjected to compulsory licenses. 03 This
was true for all industries, including pharmaceuticals. Because Scherer
only had data from one year, he was not able to determine whether the
R&D expenditures of the firms affected by compulsory licensing had
fallen from previously higher levels.' 0 4 Nevertheless, he concluded that
compulsory licensing had not forced 0 5
firms to invest in R&D at a level be-
low the norms in their industries.'
100. See Testimony of the Biotechnology Industry Organization Before the Federal
Trade Commission and the Department of Justice, at 2 (Feb. 26, 2002), available at
http://www.bio.org/ip/pdf/ftc022002.pdf (last visited Aug. 8, 2003).
101. See Susan DeSanti, The Intersection of Antitrust and Intellectual Property Is-
sues: A Report from the FTC Hearings, Remarks for the Business Development Associa-
tion Conference on Antitrust for High-Tech Companies (Feb. 2, 1996), available at
http://www.ftc.gov/speeches/other/desanti l.htm.
102. See SCHERER, supra note 31, at 67-68, 74.
103. Id. at 75 ("To sum up, the analysis of 1975 research and development spending
patterns provides no significant indication that 44 companies subjected to compulsory
patent licensing under antitrust decrees sustained less intense R&D efforts than other
firms of comparable size and industry origin. If anything the opposite tendency is re-
vealed.").
104. In an earlier survey of thrity-eight companies affected by compulsory licenses,
Scherer observed a statistically significant decline in patenting by companies. Based on
interviews, he concluded that this was due to a statistical fluke or shift toward trade se-
crecy. Id. at 66-67.
105. Id. (observing a statistically significant simple average decline of 15% in abso-
lute patenting).
Scherer's study focused on antitrust licensing decrees. These decrees

mandated compulsory licensing as part of case settlements.' 06 The major-
ity of these licenses did not require future licensing of patents issuing from
the year studied, but covered past inventions.10 7 Although Scherer hy-
pothesized that specific past experience with compulsory licenses or the
general threat of licenses might produce an adverse impact on R&D be-
havior, he found no statistical results to support these hypotheses. 0 8 In the
short-term, the largely unpredictable licenses studied did not appear to im-
pact behavior in the year studied. The view that firms focused on the long-
term was expressed in an earlier study conducted by Scherer that focused
on companies that either had been or were on the verge of being forced to
license patents in the antitrust context. ° 9 The most common explanation
provided by the firms for not changing R&D was their long-term interest
in the impacted business, and the view that they needed to continue R&D
in order to stay competitive. 110
2. Compulsory Licensing of Drugs in Canada
While the short-term, unpredictable nature of the antitrust licenses
studied by Scherer may explain in part why he observed no negative im-
pact on innovation, Canada's experience with compulsory licensing pro-
vides a useful example of the opposite extreme-completely predictable
licenses.' From 1923 to 1993, Canadian legislation authorized compul-
sory licensing over medicines under sections 4(1) and 39(4) of the Cana-
dian Patent Act. Canada's policy of issuing compulsory licenses for drugs
became so routine
2
that it led to the development of a domestic generic
drug industry. 1
In 1985, the Eastman Commission reported the effects of Canada's
broad compulsory licensing system on innovation, focusing especially on
the pharmaceutical industry."13 From 1969 to 1983, the period studied by
the Commission, almost 80% of the applications for licenses were granted,
106. Id. at 63 (describing compulsory licensing in antitrust decrees generally).

107. Id. at69.
108. Id. at 69, 74.
109. Id. at 62-63.
110. Id. at62.
111. See REICHMAN & HASENZAHL, supra note 7, at 18-20.
112. See SCHERER, supra note 31, at 83. The Canadian Parliament abolished the pro-
gram in 1993 after intense lobbying by the U.S. government during NAFTA negotiations.
Objecting vigorously to Canada's broad embrace of compulsory licensing, the U.S. gov-
ernment feared that other countries might follow suit. See REICHMAN & HASENZAHL,
supra note 7, at 21-22.
113. See McFetridge, supra note 95, at 83.
resulting in an average of approximately twenty compulsory licenses per

year." 14 Comparing R&D intensities in Canada to intensities in other
small, developed countries, the Commission concluded that 15
compulsory
licensing did not significantly affect innovation in Canada.
One reason for this result may be the relative insignificance of the Ca-
nadian market to the worldwide market for pharmaceuticals. Researchers
noted that for the most part, "Canadian R&D ... expenditures consti-
tute[d] a very small fraction of [corporate parent] R&D and... remain[ed]
below the minimum efficient scale for in-house R&D in this industry. '16
As a result, the lack of patent protection in Canada had little influence on
R&D decisionmaking.
Thus the Scherer and Eastman Commission studies both concluded
that compulsory licensing had little adverse impact on licensor innovation,
but probably for different reasons. In the case of U.S. antitrust licenses, the
unpredictability and short-term nature of the licenses may explain why
they did not greatly affect innovation. In the case of Canadian drug li-
censes, the relative insignificance of the Canadian market may have ac-
counted for the lack of a noticeable adverse impact.
3. Licensing of U.S. Government Inventions
In contrast to the Eastman and Scherer studies, a study conducted by
Harbridge House reported that in some cases, a loss of patent rights might
result in negative effects on innovation and commercialization. In the
1960s, the Federal Council for Science and Technology commissioned
Harbridge House to investigate whether or not contractors based their de-
cision to commercialize inventions they made for the government on ex-
clusivity grounds. 1 7 Under the contracts studied, when a contractor cre-
ated a patented invention for the U.S. government, the government could
either take a license to the invention or take title to the invention itself.
When the government merely took a license, the contractor had exclusive
civilian use of the invention. However, if the government took title to the
invention, the contractor had no assurance of exclusivity. The contractor
would then effectively be subject to the threat of a compulsory license in
114. id. at 82tbl.1.

115. Id. at 88.
116. McFetridge reports that with the exception of Merck, 1994 Canadian R&D ex-
penditures as a percentage of worldwide expenditures were less than 2%; e.g. 1.3% for
Glaxo, 1.0% for Hoffman LaRoche, 0.7% for Pfizer, 1.2% for Sandoz, 1.4% for Ciba,
1.7% for Eli Lilly, and 6.1% for Merck. Id.at 84 n.24.
which the government or its potential 8 licensees would have a complete,

royalty-free right to use the patent."
From its study of 1,720 contractor inventions, Harbridge House found
a substantial difference in contractor utilization of patents depending on
whether or not contractors had exclusive rights in civilian markets, al-
though prior commercial experience proved to be the most significant fac-
tor.1 19 Among contractors with commercial experience, 23.8% who had
exclusive rights chose to commercialize their inventions. The figure was 120
only 13.3% among contractors who did not have exclusive rights.
Among those without prior experience, there was also a demonstrable dif-
ference, although the shares are small-while 6.6% of those with exclu-
sive rights chose to develop the technology, only 2.2% of those without
exclusive rights did. 12 1 In-depth interviews revealed that small firms, new
entrants, or firms facing substantial development and technological risks
were the most sensitive to the presence or lack of exclusivity. These firms
were generally unwilling to invest in commercialization without an assur-
ance of exclusivity.' 22 On the other hand, where contractors perceived that
they had an advantage in the relevant market or that marginal costs were
small relative to potential revenues, development was likely even despite a
23
lack of exclusivity. 1
The Harbridge study suggests that the relative importance of the mar-
kets implicated by a compulsory license matters with respect to innova-
tion. In the case of the "licenses" analyzed, the contractors faced the loss
of exclusivity in the civilian sector, which was their primary, most impor-
tant market. In contrast, the licenses studied by the Eastman Commission
implicated the Canadian market, which was viewed as less important by
pharmaceutical patent holders. The implication of these two data points is
that where the impacted market is important (as in the Harbridge study),
an adverse impact on development may be more likely than where the im-
pacted market is unimportant (as in the Eastman study).
In terms of predictability, commercialization was expected, as reported
in the Harbridge study, because the government's election to take title to a
patent signaled its intent to commercialize that patent in the future. The
threat posed by government utilization was thus more similar to the regu-
118. Id. at 78-79.

119. Id. at 79-81.
120. Id. at 80.
121. Id.
122. Id. at 82.
123. Id.
lar licensing regime of Canada in the Eastman Commission study than the
sporadic licensing of the U.S. antitrust decrees in the Scherer study.
Thus, the Harbridge licenses were both predictable and covered a
market significant to the patentee. Although appearing to discourage
commercialization, the Harbridge licenses can be distinguished from the
compulsory licensing schemes studied by Scherer and the Eastman Com-
mission, where no negative impact on innovation was observed. The
Scherer licenses were generally not predictable, issuing as part of investi-
gative probes by the government. Although the Eastman licenses were
predictable, they did not cover an important market for the patentees. The
implication of these results appears to be that where licenses are unpre-
dictable (Scherer) or implicate insignificant markets (Eastman), there will
not necessarily be an adverse impact. However, licenses that are both pre-
dictable and affect significant markets, such as the Harbridge licenses, po-
tentially are more risky, and appear to have a greater chance of being ac-
companied by a negative impact on innovation.
4. Hypothetical Licensing ofAll PharmaceuticalInventions
Research conducted by Taylor and Silberston in the form of opinion
surveys is consistent with this conclusion. The researchers asked officials
from British industries, including the pharmaceutical industry, to predict
the impact of a hypothetical system in which all patents, both domestic 24
and foreign, were made available for licensing at reasonable royalties.
This extreme form of licensing would be predictable in its reach on future
patents and would cover all, and therefore significant, markets. Executives
from all industries were asked to evaluate this hypothetical system. The
pharmaceutical industry respondents were the most concerned. On aver-
age, they predicted that 64% of R&D would be displaced without effective
patent protection, as compared to a weighted average of 8% among all in-
25
dustries. 1
C. Summary of Results
In summary, research to date indicates that, at a minimum, the pres-
ence of two factors may be required in order for compulsory licenses to
impact innovation. These factors are the predictability of the license being
granted and the significance of the market affected by the license. Where
either factor is absent, little measurable effect on R&D expenditures has
been observed, as shown in the studies performed by Scherer and the
Eastman Commission. However when predictable licenses actually (Har-
124. Id. at 61.

