<

k
Ivory Tower and Industrial Innovation
\
I N NOVATION and TECH NO LOGY in the WORLD ECONOMY

Editors
MARTIN KENNEY

University of California, Davis/Berkeley

Round Table on the International Economy

BRUCE KOGUT

Wharton School, University of Pennsylvania

Other titles in the series

Territories of Profit: Communications, Capitalist Development,

and Innovation at G.F. Swift and Dell Computer
GARY FIELDS

2003

The Triumph of Ethernet: Technological Communities

and the Battle for the LAN Standard
URS VON BURG

2001
,
Ivory Tower and Industrial Innovation
University-Industry Technology Transfer Before and
After the Bayh-Dole Act in the United States

DAVID C. MOWERY, RICHARD R. NELSON,

BHAVEN N. SAMPAT, and ARVIDS A. ZIEDONIS

Stanford Business Books

An Imprint of Stanford University Press
Stanford, California 2004
Stanford University Press
Stanford, California

© 2004 by the Board of Trustees of the Leland Stanford Junior

University. All rights reserved.

Printed in the United States of America

On acid-free, archival-quality paper

Library of Congress Cataloging-in-Publication Data

Ivory tower and industrial innovation : university-industry

technology transfer before and after the Bayh-Dole Act in
the United States / David C. Mowery .. . [et al.].
p. cm. — (Innovation and technology in the world economy)
ISBN 0-8047-4920-5 (cloth : alk. paper)
t. Technology transfer—United States. 2. Academic-industrial
collaboration—United States. I. Mowery, David C. II. Series.
T174.3 196 2004
338-973'°6—dc22
2003025632

Typeset by C&S Typesetters, Inc. in 10/12.5 Electra

Original Printing 2004

Last figure below indicates year of this printing:

13 12 n 10 09 08 07 06 05 04
 Contents

List of Tables and Figures ix

Acknowledgments xiii

1 Introduction: The Ivory Tower and Industrial Innovation 1

2 Historical Overview: American Universities •

and Technical Progress in Industry 9

3 University Patent Policies and University Patenting

Before the Bayh-Dole Act 35

4 The Research Corporation and University

Technology Licensing, 1912-80 5^

5 A Political History of the Bayh-Dole Act of 1980 85

6 The Bayh-Dole Act and Patenting and Licensing

at the University of California, Stanford University,
and Columbia University 99

7 The Effects of Entry and Experience on U.S.

University Patenting Before and After the Bayh-Dole Act 129

8 What Happens in University-Industry Technology

Transfer? Evidence from Five Case Studies
Robert A. Lowe, David C. Mowery, and Bhaven N. Sampat 152

9 Conclusion x79

Notes 19 3

References 215

Index 229
 Tables and Figures

TABLES

2.1 Federal Support for Academic R&D, 1935 and 1960-2000 24

2.2 Funding of U.S. Academic Research

by Federal Agencies, 1954-2001 25

2.3 Industries Rating University Research

as “Important” or “Very Important” 30

2.4 The Relevance of University Science

to Industrial Technology 31

2.5 Rankings of Importance of External Information

Sources in Suggesting New R&D Projects and
in Contributing to the Completion of R&D Projects 32

2.6 Importance to Industrial R&D of Sources

of Information on Public R&D 33

3.1 University Patent Policies and Procedures, 1940 41

3.2 Carnegie Research Universities with

Research Corporation LAAs 42

3.3 Concentration of Federal R&D Funds in Universities

Receiving the Largest Federal Obligations, 1963-75 44

3.4 Patenting by U.S. Universities, 1925-45 49

4.1 Research Corporation Gross Income from “Home Run”

Inventions, 1945-85 7°

6.1 Selected Data on University of California,

Stanford University, and Columbia University
Licensing Income, FY1970-95 102
Tables and Figures

X 6.2 Comparative Evidence on Invention Disclosures

and Licenses at Stanford University, Columbia
University, and the University of California, 1986-90 n3

6.3 University of California and Stanford

University Invention Disclosures, Patents,
and Licenses, 1975-79 and 1984-88 115

6.4 Number of Patents in the University of California,

Stanford, and Columbia Datasets 120

6.5 Mean Values for “Forward Citations”; University

of California, Stanford, and Columbia Patents
and Corresponding Control Group Samples 122

6.6 Mean Values for “Generality”; University of

California, Stanford, and Columbia Patents
and Corresponding Control Group Samples 124

7-1 Number of Patents by Year of Issue: “High-intensity

incumbents,” “Low-intensity incumbents,”
and “Entrants,” 1975-92 133

7-2 Regression Coefficients for High-Intensity

Incumbent X Application Year U4

7-3 Regression Coefficients for Low-Intensity

Incumbent X Application Year D5

7-4 Regression Coefficients for Entrant

University X Application Year 136

7-5 Regression Coefficients for All

Patents (1981-92 Sample) 140

7.6 Regression Coefficients for Biomedical

Patents (1981-92 Sample) Hi

7-7 Regression Coefficients for Nonbiomedical

Patents (1981-92 Sample) 142
7.8 (UNIVt + ENTt) Coefficients (1981—92 Sample) H3

FIGURES

3.1 Proportion of Carnegie RUis with > 0.5 FTE

technology transfer personnel, 1965-95 46

3.2 University patents, 1925-80 47

3.3 University patents per R&D dollar, 1963-93 48

3.4 University patents by university category, 1948-80

3.5 Patenting in 1970s by entrants and incumbents

(Carnegie RUis and RU2s) 52
 Tables and Figures

3.6 Patenting by Carnegie Research

xi
Universities by IPA status
53
3.7 Distribution among technology classes
of university patents, 1925-45
54
3.8 Distribution among technology classes
of university patents, 1948-80
55
4.1 Research Corporation Invention Administration
Agreements, 1946-81 66
4.2 Invention disclosures to the Research
Corporation, 1955—81 67
4.3 Patents issued to the Research Corporation, 1947-81 68
4.4 Research Corporation gross royalties, r95r-8r 68
4.5 Research Corporation expenditures
on consultants and travel, 1953—81 73
4.6 Research Corporation legal and personnel
expenditures, 1953—81 74
4.7 Research Corporation total operating costs, 1951— 81 78
4.8 Total costs per Invention Administration
Agreement and disclosure, 1951— 81 78
4.9 Research Corporation net royalty income, 1947-81 79
4.10 Research Corporation share of U.S.
university patents, 1969-87 80
6.r University of California invention disclosures, 1975-90 100

6.2 University of California patents by year of issue, 1975—95 101

6.3 University of California biomedical disclosures as a

percentage of total disclosures, 1975-90 103
6.4 Stanford University invention disclosures, 1975—90 106

6.5 Stanford University patents by year of issue, 1975-95 106

6.6 Stanford University biomedical disclosures as a

percentage of total disclosures, 1975-90 107

6.7 Biomedical technology share of Columbia, Stanford,

and University of California license agreements 107

6.8 Columbia University invention disclosures, 1981-95 110

6.9 Columbia University patents, 1981-95 111

8.1 Assignees for U.S. GaN-related patents, 1972-98 161

\
 Acknowledgments

We are indebted to the staffs of the technology licensing offices of Columbia

University, Stanford University, and the University of California for invaluable
assistance with the collection and analysis of these data. Our access to the data
at the University of California was made possible through the support of
Jay Stowsky and Suzanne Quick from the Office of the President of the Uni¬
versity of California. Kathryn Ku of Stanford University made the data from
Stanfords Office of Technology Licensing available to us for research use.
The research on the Columbia University data underpinning this paper
also benefited from the efforts of Michael Crow, Holly Raider, and Annetine
Gelijns of Columbia University. Michael Barnes and Lynn Fissell of the
University of California assisted in the collection and analysis of the Univer¬
sity of California data, and the research on the Stanford data benefited from
the assistance of Sandra Bradford. Special thanks to Michael Barnes for the
use of his university patenting data and to Adam Jaffe of Brandeis University,
Bronwyn Hall of UC Berkeley, and Manuel Trajtenberg of Tel Aviv University
and the NBER for making the NBER-Case Western Reserve patent data
available to us. We thank the staffs at the archives of Columbia University,
MIT, the National Research Council, and the Smithsonian Institution for ac¬
cess to their records. In addition, we are indebted to W. Stevenson Bacon for
access to the Research Corporations archives and discussions about its patent
management activities.
Bronwyn Hall, Scott Stern, Rosemarie Ziedonis, Josh Lerner, Stephen
Cameron, Rebecca Henderson, David Roessner, Scott Shane, Ken Leonard,
Frank Lichtenberg, Adam Jaffe, and seminar participants at UC Berkeley,
Columbia University, the University of Pennsylvania, the University of Mary¬
land, the University of Reading, and the Federal Reserve Bank of Philadelphia
Acknowledgments

xiv gave us many helpful comments on earlier versions of much of the research
contained in this volume. We are especially indebted to Nathan Rosenberg
of Stanford University, who contributed to portions of Chapter 2 and who
played a central role in the early stages of the research project that produced
this volume.
Support for this research was provided by the Andrew W. Mellon Founda¬
tion; the Alfred R Sloan Foundation; the California Policy Seminar; the UC
Presidents Industry-University Cooperative Research Program; the Office of
the Provost at Columbia University; the Division of Research at the Harvard
Business School; the Institute of Innovation, Organization, and Management
at the Haas School of Business, UC Berkeley; and the Mack Center of the
Wharton School at the University of Pennsylvania.
Ivory Tower and Industrial Innovation
I
 1

Introduction: The Ivory Tower

and Industrial Innovation

This monograph deals with the relationship between U.S. universities and in¬
dustrial innovation, focusing in particular on the role of patenting and licens¬
ing of academic inventions in supporting “technology transfer” between uni¬
versities and industry. The role of universities in industrial innovation and
economic growth has received considerable fanfare in recent years, as U.S.
universities have expanded their patenting and licensing activities since the
early 1980s. Many observers have attributed this expanded patenting and li¬
censing to the Bayh-Dole Act of 1980, although little hard evidence has been
provided in support of this conclusion. Nor has much evidence been pro¬
duced to support the argument that patenting and licensing of university in¬
ventions are necessary to support the transfer to industry and commercial de¬
velopment of these inventions. Other, more critical accounts of the Bayh-Dole
Act have suggested that the growth in academic patenting and licensing has
changed the “research culture” of U.S. universities, leading to increased se¬
crecy, less sharing of research results, and a shift in the focus of academic re¬
search away from fundamental to more applied topics.
The evidence presented in this volume suggests that the Bayh-Dole Act was
one of several factors that contributed to the growth of patenting and licensing
by U.S. universities during the 1980s and 1990s. The Act provided a strong con¬
gressional endorsement for academic institutions’ involvement in patenting
and licensing research discoveries and simplified the formerly complex ad¬
ministrative processes through which U.S. universities gained title to the in¬
tellectual property resulting from publicly funded research. Even without the
Bayh-Dole Act, however, we believe that university patenting would have
grown significantly during the 1980s and 1990s. As we point out below, many
INTRODUCTION

U.S. universities were active patenters and licensors long before the passage of
the Bayh-Dole Act, and they expanded these activities during the 1970s and
early 1980s in response to advances in biomedical research and changes in
the legal treatment of patents on life forms. It is likely that these and other
universities with less experience in patenting and licensing would have initi¬
ated or expanded these activities even in the absence of the Bayh-Dole Act. In¬
deed, U.S. research universities were an important source of support for the
passage of the Act in 1980, as we note in Chapter 5. Nonetheless, there is little
doubt that U.S. universities now are more heavily and directly involved in
patenting and licensing of research results than at any previous time in their
history.
The Bayh-Dole Act was motivated by the belief that university patenting
would spur and facilitate the transfer of university discoveries to industry for
commercial development. Although the evidence on this point presented in
this volume is at best suggestive rather than definitive, it indicates that research
results and knowledge flow between universities and industry (and we stress
the two-way nature of this flow) through publications by academic researchers,
conference presentations, faculty consulting, and the movement of personnel
between universities and industry, to name but a few channels. Before and
since the Bayh-Dole Act, much of the interaction between universities and in¬
dustry has not involved patenting and licensing.
In some cases, university patenting may indeed aid technology transfer, but
in many cases patenting of an invention by a university is not necessary to sup¬
port the transfer and commercialization of an invention. In these cases, pat¬
ents may produce income for the university (although most patents yield little
licensing income), while creating some risk that restrictive licensing policies
can limit the diffusion and use of an invention or related knowledge. Indeed,
in some fields of research, patenting and licensing could limit the operation
of other channels of mutual influence and interaction.
In short, the issues raised by university patenting and licensing are com¬
plex, and the course of action that is likely to yield the greatest public benefit
(the overriding goal of Bayh-Dole) varies among inventions and fields of tech¬
nology. Nevertheless, for the foreseeable future, U.S. universities (and, it
seems, more and more non-U.S. universities) will be patenting more inten¬
sively. A key challenge for policy concerns the appropriate design of licensing
policies for universities engaged in patenting of inventions financed with pub¬
lic funds to ensure that publicly funded research yields the greatest possible
societal benefit.
The growth in U.S. university patenting and licensing has been con¬
centrated in a relatively narrow set of research fields, notably the biomedical
 INTRODUCTION

sciences and some areas of engineering. It is unlikely, therefore, that increased

patenting and licensing have altered the “culture” of academic research
throughout U.S. universities, simply because these activities have focused on
relatively few academic disciplines. Even within the fields where licensing
and patenting are concentrated, however, there is little evidence of a signifi¬
cant shift in the focus or output of academic research, although the available
evidence on this point is limited and is likely to indicate any such shift only
well after the fact.
Another factor behind the growth of university patenting is the extension of
definitions of “patentable material” into the realm of science. Since scientific
research in many fields relies heavily on a relatively liberal exchange of mate¬
rials and other intellectual artifacts, greater assertion of private property rights
over these artifacts and materials potentially could increase the transactions
costs of scientific research and impede progress. But this extension of the
realm of the patentable reflects developments in a much broader set of poli¬
cies and institutions (notably, in the U.S. Patent and Trademark Office) than
university patent and licensing policies or the Bayh-Dole Act.

ORGANIZATION OF THE VOLUME

Although the Bayh-Dole Act and its effects are central issues in the chapters
that follow, our examination of the Acts origins and consequences requires a
broader consideration of the historical evolution of university-industry rela¬
tions and technology transfer in the United States, which we undertake in
Chapters 2-4. American universities have made significant contributions to
technological innovation in industry for well over a century. The Morrill Act
of 1862 was emblematic of an American commitment that universities should
serve the democracy and its citizens. Research at university agricultural ex¬
perimentation stations provided the scientific base for the hybrid corn revolu¬
tion of the 1930s and the postwar years and played a central role in developing
hybrid seed varieties tailored to the particular conditions of individual states.
But U.S. universities’ contributions to the development of the American
economy went well beyond agriculture. Edwin Armstrongs work at Columbia
University laid the basis for vacuum tube-based amplification of electronic
signals, a key contribution to the development of modern radio technology.
Much of the early work in developing computers and lasers, as well as the In¬
ternet, was performed in universities. And universities played leading roles in
the development of a number of important pharmaceuticals.
These historic contributions of U.S. universities to industrial innovation
were made through a diverse array of channels of interaction, all of which
INTRODUCTION

continue to figure prominently in university-industry knowledge exchange

and technology transfer. Although a number of U.S. universities did pursue
patents in the pre-1980 period (see Chapters 3 and 4), many university contri¬
butions to technical advance did not rely on patenting of the university dis¬
coveries. One particularly important factor supporting the development of re¬
search links and two-way flows of knowledge and technology between U.S.
universities and industry throughout the twentieth century is the structure of
the U.S. university system, which differs in important ways from those of other
industrial economies. We discuss this issue in more detail in Chapter 2.
Chapter 3 analyzes the debates of the 1920s and 1930s over the role of uni¬
versity patenting and licensing by way of highlighting the remarkable similar¬
ities, as well as some important contrasts, between these debates and those
over the Bayh-Dole Act in the late 1970s. Through much of the twentieth cen¬
tury U.S. universities were ambivalent about direct involvement in patenting
and licensing. In part, this ambivalence reflected concerns that any appear¬
ance of profiteering at public expense would be politically embarrassing. As a
result, a number of leading research universities limited faculty patenting, pri¬
marily in the biomedical arena, for much of the pre-1970 period. Even where
such leading research universities as MIT encouraged patenting by faculty,
they avoided direct involvement in the management of these patents and as¬
sociated licenses.
Indeed, concern over the effects of direct university involvement in patent
management was one motive for University of California professor Frederick
Cottrell’s founding of the Research Corporation in 1912 to manage patenting
and licensing for academic institutions. As we point out in Chapter 4, many
U.S. research universities “outsourced” these activities to the Research Cor¬
poration or other patent management organizations during the postwar
era. The corporations growth and eventual eclipse by universities’ indepen¬
dent management of patenting and licensing during the 1970s reflect many of
the same challenges that contemporary university technology licensing man¬
agers face.
Chapter 5 describes the political origins and passage of the Bayh-Dole Act,
and Chapters 6 and 7 analyze previously unexamined data dealing with the
Act’s effects on patenting and licensing by U.S. universities. Chapter 6 focuses
on three leading research universities, and Chapter 7 examines broader na¬
tional trends in university patenting and licensing before and after the Bayh-
Dole Act. Our quantitative analyses shed light on the effects of Bayh-Dole but
do not address the validity of a key premise of the Bayh-Dole Act, the impor¬
tance of patenting for the commercialization of individual university inven¬
tions. In order to examine this issue, Chapter 8 presents detailed case studies
of five university inventions that highlight the diversity of circumstances that
 INTRODUCTION

influence the process of technology transfer and commercialization. Like the

surveys of industrial managers that are discussed in Chapter 2, our case stud¬
ies reveal substantial contrasts among fields of research and innovation in the
role of patents and licenses. Moreover, these cases suggest considerable het¬
erogeneity among inventions within the biomedical field. The final chapter
summarizes our findings and discusses their implications for further research.

WHAT HAVE WE LEARNED?

Although a number of U.S. universities were patenting and licensing faculty

inventions during much of the twentieth century, academic patenting and li¬
censing are not the primary channels for technology transfer and knowledge
exchange with industry. Surveys of industry R&D managers during the 1980s
and 1990s consistently suggest that patents and licenses are less important than
other channels for knowledge flow and interaction with university researchers
(for example, faculty publications or conference presentations) in all fields,
including the biomedical sciences.
The passage of the Bayh-Dole Act in 1980 coincided with several other de¬
velopments of great importance for the growth in U.S. university patenting
and licensing. The U.S. Supreme Court affirmed the validity of patents on life
forms in the same year, and the 1970s and 1980s were a period of significant
scientific advances in biomedical research (much of which was supported by
the U.S. National Institutes of Health) that proved to have considerable po¬
tential for industrial applications. Indeed, two large research universities that
had long been active patenters and licensors (Stanford University and the Uni¬
versity of California system) shifted their patent and license portfolios in favor
of biomedical inventions before the passage of the Bayh-Dole Act, and a lead¬
ing post-1980 “entrant” into patenting, Columbia University, filed an applica¬
tion for its most lucrative single licensed invention, also in the biomedical
field, before the Acts passage.
Passage of the Bayh-Dole Act was followed by increased patenting at both
of the “experienced” institutional patenters in our dataset and a decline in the
share of these patents finding industrial licensees, consistent with a shift to¬
ward patenting less significant inventions in the immediate aftermath of the
Acts passage. But overall, there is no evidence at any of these three leading
research universities that the frequency of citations by other patents to their
post-1980 patents (widely interpreted as a measure of the economic or tech¬
nological “importance” of a. patent) declined significantly by comparison
with nonacademic patents. These universities’ patents were more heavily
cited on average than nonacademic patents before as well as after the Bayh-
Dole Act.
INTRODUCTION

Before and after the Bayh-Dole Act, the licensing revenues of the experi¬
enced academic patenters were dominated by a very small number of “home
run” inventions, most of which were biomedical inventions. This characteri¬
zation also accurately describes the 1985—95 licensing revenues of Columbia
University, which had much less historical experience in patenting and li¬
censing. The unpredictable and infrequent appearance of such “home runs”
means that many U.S. universities active in patenting and licensing of faculty
inventions have found these activities to be unprofitable. Although income is
only one of several motives for pursuing these technology transfer activities, it
figures prominently among the reasons for entry by many universities into
patenting and licensing after 1980. Moreover, greater institutional experience
in managing patenting and licensing may not increase the probability that
such a “home run” will be discovered by a university’s faculty or other re¬
searchers and be disclosed to its licensing office.
The effects of the Bayh-Dole Act on the patenting activities of experienced
academic patenters thus were more modest than many assessments suggest.
Nevertheless, the Bayh-Dole Act facilitated the entry into patenting of a num¬
ber of institutions with little experience in managing patenting and licensing
activities. These novice patenters did indeed receive less heavily cited patents
initially, but by the end of the 1980s their patents were cited no less intensively
than those assigned to “experienced” academic patenters and (like the patents
assigned to “experienced” universities) were more heavily cited than nonaca¬
demic patents in the same classes.
As we noted earlier, the case studies in Chapter 8 seek to examine an im¬
portant assumption animating the drafting and passage of the Bayh-Dole Act,
viz., the belief that patenting was both necessary and sufficient to facilitate the
transfer to industry and commercial development of university inventions. The
case studies indicate that the processes of knowledge exchange and technolog)'
transfer are complex, and the channels through which these processes operate
most effectively differ significantly among different fields of technology. More¬
over, these processes also require a strong “relational” component for their
smooth operation — as we note in Chapters 4 and 8, centralization in university
licensing operations historically has been difficult, because of the need for fre¬
quent interaction between licensing professionals and the faculty inventors.
Another set of assumptions underpinning the design of the Bayh-Dole Act
dealt with the characteristics of the intellectual property for which universities
would seek patent protection. The Act saw patented intellectual property as a
key enabling factor in the process of technology development—without clear
property rights on intellectual property, commercial developers would not
make the investments necessary to bring these inventions to the market. But a
 INTRODUCTION

portion of the growth in university patenting since 1980 appears to involve an

extension of patents to cover inputs to science, rather than technological arti¬
facts that are candidates for commercial development. The supporters of
Bayh-Dole devoted little consideration to any negative effects of university
patenting and licensing on “open science,” partly because the growth of
patenting in scientific, as opposed to technological, fields was not apparent in
the early 1980s.
These and other developments, however, underscore the need to place
the Bayh-Dole Act in context. The growth of university patenting is in part
a response to the extension during the 1980s and 1990s of intellectual prop¬
erty rights to cover new types of artifacts (for example, research tools) and
the willingness of the U.S. Patent and Trademark Office to grant broad claims
that may affect a much broader set of future “prospects” for development of a
given technology. The Bayh-Dole Act was not itself responsible for this exten¬
sion, but its effects cannot be understood without considering these broader
developments.
Another reason for keeping this broader context in the foreground of
any evaluation of Bayh-Dole s effects is the recent enthusiasm among other
industrial-economy governments for adopting policies for university technol¬
ogy transfer that closely resemble the Bayh-Dole Act. Other governments’ em¬
ulation of the Bayh-Dole Act is motivated by their desire to increase and en¬
hance research collaboration and technology transfer between their national
university systems and industry. But these “reforms” overlook the long history
of university-industry collaboration and technology transfer in the United
States that predates the Bayh-Dole Act. Especially in the absence of broader
structural reforms in their national university systems, emulation of Bayh-Dole
is likely to accomplish little and could well prove to be counterproductive.
Our overall verdict on the Bayh-Dole Act thus is a mixed one. Much of the
post-1980 upsurge in university patenting and licensing, we believe, would
have occurred without the Act and reflects broader developments in federal
policy and academic research. This increase in patenting and licensing, how¬
ever, is concentrated in a few fields of academic research, and any “cultural
impacts” associated with such expanded patenting and licensing therefore are
also relatively localized within U.S. research universities. At the same time,
any evidence of significantly detrimental effects on the norms or direction of
academic research, disclosure of results, conflicts of interest, and the like will
take time to emerge, and such evidence will be apparent only after any dam¬
age has occurred. The Acts emphasis on patenting and licensing as a critically
important vehicle for the transfer to industry of academic inventions lacked a
strong evidentiary foundation at the time of its passage, and evidence on the
INTRODUCTION

role of patenting and licensing as indispensable components of technology

transfer remains mixed. U.S. universities have long collaborated with industry
in applied and fundamental research, and channels other than patenting have
been critical to this collaboration. Nonetheless, patenting per se is less critical
to any assessment of the Acts effects on public welfare than are the types of li¬
censing policies adopted by universities. Finally, any encouragement from the
Bayh-Dole Act for greater patenting of scientific, as opposed to technological,
artifacts remains an issue for concern.
As we point out in our concluding chapter, U.S. universities have enjoyed
a charmed political life during most of the post-1940 era, benefiting from a
surge in federal financial support for academic research that was rooted in key
public missions, such as national defense and public health. Federal support
for U.S. academic research grew in parallel with the rise of U.S. research uni¬
versities to their current positions of international preeminence. The remark¬
able postwar advance of U.S. research universities to global research leader¬
ship from positions of parity in some fields of research and laggards in many
others reflects the reliance of federal support for university research on prin¬
ciples of interinstitutional competition and autonomy. As they face a new cen¬
tury of challenges and opportunities, it is essential for the preservation of their
preeminence that U.S. research universities maintain their historic commit¬
ment to the free flow of knowledge and to research in the service of the global
public interest. As denizens of universities, we are confident that robust debate
and discussion can only contribute to the pursuit of these goals.
 2

Historical Overview: American Universities

and Technical Progress in Industry

In the opening chapters of this monograph, we consider the contributions of

universities to technical progress in industry and the role of patenting in these
contributions during the 1890-1980 period.1 This chapter focuses on the
role of American universities in industrial innovation prior to and immedi¬
ately after World War II; Chapters 3 and 4 examine aspects of U.S. universities’
patenting activities during the pre-1980 period. We also seek in this chapter
to present a broader perspective on the nature of the contributions by U.S.
universities to industrial innovation, the channels through which these con¬
tributions have been realized, and the factors that have encouraged these
contributions.

DISTINCTIVE STRUCTURAL CHARACTERISTICS

OF U.S. HIGHER EDUCATION

Their focus on training for farmers and workers, along with research oriented
to regional economic development, differentiated American land-grant uni¬
versities from European universities of the late nineteenth and early twentieth
centuries. U.S. public universities, especially those established under the
terms of the Morrill Act, affected the direction of the academic research en¬
terprise during this period to a greater extent than the private Ivy League in¬
stitutions. These characteristics were anticipated in Alexis de Tocquevilles
discussion of attitudes toward science in the young republic:
In America the purely practical part of science is admirably understood and care¬
ful attention is paid to the theoretical portion, which is immediately requisite to
application. On this head, the Americans always display a clear, free, original, and
HISTORICAL OVERVIEW

inventive power of mind. But hardly any one in the United States devotes himself
to the essentially theoretical and abstract portion of human knowledge. (i99°>
vol. 2, p. 42)

Rather than focusing on the “essentially theoretical and abstract portion of

human knowledge,” de Tocqueville argued that “science” in the young re¬
public concerned itself with applications:
Every new method which leads by a shorter road to wealth, every machine which
spares labor, every instrument which diminishes the cost of production, every dis¬
covery which facilitates pleasure or augments them, seems [to such people] to be
the grandest effort of the human intellect. It is chiefly from these motives that a
democratic people addicts itself to scientific pursuits ... In a community thus or¬
ganized, it may easily be conceived that the human mind may be led insensibly to
the neglect of theory; and that it is urged, on the contrary, with unparalleled energy,
to the applications of science, or at least to that portion of theoretical science which
is necessary to those who make such applications. (1990, vol. 2, p. 45)

This utilitarian orientation to science influenced the research and other

activities of American universities. British visitors (and at least a few U.S. ob¬
servers, including Flexner [1930] and Veblen [1918]) disdained the “vocation-
alism” of the nineteenth- and early-twentieth-century American higher edu¬
cational system. Long before their British and other European counterparts,
U.S. universities assumed responsibility for teaching and research in fields
such as agriculture and mining; commercial subjects such as accounting, fi¬
nance, marketing, and management; and an ever-widening swath of subjects
in civil, mechanical, electrical, chemical, aeronautical, and other engineering
disciplines.
There were a number of reasons for the “practical” orientation of U.S. uni¬
versities. In contrast to most European systems of higher education before
and after 1945, the American university system has always lacked any central¬
ized control. Although the Morrill Acts of 1862 and 1890 provided federal sup¬
port for the establishment of the land-grant universities, administrative over¬
sight of these institutions, as well as primary responsibility for their operating
budgets, remained under the control of state governments in the United
States.2 Throughout the twentieth century, U.S. universities retained great au¬
tonomy in their administrative policies.
The U.S. university system also was much less oriented to training gradu¬
ates for governmental service than were the university systems of Germany,
France, or Japan. Although some “finishing” and religious preparatory schools
such as Harvard and Yale originally were modeled on European institutions,
many U.S. universities and colleges chose their missions and research agenda
 HISTORICAL OVERVIEW

on the basis of the special needs of their local environment. One consequence
of this approach was that the funding and enrollment of these schools became
heavily dependent on the mores and needs of the local community.3 And as
de Tocqueville indicated, these mores tended strongly to the practical. Partly
because of these utilitarian motives, as well as the demand by many commu¬
nities for colleges or universities, the United States developed a higher educa¬
tion system” that was larger than those of contemporary European nations.
As Trow (1979) points out,

America had established 9 colleges by the time of the Revolution, when 2— Oxford
and Cambridge — were enough for the much larger and wealthier mother country.
The United States entered the Civil War with about 250 colleges, of which over 180
still survive. Even more striking is the record of failure: between the American Rev¬
olution and the Civil War perhaps as many as 700 colleges were started and failed.
By 1880 England was doing very well with 4 universities for a population of 23 mil¬
lion, while the single state of Ohio, with a population of 3 million, already boasted
37 institutions of higher learning ... By 1910 we had nearly a thousand colleges and
universities with a third of a million students —at a time when the 16 universities of
France enrolled altogether about forty thousand students, a number nearly equaled
by the faculty members of the American institutions, (pp. 271-72)

Not surprisingly, the U.S. higher education system enrolled a larger frac¬
tion of the eighteen-to-twenty-two-year-old population than those of any Eu¬
ropean nations throughout the 1900-1945 period. According to Geiger (1986),
roughly 12 percent of this age group was enrolled in U.S. universities and col¬
leges as early as 1928, a sharp increase from 8 percent in 1920. Although the
fraction of this age group enrolled in colleges and universities probably de¬
clined during the Great Depression, Graham and Diamond (1997, p. 24) esti¬
mate that the share reached 12 percent once again by 1940, three times the
4 percent of Europeans in this age cohort enrolled in colleges and universities
in the same year. Not until the 1960s did European enrollment rates exceed
10 percent of the relevant age cohorts, by which time U.S. enrollment rates
within this group were reaching 50 percent (Burn et ah, 1971).
The Morrill Act of 1862 was rooted in these American views about the role
of university research and teaching. The Acts intent was eminently practical,
inasmuch as it was dedicated to the support of agriculture and the mechani¬
cal arts by encouraging the establishment of state-controlled universities
whose long-term prosperity and success depended on their responsiveness to
the demands of the local community. But private universities also depended
on local support throughout much of the nineteenth and twentieth centuries
for their sustenance and growth. Their dependence on local sources of finan¬
cial and political support, combined with the lack of centralized control or
HISTORICAL OVERVIEW

stipulations concerning curriculum, meant that U.S. universities generally

were quick to introduce new courses or disciplines into their curricula in re¬
sponse to community demands. Ben-David (1968) contrasted this dependence
on local sources of support with the situation in most European university sys¬
tems through much of the twentieth century, arguing that
For the European civil servant in charge of the matter, the problem of financing
universities posed itself in the form of how to spend as little as possible given an es¬
tablished set of objectives and a scale of priorities. The American university presi¬
dent, however, had to ask himself how to increase income by convincing a variety
of donors about the importance of the university and by finding new markets and
extending existing ones for its services. For the university president the range of ob¬
jectives and the scale of priorities were not regarded as given. It was one of his tasks
to watch for circumstances justifying their change, and to push for changes if rec¬
ognized as necessary or useful. This entrepreneurial leadership, the only one suited
to the age of organized and rapidly changing research and training, was also adopted
by the state universities, (p. 36)

A primary activity of early American universities was the provision of voca¬

tional skills for a wide range of professions of concern to local economies. In
many cases, training was combined with research concerned with the prob¬
lems of local industry. For example, the University of Akron supplied skilled
personnel for the local rubber industry, became well know n for research in the
processing of rubber, and subsequently achieved distinction in the field of
polymer chemistry. Land-grant colleges combined training with research on
the needs of the local agricultural community —the Babcock test for measur¬
ing the butterfat content of milk was developed by an agricultural research
chemist at the University of Wisconsin. Introduced in 1890, the test provided
a new and reliable method to determine the adulteration of milk, a matter of
no small consequence in a state of dairy farms.
Public universities also developed new degree programs to address local
vocational and economic interests. After World War I, U.S. colleges of en¬
gineering at public universities offered undergraduate degrees in a bewilder¬
ing array of specialized engineering subjects; at the University of Illinois, this
included architectural engineering, ceramic engineering, mining engi¬
neering, municipal and sanitary engineering, railway civil engineering, rail¬
way electrical engineering, and railway mechanical engineering. An ob¬
server noted that during this period, “Nearly every industry and government
agency in Illinois had its own department at the state university in Urbana-
Champaign” (Levine, 1986).
On occasion, university research on problems of industry involved large-
scale, long-run commitments to the solution of a particular problem. One of
 HISTORICAL OVERVIEW

the most important such projects was conducted at the Mines Experiment
Station of the University of Minnesota, beginning before World War I and
continuing through the early 1960s. This long-term applied research project
addressed the consequences of the gradual exhaustion of the high-yielding
iron ores in the Mesabi Range. As the supply of these ores declined, re¬
searchers focused on ores of lower iron content, specifically the abundant
deposits of low-quality taconite ore. Although it did not rely on new scientific
knowledge, the solution to innumerable engineering and processing problems
associated with taconite extraction and refinement required decades of exper¬
imentation at the Mines Experiment Station (Davis, 1964).
A final characteristic of the U.S. higher education system that distinguished
it from those of other industrial economies throughout the twentieth century
was the emergence of a unified national market for faculty at U.S. research
universities. The departmental structure of most U.S. universities and colleges
and the emergence of strong disciplinary degree programs and societies by the
late nineteenth century meant that faculty qualifications were established on
the basis of their contributions to disciplinary research rather than their con¬
tributions to a specific institution.4 This interinstitutional mobility, combined
with the strong competition among U.S. universities for prestige, resources,
and students, meant that faculty moved among universities, especially the
most prestigious public and private research universities, more frequently
than was true of faculty in other national systems of higher education. Such
mobility provided a powerful mechanism for the diffusion of new ideas, cur¬
ricula, and research approaches among U.S. universities that was less highly
developed in other national systems of higher education.
Throughout the twentieth century, the U.S. system of higher education has
been distinguished from those of other industrial economies by its large scale,
the high level of autonomy enjoyed by individual universities and colleges, the
dependence by these institutions on local sources of financial and political
support, and the strong competition among universities and colleges for
funds, prestige, faculty, and students. These structural characteristics of U.S.
higher education created powerful incentives for university researchers and
administrators to establish close relationships with industry. They also moti¬
vated university researchers to seek commercial applications for university-
developed inventions, regardless of the presence or absence of formal patent
protection. Finally, the large scale and vocational orientation of many U.S.
universities, combined with the conduct of research within these universi¬
ties, created an effective channel for the rapid dissemination of new research
findings into industrial practice —the movement of graduates into industrial
employment.
HISTORICAL OVERVIEW

THE INSTITUTIONALIZATION OF ENGINEERING

14
AND APPLIED SCIENCES

The orientation of U.S. university research toward the resolution of “practi¬

cal” problems of concern to local industry can be illustrated by a selective dis¬
cussion of the contributions of U.S. university research to the institutionaliza¬
tion of new disciplines in engineering and the applied sciences. Among other
things, these disciplines provided a systematic basis for research and training
that provided intellectual links among the individuals and universities en¬
gaged in such activities.
During the early twentieth century, fields such as chemical engineering,
electrical engineering, and aeronautical engineering became established in
American universities. Each of these fields developed programs of graduate
studies with certified professional credentials, professional organizations, and
associated journals. These new disciplines and professions both reflected and
solidified new connections between American universities and a variety of
American industries. Growth in these new disciplines and training programs
in American universities responded to increased use of university-trained en¬
gineers and scientists in industry, especially the rise of industrial research in
the chemical and electrical equipment industries in the early years of the
twentieth century (see Hounshell and Smith, 1988; Mowery, 1981; Noble,
1977; Reich 1985).
Engineering education hardly existed in the United States before the Civil
War. Many schools offered vocational engineering education, but systemic
training of professional engineers was nearly unknown until the latter part of
the century. Although the first U.S. engineering college, Rensselaer Polytech¬
nic Institute (RPI), was founded in 1824, the U.S. Military Academy at West
Point, founded in 1802, trained many of the first professional engineers in the
United States. Graduates of West Point made major contributions to the vast
construction enterprises associated with the building of an extensive, ulti¬
mately transcontinental, railroad system beginning in the 1830s. The needs of
the railroad, telegraph, and, later, an expanding succession of new products
and industries increased the demand for engineers. In response, new schools
such as MIT (1865) and Stevens Institute of Technology (1871) were estab¬
lished, and engineering courses were introduced into the curricula of older
universities. Here again the American experience in higher education differed
from that of Europe. In Great Britain, France, and Germany, engineering sub¬
jects often were taught at separate institutions; but these subjects were intro¬
duced at an early date into elite U.S. universities. Yale introduced courses in
mechanical engineering in 1863, and Columbia University opened its School
of Mines in 1864 (Grayson, 1977).
 HISTORICAL OVERVIEW

Electrical Engineering

The response of the American higher education system to the emergence of

electricity-based industries was swift. Although many historians cite 1882,
when Edison opened the New York City Pearl Street generation station, as the
year the U.S. electric equipment and generation industries were founded, in
fact crude versions of the telephone and electric light had already been intro¬
duced by 1882, and the demand for well-trained electrical engineers was grow-
ing rapidly. Firms such as General Electric and Westinghouse had limited
success in training their own employees in this new field.
U.S. universities responded rapidly to this new demand for engineering
training. MIT introduced its first course in electrical engineering in 1882. Cor¬
nell introduced a course in electrical engineering in 1883 and awarded the first
doctorate in the subject as early as 1885. By the 1890s, “schools like MIT had
become the chief suppliers of electrical engineers” in the United States
(Wildes and Lindgren, 1985). In contrast to the research of such industrial pi¬
oneers as Edison, Westinghouse, and Bell, university-based research and edu¬
cation in the emergent discipline of electrical engineering defined a commu¬
nity of technically trained professionals with well-developed links among
universities and between universities and industry. Throughout the twentieth
century, U.S. schools of engineering provided the research in engineering and
applied science on which the electrical industries were based.
University research in electrical engineering and physics generated more
than research advances and trained graduates. The establishment of new com¬
panies by university professors intent upon commercializing their research
findings has occasionally been portrayed as a uniquely post-World War II
development, but this practice has ample earlier precedents. The Federal
Company of Palo Alto, California, was founded by Stanford University faculty
and became an important supplier of radio equipment during World War I
(Bryson, 1984). The klystron, a thermionic tube for generating and amplifying
microwave signals for high-frequency communication systems, was the sub¬
ject of a 1937 agreement between Hal and Sigurd Varian, inventors of the
klystron, and the Stanford Physics Department. Stanford University provided
the Varians with access to laboratory space, faculty, and a $100 annual al¬
lowance for materials. In exchange, Stanford received a one-half interest in
any patents, an arrangement that yielded handsome returns for the university.
The development of electrical engineering within American higher edu¬
cation responded to a national need, the emerging electricity-based industries,
in contrast to the more provincial needs that had motivated many previous
university researchers. Training electrical engineers became the responsibility
of public and private U.S. universities, whose training activities strengthened
HISTORICAL OVERVIEW

i6 and supported a fertile interface between university research and technical

advance in industry. Throughout the twentieth century, university research
also contributed to industrial innovation through the establishment of con¬
sulting relationships with industry and the occasional foundation of firms by
faculty.