125. Id. at 62.
bridge) or hypothetically (Taylor and Silberston) issue over important

markets, the risk of a negative impact is greater. It should be emphasized
that although these factors emerged from comparing these studies, other
factors such as level of compensation may be just as important. Even if
licenses are predictable and affect significant markets, if the price of the
license is set at market rates, the license probably will not harm innova-
tion. The factors of predictability and significant market impact may thus
be necessary but not sufficient for producing a negative impact on innova-
tion.
V. CASE STUDIES OF INVESTMENT IN INNOVATION

AFTER SIX ANTITRUST CONSENT DECREES
To test the hypothesis that only licenses that are both predictable and
threaten a significant market adversely impact investment in innovation, I
studied six cases from the 1980s and 1990s where the FTC issued pharma-
ceutical compulsory licenses. These cases were the only ones I could find
in which licensing, rather than divestiture or other remedies, was pre-
scribed and which were recent enough that data concerning the R&D be-
havior of the affected firms was available. However, these incidents of
licensing are imperfect proxies for compulsory licensing in the interna-
tional public health sphere for several reasons. The primary objective of
these antitrust licenses was to preserve competition, not to increase con-
sumer access to drugs per se. In addition, the licensing events were limited
in scope in that they affected specific products produced by specific firms,
rather than affecting broad therapeutic areas in entire industries. Such a
broad license could be implicated if, for instance, compulsory licenses
were made available to all African countries over AIDS vaccines drugs.
Nonetheless, the licenses are relevant to the question of whether past com-
pulsory licenses have been accompanied by,.a decline in innovation.
The licenses I studied were ordered under antitrust consent decrees is-
sued by the FTC. Of the six licenses, four were sporadic and two were
predictable, and three covered nascent and therefore relatively less impor-
tant markets, whereas three jeopardized already developed, and therefore
important, markets. Within this modest data set, I considered the relevance
of the predictability and market significance of licenses to the R&D out-
puts of the affected companies. While general trends are reported below,
case studies of each FTC order that analyze the license, overall business
environment, and subsequent record of innovation by the licensor in the
relevant market can be found in the Appendix.
While building on past work, this study introduces several new con-
siderations. First, rather than analyze company-level activity, as did
Scherer, I concentrate solely on company activity within an affected
therapeutic area. This focus seems appropriate given the size and diversi-
fication of pharmaceutical companies-a substantial change to one thera-
peutic area may not be reflected in the activity of a company as a whole.
However, this focus could overstate any impact on net innovation to the
extent that a shift in activity away from the affected therapeutic area to
another within a company that does not reduce overall R&D-would appear
as a decline in activity. Second, this study contemplates drug patents from
the 1980s and 1990s. Selecting data from this period allowed me to test
the robustness of previous findings in light of the trend toward strength-
ened patent protection, generally and over biological inventions. 126
A. The Antitrust Drug Licenses
With the exception of the Eli Lilly license (see Table 1), all of the li-
censes I studied arose in the acquisition or merger context. 127 Although
each order resulted from negotiations with the FTC, four were "sporadic"
in that they occurred only once to remedy a specific concern, and left little
discernable expectation of future licenses in the near-term. In contrast, the
licenses ordered in the Eli Lilly and Merieux cases both covered future in-
novation. 128 In the Eli Lilly case, the order called for all patents issued or
applied for in the five-year period following the order to be subject to a
royalty-free license. In the Merieux case, the order required that the ac-
quirer, Institut Merieux S.A. ("Meriuex"), lease Bioscience Connaught's
("Connaught") rabies vaccine29 business long-term, and that it retain no fu-
ture interest in the business.1
126. See also Diamond v. Chakrabarty, 447 U.S. 303 (1980) (expanding patent pro-
tection to human made microorganism). See generally Jon F. Merz & Nicholas M. Pace,
Trends in PatentLitigation: The Apparent Influence of StrengthenedPatents Attributable
to the Court of Appeals for the Federal Circuit, 76 J. PAT. & TRADEMARK OFF. SOC'Y
579 (1994) (noting the increase in judgments finding patents either valid or infringed, as
opposed to invalid).
127. See In re Ciba-Geigy Ltd., 123 F.T.C. 842 (1997); In re Baxter Int'l Inc., 123
F.T.C. 904 (1994); In re Dow Chem. Co., 118 F.T.C. 730 (1994); In re Roche Holding
Ltd., 113 F.T.C. 1086 (1990); In re Institut Merieux S.A., 113 F.T.C. 742 (1990); In re
Eli Lilly & Co., 95 F.T.C. 538 (1980).
128. See infra Part VIII.C-D.
129. See infra Part VIII.D.
Table 1: Drug License Orders under Antitrust Decrees

Licensor, Trigger- S(poradic) E(arly) Nature of Subject Compensa-
Invention ing Event v. v. M(id) Market of tion
(Year of order) G(eneral) v. Affected License
license L(ate)
Stage of
Drug
Dev't
Baxter, Fibrin Merger S M Mature Patents +, 0

Sealant 130 manufac-
(1997) turing
Marion Merger S L Mature Patents +, 0

Merrell Dow, manufac-
Di- 31
turing
cyclomine'
(1994)
Ciba- Merger S E Nascent Patents + "no mini-

Geigy/Chiron,
132
mum"
HSV-tk
(1997)
Roche, CD4133 Merger S E Nascent Patents + 1-3% of net

(1990) sale
Eli Lilly, Illegal G E Nascent Future Reasonable

Insulin13 4 Conspir- Patents+ share of
(1980) acy R&D
expenses
Connaught, Merger G L Mature Entire Reason-

Rabies Vac- Business able sum
1 35
cine
(1990)
The stage of development of the affected technology varied among the

cases. The Chiron and Roche cases involved concerns about patents over
130. In re Baxter Int'l Inc., 123 F.T.C. 904 (1997).

131. In re Dow Chem. Co., 118 F.T.C. 730 (1994).
132. In re Ciba-Geigy Ltd., 123 F.T.C. 842 (1997).
133. In re Roche Holding Ltd., 113 F.T.C. 1086 (1990).
134. In re Eli Lilly & Co., 95 F.T.C. 538 (1980).
135. In re Institut Merieux S.A., 113 F.T.C. 742 (1990).
broad basic technologies and therefore covered some early-stage, pre-

clinical technologies.' 36 Similarly, the Eli Lilly license covered patents
over insulin produced by novel recombinant DNA methods that had not
yet undergone clinical trials at the time the order was issued. 137 The Baxter
case was prompted by concerns about the merger of two companies that
each had products in early development.' 38 On the other end of the spec-
trum, the Merieux case involved older patents, and the Dow case covered a
product market in which generic competition had already been intro-
duced.' 3 9
The stage of technology is relevant to this analysis because, as de-
scribed earlier, the drug development process is inherently uncertain. Can-
didate compounds are eliminated at each step. Taking away patent protec-
tion over an early stage technology arguably does not affect a patentee's
competitive position as much as a license over a technology that has al-
ready surpassed many major milestones. Thus, as reported in Table 1, I
characterized licenses over mid-to-late stage technologies as impacting
developed and significant markets, and compulsory licenses covering
early stage technologies as covering relatively less significant markets.
Importantly, each license involved more than just patents, and most
provided for access to know-how and other intangible assets. As stated
earlier, this licensing practice reflected the FTC's recognition of substan-
tial market barriers in the drug industry not associated with patents in gen-
eral, as well as its view that a more robust form of licensing was crucial to
the success of licensees. 140 Although individual license orders varied, most
contained a provision of either reasonable or no compensation for the
rights to use the patented invention. However, the Baxter and Dow li-
censes required the manufacturing and delivery of the patented product, so
these orders provided for additional compensation to cover the costs of
supply. 141
B. Measurement of the Impact of Drug Licenses
To determine whether licenses brought about a decline in innovation, I
looked at patent applications filed by each licensor as reported in the
Lexis-Nexis "Utility Patents" database, and where available, considered
clinical trial, product launch, and other data specific to the affected com-
136. See infra Part VILE-F.

137. See infra Part VIII.C.
138. See infra Part VIII.A.
139. See infra Part VIII.B, D.
140. See supra note 71 and accompanying text.
141. See infra Part VIII.A-B.
pany in the affected product area during the years before and after the or-
dering of a consent decree. Although several weaknesses were inherent in
this approach, as discussed below, patent applications appeared to provide
the best means for measuring licensor impact in a specific technology
area, which might otherwise be masked by either industry-level or aggre-
gate company data. Although budget information regarding R&D in spe-
cific therapeutic areas or interviews with the companies themselves would
also have been useful, I was unable to secure either source of information
To identify patent applications, I used keyword searching in the speci-
fications and claims of patent applications that eventually matured into
patents. Because queries are sensitive to the search terms used, I selected
my terms by reading about the technology area and then formulating
searches based on the original patents or patent applications licensed by
the FTC order. I also asked a medical doctor142
to review the terms I used for
the more ambiguous technological areas.
To make my analysis less sensitive to the absolute number of patents
filed for by the impacted companies during the affected timeframe, I repli-
cated my searches in the entire Utility Patents database. I compared shares
of patents filed by an affected company to patents filed for by the general
population before, during, and after the affected period so as to eliminate
any general bias due to changes in patenting. Based on the normalizations,
I saw no difference in the patterns. Because of the time lag between R&D
investment and the issuance of a patent based on the investment, I tried
wherever possible to capture activity over long pre- and post- licensing
event windows. Through this method, and by focusing on patent applica-
tions rather than patent grants, I tried to eliminate some of the time lag be-
tween innovative activity and patenting. For the older cases, I was able to
capture up to fifteen years before and after the license; however, for the
most recent cases, I was only able to capture four years of activity after the
license.
Although I took the precautions described to filter biases from my
data, other factors may affect the accuracy of my results. For example,
companies may choose not to patent or to delay patenting inventions for
strategic or other reasons. This fact tends to discredit the use of patenting
activity as a measure of company investment in innovation. However, the
patents at issue in the six cases at hand were most likely important to the
relevant companies because the FTC considered the patents important
enough to require that they be licensed. Another limitation is the possibil-
ity that the companies shifted their intellectual protection strategies to-
142. The chosen search terms are described in the case studies in the Appendix.
wards trade secrecy, thereby maintaining their pre-licensing level of R&D

while reducing patent output. If this is the case, post-licensing investment
in innovation will have been understated by the patent counts.
To address the problems presented by using patents as a measure of
investment in innovation, I used other measures of company commitment
to each therapeutic area. I searched BioSpace Inc.'s Clinical Competitive
Intelligence System for clinical trials that had been in progress sometime
in the 2000 to April 2002 period. 14 3 The BioSpace database covers about
50% to 60% of all private and public clinical trials, and is reportedly the
most comprehensive of all clinical trials database, including the develop-
ment database offered on the Pharmaceutical Research and Manufacturers
of America website. I also looked at each company's annual reports and
websites for new drug announcements, infrastructure commitments, and
other clues about each firm's commitment to innovation in the affected
area.
The main limitation of this study is that it comprises only six data
points and therefore cannot support any statistical conclusions. At most,
the case studies analyzed in this Article provide anecdotal illustrations of
how compulsory licensing might impact investment in innovation.
C. Impact of Compulsory Licensing
1. Sporadic Licensing
Based on patent application activity (see Figure 1) and both qualitative
and clinical trial evidence (see Appendix), it appears that none of the four
"sporadic" licenses were accompanied by a reduction in innovation. This
is in line with both the existing literature and common-sense expectations;
the element of surprise and the unpredictable nature of the licenses pre-
sumably made it impossible for any of the licensors to change their behav-
ior in anticipation of the license.
For each licensing event, Figure 1 shows the absolute number of pat-
ents in the therapeutic area affected by the license filed by a licensor in the
years preceding and following the FTC order. For the most part, I counted
applications filed in twelve-month increments beginning with the month
following the order, rather than based on the calendar year. 44 In the in-
143. See generally BioSpace: Competitive Clinical Intelligence System, (Biospace

2003) at http://www.biospace/com/ccis/index.cfm (providing a searchable database of,
inter alia,clinical trials).
144. For the last two years of the Baxter data, I used international patent filing data,
which captures all applications that have been on file for at least eighteen months, to sup-
plement U.S. filing data given the long lags between filing and issue of fibrin sealant pat-
stances of Baxter International ("Baxter") and Roche Holding Ltd.

("Roche"), there appeared to be no interruption of the general trend of pat-
ent applications. With Chiron Corporation ("Chiron"), the absolute num-
ber of patent applications peaked before the licensing order, but the com-
pany continued to steadily file applications in successive years. The oppo-
site is true for Merion Merrell Dow ("Dow"), where the twelve-month pe-
riod following the order was the most productive in terms of the number
of applications filed. Based on the few data points provided, no systematic
negative impact on patent applications was observed. This result is in line
with earlier research.
Figure 1: Sporadic Licenses - Absolute Patenting
Baxter T = 0 (Licensing Event)
10 -
Chiron
5
10
5-
10 -
Roche
-5 -4 -3 -2 -1 1 2 3 4 5
Years from Licensing Event
Although the graphs do not reveal clear trends, compulsory licensing

did not cause dramatic reductions in R&D according to measures besides
patent counts. In the instance of Baxter, marketing considerations seemed
to encourage the firm to continue investing in its fibrin sealant product
line. After several years of competition with Haemacure Corporation
("Haemacure"), its licensee, Baxter still retained a market share of 75%
and enjoyed a high revenue growth rate, particularly with respect to its
ents. As a result I relied on calendar twelve-month increments in order to capture as much

application activity as possible.
other blood products. 145 With Chiron and Roche, each company currently
captures a considerable share of clinical trials, signaling long-term com-
mitments to the affected product lines. As for Dow, the dwindling impor-
tance of dicylomine as a treatment for irritable bowel syndrome ("IBS")
makes it probable that factors other than the license influenced the com-
pany's decisionmaking. Each of these examples is explored more fully in
the Appendix.
2. PredictableLicensing
The two instances of predictable licensing present a more complex
problem. While both Eli Lilly and Connaught/Merieux were subject to li-
censes covering future innovation, Eli Lilly flourished during the time of
its compulsory license while Connaught claimed to be adversely impacted.
This difference is reflected in Figure 2, which shows the absolute patent
filings of Eli Lilly and Connaught before, during, and after the compulsory
licensing period.
In the Eli Lilly example, the data indicate that the licensing event was
not coupled with a negative change in the company's patenting activity.
Innovative outputs actually rose rather than declined. In the Connaught
example, patent application counts are of limited value because there was
very little patenting activity in the relevant therapeutic area. Although
Connaught did not file any patents during the four years it was affected by
a licensing agreement, it also did not file any patents prior to the licensing.
However, evidence other than patenting suggests that Connaught's inven-
tive activities declined after licensing. For example, Connaught reported to
the FTC that the license prevented it from upgrading its facilities. 146 Also,
Merieux's continued patenting activity before, during and after the licens-
ing (see Figure 2), suggest that Merieux, which was not subject to the li-
cense continued to innovate even while Connaught did not. A more de-
tailed discussion of these two case studies follows.
145. See infra Part VIII.A.

146. In re Institut Merieux S.A., 113 F.T.C. 742 (1990).
Figure 2: Predictable Licenses - Absolute Patenting
f >.
.-
O= Q'
C~ CL 1 Eli Lilly E
0)0
0, 0..
5. . .
0 5 --
O.2
C..L 10- Connaugh tu

ILI
M.10-- Connaught+ )
Merieux '
5 -. (Connaught Parent)
Z c
-3 -2 -1 0 0 1 2 3
Periods from Licensing Event
For Eli Lilly, Period = 2.5 years, For Connaught & Merieux, Period = 2 years
(scaled according to duration of licensing event such that event = 2 periods)
The consent decree in Eli Lilly was very broad. It provided access to
Eli Lilly's intangible assets for all who, within five years of the decree,
stated a bona fide intention to produce and sell insulin products in the
United States. Included in the intangible assets made available by the de-
cree were all patents issued to and applied for by Eli Lilly during the five-
year period. 147 One limitation on the broad decree was a provision requir-
ing that a licensee contribute to Lilly's R&D expenses if asked to do so. 148
Because the order was so broad, providing for an unlimited amount of li-
censing covering both extant and future patents on any insulin technology,
it effectively prevented Lilly from exercising its patent rights over insulin
technology during the affected period. Faced with this severe version of
compulsory licensing, one might expect Lilly to have been discouraged
from further developing its insulin product for the five-year period set
forth in the order, or perhaps to have delayed patent applications until after
the period had expired, relying on trade secrets or other forms of protec-
tion in the interim.
However, Eli Lilly continued to dominate the emerging human insulin
market in performing R&D, surpassing major milestones during the period
from 1980 to 1985 covered by the consent decree. Following the initial
147. In re Eli Lilly & Co., 95 F.T.C. 538, 1980 FTC LEXIS 85, *17 (1980).
148. Id.
production of human insulin through recombinant DNA techniques in