Chemical Engineering

Another illustration of the critical role of U.S. university research in engi¬

neering is the emergence of the discipline of chemical engineering in the
United States in the early twentieth century. The development of this disci¬
pline was associated to a striking degree with a single institution: MIT (see Ser¬
vos, 1980). The discipline of chemical engineering was created to address the
challenges resulting from the fact that the knowledge generated by major sci¬
entific breakthroughs in chemistry provided little or no guidance for the man¬
ufacture of new products on a commercial scale. Chemical engineering is not
applied chemistry and cannot be adequately characterized as the industrial
application of scientific knowledge generated in the chemical laboratory.
Rather, it involves a merger of chemistry and mechanical engineering, that is,
the application of mechanical engineering to the large-scale manufacture of
chemical products (see Furter, 1980).
The complexities of this transition from laboratory to commercial produc¬
tion help explain the gap of several or many years that separates the discovery
under laboratory conditions of many important new chemical entities from
their commercial production. An entirely new methodology, one distinct from
chemistry, was developed to manage the transition from test tubes to a manu¬
facturing operation for which output was measured in tons rather than
ounces. This new methodology exploited the concept of “unit operations,” a
term coined by Arthur D. Little at MIT in 1915. Unit operations provided the
basis for a rigorous approach to large-scale chemical manufacturing and thus
may be taken to mark the origins of chemical engineering as a unique disci¬
pline.5 Unit operations also provided the basis for the systematic, quantitative
instruction of future practitioners-in other words, a form of generic knowl¬
edge that could be taught at universities.
Arthur D. Little, Warren Lewis, and other faculty encouraged the develop¬
ment of collaborative relationships with U.S. industry that involved research
and teaching, the exchange of students in cooperative education, and the
foundation at MI T of the school of chemical engineering practice. Organized
research in chemical engineering developed in parallel at MIT and in indus¬
try and was especially influenced by the symbiotic relationship between engi-
 HISTORICAL OVERVIEW

neers at Standard Oil of New Jersey and MIT faculty who worked to codify,
advance, and disseminate the key tenets of the emergent discipline.6
Much of the collaboration during this period combined joint development
of these new practices in academic and industrial laboratories with relatively
widespread dissemination, particularly through teaching and textbooks. The
Standard Oil refinery in Baton Rouge, Louisiana, also played a key role as an
unofficial external laboratory and employer of a great many of the MIT grad¬
uates and (as consultants) a number of the faculty at MIT in the school of
chemical engineering. In many respects, this collaboration culminated in the
development of fluidized bed catalysis in 1941. Research conducted at MIT
complemented research done in the Baton Rouge refinery; although patents
were an important output of this research activity, the university had no direct
role in managing or licensing this intellectual property.
The key to this style of collaboration was personnel exchange between MIT
and industry through faculty consulting, faculty rotations to and from indus¬
try, and placement of graduates. Personnel exchange brought expertise from
MIT to industry and transferred practical knowledge from industry to acade¬
mia, where it was refined and codified, supporting the development of a
broader engineering discipline. As in many other areas of engineering or sci¬
entific research, access by faculty to industrial facilities was important, as the
scale and type of equipment in industry often were unavailable within the uni¬
versity. The industrial collaborators obtained the ownership of or were as¬
signed the intellectual property resulting from collaboration, and a great deal,
although not all, of the results of the research by academics in the industrial
context was published.

Aeronautical Engineering

The contribution of American higher educational institutions to the progress

of aircraft design before World War II is another example of the production by
universities of information of great economic value to the development of a
new industry. An excellent illustration of university engineering research that
yielded valuable design data and techniques for the acquisition of new knowl¬
edge was the propeller tests conducted at Stanford University by W. F. Durand
and E. P. Lesley from 1916 to 1926 (Vincenti, 1990, chap. 1 and p. 137). Ex¬
tensive experimental testing was necessary because of the absence of a body of
scientific knowledge that would permit a more direct determination of the op¬
timal design of a propeller, given the fact that “the propeller operates in com¬
bination with both engine and airframe . . . and it must be compatible with
the power-output characteristics of the former and the flight requirements of
HISTORICAL OVERVIEW

18 the latter” (Vincenti, 1990, p. 141). The method of experimental parameter

variation was necessary because a useful quantitative theory did not exist. Vin¬
centi points out that the Stanford experiments accomplished more than just
data collection but something other than science. Instead, they contributed
to the development of a specialized methodology that could not be directly
deduced from scientific principles, albeit one that was consistent with those
principles.7
The Stanford experiments led to a better understanding of aircraft design
by producing a form of generic knowledge that lies at the heart of the modern
discipline of aeronautical engineering.8 The Stanford experiments made im¬
portant contributions to American aircraft design in the 1930s, most notably in
the emergence of the DC-3 the second half of the decade. But the great suc¬
cess of the DC-3 also owed a large debt to another educational institution, the
California Institute of Technology. Cal Techs Guggenheim Aeronautical
Laboratory, funded by the Guggenheim Foundation, performed research that
was decisive to the commercial success of Douglas Aircraft, located in nearby
Santa Monica. The technical breakthroughs associated with the DC-3 anc^ *ts
predecessors, such as durable and reliable components, and the aircrafts in¬
creased passenger capacity, which lowered its seat-mile operating costs, were
largely the product of the Cal Tech research program, highlighted by their use
of multicellular construction and the exhaustive wind tunnel testing of the
DC-i and DC-2.

Computer Science and Engineering

Computers are among the most remarkable technical contributions of Amer¬

ican universities in the last half of the twentieth century. Important develop¬
ment work on computers had been performed in Europe (Alan Turing in
Great Britain and Konrad Zuse in Germany were among the pioneers), but
the emergence of electronic, digital computers was largely the product of
wartime research and development activities conducted at American univer¬
sities. This research was concentrated in schools of engineering and trans¬
formed a logical possibility into a technical reality. University-based research
created a new discipline, computer science, which was influenced by the his¬
torical development of disciplines such as electrical engineering and physics
but nurtured its own research methodology.
The first fully operational electronic digital computer, the Electronic Nu¬
merical Integrator and Computer (ENIAC), was built at the Moore School of
Electrical Engineering at the University of Pennsylvania during 1943-46 un¬
der the direction of Presper Eckert and John Mauchly. Eckert and Mauchly’s
work drew on research at other American universities, particularly work by
 HISTORICAL OVERVIEW

John Atanasoff, a mathematician and physicist at Iowa State University, and 19

Vannevar Bush, an electrical engineer at MIT.
Atanasoff’s device was designed for the solution of systems of linear equa¬
tions, although he appears to have given a good deal of thought to the possi¬
bility of a general-purpose electronic digital computer. But Atanasoff s ma¬
chine never became operational and existed only in crude prototype form (see
Stern, 1981). Another important predecessor of the ENIAC was the differential
analyzer developed at MIT by Vannevar Bush and his associates. Consistent
with our discussion above of the practical motivations for much of the engi¬
neering research at MIT and other U.S. universities during this period, Bush’s
work grew out of problems associated with the transient stability of electric
power transmission in interconnected, large-scale electric power networks.
The Moore School’s construction of a differential analyzer in 1939 led to a
close relationship between the school and the U.S. Army's Ballistics Research
Laboratory at the Aberdeen Proving Ground. The army financed the ENIAC
project in order to develop equipment to calculate solutions to ballistics prob¬
lems more rapidly. By the time the ENIAC was ready for testing in 1945, the
war had ended, but the intercession of John Von Neumann preserved the
ENIAC for eventual use in the extensive calculations on the design of a hy¬
drogen bomb (Stern, 1981, p. 62).
How should the university research that led to the postwar emergence of
the digital electronic computer be categorized? The early participants were
trained in engineering, mathematics, and physics. Mauchly and Bush taught
and performed their research in schools of engineering. Atanasoff taught
physics and mathematics at Iowa State. Howard Aiken, who conducted re¬
search on computers at Harvard during World War II, was a mathematician
who had previously worked in engineering.9 But their research is difficult to
place in the conventional R&D boxes of “basic research,” “applied research,”
or “development.” Although the term “computer science” is common enough
in university curricula today, the discipline, if it is indeed a science rather than
engineering, is a distinctly different kind of science. It is certainly not a natu¬
ral science but may be appropriately regarded, in Herbert Simon’s apt phrase,
as a “science of the artificial” (Simon, 1969, p. xi). Much of the research in
computer science, after all, resembles engineering in that it deals with the de¬
sign and construction of artifacts, or machines.

The Applied and Engineering Sciences in U.S. Universities

Simon’s appellation also applies to other engineering disciplines developed

within U.S. universities. These “sciences of the artificial” consist of purposive,
goal-directed activities. Their explicit design orientation excludes them from
HISTORICAL OVERVIEW

20 the usual definition of basic research, which involves a quest for fundamental
understanding. In the traditional natural sciences, such a quest has often been
identified with research with no immediate concern with practical appli¬
cations. But much research in the applied sciences and engineering is quite
basic, since it involves a search for fundamental understanding. Most of the
research in the medical sciences is undertaken with specific practical appli¬
cations in view. Medical studies of carcinogenic processes necessarily involve
research into fundamental aspects of cell biology. All of these lines of inquiry
were classed by the late Donald Stokes (1997) as research in “Pasteurs Quad¬
rant," defined as research that seeks to understand the fundamental physical,
biological, or chemical processes that underpin specific problems, solutions,
or applications.
This selective review of the development of a number of important engi¬
neering disciplines suggests that engineering education in the United States
has consistently attempted to provide reference points for inquiry into the de¬
tails of practical problems. At the same time, however, university research has
provided an intellectual framework for training professional decision-makers,
as Herbert Simon (1969) reminds us:

The intellectual activity that produces material artifacts is no different fundamen¬

tally from the one that prescribes remedies for a sick patient or the one that devises
a new sales plan for a company or a social welfare policy for a state. Design, so con¬
strued, is the core of all professional training; it is the principal mark that distin¬
guishes the professions from the sciences. Schools of engineering, as well as schools
of architecture, business, education, law, and medicine, are all centrally concerned
with the process of design, (pp. 55-56)

A number of other academic disciplines resemble engineering in their ori¬

entation to specific useful goals, such as the improvements in agricultural pro¬
ductivity that are the focus of much research in the life sciences in schools of
agriculture. Statistics, surely one of the most useful of disciplines, is another
example of such a discipline that achieved curricular and department status
in the United States long before it did so in Europe. And consistent with our
earlier discussion of their role in promoting research of interest to their re¬
gions, among the most important institutions in the early development of sta¬
tistics were Iowa State University and North Carolina State University, both of
which applied statistics to the analysis of agricultural yields and prices.
By the start of World War II, the applied sciences and engineering disci¬
plines were well established within U.S. higher education, especially at the
land-grant universities, which accounted for a large share of American uni¬
versity research. The presence of the engineering disciplines and the applied
 HISTORICAL OVERVIEW

sciences extended but did not replace the longer-standing tradition in Ameri- 21
can universities of research and education in the service of local industry and
agriculture.

The Medical Sciences

The medical sciences exhibit a similar pattern of basic research oriented to

practical ends. Although U.S. universities’ basic research capabilities in the
medical sciences emerged only after World War II, the rpio report by Abra¬
ham Flexner, Medical Education in the United States and Canada, began the
processes that greatly expanded the role of biomedical research within Amer¬
ican medical schools. Flexner argued that medical students needed better
training in the natural sciences, that medical schools should undertake basic
biomedical research, and that these schools should strengthen their affiliations
with teaching hospitals. This combination of teaching, research, and medical
practice is the hallmark of the American academic medical centers that have
made significant contributions to technical progress in medicine since 1945.
Research on the causes of human illness has much in common with re¬
search in the applied physical sciences and engineering disciplines. Much
of it aims for deep understanding of fundamental, scientific principles, but
it is motivated by the very practical objective of relieving humans of the
scourge of disease. This characterization of motives and results accurately
characterizes the work of Koch and Pasteur in the nineteenth century that
identified bacteria as the cause of many human illnesses, as well as fields such
as endocrinology, neurology, and biochemistry that emerged in the late nine¬
teenth and early twentieth centuries.10 Although most of the fundamental re¬
search in these fields was dominated by European research centers until
World War II, American medical schools enhanced their research capabilities
in these fields during the rp20s and 1930s. The great expansion in federal fund¬
ing of research at U.S. academic medical centers that began after 1945 built on
the institutional research capability that had been created in the preceding
four decades.

THE IMPACT OF WORLD WAR II ON FEDERAL

FUNDING OF ACADEMIC RESEARCH

Our discussion thus far has emphasized the distinctive structure of the U.S.
higher education system throughout the late nineteenth and twentieth centu¬
ries and the contrasts between this structure and those of other industrial na¬
tions. In this section, we discuss the dramatic changes in the financing of U.S.
HISTORICAL OVERVIEW

22 academic research that occurred in the space of less than a decade in the mid¬
twentieth century. These shifts in the sources of funding for U.S. academic re¬
search created a research infrastructure in U.S. higher education whose scale
and other features also contrasted with those of other industrial economies.
War preparations and the U.S. entry into World War II in December 1941
transformed federal R&D programs and priorities. Overall federal R&D ex¬
penditures (in 1996 dollars) soared from $784.9 million in 1940 to a peak of
$12.4 billion in 1945, including an increase from $279.2 million to $4 billion
(1996 dollars) in Defense Department R&D spending. The success and the or¬
ganizational structure of the massive federal wartime R&D program yielded
several important legacies. The successful completion of the Manhattan Proj¬
ect, whose research budget in the peak years 1944 and 1945 substantially ex¬
ceeded that of the Department of Defense, created a research and weapons
production complex that ushered in the age of truly “big science.” Ironically,
the Manhattan Projects success in creating weapons of unprecedented de¬
structive power contributed to rosy postwar perceptions of the constructive
possibilities of large-scale science for the advance of societal welfare.11
Far smaller in financial terms, but highly significant as an institutional in¬
novation, was the Office of Scientific Research and Development (OSRD), a
civilian agency directed by Vannevar Bush that relied on research contracts
with private firms and universities. The largest single recipient of OSRD
grants and contracts during wartime (and the inventor of institutional over¬
head) was MIT, with seventy-five contracts for a total of more than $886 mil¬
lion (1996 dollars). The largest corporate recipient of OSRD funds. Western
Electric, accounted for only $130 million (1996 dollars) (Pursell, 1979, p. 364).
The contractual arrangements developed by OSRD during the Second World
War allowed the OSRD to tap the broad array of academic and industrial
R&D capabilities that had developed during the interwar period. Members of
the scientific community were called upon to recommend and to guide, as
well as to participate in, scientific research with military payoffs.
The OSRD and other wartime programs transformed the scale and sources
of funding for academic research in the United States. Indeed, the postwar
federal presence within academic research funding assumed a shape that dif¬
fered dramatically from that envisioned by one of the most famous and
influential figures in U.S. science policy during this century, Vannevar Bush.
In response to a request that he had solicited from President Franklin D. Roo¬
sevelt, Bush drafted the famous 1945 report on postwar federal science policy,
Science: The Endless Frontier. Anticipating the analysis of later economists,
Bush argued that basic research was the ultimate source of economic growth
and advocated the creation of a single federal agency charged with responsi-
 HISTORICAL OVERVIEW

bility for funding basic research in all defense and nondefense areas, includ- 23
ing health. The complexities of postwar domestic politics, as well as Bushs re¬
sistance to congressional oversight of his proposed agency, ultimately doomed
his proposal. Rather than a civilian agency overseeing all of federal science
policy and funding, various mission agencies, including the military and the
National Institutes of Health (NIH), assumed major roles in supporting basic
and applied research. By 1953, more than 86 percent of federal R&D spend¬
ing (5 percent of which supported academic research) was controlled by the
Defense Department and the Atomic Energy Commission.

THE POSTWAR STRUCTURE OF FEDERAL

SUPPORT FOR ACADEMIC R&D

The transformation of the postwar U.S. research system expanded and trans¬
formed the role of publicly funded research in U.S. universities. Although
Bush s recommendations of a single federal funding agency for basic research
were not implemented and his advocacy of institutional, rather than project,
funding also was ignored, U.S. universities enjoyed significant increases in
federal R&D support during this period. From an estimated level of less
than $150 million in 1935-36, federal support for university research (exclud¬
ing Federally Funded Research and Development Centers [FFRDCs] at uni¬
versities and colleges) grew to more than $2.1 billion in i960 and nearly $14 bil¬
lion in 1995 (Table 2.1; all amounts in 1996 dollars). Federal funding of
academic research, which amounted to no more than 25 percent of total aca¬
demic research support in the mid-i930S, by i960 accounted for more than
60 percent of the total.
The overall academic research enterprise increased almost six-fold in con¬
stant dollars between 1935 and i960 and more than doubled again by 1965 (see
Table 2.1). In 1953, less than a third of all U.S. basic research was performed in
universities and FFRDCs at universities and colleges. By 1995’ however, these
institutions performed 60 percent of U.S. basic research (National Science
Foundation, 1996).12 Increased federal support for university research trans¬
formed major U.S. universities into worldwide centers for the performance of
scientific research, a characterization that applied to only a few U.S. universi¬
ties in a limited number of fields during the prewar years.
In addition to financing an expanded academic research enterprise, federal
support for graduate education and university facilities, especially after the
1958 Sputnik “crisis,” enlarged the pool of scientific personnel and supported
the acquisition of the physical equipment and facilities essential to the per¬
formance of high-quality research. In the case of computer science, federal
HISTORICAL OVERVIEW

TABLE 2.1
24
Federal Support for Academic R&D, 1935 and i960 -2000
(millions of 1996 dollars)

Federally Federal share

Total academic supported of total academic
Year R&D ($) R&D ($) research funding

1935 575 138 24%

1960 3,418 2,143 63
1965 7,333 5,338 73
1970 9,453 6,668 71
1975 9,939 6,671 67
1980 11,575 7,817 68
1985 14,120 8,828 63
1990 19,551 11,570 59
1995 22,827 13,726 60
2000 27,379 15,932 58

SOURCES: Data for 1935, National Resources Committee (1938); data for
i960 and after. National Science Foundation (2001).

support for university purchases of large mainframe computers was in¬

dispensable to the institutionalization of a new academic discipline in U.S.
universities. Federal programs also increased financial aid for students in
higher education during the late 1950s.13 By funding both university educa¬
tion and research, the federal government strengthened the university com¬
mitment to research and reinforced the link between research and teaching.
The combination of research and teaching in higher education has been car¬
ried much further in the United States than elsewhere. In much of Europe
and Japan, for example, a larger fraction of research is carried out in special¬
ized research institutes not connected directly with higher education and in
government-operated laboratories.14
Although the historically unprecedented federal investment in academic
research helped propel American universities to international preeminence
in basic research and graduate education, the bulk of federal funding of aca¬
demic research was predicated on the expectation that the research would
yield practical benefits for federal agency missions. The National Science
Foundation (NSF) was established in 1950 to fulfill the Bush vision of federal
support for basic research that sooner or later would yield social benefits. But
the NSF has never accounted for even a fifth of federal support for university
research during the postwar period. Instead, agencies concerned with two key
postwar federal missions, defense and public health, have dominated federal
support for academic research.
I he Department of Defense and two other agencies with significant
defense-related responsibilities, the National Aeronautics and Space Admin-
 HISTORICAL OVERVIEW

istration (NASA) and the Atomic Energy Commission (AEC; later the De¬
partment of Energy), accounted for more than 80 percent of federal support
for academic research in 1954, a share that dropped below 30 percent of the to¬
tal after 1970 (see Table 2.2). During the 1953-60 period, the NIH provided
roughly a third of total federal academic research funding, and since i960
NIEI funding of university research has substantially increased. By the early
twenty-first century, the NIH accounted for more than 60 percent of federally
funded university research.
This enormous postwar federal investment in academic biomedical re¬
search forged stronger links between basic science and clinical applications in
U.S. biomedical research. By combining scientific research with clinical prac¬
tice, the U.S. academic medical center has been able to link science and in¬
novation to a remarkable degree, enabling the rapid collection by scientists of
feedback from practitioners in the development of new medical devices and
procedures, facilitating clinical tests of new pharmaceuticals, and contribut¬
ing powerfully to innovations in both pharmaceuticals and medical devices.
The combination of science and clinical applications in one institution is un¬
usual—as Henderson, Orsenigo, and Pisano (1999) and Gelijns and Rosen¬
berg (1999) point out, most western European medical institutions emphasize
clinical practice and applications more heavily than scientific research. In
contrast, U.S. universities and academic research facilities have maintained
an important presence in the R, as well as the D, of R&D throughout the post¬
war period.
The mission orientation of the major federal funders of academic research
is reflected in the distribution of research funding among fields of science and
engineering. By 1989, for example, more than half of academic research in

TABLE 2.2

Funding of U.S. Academic Research by Federal Agencies, 1954-2001

Share of federal research funds for academic R&D

originating within particular agencies

Year NIH NSF DOD NASA DOE USDA Other

1954 n/a 1.3% 51.0% n/a 32.5% 5.2% 10.0%

1971 36.7% 16.2 12.8 8.2% 5.7 4.4 16.0

1976 46.4 17.1 9.4 4.7 5.7 4.7 12.0

1981 47.0 15.7 12.8 3.8 6.7 5.4 11.0

1986 49.4 15.1 16.7 3.9 5.3 4.2 8.4

1991 54.3 14.1 11.3 5.2 6.1 3.8 13.7
1996 55.3 14.5 12.1 5.5 5.0 3.1 4.5
2001 60.5 14.9 8.7 4.4 4.0 2.8 4.7

SOURCE: National Science Board (2002). Data for 2001 are based on preliminary NSF
estimates.
HISTORICAL OVERVIEW

26 science and engineering was in the life sciences. Much of the research funded
by these agencies is appropriately defined as basic research, in that it aimed for
fundamental understanding of the object of study, but it was also positioned
in “Pasteurs Quadrant,” inasmuch as the research was motivated by the desire
to solve practical problems. Since the early 1980s, the central role of the fed¬
eral government in supporting academic research has been supplemented by
increased funding from industry, and university-industry research linkages
have attracted considerable comment. But as we pointed out earlier, these
linkages were well established before World War II. Indeed, the share of uni¬
versity research expenditures financed by industry appears if anything to have
declined during the early postwar period.15 During 1953-58, industry sup¬
ported 8 percent of annual academic R&D spending, on average, a share that
declined to 2.7 percent by 1970, in part as a result of increased federal govern¬
ment funding of academic research. By 1980, industrial support for university
research (excluding university-based FFRDCs) had rebounded to account for
more than 4 percent of academic research spending, and this share increased
further to approximately 7.4 percent by 1998 (National Science Board, 2002).
The structure of federal programs for support of academic R&D reinforced
many of the internationally distinctive characteristics of the U.S. higher edu¬
cation system that were apparent before 1940. The large scale of postwar fed¬
eral funding of academic research and the pluralistic, decentralized structure
of federal R&D programs (even within a single, large R&D-supporting agency
such as the Department of Defense) meant that numerous alternative paths of
R&D were supported in such key technological areas as information technol¬
ogy, biomedical sciences, and materials science during the postwar period.
The ability of federal R&D programs to support broad exploration of alter¬
native applications of fundamentally uncertain technologies proved to be an
important source of U.S. competitive advantage in such embryonic areas as
computer hardware, semiconductors, and, eventually, the Internet (Mowery
and Simcoe, 2002). Equally important, however, was the emphasis on peer re¬
view and interinstitutional competition in virtually all federal programs sup¬
porting academic R&D. The availability of funding from multiple federal
sources, combined with the competitive processes allocating the bulk of this
academic R&D support, powerfully reinforced the interinstitutional auton¬
omy and competition for faculty, students, resources, and prestige that char¬
acterized the pre-1940 U.S. system of higher education. Moreover, the size of
the federal academic R&D budget, as well as the reliance by most federal
agencies on extramural research support rather than public laboratories,
meant that U.S. universities’ research enterprises dwarfed those of other in¬
dustrial economies throughout the postwar period.
 HISTORICAL OVERVIEW

The “new structure” of U.S. academic R&D in the postwar period had im¬ 27
portant effects for industrial innovation and changed a number of its key char¬
acteristics. Although U.S. universities had played a significant role in indus¬
trial innovation in the pre-1940 U.S. economy, many of their important
contributions were exploited by large, established industrial firms such as
Standard Oil of New Jersey or Du Pont. During the postwar period, however,
relatively new firms, many of which drew on universities for personnel or sci¬
entific and technological knowledge, played central roles in the commercial
exploitation and growth of such “new industries” as computer hardware, semi¬
conductors, computer software, and biotechnology. The economic role of
these new firms in these postwar U.S. industries outstripped the importance
of new firms in other industrial economies, such as Germany and Japan.
Moreover, regions such as California’s Silicon Valley or Massachusetts’s Route
128 enjoyed significant advances in income and employment that were attrib¬
utable in part to the presence of major research universities in these regions.

INDUSTRIAL INNOVATION AND UNIVERSITY RESEARCH:

RECENT STUDIES OF THEIR INTERACTION

Thus far we have described the historical origins and development of the U.S.
system of higher education, focusing on the ways in which its scale and struc¬
ture created strong incentives for collaboration between university and indus¬
trial researchers well before the recent growth of university patenting and
licensing. In this section, we summarize a number of recent studies of the
contemporary relationship between university research and industrial innova¬
tion. To what extent do U.S. firms utilize the results of university research as
the source of the technological innovations that they develop and commer¬
cialize, a view of industrial innovation that influenced the drafting of the
Bayh-Dole Act? What importance do industrial managers assign to patenting
and licensing as key channels for the flow of university research findings to in¬
dustrial innovation? How, if at all, do managers’ assessments of these issues
vary among technologies and industries? The studies summarized in this sec¬
tion hed light on these key questions and provide a useful basis for examining
university-industry collaboration and technology transfer.
All of the studies reviewed in this section relied on responses through in¬
terviews or surveys from senior industrial managers in industries ranging from
pharmaceuticals to electrical equipment. The National Research Councils
Government-University-Industry Research Roundtable (GUIRR) (1991) ex¬
amined the contributions of university research to technological innovation,
and Mansfield (1991) surveyed industry managers about the number of their
HISTORICAL OVERVIEW

28 recent innovations that either could not have been developed, or could have
been developed only after significant delay, in the absence of recent academic
research. Two other studies relied on large-scale surveys of industrial R&D
managers. The “Yale survey” (results of which are summarized in Levin et ah,
1987) and the more recent “Carnegie-Mellon survey” (summarized in Cohen
et ah, 2002) asked industrial research managers about the nature and scope of
the influence of university research on industrial R&D. The Carnegie-Mellon
survey also asked respondents to describe the most important channels
through which their firms gained access to the results of university research
for application in their industrial innovation strategies. The Yale survey fo¬
cused mainly on large U.S. firms in the late 1970s and early 1980s, while the
Carnegie-Mellon survey focused on a broader range of large and smaller U.S.
firms in the early 1990s, thereby capturing insights from managers well after
the passage of the Bayh-Dole Act of 1980.
These studies also highlight the interindustry differences in the relation¬
ship between university and industrial innovation. Respondents in all four of
these studies characterize the biomedical sector, especially biotechnology and
pharmaceuticals, as unusual —university research advances affect industrial
innovation more significantly and directly in this field than is true of other sec¬
tors. Biotechnology firms’ managers quoted in the GUIRR study stated that
they relied on university research as a source of inventions. Managers inter¬
viewed in the GUIRR study, however, differentiated between pharmaceuticals
based on biotechnology and other drugs, stating that university research rarely
was the source of new drugs not based on biotechnology, for which the key
work took place in industry. But university research affected the development
of these nonbiotechnology drugs as well — managers interviewed in the
GUIRR study highlighted a number of cases in which academic research had
illuminated the specific biochemical reactions that pharmaceutical firms
needed to find in searching for new drugs. In other cases, university research
advances permitted companies to make a more efficient assessment of possible
uses for drugs they were testing.
Mansfield (1991) found a similarly strong dependence on academic re¬
search in his survey of pharmaceutical industry managers, who stated that
more than a quarter of the new drugs commercialized by the companies could
not have been developed, or would have been developed only with substantial
delay, without academic research. These managers further asserted that the
development of an additional 20 percent of the drugs introduced by their firms
was substantially aided by academic research.
The nature of the relationship between university and industrial research
 HISTORICAL OVERVIEW

and innovation in the biomedical sector contrasted with that in other indus¬ 29
tries in both the GUIRR and Mansfield studies. Interviewees from electronics
firms reported in the GUIRR study that universities occasionally made rele¬
vant “inventions” but opined that most such inventions came from nonaca¬
demic research. University research did contribute to technological advances,
but its contributions were largely in the form of knowledge of the fundamen¬
tal physics and chemistry underlying manufacturing processes and product in¬
novation, an area in which training of scientists and engineers figured promi¬
nently, and experimental techniques. Mansfield found that the reported
percentage of new products that were “heavily dependent” on academic re¬
search for their introduction was significantly lower in areas other than phar¬
maceuticals. The executives from the information-processing equipment and
instruments industries reported that 10-15 percent of their innovations
depended on academic research (in the sense defined above in our discussion
of Mansfields work). Respondents from the metals industry estimated that
slightly more than 10 percent of new products and processes would not have
been developed in the absence of recent academic research. Even more strik¬
ing is Mansfields finding that in three industries-electrical equipment,
chemical products, and metal products —at most 6 percent of new products
depended on recent academic research.
The Yale and Carnegie-Mellon surveys of industry R&D executives cor¬
roborate the findings of these other studies concerning the differences in the
relationship between university and industrial research in the biomedical and
other industry sectors. The Yale survey queried R&D managers in industry
about the sources of knowledge that affected innovation in their industries.
Only fifteen of the fifty industries with three or more respondents in the Yale
study rated university research as “important’ or “very important to techni¬
cal advance in their line of business (Table 2.3). Once again, pharmaceuticals
(which had only begun to recognize the significance of biotechnology at the
time the survey was administered in the late 1970s) figures prominently on this
list, but R&D managers in engineering and scientific instruments, semicon¬
ductors, and synthetic rubber also reported that university research was “im¬
portant” or “very important” to their innovative activities. Interestingly, many
of the other industries reporting that university research was important to their
innovative activities are related to agriculture and forestry, prominent bene¬
ficiaries of federally funded university research for much of the twentieth
century.
Another set of questions in the Yale survey asked R&D managers to as¬
sess the importance of specific fields of university science in their industries’
HISTORICAL OVERVIEW

30 TABLE 2.3
Industries Rating University Research as
“Important” or “Very Important”

Fluid milk
Dairy products except milk
Canned specialities
Logging and sawmills
Semiconductors and related devices
Pulp, paper, and paperboard mills
Farm machinery and equipment
Grain mill products
Pesticides and agricultural chemicals
Processed fruits and vegetables
Engineering and scientific instruments
Millwork, veneer, and plywood
Synthetic rubber
Drugs
Animal feed

SOURCE: Previously unpublished data from the Yale Survey

on Appropriability and Technological Opportunity. For a descrip¬
tion of the survey, see Levin et al. (1987).

innovative activities (Table 2.4). Virtually all of the fields of university research
that were rated as “important” or “very important” for their innovative activi¬
ties by survey respondents are related to engineering or applied sciences. As
we noted previously, these fields of U.S. university research frequently devel¬
oped in close collaboration with industry.
With the exception of chemistry, very few basic sciences appear on the list
of university research fields deemed by industry respondents in the Yale sur¬
vey to be highly relevant to their innovative activities. But the absence of fields
such as physics and mathematics in Table 2.4 should not be interpreted as in¬
dicating that academic research in these fields does not contribute directly to
technical advance in industry. Instead, these results reflect the fact that fun¬
damental advances in physics, mathematics, and related sciences penetrate
industry gradually. I heir effects on industrial innovation are realized only af¬
ter the passage of considerable time, a characterization very similar to that of
the GUIRR interviewees in the electronics industries. Indeed, in many cases
the effects of advances in these areas of science on industry are realized
through the incorporation of such advances into the applied sciences, such as
chemical engineering, electrical engineering, and material sciences.
I he findings of the Carnegie-Mellon survey (Cohen et al., 2002) concern¬
ing interindustry differences in the importance of university research are sim¬
ilar to those of the Yale survey.16 Although most industries do not assign great
importance to university research as a contributor to their innovation activi-
 HISTORICAL OVERVIEW

ties, pharmaceuticals stands out among the industries that view university re- 31
search results as important, as do some of the electronics industries. The find¬
ings of the Carnegie-Mellon survey regarding the fields of university research
that industry considered important for innovation also were similar to those of
the Yale survey.
The Carnegie-Mellon survey included a number of additional questions
concerning the ways in which university research influenced the industrial
R&D agenda and the channels through which industry gained access to such
research results. On the first point, the responses summarized in Table 2.5 in¬
dicate that in most industries, university research results play little if any role
in triggering new industrial R&D projects; instead, the stimuli originate with
customers or from manufacturing operations. Here as elsewhere, pharmaceu¬
ticals is an exception —university research results trigger industrial R&D proj¬
ects in a significant number of cases. But industry R&D managers reported
that university research results most often aided problem solving in the course
of a R&D project, rather than affecting the decision to initiate such projects.
The Carnegie-Mellon survey also queried R&D managers about their
use of different types of research from universities and government research

table 2.4
The Relevance of University Science to Industrial Technology

No. of industries
with “relevance”
scores of Selected industries for which the reported
“relevance” of university research was
Field ^5 s=6 large (3s 6)

Biology 12 3 Animal feed, drugs, processed

fruits/vegetables
Chemistry 19 3 Animal feed, meat products, drugs
Geology 0 0 None
Mathematics 5 1 Optical instruments
Physics 4 2 Optical instruments, electronics
Agricultural science 17 7 Pesticides, animal feed, fertilizers,
food products
Applied math/ 16 2 Meat products, logging/sawmills
operations research
Computer science 34 10 Optical instruments, logging/sawmills,
paper machinery
Materials science 29 8 Synthetic rubber, nonferrous metals
7 3 Surgical/medical instruments, drugs, coffee
Medical science
21 6 Nonferrous metals, fabricated metal products
Metallurgy
Chemical engineering 19 6 Canned foods, fertilizers, malt beverages

Electrical engineering 22 2 Semiconductors, scientific instruments

Mechanical engineering 28 9 Hand tools, specialized industrial machinery

SOURCE: Previously unpublished data from the Yale Survey on Appropriability and Technological
Opportunity in Industry. For a description of the survey, see Levin et al. (1987).
HISTORICAL OVERVIEW

32 laboratories in their internal innovation activities, and these findings are strik¬
ingly similar to the other data discussed above on the types of knowledge out¬
puts rated by managers as most important for industrial innovation. Respon¬
dents to the Carnegie-Mellon survey reported that general research findings
from “public research” performed in government labs or universities were
used more frequently (on average, in 29.3 percent of industrial R&D projects
in respondents’ firms) than prototypes emerging from these external sources
of research (used in an average of 8.3 percent of industrial R&D projects). In¬
terestingly, research techniques and instruments from these external research
sources were rated as more important in their contributions to industrial
R&D, used in an average of 22.2 percent of projects, than were prototypes.
Respondents from the pharmaceutical industry reported that more than
40 percent of their R&D projects used research findings from universities and
government laboratories, and more than 35 percent used techniques and in¬
struments developed at these sites. But pharmaceutical industry respondents
reported that only 12.3 percent of their industry projects relied on prototypes
developed in university or government research facilities.
A similar portrait of the relative importance of different outputs of univer¬
sity and public-laboratory research emerges from the responses to questions
about the importance to industrial R&D of various information channels
(Table 2.6). Although pharmaceuticals once again is unusual in its assignment
of considerable importance to patents and license agreements involving uni¬
versities and public laboratories, respondents from this industry still rated re¬
search publications and conferences as more important sources of informa-

TABLE 2.5
Rankings of Importance of External Information Sources in
Suggesting New R&D Projects and in Contributing to the
Completion of R&D Projects

% of respondents % of respondents
indicating indicating that
that source source contributed
suggested new to project
Information source R&D project completion

Consultants 22.8% 34.2%

Joint or cooperative ventures 49.6 47.2
Competitors 40.5 11.7
Independent suppliers 45.6 60.6
Internal manufacturing operations 73.7 78.2
Customers 90.4 59.1
Universities/government labs 31.6 36.3

SOURCE: Cohen, Nelson, and Walsh (2002).

 HISTORICAL OVERVIEW

TABLE 2.6
33
Importance to Industrial R&D of Sources
of Information on Public R&D
(including university research)

% of respondents
rating a source as
“moderately” or
“very” important
Information source for industrial R&D

Publications &r reports 41.2%

Informal interaction 35.6
Meetings & conferences 35.1
Consulting 31.8
Contract research 20.9
Recent hires 19.6
Cooperative R&D projects 17.9
Patents 17.5
Licenses 9.5
Personnel exchange 5.8

SOURCE: Cohen, Nelson, and Walsh (2002).

tion. For most industries, patents and licenses involving inventions from uni¬
versity or public laboratories were reported to be of very little importance,
compared with publications, conferences, informal interaction with univer¬
sity researchers, and consulting.

CONCLUSION

The U.S. “system” of higher education is internationally unique among the

Organization for Economic Cooperation and Development (OECD) econo¬
mies in its lack of strong central governmental controls of policy, administra¬
tion, or resources; its large scale; its dependence on local sources of political
and financial support; and its strong interinstitutional competition for re¬
sources, faculty, and prestige. These characteristics have distinguished U.S.
higher education from the higher education systems of most other industrial
economies through much of the twentieth and into the twenty-first centuries.
More important for the concerns of this volume, these structural characteris¬
tics of U.S. higher education created strong incentives for faculty and univer¬
sity administrators to develop links with industrial research. Over the course of
the past century, these close research links between U.S. universities and in¬
dustry have produced important industrial innovations in fields ranging from
pharmaceuticals to mining, as well as in agriculture. But these links also have
influenced the development, indeed the creation, of new areas of engineering
HISTORICAL OVERVIEW

34 and scientific research within U.S. universities. Knowledge, technology, and

personnel have moved from universities to industry, and vice versa, through¬
out the history of this interaction between academic and industrial research.
Collaboration between university and industrial researchers in the United
States meant that many U.S. universities and university faculty were actively
patenting inventions and licensing these patents to industry well before the
1980s. But the flows of knowledge and technology transfer from U.S. univer¬
sities to industrial innovation moved through many other channels in addi¬
tion to those of patents and licensing. Indeed, as we have pointed out in this
chapter, experienced managers of industrial research regard patents and li¬
censes as important sources of industrial innovation in only a few industries.
Instead, other types of interaction, ranging from publication of papers to the
employment within industry of university-trained scientists and engineers
with experience at the frontiers of research, are of greater importance for in¬
novation in many technology-intensive and other industries.
Not only do channels other than patenting and licensing figure more
prominently as sources of knowledge for industrial innovation in these indus¬
tries, the evidence from expert surveys and other sources also highlights the
substantial differences among industries and fields of research and innovation
in the importance of different channels. Pharmaceuticals and the biomedical
sector generally display a relationship between academic research and indus¬
trial innovation that most nearly approximates the “linear model” of inno¬
vation that influenced Vannevar Bush and numerous other leaders of U.S.
science policy during the early postwar period. Not only do advances in fun¬
damental scientific understanding influence the direction of industrial inno¬
vation, but in many cases the task of innovation involves the development and
commercialization by industry of the fundamental advances achieved in aca¬
demia. But the electronics, materials, and chemicals industries all exhibit very
different patterns of interaction between university research and industrial in¬
novation, a topic that we examine at some length in Chapters 6-8.
 3

University Patent Policies and University

Patenting Before the Bayh-Dole Act

This chapter examines the evolution of U.S. university patent policies and
university patenting during the pre-Bayh-Dole era, spanning 1925-80. Al¬
though many of the issues in recent debates about the costs and benefits of uni¬
versity patenting had been articulated by proponents and opponents of uni¬
versity patenting as early as the 1930s, several arguments in favor of university
patenting that figured prominently in these early debates were absent from the
debates over the Bayh-Dole Act in the 1970s and subsequently. Interestingly,
many proponents of university patenting argued in the debates of the 1930s
that universities should avoid a direct role in managing patents and licenses,
an argument that was rarely heard in the debates of the 1970s. The growth of
university patenting and licensing throughout the 1925-80 period also exhib¬
its many of the characteristics of the diffusion of a new organizational phe¬
nomenon, including emulation by universities of one another’s approaches to
patenting and licensing and considerable responsiveness to the demonstration
by some universities of financial gains from licensing of faculty inventions.
Our examination of U.S. university patenting during the 1925-80 period
focuses on the level and technological composition of university patenting
and the characteristics of the universities active in patenting. Although some
universities had begun to patent faculty inventions as early as the 1920s, few
institutions had developed formal patent policies prior to the late 1940s, and
many of these policies embodied considerable ambivalence toward patenting.
Relatively few universities managed their patent portfolios themselves during
the 1925-70 period, but this situation began to change during the 1970s. In
particular, the 1970s are characterized by considerable growth in patenting by
private U.S. universities. Our data also highlight steady growth in biomedical
UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

36 patents as a share of overall university patenting during the postwar period. As

we show in Chapters 6 and 7, the Bayh-Dole Act accelerated the growth of uni¬
versity patenting and resulted in the entry into patenting and licensing by
many universities during the 1980s. But the “transformation wrought by the
1980 Act followed trends that were well established by the late 1970s.

A BRIEF HISTORY OF U.S. UNIVERSITY PATENT POLICIES

The Pre-World War II Debate

Just as the post-1970 surge in U.S. university patenting coincided with in¬
creased research collaboration between universities and industry (Henderson,
Jaffe, and Trajtenberg, 1998a), so did the first wave of university involvement
in patenting, which began after World War I. The expanding links between
university and industrial research during the 1920s and 1930s that we discussed
in Chapter 2 triggered a debate among U.S. research university administrators
over patent policy (McKusick, 1948; Palmer, 1934).1
In 1933, the American Association for the Advancement of Science (AAAS)
Committee of Patents, Copyrights, and Trademarks surveyed the different po¬
sitions on the “patent problem,” as it was then known, that faced university sci¬
entists. Among the questions addressed by the committees report (AAAS,
1934) were: “Should [scientists] proceed to obtain patents? What are the ad¬
vantages in doing this? What are the disadvantages?” (p. 7).2
One of the issues the committee considered was whether patenting was
necessary for “technology transfer.” Although a common criticism of aca¬
demic patenting was “that publication or dedication to the public is sufficient
to give the public the results of work of scientists” (AAAS, 1934, p. 9), the re¬
port concluded that this position was naive, for several reasons. Anticipating
an argument made during the 1970s in support of the Bayh-Dole Act, the com¬
mittee opined that “discoveries or inventions which are merely published and
thus thrown open equally to all, unless of great importance to the industry, are
seldom adopted” (p. 9) and that “ordinarily no manufacturer or capitalist
would be willing to-day to risk his money, and expend time and energy in de¬
veloping on a commercial scale a new product or process without being as¬
sured that his investment in developing the invention would be protected in
some measure” (p. 10).3
The committee argued that a scientist could not expect that publication of
the technical details of his invention would yield social benefits, because of
the presence of “patent pirates” who would “wrongfully appropriate his work”
and “deny the public what he thought he gave it,” either by charging monop-
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

oly prices or withholding the invention from use (AAAS, 1934, p. 10). Were the 37
inventor or his university, rather than a “pirate,” to obtain a patent on the in¬
vention, this risk would be reduced.4
A related argument stated that patents on university inventions were nec¬
essary for “quality control” reasons. Patenting of university research advances
would prevent the unsuccessful or even harmful exploitation of university re¬
search advances by unqualified individuals or firms. Patenting of such ad¬
vances by the university, according to this argument, prevented incompetent
exploitation of academic research that might discredit the research results and
the university.5
Proponents of university patenting thus concluded that doing so enhanced
the public good. University patents advanced social welfare by inducing pri¬
vate parties to develop and commercialize university research results, by pre¬
venting “patent pirates” from patenting the research and charging monopoly
prices, and by allowing institutions to ensure that only reputable parties de¬
veloped the research, protecting the reputations of the invention and the uni¬
versity itself.
As we note in Chapter 5, the first of these three arguments (“patents induce
development”) figured prominently in debates about university patent policy
in the 1960s and 1970s (Eisenberg, 1996), which often asserted that commer¬
cialization of a university invention might require that any patented research
results be licensed on an exclusive or semiexclusive basis. Prevention of “pat¬
ent piracy” and protection of universities’ reputations, however, require only
that an institution obtain a patent and license it widely at low or no royalties.
The committee s report also addressed the potentially negative effect of pat¬
ents on the progress of research in fields where advances are cumulative, that
is, the possibility “[t]hat patents will place unfortunate strictures on other men
who subsequently do fundamentally important work in the same field”
(AAAS, 1934, p. 12). The committee acknowledged such dangers, but argued
that universities could avoid them “by permitting the use of patents on liberal
terms .. . [which] is particularly necessary in the case of broad or basic inven¬
tions” (p. 12).
A third class of issues in the committees report dealt with the tension be¬
tween university patenting and the “open science” norms and institutions of
academe (Merton, 1973). Objections to patenting based on this view included
the assertion “that it is unethical for scientists or professors to patent the results
of their work” (AAAS, 1934, p. 8; the committees report indicated that “this ob¬
jection is probably the vaguest and most frequent one raised”). Although the
committee considered the risk that patenting by university scientists would
UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

38 hinder communalisin and bias academic inquiry away from basic research, a
concern echoed in more recent debates over university patenting (Dasgupta
and David, 1994), its report dismissed these fears.6
The debates over university patent policies in the 1930s treated medical pat¬
ents as a special case. Opposition to medical patents was widespread, based on
the argument that patents restricted the use of new discoveries and therefore
had no place in the medical community (Weiner, 1986).7 Opponents of med¬
ical patents also expressed concern over perceptions of university profiteering
at public expense in the field of public health (McKusick, 1948)- The AAAS
committee acknowledged the “special” nature of patents in the field of public
health (with little elaboration) but suggested that the benefits of patenting dis¬
cussed above, particularly those associated with using patents as a “quality
control” mechanism, were sufficient to warrant patenting of such discoveries.
Although the AAAS report favored university patenting of faculty inven¬
tions, the committee did not endorse direct involvement by universities in pat¬
ent management, arguing instead that “this can preferably be done by inde¬
pendently organized foundations, holding companies or other suitable
experienced organizations” (AAAS, 1934, p. 14). This position reflected the
reluctance of many U.S. universities during the 1925-70 period to directly
manage patenting and licensing, a reluctance that faded somewhat during
the 1970s.
This review of the early debates about university patenting underscores
the widespread awareness of and interest in patenting by U.S. university sci¬
entists and administrators well before the 1970s and 1980s. The primary moti¬
vation for patenting was the protection of the public interest and the preser¬
vation of academic institutions’ reputations. The prospect of licensing income
influenced the entry by several universities into patenting and licensing dur¬
ing the 1930s and subsequently; but through much of the 1925-80 period,
many academic scientists and administrators preferred to avoid direct in¬
volvement in the management of these patents. These tensions were reflected
in the patent policies adopted by leading pre-1940 institutional patenters and
licensors.

Pre-World War II University Patent Policies: An Overview

Since one of the forces that sparked the debate over university patenting was
the growth of research collaboration between U.S. universities and industry, it
is not surprising that the first group of universities to become involved with
patents were land-grant institutions, which, as we noted in Chapter 2, con¬
ducted a good deal of applied research of interest to industry and agriculture.
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

A number of these public universities sought patents for faculty inventions in 39

the expectation that state taxpayers and the local economy would benefit from
their research. But land-grant institutions frequently avoided a management
role in patenting and licensing (Committee on Uniform Patent Policies in
Land Grant and Engineering Experiment Stations, 1922). Instead, several
land-grant and other universities “outsourced” their patent management to af¬
filiated but legally separate foundations or to a third-party technology transfer
agent, the Research Corporation (see Chapter 4 for more discussion of the Re¬
search Corporation).
The first university-affiliated research foundation was the Wisconsin
Alumni Research Foundation (WARF). In 1924, Dr. Harry Steenbock of the
University of Wisconsin developed a method for increasing the vitamin D
content of food and drugs via the process of irradiation. Despite criticism by
many in the medical community and his colleagues at the university, Steen¬
bock decided to patent his findings, based partly on his desire to protect the
public from unscrupulous or incompetent firms and from monopolization of
the industry by a private patentee (Apple, 1996).8 Steenbock offered to assign
his patent to the University of Wisconsin, but the university thought that the
creation of a specialized office to handle patents was not worth the necessary
investment (Apple, 1996). Instead, Steenbock convinced several alumni to cre¬
ate the WARF, an entity affiliated with but legally separate from the University
of Wisconsin that would accept assignment of patents from university faculty,
license these patents, and return part of the proceeds to the inventor and the
university. According to Apple (1996, p. 42), “[wjith this structure, business
matters would not concern or distract the university from its educational man¬
date; yet academe could reap the rewards from a well-managed patent whose
royalties would pay for other scientific work.”
In an action that illustrates the tension between pecuniary and other mo¬
tives (such as those articulated by Steenbock) in university licensing policy,
WARF licensed the Steenbock patents to Quaker Oats on an exclusive basis
for the production of vitamin-enriched cereal products. The foundation also
negotiated a limited number of licenses with pharmaceutical firms in other
fields of use covering the development of vitamin D supplements. These li¬
censing agreements were criticized by some observers, who argued that they
raised the cost of enriched food products that were beneficial to the public
health. Such criticism received additional impetus from the considerable in¬
come that the Steenbock patents earned for the University of Wisconsin
(Weiner, 1986).9
WARFs success and substantial licensing income influenced the develop¬
ment of patent policies at other U.S. universities during the 1930s, when many
UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

40 institutions revised and formalized their patent policies in response to several

forces. First, the Depressions devastating effects on university finances forced
many institutions to seek new sources of income, and the widely acknowl¬
edged profitability of WARF offered an attractive model for emulation. Sec¬
ond, the post-World War I growth in industry-funded research at universities
had led some universities to clarify or codify policies affecting ownership of
faculty inventions. Land-grant universities were the most active in this area,
and Purdue University, the University of Minnesota, and Cornell University
established affiliated but legally separate research foundations that were simi¬
lar to WARF during the late 1920s and early 1930s. In some cases, these re¬
search foundations also sought to increase university research funding from
industry: McKusick (1948, p. 2) noted that “patent administration was only in¬
cidental to [their] primary purpose.” Although this group of schools typically
did not assert rights to all university inventions, faculty with patentable in¬
ventions frequently asked their university’s research foundation to hold and
administer the inventions. Still other universities referred their faculty to
third-party patent management agents, including the Research Corporation.
During the 1930s and 1940s, several universities, including MIT, Princeton,
and Columbia, signed formal Invention Administration Agreements (LAAs)
with the corporation.
Table 3.1 summarizes the 1940 patent policies of the sixteen institutions
classified by Geiger (1986) as the leading U.S. research universities before
World War II: Illinois, Michigan, Minnesota, Wisconsin, California, Colum¬
bia, Harvard, Pennsylvania, MIT, Cornell, Johns Hopkins, Princeton, Yale,
Stanford, Chicago, and the California Institute of Technology. As of 1934, only
two of these institutions had adopted formal patent policies. By the end of the
1930s, however, twelve had done so, an indication of increased institutional in¬
terest in patent management.10
Several elements of these policies deserve mention. First, despite increased
interest in patenting among the leading U.S. research universities and a few
other institutions, most U.S. universities still lacked patent policies before
World War II (Palmer, 1957). Several of the universities that had adopted pat¬
ent policies, including Harvard, Penn, Chicago, and Johns Hopkins, discour¬
aged patenting or prohibited university faculty from patenting, especially in
the field of medicine. Nonetheless, with the exception of the University of
Chicago, each of these “antipatent” institutions acknowledged that, under
special circumstances, patents might be necessary, even in medicine, for
“quality control” purposes of the sort discussed earlier. A number of the in¬
stitutions that allowed faculty patenting used research foundations or the
Research Corporation for patent management (Table 3.1), reflecting their
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

TABLE 3.1

University Patent Policies and Procedures, 1940

University
Formal asserts
systemwide rights to Restrictions
patent nonsponsored on medical
Institution policy research patents Patent management

Illinois Yes Yes No

Michigan No No No
Minnesota No No No University of Minnesota
Research Foundation
Wisconsin Yes No No Wisconsin Alumni
Research Foundation
California Yes No No
Columbia Yes No No University Patents Inc;
Research Corporation
Harvard Yes No Yes
Penn Yes No Yes
MIT Yes Yes No Research Corporation
Cornell No Yes No Cornell Research
Foundation, Inc.
Johns Hopkins Yes No Yes
Princeton Yes No No Research Corporation
Yale Yes No No
Stanford Yes Yes No
Chicago Yes No Yes
California Institute of No No No
Technology

SOURCES: Spencer (1939); Potter (1940).

aversion to in-house patent management. The onset and aftermath of World

War II changed this landscape, as we discuss in the next section.

The Impact of World War II

As we noted in Chapter 2, World War II and the Cold War that followed trans¬
formed the structure of the U.S. national innovation system. Nowhere was this
transformation more dramatic than in U.S. universities. Formerly funded
largely by state governments, the U.S. Department of Agriculture, and indus¬
try, academic research experienced a surge of federal funding. As the growth
in university-industry research links had done during the 1920s and 1930s, in¬
creased federal funding of university research strengthened two motives for
university involvement in patenting. First, the expanded scale of the academic
research enterprise increased the probability that universities would produce
patentable inventions. Second, many federal research sponsors required the
development of a formal patent policy.
UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

42 By the late 1940s, virtually all major U.S. universities had developed pa¬
tent policies, a considerable increase from the situation in the late 1930s
(McKusick, 1948). Summarizing his survey of U.S. universities in the late
1950s, Palmer (1962) found that eighty-five institutions had adopted or revised
patent policies during the 1940-55 period; more than half of these new or
revised policies were announced during 1946—55.11 Typically, patents from
government-funded work were governed by the policies of the sponsoring
agency (see below); privately sponsored research was managed on a case-by¬
case basis, although research sponsors often received preferential treatment.
Significant differences nonetheless remained among U.S. universities in their
assertion of rights to faculty inventions and their willingness to pursue patent
licenses.
As in the prewar period, many universities during the 1950s and 1960s out¬
sourced patent management (McKusick, 1948). Data on Research Corpora¬
tion IAAs reveal the dimensions of this trend: as of 1940, only three of the na¬
tions eighty-nine “Research Universities” (as classified by the Carnegie
Commission on Higher Educations 1973 taxonomy; see below for details) had
signed IAAs with the Research Corporation. By 1950, this number had in¬
creased to twenty (Table 3.2), and by the mid-1960s nearly two-thirds of the
Carnegie Research Universities were Research Corporation clients.
Well into the 1960s, U.S. university patent policies and procedures re¬
flected the ambivalence toward patents that was revealed in the debates of the
1930s. Many institutions continued to avoid direct involvement in patent ad¬
ministration, and others maintained a hands-off attitude toward patents alto¬
gether. Columbia’s policy left patenting to the inventor and patent adminis¬
tration to the Research Corporation, stating that “it is not deemed within the
sphere of the University’s scholarly objectives” to hold patents, and Harvard,

table 3.2
Carnegie Research Universities with Research Corporation IAAs

Number with IAAs with the Research Corporation

Percentage of category with IAAs with the Research Corporation

Type of University 195° 1955 1960 1965 1970 !975 1980

RU1 and RU2 20 39 51 59 62 68 68

(N = 89) 22% 44% 57% 66% 70% 76% 76%
Public 11 22 33 40 40 43 43
(N = 54) 20% 41% 61% 74% 74% 80% 80%
Private 9 17 18 19 22 25 25
(N = 35) 26% 49% 51% 54% 63% 71% 71%
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

Chicago, Yale, and Johns Hopkins adopted similar positions. All of these uni¬ 43
versities, as well as Ohio State and Pennsylvania, discouraged or prohibited
medical patents. Other universities allowed patents on biomedical inventions
only if it was clear that patenting would be in the public interest.12 This insti¬
tutional ambivalence toward patenting began to change during the 1960s, al¬
though the prohibitions on medical patenting at Columbia, Harvard, Johns
Hopkins, and Chicago were not dropped until the 1970s. The pace of change
accelerated during the 1970s in response to federal initiatives in R&D funding
and patent policy.

The Growth of University Involvement in Patent

Management, 1960-80

The Rise of Federal Support for Academic Research

Although federal funding of academic research grew during the first postwar
decade, the rate of growth in such support accelerated in the late 1950s, and
the composition of this support shifted in favor of basic research. U.S. univer¬
sities’ share of total U.S. basic research performance (excluding university-
operated Federally Funded Research and Development Centers [FFRDCs])
nearly doubled from 27 percent in 1953 to 50 percent in 1968, while federal
funds for basic research in U.S. universities increased more than five-fold dur¬
ing the 1958-68 period (National Science Board, 2002). The greatest in¬
creases in funding in the late 1950s and 1960s occurred in the biomedical sci¬
ences, funded largely by the National Institutes of Health (NIH) within the
Department of Health, Education, and Welfare (HEW). By 1971, NIH sup¬
ported nearly 37 percent of academic R&D, a share that grew to more than
46 percent by 1981 (Table 2.2). These increases in federal research funding,
especially in biomedical research, transformed the content of academic
research and elevated the salience of patent management issues at many
universities.
Increased federal research funding also broadened the number and char¬
acteristics of U.S. research universities. The interinstitutional concentration
of federal support for university research was high during the 1940s and early
1950s but declined during the 1950-80 period. The share of federal R&D
funds accounted for by the ten leading academic recipients of these funds
shrank from 43 percent in 1952 to 37 percent in 1958 and declined further still
during the 1960s and 1970s (Table 3.3). In the life sciences, where funding was
less institutionally concentrated than overall federal academic research fund¬
ing (Geiger, 1993), the dispersion of funding also grew during the 1960s and
UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

TABLE 3.3

Concentration of Federal ROD Funds in Universities

Receiving the Largest Federal Obligations, 1963-75

Year Top 10 Top 20 Top 50 Top 100

1963 32.8% 49.5% 71.9% 89.5%

1964 30.8 47.1 71.6 88.9
1965 30.0 45.8 70.5 88.8
1966 29.6 45.4 69.9 88.7
1967 29.1 44.9 69.6 88.1
1968 27.7 43.1 67.4 86.7
1969 27.4 43.2 67.3 86.1
1970 27.8 43.3 67.6 86.1
1971 27.1 42.7 67.2 85.8
1972 27.4 42.8 67.0 85.3
1973 27.1 42.4 67.1 85.1
1974 27.3 42.8 67.2 85.7
1975 25.8 41.4 66.2 85.0

SOURCE: Smith and Karlevsky 1975.

1970s. Reduced concentration in federal support for academic research owed

much to the redistributive instincts of Congress and the actions of research
agency administrators who sought to accommodate these instincts (Graham
and Diamond, 1997).

Changes in Policy, Appropriability, and Opportunity

During the 1960s and 1970s

U.S. universities’ patent policies and practices changed significantly during

the 1970s. First, the Research Corporation, which by 1970 was administering
inventions for more than 200 institutions, began to encourage and assist its
client universities in developing capabilities to manage the early stages of the
technology transfer process, particularly invention screening and evalua¬
tion. 13 A second development resulting from the growth in federal support for
biomedical research was the emergence during the 1970s of molecular biology
as a field characterized by numerous advances in basic science that promised
commercial applications of considerable interest to industry. The growth of
biomedical research increased universities’ interest in capturing revenues
from licensing of biomedical patents in an era of slower growth in overall fed¬
eral research funding and accelerating growth in research costs. A growing
number of U.S. universities began to seek patents for faculty inventions and
(their former ambivalence notwithstanding) began to manage their patenting
and licensing activities themselves (Weiner, 1986).
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-

Change also occurred during this period in federal patent policies. Before
the 1960s, federal agencies lacked any policy beyond case-by-case negotiations
for dealing with requests from universities for title to inventions resulting from
federally funded research. In the mid-1960s, the Department of Defense
(DOD) began to grant title to inventions resulting from DOD-financed re¬
search to universities with “approved” patent policies, that is, policies that
required faculty to report inventions resulting from sponsored research and
allowed universities to obtain patents on these inventions. As we point out in
Chapter 5, two other federal sources of academic research funding, HEW and
the National Science Foundation (NSF), began to negotiate Institutional Pat¬
ent Agreements (IPAs) with universities in 1968 and 1973, respectively. IPAs
eliminated the need for case-by-case reviews of the disposition of individual
academic inventions resulting from federally funded research and facilitated
licensing of such inventions on an exclusive or nonexclusive basis.
A fourth impetus for increased university involvement in patenting and
licensing resulted from the early successes of Stanford University’s Office of
Technology Licensing (OTL). Niels Reimers, the associate director of Stan¬
ford University’s Sponsored Projects Office and the founding director of
Stanford’s OTL, noted that shortly after he was hired by the university in
1968: “I looked up the income we had from Research Corporation from ’54
to ’67, and it was something like $4,500. I thought we could do a lot better
licensing directly, so I proposed a technology licensing program” (Reimers,
i998)-
Reimers s technology licensing program, which had begun in 1968 as an ex¬
periment, focused more on marketing and less on the administrative and legal
details of patent management. Reimers staffed his licensing program with in¬
dividuals skilled in technology evaluation and marketing, rather than the at¬
torneys who dominated the staffs of most U.S. university licensing programs in
the 1960s. The pilot program increased university licensing income to $55,000
in its first year (more than twelve times what it had earned via its contract with
Research Corporation in the previous thirteen years), the experiment was
judged a success, and Stanford created its OTL in 1970.
All of these developments sparked considerable interest among U.S. uni¬
versities in expanding or establishing university technology licensing offices.
The dimensions of this increased university involvement in patent manage¬
ment and technology transfer during the 1970s are apparent in Figure 3.1,
which shows “entry” into technology transfer activities by year, based on data
from the Association of University Technology Managers (AUTM, 1998).14 As
the figure shows, the number of universities establishing technology transfer
UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

46 offices and/or hiring technology transfer officers began to grow in the late
1960s, well before the passage of the Bayh-Dole Act. Although the Act was fol¬
lowed by a wave of entry by universities into management of patenting and li¬
censing, growth in these activities was well established by the late 1970s.

UNIVERSITY PATENTS, 1925-80

A number of scholars have documented the role of the Bayh-Dole Act in the
growth of patenting and licensing by universities since 1980 (Henderson, Jaffe,
and Trajtenberg, 1998a,b). But Bayh-Dole is properly viewed as initiating the
latest, rather than the first, phase in the history of U.S. university patenting.
And this latest phase is characterized by a higher level of direct involvement
by universities in management of their patenting and licensing activities, in
contrast to the reluctance of many U.S. universities to become directly in¬
volved in patenting prior to the 1970s. Public universities were more active in
patenting than private institutions during much of the pre-Bayh-Dole era, re¬
flecting the strong incentives that they faced to reap the benefits of university
research for local taxpayers and the importance of applied research at many of
these institutions. By the 1970s, however, both public and private universities
had become directly involved in patenting.
In this section, we examine the growth of U.S. university patenting during
1925-80. We provide an overview of the characteristics of active university
patenters and changes in the technological fields of university patents during

figure 3.1 Proportion of Carnegie RUis with > 0.5 FTE technology transfer per¬
sonnel, 1965-95
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

this period. We also consider some reasons for the increased direct involve¬ 47
ment of universities in patent management that became apparent in the 1970s.

Basic Trends

Figure 3.2 plots the number of patents assigned to American universities from
1925 to 1980. Our data cover university patents for every fifth year during 1925-
45 and annual counts for 1948-80.15 During the 1925-80 period, U.S. univer¬
sities’ share of all domestically assigned U.S. utility patents grew from zero to
slightly less than 1 percent. After growing during the 1940s and 1950s, the num¬
ber of university patents remained roughly constant until the 1970s, when uni¬
versity patenting increased significantly. The total number of university pat¬
ents issued during the 1970s alone is 1.5 times the total number of university
patents issued in the previous two decades.16 Although the surge in U.S. uni¬
versity patenting after the passage of the Bayh-Dole Act in 1980 has been
widely noted, Figure 3.2 suggests that university patenting began to grow as a
share of U.S. patenting during the decade prior to Bayh-Dole.

400 -

350 -

300 -

250 -

1-1
<L>

| 200 -
Z

150 -

100 -

50 -

0
o
O'

Issue Year

figure 3.2 University patents, 1925-80

UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

48 Further evidence that the growth in university patenting predates Bayh-

Dole is seen in Figure 3.3, which shows the ratio of total U.S. university patent¬
ing to total prior-year academic R&D expenditures for application years 1963—
93.17 Aggregate university “patent propensity” does increase after 1981 (as
pointed out by Henderson, Jaffe, and Trajtenberg, 1998a,b), but the trend rate
of growth is stable during and after the 1970s; there is no evidence of a sharp
increase in the rate of growth in the patent propensity of universities after
Bayh-Dole.
The heterogeneity among U.S. universities’ patent policies that we noted
earlier is reflected in considerable differences in the size and characteristics of
the patent portfolios that each university acquired during the 1925-45 period.
Patents issued to the nations leading pre-1940 research universities (the
“Geiger 16”) accounted for just over half of the academic patents (excluding
Research Corporation patents) issued during the 1925-45 period (Table 3.4).
Ivy League institutions such as Harvard and Yale are missing from the list
of “Geiger 16” universities in Table 3.4 that obtained patents during 1925-45,
reflecting their aversion to direct institutional involvement with patents that
was discussed earlier. Individual faculty from these universities, however, used
the Research Corporation for patent administration beginning in the late
1920s.18 Since MIT signed an IAA with the Research Corporation in 1937 (see
Chapter 4), the data in Table 3.4 omit considerable patent activity at MIT af¬
ter this year.19 The table shows that several other private universities active in
both engineering and the applied sciences before World War II, such as the

figure 3.3 University patents per R&D dollar, 1963-93

 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

TABLE 3.4
49
Patenting by U.S. Universities, 1925-45

Total Total
Geiger r6 patents Others patents

California Institute of
Technology 10 Purdue University 13
University of Illinois 7 Washington State University 7
University of Minnesota 7 Iowa State University 6
Stanford University 5 Illinois Institute of Technology 4
University of Michigan 4 Carnegie Mellon University 3
University of Pennsylvania 4 Ohio State University 2
University of Wisconsin 4 University of Kansas 2
University of California 3 Fordham University 1
Cornell University 2 Louisiana State University 1
MIT 2 Saint Louis University 1
Tennessee State University 1
University of Cincinnati 1
University of Iowa 1
University of New Hampshire 1
Other 2
Total 48 Total 46

NOTE: Data collected at five-year intervals.

California Institute of Technology and Stanford University, accounted for a

substantial number of patents during the 1925-45 period.20
Public universities were more heavily represented in patenting than private
universities during the 1925-45 period, both within the top research universi¬
ties and more generally (Table 3.4). As we noted earlier, a number of public
universities had established affiliated but legally separate research foundations
that were similar to WARF in the late 1920s and early 1930s. Patents assigned
to these university-affiliated foundations show up as “university patents” in
Table 3.4, although the parent university was not directly involved in man¬
agement of patenting and licensing.
This dominance of overall academic patenting by state universities per¬
sisted through much of the postwar era. But private universities increased their
share of academic patenting from 14 percent in i960 to 39 percent in 1970 and
45 percent in 1980.21 This growth in the share of private-university patents is
particularly noteworthy, inasmuch as it occurred during a period of rapid
growth in overall U.S. university patenting. The era of increased public fund¬
ing for academic research thus was associated with an expanded role for pri¬
vate universities in managing the patents received by their faculty on publicly
and privately funded research.
 POLICIES AND PATENTING BEFORE BAYH-DOLE

Dispersion and Entry in University

Patenting During the igyos
The 1970s represented the most dramatic period of change in U.S. university
patenting during the 1945—80 period, and arguably during the entire 1925 — 80
period. Overall university patenting grew significantly and became less con¬
centrated; more and more universities chose to manage their patents them¬
selves; and biomedical inventions increased in importance within university
patenting and licensing. Here, we analyze the sources of expanded university
patenting, focusing on the increased dispersion of patenting among universi¬
ties and the sources of entry into patenting by universities with limited expe¬
rience in these activities.
The dynamics of federal support affected growth in the number of univer¬
sity patents and change in the characteristics of universities that were active
patenters. Increased federal funding of academic R&D was a necessary con¬
dition for the growth of university patenting depicted in Figure 3.2. But in¬
creases in the interinstitutional dispersion of federal research funds during the
1960s and 1970s also appear to have affected the concentration of academic
patenting among U.S. universities.
In order to explore the existence and extent of such dispersion in patenting,
we examined the distribution of university patents among the institutional cat¬
egories established by the Carnegie Commission on Higher Education in
1973. The Carnegie Commission classified the nations 173 doctorate-granting
institutions as “Research Universities” or “Doctoral Universities.” Institutions
that awarded at least fifty doctorates in 1969-70 and were among the fifty lead¬
ing recipients of federal financial support in at least two of the three years
1968-69, 1969-70, and 1970-71 were placed in the “Research University 1”
(RUi) category. An institution that awarded at least fifty doctorates in 1969-70
and ranked between fifty and one hundred in federal financial support in two
of the three years was categorized as a “Research University 2” (RU2). The
commissions Doctoral Universities category includes all other institutions
that granted more than ten doctorates in the 1969-70 period.22
Figure 3.4 shows postwar patenting by Research Universities, Doctoral Uni¬
versities, and Other Institutions (including those not classified by the
Carnegie Commission). The Research Universities account for the majority
of university patents throughout the 1948-80 period. But the share of patents
accounted for by RU2S increased from 5 percent to 12 percent during this pe¬
riod, and the concentration of patents within the overall Research Universities
category decreased considerably, especially after 1970. Increases in the level
and interinstitutional dispersion in university patenting thus appear to track
the growth and dispersion of federal funding of U.S. academic research.
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

Growth in university patenting during the 1970s reflected expanded patent¬

ing by Carnegie Research Universities (RUis and RU2S) with little or no pre-
\ ious experience in patenting. We defined Carnegie Research Universities
with ten or more patents in the 1950-69 period as “incumbents” and those
with fewer than ten patents in the 1950-69 period as “entrants.”23 Of the
seventy-seven Carnegie Research Universities that obtained patents during
the 1970s, fifty-three were entrants according to our definition. Figure 3.5
shows that growth in university patenting during the 1970s is due in large part
to increased patenting by entrants.
These entrants came disproportionately from the ranks of the second tier of
the Carnegie Research University category, the RU2 institutions, which are
more heavily represented among entrant institutions than among incumbents.
This increased patenting by relatively inexperienced academic patenters may
reflect increased dispersion of federal research funding (especially biomedical
research funding) among institutions. Overall and within the two classes of
universities, entrants’ patents include a higher share of biomedical patents
than do those of incumbents.
A second factor that affected growth in patenting by universities during the
1970s was the negotiation of IPAs with federal research funding agencies by

Carnegie DU1, DU2 Carnegie RU1, RU2 -**- Other

figure 3.4 University patents by university category, 1948-80

UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980
Issue Year
* Entrants n Incumbents

figure 3.5 Patenting in 1970s by entrants and incumbents (Carnegie RUis

and RU2S)

academic institutions. As we noted earlier, IPAs lowered the costs of patenting

university inventions that resulted from federally funded research, albeit less
significantly than the subsequent Bayh-Dole Act. Slightly more than half
(52 percent) of the seventy-seven Carnegie Research Universities receiving
patents during the 1970s had IPAs with either the NSF or HEW.24 IPAs were
somewhat more common for incumbent institutions (68 percent had IPAs)
than entrants (46 percent) and much more common for RUis (64 percent had
IPAs) than RU2S (31 percent). Figure 3.6 shows that institutions with IPAs
dominated the growth of university patenting during the 1970s. Nonetheless,
although IPAs may have encouraged entry by lowering the costs of patenting
and licensing, fewer than half of entrant institutions had IPAs; patenting
during the 1970s grew for both entrants with IPAs and entrants without IPAs.
The diffusion of IPAs alone thus does not explain entry by universities into
patenting.
These data on the changing sources of university patenting during the
1970s highlight three aspects in which this decade departed from historical
trends and prefigured the 1980s and 1990s. First, during the 1970s private uni¬
versities expanded their patenting and licensing of faculty inventions. A sec¬
ond and closely related shift was the growth in the role of public and private
universities in directly managing these activities. Finally, the 1970s were char¬
acterized by increased entry into patenting and licensing by universities with
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

little experience in these activities, a cohort of entrants drawn primarily from 53

RU2 institutions that specialized in biomedical patenting to a greater degree
than incumbents. All of these trends in university patenting were influenced
by increased interinstitutional dispersion of federal research funding, the
growth of IPAs, and the success of entrants such as Stanford University. De¬
termining the relative importance of these various factors in affecting the
three key trends discussed above remains an important task for future re¬
search. Interestingly, only one of these factors (the IPAs) represented a change
in federal policy toward the patenting of publicly funded research. It is likely
that a similarly diverse range of factors, and not the Bayh-Dole Act alone, un¬
derpinned the continued growth of U.S. university patenting after 1980.

University Patents by Technological Field

A final issue of interest in academic patenting is the distribution among tech¬

nology fields of university patents during 1925-80. Figure 3.7 displays this in¬
formation for university patents during 1925-45, and Figure 3.8 displays simi¬
lar data for the 1948-80 period (both figures exclude Research Corporation
patents). The data in Figure 3.7 highlight the importance of “chemical” pat¬
ents during the pre-1940 period. The prominent role of these patents reflects

250 -

■ Institutions w/o IPAs

□ Institutions w/ IPAs

200

I 150

I 100
z

50

iJiJiJiJi
1970 1971 1972 1973
IiJiIi11
1974 1975
Issue Year
1976 1977 1978 1979 1980

figure 3.6 Patenting by Carnegie Research Universities by IPA status

UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

Issue Year
□ Other □ Drugs and Medicine
a Mechanical ■ Chemicals
13 Electronic, Optical, Nuclear

figure 3.7 Distribution among technology classes of university patents,

x925—45

interwar academic research in fields including agricultural chemistry, indus¬

trial chemistry, and chemical engineering, as well as the strong university-
industry linkages in those fields of research (Rosenberg and Nelson, 1994).
Nearly 60 percent (22/37) °f chemical patents issued to universities during
this period covered organic compounds. Some of the “organic compounds”
that are classified as chemicals in Figure 3.7 are vitamins and could also be
classified as biomedical inventions.25 During the interwar period, U.S. uni¬
versities made significant contributions to vitamin research and synthesis
(Apple, 1996), and the most lucrative academic patent licenses in this period
involved vitamin-related inventions. These included the Williams-Waterman
method for synthesizing vitamin BL (developed at the University of California,
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

Berkeley and administered by the Research Corporation), Milass vitamin A

and B formulations (developed at MIT and administered by the Research Cor¬
poration), and Steenbocks method for producing vitamin D (administered by
WARF).
Although data on the distribution of licensing revenues among university
patents are limited, biomedical inventions (including vitamins) appear to
have accounted for the majority of licensing revenue from university patents
issued before Bayh-Dole. Blumenthal, Epstein, and Maxwell (1986) suggest

400 -

350 -

<D

Issue Year

□ Other □ Drugs and Medicine

Hi Mechanical ■ Chemicals
E3 Electronic, Optical, Nuclear

figure 3.8 Distribution among technology classes of university patents,

1948-80
UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-DOLE

56 that throughout its history, WARF earned most of its revenues from a handful
of biomedical inventions, and Chapter 4 shows that this was also true for the
Research Corporation from the 1940s to the 1970s. In spite of the reluctance
of many universities to seek such patents prior to the 1970s, medical patents
have been important revenue sources for academic licensors throughout the
twentieth century.
Increased federal financial support for biomedical research during the
postwar era is reflected in the growing share of biomedical patents within
our university patent data during and after the 1970s. Much of the growth in
university patenting during the 1970s occurred in the biomedical area (Fig¬
ure 3.8). Nonbiomedical university patents increased by 90 percent from the
1968-70 period to the 1978-80 period, but biomedical university patents in¬
creased by 295 percent. This rapid growth in biomedical patents also reflected
the expansion of the IPA program associated with the major biomedical fund¬
ing agency (HEW) during the 1970s. The increased share of the biomedical
disciplines within overall federal academic R&D funding, the dramatic ad¬
vances in biomedical science that occurred during the 1960s and 1970s, and
the strong industrial interest in the results of this biomedical research all af¬
fected the growth of university patenting during this period.