1978, Lilly initiated clinical trials of its human insulin product, "Huma-
lin," in the United States in 1980, and invested in additional research fa-
cilities. 149 In 1982, Lilly was rewarded for its efforts, receiving the first
FDA approval for human insulin in the United States.150 Eli Lilly was ac-
tually more active in filing for patents during the five-year period of the
decree than it was during the previous and subsequent five years com-
bined.
Several factors seem to have motivated Eli Lilly's continued innova-
tion during this period. First, and perhaps most importantly, Lilly was ex-
tremely well-positioned to exploit and benefit over the long-term from the
genetics revolution emerging at the time. The company maintained its
early lead into the testing and commercialization phases of insulin, and
over the decade following expiration of the order was usually first or sec-
ond to introduce products of increasing purity on the market. 1 ' In addi-
tion, Lilly historically enjoyed a position of market leadership. As it stated
in the 1984 Annual Report: "With our historical position in diabetes, and
the patients we serve, it is clear we have to aggressively go out and look at
proinsulin. If it is potentially better, then we have
152
an obligation to bring it
forward. We owe this to society and humanity."'
Furthermore, insulin has long been one of Eli Lilly's most important
products. Shortly after Lilly took the first license to insulin in 1923, insu-
lin accounted for half of the company's profits, and it was the company's
second largest revenue producer in 1994.153 Finally, Lilly has faced con-
tinuous pressure from its main competitor, Novo Nordisk. In 1980 the two
companies together controlled 77% of the insulin market-53% by Eli Lilly
and 24% by Novo Nordisk. By 1995, the figure rose to over 90%, with Eli
149. See ELI LILLY, 1981 ANNUAL REPORT 5-6 (1982) [hereinafter ELI LILLY 1981
ANNUAL REPORT].
150. See, e.g., A Market Face-offfor Two Insulin Pioneers, Bus. WK., Nov. 1, 1982.
151. See, e.g., CLAYTON M. CHRISTENSEN, ELI LILLY AND COMPANY: INNOVATION IN
DIABETES CARE 4-5 (Harv. Bus. Sch., Case Study no. 9-696-077, 1996).
152. See ELI LILLY, 1984 ANNUAL REPORT 17 (1985) [hereinafter ELI LILLY 1985
ANNUAL REPORT]. Eli Lilly's position of market leadership began in 1923, with its exclu-
sive license over the manufacture of insulin with the University of Toronto, where Nobel
Prize Winner Frederick Banting did his ground-breaking work. See Irving S. Johnson,
Human Insulin from RecombinantDNA Technology, 219 SCIENCE 632 (1983).
153. See CHRISTENSEN, supra note 151, at 1, 3. Insulin has continued to be a high
revenue generator, despite being viewed as a commodity product because of significant
barriers to entry including the high cost of clinical trials for new biotechnology products
and the cost of efficient manufacturing facilities. See id. at 4.
1 54
Lilly capturing 46% and Novo Nordisk capturing 45% of the market.
The pressures generated by market leadership, a desire for market domi-
nance, and competition provided Lilly significant motivations to keep in-
novating, notwithstanding the temporary suspension of patent rewards.
The case study of Connaught's business also involves predictable li-
censing. At the time it received the FTC order, Merieux was the sole sup-
plier of rabies vaccines in the United States. Prior to the order, Merieux
had acquired the company Connaught, one of two potential entrants into
the market. 155 Worried that Merieux's monopoly would remain unchal-
lenged, the Commission called upon Merieux to lease Connaught's entire
rabies vaccine manufacturing business, including both tangibles and in- 156
tangibles, to an approved lessee for a minimum of twenty-five years.
However, Merieux was unable to find a suitable buyer for Connaught's
manufacturing business. 157 Nearly four years after the decree
58
had issued,
the FTC withdrew the leasing requirement from the order.1
The requirement that Merieux lease Connaught presumably reduced
Merieux's incentive to invest in Connaught's facilities. Merieux stated as
much in its request to the FTC that the leasing requirement be dropped.
The company claimed that "the continuing lease requirement may be
harmful to competition ...because it adversely affects Connaught's abil-
ity to respond to the increased demand for vaccine with capital invest-
ments to upgrade and expand the business's productive capacity."' 59 Al-
though evidence of this decline was not provided in the consent order,
Connaught did not file any patents for rabies vaccine inventions during the
contested period. In contrast, Merieux (which ultimately became Aventis)
filed five such patents in the subsequent years.
One possible reason that Connaught temporarily discontinued patent-
ing is its potential entrance in the U.S. market, over which Merieux had a
stronghold. In this light Merieux may have viewed any enrichment of
Connaught's business as tantamount to enriching a potential competitor in
the same market. Even though the consent decree was flexible enough to
enable Merieux to recoup any improvements it made to the Connaught
business, given that the decree called for a reasonable lump sum payment
by the licensee, Merieux's competitive interests arguably created an incen-
154. Id.at 17 exh.9.

155. In re Institut Merieux S.A., 113 F.T.C. 742, 1990 FTC LEXIS 291, *3-4 (1990).
156. Id.at 7-9.
157. In re Institut Merieux S.A., 117 F.T.C. 473, 474-75 (1994) (modifying 1990
order).
158. Id.at 482.
159. Id. at 477.
tive to neglect Connaught's rabies vaccine business while enriching its

own. Indeed, during the same period in which Connaught did not file for
any patents, Merieux sustained its lead in the rabies vaccine business, fil-
ing for a patent in late 1991, and launching a new product, Raboral, in
1992.16°
D. Results
These results, although limited, lend support to the theory that only
drug licenses that issue predictably in significant markets are likely to
harm innovation. Of the six companies subjected to compulsory licensing,
Merieux was the only one that exhibited a decline in patenting. Merieux's
licensing event was the only one that was both anticipated and affected a
market that was significant to the company. Although the data used in this
study cannot prove that the licensing event caused Merieux's decline in
patenting, it does indicate how pharmaceutical companies might react to
these types of compulsory licenses.
Table 2: Summary of Antitrust Case Results
Antitrust Licenses Sporadic Nature of Market Perceived or Actual

v. General Affected Negative Impact?
License
Baxter Sporadic Developed No
Dow, Roche Sporadic Developed No
Chiron/Ciba-Geigy Sporadic Nascent No
Roche Sporadic Nascent No
Eli Lilly General Nascent No
Merieux/Connaught General Developed Yes
The results of this study are contrary to the prevalent assumption that
compulsory licensing categorically harms innovation. Were the assump-
tion true, all six cases would reveal a drop in investment in innovation
subsequent to licensing, yet no such uniform downward trend was ob-
160. See Merieux Doubles Profits,Sees Growth in Rabies Vaccine, AGENCE-FRANCE-

PRESSE, Mar. 16, 1992 [hereinafter Merieux Doubles Profits], available at 1992 WL
8462395.
served. In fact, the opposite seems to be true-in all cases but one, activi-
ties of innovation continued at the same or even higher pace than before
the advent of a license. These results cast doubt on concerns that compul-
sory licensing is uniformly deleterious.
The study also suggests that, notwithstanding the absence of a uniform
downward trend, the circumstances surrounding a compulsory licensing
event can impact innovation. Where a license is predictable and the market
it affects is significant, a negative impact on innovation may be possible.
More caution may be in order when such licenses are contemplated over
patents held by companies or individuals who depend on patent profits.
VI. IMPLICATIONS FOR DRUG LICENSING IN

DEVELOPING COUNTRIES
As discussed in the preceding section, at least two factors may influ-
ence whether compulsory licenses impact pharmaceutical innovation.
These factors, namely the predictability of the license and the significance
of the affected market, have implications for the compulsory licensing of
drugs by developing countries.
An important consideration in determining whether compulsory li-
censes taken by developing countries will impact innovation is the type of
drug licensed. Developing countries care about two categories of drugs,
each with its own set of incentives.16
" ' First, there are "global" drugs that
are created for rich markets, but are also useful in developing countries.
Examples of these are cancer drugs and AIDS therapeutics.' 62 Second,
there are drugs specific to developing countries. Examples of these include
drugs to treat malaria or tuberculosis, or an AIDS vaccine specific to
strains of the virus found primarily in Africa.' 63 Historically, such drugs
have not been the priority of pharmaceutical companies. For example, a
2001 Harvard School of Public Health survey of twenty large pharmaceu-
tical firms found that "[o]f 1 responders, eight had done no research over
the past year in tuberculosis, malaria, African sleeping sickness, leishma-
niasis, or Chagas disease; seven spent less than 1% of their research and
development budget on any of these disorders."' 64
Funds for researching diseases specific to developing countries often
come from public or philanthropic resources such as the Centers for Dis-
161. See LANJOUW, supra note 91.

162. See id.
163. See id.
164. Ricki Lewis, Fighting the 10/90 Gap, SCIENTIST, May 13, 2002, http://www.the-
scientist.com/yr2002/may/lewis_p22_020513.html (last visited July 22, 2003).
ease Control or public-private partnerships like those created by the Inter-