CONCLUSION

Patenting and licensing by universities of faculty inventions have a long his¬

tory in the United States. For much of the 1925-80 period, the appropriateness
of this activity as one of the many missions of U.S. universities was itself a sub¬
ject of debate, the terms of which anticipated many of the arguments articu¬
lated during and after the framing and passage of the Bayh-Dole Act of 1980.
As we point out in our discussion of the origins of the Bayh-Dole Act in Chap¬
ter 5, the terms of the debate over university technology licensing shifted
somewhat between the 1930s and the 1970s from an emphasis on the regional
economic benefits associated with the licensing by land-grant universities of
faculty inventions to a focus on the benefits to the national economy associ¬
ated with the more rapid and successful technology transfer that (according to
supporters of the Bayh-Dole Act) could be attained by patenting and licensing
faculty inventions. This shift in the terms and geographic delimitation of the
economic benefits highlighted in debates over university technology licensing
reflected, of course, the shift in the scale and sources of financial support for
university-based research.
Concern among many university administrators about their exposure to po¬
litical criticism appears to have limited the willingness of at least some U.S.
 UNIVERSITY POLICIES AND PATENTING BEFORE BAYH-

universities to assume a direct role in the management of patenting and li¬

censing for much of the 1925—70 period. But the 1970s witnessed an upsurge
in entry by universities, particularly private universities, into direct manage¬
ment of patenting and licensing activities. The growth in university patenting
during the 1970s reflected shifts in the underlying content of academic re¬
search, the effects of the longer-term tendency for federal funding of academic
research to become more concentrated in biomedical research and simulta¬
neously less concentrated among institutions, and a shift (before Bayh-Dole)
in federal policy toward university patenting and licensing. Nonetheless, prior
to 1980, federal policy remained ambivalent toward university licensing, evi¬
denced in the debates over the appropriateness of exclusive licensing under
IPAs that we discuss in Chapter 5.
The historical evolution of university patenting behavior in the United
States reflects the unanticipated interaction of an array of federal and state
policies, rather than intellectual property regimes alone. Indeed, many of the
features of university patenting commonly associated with the post-Bayh-Dole
period have deep roots in the pre-1980 period. And as we emphasized in Chap¬
ter 2, this lengthy involvement in patenting has been influenced by the un¬
usual structure of the U.S. university system. The decentralized system for
funding of universities created strong incentives for public universities to pur¬
sue research of interest to local firms and to patent their inventions for the
benefit of state taxpayers. But the broader environment of interinstitutional
competition, administrative autonomy, and the constant search for resources
created strong incentives for both public and private universities to develop
strong links with industrial research. The changing structure of these incen¬
tives, especially change in the sources and structure of funding for university
research, has been the most powerful influence on the long history of U.S.
university patenting and licensing, before and after the Bayh-Dole Act.
 4

The Research Corporation and University

Technology Licensing, 1912-80

We argued in Chapter 3 that U.S. universities' reluctance to become directly

involved in patenting and licensing during much of the 1925—80 period led
many of them to contract with the Research Corporation to manage these
activities. This chapter discusses the history of the Research Corporation in
detail. The corporation played a central role in university patenting and li¬
censing before 1980, and its history anticipates many of the issues currently
confronting contemporary university technology transfer offices. These in¬
clude the high costs of patent management, the severe uncertainty over the in¬
frequent and unpredictable “arrival rate” of the “home run” patents that ac¬
count for the majority of licensing revenues, the need for close interaction
with faculty inventors, and the difficulties in balancing licensing revenues
with other objectives of university patent and licensing programs.
Immediately below, we discuss the origins of the Research Corporation in
the patenting activities of Frederick Cottrell of the University of California.
The next section describes the corporations evolution by 1945 into a manager
of patenting and licensing of university inventions for many of the leading re¬
search universities in the United States. The section that follows examines the
zenith during the 1950s and 1960s of the Research Corporations importance
as a technological intermediary operating on behalf of academic inventors,
with the foundation and growth of the corporations Patent Management Di¬
vision. But even during its period of greatest activity and influence, the Re¬
search Corporation experienced low profitability and high costs, and the final
section examines its decline during the 1970s and 1980s.
 THE RESEARCH CORPORATION AND LICENSING

THE ORIGINS OF THE RESEARCH CORPORATION

59

The Research Corporation originated from the research of Frederick Gardner

Cottrell of the University of California at Berkeley during the early twentieth
century. Consistent with Chapter 2's portrayal of the practical motives for
much U.S. university research, Cottrell’s research was motivated in part by his
interest in problems of industrial technology and practice (Cameron, 1993).
Beginning in 1905, Cottrell conducted research on industrial air pollution
control, responding to a growing nuisance in his native San Francisco. Build¬
ing on the work of the British scientist Sir Oliver Lodge, Cottrell invented the
electrostatic precipitator, a device that removed dust and fumes from smoke¬
stacks by electrically charging them and collecting them on an oppositely
charged plate.
Cottrell received the first of six patents on the electrostatic precipitator in
1907. Although he was not interested in amassing a personal fortune from his
inventions, Cottrell opposed placing his intellectual property in the public do¬
main, arguing that

a certain minimum amount of protection is usually felt necessary by any manufac¬

turing concern before it will invest in machinery or other equipment, to say noth¬
ing of the advertising necessary to put a new invention on the market. Thus a num¬
ber of meritorious patents given to the public absolutely freely by their inventors
have never come upon the market chiefly because “what is everybody’s business is
nobody’s business.” (Cottrell, 1912, p. 865)

In other words, Cottrell believed that clearly defined intellectual property

rights were needed to provide firms with the incentives to undertake the costly
research, development, and marketing activities necessary to bring an em¬
bryonic invention such as the precipitator to commercialization. In this
argument, Cottrell anticipated subsequent work by scholars such as Kitch
(1977) and the arguments set forth by the supporters of the Bayh-Dole Act (see
Chapter 5).
Cottrell intended to license his patents and use the proceeds to support sci¬
entific research. Implementation of this plan, however, required an organiza¬
tion to manage the licenses. Cottrell first considered using the University of
California as a licensing manager but rejected this possibility because of his
belief that the involvement of university administrators in licensing manage¬
ment could have detrimental consequences for the culture of scientific re¬
search at the university, anticipating some of the concerns discussed in the re¬
port of the American Association for the Advancement of Science (AAAS)
Committee on Patents, described in Chapter 3:1
THE RESEARCH CORPORATION AND LICENSING

6o A danger was involved, especially should the experiment prove highly profitable to
the university and lead to a general emulation of the plan. University trustees are
continually seeking for funds and in direct proportion to the success of our experi¬
ment its repetition might be expected elsewhere . . . the danger this suggested was
the possibility of growing commercialism and competition between institutions
and an accompanying tendency for secrecy in scientific work. (Cottrell, 1932, p. 222)

Cottrell approached the Smithsonian Institution, where his plan was em¬
braced by Secretary Charles Walcott but rejected by the Institutions regents,
reflecting their desire to avoid any involvement in patent administration. In¬
stead, a corporation, in which Walcott played an advisory role, was formed to
oversee patenting and licensing. The Research Corporation was founded in
1912 with the assistance of none other than William Howard Taft, then a
Smithsonian regent, in drafting its charter.
Although he originally intended to assign the corporations licensing royal¬
ties to the Smithsonian for the “increase and diffusion of knowledge,” Cottrell
subsequently decided to use the organizations earnings to support scientific
research at institutions other than the Smithsonian. The corporation also
managed the licensing and development of the electrostatic precipitator pat¬
ents, and Cottrell argued that the corporation should assume responsibility for
managing the licensure of donated patents in other technological areas:

The ever growing number of men in academic positions who evolve useful and
patentable inventions from time to time in connection with their regular work and
without looking personally for any financial reward would gladly see these further
developed for the public good, but are disinclined either to undertake such devel¬
opments themselves or to place the control in the hands of any private interests.
(Cottrell, 1912, p. 865)2

In Cottrells view, the purpose of the Research Corporation “was not merely
to produce revenue for scientific research, but to act as a sort of laboratory of
patent economics and to conduct experiments in patent administration” (as
cited in McKusick, 1948, p. 208). From its inception, he envisioned the Re¬
search Corporation as an entity that would develop and disseminate tech¬
niques for managing the intellectual property of research universities and sim¬
ilar organizations.
During its first twenty-five years, the Research Corporation focused on
three activities: (1) designing, manufacturing, and installing electrostatic pre¬
cipitation and other equipment arising from patents that it owned, adminis¬
tered, or purchased; (2) managing patentable inventions from educational or
research institutions that were donated to the corporation; and (3) providing
grants for scientific researchers. The first of these three tasks, however, ac¬
counted for the bulk of the corporations activities during this period.
 THE RESEARCH CORPORATION AND LICENSING

THE EVOLUTION OF THE RESEARCH CORPORATION, 1912-45 6l

Early Activities: 1912-36

Frederick Cottrell played no official role in the Research Corporation once it

was founded, although he exerted considerable influence on its early deci¬
sions and took part in a number of the corporations activities until his death
in 1948.3 The corporations directors and managers quickly learned that ad¬
ministration of the Cottrell patents was more difficult than they had antici¬
pated. Merely licensing the precipitation patents to firms resulted in minimal
technology adoption and modest royalty income, since application of the pre¬
cipitator technology required customization to meet the technical needs of
different industries. Few engineers outside the corporation had the skills nec¬
essary to adapt the electrostatic precipitator to these varied applications. As a
result, the Research Corporation extended its activities beyond those of a
“pure” technology intermediary and entered the precipitator installation and
design business. Research Corporation staff (which by 1917 included forty-five
engineers) performed research and development to adapt precipitator tech¬
nologies to licensees’ plants.
Despite these unanticipated investments, the electrostatic precipitation
business became profitable by 1915, and the Research Corporation began to
award research grants in 1924. During the next several decades, the Research
Corporation dominated the electrostatic precipitator market (White, 1963).
Since the corporations “endowment” of Cottrell’s patents proved insufficient
to support its licensing activities, the organization acquired additional patents
from other inventors and pursued R&D in this technology field.
As a by-product of its work in the precipitation industry, the Research Cor¬
poration developed considerable expertise in patent management and litiga¬
tion. As early as 1913, it began to pursue infringers of the Cottrell patents. By
1918, the Research Corporation was “making the securing of business second¬
ary to the protection of patent rights” in the precipitation field, devoting more
effort to the management of its patent portfolio than to the installation of pre¬
cipitators.4 Their experience with the precipitation patents expanded the skills
of Research Corporation staff in seeking out potential licensees, negotiating li¬
cense agreements, and overseeing the development and marketing of inven¬
tions.5 The corporation also developed numerous contacts with patent attor¬
neys throughout the country and with scientists at several university
laboratories, the site of much of the development work on the electrostatic
precipitators.
The successes of the Research Corporation in managing its precipitation
patent portfolio drew the attention of a number of university scientists and
THE RESEARCH CORPORATION AND LICENSING

62 independent inventors who sought assistance with the patenting and licensing
of their inventions in fields other than precipitators. Following Cottrells ex¬
ample, several of these scientist-inventors wished to assign their licensing roy¬
alties to the Research Corporation to support its philanthropic activities. The
expanding research collaboration between U.S. universities and industry and
the related growth of science-based industry increased the volume of com¬
mercially valuable academic research in the 1920s and 1930s, resulting in more
and more requests from academic inventors to the Research Corporation for
assistance in the management of patenting and licensing.
Beginning in the late 1920s, the Research Corporation began to accept do¬
nations of patents outside of the precipitation field and received a number of
patents in a diverse array of technical fields.6 Among the most important of
these inventions was the Williams-Waterman process for synthesis of vitamin
B,. Developed in Robert R. Williams’s laboratory at the University of Califor¬
nia in 1932, the Williams-Waterman process was important in the fight against
malnutrition and was immediately recognized to have significant income
potential. Williams and Robert E. Waterman filed for a patent on their new
process with the aid of the Research Corporation, which also assisted the in¬
ventors in the successful prosecution of an interference hearing. The patent
issued in 1935 and was assigned to the Research Corporation.
The Williams-Waterman invention was an important development in the
history of the Research Corporation for several reasons. One of the first in a se¬
ries of public health-related inventions that accounted for a growing share of
the Research Corporations licensing revenues during the next two decades,
the patent was the first “home run” in a field other than electrostatic precipi¬
tation. Income from this and other biomedical patents proved indispensable
to the corporations growth. The corporations involvement in the Williams-
Waterman interference hearing also expanded its activities in nonprecipita¬
tion technologies from administering patents to helping secure them. Like
contemporary university technology licensing offices, the corporation recog¬
nized that it had to assume responsibility for a number of activities in addition
to the administration of patents and licenses.
The Research Corporations reputation in academic patent management
also benefited from its successful management of the precipitation patents, as
well as the marketing efforts of Cottrell, who frequently visited leading col¬
leges and universities to advise administration and faculty on matters of
patenting and to promote the Research Corporations services. In addition, the
Research Corporations research grants to university researchers had produced
a number of potentially patentable inventions by the 1930s. Although they
 THE RESEARCH CORPORATION AND LICENSING

were not obliged to do so, many of these inventors turned to the Research Cor- 63
poration for patent assistance.
Another development that aided the Research Corporations growth was
the increased interest of U.S. universities during the 1930s in the financial re¬
turns from patenting and licensing of faculty inventions. But as we pointed
out in Chapter 3, many U.S. universities in the 1930s avoided direct manage¬
ment of patent prosecution and licensing, in contrast to their embrace of this
role during the 1970s and 1980s. This reluctance was partly due to the politi¬
cal controversy that had erupted over the University of Wisconsin’s patenting
and licensing activities (although these were managed by the Wisconsin
-Alumni Research Foundation), as well as the fact that the scale of the research
enterprise at most universities during this period was too small to justify direct
involvement in patent administration. As the Research Corporation noted,
“University staffs were not equipped to handle such problems —the number
[of patents] arising in any one year at a single institution would be too small to
justify having expert patent personnel” (Research Corporation, 1972, p. 15).
Faculty inventors therefore were encouraged by administrators, colleagues,
and the broader scientific and technological community to work through the
Research Corporation (Palmer, 1948).7
Nevertheless, despite expanded discussions with U.S. universities and in¬
ventors over management of their patents, management of nonprecipitation
patents accounted for a relatively small part of the Research Corporations ac¬
tivities during the 1930s. Although it handled inventions for individual faculty
members and universities on a case-by-case basis, the corporation had no for¬
mal agreements with universities. This situation began to change, however,
with the negotiation of the Research Corporations first Invention Administra¬
tion Agreement (IAA) with MIT in 1937.

The MIT Invention Administration Agreement

As key precipitation patents expired and the market for electrostatic precipita¬
tors declined during the Depression, the Research Corporation sought to de¬
velop new revenue sources.8 In addition to its work with precipitator patents,
the corporation’s management of its portfolio of nonprecipitation patents had
contributed to the development of unique competencies that could be ex¬
ploited for the mutual benefit of the Research Corporation and academic in¬
stitutions, while sustaining Cottrell’s original vision for the corporation as a
source of funding for scientific research. In addition, as the Research Corpo¬
ration argued in a 1972 document, centralized management of patenting and
licensing on behalf of universities by a specialist could be less costly for each
THE RESEARCH CORPORATION AND LICENSING

64 institution than independent management of their faculty patents: “if there

were an inter-university clearing house for handling patents, experts could be
engaged, contacts with industry could be maintained, legal costs could be
spread, and many other advantages gained for the whole group” (Research
Corporation, 1972, p. 15).
The Research Corporations search for new income sources, along with the
increased interest of U.S. universities in licensing income, led it to expand its
activities in patent management by negotiating agreements with U.S. univer¬
sities. The Research Corporations 1937 IAA with MIT laid the groundwork for
the corporations emergence as a national technology licensing intermediary
for U.S. universities.9
Prior to the arrival of Karl Compton as its president in 1930, MIT did not
have a well-defined institutional patent policy. Instead, the Institute either
ceded rights to the individual inventor or deferred to the requirements of re¬
search sponsors. Shortly after becoming president of MIT, Compton orga¬
nized an internal study of all aspects of its industrial relations, including its
patent policy. After some debate, the Institutes Committee on Patent Policy
(chaired by Vannevar Bush) decided that MIT would assert rights to any fac¬
ulty inventions resulting from research financed by the Institute. The com¬
mittee recommended that procedures be developed “to relieve the Institute of
all responsibility in connection with the exploitation of inventions while pro¬
viding for a reasonable proportionate return to the Institute in all cases in
which profit shall ensue.”10 Accordingly, MIT sought outside parties to ad¬
minister its patents and aid in the patent application process.
The eventual choice of the Research Corporation as MIT’s patent manager
reflected a number of ties linking corporation managers with MIT staff and ad¬
ministration. Howard Poillon, the Research Corporations president since
1928, was a close friend of Karl Comptons and lobbied him for the right to
market MIT’s inventions. The Research Corporation also administered pat¬
ents that covered the inventions of several MIT faculty members, such as Rob¬
ert Van de Graaf’s electrostatic generator and Nicholas Milas’s vitamin for¬
mulas, both of which were based on research funded by the corporation.
An IAA between MIT and the Research Corporation was signed in 1937.
The agreement called for the corporation to establish a Boston office to man¬
age the review and evaluation of MIT inventions. Under the terms of the IAA,
MIT disclosed potentially patentable inventions to the Research Corporation,
which evaluated them and accepted for administration those deemed to have
commercial promise. The corporation agreed “to use its best efforts to secure
patents on inventions so assigned to it and to bring these inventions into use
and derive a reasonable income therefrom” and further to “use its best efforts
 THE RESEARCH CORPORATION AND LICENSING

to protect these said inventions from misuse and to take such steps against in- 65
fringers as [it] may deem for the best interest of the parties hereto, but with the
general policy of avoiding litigation wherever practicable.” All services were
provided at the expense of the Research Corporation, although the corpora¬
tion assumed no financial responsibility for any additional development costs
for MIT inventions. Any licensing income net of expenses was to be divided
between MIT and the corporation on a 60-40 basis. The Research Corpora¬
tion used its portion of the earnings to support its operating expenses and phil¬
anthropic activities.
The outbreak of World War II cut off the flow of inventions from MIT and
other universities, and the Research Corporation virtually ceased its patent ad¬
ministration efforts.11 Shortly after the return of peace, the corporation re¬
assessed its patent management activities, most of which had involved MIT in¬
ventions, noting that they had produced net losses during the 1937-47 period
(1947 Annual Report, p. 1).12 Although none of the forty MIT inventions and
none of the handful of inventions it was administering for other universities
had yet produced significant returns, the corporation opined that four MIT in¬
ventions had “bright futures” and asserted that “after ten years of trial of the
economic soundness of the type of educational institution patent manage¬
ment agreement initiated in 1937, the soundness of the idea, in spite of the dor¬
mant war years, seems to be confirmed” (1947 Annual Report, p. 1).
Nevertheless, this internal review argued that the scale of the Research
Corporations patent-management operations was too small to be economical,
noting that
the handling of inventions for educational institutions can be successful only if the
service can be rendered to a considerable number of such institutions so that a large
number of potentially valuable inventions can be administered . . . unless a sizable
number of such inventions is available for administration the numerical odds of un¬
successful developments to successful ones would render the management under
such circumstances too great a risk. (1947 Annual Report, p. 3)

By the time the 1947 Annual Report was published, the corporation had begun
a program to attract more university clients to its intellectual property man¬
agement services.

THE PATENT MANAGEMENT DIVISION

The MIT agreement sparked interest in patent licensing by other universities.

Although the Boston office had originally been established to handle MIT’s in¬
ventions, it also began to handle patent matters for other universities. In 1946,
the Research Corporation established the Patent Management Division to
THE RESEARCH CORPORATION AND LICENSING

66 handle all nonprecipitation patents, consolidating the work previously done in

the Boston and New York offices. Following the policies established in the MIT
agreement and in contrast to the corporations activities in precipitator tech¬
nologies, the Patent Management Division had the following responsibilities:

1. Negotiation of agreements with educational and scientific institutions for the

administration and development of their patents;
2. Evaluation of inventions submitted under these agreements, as well as those
submitted by individual inventors;
3. Submission of patent applications for promising inventions; and
4. Licensing patents to industry.

The Patent Management Division assumed no responsibility for the devel¬

opment of university inventions but worked with universities to improve their
expertise in patenting and licensing of faculty inventions. The division sought
to expand university patenting, thereby increasing the volume of business for
the Research Corporation, and its efforts contributed to growth in the number
of universities initiating formal technology transfer activities during the 1960s
and 1970s (see below for further discussion).13
The combined effects of the formation of the Patent Management Division
and the MIT agreement were significant. The expansion of military and bio¬
medical research conducted in U.S. universities during and after the war had
increased the pool of potentially patentable academic inventions, and federal
funding agencies compelled universities to develop formal patent policies
during the early postwar period. The Research Corporation negotiated IAAs,

figure 4.1 Research Corporation Invention Administration Agreements, 1946-81

 THE RESEARCH CORPORATION AND LICENSING

FIGURE 4.2 Invention disclosures to the Research Corporation, 1955-81

modeled on the MIT agreement, with several hundred other U.S. universities
during the 1940s and 1950s (Figure 4.1). Under the terms of these agreements,
university faculty submitted invention disclosures to the Patent Management
Division, which bore all costs of invention evaluation, patent prosecution, and
licensing. Any royalty income was divided among the academic institution,
the inventor, and the Research Corporation. The Research Corporations
share of the royalties, after deduction of operating expenses, funded its re¬
search grants.
Growth in the number of university clients was paralleled by expansion in
the annual number of inventions disclosed to the Research Corporation (Fig¬
ure 4.2). The average annual flow of disclosures, which never fell below
200 during this period, represented a sharp increase from the average of
roughly 50 disclosures per year received by the corporation (primarily from
MIT) at the time the Patent Management Division was established. But this
increased flow of disclosures did not produce a comparable increase in patents
(Figure 4.3) or gross royalties (Figure 4.4)7 an issue that we discuss in greater
detail below.
As was noted earlier, the Research Corporations Patent Management Divi¬
sion had hoped that its expanded activities in managing university inventions
would increase the corporations licensing income while reducing its reliance
on a small number of donated inventions for the bulk of its income. None¬
theless, the corporations 1951 Annual Report noted that institutional inven¬
tions still accounted for a small share of licensing royalties:
 figure 4.3 Patents issued to the Research Corporation, 1947—81
1996 Dollars

figure 4.4 Research Corporation gross royalties, 1951-81

 THE RESEARCH CORPORATION AND LICENSING

[Most revenues] have been derived from agreements covering specific projects; and
not from the over-all patent management contracts. . . the blanket agreements still
have not “paid off.” The corporations experience shows that this is not to be ex¬
pected except over a period of many years, and even then only in a few instances. It
is important only that there be a sufficient number of these “few instances” and to
that end, it is important that more institutional agreements be secured. (1951 An¬
nua/ Report, p. 42)

This problem was a persistent one for the Research Corporation. In 1947, it
had noted that “none of our other [other than MIT] general agreements with
educational institutions to date show any promise of financial return, but we
continue optimistic of future developments, particularly as the number and
scope of such agreements continues to increase” (1947 Annual Report, p. 2).
Although its portfolio of institutional agreements, patents, and licenses grew
during the 1950s, growth in net revenues proved elusive.

The Patent Management Division’s Invention Portfolio

From the inception of the Patent Management Division, licensing royalties

were dominated by income from a small number of inventions, most of which
were in the biomedical field (Table 4.1). Throughout the 1945-85 period, the
corporation’s top five inventions accounted for the lions share of gross reve¬
nues, a characteristic of the corporations licensing portfolio that also is appar¬
ent in contemporary data on the royalty income of leading U.S. academic li¬
censors (see Chapter 6).
By the mid-1950s, the Research Corporations heavy reliance on biomedical
inventions for income had begun to raise concerns among the corporations
management. The 1955 Annual Report noted that wide fluctuations in the
prices of vitamins had affected the corporations royalty income (see Table 4.1).
Although the report forecast that other types of inventions would “soon” begin
to contribute significantly to the royalty stream, these new revenue sources
failed to materialize.
This problem was attributable in part to the fact that the Research Cor¬
poration’s considerable expertise in prosecuting, managing, and licensing
patents in the areas of precipitation technology and the life sciences did not
translate into comparable strength in other technical fields. As a result, the cor¬
porations activities in nonbiomedical fields were limited. Indeed, Fishman
(1996) notes that one reason for MIT’s 1946 decision to stop using the Research
Corporation as the sole agent for administration of its patent portfolio was
concern over the corporation’s lack of expertise in fields other than the life

sciences.14
Another striking characteristic of the biomedical “home runs” that ac¬
counted for the bulk of the Research Corporation’s income through the early
 TABLE 4.1

Research Corporation Gross Income from “Home Run” Inventions, 1945-85

Gross Share of Share of income

income gross income from biomedical Top 5 grossing
Gross from top 5 from top 5 inventions in inventions
Year income4 earners earners prev. column (descending order)

1945 4,713,417 4,604,511 98% 100% Eschatin

Ergotrate
Merthiolate
Pantothenic acid
Vitamin B1
1950 5,528,142 5,165,811 93% 90% Panthothenic acid
Vitamin B1
Ergotrate
Electromagnetic horns*
Merthiolate
1955 7,485,868 5,996,276 80% 100% Cortisone
Vitamin B1
Vitamin A
Pantothenic acid
Ergotrate
1960 6,516,215 5,389,883 83% 100% Cortical hormones
Vitamin A
Pantothenic acid
Nystatin
Vitamin K
1965 4,155,908 3,684,411 89% 93% Nystatin
Cortical hormones
Pantothenic acid
Glass strength promoter*
Reserpine
1970 13,070,160 12,433,948 95% 43% Hybrid seed corn**
Nystatin
Cortical hormones
Plant growth regulants*
Reserpine
1975 5,823,032 4,755,064 82% 63% Quinoxaline-di-n-oxides
Plant growth regulants*
Nystatin
Maser*
Antibacterial agent
1980 9,041,497 6,547,132 72% 96% Platinum anti-tumor
compounds
Burn ointment
Quinoxaline-di-n-oxides
Radiopharmaceuticals for
skeletal imaging
Delayed release mushroom
nutrients*
1985 13,869,300 12,776,616 92% 96% Plantinum anti-tumor
compounds
Antibacterial agent
Quinoxaline-di-n-oxides
Delayed release mushroom
nutrients*
Radiopharmaceuticals for
skeletal imaging

notes: Income numbers are in 1996 dollars. 4 Nonbiomedical invention. 44 Includes back royalties paid as a result of
infringement suit.
 THE RESEARCH CORPORATION AND LICENSING

1970s was the prominence within this group of donated inventions. The most 71
important invention in the first decade of the Patent Management Division
(although it was donated to the corporation before the establishment of the di¬
vision) was the method for synthesis of vitamin Eh that we discussed earlier.
The corporations most lucrative invention after the expiration of the vitamin
Bj patents was nystatin, an antifungal antibiotic developed by Rachel Brown
and Elizabeth Hazel of the New York State Department of EEealth that the in¬
ventors had donated to the corporation.15 During the 1960s and early 1970s,
nystatin, two other donated biomedical inventions (Edward Kendall’s cortical
hormones and Robert Woodwards reserpine), and one donated agricultural
invention (the Jones-Mangelsdorf hybrid seed-growing process) were the cor¬
porations most important sources of royalty income.
Although donated inventions had been important to the Research Corpo¬
ration’s revenues since the beginning of its patent management activities, the
extent of the corporation’s reliance on donated inventions had increased by
the mid-1960s. This is apparent even if we restrict our attention to the donated
inventions among the top five gross revenue earners, or the donated “home
runs.” Having accounted for 48 percent of total gross royalty revenues in 1955
and 50 percent in i960, the donated inventions among the top five gross reve¬
nue earners accounted for 69 percent of gross royalties in 1965 and 91 percent
in 1970.16 Only after 1975, when the patents on the last of the donated “home
runs” expired, did the entirety of the Research Corporation’s income depend
on inventions administered through institutional IAAs.

The Cancellation of the MIT Invention

Administration Agreement

The Research Corporations failure to attract economically valuable institu¬

tional inventions came into sharp focus following the 1963 cancellation of the
IAA with MIT. The events that led to MIT’s termination of its agreement with
the Research Corporation merit discussion not only because of their impor¬
tance for the subsequent history of the Research Corporation but also because
they illustrate some of the difficulties of academic patent management.
The immediate catalyst for the series of events that led to the cancellation
was the dispute between the Research Corporation and MIT over the corpo¬
ration’s management of the Forrester patent on magnetic core memories for
computers (see Pugh, 1984).17 In 1951, the Research Corporation filed a patent
application on Jay Forresters invention under the MIT IAA. A similar patent
application was filed by Jan Rajchman at RCA in 1950, and an interference
was declared in 1956. In the meantime, IBM had become interested in using
the core memory technology in its mainframe computers and entered a cross-
THE RESEARCH CORPORATION AND LICENSING

72 licensing agreement with RCA in 1957. IBM also entered into negotiations
with the Research Corporation to negotiate a license should Forrester win the
case. But IBM and the corporation failed to agree on royalty rates, and the ne¬
gotiations collapsed. IBM, which had a long history of donating funds and
equipment to MIT, expressed concern to Institute administrators that the Re¬
search Corporation was taking a hard line in the negotiations. Indeed, Pugh
(1984) reports that IBM may have threatened to withhold future funding to
MIT if the Research Corporation did not compromise.
In i960, the Board of Patent Interferences declared Rajchman victorious on
ten significant claims and awarded several narrower claims to Forrester. The
Research Corporation launched a civil action to attempt to recover the ten
claims awarded to Rajchman and initiated a patent infringement suit against
RCA and IBM.18 By this time, the interests of MIT and the Research Corpo¬
ration seem to have diverged; As Pugh (1984, p. 211) describes it:

As the attorneys prepared for the court case, MIT decided that its interests were not
properly represented by Research Corporation. Research Corporation wanted pat¬
ent fees paid over time, based on use, so that a predictable revenue stream would
permit orderly operation of its business. MIT preferred a single payment to help
cover the costs of a building program. Even more importantly MIT was concerned
about its long-term relationship with IBM and the rest of the computer industry.

In 1962, MIT terminated its IAA with the Research Corporation. Under the
terms of the termination agreement, the Research Corporation received a
$1.6 million cash settlement, and MIT retained all residual rights to the pat¬
ents it had assigned to the Research Corporation.
The Forrester case was a pivotal development in the Research Corpora¬
tions history, since it led to the termination of the IAA that had been the foun¬
dation for the corporations patent management activities. But this case also il¬
lustrates the potential conflicts between the incentives of patent management
agents and those of their academic principals. Similar tensions between the
goals of maximizing licensing income and maintaining or deepening rela¬
tionships with industrial firms (relationships that ultimately may prove to yield
substantially more research funding or income) are not uncommon within
contemporary university technology licensing offices and faculties.19

The Growing Costs of Patent Management

Although the Research Corporation had expanded its operations in the early
postwar period in order to reduce the unit costs of patent management, these
cost savings proved to be elusive. As we noted above, the corporations exper¬
tise in locating licensees and negotiating licensing agreements in one tech¬
nology field appears to have had limited relevance to other; technological
 THE RESEARCH CORPORATION AND LICENSING

areas. Other costs of the corporations operations, such as legal and “visitation” 73
expenses, also grew as its activities and client list expanded. Instead, dis¬
economies of scale and scope increased the Research Corporations operating
costs and reduced net revenues. This section examines several of these sources
of increasing costs, which highlight challenges faced by many contemporary
university technology licensing offices.

Expanded Visitation and Marketing Efforts

Following the cancellation of its agreement with MIT, the Research Corpora¬
tion reorganized its patent management program. Although the corporation
maintained its support for a specialized Patent Management Division that
served multiple institutional clients, the 1963 Annua/ Report argued that the
corporation’s efforts had been “too conservative. In 1963, the corporation ex¬
panded its efforts in patent management and marketing. These intensified
efforts included an increase in staff and greater reliance on technical con¬
sultants (and outside patent counsel) for evaluating and licensing inventions.
The increased reliance on consultants was most pronounced in the mechani¬
cal and electronics fields, where the corporation had developed little exper¬
tise and (as we discuss below) generated little royalty income.20 Figures 4-5
1996 Dollars

figure 4.5 Research Corporation expenditures on consultants and travel, 1953 81

THE RESEARCH CORPORATION AND LICENSING

74 and 4.6 illustrate the rapid growth during 1964-74 in expenditures on

consultants, travel, litigation, and staff.21
As part of the expansion of its marketing and patent management activities
during the 1960s, the Research Corporation had initiated an ambitious visita¬
tion program to foster “a closer working relationship with institutional inven¬
tors and their administrators, more expeditious handling of invention disclo¬
sures submitted for evaluation, and improved understanding of the problems
involved in bringing inventions from the laboratory to the marketplace” (1965
Annual Report, p. 2). The corporation made 81 visits to academic institutions
in 1964, 116 in 1965, and 120 in 1966.22 In 1966, the Research Corporation
announced plans for annual visits to large universities and technical schools
with which it had agreements and biennial visits to smaller schools (1966 An¬
nual Report). These increased travel and visitation activities contributed to
growth in the corporations travel costs during the 1960s. Moreover, the higher
number of invention disclosures produced by expanded visitation activities
(see Figure 4.2) required the hiring of additional staff and consultants, further
increasing costs.23
These increased costs of travel and related activities reflected the impor¬
tance of face-to-face interactions between Research Corporation staff and uni-
1996 Dollars

i igure 4.6 Research Corporation legal and personnel expenditures, 1953—81

 THE RESEARCH CORPORATION AND LICENSING

versity administrators and inventors, a theme that runs through the corpora- 75
tions history. MIT had insisted that the Research Corporation establish a Bos¬
ton office as a condition of the 1937 Invention Management Agreement. Even
after the corporations Boston office was closed in 1946, Research Corporation
staff continued monthly visits to MIT to evaluate and discuss Institute inven¬
tions through the late 1950s. For the Research Corporation (as for contempo¬
rary licensing offices), patenting and licensing transactions required signifi¬
cant investments of time in the development of close relationships with
inventors and their parent institutions, reflecting the substantial “relational”
component of such transactions. Evaluation of inventions, especially those in
their embryonic stages, requires inventor cooperation, guidance, and equip¬
ment. Inventors also are often more knowledgeable about markets for licenses
and may be willing to contribute consulting assistance to transfer important
know-how along with the patent license. Inventor assistance and cooperation
may be needed to write and prosecute patent applications.24
The importance of these relational aspects of technology licensing opera¬
tions helps explain the role of geographic proximity in markets for technol¬
ogy.25 Indeed, the need for frequent interaction between licensing profession¬
als and university inventors has provided one impetus for the decentralization
of technology licensing offices in large, multicampus university systems such
as the University of California.

Training Universities in Patent Management

During the 1970s, the Research Corporation initiated an ambitious training

program to teach university administrators and faculty to evaluate the
patentability and licensing prospects for faculty inventions. 1 he outreach
program” included a newsletter, Research and Invention, which began publi¬
cation in 1971; an instructional pamphlet for university inventors and admin¬
istrators entitled “Evaluating and Patenting Faculty Inventions”; and sponsor¬
ship of conferences and seminars on university-industry technology transfer.
In addition, with funding from the National Science Foundation, the corpo¬
ration began a “Patent Awareness Program” in 1973 that sought to “increase
the ability of faculty members and administrators alike to recognize inventions
developed at educational institutions” (Marcy, 1978, p. 156).
The expanded training program and associated efforts to shift more re¬
sponsibility for invention evaluation and patent management to client uni¬
versities represented a change from the Research Corporation’s previous posi¬
tion that universities should not get involved in patenting.26 The corporation
now argued that the early stages of invention screening and evaluation (that is,
persuading faculty members to disclose valuable inventions and collecting
THE RESEARCH CORPORATION AND LICENSING

76 properly written invention disclosures) were best done by the universities

themselves. This shift in policy may have reflected the inability of the ex¬
panded visitation program to support the close interactions between inventor
and patent agent needed to facilitate the early stages of the patenting and li¬
censing process. But these policy changes also were a response to the Research
Corporations deteriorating financial situation.27
In addition to their costs, the extensive visitation and outreach activities as¬
sociated with the more aggressive patent management activities had another
unanticipated effect. As we noted in Chapter 3, by the 1970s the corporations
activities and other developments in federal policy and the content of univer¬
sity research had led a growing number of U.S. universities to assume a larger
role in patent management. Paradoxically, the efforts of the Research Corpo¬
ration to train a larger group of faculty and administrators in patenting and li¬
censing eroded its competitive advantage in patent administration.

The Paucity of “Home Runs”

The Research Corporations efforts during the postwar period to expand the
number of universities with whom it worked, as well as its more intensive mar¬
keting activities of the 1950s and 1960s, do not appear to have offset the detri¬
mental financial effects of higher operating costs, declining royalties, and the
expiration of key donated patents on home runs. In the face of the imminent
expiration of the last of its lucrative donated patents in 1975, the corporation
noted that

while it is possible that other substantial income producing inventions will become
available in the future, it will be a rare occurrence. Therefore, if the invention ad¬
ministration operation is to continue, total reliance must be put on finding and de¬
veloping commercially viable shared inventions. One consequence of this situation
is that a major effort will be required in the future to develop and enhance rela¬
tionships with universities and faculty inventors. (1974 Annual Report, p. 3)

Unfortunately, this statement overlooked the fact that such efforts had been
under way for more than a decade, with little effect. The “arrival rate” of
highly profitable inventions proved to be unresponsive to the Research Cor¬
poration’s efforts to increase it. This inability to obtain and license additional
“home runs” in the face of growing needs for revenues foreshadows the chal¬
lenges that technology licensing offices and administration at a number of
prominent universities face as patents on key inventions from the early 1980s,
many of which are associated with the rise of biotechnology, begin to expire.
Did the Research Corporation fail to capture more profitable institutional
inventions because its client universities began to “cherry-pick” their faculty
 THE RESEARCH CORPORATION AND LICENSING

inventions, choosing to license their most valuable inventions independently? 77

To examine this possibility, we collected data from the annual reports of the
U.S. Patent and Trademark Office on independent patenting by the corpora¬
tion’s clients in each of six years (1950,1955, i960,1965,1970, and 1975) during
the 1950-75 period.28 These data suggest that there was little cherry-picking in
the first two decades of the period, although this began to change during the
late 1960s and 1970s. Specifically, fewer than 10 percent of institutional clients
received patents (patents assigned to themselves rather than to the Research
Corporation) before 1970. Fully 95 percent of these “independent” patents
were in fields outside of the life sciences, and the apparent decision of its in¬
stitutional clients to bypass the Research Corporation may have reflected the
corporations lack of experience in fields outside of biomedical technology.29
But cherry-picking appears to have expanded during the 1970s. The number
of patents assigned to Research Corporation clients nearly doubled between
1970 and 1975, and much of this growth resulted from increased independent
patenting by Research Corporation institutional clients in the life sciences.
The corporations 1975 Annuel Report noted that many universities were be¬
ginning to withhold valuable inventions.50

The Limits of Scale Economies

The failure to attract highly profitable inventions from its client universities
suggests that the benefits for the Research Corporation resulting from di¬
versification among technologies and institutional clients were limited. De¬
spite increased expenditures on consultants and specialized staff, the corpora¬
tion patented few lucrative inventions. This failure was surprising in view of
the fact that during the 1950s and 1960s, the Research Corporation had access
to the patentable inventions of many of the nations most prominent research
universities.
The effects of slower revenue growth were exacerbated by continued
growth in the corporations total operating costs during the 1960s (Figure 4.7).
Beginning in the mid-1970s, the visitation program and the more aggressive
patent management program were curtailed, and the rate of growth of expen¬
ditures on consultants, legal services, personnel, and travel declined (revealed
in the downward trend after 1975 in consulting, legal, and personnel expenses
in Figures 4.5 and 4.6). . ,
Another source of scale economies that the Research Corporation had
hoped to exploit, dating back to Cottrell’s original vision, was the spreading
of “fixed” costs over a large number of institutions and inventions. Nonethe¬
less its growing client roster failed to prevent the corporation’s total costs per
 $7,000,000 -
1996 Dollars

$0 i i i i i i i i i i i i i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r
rv~s lt\ 0n
LT\ LT\ LT\ LT\
O' O O' O' o O' O' O' O' O' O' O' O' O' O' O'

Fiscal Year

figure 4.7 Research Corporation total operating costs, 1951-81

1996 Dollars

" Total Costs/IAA (left axis) —Total Costs/Disclosure (right axis)

FIGURE 4.8 Total costs per Invention Administration Agreement and disclosure,
1951-81
 THE RESEARCH CORPORATION AND LICENSING

invention agreement and per disclosure from increasing, rather than decreas¬ 79
ing, in the late 1960s (Figure 4.8).
By the mid-1970s, the Research Corporations net royalty income had de¬
clined sharply (Figure 4.9), and a “Committee on Goals and Objectives of the
Corporation” was convened to conduct a strategic review. The committees
report concluded in 1979 that “if in the years 1980-1983 forecasts of enhanced
income do not begin to be realized, significant changes in programs will be
necessary” (Research Corporation, 1979, p. 5). In the event, these “significant
changes” included closing the Patent Management Division.31

FEDERAL POLICY CHANCES, LOSS OF MARKET SHARE,

AND THE DECLINE OF THE RESEARCH CORPORATION

The Research Corporation played an important role in assisting U.S. univer¬

sities in responding to the changes in federal policy toward university patent¬
ing during the 1960s and 1970s that were described in Chapter 3. For example,
the Research Corporation helped universities to bring their policies and pro¬
cedures into conformity with new guidelines for patenting developed by the
Defense Department, the National Science Foundation, and the Department
of Health, Education, and Welfare during the 1970s, in hopes that doing so
1996 Dollars

figure 4.9 Research Corporation net royalty income, 194^

THE RESEARCH CORPORATION AND LICENSING

8o would ultimately increase the demand for its services.32 But the corporations
assistance to U.S. universities in responding to these changes in federal policy
may have hastened its eclipse, as U.S. universities expanded their independent
management of patenting and licensing activities.
In responding to requests for assistance from U.S. universities, the Research
Corporation supported the growth of independent technology transfer opera¬
tions at a number of institutions. By teaching university administrators about
the processes of invention evaluation and administration, the Research Cor¬
poration helped academic institutions develop the capabilities needed to
manage patenting and licensing.33 The corporations 1974 Annual Report
noted that almost every major U.S. university was considering the establish¬
ment of an internal technology transfer office, and the 1975 Annual Report
stated that a growing number of client universities no longer asked the corpo¬
ration to evaluate particularly valuable inventions, choosing instead to man¬
age them internally.34 The potential for significant royalty income in these
fields thus appears to have contributed to growth in cherry-picking by univer¬
sities after the early 1970s, ironically reducing the Research Corporations ac-

figure 4.10 Research Corporation share of U.S. university patents, 1969-87

 THE RESEARCH CORPORATION AND LICENSING

cess to university inventions at the same time that the pool of commercially 8i
valuable academic inventions was expanding.
The passage of the Bayh-Dole Act, which the Research Corporation sup¬
ported, was another important factor in the corporations decline. Growth in
independent university technology licensing operations, which accelerated
after the passage of Bayh-Dole, reduced the Research Corporations share of
the market for academic patents. As the number of universities that entered
into patenting and licensing activities increased (see Chapter 3), the Research
Corporations share of all university patents decreased dramatically, as seen in
Figure 4.10.
In a recent review of its patent management activities, the Research Cor¬
poration noted that “the Bayh-Dole Bill, which the foundation strongly sup¬
ported and testified in favor of, curtailed this pioneering initiative as the na¬
tions colleges and universities developed their own technology transfer offices
and staffs of legal and technical experts” (1997 Annual Report, p. 16). Our
analysis above supports this view but suggests that the internal operating and
financial difficulties that ultimately caused the Research Corporation to cur¬
tail its patent management activities predated the passage of Bayh-Dole.

The Research Corporation After ig8o

The Research Corporation remained active during the 1980s in the face of
Bayh-Dole and the growth of independent university-based technology trans¬
fer operations, but its importance declined. The ability of the corporation to
compete in this arena was further hindered by its status as a charitable foun¬
dation. In the early 1980s, the Internal Revenue Service ruled that the Re¬
search Corporations technology transfer activities were an “unrelated trade or
business.” As a result, the corporation was prohibited from activities it viewed
as essential to technology transfer, such as investing in start-up companies and
early-stage funding of development activities. In 1987* ^*e corporation closed
the Patent Management Division and transferred all of its activities to an in¬
dependent for-profit organization, Research Corporation Technologies
(RCT), whose purpose is “identifying, developing, and commercializing in¬
ventions for universities, colleges, medical research organizations, and other
non-profit laboratories” (1987 Annual Report, p. 20).
Although the RCT has the same mission as the corporation’s former Patent
Management Division, it uses retained earnings to fund technology transfer
activities rather than scientific research. RCT received half of the Research
Corporations endowment (about $45 million in 1996 dollars) at the time of the
spin-off, as well as the Research Corporations patent portfolio and technology
THE RESEARCH CORPORATION AND LICENSING

82 transfer staff. The spin-off agreement stipulated that the Research Corporation
would receive a portion of RCT’s net royalty income until the $45 million
transfer in endowment was paid back. Today, the Research Corporation fo¬
cuses exclusively on research grants, which are funded from the returns on its
endowment, the repayment by the RCT of its endowment grant, and philan¬
thropic contributions.
The RCT s Cooperative Program allows academic institutions to use RCT
to evaluate inventions, similar to the Research Corporations Invention Ad¬
ministration Program. The RCT also operates the Benchmark Program,
which allows universities and other institutions to use RCT services on an “as-
needed” basis. According to Gilles (1991), the Benchmark Program is used pri¬
marily for evaluating the inventions of academic clients, with no commitment
to managing any patents, and management of patenting and licensing for in¬
ventions in fields in which the university technology licensing offices lack ex¬
pertise. Very few universities have used RCT s Cooperative Program, and both
the Cooperative and Benchmark Programs are being phased out.
By some measures, the RCT remains a successful technology transfer or¬
ganization: the Association of University Technology Managers s 1996 AUTM
Licensing Survey (attachment E, p. 53) reports that its gross licensing income
amounted to $70 million, more than any U.S. university in that year. About
$45 million of this amount was distributed to the inventors and universities re¬
sponsible for the inventions (see Research Corporation Technologies, 1996,
“Corporate Update”). Nonetheless, the importance of RCT as a technolog)'
transfer agent has diminished considerably, and its activities differ significantly
from those of the Research Corporation.

CONCLUSION

I his brief history of the Research Corporation underscores the points made in
Chapters 2 and 3 concerning the history of U.S. universities’ patenting activi¬
ties during the twentieth century. I he Research Corporation played an im¬
portant role in the growth of these activities during the middle years of this
century and created a portion of the institutional and administrative infra¬
structure for the post-1980 upsurge in patenting and licensing by individual
U.S. universities. Discussions surrounding the foundation of the Research
Corporation, and much of the debate over its early years of operation, antici¬
pated many of the contemporary comments and debates over the desirability
and effects of university technology licensing. Indeed, Cottrell articulated the
prospect argument (Kitch, 1977) for patenting of university inventions in his
1912 assertion that “some minimum level” of protection for the intellectual
 THE RESEARCH CORPORATION AND LICENSING

property resulting from academic research was necessary in order for industry 83
to invest in its subsequent development and commercialization. But Cottrell
was ambivalent about the effects on academic norms and behavior of locating
the responsibility for management of these activities within the university,
fearing the consequences of conflicts of interest and pressures by university ad¬
ministrators to increase the income flowing from patent licenses, an ambiva¬
lence that is notably absent today.
The Research Corporations history illustrates other features of markets for
intellectual property that manifest themselves in many U.S. universities’ cur¬
rent patenting and licensing operations. Although there appear to be advan¬
tages to specialization in patent and license management that reflect the high
fixed costs, benefits of portfolio diversification, and possibilities for efficiency
gains through learning, the history of the Research Corporation suggests that
these benefits are counterbalanced by significant limitations.
In particular, the requirements for specialized knowledge about individual
technology fields and (of equal importance) about the market for licenses in a
given technology field limit the applicability of know-how acquired from
patenting and licensing in one technology field to other fields. The impor¬
tance of geographic proximity in even arms’ length licensing arrangements
also limits the cost savings attainable through centralization of activities in
a single site or organization. These factors contributed to the declining fi¬
nancial performance of the Research Corporation in the 1970s, and they are
apparent in the operation of many contemporary university technology licens¬
ing offices. Licensing managers tend to focus on specific technology areas,
rather than being generalists. A number of universities’ patenting and licensing
offices have been reorganized so as to decentralize their operations, while at
least a few other universities have maintained independent licensing efforts in
different organizational units, such as their medical and engineering schools.
The study also underscores the limits of “pure” licensing agreements to
support the exploitation or commercialization of inventions. From its earliest
days, the corporation had to invest in extensive applications engineering and
in customization of its precipitation technologies, and it found it necessary to
acquire an extensive portfolio of complementary patents. The Research Cor¬
poration also relied on investments in additional development and other com¬
plements to many of the academic patents that it sought to commercialize,
driving up costs and reducing profits. Contemporary university licensing of¬
fices have also found it necessary to make additional investments, in some
cases taking equity stakes in firms founded to exploit their inventions, in order
to pursue commercial development. Needless to say, this expansion of their ac¬
tivities increases costs and risks.
THE RESEARCH CORPORATION AND LICENSING

84 Finally, the history of the Research Corporation illustrates the difficulty of

managing a patent licensing portfolio solely with a view to maximizing in¬
come, another issue that has bedeviled the operation of many contemporary
university technology licensing offices. Licensing income tends to be domi¬
nated by a small number of “home runs,” often based on biomedical research.
The appearance of such “home runs” is difficult to predict, they typically ac¬
count for a small share of all inventions, and their contribution to licensing in¬
come is limited by the term of their patents. As a result, licensing income is
unstable and difficult to forecast.
The high costs of patent management also limit the financial returns from
patent licensing. Trune and Goslin (1998) estimate that only 41 percent of uni¬
versity technology transfer offices yield a positive net income for their insti¬
tutions, and we believe that their estimate may be high. But many of these
contemporary licensing offices pursue a broader set of goals, including the de¬
velopment of closer links with regional industry for the support of university-
based research. The single-minded pursuit of licensing income above all other
goals led the Research Corporation into a dispute with MIT over the terms
of the license to IBM for MIT’s Forrester memory patent. The dispute among
the Research Corporation, MIT, and IBM over this license vividly illustrates
the potential conflicts between the goals of the licensing agent, the Research
Corporation, and its principal, MIT. Similar conflicts have emerged recently
within U.S. universities that manage their patenting and licensing internally.
The history of the Research Corporation anticipates many contemporary
challenges faced by U.S. universities in their efforts to manage licensing and
research relationships with industry. The ultimate withdrawal of the Research
Corporation from its major role in managing U.S. universities’ intellectual
property reflected shifting federal policies toward philanthropic organiza¬
tions, as well as the Bayh-Dole Act. But this chapter has pointed out that the
seeds of decline had been planted well before 1980 and were apparent in the
growing deficits of the corporation during the 1970s. Indeed, the same factors
that contributed to the withdrawal of the Research Corporation from its dom¬
inant role as a manager of patent licensing for U.S. universities may well pro¬
duce cutbacks in the independent patenting and licensing offices of some of
the U.S. universities that entered into these activities during the 1970s and
1980s. I he university technology licensing operations that survive are likely to
pursue a broader set of goals than royalty income alone.35
 5

A Political History of the

Bayh-Dole Act of 1980

The effects of the Bayh-Dole Act of 1980 on university patenting and licensing
cannot be understood without some consideration of the political origins of
the Act itself. The Bayh-Dole Act is rooted in a debate that extends back to the
late 1940s over the disposition of intellectual property resulting from federally
funded research. But an important factor in the drafting and passage of the Act
in 1980 was the growing role of U.S. universities as patenters and licensors. In¬
deed, U.S. universities themselves played a significant role in the passage of
the Act, which can be viewed in part as a response to increased academic
patenting and licensing activity, even as it provided an impetus to expansion
in these activities.
In the following section, we argue that the debates during the late 1970s
over Bayh-Dole included little evidence that patents and exclusive licenses fa¬
cilitate the transfer and commercialization of federally funded academic re¬
search. Moreover, in contrast to the debates about university patenting earlier
in the century that were discussed in Chapter 3, the Bayh-Dole debates de¬
voted little consideration to potential effects of expanded university patenting
and licensing on “open science” and on academic research more generally.
The overriding concern of Bayh-Dole supporters was increased commer¬
cialization of university inventions, and numerous subsequent journalistic ac¬
counts of the “effects of Bayh-Dole,” as well as a number of analyses from gov¬
ernmental agencies, argue that these goals have been met. As we point out
later in this chapter, however, important questions about the positive and neg¬
ative effects of Bayh-Dole, and about what might have happened in the ab¬
sence of the Act, remain unanswered. Subsequent chapters in this volume ad¬
dress a number of these issues.
POLITICAL HISTORY OF THE BAYH-DOLE ACT

86 Partly because of the widespread journalistic and political portrayal of the

Bayh-Dole Act as a success, a number of other industrial-economy govern¬
ments are considering or have adopted similar policies. The penultimate sec¬
tion of this chapter considers foreign-government “emulation” of the Bayh-
Dole Act. Although it is hardly an isolated example, the selective nature of
foreign governments’ emulation of Bayh-Dole could produce limited or even
counterproductive results. These efforts at emulation overlook the fact that
U.S. university-industry interactions incorporate many channels beyond those
of patenting and licensing and predate Bayh-Dole, as we pointed out in Chap¬
ters 2-4.

A POLITICAL HISTORY OF THE BAYH-DOLE ACT

The Debate over Patent Rights for

Federally Funded Inventions

Congress had debated the issue of ownership of patents resulting from pub¬
licly funded research for decades before the passage of Bayh-Dole, and federal
patent policy was a central point of contention during the debates of the 1940s
over the organization of postwar U.S. science and technology policy. One side
of the debate over patent policy was represented by Senator Harley Kilgore
(D-W.Va.), who argued that the federal government should retain title to pat¬
ents resulting from federally funded research and place them in the public do¬
main (Kevles, 1978). According to Kilgore, allowing private contractors to re¬
tain patents represented a “giveaway” of the fruits of taxpayer-funded research
to large corporations, reinforcing the concentration of technological and eco¬
nomic power. The opposing position was articulated by the director of the
wartime Office of Scientific Research and Development, Vannevar Bush, who
argued that allowing contractors to retain patent rights would preserve their in¬
centives to participate in federal R&D projects and to develop commercially
useful products based on government-funded research.1
The postwar debate highlighted the central questions in government pat¬
ent policy for the next three decades. Supporters of the retention of intellec¬
tual property rights by government agencies argued that allowing contractors,
rather than government agencies, to retain title to patents resulting from fed¬
erally funded research favored large firms at the expense of small business.
Moreover, they asserted, such a policy would raise prices for the fruits of
taxpayer-funded research. Supporters of allowing contractors to retain title to
patents resulting from federally funded research argued that failure to do so
would make it difficult to attract qualified firms to perform government re¬
search and that the absence of clear title to intellectual property resulting from
 POLITICAL HISTORY OF THE BAYH-DOLE ACT

such research would reduce incentives to invest in commercial development 87

of these inventions.
Another contentious issue in these debates about government patent pol¬
icy was the desirability of a uniform patent policy across all federal agencies.
Each of the major federal R&D funding agencies had established its own pat¬
ent policy following World War II, and the resulting mix of agency-specific
policies created ambiguities and uncertainties for contractors and for govern¬
ment employees.2 Despite numerous congressional hearings on this issue, no
legislation was adopted during the 1950-75 period because of the inability of
supporters of the opposing positions outlined above to resolve their differ¬
ences. I his legislative deadlock was reinforced by statements on federal agen¬
cies’ patent policies issued by Presidents Kennedy and Nixon in 1963 and 1971,
respectively. Both presidents’ statements asserted that agency-specific differ¬
ences in patent policy were appropriate and reflected differences in their mis¬
sions and R&D programs.3

Government Patent Policy and University Research

The debates over federal patent policy largely ignored U.S. universities during
the 1940s and 1950s. After all, U.S. universities have.never been responsible for
performance of more than 34 percent of federally funded R&D (excluding
university-based Federally Funded Research and Development Centers
[FFRDCs]) during the postwar period and their share of federally funded
R&D exceeded 20 percent only after 1991.4 Moreover, as we noted in Chap¬
ter 3, U.S. universities historically had limited their direct involvement in
patenting and licensing activities.
Federal policy toward patents resulting from publicly funded university re¬
search became a topic of debate after the release in 1968 of reports on the Na¬
tional Institutes of Health’s (NIH) Medicinal Chemistry program by the U.S.
General Accounting Office (GAO) (1968) and by Harbridge House (1968a), a
consulting firm that the Federal Council for Science and Technology (FCST)
commissioned to conduct a study on government patent policy as part of a re¬
view of this issue by the FCST itself.5 Both reports examined the effects of fed¬
eral patent policy on research collaboration between U.S. pharmaceutical
firms and academic researchers in medicinal chemistry. During the 1940s and
1950s, these pharmaceutical firms had routinely screened compounds de¬
veloped by NIH-funded university researchers at no charge.6 In some cases
(depending on the patent policies of particular universities), these pharma¬
ceutical firms received exclusive rights to develop and market the compounds.
In 1962, the Department of Health, Education, and Welfare (HEW) notified
POLITICAL HISTORY OF THE BAYH-DOLE ACT

88 universities that firms screening compounds must sign formal patent agree¬
ments that prevented the firms from obtaining patents on technologies result¬
ing from NIH funding (GAO, 1968, p. 10). Indeed, the HEW policy further
stated that firms could not obtain patents on inventions in the “field of re¬
search work” supported by NIH funds at a given institution.
The GAO and Harbridge House reports criticized HEW’s patent policy,
arguing that pharmaceutical firms had stopped screening NIH grantees’ com¬
pounds because of the firms’ concerns that the HEW policies could compro¬
mise their rights to intellectual property resulting from their in-house research
(Harbridge House, 1968a, p. II-21; GAO, 1968, p. 11). Both reports recom¬
mended that HEW change its patent policy to clarify the circumstances un¬
der which rights reverted to the government and those under which universi¬
ties could retain title to patents and issue exclusive licenses to firms.
HEW responded to these critical reports in 1968 by establishing Institu¬
tional Patent Agreements (IPAs) that gave universities with “approved tech¬
nology transfer capability” the right to retain title to agency-funded patents.
In addition, the agency began to act more quickly on requests from universi¬
ties and other research performers for title to the intellectual property result¬
ing from federally funded research. Between 1969 and 1974, HEW approved
90 percent of petitions for title, and between 1969 and 1977 the agency granted
IPAs to seventy-two universities and nonprofit institutions (Weissman, 1989).
The National Science Foundation (NSF) instituted a similar IPA program in
1973, and the Department of Defense (DOD) began in the mid-1960s to allow
universities with approved patent policies to retain title to inventions resulting
from federally funded research.
By the beginning of the 1970s, U.S. universities thus were able to patent the
results of federally funded research via agency-specific IPAs or similar pro¬
grams at the DOD as well as through case-by-case petitions. Nevertheless, in
the late 1970s U.S. universities became concerned that HEW might limit their
ability to negotiate exclusive licenses in the department’s IPA programs. These
concerns, along with growing dissatisfaction within Congress and the indus¬
trial community over the lack of uniformity in patent rights to inventions re¬
sulting from federally funded research, provided the immediate impetus for
the introduction in 1978 of the bill that eventually became the Bayh-Dole Act.

The Bayh-Dole Bill

As we noted above, HEW’s 1968 revisions in its patent policy expanded the
number of IPAs negotiated with universities and other research institutions
and allowed more universities to obtain patents on NIH-funded inventions on
a case-by-case basis. U.S. universities had become more active in patenting
 POLITICAL HISTORY OF THE BAYH-DOLE ACT

and licensing faculty inventions during the 1970s, especially in the biomedical 89
field (see Chapters 3 and 4), and therefore welcomed these changes in HEW
patent policies.
But these patent policies triggered considerable debate within HEW. In
August 1977, HEWs Office of the General Counsel expressed concern that
university patents and licenses, particularly exclusive licenses, could con¬
tribute to higher health-care costs (Eskridge, 1978). The department ordered
a review of its patent policy, including a reconsideration of whether universi¬
ties’ rights to negotiate exclusive licenses should be curtailed.8 During the en¬
suing twelve-month review by HEW of its patent policies, the agency deferred
decisions on thirty petitions for patent rights and three requests for IPAs.
Thus, although Bayh-Dole is often regarded as being the catalyst for the
growth of university patenting and licensing in the United States, in fact U.S.
universities that were active in patenting and licensing, as well as other uni¬
versities, lobbied for the introduction and passage of the bill. Broad (1979a,
p. 476) notes that in response to HEWs review of its patent policies, “[univer¬
sities got upset and complained to Congress.” And Heaton et al. (2000) note
that a patent attorney from Purdue University and a congressional staffer who
previously worked at the University of Arizona, both of which sought more lib¬
eral policies toward patenting publicly funded research, respectively asked
Senators Birch Bayh (D-Ind.) and Robert Dole (R-Kans.) to introduce the bill.
These and other university licensing officials also aided in drafting portions of
what became the Bayh-Dole Act.9
In September 1978, Senator Dole held a press conference where he criti¬
cized HEW for “stonewalling” university patenting (commenting, “rarely
have we witnessed a more hideous example of overmanagement by the bu¬
reaucracy”) and announced his intention to introduce a bill to remedy the sit¬
uation (Eskridge, 1978, p. 605). On September 13, 1978, Senators Bayh and
Dole introduced S. 414, the University and Small Business Patent Act. The Act
proposed a uniform federal patent policy that gave universities and small busi¬
nesses rights to any patents resulting from government-funded research.10 The
bill lacked provisions that had been included in most IPAs, including the re¬
quirement that a participating university must have an “approved technology
transfer” capability. In contrast to the language of some IPAs between univer¬
sities and HEW, the bill imposed no restrictions on the negotiation by uni¬
versities and other research institutions of exclusive licensing agreements.11
As we noted earlier, many members of Congress had long opposed any
federal grant of ownership of patents to research performers or contrac¬
tors (Broad, 1979b). The Bayh-Dole bill nevertheless attracted little opposi¬
tion. The bills focus on securing patent rights for only universities and small
POLITICAL HISTORY OF THE BAYH-DOLE ACT

90 business weakened the argument (a la Kilgore) that such patent-ownership

policies would favor big business.12 The bills introduction in the midst of de¬
bates over U.S. economic competitiveness also proved crucial to its passage.
An article in Science discussing the debate on the Bayh-Dole bill observed:
The critics of such legislation, who in the past have railed about the “giveaway of
public funds” have grown unusually quiet. The reason seems clear. Industrial in¬
novation has become a buzzword in bureaucratic circles . . . the patent transfer
people have latched onto this issue. It’s about time, they say, to cut the red tape that
saps the incentive to be inventive. (Broad 1979b, p. 479) .

A number of universities, including Harvard University, Stanford University,

the University of California, and MIT, lobbied for passage of the bill and
throughout the debates were active in “commenting and helping to develop
the final language” of the House and Senate versions of the bill (Barrett, 1980).
Not surprisingly, witnesses from active institutional patenters (including Stan¬
ford, Purdue, and Wisconsin) testified in support of the bill, as did represen¬
tatives from various university associations (including the American Council
on Education, the Society for University Patent Administrators, and the Na¬
tional Association of College and University Business Officers) and the Re¬
search Corporation. The support of these groups was supplemented by posi¬
tive statements from witnesses representing various small businesses and small
business trade groups, like the National Small Business Association, the Small
Business Legislative Council, and the American Society of Inventors.
Considerable testimony and commentary during these hearings focused
on lagging U.S. productivity growth and innovativeness, suggesting that gov¬
ernment patent policy contributed to these woes.15 In their opening state¬
ments in the Senate Judiciary Committee hearings on the bill, Senators Bayh
and Dole each pointed to two problems with federal patent policy as of 1979:
(1) the “policy” in fact consisted of more than twenty different agency-specific
patent policies; and (2) most federal agencies made it difficult for contractors
to retain title to patents.
Witnesses supporting the Bayh-Dole bill cited another section of the
1968 Harbridge House report (1968b) in claiming that rates of utilization of
government-funded patents were higher when contractors rather than agen¬
cies held title to these patents.14 Another frequently cited statistic was drawn
from the FCSTs 1978 Report on Government Patent Policy (FCST, 1978),
which concluded that fewer than 5 percent of the 28,000 patents owned by the
federal government as of 1976 were licensed.15 Legislators and witnesses used
this finding to argue that giving patent rights to contractors would create in¬
centives for development and commercialization that were lacking under the
current system.
 POLITICAL HISTORY OF THE BAYH-DOLE ACT

As Eisenberg (1996) has pointed out, however, the data in the Harbridge 91
House and FCST reports did not support this conclusion. The patents exam¬
ined in the Harbridge and FCST studies were drawn primarily from research
funded by the DOD (83 percent of the patents from the Harbridge House
sample and 63 percent of those from the FCST sample), an agency that read¬
ily granted patent rights to research performers. Since these defense contrac¬
tors could and often did seek the rights to patents that they believed might re¬
sult in profitable innovations, it is likely that the patents for which they did not
seek ownership rights had limited commercial potential. It therefore is not sur¬
prising that these patents were not commercialized.
The data in these reports also were based primarily on patents resulting
from government-funded R&D carried out by private firms. As such, they
shed little light on the importance for commercial exploitation of patenting
by universities of inventions funded by the federal government. It is likely, for
example, that the characteristics of federally funded university inventions
differ from those associated with federally funded inventions patented by con¬
tractors, reflecting the fact (among other things) that federally funded aca¬
demic research typically includes more fundamental research and less de¬
velopment activity than do research contracts with industry.
In addition to citing this statistical evidence, supporters of the bill argued
that the characteristics of university inventions made patents and exclusive li¬
censes important for the commercialization of these inventions. According to
these witnesses, most university inventions are embryonic when first disclosed,
requiring significant additional development before they can be commer¬
cially useful. In this view, firms would not invest in these costly development
activities without clear rights to the relevant intellectual property, something
that in many instances would require an exclusive license.16 Other witnesses
suggested that giving title to universities would create incentives for inventors
and institutions to become actively involved in the development and com¬
mercialization of embryonic inventions, anticipating arguments developed
more formally by Jensen and Thursby (2001).
The Bayh-Dole Act was passed overwhelmingly in both the House and the
Senate in the winter of 1980 with minimal floor debate, and President Jimmy
Carter signed the Act into law in December 1980.18 The Acts provisions facil¬
itated university patenting and licensing in at least two ways. F irst, they re¬
placed the web of IPAs and case-by-case petitions with a uniform policy. Sec¬
ond, the Act expressed congressional support for the negotiation of exclusive
licenses between universities and industrial firms for the results of federally
funded research.
The broad political support for university patenting of intellectual
POLITICAL HISTORY OF THE BAYH-DOLE ACT

92 property, especially that resulting from federally funded research, that was ex¬
pressed in hearings and floor debate on the Bayh-Dole Act contrasts with the
widespread expressions of concern by university administrators and others in
the debates of the 1930s over university patenting and licensing that were dis¬
cussed in Chapter 3. For example, none of the witnesses in these hearings dis¬
cussed the potential risks created by university patenting and licensing for the
“disclosure” and other norms of academic science, nor were any potentially
detrimental effects of patenting and licensing for other channels of university-
industry technology transfer considered.19 Nor did the hearings and floor de¬
bate over Bayh-Dole discuss the political risks for universities of direct in¬
volvement in the management of patenting and licensing that loomed larger
in the debates of the 1930s and 1940s.
In many respects, the central justifications for the Bayh-Dole Act elevated
to the national level the “taxpayer benefit” arguments that were widely used
during the 1920s and 1930s to justify state universities’ involvement in patent¬
ing and licensing. These arguments were especially salient in the economic
environment of the late 1970s, when U.S. global competitiveness (or the lack
of same) had become an important political issue in domestic debate. In a
global economy in which scientific and technological knowledge moved
across national boundaries and could be exploited more easily by non-U.S. en¬
terprises than in previous decades, supporters of the Bayh-Dole Act argued
that U.S. taxpayers would benefit from federal support of academic R&D only
if the results of this research were commercialized, an outcome widely be¬
lieved to be more likely if these results were patented.
1 his argument repeated the justifications of the prewar period for the wide¬
spread involvement of public universities in patenting and licensing —patent¬
ing was essential to the capture by state taxpayers of the benefits of their uni¬
versities’ research activities. But the majority of the inventions whose
licensure was debated during the 1930s had been funded by industrial or state
government sources rather than by the federal government. The invocation by
supporters of Bayh-Dole of these justifications at the national level reflected
the rise to dominance of federal agencies as sources of academic research sup¬
port during the postwar period and implied that both private and public uni¬
versities faced a similar “public service” obligation to patent and license fac¬
ulty inventions in the wake of Bayh-Dole.

The Bayh-Dole Act

The Bayh-Dole Act became effective on July 1, 1981, creating a uniform fed¬
eral patent policy for universities and small businesses that gave them the
rights to any patents resulting from grants or contracts funded by any federal
 POLITICAL HISTORY OF THE BAYH-DOLE ACT

agency. The federal government retained a nonexclusive royalty-free license 93

to any such patents and retained “march-in” rights to compel licensing or to
utilize the invention itself when contractors’ licensing policies failed to pro¬
mote utilization or where doing so is necessary for public health or safety. The
Act and subsequent implementing regulations also incorporated policies gov¬
erning the timetable for disclosures of inventions to the funding agency and
for filing patent applications.20 These regulations required that universities
share any licensing royalties with inventors and mandated a preference for
small businesses in the award of licenses by universities and other research
performers. The Act also limited the duration of any exclusive licenses that
universities could negotiate with large businesses.
The Acts limitations on the rights of large firms to obtain title to patents re¬
sulting from federally funded research were subsequently nullified. In 1983,
President Ronald Reagan circulated an Executive Memorandum to agencies
instructing them to allow large businesses as well as universities and small
businesses to retain title to federally funded patents, and he issued an Execu¬
tive Order to this effect in 1987.21 In 1984, an amendment to Bayh-Dole re¬
moved the time limits on the length of exclusive licenses universities could of¬
fer to large businesses.22
The Bayh-Dole Act was one part of a broader shift in U.S. policy toward
stronger intellectual property rights during the 1980s.23 Among the most im¬
portant of these policy initiatives was the establishment of the Court of Ap¬
peals for the Federal Circuit (CAFC) in 1982. Established to serve as the court
of final appeal for patent cases throughout the federal judiciary, the CAFC
soon emerged as a strong champion of patentholder rights. Even before the es¬
tablishment of the CAFC, however, an important U.S. Supreme Court deci¬
sion in 1980, Diamond v. Chakrabarty, upheld the validity of a broad patent in
the new industry of biotechnology, opening the door to patenting the organ¬
isms, molecules, and research techniques emerging from biotechnology. The
origins and effects of Bayh-Dole must be viewed in the context of this larger
shift in U.S. policy toward intellectual property rights, and the effects of
the Act are easily confounded with those of these other intellectual property
initiatives.

THE EFFECTS OF BAYH-DOLE

During the late 1990s and early twenty-first century, many commentators and
policymakers portrayed the Bayh-Dole Act as the critical catalyst to growth in
U.S. universities’ innovative and economic contributions. Indeed, the Orga¬
nization for Economic Cooperation and Development (OECD) went so far
as to argue that the Bayh-Dole Act was an important factor in the remarkable
POLITICAL HISTORY OF THE BAYH-DOLE ACT

94 growth of incomes, employment, and productivity in the U.S. economy of the

late 1990s.24 Implicit in many if not all of these characterizations is the argu¬
ment that university patenting and licensing were indispensable to these as¬
serted increases in the economic contributions of U.S. university research. For
example, a recent article in the Economist opined that:
possibly the most inspired piece of legislation to be enacted in America over the past
half-century was the Bayh-Dole Act of 1980. Together with amendments in 1984 and
augmentation in 1986, this unlocked all the inventions and discoveries that had
been made in laboratories throughout the United States with the help of taxpayers’
money. More than anything, this single policy measure helped to reverse America’s
precipitous slide into industrial irrelevance. Before Bayh-Dole, the fruits of research
supported by government agencies had gone strictly to the federal government. No¬
body could exploit such research without tedious negotiations with a federal agency
concerned. Worse, companies found it nigh impossible to acquire exclusive rights
to a government owned patent. And without that, few firms were willing to invest
millions more of their own money to turn a basic research idea into a marketable
product.

But after Bayh-Dole,

Overnight universities across America became hotbeds of innovation, as entrepre¬

neurial professors took their inventions (and graduate students) off campus to set up
companies of their own. Since 1980, American universities have witnessed a ten¬
fold increase in patents they generate. (“Innovations Golden Goose,” p. T3)

Similar characterizations of the effects of the Bayh-Dole Act have been artic¬
ulated by the president of the Association of American Universities,25 the com¬
missioner of the U.S. Patent and Trademark Office,26 and the Technology Re¬
view, edited and published at MIT.27
Remarkably, none of these characterizations of the positive effects of the
Bayh-Dole Act cite any rigorous evidence in support of their claims. This
cavalier treatment of evidence, of course, is not unprecedented in the po¬
litical history of the Act. We noted above that the legislative debates over the
bill’s drafting and passage relied on little if any evidence that patents and li¬
censes are necessary for the commercialization and development of university
technologies.
Scholarly research on the effects of Bayh-Dole has been slower to appear,
reflecting the need for evidence to support such empirical research. We dis¬
cuss this research in greater detail in subsequent chapters as part of our em¬
pirical investigation of U.S. university patenting and licensing before and af¬
ter 1980. Here, we simply highlight the central findings and unresolved issues
raised by this research.
There is little evidence of substantial shifts since Bayh-Dole in the content
 POLITICAL HISTORY OF THE BAYH-DOLE ACT

of academic research, a finding that is not surprising in view of the blurry 1 ines 95
between “basic” and “applied” research in the biomedical sciences that have
been the focus of most of the increased university patenting and licensing
since the 1970s. But current research, which remains quite preliminary in its
findings on this issue, provides mixed support for the argument that patenting
and licensing are necessary for the transfer and commercial development of
university inventions. Other scholarly work (for example, Louis et ah, 2001)
suggests that university patenting and licensing have negatively affected “dis¬
closure norms” of academic research in specific fields, leading to higher lev¬
els of secrecy and less sharing among researchers of early results, but more re¬
search on this issue is needed. Finally, the effects on the research enterprise
itself of any increased assertion by institutional and individual inventors of
property rights over inputs to scientific research have only begun to receive se¬
rious scholarly attention.
Journalistic and policy-oriented accounts of the Bayh-Dole Act thus pre¬
sent little evidence in support of their claims, while scholarly research on the
Act’s effects by other scholars and ourselves reaches a much more tentative and
guarded verdict on the Bayh-Dole Act. Nevertheless, the limited evidence on
the Acts effects (both positive and negative) has not prevented a number of
other OECD governments from pursuing policies that closely resemble the
Bayh-Dole Act.

INTERNATIONAL “EMULATION” OF THE BAYH-DOLE ACT

A recent OECD survey reports “a general trend across OECD countries to

emulate the Bayh-Dole patent legislation that allows performers of govern¬
ment research, including universities and small businesses, to patent and li¬
cense inventions” (OECD 2002, p. 10). These initiatives are based on the be¬
lief that Bayh-Dole was essential to increased university-industry interaction
and technology transfer in science-based industries in the United States. But
as the review above suggests, these attempts are based on a misreading of the
limited evidence concerning the effects of Bayh-Dole and on a misunder¬
standing of the factors that have encouraged the long-standing and relatively
close relationship between U.S. universities and industrial innovation.
International interest in and emulation of the Bayh-Dole Act are hardly
unique to this policy initiative. Such emulation has been especially wide¬
spread in the field of technology policy, most notably in the area of collabora¬
tive R&D policies. Research collaboration was cited by U.S. and European
policymakers during the 197°$ und 1980s as a key policy underpinning Japans
rapid technological advance. Accordingly, both the European Union and the
POLITICAL HISTORY OF THE BAYH-DOLE ACT

96 U.S. during the 1980s implemented policies and programs to encourage

such collaboration, with mixed results. One of the best-known examples
of such R&D collaboration is the SEMATECH (Semiconductor Manufac¬
turing Technology) R&D consortium established in Austin, Texas, in 1987
with public and private funding. In response to the perceived success of the
SEMATECH collaboration, Japanese managers and policymakers initiated
publicly and privately funded research consortia (ASET and SELETE) in the
late 1990s. Japan, which initially provided the model for emulation by the
United States and the European Union, now is emulating the programs that
allegedly were initially based on Japanese programs.
International policy emulation of this sort is characterized by two key fea¬
tures: (1) the “learning” that underpins the emulation is highly selective; and
(2) the implementation of program designs based on even this selective learn¬
ing is affected by the different institutional landscape of the emulator. Both of
these characteristics of international emulation are readily apparent in the
case of SEMATECH. They are even clearer in the international emulation of
the Bayh-Dole Act that began during the 1990s.
The policy initiatives that have been debated or implemented in most
OECD economies have either sought to shift ownership of the intellectual
property rights for inventions resulting from publicly funded research from
the government funding agency or inventor to the entity performing the re¬
search, or in some cases from the government to the academic inventor. In
university systems such as those of Germany or Sweden, researchers have long
had ownership rights for the intellectual property resulting from their work,
and debate has centered on the feasibility and advisability of shifting these
ownership rights from the individual to the institution. In Italy, legislation
adopted in 2001 shifted ownership from universities to individual researchers.
In Japanese universities, ownership of intellectual property rights resulting
from publicly funded research is determined by a committee, which on occa¬
sion awards title to the researcher. No single national policy governs the own¬
ership of intellectual property rights within the British or Canadian university
systems, although efforts are under way in both nations to grant ownership to
the academic institution rather than to the individual researcher or the fund¬
ing agency. In addition, the Swedish, German, and Japanese governments
have encouraged the formation of external technology licensing organiza¬
tions, which may or may not be affiliated with a given university.
These foreign-government emulation initiatives differ from one another
and from Bayh-Dole itself. The policy proposals and initiatives display the
classic signs of international emulation described above — selective borrowing
from another nations policies for implementation in an institutional context
that differs significantly from that of the nation being emulated.
 POLITICAL HISTORY OF THE BAYH-DOLE ACT

But patenting and licensing were only one of many channels through 97
which U.S. universities contributed to industrial innovation throughout
the twentieth century, and the surveys of industrial managers discussed in
Chapter 2 suggest that these channels are not the most important ones in most
technological fields. Inasmuch as patenting and licensing are of secondary
importance in most fields, emulation of the Bayh-Dole Act is insufficient and
perhaps even unnecessary to stimulate higher levels of university-industry in¬
teraction and technology transfer. Instead, reforms to enhance interinstitu-
tional competition and autonomy within national university systems, as well
as support for the external institutional contributors to new-firm formation
and technology commercialization, such as venture capital funding, labor
mobility, and other important catalysts for technology commercialization, ap¬
pear to be more important.
Moreover, the impact of Bayh-Dole on the other channels of knowledge
and technology transfer remains unclear, as we have argued above. There are
potential risks to the university research enterprise that accompany increased
involvement by university administrators and faculty in technology licensing
and commercialization, and uncritical emulation of Bayh-Dole in a very dif¬
ferent institutional context could intensify these risks.

CONCLUSION

The Bayh-Dole Act of 1980 was an effect as well as a cause of increased uni¬
versity patenting and licensing in the United States. The debates leading to
passage of the Act included little if any of the substance of previous debates
among university administrators over the desirability of direct institutional in¬
volvement in patenting and licensing. Nor did parties to the debate over the
Bayh-Dole Act draw on the lengthy controversy within the U.S. Congress over
the appropriate policy toward ownership of the intellectual property rights cre¬
ated in federally funded research programs. The relatively noncontroversial
character of the bill and the limited debate surrounding it both seem to have
reflected the broad bipartisan concern within the Congress and executive
branch in the late 1970s over “international competitiveness,” as well as the ris¬
ing tide of faith in the role of strong intellectual property rights to strengthen
the United States in international competition.
A remarkable feature of the debate over Bayh-Dole and much of the cele¬
bratory “analysis” of its effects in the popular press and elsewhere is the lack
of hard evidence to support the claims of proponents and opponents. This
evidentiary vacuum, however, has not prevented the widespread adoption
or serious consideration of similar policies by other industrial-economy
governments.
POLITICAL HISTORY OF THE BAYH-DOLE ACT

98 In the following chapters, we analyze the effects of the Bayh-Dole Act, ar¬
guing that much of what has occurred since 1980 in U.S. university patenting
and licensing might well have occurred without the Act. The surge in univer¬
sity patenting and licensing is as much a result of the long-standing relation¬
ships between U.S. universities and industry, the broader shift toward stronger
intellectual property rights during the 1980s in the United States, and the
transformation of biomedical science as it is a result of the Bayh-Dole Act.
Moreover, the Acts proponents have exaggerated the necessity of patenting
and licensing to support such interactions and technology transfer, even as
they have overlooked potential risks for U.S. universities and for the broader
economy flowing from increased patenting of university research results.
 6

The Bayh-Dole Act and Patenting and

Licensing at the University of

California, Stanford University,

and Columbia University

Although the Bayh-Dole Act of 1980 is widely cited as the central factor in the
growth of U.S. academic patenting and licensing during the 1980s, its effects
on U.S. research universities and on the U.S. innovation system have been the
focus of little empirical analysis (Henderson, Jaffe, and Trajtenberg 1998a,b,
are important exceptions). In this chapter, we examine the patenting and li¬
censing activities of three academic institutions that were the leading recipi¬
ents of licensing and royalty income for much of the 1990s: the University of
California (UC), Stanford University, and Columbia University.1 A combined
analysis of data from these three universities allows us to consider the effects
of these new federal policies on universities such as Columbia, which became
large-scale patenters and licensers only after 1980, and universities that were
active in patenting and licensing well before 1980, such as the University of
California and Stanford. Chapter 7 complements this analysis of individual in¬
stitutions with an examination of the characteristics of the patents issued to all
U.S. universities before and after 1980.