national AIDS Vaccine Initiative. The Medicines for Malaria Venture, for
instance, matches academic researchers with private firms to generate col-
laborations in malaria medicines, an area that has largely been overlooked
by industry. 65 The Global Alliance for Tuberculosis Drug Development,
partly sponsored by the Rockefeller Foundation, similarly tries to shift
product development risk away from drug companies by conducting costly
clinical trials for promising drug candidates. 66 Efforts to develop67an AIDS
1
vaccine for countries in Africa have likewise been collaborative.
Research to date suggests that if compulsory licenses are taken in less
significant markets, their impact on innovation should be marginal. For
global drugs such as AIDS therapy, this would imply that compulsory li-
censes that are limited to developing countries (i.e. ancillary markets) and
do not impact the target markets for the drugs (i.e., rich countries) might
not be detrimental to research efforts in the rich developed countries. This
is in accord with common sense. For global drugs, companies are respon-
sive to the incentives provided by wealthy markets and consumers. If these
incentives stay intact, selective compulsory licensing for developing na-
tions should have little impact on overall R&D investment as long as the
affected market is limited to developing countries.
On August 30, 2003, a historic accord on compulsory licensing was
announced addressing this concern. After several days of negotiations, the
United States and other WTO countries effectively agreed to allow poor
countries to import generic drugs through compulsory licenses as long as
measures to prevent re-exportation of the drugs to other, rich markets are
taken.' 6 8 For example, such measures include special packaging or color-
ing to clearly delineate drugs that have been exported under compulsory
165. Martin Enserink MalariaResearchers Wait for Industry to Join the Fight, 287
Sci. 1956, 1958 (2000).
166. Exotic Pursuits,ECONOMIST, Jan. 30, 2003, availableat 2003 WL 6244875.
167. See, e.g., Alexandra Bojak et al., The Past, Present,and Future ofHIV- Vaccine
Development: A Critical View, 7 DRUG DISCOVERY TODAY 36, 41-43 (2002) (describing
the funding of basic research by the European Union and other governments and the
continued need for collaboration between rich and poor countries and for money from
donor organizations such as the European Vaccine Efforts); Paul J. Weidle, et al.,
HIV/AIDS Treatment and HIV Vaccinesfor Africa, 359 LANCET 2265 (2002) (describing
HIV vaccine trials in Africa as sponsored by the NIH, CDC, lAVI, and other public and
philanthropic organizations).
168. See, e.g., WTO, The General Council Chairperson's Statement, Aug. 30, 2003
[hereinafter WTO Aug. 30 Statement], at http://www.wto.org/english/newse/news03_e/
tripsstat_28aug03_e.htm (last visited Sept. 6, 2003); Becker, supra note 87, § 1, at 14.
licenses from drugs sold in rich countries.1 69 The United Staies has de- 170
manded that the scope of the accord cover life-threatening diseases.
While some details about its coverage are yet to be resolved,17 the accord
is indisputably intended to reach AIDS therapy drugs..
The implication is somewhat different for drugs developed to treat dis-
eases endemic to developing countries, such as malaria. As discussed
above, much of the research on these diseases is carried out or facilitated
by public or philanthropic institutions, for whom patent protection and the
promise of a patent monopoly are less, if at all, important. In addition, the
potential for monopoly abuse which compulsory licenses are designed to
counter could also be less likely. If pharmaceutical companies, on the
other hand, begin investing significantly in such disease areas due to the
introduction of patent protection, as is hoped, a compulsory license cover-
ing all developing country markets might well usurp the primary target
markets. Threatening or implementing licenses on a regular, predictable
fashion may deter pharmaceuticals from initiating and carrying out R&D
investments.
Based on these observations, and focusing exclusively on innovation
concerns, one can make a preliminary case for employing different ap-
proaches to compulsory licensing depending on whether global or devel-
oping country-specific drugs are licensed. Because the relative importance
of developing country markets is small when it comes to global drugs, the
incentive to develop these drugs, which comes from the developed world,
is not substantially impacted. This means that allowing developing coun-
tries to take compulsory licenses to AIDS therapy drugs should not pro-
duce a negative impact on AIDS therapy research and development. The
recent WTO accord is entirely appropriate in this regard.
The picture is different when it comes to drugs being developed spe-
cifically to treat developing country diseases, such as AIDS strains en-
demic to Africa. Compulsory licenses for developing countries could
cover the entire target market of local and international pharmaceuticals.
The threat of systematic compulsory licensing of these drugs may make a
difference and could cause some companies to avoid these markets alto-
gether. To the extent that the compulsory licensing framework that devel-
169. WTO Aug. 30 Statement, supra note 168.

170. Becker, supra note 87, § 1,at 14.
171. Editorial, WTO Takes a First Step, 362 LANCET 753 (Sept. 6, 2003) (arguing for
an interpretation of the agreement broader than including only drugs to treat HIV/AIDS,
tuberculosis, and malaria), available at http://pdf.thelancet.com/pdfdownload?uid=llan.
362.9386.editorialandreview.27086.1 &x=x.pdf.
1 72
ops under the WTO accord also covers such diseases, as it likely will,
special care should be taken to ensure that incentives remain intact. To
date, the patent incentive has arguably not successfully prompted R&D in
these medicines. Therefore, the importance of preserving the current pat-
ent incentive should not be overstated. This is particularly true because of
the strong role played in this area by public and philanthropic institutions,
which presumably are not motivated by monopoly profits.
Compulsory licensing is far from an easy solution; exploiting it fully
requires political will and technical capability. In the past, countries that
have elected to take licenses have had to endure lawsuits, pressure, and
threats of trade sanctions from the United States.' 73 In addition, producing
drugs pursuant to a license requires a level of technical and manufacturing
capability possessed by few countries. 74 The August 2003 WTO accord
significantly deals with these issues. Still, meeting the accord's require- 75
ments for licensing could prove challenging, or at least bureaucratic.
Over-reliance on compulsory licensing may also produce unintended
negative downstream impacts on society.
While high drug prices comprise only one aspect of the AIDS prob-
lem,176 the WTO accord evidences the growing realization that increasing
access to drugs must be a part of the solution. This is partly due to a num-
ber of factors that have shifted attention towards affordable treatment and
vaccination rather than prevention alone.' 77 The initial push for prevention
172. See id.

173. The U.S. and western pharmaceutical companies have routinely used the Special
301 mechanism for authorizing trade sanctions and lawsuits at the WTO and in domestic
courts to oppose policies implemented by other countries that are unfavorable to pharma-
ceutical company interests. See, e.g., Sarah Boseley, How the U.S. Wields a Big Stick for
Big Pharm, GUARDIAN, Feb. 18, 2003 (describing actions against Thailand);
Carin Hkansta, The Battle on Patents and AIDS Treatment, 16 BIOTECH. AND DEV.
MONITOR 34 (1998) (describing early battles against India in the TRIPS court), available
at http://www.biotech-monitor.nl/3406.htm (last visited Aug. 27 2003).
174. See Attaran & Gillespie-White, supra note 5 (describing non-patent barriers to
drugs,. such as insufficient finances, lack of political will, poor medical care and infra-
structure, inefficient drug regulatory procedures, and high tariffs and sales taxes).
175. Becker, supra note 87, § 1, at 14 (describing the concerns of some groups that
"red tape" will discourage use of compulsory licenses).
176. Some have suggested that cheap drugs might actually aggravate the problem by
diverting attention away from prevention. See, e.g., Michael Specter, India's Plague,
NEW YORKER, Dec. 17, 2001, at 74.
177. See, e.g., Hope for the Best. Preparefor the Worst-the Future of Aids.,
ECONOMIST, July 11, 2002 (describing the growing intensity in AIDS vaccine research in
light of the limitations of prevention and treatment), availableat 2002 WL 7246756; Mi-
chael Specter, The Vaccine, NEW YORKER, Feb. 3, 2003, at 54.
was based on the conventional wisdom that prevention (through education,

the empowerment of women, and distribution of condoms) is the best cure
for the AIDS problem, and that AIDS therapy regimes were too expensive
and complicated to be suitable for developing countries. However, in the
past two years, experience has shown that people in developing countries
can and will comply with drug regimes at levels equal to or greater than
their Western counterparts. At the same time, there has been a realization
that attitudes and culture are hard to change and that solutions other than
behavioral transformation must be explored. 178 In addition, the availability
of drugs is crucial not only for treating sick patients but also for diagnos-
ing and stopping the spread of AIDS because "people who are infected
and cannot be treated have little incentive to get tested; that, in turn, means
they do not know they are infected, and so do not take precautions against
infecting others.' ' 179 All of these factors
18
have made access to AIDS drugs a
more pressing and realistic objective. 0
Still, the focus on cheap drugs for therapy today should not draw atten-
tion away from the hope of an AIDS vaccine tomorrow. How compulsory
licensing programs are designed and implemented matter in this regard,
and the emerging regime of compulsory licensing deserves continued at-
tention in this respect.
VII. CONCLUSION
Although modest, the data analyzed in this study yield potentially
surprising implications for the current debate over compulsory licensing.
At a minimum, they challenge the wholesale rejection of licensing
schemes for AIDS drugs based on their perceived negative impact on
178. See Michael Grunwald, All-Out Effort Fails to Halt AIDS Spread; Botswana's
Program Makes Progress, But Old Attitudes Persist,WASH. POST, Dec. 2, 2002, at A1;
Rosenberg, supra note 7.
179. Hope for the Best. Preparefor the Worst-the FutureofAids, supra note 177.
180. Even before the August 2003 WTO accord, the pharmaceutical industry volun-
tarily reduced prices on a number drugs in recognition of the humanitarian crisis. See,
e.g., Geoff Dyer, How Do You Price AIDS Treatment?, FIN. TIMES, Mar. 26, 2003, at 13
(describing Roche's statement that it will not enforce intellectual property rights on its
AIDS drug Fuzeon, priced at $20,000 a year per user, in sub-Saharan Africa); Grunwald,
supra note 178, at Al (describing Merck's offer of an unlimited supply of antiretroviral
drugs to Botswana); Paul Jacobs, Gilead Unveils AIDS Drug Plan, SAN JOSE MERCURY
NEWS, Apr. 4, 2003 (describing how Gilead Sciences plans to offer its successful AIDS
drug Viread to 68 developing countries at substantially reduced prices), available at 2003
WL 14985084; see also Geoff Dyer, Investors Warn Drugs Industry of Backlash over
Health Crises, FIN. TIMES, Mar. 24, 2003, at 25 (describing investor pressure for price
cuts).
AIDS innovation. They also invite consideration of how compulsory li-

censes are designed and implemented. They suggest that, based on innova-
tion concerns, the use of different kinds of licenses over global and devel-
oping country diseases may be appropriate given the different incentives
driving innovation in these areas.
VIII. APPENDIX: ANTITRUST LICENSE CASE STUDIES