THE EFFECTS OF THE BAYH-DOLE ACT: EVIDENCE

FROM THE UNIVERSITY OF CALIFORNIA, STANFORD

UNIVERSITY, AND COLUMBIA UNIVERSITY

The University of California

Unlike Columbia, the University of California established policies requiring
faculty disclosure of potentially commercially useful research results long be¬
fore Bayh-Dole. Mechanisms for supporting the commercial exploitation of
any resulting patents were put in place in 1943, and assignment by faculty of
PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

too their inventions to the university was determined on a case-by-case basis.

Patenting and licensing were the responsibility of the UC General Counsel’s
office, which oversaw the creation and gradual growth of the UC Patent Of¬
fice. The UC Board of Regents established the University Patent Fund in 1952
to invest the earnings from university-owned inventions in the UC systems
General Endowment Pool: earnings from the fund also supported the ex¬
penses of UC patenting activities and faculty research.2 In 1963, the UC Board
of Regents adopted a policy stating that all “[m]embers of the faculties and em¬
ployees shall make appropriate reports of any inventions and licenses they
have conceived or developed to the Board of Patents.”3
In 1976, the Patent Office was reorganized into the Patent, Trademark, and
Copyright Office (PTCO). Only in 1980, however, was the PTCO staffed with
experts in patent law and licensing, as part of a broader expansion in UC
patenting and licensing activities. The Board of Patents was abolished in 1985,
and new policies allowing for sharing by campuses in patent licensing reve¬
nues were adopted by the Office of the President and the campus chancellors
in 1986. Staff employment in the PTCO grew from four in 1977-78 to forty-
three in 1989-90, and the PTCO was renamed the Office of Technology
Transfer (OTT) in 1991. In 1990, however, UC Berkeley and UCLA estab¬
lished independent patenting and licensing offices, relying on the systemwide
Office of Technology Transfer selectively for expertise in patent and licensing
regulations. By 2002, six UC campuses (UC Berkeley, UCLA, UC San Diego,
UC Irvine, UC Davis, and UC San Francisco) had established independent
licensing offices.4

FIGURE 6.1 University of California invention disclosures, 1975—90

 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

figure 6.2 University of California patents by year of issue, 1975-95

Since the University of California was active in patenting and licensing

well before the passage of the Bayh-Dole Act, a comparison of the 1975-79 pe¬
riod (prior to Bayh-Dole) and 1984-88, following the passage of the bill, pro¬
vides a “before and after” test of the Acts effects. The average annual number
of invention disclosures during 1984-88 is almost 237, substantially above
their average level (140 annual disclosures) for the 1975-79 period. Similarly,
the average annual number of patents issued to the University of California in
the earlier period is 22, compared to 58 for 1984-88. The period following the
Bayh-Dole Act thus is associated with a higher average level of annual inven¬
tion disclosures and patents (confirmed in Figures 6.1 and 6.2); but the timing
of the increase in annual disclosures suggests that more than the Bayh-Dole
Act affected this shift.
The increase in the average annual number of invention disclosures in Fig¬
ure 6.1 predates the passage of the Bayh-Dole Act; indeed, the largest single
year-to-year percentage increase in disclosures during the entire 1974-88 pe¬
riod occurred in 1978—79, before the Act s passage. This increase in disclosures
may reflect the important advances in biotechnology that occurred at UC San
Francisco during the 1970s or other changes in the structure and activities of
the UC patent licensing office that were unrelated to Bayh-Dole. For example,
the Cohen-Boyer DNA splicing technique, the basis for the license that ac¬
counted for more licensing revenues than any other invention at either the
UC system or Stanford University, was disclosed in 1974, and the first of sev¬
eral patent applications for the invention was filed in 197^’ before Bayh-
Dole (this patent issued in 1980).
 32 681.0
00 Tf- '-j
LTV
LA .—I VO O' r-i la LTV

94
O o O' r-

91
S NrHVO o o -H VO 00 O' o
0s rsl ~-* O
£ pa" rsf
NO ^f
fc O'" pa"
pa co 't-
PA
LA,
ON
I
o
r\
On

6,950.5
o NO OO
o
00 o
o NO
Tf r-i ltv O' o o no o

87
92
O'
£ O' la O' la O' *—1 pa
'•t*
00 O'
NO
PA
°° •—^ rN|_ 10

S E £ no" P'f £
o
o
jg

OJD
•£

584.7
la pa >—1 LA
£5 00 OO
rA LA [v. o

99
o

81
$G O' O' O' O' O'
d> Tf rt- NO fA. VO NO
o r-J © PA VO
E ^ rsT E ltv rrC
><
£"

o NO CO o go r- o
e 00
O' O' p'i ■—1 NO
00
O' co coNO o
co
ID N co ia
CO <
LTV ^3" CO N 00 ^t-
O'

E rsf —■T E
-o

o
O LA NO p LA P LA
"gj f'- LA CO CO r- r-
O' O' o O' O' P'i O' r- O'
c o ■—1 CA NO 00
gj NO —^ O' NO
vd £ E E
w
j
CQ d>
< •d
H
’£ o p p o
r- r-»
O' LA Ovj O' r- O'
1- OO I\ PA LA
Osj O'
£ E
e
-S
00

•S'
Gross income from top 5 earners (1996 dollars: 000s)

'£
c
Share of income of top 5 earners associated with

vg
Share of gross income from top 5 earners (%)

,-~G GO JsJJ
GS
_E5
g^'-o<u GO
d>
O JZ3 v> -*-•
O 6 -2 o c .a
T)
vo
r O
®
T3 G
t
0
C
NO
£> O' 2 8 o^ 03
« Vi
g
O' 0 OJ 0J ^
O' W 03
•-I LA co p-^LA v,
d>
.> c/j
w, o
SC (D
G x-v G O c o
S d> 4“* • ■
biomedical inventions (%)

o o
D s^
EB. £ ^
d> : d> '—
1 ^ Vi
E E o

£
c ^ G
S s
o
^ i> g\2 Oh O ^ <u S-.2
S S'-g O o
*s C B o o c^ I d>
T3
NO
I"! ■? ° ^3
D NO
Q d> c d> G O'
r- °
<u c O'

ouES03
O'
^3
O 03
E^ O'
«S O CO
O O 0
o g C B d) d)
d> toe .S _c
——1 E E hJO ^3
d> G-. G_ <u u- 3 4h d»
0 0
00 c
o
O C
<l> <u
0 0
S
<U
0 0 O 0
E
1— .2 Lh So _G c (U d) 0
1-4 V-
03 15 cd 03 03 03 IS
3 Vi co
JG JG JG Vi CA _G JG
CO CO 00 0 C CO OO o
Lh 1— t-M
o O O o
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

103

FIGURE 6.3 University of California biomedical disclosures as a percentage

of total disclosures, 1975-90 (3-year moving average)

Since biomedical inventions account for the lions share of UC patenting

and licensing after 1980, our assessment of trends before and after Bayh-Dole
focuses on biomedical inventions, patents, and licenses. Figure 6.3 reveals that
the shares of biomedical inventions within all UC invention disclosures began
to grow in the mid-1970s, before the passage of Bayh-Dole. Moreover, these
biomedical inventions accounted for a disproportionate share of the patenting
and licensing activities of the University of California during this period: bio¬
medical invention disclosures made up 33 percent of all UC disclosures dur¬
ing 1975-79 and 60 percent of patents issued to the University of California
for inventions disclosed during the period. Biomedical patents accounted for
70 percent of the licensed patents in this cohort of disclosures, and biomed¬
ical inventions accounted for 59 percent of the UC licenses in this cohort that
generated positive royalties. Biomedical inventions retained their importance
during 1984-88, as they accounted for 60 percent of disclosures, 65 percent of
patents, 74 percent of the licensed patents from this cohort of disclosures, and
73 percent of the licenses for this cohort of disclosures with royalty income.
Growth in the number of biomedical disclosures, which dominated UC
patenting and licensing throughout 1975—9°? thus predates the passage of the
Bayh-Dole Act. Additional evidence on the shifting composition of the UC
technology licensing portfolio is displayed in Table 6.1 and Figure 6.7. The
UC data in Table 6.1 reveal the high concentration of licensing revenues
among a small number of inventions before and after the Bayh-Dole Act and
highlight the remarkable growth (more than fifty-fold) in constant-dollar gross
PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

104 revenues during 1970-95. The share of gross licensing revenues accounted for
by the UC systems “top five” inventions actually decreases throughout the
1970-95 period, from nearly 80 percent in fiscal 1970 to 66 percent in fiscal
1995, having reached a low point of 47 percent in fiscal 1985.
Equally remarkable is the illustration in Table 6.1 of the shift in the UC sys¬
tems “top five” inventions from agricultural inventions (including plant vari¬
eties and agricultural machinery) to biomedical inventions. Among the three
universities in this comparative analysis, only the University of California
maintained a large-scale agricultural research effort. During the 1970s, agri¬
cultural inventions accounted for a majority of the income accruing to the
“top five” UC money-earners. Beginning in fiscal 1980, however, this share be¬
gan to decline, and by fiscal 1995 100 percent of the UC systems licensing in¬
come from its “top five” inventions, accounting for over $40 million in reve¬
nues (in 1996 dollars), was derived from biomedical inventions, up from
19 percent in fiscal 1975. Moreover, and consistent with the discussion of the
previous paragraph, this share increased before the passage of Bayh-Dole in
late 1980: the share of “top five” licensing revenues associated with biomedical
inventions jumped from less than 20 percent in fiscal 1975 to more than 50 per¬
cent in fiscal 1980. The depiction in Figure 6.7 of the share of UC licensing
agreements accounted for by biomedical inventions during 1975—95 as a
three-year moving average similarly shows an increase in this share before the
effective date of Bayh-Dole to a higher level that is sustained through much of
the 1980-95 period.

Stanford University

As we noted in Chapter 3, Stanford University’s Office of Technology Licens-

ing (OTL) was officially established in 1970, following the success of Niels
Reimers’s “experiment” with direct management of patenting and licensing,
and Stanford was active in patenting and licensing throughout the 1970s. Stan¬
ford s patent policy, adopted in April 1970, stated that “except in cases where
other arrangements are required by contracts and grants or sponsored research
or where other arrangements have been specifically agreed upon in writing, it
shall be the policy of the University to permit employees of the University,
both faculty and staff, and students to retain all rights to inventions made by
them” (Stanford University Office of Technology Licensing, 1983, p. 1). Dis¬
closure by faculty of inventions and their management by Stanford’s OTL thus
was optional for most of the OTL’s first quarter-century.
In 1994, Stanford changed its policy toward faculty inventions in two im¬
portant aspects. First, assignment of title to the university of inventions “de-
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

veloped using University resources” was made mandatory.5 Second, the uni¬ 105
versity established a policy under which “[cjopyright to software developed for
University purposes in the course of employment, or as part of either a spon¬
sored project or an unsponsored project specifically supported by University
funds, belongs to the University” (Stanford University Office of Technology
Licensing, 1994a).6
Faculty disclosure of inventions to university administrators thus was no
more mandatory at Stanford before 1994 than at Columbia before 1984. Nev¬
ertheless, especially during the 1970-80 period, Stanford operated a much
more elaborate administrative apparatus for the patenting and licensing of in¬
ventions than did Columbia. The expanding scale of Stanford’s licensing op¬
erations during the 1970s and 1980s suggests that a substantial share of faculty
inventions were disclosed to the OTL.
Data from the Stanford OTL provide some insight into the patenting and
licensing activities of a major private research university before and after
Bayh-Dole. And similar to the situation at the University of California, these
data suggest that the growth of Stanford s patenting and licensing activities re¬
flected influences other than Bayh-Dole. Figure 6.4 displays trends during
1975-90 in Stanford invention disclosures. The average annual number of dis¬
closures to Stanfords OTL increased from 74 during 1975-79, prior to Bayh-
Dole, to 149 during 1984-88. Figure 6.5 shows the number of patents issued
to Stanford during the 1975-95 period; the average annual number of issued
patents also increased from thirteen during 1975-79 to forty-two during 1984-
88. Although the evidence of a “Bayh-Dole effect” in the annual number of
disclosures (such as the jump in disclosures between 1979 and 1980) is stron¬
ger in the Stanford data than in the UC data, the increase in disclosures be¬
tween 1977 and 1978 indicates that the annual number of invention disclosures
was growing prior to Bayh-Dole.
The data in Figures 6.4 and 6.6 also suggest that the importance of bio¬
medical inventions within Stanfords invention portfolio advances had begun
to expand before the passage of Bayh-Dole. Figure 6.4 indicates that the an¬
nual number of biomedical invention disclosures began to increase sharply
during the 1978-80 period, and the share of all disclosures accounted for by
biomedical inventions (see Figures 6.4 and 6.6) increased steadily from 1977
to 1980, leveling off after 1980 and declining after 1983. The magnitude of
these increases in biomedical inventions prior to Bayh-Dole is more modest
than at the University of California, but the trend is similar.
The Stanford data in Figure 6.7 highlight trends in licensing agreements
that resemble those for the UC system during the same period; biomedical
inventions increased as a share of Stanfords (nonsoftware) licenses during
PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

FIGURE 6.4 Stanford University invention disclosures, 1975-90

Year of Issue

figure 6.5 Stanford University patents by year of issue, 1975-95

1975-90, although the upward trend was less pronounced and fluctuated
more widely than in the UC data.7 Table 6.1 indicates that as of fiscal 1980,
slightly more than 40 percent of the income from Stanfords “top five”
inventions was attributable to biomedical inventions, suggesting the con¬
siderable importance of these inventions prior to Bayh-Dole. This share in¬
creased to more than 96 percent by fiscal 1995. Stanfords licensing revenues
grew by almost 200-fold (in constant dollars) during 1970-95, and its “top five”
 figure 6.6 Stanford University biomedical disclosures as a percentage of
total disclosures, 1975—90 (3-year moving average)

—Columbia Stanford UC

figure 6.7 Biomedical technology share of Columbia, Stanford, and Uni¬

versity of California license agreements, excluding Cohen-Boyer and software
licenses, 1975-95 (3-year moving average). Columbia shares reported for
5^983—95 (100% of licenses at Columbia were for biomedical technologies in
1981, but Columbia did not execute any licenses in 1982). Stanford shares re¬
ported for 1975-90.
PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

108 inventions accounted for a larger share of gross income for the 1980-95 period
than do the “top five” UC inventions.
Both Stanford and the UC system thus experienced a shift in the composi¬
tion of their invention and licensing portfolios in favor of biomedical inven¬
tions prior to Bayh-Dole. Bayh-Dole was an important but not a determinative
factor in the growth and changing composition of patenting and licensing ac¬
tivity at these institutions.
In contrast to the data from the University of California, Stanfords inven¬
tion disclosures include a number of software inventions, which account for
10-15 percent of annual disclosures. As is the case at Columbia during the
1980s, most of these inventions were not patented and therefore cannot be
traced through annual patent counts. The importance of software disclosures
in Stanford’s licensing activity expanded during the 1970s and 1980s. Only two
of the forty-one inventions disclosed during 1974-79 (less than 5 percent) that
were licensed within eight years of their disclosure were software inventions,
but this figure increased to more than 20 percent for the 1984-88 period.
Many of these software inventions (for example, the WYLBUR operating sys¬
tem) were licensed on a nonexclusive basis to academic institutions through
Stanfords Software Distribution Center during the 1980s. The majority of
these licenses involved a small, one-time payment by the licensee institution.8
Partly because of the large number of such “site licenses,” the coverage by our
data of Stanford software licensing agreements is spotty, and our estimate of
the share of all Stanford licensing agreements accounted for by software is less
accurate. Nonetheless, like Columbia University, a significant portion of Stan¬
ford’s licensed inventions (at least 10-20 percent of annual licensing agree¬
ments and a smaller share of gross revenues) covered nonpatented inventions
during the 1980s, Bayh-Dole notwithstanding.

Columbia University

The essence of Columbia’s pre-Bayh-Dole patent policy is contained in the

University’s 1944 “Statement of Research Policy and Patent Procedures”
(Trustees of Columbia University, 1944).9 This policy prohibited patents on
the results of research (regardless of the source of funding for such research)
at the medical school, although other faculty members at the university were
free to patent inventions resulting from their research. In many cases, non¬
medical faculty turned to the Research Corporation for assistance in patent¬
ing and administering their inventions. The 1944 policy remained substan¬
tially intact until 1975, when the stipulation against patenting medical
inventions was dropped. Discussions within Columbia’s administration and
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

faculty over institutional assertion of rights to faculty inventions began in the 109
late 1960s and gained momentum in the 1970s but produced no formal change
in policy until the passage of Bayh-Dole (for a fuller discussion, see Crow
et ah, 1998).
Columbia’s laissez-faire policy toward patenting the results of faculty re¬
search during the three decades following World War II meant that the uni¬
versity had no administrative structure for managing patent matters. Begin¬
ning in the late 1970s, however, the university expanded its involvement in
patent transactions between its inventors and the Research Corporation, cre¬
ated a central archive of potentially patentable inventions, and petitioned gov¬
ernment agencies for title to faculty inventions resulting from federally
funded research. Indeed, Columbia filed a patent application on the Axel co¬
transformation invention, which accounted for the largest share of its licens¬
ing revenues during the post-1981 period, before the passage of Bayh-Dole or
the establishment of a technology transfer office (see Chapter 8 for a more ex¬
tensive discussion of the Axel cotransformation invention and patent). Never¬
theless, fewer than ten patents were issued to Columbia University during
1975-8i-
Columbia changed its policies toward faculty patents and created an ad¬
ministrative apparatus for managing their prosecution and licensing only after
the passage of Bayh-Dole. A new patent policy, which took effect on July 1,1981
(the effective date of Bayh-Dole), stated that Columbia could assert rights to
faculty inventions created within university laboratories or research facilities,
mandated the disclosure of such inventions to the university, and provided for
royalty sharing with the inventor and his or her department. In 1984, this pol¬
icy was published in the university Faculty Ftandbook. In 1989, Columbia’s
policy on reserving rights to the university for faculty inventions was extended
to cover software. Inventions were to be disclosed to Columbia’s technology
transfer office, the Office of Science and Technology Development (OSTD),
which was founded in 1982. OSTD was renamed the “Columbia Innovation
Enterprise” in 1994, and in 2001 its name was changed again, to “Science and
Technology Ventures.”
Figure 6.8 shows the rapid “ramping up” of Columbia invention reports
during the 1980s. Since most academic research programs change only grad¬
ually, the initial surge of invention reports may reflect increased identification
by university administrators (based on a more intensive canvassing of the fac¬
ulty) of potentially valuable inventions derived from research projects already
under way. Almost 75 percent of the 877 invention reports disclosed between
1981 and 1995 are biomedical technologies (that is, biotechnology, medical
devices, pharmaceuticals, and biochemical compounds), and biotechnology
PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

figure 6.8 Columbia University invention disclosures, 1981—95

inventions accounted for 60 percent of these biomedical invention reports.10

Although our Columbia data on inventions and patenting do not extend back
into the pre-1980 period, the timing of the changes in this university’s patent
policies, as well as the filing of a patent application in 1980 by Columbia ad¬
ministrators on a major biotechnology invention, suggests that the growth of
Columbia’s post-1980 patenting and licensing activity in biomedical tech¬
nologies was caused by more than the passage of Bayh-Dole.
Outside of the medical school, Columbia’s inventive activity was concen¬
trated in a few departments and research institutes that, like the medical
school, rely heavily on federal R&D funding. During the 1981-95 period, over
60 percent of the nonbiomedical invention reports and over 65 percent of the
patenting associated with nonbiomedical invention reports originated in two
departments, electrical engineering and computer science, and two research
centers, the Center for Telecommunications Research and the Lamont-
Doherty Earth Observatory. Software inventions accounted for a significant
share of Columbia faculty disclosures, increasing to more than 10 percent of
disclosures by the 1990s; software inventions also account for a large share of
Columbia licensing agreements. The data in Figure 6.9 suggest that increased
invention reports generated increased patenting with a slight lag.
As the comparative data in Table 6.1 indicate, Columbia University’s tech¬
nology licensing activities were associated with a surge in constant-dollar gross
licensing revenues, which grew almost sixty-fold in the decade between 1985
and 1995. This income was highly concentrated among a small number of in¬
ventions: the “top five” accounted for more than 90 percent of gross revenues
throughout this period. Biomedical inventions accounted for more than
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

8o percent of the revenues of the “top five” inventions during the 1985-95 pe- 111
riod (Table 6.1).11 Moreover, biomedical technologies accounted for most of
the nonsoftware licenses at Columbia between 1983 and 1995 (Figure 6.7).
In assessing the effects of Bayh-Dole on Columbia University, we lack a
compelling counterfactual: what would have happened in the absence of this
federal law, given the other trends operating in university finances and re¬
search after 1980? We believe industrial interest in Columbia's research re¬
sults, especially in the biomedical area, combined with the prospect of large
licensing revenues, would have led Columbia to develop some administrative
machinery for patenting and licensing in the absence of the Bayh-Dole Act.
Indeed, as we noted earlier, Columbia appears to have been moving in that di¬
rection before Bayh-Dole. Nevertheless, the change in federal policy embod¬
ied in Bayh-Dole may well have catalyzed a more dramatic change in policies,
procedures, and rules than would otherwise have occurred.
Another piece of evidence relevant to an assessment of the effects of Bayh-
Dole at Columbia concerns the significant role of software in Columbia’s post-
1980 licensing activities. Virtually all of the software inventions licensed by
Columbia during the 1980s were protected by copyright, a form of intellectual
property never affected by Bayh-Dole, rather than by patents, the focus of this
federal law. Software licensing was a new form of technology marketing that
resulted from the university’s creation of a technology marketing operation

figure 6.9 Columbia University patents, 1981-95

PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

112 and (like biomedical research) from exogenous changes in the academic
research agenda, rather than from the specific policy shifts embodied in Bayh-
Dole. Moreover, although the arguments for Bayh-Dole stressed the impor¬
tance of exclusive licensing in effective technology transfer and commercial¬
ization, many of Columbia’s licensed inventions, including its biggest single
source of revenues, have been licensed widely on a nonexclusive basis.

Comparing Invention Disclosures and Licenses

at All Three Universities in the Late 1980s

In addition to comparing the periods before and after Bayh-Dole for two of
these three universities, we compared data on the disclosure, patenting, and
licensing activities of all three universities for the 1986-92 period to assess in-
terinstitutional similarities and differences after the passage of the Bayh-Dole
Act.12 The data in the top panel of Table 6.2 suggest considerable similarity
among these three universities in the characteristics of their invention disclo¬
sures, though a larger share of Stanford University’s disclosures are licensed
and a larger share of Stanford’s invention disclosures yield positive licensing
income than is true of either Columbia or the University of California.13 Re¬
stricting the focus to biomedical inventions does little to change the conclu¬
sions of this comparison among our three universities (the second panel of
Table 6.2).
The comparison of software inventions for the two universities within our
sample (Stanford and Columbia) that have been active licensors of these tech¬
nologies suggests that software licenses rarely involved patented inventions in
the early post-Bayh-Dole period: 100 percent of the software inventions dis¬
closed at Stanford during 1986-90 and 83 percent of Columbia’s software in¬
ventions disclosed during this period that were licensed within six years of dis¬
closure were not patented.
Another issue for consideration in our comparative evaluation of licensing
in the post-Bayh-Dole era at these three universities concerns the share of in¬
ventions that were licensed through “exclusive” contracts, which we define
here as contracts that are globally exclusive or that contain specific field-of-use
or market restrictions. A relatively high percentage of all inventions that are li¬
censed—as high as 90 percent for UC licenses and no less than 58.8 percent
for Stanford licenses of “all technologies” during this period —are licensed on
an exclusive basis under this definition, and these shares are similar for bio¬
medical inventions. Perhaps because of the weaker formal protection for this
technology (during much of this period, the lack of patent protection), soft¬
ware inventions are less frequently licensed on an exclusive basis: 46 percent
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

TABLE 6.2 113

Comparative Evidence on Invention Disclosures and Licenses at Stanford University,

Columbia University, and the University of California, 1986-90

Institution

Technologies Stanford Columbia UC

1986-90 (6-year “trailing window”):

All Technologies
% of disclosures yielding patents 23.2 18.6 20.4
% of disclosures that are licensed 33.2 16.4 12.3
Disclosures with licensing income
>0/all disclosures (%) 22.4 12.3 7.4
Licensed disclosures with licensing income
>0/number of disclosures that are licensed (%) 67.4 75.0 60.6
% of licensed disclosures that are licensed exclusively 58.8 59.1 90.3

1986-90 (6-year “trailing window”):

Biomedical Technologies
% of disclosures yielding patents 17.5 15.3 15.7
% of disclosures that are licensed 38.7 17.3 14.8
Disclosures with licensing income >0/all disclosures (%) 33.5 13.9 10.0
Licensed disclosures with licensing income
>0/number of disclosures that are licensed (%) 86.6 80.0 67.2
% of licensed disclosures that are licensed exclusively 54.9 62.9 90.3

1986—90 (6-year “trailing window”):

Software Technologies
% of disclosures yielding patents 0 17.6 NA
% of disclosures that are licensed 53.6 35.3 NA
Disclosures with licensing income > 0/all disclosures (%) 45.5 23.5 NA
Licensed disclosures with licensing income
>0/number of disclosures that are licensed (%) 84.7 66.7 NA
% of licensed disclosures that are licensed exclusively 46.3 16.7 NA

and 17 percent, respectively, of software invention disclosures at Stanford and

Columbia were licensed exclusively during 1986-90.
Nevertheless, the licenses accounting for the largest share of revenues at all
of these universities are nonexclusive licenses. The Stanford-UC Cohen-
Boyer patents were licensed widely and nonexclusively.14 Columbia Univer¬
sity’s Axel cotransformation patent also was licensed on a nonexclusive basis.
Although many proponents of patent protection for university inventions dur¬
ing and after the debate over the Bayh-Dole Act argued that would-be com¬
mercial developers of these technologies needed exclusive title to this intel¬
lectual property in order to obtain a clear “prospect” for their significant
investments in this activity, these cases suggest that for inventions of broad
promise and potentially widespread use, nonexclusive licenses accommo¬
dated both universities’ interest in revenues and the needs of commercial users
 oo <. rr> Qv r^> rsj
Qv Z K Z fVN VC
fvj .—I O h-

oo
oo oo -o
3
OO l/n ^ o
oo rvi o’ Z rvj o
NO o
CN

K
^3
Os H<\qo <! t~^ r-
r\ iri 4 O N Z fA O
NO O
LO
on
co ~o
3
C> < O CO < Tf-
rj- Z fv> Qv Z rvi r—*
vO O'
O

C O; N VO O
^3
Z Z rvj iX CAV LAN
^ C
N ro f-H "=t" r<v Tt-
C S ,o
<U ^
•3 CO
c2
5U <n o o o
£ 6 "t LA i*' OvJ LAN tJ-
_ o rsj r^j VO N OO
S "o
o .£
rr> o ^
oow~*
MD
W o
K co
CQ O
< •2 |
H c E-<
£ *" c
kH ^ o
£
oo : ~o
1 c
<5k) o
"T3 '.£ J3
C C ^
'5 n <D
V
x-v
r—i
ID

"S
v&
K g
O 3
CO
SJ •— o
CO &
C .2 JS Vi
^3 JD
G
C "ca
^3 Oh G >
Oh rc .2 -o
o «
"T3 ^ •- <D n ^
•2 .y CO3
'2
c
c c G
<D <u <D --1 ■
Oh'-
nJ > CJ Oh 42
vg. 03 -g
bjDi^T tuD C -2
3 a w>
,o _G
■s^ +H E e K-
as <y 03 -+-j
O 03 v—✓
I—
o "3 2
<D G <D a
Oh ^
C CO
•—H 3 G <u
G \ '-o
■fr <U <D O
y <D "rj
co
£n W)G3 d£ 2 cd G
s « CD
G <D
c J2 15 co O
*5 •Si 3 O DJ
ID
J= 8
on •—
q 42 42
cyo o a o o CkJO CD G G
£* CO OO CO CD
2 C H-H03
H—•
03
■■§ 5 Q D I Oh Oh
if "-r^r^r
'■a T3 <—I <N fA g LA VO h
c
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

for access to the essential intellectual property (see Chapter 8 for additional
discussion of the role of exclusive licenses in specific cases of university-
industry technology transfer).

The Yield of Patenting and Licensing at the University

of California and Stanford, 1975-79 an<^ 1984-88

In this section, we use the data on UC and Stanford disclosures, patenting,

and licensing to discuss the technology marketing efforts of these universities
before and after Bayh-Dole. The data in Table 6.3 provide measures for the
1975~79 an<^ 1984—88 periods of (1) the intensity of these institutions’ tech¬
nology marketing efforts, reflected in their propensities to patent or license
faculty invention disclosures, and (2) the yield of these marketing efforts,
based on such indicators as the share of patent applications that result in is¬
sued patents, the share of issued patents that are licensed, and the share of li¬
censes that generate positive income.15
The first row in Table 6.3 displays “before and after” data for the share of in¬
vention disclosures resulting in issued patents, which scarcely changes be¬
tween the two periods at the University of California and increases from nearly
15 percent to slightly more than 25 percent (excluding software, from 15.5 per¬
cent to 29.8 percent) at Stanford. The second row of Table 6.3 shows that the
UC system also increased its patent application rates in the wake of Bayh-
Dole—the share of disclosures that generate patent applications rose from
24 percent to slightly more than 31 percent. But this increased tendency to
file for patents was associated with a decline in the “yield” of these efforts,
as the share of patent applications resulting in issued patents dropped from
more than 62 percent in the 1975-79 period to less than 44 percent in 1984-
88 (row 5 of Table 6.3). These data on marketing intensity and patent yield sug¬
gest that UC administrators intensified their efforts to protect and promote fac¬
ulty inventions after the Bayh-Dole Act. Their intensified efforts nevertheless
appear to have produced a lower “yield,” as measured by the share of applica¬
tions resulting in issued patents. The increase in the share of Stanford disclo¬
sures yielding issued patents after 1980 is more difficult to interpret, since we
lack data on patent applications that are comparable to those from the Uni¬
versity of California. But this increase is broadly consistent with a more ag¬
gressive effort to patent faculty inventions.
Rows 3, 4, 6, and 7 in Table 6.3 provide additional evidence of change in
the “intensity” and “yield” of the UC and Stanford technology-marketing
efforts after Bayh-Dole. The share of UC disclosures resulting in licenses
(row 3) increased from 5.6 percent in 1975-79 to 12.6 percent in 1984-88,
PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

n6 while the share of UC disclosures resulting in licenses that generated positive

royalties (row 4) grew from 4.6 percent in 1975-79 to 5 percent in 1984-88. In
contrast, the share of Stanford disclosures that were licensed nearly tripled be¬
tween 1975—79 and 1984—88. The share of Stanford disclosures yielding posi¬
tive licensing income more than doubled (growing from 9.8 percent to 22 per¬
cent) when software invention disclosures are included and nearly doubled
when these are excluded (increasing from 9.5 percent to almost 18 percent).16
At both the University of California and Stanford, the proportionate increase
in the share of disclosures that were licensed after 1980 (a measure of the in¬
tensity of institutional technology marketing efforts) exceeded the propor¬
tionate increase in the share of disclosures yielding positive licensing income
(an indicator of the yield of these marketing efforts). We interpret these data
as indicating some decline in the yield of licensing revenues during the post-
Bayh-Dole period in both universities’ licensing operations.
Two additional measures of the yield of these institutions’ technology mar¬
keting efforts, the share of issued patents that were licensed (row 6 of Table 6.3)
and the share of these licenses yielding positive income (row 7 of Table 6.3),
also suggest some decline in yield after the Bayh-Dole Act. The share of UC
patents that were licensed grew from slightly more than 25 percent to more
than 35 percent, but the share of these licenses yielding positive income
dropped by almost half, from roughly 84 percent to 47 percent. The “before
and after” data at Stanford reveal a similar trend. The share of issued Stanford
patents that were licensed (an indicator that is comparable to the UC data only
for the far-right columns that exclude software inventions) is essentially un¬
changed, increasing from almost 63 percent to almost 64 percent. The share
of Stanford licenses yielding positive income17 dropped from 90 percent in
1975-79 roughly 76 percent in 1984-88 when software inventions are ex¬
cluded (a more meaningful basis for comparison with the UC licensing
effort).18
Overall, these data suggest an intensification in technology marketing ac¬
tivities after the passage of Bayh-Dole. The post-Bayh-Dole period of inven¬
tion disclosures is associated with more extensive licensing of the patents re¬
sulting from these universities’ disclosures. The more extensive licensing
efforts of the post-1980 period also appear to have produced some decline in
the “yield” of these licenses at both the University of California and Stanford,
measured as the share of licenses yielding positive royalties. But these indica¬
tors of decline in the average productivity of the marketing efforts of the UC
and Stanford technology licensing offices need not imply inefficient or eco¬
nomically irrational behavior —after all, the marginal, rather than the average,
returns to licensing activities are most important in evaluating the returns to
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

these institutions licensing activities. Moreover, our measure of the “yield” of

these universities’ licensing activities does not capture the size of the revenue
streams associated with the average or the marginal license before and after
Bayh-Dole.19 Nevertheless, these effects appear to be relatively small and sug¬
gest that Bayh-Dole had modest effects on the “technology transfer” activities
of these established academic patenters and licensors.

THE “IMPORTANCE” AND “GENERALITY” OF STANFORD,

COLUMBIA, AND UC PATENTS, 1975-92

The data on the intensity and yield of the technology marketing activities at
the University of California and Stanford suggest that in the first decade after
the passage of the Act, these “incumbent” universities’ technology transfer
efforts intensified moderately and their yield declined. Did these changes in
their patenting and licensing activities reflect a shift in the characteristics of
Stanford and UC research in response to the Bayh-Dole Act? The possible
effects of Bayh-Dole on academic research, especially on the mix of funda¬
mental and more applied research activities within universities, has been the
focus of a growing body of empirical research.
One analysis of this issue by Thursby and Thursby (2002) used nonpara-
metric linear programming techniques to disaggregate the contributions to
growth in university commercial activity between 1994 and 1999 into those at¬
tributable to changes in the nature of academic research and those attributa¬
ble to changes in other factors, including universities’ propensities to file pat¬
ent applications and firms’ propensities to license university inventions. These
authors cannot reject the hypothesis that the content of faculty research
shifted during the 1994-99 period, but their analysis suggests that most of the
recent growth in university patenting and licensing is largely attributable to
other changes, particularly the increased propensity of universities to file for
patents, instead of some shift in academic research toward applied topics.
Another body of empirical work on university patenting and licensing com¬
pares the characteristics of individual researchers who do and do not patent. If
academic patenting has indeed “crowded out” basic research, faculty who are
relatively intensive patenters should have fewer (or fewer significant) publica¬
tions in leading basic research journals. But the limited empirical research on
this issue reveals few such differences between faculty who patent and faculty
who do not. The analysis of fifty-five “star” academic scientists in biotechnol¬
ogy by Zucker, Darby, and Armstrong (1994) found that faculty with patents
were the authors of a significantly larger number of publications that were
cited by other papers (and therefore deemed more scientifically important)
PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

118 than faculty who did not patent. Similar results for faculty in nonbiomedical
research are reported byAgrawal and Henderson (2002) in their analysis of pat¬
ents and publications by faculty members in two MIT engineering depart¬
ments. Agrawal and Henderson found no evidence of a negative relationship
between the extent of patenting and publication by the faculty members in
their sample. Indeed, the number of patents obtained by an individual faculty
member was positively related to the quality of his or her publications, as mea¬
sured by the number of times they are cited in subsequent publications. Pre¬
liminary work by Stephan etal. (2002), based on data from a 1995 National Sci¬
ence Foundation survey of doctoral-degree recipients, also finds a positive
relationship between the number of patents and publications by individual
faculty members.
Still other scholars have examined various characteristics of academic pat¬
ents to determine any changes in the content of academic research activities
after the Bayh-Dole Act. The analysis of post-1980 academic patenting by Hen¬
derson, Jaffe, and Trajtenberg (1995, 1998a) found that these intensified post-
Bayh-Dole efforts to market faculty inventions were associated with the issue
to U.S. universities of patents that were less “important” and less “general,”
based on the patterns of citations to these patents. Moreover, these scholars
also found that experienced academic licensors’ post-1980 patents declined
somewhat in quality, as measured by the “importance” and “generality” (de¬
fined below) of these patents. As Henderson, Jaffe, and Trajtenberg (1995) ar¬
gue, these changes in the characteristics of U.S. universities’ post-1980 patents
“may reflect changing motives within the university research community or
shifts in the focus of university research from basic to more applied work” (p. 1;
see also Foray and Kazancigil, 1999).
Our analysis of this issue in this chapter follows the approach of Hender¬
son, Jaffe, and Trajtenberg (1995, 1998a) and Trajtenberg, Henderson, and
Jaffe (1997) in examining the importance and generality of UC and Stanford
patents before and after Bayh-Dole and comparing these characteristics of
UC, Stanford, and Columbia patents after 1980. Although we examine a sim¬
ilar set of characteristics of these universities’ patents, we employ a slightly dif¬
ferent “control population” of patents (see below). Our “before and after”
analysis of UC and Stanford patents uses the year in which the invention was
first disclosed as the key datum in classifying faculty disclosures and any asso¬
ciated patents as falling into the pre-Bayh-Dole or post-Bayh-Dole eras. We
further categorize patented disclosures by the date of the patent application.
Our data for the University of California and Stanford thus contain: (1) dis¬
closures and patents for which the disclosures and patent application occurred
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

during and after 1970 and the issue of the patent occurred before 1981 (the year
during which the Bayh-Dole Act went into effect); (2) disclosures and patent
applications that occurred during or after 1981 and before 1993; and (3) inven¬
tions disclosed after 1970 and prior to 1981, for which patent applications were
filed during or after 1981 and before 1993.20
We also include Columbia University patents in our analysis of the post-
Bayh-Dole era. Although we define Columbia as an “entrant” academic
patenter, reflecting the fact that this university developed an active patenting
and licensing policy only after 1980, as mentioned earlier, Columbia did
accumulate a modest portfolio of fewer than ten patents during the 1975-80
period.
Henderson and colleagues use patent-citation measures in their analysis of
university patents. When the U.S. Patent and Trademark Office grants a pat¬
ent, the patent examiner includes a list of all previous patents constituting
“prior art.” This list is made public as part of the publication of the patent at
the time it issues. The examiner is aided in compiling this list by the patent
applicant, who is legally bound to provide with the application a list of all
“prior art” of which he or she is aware. Citations of prior patents can thus serve
as an indicator of the technological lineage of new patents, much as biblio¬
graphic citations indicate the intellectual lineage of academic research.
Our analysis of UC, Stanford, and Columbia patents focuses on “forward
citations” to these patents, defined as the number of citations received by each
patent following its issue. Citations to patents typically peak four to five years
after the date of issue of the cited patent. As a result, data on citations to more
recently issued patents will be “right-truncated,” that is, more recent patents
will be underrepresented in the citations data. In order to address this prob¬
lem, our dataset includes only citations to patents that occurred within six
years of the year of issue of the patent, and our sample includes only patents
issued between 1975 and 1992. Our dataset also includes a control sample of
nonacademic patents for each of these three universities, spanning the same
time period and replicating the distribution of the UC, Stanford, and Co¬
lumbia patents among patent classes. We matched one nonuniversity patent
with each university patent, based on the patent class and application date of
the relevant university patent.21 Citations to patents in the control sample also
are restricted to those occurring within six years after the year of issue.
We used the number of citations to a patent during the six years following
its issue as a measure of the importance of the patent, based on the assump¬
tion that citations form an index of the influence over subsequent inventive
activity of the cited patent. The number of observations for our UC, Stanford,
 O'
Tt_

r- o -t-
OO N OO
rg

u Tf O' ^
C^)
co
D ir\ O'

-S
.2
Q Zg
•2 r^.
^5 JB3 vO
O
g U
-5o
T3
O u2 jv\ o rw
^3 c
ca r^4 —< O'
C o
tn
cq

K U rr\
t
r<v
N
xj-
ia
5 D
cn
.g*
£
v— .2
vd
w E i 2
__p
CQ O o
< U
H o
2<u "H
>
£c o
cr\
r-
r-
O OJ rv\

K C/5
50
•js
O C^i
rrv LT\
t}-
•£ D O'
rvl
i*

C
o
v*
ii, °° «
v2^
£
=a v §:
£ v/ =a "
R v/ | cx
cx
O' IA CU OJ
"—' Q-rvi ■+- rvj
c®; ••o' C O'
O O'
• 2 ^ 00 _

SS2v/ S.w
^.av
£ “ T3 _ ^ '-O r—H

-*—1™ <L>G O£ O^
“O ^ OO
J3 2
I s.’S H « A\
c£ Q Q
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

and Columbia samples of patents is displayed in Table 6.4 (because Columbia

has very few pre-1981 patents, we exclude this university from the pre-1981
data). We also divide the Columbia, UC, and Stanford patents into biomed¬
ical and nonbiomedical categories in order to test for differences in the effects
of Bayh-Dole between these broad technology classes.22 Clearly, for some cells
of the dataset (for example, Stanford biomedical patents in the “disclosed be¬
fore 1981 and patented during or after 1981” subset), the number of observa¬
tions is small (far smaller, for example, than the university patent samples for
individual years in Trajtenberg, Henderson, and Jaffe, 1997). Because of the
small number of observations within each, the standard errors for these sub¬
samples are large, and the statistical significance of our tests is low. Neverthe¬
less, we believe that measures of the importance and generality of this cate¬
gory of our patents are worth computing and analyzing separately and
therefore have retained these small subsamples in the results that follow.
The upper panel of Table 6.5 displays means and standard errors for the
number of citations to the UC, Stanford, and Columbia patent samples and
their disaggregated subsamples. We also present means and standard errors for
the control patent samples and subsamples corresponding to each of the UC,
Stanford, and Columbia categories in the lower panel of Table 6.5. Without
exception, the number of citations received in the six years following issue by
the patents of all three universities is larger than those received during the
identical period by the control samples of patents, for all time periods and dis¬
aggregated categories. There is no apparent time trend in the absolute magni¬
tude of differences between the number of citations received by the Stanford
and UC patents and those in the control sample. With the exception of bio¬
medical patents in the second and third of our three subperiods, the mean
number of citations received by the Stanford patents exceeds the mean num¬
ber of citations received by UC patents, although these differences frequently
are small. With one exception (Stanford biomedical patents), UC and Stan¬
ford patents also are more heavily cited than Columbia patents during the
post-1980 period.
The upper panel of Table 6.5 also reports the results of tests for the statisti¬
cal significance of differences between the mean values of forward citations to
the UC, Stanford, and Columbia patents and forward citations to patents in
our control samples.23 We find no evidence of a statistically significant decline
in the importance of Stanford and UC patents, relative to the control samples,
after 1980 (columns 2 and 3).24 Tests for the significance of the difference in
mean citations between the Columbia patents and the Columbia control
sample, however, indicate that the difference is significant only for the overall
Columbia patent sample at the 10 percent level. The absence of significantly
 O' PsJ LAs Al
rA vq SO LAN
| | | |

NOTES: Standard errors in parentheses. Unequal variances assumed. Asterisks denote significance in difference in means tests (t-statistics) between university and
* *• Vr
* r-s Vr Vr
VO rA 0 VO rA AJ LA O O LA
rA rs| pH vq r-; O Al O aa OO
LA t< O' r4 OO Al’ rA rA rA.

C© VO ■—I r<\ N rvv rA ai

Tj~ VO '—i LA Tf- CO co q ^
rv\ vO OO
'I- O'
tJ- o' a vd la" Al Tf‘ r—i ai’
^3
C
^3

*tf
V-

O' r-. ai rf-

c rt- vq O' VO
-S i i
^t* ia i I
C/5 £
tj
oo
•S'
'£
K- 3
^^ x ,_^ ^
£ 03
r- 0 r- rxj LA O' r- 0 CNj LA r-
■—; ■—_ vq r^ rA ca Tj- rA >—|
o t-2
c
•“-Hi vd AA rj- rA -rf- rrs r4 r4 AsJ rvi rA LA
«
3
s Zn

e
o

O
OJO O O rA VO O Al Al a vo Al o o
ia rA LA tJ- r—( LAN rs)
.S O tJ- O' rA
'O ~5 rA rA LA LA Ml) rA t}- ai Tf
c
w o
J
CQ
<
O LAN ^1“ Tf- O'
Tf- rsj —(
o I I I i
^t- vd
^3
o C
^3
~2 42
^3 # *
* *
§ LA LA rA VO Al LA 00 A O r- r-
vq O' O' rA Al cq O O' c>
£ eg LA* 06 r4 VO vd A A4 Al rA rA. rj-
p > 0.05; * p > 0.10.

•S
£5

# *
,—v «• #
VO so rA 1—1 't- Al rA rA LAS O'
'"—w O LA O' rA Al 0^ A LA —* r-

£ O rA A- LA A LA vd^ A4 A rA rA
c
^3
c
0
Al
O'
"ct
corresponding control group samples:

O'
LA
IA
la <U
> *v
QO ^
_>
•4-J Cn “O A A A. °<3
C
<D
v/
. ,
Cu
ex
Co
C a •
‘
o
A

• _-.
‘5 —•
OO
c
A\
'" d)
03 O' H3 (U
o °° D
k-. o
&jDr-Ho
O' ^
CX„ a — CD _C
C 3 • •* OO
c
o ai «-A\ 33
co
O
-*- 43 A\ nj
A\ ^
*5 2 O'
3-C
CO 03 C £
(L)
c <u ■ ?? c <u rrs
O —
•is'
co
3 c
<D
A\ 1-2
•.5 c — 03 • j2 O ex-2

c°a V/ al D ,—i ^ 03 03 3 a t 3
CX '
CX•
■ ra O
_ .y to oo ^ e o o c o

03
ex •2~g^
e-o. ^2 ab^vE &
■g - V cxai
WO O-

:§ .y -2 oo
rt £ _,
ts cx^LT v 42 42
O .3 00 Sx C TO O' C
U 03 o
S O? O O'
ex S3 ■ i i O' <0 <L) -*-< Ca <l)
c c ^
£ V V/ Q £5 A\ .22 V V/ D £03
V/ s.103 V 1l
Oh Oh
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

greater citation rates for Columbia patents could reflect a less selective ap¬ 123
proach to patenting during the early years of Columbia’s licensing activities,
reflecting its limited experience in patenting in the early years of the Office of
Science and Technology Development.25 We discuss the possible effects of en¬
try following the Bayh-Dole Act by inexperienced institutional patenters on
post-1980 overall academic patent “quality” in Chapter 7.
The results of our tests for the significance of differences between the mean
citation rates of the UC and Stanford biomedical patents and their respective
nonacademic patent control samples indicate that the higher citation rates for
these academic patents are statistically significant (at the 5 percent level) in
the post-1980 period only for the University of California (the higher number
of citations to UC biomedical patents for the “middle period” is weakly signifi¬
cant, at the 10 percent level). The mean number of forward citations for Stan¬
ford nonbiomedical patents significantly exceeds that for their control samples
for all three periods, while the citation frequencies for UC nonbiomedical pat¬
ents are significantly greater (at the 5 percent level) than those in the control
sample only once, for the post-1980 period.26 The results for Columbia during
the post-1980 period indicate no significant differences in citation frequencies
with the patents in the Columbia control sample, but there is no evidence that
these Columbia patents are significantly less heavily cited than those in its
control sample.
These results must be interpreted with considerable caution, in view of the
small number of observations for some time periods. The fact that significant
differences in importance between the university and control sample bio¬
medical patents’ citations are relatively infrequent is interesting in view of the
importance of biomedical patents within the patenting and licensing activities
of Stanford and the University of California before and after Bayh-Dole. But
these data provide no indication of any decline in the importance of these uni¬
versities’ patents, relative to our control samples of nonacademic patents, after
the Bayh-Dole Act. If anything, these data suggest that the UC and Stanford
patents’ relative importance increased, rather than declined, after the Bayh-
Dole Act. Although these results do indicate that the patents applied for dur¬
ing the 1980s by Columbia, a university that did not patent significantly prior
to Bayh-Dole, were less “important,” relative to all nonacademic patents, than
those of Stanford and the University of California during this period, they do
not suggest that Columbia’s patents were significantly less important than
those in its control sample.27
Table 6.6 presents data on the generality of UC, Stanford, and Columbia
patents before and after Bayh-Dole. The more widely cited a patent outside
of its “home” patent class, the greater its generality and, arguably, the more
 r—I LTN
^ It
lr\ Tf-
I I I I 't Tt
o o c6 o

NOTES: Standard errors in parentheses. Unequal variances assumed. Asterisks denote significance in difference in means tests (t-statistics) between university and
3
00
-Cl

2 o IA o OO Tf O'
O la ^ h- rA VO rr\ c<\ rf- rj-
o o o o o o o o o o o’ o
is
c
o
O
cuo VO O O' t"- rr\ >—i rvj cr\ \o -OsJ r^J r/>
rrv tJ- vo cr\ \Q tJ- Tt" Tf- rf- -f ~t~ ^
.£
£3 OOOOOO o o o o o o’
c
o
-CL
CO
2
L.
o I I
r-i Cf\
LTN Tt“ I I cf'
rTN
Tt“
O
o o
K
CJ

c
r- .—' h- 0s rw vo N ^ -t h- N
-+-J
l/-\ c-g -rj- T1“ \0 t}~ r<v tJ-
r*> Tf
c2 o o o o o o 666666

£
"o ■— ’t O' N N t O' vO ltn rv\ r^\
tJ- -q- LTN rt- LTN
TABLE 6.6

O
O CO O O O O 0 0*0000

co
c I I LTs r^- I I c^v Tj-
-S o o O O
oo

VO <—I VO LTV Tf
LTN^L^^-VOrJ- Tt- Tf Th
666666
O
responding control group samples: *’s p > 0.05; * p > 0.10.

o
£>
CO N N O VO rA ^ 1" <-• N r^.
rrv rj~ *0 u^n ^~ ^ ^ 't
666666 666666
c
ID

V O'
LTN o
<31 cc °o r^. Q^J
o ^ W coj;
c 43 ^ oo
C O'T) C o& CL O'
<^> <—1 QJ <L> CL
O 03
£A g 2 oo 4-- r-sj S;a\ § A\ A\
0-.QS
■?- C .K C O 3 C -2 c
£\! O 4M OJ C> P O 4- o
OO A\ *XS C3
£ aj OO O'
S.w - <sj 03 <U r—H
^ o
,> ^ .2 to OO
3
£ CL S.2 V •2 ^
CL CL
T3 TO CL^nJ CLOvl
I ^ & sr— V (D O ^ 00
a OO ^ £o
v C30 ^ 03 O 03 O'

c ■g o' °d £ o 2" O -g o °o 4-> O'

c —<
.*_. O'
c —
CJ cj CL
Q» 2V V/ 03 ^ V V/ ^ V/ £ VI
03 03
Pm Pm

O
o
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

significant the advance in knowledge represented by the patent. Following 125

Henderson, Jaffe, and Trajtenberg (1995, 1998a), we compute generality as
follows:
IV j
NCITINGik\2
GENERAL = 1 - 2
*=1 NCITINGi )

where for patent i, k is the index of patent classes and N, the number of differ¬
ent classes to which the citing patents belong. This measure can be computed
for only those patents with at least one citation, and the sample in Table 6.6
accordingly differs slightly from that in Table 6.5. As in our measure of im¬
portance, we restrict forward citations to those occurring during the first six
years following the issue of the patent. Higher values of GENERAL, indicate
that a patent is relevant to subsequent inventive activities in a broader range of
technological fields.
As a measure of “dispersion” of patent citations, GENERAL, is inherently
biased —the more citations a patent receives, the greater the likelihood that
these citations will be spread across a larger number of patent classes. More¬
over, this bias is greater for patents with a small number of citations. There¬
fore, we employ the following correction (Hall, 2002):

g, = ([jtzj)general

where N, is the number of forward citations to patent i during the first six years
following the issue of the patent. All reported results for generality incorporate
this correction for bias.
The data in Table 6.6 indicate that the mean generality measures for over¬
all UC, Stanford, and Columbia patents are higher than those for their re¬
spective control sample patents, with the exception of UC patents applied for
and issued before 1981.28 There is no evidence in Table 6.6 of any decline in
generality, relative to the control sample of nonacademic patents, in the UC
and Stanford patents in the post-Bayh-Dole era. The differences in mean gen¬
erality between the overall UC and Stanford patents and their respective con¬
trol samples are statistically significant (at the 5 percent level) for the post-1981
period. The mean generality score of the post-1981 Columbia patents also is
significantly higher (at the 5 percent level of significance) than the similar
score for the patents in its control sample.
Overall, the results of this analysis of the importance and generality of Stan¬
ford, UC, and Columbia patents yield conclusions that differ from those of
Henderson, Jaffe, and Trajtenberg (1995^ x99^a)’ w^° analyzed a larger sample
of U.S. university patents. Importantly, these scholars also found a decline in
the importance and generality of the patents assigned to the universities that
PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

126 occupied the top decile of the post-1980 patent distribution (using patents
from 1988), a group that included the University of California and Stanford as
well as other U.S. universities with considerable pre-1980 patenting experi¬
ence. Why do we find little or no evidence of declines in the importance or
generality of these universities’ patents after the Bayh-Dole Act? First, our
sample of university patents is small, although this should limit the statistical
significance of any differences between the university and nonacademic pat¬
ent samples. Another possible explanation for the different results is the use
by Henderson and colleagues of a shorter time series of forward citations than
we use in this analysis. Their results thus may be affected more heavily by
“truncation bias.”29 Our empirical results also ignore the effects of entry by
other universities, other than Columbia, into patenting and licensing after the
passage of the Bayh-Dole Act on the characteristics of the overall U.S. aca¬
demic patent portfolio. Although Henderson, Jaffe, and Trajtenberg (1998a)
acknowledge that entry into patenting by less experienced institutions could
account for their results, they are not able to test for the effects of entry after
1980 by inexperienced universities on overall academic patent quality. We un¬
dertake such a test in the next chapter.

CONCLUSION

The effects of the Bayh-Dole Act on U.S. research universities have received
extensive rhetorical attention but modest empirical analysis. In this chapter,
we have used an invaluable byproduct of the Act, the systematic records of the
faculty inventions, patents, and licenses compiled by three leading U.S. re¬
search universities, in an analysis of some of the Acts effects at these institu¬
tions. Our data on the University of California and Stanford University suggest
that for universities already active in patenting and licensing, Bayh-Dole re¬
sulted in expanded efforts to market academic inventions. The Act also led
Columbia University, along with many other research universities formerly in¬
active in this area, to enter into large-scale patenting and licensing of faculty
inventions.
But several additional factors in addition to Bayh-Dole stimulated the post-
1980 upsurge in patenting and licensing at U.S. research universities, and it is
difficult to separate their effects from those of the Act. These additional factors
were especially influential in biomedical research. In particular, by the mid-
1970s, biomedical technology, especially biotechnology, had increased signifi¬
cantly in importance as a productive field of university research that yielded
research findings of great interest to industry. The feasibility of technology li¬
censing in biotechnology was advanced by the Diamond v. Chakrabarty deci-
 PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

sion, and broader shifts in U.S. intellectual property rights policy increased 127
the economic value of patents and facilitated patent licensing. The expanded
patenting and licensing activities of a number of U.S. universities during the
1970s provided an important political impetus for the drafting and passage of
the Bayh-Dole Act, as we pointed out in Chapter 5.
An array of developments in research, industry, and policy thus combined
to increase U.S. universities’ activities in technology licensing, and Bayh-
Dole, while important, was not determinative. Even without the Bayh-Dole
Act, the evidence presented in this chapter suggests that U.S. universities
would have increased their patenting and licensing activities during the 1980s
and 1990s. Indeed, our data indicate such growth before Bayh-Dole at the two
universities in our study active in patenting and licensing before its passage. At
Columbia, internal discussions of the feasibility of expanded institutional
patenting of faculty research results were well under way by 1980, and a key
patent application was filed before the passage of Bayh-Dole. The act never¬
theless hastened the entry into patenting and licensing by many universities
(such as Columbia) that had formerly avoided these activities.
By the end of the first decade of Bayh-Dole, these three universities display
remarkable similarities in their patent and licensing portfolios, as illustrated
by the data in Tables 6.1 and 6.2 and Figure 6.7. A very small share of all
patented inventions account for the majority of gross licensing revenues at all
three universities. Moreover, these leading earners are concentrated in the
biomedical area, a technology field characterized by relatively strong patents
that are economically significant (Levin et ah, 1987). A second important area
of licensing at two of these three universities, however, is software, for which
formal patent protection is less important.
Our analysis of the effects of Bayh-Dole on the content of academic re¬
search and the importance of the patents assigned to the University of Cali¬
fornia and Stanford University produced no evidence of significant change in
the characteristics of the academic research disclosures received by these uni¬
versities’ technology licensing oEces. Nor did our analysis of the pre- and post-
Bayh-Dole patents assigned to these two universities reveal any decline in the
importance or generality of these universities’ patents after Bayh-Dole, in con¬
trast to the findings of Henderson, Jaffe, and Trajtenberg (1995, 1998a) for a
larger sample of U.S. academic patents. The patents of a major post-1980 en¬
trant,” Columbia University, also are no less important or general than those
within a matched sample of nonacademic patents.
How do we reconcile a finding that citation-based measures of UC and
Stanford patents reveal no decline in importance after Bayh-Dole with our
conclusion that both the University of California’s and Stanford’s technology
PATENTING AND LICENSING AT U.C., STANFORD, AND COLUMBIA

128 licensing operations appear to have experienced a decline in “yield,” that is, a
decline in the share of licenses yielding positive revenues? Fundamentally,
these two sets of indicators measure different characteristics of the invention
and patent portfolios of these universities. Along with other scholars, we in¬
terpret patent citations as measures of the importance of the contribution to
inventive knowledge of a given patent. But this contribution may or may not
be correlated with the willingness of an industrial firm to pay for a license for
this patent. The extent of correlation between licensing revenues and patent
citations remains an important question for future research.
A number of observers have expressed concern that the upsurge of patent¬
ing and licensing has been associated with significant change in the character
of university research toward applied questions and away from basic research.
As we noted in Chapter 3, however, the shift in U.S. universities’ patenting ac¬
tivity toward biomedical research on its face cannot be interpreted as a shift
away from fundamental research — after all, much of the research that gener¬
ated these biomedical patents is fundamental in nature. In addition, the ef¬
fects of any change in university research “culture” and norms wrought by
Bayh-Dole appear to be limited to relatively few units of these research uni¬
versities. Increased post-1980 patenting and licensing activity has affected a
relatively small number of departments and institutes at Columbia University,
the only one of our three universities for which we have data on the depart¬
mental origins of invention disclosures. Were they available, similar data from
Stanford and the UC system would likely support a similar conclusion.
 7

The Effects of Entry and Experience

on U.S. University Patenting Before

and After the Bayh-Dole Act

The preceding chapters analysis of the effects of the Bayh-Dole Act on the
patenting and licensing activities of the University of California (UC), Stan¬
ford University, and Columbia University found that the “importance” and
“generality” of the post-Bayh-Dole patents of our two universities with consid¬
erable pre-1980 experience in patenting (Stanford University and the Univer¬
sity of California) was stable after the Bayh-Dole Act, a conclusion that con¬
trasts with the results of Henderson, Jaffe, and Trajtenberg (1995,1998a). These
different conclusions from similar analyses of university patenting motivate
this chapters analysis, which examines the effects on overall U.S. university
patenting of expanded patenting by less experienced academic institutions af¬
ter the passage of the Bayh-Dole Act in 1980.
Although Henderson, Jaffe, and Trajtenberg (1998a) suggest that increased
patenting after 1980 by smaller institutions may have been partly responsible
for declines after 1980 in the importance and generality of overall U.S. aca¬
demic patents, they are not able to control explicitly for the effects of expanded
patenting by such institutions after 1980. And one of the most important char¬
acteristics of U.S. university patenting during the 1980s was precisely the en¬
try into these activities of many universities that had been inactive or mini¬
mally active patenters before 1980.
Since Chapter 6 suggests that Bayh-Dole had modest effects on the patent¬
ing activities of two leading “incumbent” institutions, this analysis of the
effects of entry on overall U.S. university patenting enables us to consider
an area in which the effects of Bayh-Dole may be more pronounced. This
chapter also extends our analysis of the importance and generality of the
patents assigned to experienced university patenters before and after 1980
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

130 beyond the two research universities discussed in Chapter 6 to a broader

sample of academic patenters.
Our analysis of this issue compares the importance and generality of pat¬
ents assigned to universities with varying levels of experience in patenting be¬
fore 1980. Parallel declines in the importance and generality of the post-1980
patents of both “incumbent” and “entrant” universities could suggest that the
Bayh-Dole Act affected the incentives of academic researchers and adminis¬
trators to disclose and patent inventions of lower importance and generality in
all U.S. academic institutions, regardless of their pre-1980 experience in
patenting. But if the importance and generality of entrants’ patents are signifi¬
cantly lower than those of more experienced institutional patenters after 1980,
a different interpretation of the effects of Bayh-Dole is plausible. For example,
inexperienced university patent offices may seek patents on a broad cross sec¬
tion of faculty discoveries and thereby accumulate a portfolio of less signifi¬
cant patents (some anecdotal evidence supports this characterization). But as
these entrants acquire experience in patenting and learn the complexities of
protecting and marketing intellectual property, they may become more selec¬
tive in their patenting, reducing the gaps between the characteristics of their
patents and those of the experienced patenters.
The first interpretation of Bayh-Dole’s effects emphasizes lasting changes
in incentives and behavior throughout U.S. universities, while the second one
views the 1980s as a period of learning and adjustment by entrants inexperi¬
enced in patenting and licensing. Needless to say, these explanations are not
mutually exclusive, and the development of U.S. academic patenting during
the 1980s may reflect both effects. But our empirical analysis allows for an as¬
sessment of the relative strength of these two effects.
Immediately below, we conduct a statistical examination of the character¬
istics of the patents assigned to a comprehensive sample of U.S. universities in
order to assess the effects on the overall importance and generality of U.S. aca¬
demic patents of the entry of less experienced academic patenters after 1980.
Our findings indicate that less experienced academic patenters did indeed re¬
ceive less important or general patents in the aftermath of Bayh-Dole. The ini¬
tial effects of the Bayh-Dole Act on entry thus may underpin any observed
change in the importance and generality of overall U.S. academic patents af¬
ter 1980.
Consistent with the second interpretation of the effects of Bayh-Dole on
patenting by U.S. universities, we find that the importance of entrant institu¬
tions’ patents improved during the 1980s and 1990s, closing the gap with in¬
cumbents during a period in which the average importance of overall aca-
 EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

demic patents improved relative to nonacademic patents. The second part of 131
this chapter examines such “university learning” in greater detail, seeking to
understand whether and why the importance (based on citations to these pat¬
ents) of the post-1980 patents issuing to less experienced academic patenters
has improved during the 1980s and 1990s.
We find little evidence of strong “learning curve” effects, as neither cumu¬
lative patenting nor the (relatively) early establishment of a technology trans¬
fer office explains these improvements. Links with the Research Corporation
during the pre-Bayh-Dole era also exercise little influence over changes dur¬
ing the 1980s and 1990s in these characteristics of incumbent and entrant in¬
stitutions’ patents. Inasmuch as these observable sources of learning exercise
little influence, we conclude that a broader process of learning, based on
spillovers among universities and the strengthening of professional networks
among university licensing officers, may account for the convergence in im¬
portance between the patents of incumbent and entrant universities.

ENTRY, IMPORTANCE, AND GENERALITY

IN U.S. ACADEMIC PATENTING, 1 975~92

We first examine the characteristics of all U.S. university patents during 1975—
92 in an analysis of the effects on academic patenting of entry into patenting
activities after 1980 by universities with limited patenting experience. As in
Chapter 6, we use forward citations to U.S. university patents, defined as the
number of citations received by each patent following its issue. We include
only citations to patents that occurred within six years of the year of issue of
the patent and examine patents issued between 1975 and 1992. We also con¬
structed a control group of patents that included one nonacademic patent for
each patent in our academic patent database. Each nonacademic control pat¬
ent was chosen so as to match as closely as possible the corresponding aca¬
demic patents U.S. Patent and Trademark Office technology class (at the
three-digit level) and its date of application.1 Citations to patents in the con¬
trol sample were restricted to those occurring within six years after the year
of issue.
Our academic patent dataset includes all patents assigned to U.S. universi¬
ties other than Stanford, the University of California, and Columbia during
1975-92. Within this dataset, we distinguished among three categories: (1) uni¬
versities with at least ten patents that were applied for after 1970 and issued
during 1975-80 (“high-intensity” incumbents); (2) universities with at least
one but less than ten patents applied for after 1970 that issued during 1975-80
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

132 (“low-intensity” incumbents); and (3) universities with no patents issued dur¬
ing the 1975-80 period and at least one patent issued during 1980-92 that was
applied for after 1980 (“entrants”). Our definitions of “entrant” and “incum¬
bent” universities are intended to separate the effects on patent importance
and generality of increased patenting after Bayh-Dole by active pre-1980
patenters from increased post-1980 patenting by other institutions.
Table 7.1 contains information on the annual number of patents issuing to
these three groups of universities during the 1975-92 period that were applied
for after 1970, demonstrating the shrinking share of the “high-intensity” pre-
1980 academic patenters after the passage of Bayh-Dole. The “high-intensity”
patenters’ share declines from more than 85 percent during 1975-80 to less
than 65 percent by 1992. The “low-intensity” pre-1980 patenters, by contrast,
increase their share of all academic patents from 15 percent in 1981 to almost
30 percent in 1992. And entrants increase their share of overall academic
patenting from zero in 1980 to more than 6 percent by 1992.
Tables 7.2-4 present the results of separate regressions of importance and
generality for the patents assigned to academic institutions in each of these
three categories, covering 1975-91 for the “high-” and “low-intensity” incum¬
bents and 1981-91 for the entrants.2 Negative binomial regression results are
reported for models in which importance is the dependent variable. Our
analysis of generality uses a tobit specification, since the dependent variable
(generality) is truncated at a lower limit of zero and an upper limit of one.3
Each specification is estimated for a dataset covering the patents of the rele¬
vant academic institutions and those in the control sample of nonacademic
patents. We control for year effects and interact a dummy variable denoting
academic patents with a dummy for the application year —the reported coef¬
ficients are those for the interaction terms for 1975-91. We present results for
overall academic patents as well as for biomedical and nonbiomedical patents.
The results for our analysis of each of these three samples of academic pat¬
ents display contrasting patterns of importance and generality.4 The “high-
intensity” incumbents exhibit consistently more important and more general
patents, relative to nonacademic patents, throughout the 1975-91 period than
do the other two groups of universities (Table 7.2). Thirteen of the seventeen
interaction coefficients in column 2 for the overall patent sample associated
with this group of universities are positive and statistically significant at the
5 percent level, indicating that for most of this period (and increasingly so dur¬
ing the 1980s), these institutions’ patents were cited more intensively than
those in the matched industrial sample. As was true in Chapter 6’s analysis of
UC, Stanford, and Columbia patents, the differences in importance between
 EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

TABLE 7.1 133

Number of Patents by Year of Issue: “High-intensity incumbents,” “Low-intensity
incumbents,” and “Entrants,” 1975-92

High-intensity Low-intensity Entrants

incumbents (n = 51) incumbents (n = 92) (n = 81)
Patent
issue %of % of % of
year Number total Number total Number total Total

1975 213 86.9 32 13.1 0 0.0 245

1976 248 89.2 30 10.8 0 0.0 94
1977 243 88.4 32 11.6 0 0.0 239
1978 262 89.7 30 10.3 0 0.0 280
1979 193 86.9 29 13.1 0 0.0 216
1980 295 87.8 41 12.2 0 0.0 332
1981 291 84.6 53 15.4 0 0.0 344
1982 319 88.1 42 11.6 1 0.3 361
1983 259 80.9 58 18.1 3 0.9 320
1984 320 76.2 85 20.2 15 3.6 419
1985 332 71.4 111 23.9 22 4.7 465
1986 384 71.0 130 24.0 27 5.0 541
1987 477 72.4 154 23.4 28 4.2 659
1988 496 74.8 138 20.8 29 4.4 663
1989 743 68.4 281 25.9 62 5.7 1,086
1990 738 68.2 282 26.1 62 5.7 1,082
1991 819 67.2 307 25.2 92 7.6 1,218
1992 932 64.0 431 29.6 93 6.4 1,456

Total
patents 7,564 2,266 434 10,264

NOTES: High-intensity incumbents have 10 or more issued patents applied for prior to 1981 and after 1970;
low-intensity incumbents have fewer than 10 issued patents applied for prior to 1981 and after 1970; and
entrants have no issued patents applied for prior to 1981 and after 1970.

the patents of these experienced academic patenters and the nonacademic

patent control sample are most significant for the nonbiomedical patents.
A broadly similar pattern appears in the results of our specifications com¬
paring the generality of the entire sample of patents accounted for by these
“high-intensity” incumbents with those in the control sample —all of the in¬
teraction coefficients are positive wherever significant, and ten of the seven¬
teen positive interaction coefficients are statistically significant at the 5 per¬
cent level. Once again, these differences in generality are more frequently
significant for the nonbiomedical patents.
The results of our regressions for the “low-intensity” pre-1981 patenters
(Table 7.3) reveal only five years of patents that are more intensively cited (at
the 5 percent significance level) than those in the nonacademic control
sample. The results of the regressions for generality of their overall patents for
 TABLE 7.2

Regression Coefficients for High-Intensity Incumbent X Application Year

Negative binomial model Tobit model

(dependent variable: Importance) (dependent variable: Generality)

Year Overall Biomed Nonbiomed Overall Biomed Nonbiomed

1975 0.36* * 0.27 0.39** 0.10 0.14 0.08

(0.11) (0.21) (0.13) (0.13) (0.27) (0.16)
1976 0.13 0.29 0.07 0.39** 0.28 0.42**
(0.11) (0.22) (0.12) (0.13) (0.27) (0.15)
1977 0.19* 0.19 0.19 0.15 0.21 0.13
(0.11) (0.19) (0.13) (0.12) (0.23) (0.04)
1978 0.19* 0.41** 0.09 0.30‘* 0.20 0.35**
(0.11) (0.19) (0.13) (0.13) (0.23) (0.15)
1979 0.27** 0.01 0.40** 0.17 0.38** 0.05
(0.10) (0.17) (0.12) (0.11) (0.19) (0.13)
1980 0.32** 0.20 0.38** 0.23** 0.19 0.23**
(0.09) (0.15) (0.11) (0.11) (0.18) (0.13)
1981 0.38** 0.26* 0.45** 0.45** 0.63** 0.37**
(0.09) (0.15) (0.12) (O.H) (0.18) (0.13)
1982 0.20** -0.12 0.33** 0.12 -0.51** 0.52**
(0.09) (0.15) (0.12) (0.10) (0.17) (0.13)
1983 0.27** 0.17 0.35** 0.17* 0.00 0.30**
(0.09) (0.14) (0.12) (0.10) (0.15) (0.14)
1984 0.32** 0.04 0.54** 0.15* 0.22* 0.09
(0.08) (0.12) (0.11) (0.09) (0.13) (0.12)
1985 0.31** 0.02 0.54** 0.35** 0.22** 0.43**
(0.08) (0.12) (0.10) (0.09) (0.13) (0.12)
1986 0.37** 0.09 0.57** 0.28** 0.15 0.39**
(0.08) (0.11) (0.10) (0.08) (0.13) (0.11)
1987 0.20** 0.03 0.31** 0.22** 0.25** 0.19**
(0.06) (0.10) (0.08) (0.07) (0.11) (0.09)
1988 0.15** -0.09 0.29** 0.25** 0.21** 0.27**
(0.06) (0.10) (0.08) (0.07) (O.H) (0.08)
1989 0.29** 0.24** 0.32** 0.17** 0.15 0.18**
(0.06) (0.10) (0.08) (0.07) (O.H) (0.08)
1990 0.28** 0.19* 0.32** 0.10 -0.01 0.15*
(0.07) (0.12) (0.08) (0.07) (0.13) (0.09)
1991 -0.06 -0.23 0.03 0.34** 0.67** 0.20
(0.10) (0.18) (0.12) (0.11) (0.20) (0.13)
Constant 2.10 2.54 1.73 0.22 -0.09 0.96
(0.19) (0.63) (0.46) (0.22) (0.90) (0.51)
Number of
observations 15125 5569 9556 11716 4313 7403
Log
likelihood -36170 -13280 -22836 -12468 -4456 -7939
Pseudo R2 0.02 0.01 0.02 0.01 0.01 0.01

NOTES: Standard errors in parentheses. Year dummies not reported. Sample excludes patents from the
University of California, Stanford, and Columbia. ** p < 0.05 * p < 0.10
 TABLE 7.3
Regression Coefficients for Low-Intensity Incumbent X Application Year

Negative binomial model Tobit model

(dependent variable: Importance) (dependent variable: Generality)

Year Overall Biomed Nonbiomed Overall Biomed Nonbiomed

1975 0.65** 0.38 0.71** -0.30 -1.01 -0.15

(0.32) (0.77) (0.34) (0.35) (0.84) (0.38)
1976 -0.11 -0.49 0.24 0.84** 1.85** 0.23
(0.33) (0.55) (0.42) (0.39) (0.69) (0.49)
1977 0.63** 1.69** 0.45 0.36 — 0.19
(0.30) (0.67) (0.34) (0.33) (0.37)
1978 0.01 0.18 -0.05 -0.29 -0.15 -0.35
(0.27) (0.47) (0.32) (0.29) (0.57) (0.33)
1979 0.66** 0.35 0.79** 0.62** 0.69 0.59*
(0.25) (0.45) (0.30) (0.29) (0.54) (0.34)
1980 0.35 0.13 0.51* -0.01 0.23 0.14
(0.24) (0.38) (0.31) (0.25) (0.41) (0.31)
1981 -0.12 -0.15 -0.10 0.31 1.01** 0.00
(0.20) (0.34) (0.26) (0.22) (0.40) (0.27)
1982 -0.06 -0.21 0.10 0.38** 0.35 0.42*
(0.18) (0.25) (0.25) (0.18) (0.25) (0.26)
1983 0.16 0.50* 0.03 0.18 0.63** 0.02
(0.15) (0.27) (0.18) (0.16) (0.32) (0.19)
1984 -0.18 -0.25 -0.11 0.04 -0.13 0.19
(0.15) (0.22) (0.20) • (0.16) (0.24) (0.22)
1985 -0.05 -0.04 -0.06 -0.07 -0.19 0.00
(0.14) (0.23) (0.18) (0.16) (0.25) (0.19)
1986 0.40** 0.38** 0.41** 0.41** 0.38** 0.43**
(0.13) (0.19) (0.17) (0.14) (0.20) (0.19)
1987 0.11 0.04 0.15 0.28** 0.06 0.40**
(0.11) (0.19) (0.14) (0.12) (0.20) (0.15)
1988 0.17* 0.02 0.24** 0.15 0.15 0.16
(0.10) (0.17) (0.12) (0.10) (0.18) (0.12)
1989 0.38** 0.22 0.46** 0.29** 0.26 0.29**
(0.10) (0.17) (0.11) (0.10) (0.18) (0.12)
1990 0.15* 0.17 0.14 0.20** -0.05 0.29**
(0.09) (0.18) (0.11) (0.10) (0.18) (0.11)
1991 0.11 -0.19 0.28 0.53** 0.38 0.57**
(0.16) (0.30) (0.18) (0.16) (0.31) (0.19)

Constant 1.10 2.30 -16.41 0.69 0.67 0.20

(0.38) (1.13) (3661.49) (0.38) (1.01) (0.52)
Number of
Observations 4535 1627 2908 3528 1242 2286

Log
Likelihood -10980 -3893 -7056 -3745 -1285 -2436

Pseudo R2 0.01 0.01 0.01 0.01 0.02 0.01

NOTES: Standard errors in parentheses. Year dummies not reported. Sample excludes patents from the University of
California, Stanford, and Columbia. ** p < 0.05 * p < 0.10
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

136 this group of institutions yield positive and statistically significant (at the
5 percent significance level) year-coefficients for eight of the seventeen years.
Finally, the results for 1981-91 for the entrant population of institutions also in¬
dicate lower levels of importance (only one year interaction is positive and sta¬
tistically significant at the 5 percent level in Table 7.4) and generality (also one
significant interaction coefficient), relative to nonacademic patents, than
those accounted for by the “high-intensity” pre-1981 patenters.
These results broadly corroborate our findings in Chapter 6 concerning the

TABLE 7.4
Regression Coefficients for Entrant University X Application Year

Negative binomial model Tobit model

(dependent variable: Importance) (dependent variable: Generality)

Year Overall Biomed Nonbiomed Overall Biomed Nonbiomed

1981 0.39 0.51 0.18 0.38 0.35

(0.51) (0.81) (0.68) (0.57) (0.78)
1982 0.80 1.56* 0.11 -0.81 -0.42 -1.09
(0.56) (0.86) (0.72) (0.54) (0.77) (0.74)
1983 0.18 0.53 -0.20 -0.53 0.18 -0.95**
(0.36) (0.60) (0.44) (0.34) (0.52) (0.44)
1984 -0.19 -0.28 -0.03 -0.11 0.00 -0.22
(0.32) (0.44) (0.45) (0.32) (0.39) (0.49)
1985 -0.03 -0.63 0.11 0.06 -0.72 0.33
(0.35) (0.65) (0.40) (0.35) (0.61) (0.45)
1986 0.23 0.12 0.29 0.28 0.49 0.13
(0.29) (0.48) (0.35) (0.28) (0.42) (0.37)
1987 -0.16 -0.62** 0.50* 0.35 0.33 0.37
(0.23) (0.31) (0.31) (0.22) (0.27) (0.33)
1988 0.19 -0.20 0.46* 0.37* 0.30 0.43
(0.23) (0.34) (0.29) (0.21) (0.31) (0.28)
1989 0.51** 0.11 0.85** 0.51** 0.56** 0.44
(0.19) (0.26) (0.25) (0.18) (0.3) (0.27)
1990 0.26 0.10 0.47** 0.19 0.27 0.10
(0.20) (0.28) (0.28) (0.20) (0.26) (0.29)
1991 0.32 0.53 0.24 0.38 0.31 0.42
(0.32) (0.55) (0.38) (0.20) (0.50) (0.38)
Constant 1.70 1.32 0.25 -0.52 1.36 1.46
(0.83) (0.59) (0.52) (0.42) (0.67) (0.80)
Number of
Observations 868 367 501 666 284 382
Log
Likelihood -2105 -919 -1163 -700 -290 -399
Pseudo R2 0.01 0.02 0.02 0.03 0.04 0.04

NOTES: Standard errors in parentheses. Year dummies not reported. Sample excludes patents from the
University of California, Stanford, and Columbia. p<o.o5 * pco.io
 EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

importance and generality of post-1980 patents issuing to the University of 137

California and Stanford University. Taken together, these results indicate that
any deterioration in the importance and quality of overall U.S. academic pat¬
ents after 1980 may have resulted from the Bayh-Dole Acts encouragement of
entry into patenting by academic institutions with relatively little experience
in this activity.5

INSTITUTIONAL EXPERIENCE, LEARNING, AND THE

CHARACTERISTICS OF UNIVERSITY PATENTS, 1981-92

Why might universities with less experience in patenting and licensing tend
to receive patents that are less heavily cited? Conversations with licensing of¬
ficers at several research universities and other field research suggested the
possibility that inexperienced universities adopted an indiscriminate policy
toward patenting as they entered into this activity after passage of the Bayh-
Dole Act. In this view, many entrant universities sought patents for faculty in¬
ventions with little evaluation of the market for licenses for these inventions.
As these entrant institutions encountered rising costs and stagnant licensing
revenues, many of them became more selective in their patenting activities.
Other research showing a correlation between licensing revenues and cita¬
tions for licensed university patents (Sampat and Ziedonis, 2003, and the work
cited therein discuss the links between citations and other measures of the pri¬
vate economic value of patents) suggests that patents that are more heavily
cited are more likely to yield positive licensing revenues.
Accordingly, the analysis in this section tests the possibility that entrant uni¬
versities, possibly motivated by the prospect of higher licensing revenues, grad¬
ually became more selective in their patenting activities, effectively “learning
to patent.” Any such increase in the selectivity of their patenting should be re¬
vealed in shifts in the patent portfolios of entrant universities to include more
heavily cited patents. Our analysis in this section examines post-Bayh-Dole
patents issued to all institutions designated as Research or Doctoral Universi¬
ties in the Carnegie Commission on Higher Educations 1973 and 1993 reports
(Carnegie Commission on Higher Education, 1973, 1993). Specifically, our
dataset includes the 10,881 patents issued to these universities that were ap¬
plied for between 1981 and 1992 and granted before 1994. We compare these
patents to one another and to a matched control sample of patents, con¬
structed as in the previous section.6 Our analysis tests for any narrowing dur¬
ing the 1981-92 period in the gaps in importance between the patents of more
and less experienced pre-1980 academic patenters.
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

138 We construct measures of the "importance” of these patents based on

counts of citations to these patents appearing within five years of their issue
dates. Limiting the citation window to five years (rather than six years, as in the
previous section) enables us to increase the number of “patent cohorts” in our
analysis, which is desirable since “learning to patent” presumably takes time.7
We also exclude “self-citations” (that is, citations by an assignee of its previous
patents), which may be less accurate indicators of the importance of a patent
than citations by others.