A. Baxter/Fibrin Sealant
1. The Order
In early 1997, the FTC ordered Baxter to license its rights to fibrin
sealant, a topical agent used to control surgical bleeding, in connection 18
with Baxter's acquisition of Immuno International AG ("Immuno"). 1
Although the product had been available in Europe for several years be-
fore the order, Baxter and Immuno were two of just a handful of compa-
nies seeking FDA approval for the first product launch in the U.S. mar-
ket,' 82 estimated
83
shortly after the order to be worth up to $200 million an-
nually. 1
The part of the order pertaining to fibrin sealant required Baxter to
provide a license to all of Immuno's intangible assets and rights (including
patents, trade secrets, technology, know-how, specifications, customer
lists, and FDA approval data) related to the R&D, manufacture, and sale
of fibrin sealant. 184 The order mandated one commission-approved licen-
see.185 Once Baxter obtained FDA approval, the order required the com-
pany to supply the licensee with Immuno's fibrin sealant product until the
licensee received approval for its own product.' 86 In exchange, the order
required the licensee to reimburse Baxter for the costs of manufacturing
the fibrin sealant product while demonstrating a continuing commitment 87
to obtain approval from the FDA for its own fibrin sealant product.'
Other portions of the order called for the divestiture of Immuno's Factor
VIII assets. 188
181. In re Baxter Int'l Inc., 123 F.T.C. 904 (1997).

182. Id.at906.
183. See Gary Shepherd, To Haemacure, 'Scab' is Not a Four-Letter Word, TAMPA
BAY Bus. J.,Sept. 11, 1998, available at 1998 WL 33483931.
184. Baxter, 123 F.T.C. at910, 921.
185. Id.at 921.
186. Id.
187. Id.
188. Id.at 910-916.
BERKELEY TECHNOLOGY LAW JOURNAL (Vol. 18:853
Within six months of the order, the FTC approved Haemacure as

the fibrin sealant licensee.' 8 9 A little over a year after the order was issued,
in May 1998, Immuno's fibrin sealant received FDA approval, and both 90
Baxter and Haemacure introduced the product into the U.S. market.
Over the next few years, Baxter and Haemacure were the only sellers of
fibrin sealant in the U.S. market, with Baxter capturing 75% of the market
in 1999.19' The FTC approved several requests by licensee Haemacure to
extend the license, and in 2002, Haemacure estimated that the license
could expire in 2004.192
2. Impact on Innovation
Baxter's interest in pursuing follow-on innovation and products
could have hypothetically declined with their loss of exclusive control
over the market because the license required them to share late-stage tech-
nology and profits with Haemacure. On the other hand, the late stage of
the technology possibly reduced other uncertainties associated with the
technology, and the license potentially provided the chance to capture ad-
ditional revenue with little investment.
After the order, Baxter continued to invest in fibrin sealant and re-
lated therapeutic areas. This is shown by their patenting activity, new
product development, and clinical trials. The company filed about as many
patent applications for this technology in the five years following the order
as in all years prior to it. 193 Additionally, Baxter introduced a follow-on
application device,1 94 and worked on a patch, rather than liquid, version of
the product. 195 Also, the company conducted clinical trials on a hemostatic
189. See For Your Information, Fed. Trade Comm'n Office of Public Affairs, Aug. 1,
1997, at http://www3.ftc.gov/opa/1997/08/petapp40.htm (last visited Aug. 11, 2003).
190. See Christiane Truelove, Baxter's Bloodline, MED. AD NEWS, Sept. 1, 1999,
available at 1999 WL 12977876; Haemacure Announces Fiscal Year 2001 Financial
Results, CAN. NEWSWIRE, Jan. 11, 2002, at http://www.newswire.ca/releases/January
2002/11/c 1986.html (last visited Aug. 11, 2003).
191. See HAEMACURE CORP., 1999 ANNUAL INFORMATION FORM 8 (2000),
http://www.haemacure.com/pdf/infoform/Ai280400a.pdf (last visited Aug. 11, 2003).
192. See HaemacureAnnounces FiscalYear 2001 FinancialResults, supranote 190.
193. This finding is based on a date and assignee search of the LEXIS patent applica-
tion database using the keyword "fibrin sealant." The company successfully obtained six
patents prior to the order, and obtained five patents after it.
194. See News Release, Baxter, Baxter's New Fibrin Sealant Application System
Cleared by FDA (July 10, 2000), available at http://www.baxter.com/utilities/news/
releases/2000/07- 1Otissomat.html (last visited Aug. 11, 2003).
195. See Glenn M. Reicin & Jason H. Wittes, A Feel-GoodAnalyst Meeting, Morgan
Stanley Dean Witter, Mar. 23, 2001, at 8 (on file with author).
sealant, which it subsequently introduced on the market.' 96 In 2000, the

company announced a $400 million commitment to upgrade facilities used 97
for fibrin sealant and other plasma and recombinant DNA products.'
Overall, Baxter did not appear to reduce its investment in R&D in fibrin
sealant and related products.
Product economics may explain Baxter's interest and aggressive
marketing concerning fibrin sealant. According to a Lehman Brothers re-
port issued in late 1999, the projected revenue compounded annual growth
rate for fibrin sealants was over 35%, which was the highest rate in Bax-
ter's blood product division. 198 Additionally, Baxter was projected to cap-
ture nearly 60% of the international fibrin sealant market by 2001, of
which approximately half was estimated to come from the United
States. 199 Other companies, including Aventis, Omrix, the American Red
Cross, Vitex, and a Bristol-Meyers Squibb subsidiary also developed fi-
brin sealants. 20 0 Thus, Baxter was plausibly motivated to capture the first-
mover advantage in the years just after the license was ordered.
B. Marion Merrell Dow/Dicyclomine
1. The Order
In late 1994, the FTC ordered Dow to license its rights to di-
cyclomine, a product used for the treatment of IBS. 20 At the time of the
order, Dow's product was already on the market with only 60% of the $7
to $8 million market, due to generic competition. 20 2 Dow's acquisition of
Rugby Darby, the only generic company 203
approved to manufacture the
drug at the time, raised antitrust concerns.
Given that Rugby Darby already made dicyclomine, some key pat-
ents were presumably expired, but other barriers, described by the FTC
order as "difficult and time consuming," prevented other generic compa-
nies from entry. 204 Accordingly, the order required Dow to provide a per-
196. According to Biospace's CCIS database (based on date search and keyword
"sealant").
197. See supranote 196.
198. See David A. Gruber et al., A PotentialSource of Stability, Lehman Brothers,
Oct. 6, 1999 (on file with author); Reicin & Wittes, supra note 195, at 14.
199. See Reicin & Wittes, supra note 195, at 14.
200. See HAEMACURE CORP., supra note 191, at 8.
201. In re Dow Chem. Co., 118 F.T.C. 730, 736-38 (1994).
202. See Lannett Expects IncreasedSales andProfit with the Launch of its Third and
Most Significant GenericDrug Product,Bus. WiRE, July 21, 1994.
203. Dow, 118 F.T.C. at 732-33.
204. Id. at 732.
petual license to a commission-approved licensee, including all formula- 20 5

tions, patents, trade secrets, technology, know-how, and specifications.
The order did not specify a price, but stated that there was "no minimum
price" for the license, implying that no potential deal could be rejected on
the basis of price. The order also required Dow to provide manufactured
dicyclomine to the licensee until it received FDA approval. In exchange,
the order mandated that the licensee pay up to 48% of the wholesale price
of the dicyclomine, while attempting to obtain FDA approval for its own
manufacturing facilities.
2. Subsequent Developments
Within one year of the order, Hoechst Marion Roussel, Inc.
("HMRI"), Dow's successor through merger, requested and obtained FTC
approval to award a license to Endo Laboratories, a subsidiary of Dupont
Merck Pharmaceutical Company. 206 During this time, other generic com-
panies entered the dicyclomine market, 20 7 causing the IBS market to ma-
ture by the last half of the 1990s, with few new product introductions until
2000. 20' Although dicyclomine was one of just three main products in the
anti-cholinergic and anti-spasmodic segment of IBS medications, other
medication segments such as anti-diarrhea and constipation were more
important. 20 9 Dicyclomine's share of the overall IBS medication market
eventually dwindled to less than 2% in 2000.210
Even though generic manufacturers were producing Dicyclomine
at the time of the order, the license conferred advantages principally asso-
ciated with the head start that Dow achieved over the generic manufactur-
ers. It is unclear whether much opportunity existed for innovation in this
particular therapeutic area, because the product was already mature. Still,
assuming that Dow and its successor HMRI had to make a decision about
the IBS market, its weak market position and the relatively small size of
the dicyclomine market could potentially deter any future investment.
205. Id. at 736.

206. See For Your Information, Fed. Trade Comm'n Office of Public Affairs, July
19, 1996, http://www3.ftc.gov/opa/1996/07/petapp41.htm (last visited Aug. 11, 2003).
207. Id.
208. See Stewart Adkins et al., IrritableBowel Syndrome, Poetry in Motion, Lehman
Brothers, Sept. 2, 1999, at 19 (on file with author).
209. See id. at 6.
210. See Jeffrey Chaffkin et al., Company & Therapeutic Prescription Statistical
Update, Paine Webber, Oct. 23, 2000, at 96 (on file with author).
2 11
A year after the license issued, HMRI stopped filing for patents.
However, it is difficult to attribute this absence in activity only to the li-
cense, because the year following the license was the most productive year
in terms of number of patent applications, with six patent applications
filed. Although it is possible that these were merely the result of pre-
license innovation activities, the presence of other factors such as dwin-
dling market share and the earlier loss of patent protection could each
plausibly explain the rise and then discontinuation of patenting. Even
though little can be definitively concluded about this case, it does not ap-
pear that licensing alone entirely explains HMRI's exit from the R&D IBS
market.
C. Eli Lilly/Insulin
1. The Order
In 1980, the FTC charged Eli Lilly with involvement in a wide-
ranging conspiracy, dating back to 1952, with other manufacturers of pan-
creatic insulin. 212 The FTC ordered the firm to license the know-how and
rights relating to both its existing and future insulin-related patents. 213 Any
potential entrant who, within five years of the decree, stated a bona fide
intention to produce and sell insulin products in the United States would
obtain access to Lilly's intangible assets, including all patents issued and
applied for during the five-year period.214 Significantly, Lilly could im-
pose a charge on the licensee equal to a "[r]easonable pro rata share of the
amounts actually spent by Lilly in acquiring, or financing the research and
development... [of] such licensed patents and know-how," in addition to
a requirement to give grantbacks. The 216
order also required licensees to
keep all production in the United States.
No information is available on whether any companies came for-
ward and took advantage of the compulsory license made available by the
211. While other companies have continued to develop IBS medications, with sixteen
products in clinical trials during the 2000-2002 period according to Biospace,
Dow/HMRI has not participated in any reported drug development activities. Search con-
ducted by the author of the Biospace database using keywords "irritable bowel syn-
drome" or "dicyclomine."
212. In re Eli Lilly & Co., 95 F.T.C. 538, 1980 FTC LEXIS 85, *5 (1980).
213. Id. at *17, *23.
214. Id.at*17.
215. Id.at*24.
216. Id.at*23.
consent decree. 217 However, Lilly continued to dominate the emerging

human insulin market in both research and development, surpassing major
milestones during the five-year period covered by the consent decree. In
1980, following the initial production of human insulin through recombi-
nant DNA techniques in 1978, Lilly initiated clinical trials in the United
States of its human insulin product "Humalin" and invested in research
facilities to carry out additional work.2 18 In 1982, the FDA rewarded Lilly
for its efforts with the first approval for human insulin in the United
States.219
The broad order, covering future patents issued on any insulin
technology and allowing a potentially large number of licensees, effec-
tively prevented Lilly from obtaining patent protection over its insulin
technology during the affected period. Faced with this severe version of
compulsory licensing, the company was potentially discouraged from any
innovation in insulin technology during the five-year period. Additionally,
Lilly probably at least delayed patent applications until after the licensing
period, relying instead on trade secret or other forms of protection. The
one significant mitigating factor, however, was the license's provision that
the licensee could be asked to contribute to the R&D expenses. 220
Based on a few indicators, Lilly continued to aggressively pursue
insulin R&D during the period covered by the license. For example, pat-
enting behavior did not appear to be affected. The company filed for seven
patents over the five-year licensing period, whereas fewer than seven pat-
ents were filed during the periods five years prior and subsequent to the
licensing event combined.221
Several factors seemed to motivate Lilly's continued innovation
during the licensing period. One is historical market leadership. The com-
pany's 1984 Annual Report states that "[w]ith our historical position in
diabetes and the patients we serve, it is clear we have to aggressively go
out and look at proinsulin. If it is potentially better, then we have an obli-
217. Because the license was made available to any domestic company with a bona
fide intention to enter the insulin market, and the licensee did not require FTC approval,
the FTC did not publicly track whether any licenses were implemented. The FTC would
only have intervened had there been a complaint of non-compliance. Telephone Interview
with Kenneth Davidson, Fed. Trade Comm'n Bureau of Competition (Apr. 26, 1995).
218. See ELI LILLY 1981 ANNUAL REPORT, supra note 149, at 5-6.
219. See A Market Face-offfor Two Insulin Pioneers,supra note 150.
220. See In re Eli Lilly & Co., 95 F.T.C. 538, 1980 FTC LEXIS 85, *24 (1980).
221. Search using keyword "insulin."
2003) CHEAP DRUGS AT WHAT PRICE TO INNOVATION
gation to bring it forward. We owe this to society and humanity." 222 An-
other factor is that Lilly was an early leader in the research leading to the
production of human insulin through recombinant DNA methods in 1978.
Through subsequent testing and commercialization, the company was of- 223
ten first or second to introduce products of increasing purity to market.
Likewise, insulin was always one of Lilly's most important products.
Shortly after the company took its first license in 1923, insulin accounted
for half of all Lilly's profits, and in 1994, it was still the company's sec-
ond largest revenue producer.224
Insulin continues to be a high revenue generator, despite being
viewed as a commodity product due to significant barriers to entry such as
the high cost of clinical trials for new biotechnology products and the cost
of an efficient manufacturing facility. 225 Finally, Lilly continues to face
continuous pressure from competitor Novo Nordisk; in 1980 the two
companies together held nearly 80% of the insulin market (53% by Eli
Lilly and 24% by Novo Nordisk), and by 1995, the two virtually split 91%
of the market (Eli Lilly capturing 46% and Novo 45% of the market).226
The pressures generated by market leadership, a desire for market domi-
nance, and competition provided significant motivations for Lilly to con-
tinue to innovate, even during the compulsory licensing period.
D. Connaught/Rabies Vaccine
1. The Order
In 1992, citing concerns about increased domination of the U.S.
rabies vaccine market, the FTC ordered Merieux to lease the rabies manu-
facturing business of the company it acquired, Connaught Bioscience.227
Merieux was the sole supplier of rabies vaccine in the United States, and
Connaught was one of two potential entrants into the market. Worried that
Merieux's monopoly would remain unchallenged, the Commission called
upon Merieux to lease Connaught's entire rabies vaccine manufacturing
business, including both the production facility and technology, to an ap-
proved lessee for a minimum of twenty-five years.228 In exchange, the or-
222. See ELI LILLY 1983 ANNUAL REPORT, supra note 152, at 17.
223. See CHRISTENSEN, supra note 151, at 1.
224. See id. at 1, 4.
225. See id. at 4.
226. See id. at 17, exh.9.
227. In re Institut Merieux S.A., 113 F.T.C. 742, 1990 FTC LEXIS 291,*8-9 (1990).
228. Id.
der provided for the lessee to give a lump sum

229
payment, under customary
and reasonable terms, to Merieux/Connaught.
Despite contacting twenty-eight prospective licensees over the next
several years, including serious negotiations with a few parties, Merieux
could not find a suitable buyer for Connaught's rabies manufacturing
business. 230 The most serious offer, from North American Vaccine, Inc.,
was rejected because of a lack of relevant experience. 2 1 1 In April 1994, the
FTC modified the original order and removed the leasing requirement, cit-
ing the entry of SmithKline Beecham into the market and Merieux's bona
fide attempts to satisfy the consent decree.232
During the period between the original and modified order,
Merieux had little incentive to invest in Connaught's facilities, given that
the order required it to lease the business away at uncertain prices. In fact,
in the consent modifying order, Merieux suggested that "the continuing
lease requirement may be harmful to competition ... because it adversely
affects Connaught's ability to respond to the increased demand for vaccine
with capital investments to upgrade and expand the business's productive
capacity.''233 Although evidence of this decline is not in the consent order,
no patents were filed by the would-be leased Connaught for rabies vaccine
inventions during the contested period, while five patent applications were
filed by Connaught-Merieux, which ultimately became Aventis, in the
subsequent years.
Based on the evidence, it appears likely that Connaught's rabies
vaccine practice suffered under this most extreme version of compulsory
licensing. Given Connaught's position as a potential entrant in the U.S.
market, the potential of enriching a competitor in the same market proba-
bly served as a major deterrent. Despite the continued interest of Merieux
in the rabies business, little motivation may have existed for them to invest
in Connaught. During the same interim period in which Connaught did not
file for any patents, Merieux remained active in the rabies vaccine busi-
229. Id.
230. In re Institut Merieux S.A., 117 F.T.C. 473, 474-75 (1994) (modifying the 1990
order).
231. Id.at 476.
232. Id.
233. Id. at 477.
234. Search using keywords "rabies" and "vaccine" with incidental mentions
screened out.
ness, launching a new product in 1992235 and filing for a patent in late
1981. All of this is consistent with Merieux's statement to the Commission
that the order adversely affected its incentives to maintain and improve the
Connaught manufacturing capabilities.
E. Chiron/HSV-tk Related Therapeutics
1. The Order
In early 1997, the merger of Ciba Geigy, which owned the largest
share of Chiron, and Sandoz concerned the FTC. Believing that the com-
bination would create a "killer" patent portfolio 236 concerning the herpes
simplex virus-thymidine kinase (HSV-tk) gene, the FTC ordered the com-
panies to license their patent portfolios to an approved licensee. 237 The
FTC was concerned that combining the patent portfolios would heighten
already existing barriers to entry in the market for HSV-tk gene therapy, in
which Chiron and Sandoz were leaders. 238 Anticipating that the combined
portfolio would reduce the parties' incentives to license their patents, the
order called for the licensing of other key gene therapy patents and divesti-
tures in unrelated areas.239 Unlike the other situations discussed here, this
decision seemed to be openly motivated by protecting public health in ad-
dition to protecting competition. FTC Bureau of Competition Director
William Baer even stated, "[t]his case is about saving lives. Today there
are two firms racing to develop new gene therapies to combat deadly dis-
eases. The deal threatened to eliminate that competition.
240
Our order ensures
that this sprint to the finish line will continue."
The order required the merging parties to offer perpetual rights to
their HSV-tk patent portfolios and provide related know-how to Rhone-
Poulenc Rorer ("RPR") or another approved licensee. In order to ensure
that a license would be issued, the FTC specified that compensation could
be in the form of an equivalent cross-license or a royalty. 24 1 Within six
months of the decree, the Commission approved the licensing of Chiron's
235. See Merieux Doubles Profits,supra note 160.

236. James B. Kobak Jr. & Richard P. McGuire, FTC Looks at Merger's Antitrust
Effects on R&D, NAT'L L.J., Mar. 31, 1997, at C3.
237. In re Ciba-Geigy Ltd., 123 F.T.C. 842, 873-77 (1997).
238. Id. at 864-73, 877-86.
239. Id. at 846-47.
240. See Press Release, Fed. Trade Comm'n Office of Public Affairs, FTC Accord in
Ciba Geigy/Sandoz Merger to Prevent Slowdown in Gene Therapy Development & Pre-
serve Competition in Corn Herbicides, Flea-Control Markets (Dec. 17, 1996), available
at http://www.ftc.gov/opa/1996/12/ciba.htm (last visited Aug. 27, 2003).
241. In re Ciba-Geigy Ltd., 123 F.T.C. at 873-77.
HSV-tk portfolio to multiple companies in fulfillment of the order.242 In

exchange, one licensee, Novartis, paid Chiron $60 million in addition to
cross-licenses to some of its technologies.243
Chiron's loss of exclusivity and "killer" patents over HSV-tk tech-
nologies potentially dampened its enthusiasm and willingness to invest in
additional research. However, the opportunities presented by the cross-
licenses given by Novartis, in addition to Chiron's market leadership posi-
tion, probably mitigated any such effect. Chiron was likely not interested
in scaling back research merely based on the speculative downstream im-
pact of a licensee.
In the years following the order, Chiron continued to patent HSV-
tk technologies at a rate comparable to its filings before the order."' Addi-
tionally, in 1998, the company reported that it had two products in devel-
opment, one for graft-versus-host disease and another for hemophilia A.245
Over the 2000 to 2002 period, Chiron and Novartis were involved in two
of the fifteen trials reported in the Biospace CCIS database. 246 Meanwhile,
licensee RPR, which became part of the larger pharmaceutical entity
Aventis, appeared to make strides in the gene therapy market. The com-
pany launched RPR Gencell to develop gene therapies for cancer and
other diseases in collaboration with other companies.247 According to
these measures, the license does not appear to have significantly harmed
Chiron's innovation.
F. Roche/CD-4
1. The Order
In late 1990, the FTC ordered Roche, in connection with its acqui-
sition of Genentech, to license its future rights to pending patents covering
CD4-based technologies. 248 The order narrowly defined the relevant mar-
242. See For Your Information, Fed. Trade Comm'n Office of Public Affairs,
Sept. 12, 1997, http://www.ftc.gov/opaJ1997/09/petapp50.htm (last visited Aug. 11,
2003).
243. See Chiron to License RPR for HSV-tk Gene and Cross-License Technologies
with Ciba-Novartis,BUS. WIRE, Dec. 17, 1996.
244. Search using keywords "gene therapy", "retroviral vector", and "HSV."
245. CHIRON CORP., 1998 ANNUAL REPORT 9 (1999).
246. Search using keyword "HSV."
247. Jim Papanikolaw, Waiting for the Fruits of Gene Therapy, CHEM. MKTG. REP.,
Mar. 20, 2000.
248. In re Roche Holding Ltd., 113 F.T.C. 1086, 1990 FTC LEXIS 543, *25 (1990).
ket as "CD-4 based therapeutics for the treatment of AIDS and HIV infec-
tion. ' '249 At the time, Genentech led the market with a product in clinical
trials, with Roche following along with several patent applications. How-
ever, given the early stage of the technology, the merging
250
parties were at
most only potential competitors in the marketplace.
The order provided perpetual access to Roche's patents in ex-
change for 1% of net sales for process patents and 3% of net sales for
product patents. The license could be requested by any competitor or po-
tential entrant over the ten years following the order, subject to its continu-
ing commitment to CD4 research. 25 '
In the years following the order, Roche's patenting activity252 far
outperformed its pre-order levels. This is not surprising given that Roche
only began to file for patents shortly before the order. In addition, the
company remained committed to the investigation of CD4-based therapeu-
tics in the treatment of AIDS. During the two-year period from 2000 to
2002, the company was a partner in four of the twelve clinical trials of
drugs, three with Genentech and another with Baxter International.253 Ac-
cordingly, the order did not significantly affect Roche's CD4 HIV re-
search.
To the extent that Roche relied on its patents to secure its competi-
tive position in the CD4-based therapeutic market, the patent weakening
license potentially discouraged Roche from investing as heavily as without
the license. Even without the license, Roche presumably could have de-
cided to abandon its own efforts relying instead on the innovation of
leader Genentech. However, given the early stage of CD4 therapeutic de-
velopment, Roche most likely decided that the compulsory license posed
little threat in the ultimate therapeutic market.
249. Id. at *3.

250. Id.at *4.
251. Id at*25-26.
252. Search using keywords "CD4" and "viral."
253. According to Biospace's CCIS database.
BERKELEY TECHNOLOGY LAW JOURNAL

18 Berkeley Tech LJ 853

Uploaded by

Copyright:

Available Formats

18 Berkeley Tech LJ 853

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

18 Berkeley Tech LJ 853

Uploaded by

Copyright:

Available Formats

DATE DOWNLOADED: Sun Apr 9 13:39:11 2023

SOURCE: Content Downloaded from HeinOnline

Bluebook 21st ed.

ALWD 7th ed.

APA 7th ed.

Chicago 17th ed.

McGill Guide 9th ed.

AGLC 4th ed.

MLA 9th ed.

OSCOLA 4th ed.

© 2003 Colleen Chien

III. THE COMPULSORY LICENSING OF DRUGS ............................................................... 864

Related Aspects of Intellectual Property Rights ("TRIPS"), 6 compulsory

In five of the six cases I studied, I observed no measurable decline in

II. COMPULSORY LICENSING OVERVIEW

9. See SCHERER & WATAL, supra note 3, at 12.

tion while increasing access to innovations themselves, the Congress

acceptance of the principle of compulsory licensing--of compelling all patentees to li-

B. United States versus Developing Country Perspectives on

patent system facilitates the transfer of monopoly rents to foreigners out-

source data for fig.6-27, at http://www.nsf.gov/sbe/srs/seind02/c6/figO6-27.xls); U.S.

her invention,34 the U.S. government routinely relies on 28 U.S.C. § 1498

III. THE COMPULSORY LICENSING OF DRUGS

48. See Tamar V. Terzian, Direct-to-ConsumerPrescriptionDrug Advertising, 25

regulation has been contemplated intermittently since the 1950s. In the

61. See SCHERER& WATAL, supra note 3, at 26.

vision permits WTO member countries to authorize compulsory licenses

72. TRIPS Agreement art. 31 (h).

We stress the importance we attach to implementation and inter-

In addition, the Doha Declaration clarifies that member countries may

the Declaration as an "important achievement" that "broke new ground in guaranteeing

IV. EMPIRICAL BACKGROUND (LITERATURE REVIEW)

86. Doha Declaration, supra note 83, 6.

incentives, the challenge for policy makers will be to implement patent-

94. See LANJOUW & COCKBURN, supra note 42.

the original innovation. 100 Indeed, follow-on innovation by competing li-

Scherer's study focused on antitrust licensing decrees. These decrees

106. Id. at 63 (describing compulsory licensing in antitrust decrees generally).

resulting in an average of approximately twenty compulsory licenses per

114. id. at 82tbl.1.

which the government or its potential 8 licensees would have a complete,

118. Id. at 78-79.

124. Id. at 61.

bridge) or hypothetically (Taylor and Silberston) issue over important

V. CASE STUDIES OF INVESTMENT IN INNOVATION

Table 1: Drug License Orders under Antitrust Decrees

Baxter, Fibrin Merger S M Mature Patents +, 0

Marion Merger S L Mature Patents +, 0

Ciba- Merger S E Nascent Patents + "no mini-

Roche, CD4133 Merger S E Nascent Patents + 1-3% of net

Eli Lilly, Illegal G E Nascent Future Reasonable

Connaught, Merger G L Mature Entire Reason-

The stage of development of the affected technology varied among the

130. In re Baxter Int'l Inc., 123 F.T.C. 904 (1997).

broad basic technologies and therefore covered some early-stage, pre-

136. See infra Part VILE-F.

wards trade secrecy, thereby maintaining their pre-licensing level of R&D

143. See generally BioSpace: Competitive Clinical Intelligence System, (Biospace

stances of Baxter International ("Baxter") and Roche Holding Ltd.