Specification
We first compare the number of citations to academic patents with the cita¬
tions to patents in our nonacademic control sample. The count nature of
the dependent variable makes it necessary to utilize a negative binomial spe¬
cification in this and related analyses.8 Our base specification includes
dummy variables for application year cohorts (1981-83, 1984-86, 1987-89,
and 1990-92), university patents applied for in these years (L/N7V8l8?,
UNIVs486, UNIV^, and UNIVgo92), and patent classes. The patent class
dummy variables span the 303 classes in our sample.
The next step in our analysis is a comparison of the importance of aca¬
demic patents with patents in our nonacademic control sample. Next, we test
for differences in average citation rates of entrant universities’ patents with
those issued to incumbent universities and with our nonacademic control pat¬
ents by inserting additional dummy variables denoting patents assigned to
entrant universities in these application-year cohorts (ENT8l8y ENTS4S6,
ENT8?g9, and ENTgoi)2). Finally, to assess whether our various “learning mech¬
anisms” influence the characteristics of patents issued to entrant universities,
we identify three characteristics of entrant universities: (1) the existence of a
contractual relationship with the Research Corporation, (2) above-median lev¬
els of cumulative 1980-86 patenting, and (3) the assignment of at least a single
half-time employee (0.5 FTE) to formal technology transfer activities during
1980-86. We also include a set of dummy variables indicating patents as¬
signed to experienced entrants in each application-year cohort (EXPENT8l8 ,
EXPENT8486, EXPENT8?89, and EXPENT90g2) to compare the number of cita¬
tions to the patents of entrant universities with each of these characteristics
with those of patents assigned to entrants lacking these characteristics. A more
detailed discussion of our econometric specifications is contained in the ap¬
pendix to this chapter.
lable 7.5 reports results of negative binomial regressions for the full sample
of university and control group patents. Table 7.6 reports similar results
 EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

for biomedical patents only, and Table 7.7 reports results for nonbiomedical 139
patents. In each table, Model 1 represents the base specification, which in¬
cludes application year and class dummies, as well as the UNIV dummy
variables (UNIV8l83~UNIV9092). Model 2 adds the ENT dummy variables,
and Models 3-5 separately add each of the sets of EXPENT dummy
variables. Table 7.8 reports results from Wald tests of entrant university pat¬
ent importance relative to control group patent importance for Model 2 (see
the appendix for the description of these tests). For expositional clarity, we
do not report coefficients for the application year and patent class dummy
variables.
We first compare the importance of patents assigned to incumbent and en¬
trant universities during the early post-Bayh-Dole period (1981-86) and sepa¬
rately examine biomedical and nonbiomedical patents from these universities
during this period. We then analyze changes in the importance of patents is¬
sued to entrant universities during the 1980s and consider sources of learning
that could explain observed improvements.

Do Post-1980 Entrants Have Less Important

Patents Than Incumbents During 1981-86?
We assess incumbent-entrant differentials in patent importance by examining
the coefficients for ENT reported in Model 2 of Table 7.5. In this section, we
compare the importance of incumbent-university patents with that of entrant-
university patents for patents applied for during 1981-86. The negative and
significant coefficients in Table 7.5 for 1981-83 and 1984-86 suggest that en¬
trant universities received less important patents than incumbent universities
in the early 1980s. The point estimates of the ratio of citations to entrants pat¬
ents to citations to incumbents’ patents are 0.82 and 0.80 in 1981-83 and 1984-
86, respectively, indicating that incumbent-university patents applied for dur¬
ing 1981-83 and 1984-86 received 20 percent and 25 percent more citations
than entrant-university patents applied for during these periods.9
Biomedical patent results are reported in Table 7.6, and nonbiomedical
patent results are reported in Table 7.7. The coefficients in Model 2 of
Table 7.6 indicate that entrants’ biomedical patents did not receive a signifi¬
cantly different number of citations than incumbents’ patents applied for dur¬
ing 1981-83 and 1984-86.10 Thus, for biomedical technologies, there is little
evidence that incumbents received more important patents than entrants in
the first six years following Bayh-Dole.
The analysis of nonbiomedical patents, however, reveals significant dif¬
ferences in importance between entrant- and incumbent-university patents
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

140 TABLE 7.5

Regression Coefficients for All Patents (1981-92 Sample):

Negative Binomial Model, Dependent Variable = Number of Citations

Model 1 Model 2 Model 3 Model 4 Model 5

Add EXPENTt to Model 2

For entrants
For entrants with above For entrants
who were median with early
Add ENTt active RC patenting in establishment
Variables Base model to base model clients 1981-86 ofTLOs

UNIVna 0.237 0.262 0.262

(5.31)** (5.66)** (5.66)**
UNIV8m 0.207 0.246 0.246
(5.56)** (6.25)** (6.26)**
univ8789 0.193 0.188 0.188 0.189 0.189
(6.78)** .(6.13)** (6.13)** (6.14)** (6.15)**
univ9092 0.261 0.249 0.249 0.255 0.255
(9.98)** (8.73)** (8.73)** (8.94)** (8.94)**
EN I 8I83 -0.190 -0.095
(2.07)* (0.46)
EN 1 8486 -0.218 -0.166
(3.19)** (1.22)
EN 1 8789 0.022 -0.161 0.104 0.172
(0.47) (1.55) (1.63) (2.83)**
EN 1 9092 0.044 0.134 0.110 0.021
(1.07) (1.26) (2.02)* (0.42)
EXPEN 1 8183 -0.116
(0.52)
EXPENTm6 -0.068
(0.45)
expent8789 0.217 -0.183 -0.368
(1.95) (2.22)* (4.46)**
EXPEN 1 9092 0.102 -0.157 0.038
(0.91) (2.20)* (0.52)
Constant 0.834 0.818 0.806 1.384 1.351
(3.20) (1.47) (1.50) (1.82) (1.35)
N 21,455 21,455 21,455 14,554 14,554

NOTES: Absolute value ofz-statistics in parentheses. Application year dummy variables and patent class
dummy variables included in all specifications (not reported). Coefficient for UNIVr is fa; coefficient for
ENTr is fa; coefficient for EXPENTt is <t>v (see appendix). * significant at 5% level "“significant at 1% level

applied for during the 1984—86 period. In Table 7.7, entrants’ patents re¬
ceive significantly fewer citations than incumbents’ patents in each of the
two cohorts of patents applied for before 1987 (see below for more discussion
of improvements in the importance of entrants’ patents during the later
1980s).11
The fact that incumbents enjoy a greater “importance advantage” for
 EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

TABLE 7.6 141

Regression Coefficients for Biomedical Patents (1981-92 Sample):
Negative Binomial Model, Dependent Variable = Number of Citations

Model 1 Model 2 Model 3 Model 4 Model 5

Add EXPENTt to Model 2

For entrants
For entrants with above For entrants
who were median with early
Add ENTt active RC patenting in establishment
Variables Base model to base model clients 1981-86 ofTLOs

UNIV8m 0.175 0.153 0.153

(2.24)* (1.86) (1.87)
UNIV8m 0.070 0.092 0.092
(1.14) (1.40) (1.40)
univ8789 -0.004 -0.007 -0.007 -0.014 -0.014
(0.07) (0.13) (0.13) (0.25) (0.26)
univ9092 0.220 0.193 0.194 0.195 0.195
(4.69)** (3.82)** (3.82)** (3.80)** (3.79)’*
ENT8i88 0.131 0.452
(0.88) (1.41)
ent8486 -0.105 0.091
(0.98) (0.47)
ent8789 0.015 -0.079 0.033 0.107
(0.17) (0.43) (0.27) (0.85)
ENT9092 0.102 0.226 0.215 0.097
(1.33) (1.30) (2.02)* (0.91)
EXPENT8m -0.403
(1.16)
EXPENTm6 -0.270
(1.24)
EXPENT8789 0.112 -0.028 -0.152
(0.58) (0.18)* (0.99)
expent9092 -0.148 -0.194 0.028
(0.81) (1.44) (0.21)

Constant 0.920 0.925 0.924 1.135 1.133

(4.14) (3.01) (4.15) (3.85) (3.85)

N 7,383 7,383 7,383 4,823 4,823

NOTES: Absolute value of z-statistics in parentheses. Application year dummy variables and patent class
dummy variables included in all specifications (not reported). Coefficient for UNJVT is fo coefficient for
ENTt is &r; coefficient for EXPENTt is cfo- (see appendix). * significant at 5% level ** significant at 1% level

nonbiomedical technologies than biomedical technologies during the first six

years after Bayh-Dole may indicate that patenting experience is less important
in biomedical fields. This hypothesis receives modest support in Chapter 4,
which argued that Research Corporation client institutions (all of whom had
little experience in patenting) were more successful at “cherry-picking” bio¬
medical patents than nonbiomedical patents. These results also are consistent
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

142 TABLE 7.7

Regression Coefficients for Nonbiomedical Patents (1981-92 Sample):
Negative Binomial Model, Dependent Variable = Number of Citations

Model 1 Model 2 Model 3 Model 4 Model 5

Add EXPENTt to Model 2

For entrants
For entrants with above For entrants
who were median with early
Add ENTt active RC .patenting in establishment
Variables Base model to base model clients 1981-86 ofTLOs

UNIV8m 0.268 0.315 0.315

(4.97)’* (5.64)** (5.65)**
UNIVm6 0.290 0.333 0.333
(6.21)** (6.81)** (6.82)**
univ8789 0.286 0.280 0.280 0.282 0.282
(8.39)** . (7.64)** (7.65)** (7.73)** (7.74)**
UNIVgogz 0.286 0.279 0.279 0.288 0.288
(9.10)** (8.12)** (8.12)** (8.41)** (8.41)**
E1VI8183 -0.418 -0.596
(3.54)** (2.11)*
EN 1 8486 -0.274 -0.431
(3.06)** (2.19)*
EN J 8789 0.024 -0.198 0.114 0.163
(0.43) (1.56) (1.54) (2.37)**
t,N 1 9092 0.024 0.042 0.076 0.005
(0.48) (0.31) (1.22) (0.08)
EXPENT8m 0.211
(0.69)
expent8486 0.189
(0.88)
EXPENTmg 0.261 -0.224 -0.421
(1.93) (2.30)* (4.23)**
EXPENTgog2 -0.019 -0.160 0.002
(0.14) (1.91)* (0.03)
Constant 0.869 0.848 0.839 1.36 1.32
(3.39) (3.31) (3.27) (4.02) (3.90)
N 14,072 14,072 14,072 9,731 9,731

NOTES: Absolute value of /-statistics in parentheses. Application year dummy variables and patent class
dummy variables included in all specifications (not reported). Coefficient for UNIVT is fa; coefficient for
ENTt is fa; coefficient for EXPENTt is tfa- (see appendix). * significant at 5% level *" significant at 1% level

with Chapter 6 s analysis of citations to UC and Stanford biomedical and non-

biomedical patents before and After Bayh-Dole.

Does the Importance of Entrant-University

Patents Improve After 1986?

Our analysis thus far suggests that post-Bayh-Dole entrants into academic
patenting received less important patents, on average, during the 1981-86 pe-
 EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

riod than did more experienced incumbents. If their lack of institutional ex¬ M3
perience in patenting is responsible for the lower importance of entrant uni¬
versities’ patents, then expanded patenting by these universities could result in
institutional learning that narrowed the differentials between more and less
experienced institutional patenters. Learning by entrants could also improve
the importance of entrants’ patents relative to the control patents. Accord¬
ingly, in this section we test for changes during the 1980s and early 1990s in the
“importance differential” between entrants’ and incumbents’ patents and be¬
tween entrants’ and control-groups’ patents.
Within our overall sample (Table 7.5), the importance of entrants’ patents
was significantly lower than the importance of incumbents’ patents for patents
applied for during 1981-86 (Model 2), but these differentials are statistically
insignificant for patents applied for during 1987-92. Nonbiomedical patents,
where the incumbent-entrant importance differential was most pronounced
during the 1981-86 period, display similar changes in importance (Model 2 of
Table 7.7). These findings suggest that in nonbiomedical fields, where the ini¬
tial post-Bayh-Dole patent importance differences were most significant, en¬
trant institutions narrowed the gap in importance between their patents and
those of incumbent institutions for patents applied for during 1987-92.
We also examined changes in the “importance differentials” between en¬
trant university patents and control group patents and report these results in
Table 7.8. The results in Table 7.8 suggest that entrant-university patents were
not significantly more highly cited than control-group patents during the early
post-Bayh-Dole period (Wald test statistics are not significant for the patents
applied for in either the 1981-83 or 1984-86 period). Wald statistics also
are not significant in separate tests for differences in importance between

TABLE 7.8

(UNIVt + ENTt) Coefficients (1981-92 Sample): Negative Binomial

Model, Dependent Variable = Number of Citations

Full sample, Biomedical Nonbiomedical

Sample controls patents, controls patents, controls

1981-83 0.072 0.284 -0.103

(.62) (3.73) (0.77)
1984-86 0.028 -0.013 0.059
(.16) (.02) (0.43)
1987-89 0.21 0.008 0.304
(21.27)** (0.01) (31.34)**
1990-92 0.293 0.295 0.303
(54.05)** (16.04)** (41.10)**

NOTES: Coefficient for UNIVT is ft; coefficient for ENTt is ft Chi-square statistics for
Wald tests of (ft + ST = o) in parentheses (see appendix). * significant at 5% level
*4 significant at 1% level
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

144 entrants’ and incumbents’ biomedical patents and for these groups’ nonbio¬
medical patents. But entrant-university patents applied for during 1987-89
and 1990-92 were more highly cited on average than the corresponding con¬
trol group patents for both the overall sample and the nonbiomedical sub¬
sample. Entrant biomedical patents applied for during 1990-92 were also
more highly cited than the corresponding control group patents.
Overall, these data suggest that the “importance” of incumbent-university
and entrant-university patents applied for during the later 1980s and early
1990s converged to a considerable extent. Moreover, this convergence reflects
improvement in the importance of entrants’ patents, rather than declines in
the importance of incumbents’ patents.12 Convergence is most marked for
nonbiomedical technologies, where the differences between the importance
of incumbents' and entrants’ patents applied for during 1981-86 were greatest.
In most of our specifications, the importance of entrants’ patents also im¬
proved relative to that of the nonacademic controls in the later cohorts of pat¬
ents (those applied for after 1986). Taken together, these results suggest that en¬
trant universities do indeed seem to have learned to patent during the 1980s.

Sources of Entrant Learning

What factors influenced the ability of entrant universities to “learn to patent”

after 1980? A number of post-1980 entrants into patenting had been clients of
the Research Corporation prior to 1980, and by the 1960s the Research Cor¬
poration had begun to train its institutional clients to identify inventions that
were the best candidates for patenting and licensing (see Chapter 4). We
therefore tested the hypothesis that “active” clients of the Research Corpora¬
tion in the 1970s (that is, those that submitted five or more disclosures to the
Research Corporation in any one year during that decade) received more im¬
portant patents than other entrants during the first six years after the passage
of the Bayh-Dole Act.
Another potential source of improvement in the importance of entrants’
patents in the post-Bayh-Dole period is learning from cumulative institutional
experience in patenting. If the relevant learning results from higher volumes
of patenting activity, then entrants that patented more intensively during 1981—
86 might obtain more important patents, on average, after 1986 than would en¬
trant institutions with less patenting experience during the 1981-86 period. To
test this hypothesis, we analyzed differences in patent importance between en¬
trants receiving more than the median number of entrant-university patents
during the 1981—86 period and other entrants.
Finally, the establishment of a formal technology licensing office (TLO)
 EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

and the hiring of licensing professionals may have contributed to improve¬ M5

ments in the importance of entrant universities’ patents after 1980. We tested
the hypothesis that “early entrants” (institutions that established TLOs during
the first six years after the passage of the Bayh-Dole Act) acquired “higher-
importance” patents, on average, than entrants that were slower to establish
formal TLOs. We compared the importance of patents assigned to entrants
who reported that at least one 0.5 FTE was employed in technology transfer
activities prior to 1986 with the importance of patents assigned to entrant uni¬
versities that had not established such technology transfer offices by that date.
As we noted earlier, the coefficient for EXPENT measures the difference
in importance between patents issued to experienced entrants and those is¬
sued to less experienced entrants in a given period. Model 3 in Tables 7.5-7
reports the EXPENT variables’ coefficients for entrants who were active
clients of the Research Corporation before 1980 and for other entrants in each
period from 1981-83 to 1990-92. The existence of a contractual relationship
with the Research Corporation before 1980 is not associated with improve¬
ment in the importance of entrants’ patents during the 1980s. The EXPENT
coefficients in Column 3 for the overall sample are not significant in any pe¬
riod (Table 7.5). Similarly, the biomedical and nonbiomedical subsamples
(Tables 7.6 and 7.7) provide no evidence that entrants who were active clients
have more important patents than other entrants after Bayh-Dole.
The reasons for the Research Corporation’s lack of influence on improve¬
ment in client firms’ patents after 1980 are unclear. One reason could be the
lack of incentives for selectivity by client universities in their submission of in¬
ventions to the Research Corporation for patenting and licensing. Research
Corporation client universities bore no costs under Invention Administration
Agreements (LAAs) with the corporation (Chapter 4) and therefore lacked any
incentives to select commercially promising inventions during the 1960s and
1970s. It is also possible that their “outsourcing” of patent management activ¬
ities to the Research Corporation prevented client universities from acquiring
these skills.
A second source of learning discussed above is cumulative patenting
(“learning by doing”). In Model 4 of Tables 7.5-7, EXPENT is set equal to 1
for all entrants with an above-median number of patents within the entrant
population applied for during the 1981-86 period.13 EXPENT in Model 4
measures differences between the importance of patents applied for after 1986
by entrants that received an “above-median number of patents during 1981—
86 and the importance of patents applied for during these years by entrants
with a “below-median” number of patents for 1981-86. Surprisingly, the post-
1986 patents issued to entrants with greater patenting experience in the 1981-
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

146 86 period tended to receive fewer citations, on average, than did other en¬
trants’ patents.
For the full sample, patents issued to more experienced entrants during
1987-89 and 1990-92 received significantly fewer citations than did the pat¬
ents issued to less experienced entrants during these periods (Model 4 of
Table 7.5). Similarly, within the biomedical and nonbiomedical subsamples,
there is no evidence that entrants with higher levels of pre-1987 patenting had
more important patents after 1987, as the “learning curve” hypothesis would
imply. These findings suggest that cumulative experience in patenting does
not account for the improvements in importance of entrants’ patents observed
after the mid-1980s.
Finally, we examined the effects on the importance of entrants’ patents of
the establishment after 1980 of formal TLOs. Using Association of University
Technology Managers (AUTM, 1994) data, we coded EXPENT to equal 1 for
entrant universities that had established a formal TLO (that is, reported allo¬
cating at least one 0.5 FTE to the management of technology transfer, patent¬
ing, and licensing activities) by 1986. The results of our analysis of the effects
of establishing a TLO during 1981- 86 on the importance of entrants’ post-1986
patents are reported in Model 5 in Tables 7.5-7. Table 7.5 suggests that for the
overall sample of patents, early establishment of a TLO did not improve pat¬
ent importance relative to entrants that did not establish TLOs prior to 1987.
The EXPENTg7g9 coefficient is negative and significant in Table 7.5, and the
coefficient for EXPENT9Qg2 is not significant. For biomedical technologies, the
coefficient for EXPENT8y8g is not significant in Table 7.6. For nonbiomedical
technologies, the EXPENT8y89 coefficient is negative and significant, and the
coefficient for EXPENTgog2 is not significant in Table 7.7.
These results suggest that early establishment of a TLO after Bayh-Dole
was not correlated with receiving more important patents during the 1987-92
period. Indeed, the establishment of a TLO during 1980-86 actually is asso¬
ciated with less important patents during the 1987-89 period. Although not
definitive, these findings are consistent with anecdotal evidence that some
universities initially assigned persons with little knowledge or skill in patent
and licensing management to these activities when they established formal
TLOs and that simply bestowing the title of technology licensing officer on an
inexperienced administrator was not sufficient to generate important patents.
In some cases, the establishment of a formal office of technology licensing
may have been an outcome, rather than a cause, of “learning to patent.”
Although we are not able to test for their effects, other forms of organiza¬
tional learning may have enabled entrants to improve the importance of their
patents. Organizations can learn from the experiences of others directly (for
 EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

example, consulting, personnel transfer, and conferences) and indirectly (for 147
example, monitoring of others and informal discussions with colleagues).
Both of these channels appear to have been important means for entrant uni¬
versities to learn to manage patenting and licensing. For example, Niels
Reimers, founder of Stanford University’s Office of Technology Licensing in
1969, served as a consultant to a number of entrant and incumbent universi¬
ties’ TLOs during the 1980s and disseminated information on procedures that
were successful at Stanford to other institutions.
Anecdotal evidence of personnel transfer among universities suggests an¬
other channel for interinstitutional knowledge diffusion and learning. In the
late 1980s, Columbia University recruited the former director of the technol¬
ogy transfer program at Iowa State University, an institution with a long history
of patenting. The AUTM, founded in the mid-1970s, also appears to have dis¬
seminated information on “best practice” through its conferences and profes¬
sional meetings of university technology transfer personnel.
Entrant universities also learned from informal communication and inter¬
action within the broader community of technology licensing officials during
the 1980s. Following an initial wave of less selective patenting during 1981-85,
Columbia University adopted the practice of patenting only when a probable
licensee could be located, based on guidance from other institutions with
greater experience in patenting.14 Columbia’s adoption of this and other poli¬
cies resulted from informal discussions with directors of other university
TLOs. Similar informal communications and interactions may have con¬
tributed to improvements in the “quality” of entrants’ patents during the 1980s
and early 1990s.
Indeed, from the earliest days of their involvement in patenting and licens¬
ing, U.S. universities have drawn on and learned from the experiences of oth¬
ers in selecting the inventions they sought to patent and in the overall man¬
agement of patenting and licensing activities (Chapter 3; Sampat and Nelson
2002). This trend appears to have continued after Bayh-Dole; Feldman et al.
(2002) suggest that universities have emulated one another in the use of equity
investments as part of their technology transfer strategies during the 1990s.
Seen against these historical and contemporary accounts, it would be surpris¬
ing if universities did not “learn to patent via informal channels, professional
networks, and other mechanisms.

CONCLUSION

This chapters empirical analysis of overall U.S. academic patenting before

and after 1980 yields conclusions that are consistent with the results of the
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

148 analysis of UC, Stanford, and Columbia patenting presented in Chapter 6.

Specifically, we find no decline in the importance and generality of the post-
1980 patents assigned to universities with substantial pre-1980 patent portfo¬
lios. But the patents of institutions with little or no previous history were not
significantly more important or general than nonacademic patents. We also
examined the changing characteristics of the patents assigned to entrant and
incumbent university patenters during the post-Bayh-Dole period in an effort
to determine the existence and scope of improvements in the importance of
entrant universities’ patents.
Our results provide little evidence to support an argument that change in
U.S. universities’ internal “research culture” after Bayh-Dole triggered a de¬
cline in the importance of academic patents, at least for the first fourteen years
following the passage of the Act. This finding does not rule out the possibility,
however, that such changes in the academic research culture may be occur¬
ring gradually and could eventually be revealed in declines in these measures
of academic patents’ importance.
Although our analysis seems to suggest that universities can indeed learn to
patent, the sources of such learning are not clear. Cumulative patenting ex¬
perience, historical relationships with the nonprofit Research Corporation,
and the allocation of administrative talent to formally designated technology
transfer activities all fail to explain improvement over time in the importance
of the patents in this group of universities. All of these measures of sources of
learning are flawed, but their lack of significance suggests that a more diffuse
learning process may underpin our results.
 Appendix

Derivation of the Econometric

Specifications Used to Test for

“Learning” in Academic Patenting

In this appendix, we describe in more detail the econometric specifications

outlined in the section of the chapter dealing with “Institutional Experience,
Learning, and the Characteristics of University Patents, 1981-92.” The depen¬
dent variable in our analysis is the number of citations received by a patent.
We utilize negative binomial regressions to work with this “count” dependent
variable. The conditional mean of the negative binomial citation function for
our basic specification is:
E(Citations\App„ UNIV, Classc)
= exp{Er[a,App, + p,(App, * UNIV)] + 2cycClassc} (1)

where Appt is a dummy variable that equals 1 for all patents with application
year t, UNIV is a dummy variable that equals 1 for all university patents, Class,
is a dummy variable that equals 1 for all patents in patent class c, and at, /3t, and
yt are coefficients. Our reported results aggregate the application year dummy
variables into four groups: t = 8183, t = 8486, t — 8789, and t = 9092 for ap¬
plication years 1981—83,1984—86, 1987—89, and 1990—92, respectively. I he
patent class dummy variables span the 303 classes in our sample.
The conditional mean for the number of citations to the sample of non-
academic control patents in year t = T is:
E(Citations\Control Sample) = E(Citations\AppT = 1, UNIV = 0, Class,)
= exp (aT + 2cycClassc) (2)

Similarly, the conditional mean for the number of citations to university pat¬
ents in year t — T is:
E(Citations\University Sample) = E(Citations\AppT = 1, UNIV= 1, Class,)
— exp (aT + ft T + 2]cycClassc) (3)
EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

150 The ratio of equations (2) and (1) is the “proportional difference” between the
number of citations to academic patents and the number of citations to the
control patents in year T, or exp(^T). We test for differences between the num¬
ber of citations to university and control group patents in year T by testing
whether = o. Such a test indicates whether academic patents applied for in
year T received significantly more citations than the corresponding sample of
nonacademic control patents applied for in that year.
To compare the importance of entrant universities’ patents with those
issued to incumbent universities and with our nonacademic control patents,
we insert a dummy variable, ENT, that identifies entrant universities in equa¬
tion (1), making the conditional mean function:
E(Citations\App„ UNIV, ENT, Classc)
f li,[a,App, + p,(App, * UNIV) 1
\ + 8,(Appt * ENT)] + XcTcClassJ
(4)
where ENT equals 1 for all entrant universities and zero for incumbent uni¬
versities and controls, and St is the coefficient for the Appt*ENT interaction
term. Equations (5-7), respectively, calculate the mean number of citations to
patents in the entrant, incumbent, and control patent samples for the applica¬
tion year t = T:
E(Citations\Entrant)
= E(Citations\AppT = 1, UNIV = 1, ENT = 1, Classc)
= exp(o:r + (3T + 8T + 'Z^^lass,)

E(Citations\I ncumbent)
= E(Citations\AppT = 1, UNIV = 1, ENT = 0, Class,)
= exp(ar + (3t + 2c~ycClassc)

E(Citations\Control Sample)
= E{Citations\AppT = 1, UNIV = 0, ENT = 0, Class,)
= exp(ar + Scy ,Classc) ^

Dividing equation (5) by equation (6) yields the proportional difference be¬
tween the number of citations to issued patents applied for by entrants and ci¬
tations to those applied for by incumbents in year t = T, or exp(<5T). Similarly,
the ratio of equations (6) and (7) is the proportional difference in citations
between incumbent and control patents (exp(/?T)), and the ratio of equa¬
tions (5) and (7) is the proportional difference in citations between entrants
and control patents, or exp(/?T + Sr). Testing for the significance of the first two
differences (exp(<ST) and (exp(/?T)) requires a standard Z-test. We use a Wald
test of the linear restriction (3j + <5^ = o to analyze the difference in citation
rates between entrant and control sample patents.
 EFFECTS OF ENTRY AND EXPERIENCE ON U.S. UNIVERSITY PATENTING

To assess whether our various “learning mechanisms” influence the char¬ 151
acteristics of patents issued to entrant universities, we constructed three
dummy variables, EXPENT, that each equal r for experienced entrants cor¬
responding to: (1) the existence of a contractual relationship with the Research
Corporation, (2) above-median levels of cumulative 1980-86 patenting, and
(3) the assignment of at least one 0.5 FTE to formal technology transfer
activities during 1980-86. Adding EXPENT and the coefficient <fit to equa¬
tion (4) yields the following conditional mean function:
E(Citations\App„ UNIV, ENT, EXPENT, Classc)
' J,,[a,App, + pt(Appt * UNIV)

exp
+ 8,(App, * ENT)
+ (f),(Appt * EXPENT)]
I
,+ ^cycClcLSSc (8)

The proportional difference in the number of citations to patents applied for

by experienced entrants and those applied for by other entrants in year t = T
therefore is exp(0f).
 8

What Happens in University-Industry

Technology Transfer?

Evidence from Five Case Studies

Robert A. Lowe, David C. Mowery,

and Bhaven N. Sampat

Thus far, much of our discussion of university-industry technology transfer in

this volume has highlighted trends in various quantitative indicators of re¬
search collaboration and technology transfer. But these indicators are less re¬
vealing about the characteristics of the process of technology- transfer between
universities and industry. And as we have noted in Chapter 5 and elsewhere,
the framers and supporters of the Bayh-Dole Act held specific assumptions
about this process and, in particular, about the role of patents and licenses
within it. In order to provide a richer and more detailed account of the tech¬
nology transfer process, this chapter presents case studies of the technology
transfer process for five inventions patented and licensed by Columbia Uni¬
versity and the University of California (UC).
Supporters of the Bayh-Dole Act argued that the lack of clearly defined
property rights for university inventions impeded the commercialization of
these inventions, and some recent scholarly research presents results consis¬
tent with this assumption. A paper by Jensen and Thursby (2001), based on a
survey of technology transfer officers at sixty-two major universities, reports
that 48 percent of university inventions are “proofs of concept” at the time of
licensing. Jensen and Thursby argue that this finding is consistent with the
characterization by supporters of Bayh-Dole of university inventions as “em¬
bryonic,” requiring substantial additional investment for commercial devel¬
opment. Their survey results also suggest that inventor involvement is com¬
mon in the commercialization of such early-stage inventions.1 Based on these
results, the authors argue that by giving academic inventors a financial stake
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

in the commercial success of early-stage inventions, Bayh-Dole created in¬ 153

centives for inventor involvement in postlicense development and commer¬
cialization and thus facilitated university-industry technology transfer.
Although these survey findings are broadly consistent with the assumptions
about the characteristics of university technologies made by the drafters and
supporters of Bayh-Dole, they provide little information about the central ar¬
gument for Bayh-Dole, that is, that patents and exclusive licenses are neces¬
sary to give firms incentives to develop and commercialize university inven¬
tions. Nor do the survey results provide much detail on the importance of
patent-based incentives for inducing inventor involvement in the commer¬
cialization process or in the differences (if any) among fields of academic re¬
search in the technology transfer process.
The case studies in this chapter highlight the field-specific and invention-
specific differences in the technology transfer process and the role of patents
and licenses in this process. There is substantial variation across the cases in
the importance of patents and licenses, the role of the university, the impor¬
tance and involvement of the academic inventor, and even the directionality
and characteristics of the knowledge flows between university and industry.
For example, Columbia University’s patenting and licensing activities were
important to the development and commercialization of Xalatan, a glaucoma
treatment. University patents and licenses were less important to transfer and
commercialization, however, for two other inventions discussed in this chap¬
ter (the Axel cotransformation process and soluble CD4): firms learned about
the inventions through informal scientific and technological communities
and were willing to invest in commercialization without clearly established or
exclusive property rights to the inventions.
Two other inventions (gallium nitride and the Ames II Tests) were licensed
by inventor-founded start-ups after established firms had elected not to license
the inventions. These inventor-founders argued that protection for their intel¬
lectual property was important to the foundation of their firms, but it remains
unclear whether patent protection was in fact necessary for the commercial
development of their inventions.
Previous scholars have noted the importance of inventor cooperation in
developing embryonic technologies (Jensen and Thursby, 2001) and inven¬
tions associated with considerable know-how or tacit knowledge (Lowe, 2002,
Shane, 2002). These five cases, however, reveal considerable contrast in the
role of the university inventor in technology commercialization. In three of
the five cases, inventor-founded start-up firms played a central role in com¬
mercialization, and inventors necessarily were heavily involved. In the fourth
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

154 case, the efforts of established firms to exploit the university invention were
aided by the inventor. In the fifth case, by contrast, the licensees required no
assistance from the inventor.
The heterogeneity within this small sample of cases underscores the need
for caution in generalizations about the nature of the technology transfer
process and the role of formal intellectual property rights in that process.
This heterogeneity also highlights the importance of flexibility in the tech¬
nology management policies and practices of universities — for example, tech¬
niques or policies that are effective in biomedical fields may be less effective
in electronics.

METHODOLOGY

The five cases discussed in this chapter are not random samples of the popula¬
tion of inventions at Columbia or the University of California, reflecting their
origins in separate research projects focusing on different aspects of the uni¬
versity technology transfer process (see Colyvas et al., 2002; Lowe, 2001). These
five case studies are not a representative sample of the full array of technologies
covered by either the patent portfolios or the “invention disclosure portfolios”
of Columbia University and the University of California. Moreover, all of these
five inventions were patented and attracted industrial interest from either es¬
tablished or start-up firms, something that is not true of the majority of inven¬
tions disclosed at either institution. The cases thus overrepresent relatively
“important” inventions, patented inventions, and successful technology trans¬
fer efforts at both universities. Nonetheless, interviews with university licens¬
ing officers and inventors, as well as our previous analysis of university inven¬
tion and patenting data, led us to conclude that these cases provide an accurate
depiction of the university patenting and licensing process.
The three cases from Columbia are drawn from a broader project that over¬
sampled licensed inventions that produced positive licensing revenues (see
Colyvas et ah, 2002). Indeed, two of the three Columbia inventions discussed
in this chapter have produced more licensing revenues than all but a few in¬
ventions licensed by any U.S. university since the passage of the Bayh-Dole
Act. The Columbia sample thus is biased in favor of particularly successful
patented inventions. All of the Columbia inventions also are in the biomed¬
ical field, which, as noted in Chapter 6, accounts for the majority of Colum¬
bia licensing revenues.
The two UC cases focus on inventions licensed by inventor-founded start¬
up firms and are drawn from a sample of inventions that was selected ran¬
domly to match the distribution of all inventor-founded start-ups at the Uni-
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

versity of California in three dimensions: technology field, inventors campus, 155

and date of founding (see Lowe, 2001). These UC cases include one bio¬
medical and one electronics invention. By construction, therefore, our UC
cases include only inventions with a high level of involvement by inventors in
the technology transfer process.
Information for each case was collected from university records; interviews
with inventors, licensing officers, and others involved in the process; and
secondary sources. Each case study begins with a brief summary of the tech¬
nology, followed by a history of the research leading up to the invention, in¬
cluding research efforts both within and outside university labs. We then
describe the state of development of the invention at the time a patent appli¬
cation was filed. Included in this discussion are the factors considered in the
inventors decision to disclose the invention and the university's decision to
pursue patenting and promote the commercial development of the invention.
Following this description, we discuss the role of the inventor in the technol¬
ogy transfer process and conclude with a summary of the current status of
the technology, including products, product sales, and royalties paid to the
university.
The five case studies are the following:
1. Cotransformation: a process to transfer genes into mammalian cells (Columbia
University).
2. Gallium Nitride: a semiconductor with both military and commercial applica¬
tions (University of California).
3. Xalatan: a glaucoma treatment (Columbia University).
4. Ames II Tests: a bacteria assay for testing potential carcinogenic properties of
pharmaceuticals and cosmetics (University of California).
5. Soluble CD4: a prototype for a drug to fight AIDS (Columbia University).

THE CASE STUDIES

Cotransformation
Cotransformation is a process that allows for the transfer of genes into mam¬
malian cells. Researchers use the cotransformation process to identify regula¬
tory sequences controlling gene transfer, to define mechanisms of gene ex¬
pression, and to understand gene function. Biotechnology firms also use the
process to produce protein-based drugs. Columbia University s patents on the
cotransformation process and resulting products were based on research by
Richard Axel of Columbia’s College of Physicians and Surgeons and have
been among the most profitable academic inventions during the past two
decades, producing more than $370 million in gross royalties since 1983
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

156 (1996 dollars). The cotransformation patents have been licensed by many
firms and are used to produce an array of important genetically engineered
pharmaceuticals.

Research History

Research on gene transfer techniques dates back to the 1940s, but the field was
revolutionized in the 1970s with the development of recombinant DNA meth¬
ods by researchers at Stanford and the University of California. Using these
methods, researchers could for the first time control precisely which genes
were transferred from one cell to another. Recombinant DNA technology also
allowed researchers to tailor DNA to examine gene expression. Prior to re¬
combinant DNA, little was known about the regulatory sequences controlling
gene expression (Marx, 1980).
Progress in gene transfer research in the 1970s was impeded by the fact that
relatively few cells took up the genes transferred into them by these methods,
making it difficult to pinpoint the transformed cells. A major contribution to
the transfer problem was provided in a series of papers published between 1977
and 1979 by Axel and colleagues that described the “cotransformation process”
for inserting genes into mammalian cells.2 The method involved transferring
a gene of interest together with a marker gene that could be used by re¬
searchers to identify and isolate certain cells as the cells attached to the DNA,
enhancing the efficiency of mammalian gene transfer. The Axel cotransfor¬
mation process was widely viewed as a major advance in gene transfer (Marx,
1980).

University Patenting and Licensing

Although much of the research in gene transfer was motivated by a desire to

understand gene expression, the scientific and technological community rec¬
ognized that such methods could also be used to produce large amounts of
proteins and therefore had significant applications in commercial drug pro¬
duction. Production of proteins with the cotransformation method had dis¬
tinct advantages over the prevailing method of protein production, which used
bacterial (prokaryotic) systems (Fox, 1983).
Axel discussed the commercial potential of cotransformation with the dean
of Columbia’s College of Physicians and Surgeons, who urged Axel and his
team to disclose the invention to the Office of the General Counsel at Co¬
lumbia University.' The university filed a patent application on the invention
in February 1980, before the passage of the Bayh-Dole Act. In April 1980, Co¬
lumbia petitioned the National Institutes of Health (NIH), which had sup-
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

ported Axels research, for title to any patents that might be granted and re¬ 157
quested the authority to grant an exclusive license. The NIH granted Colum¬
bia’s petition for title but denied the university authority to negotiate an ex¬
clusive license.
As we noted in Chapter 6, Columbia established the Office of Science and
Technology Development (OSTD) in 1981, and OSTD assumed control of ad¬
ministration of the Axel patent application in 1982. The first of five patents on
cotransformation was issued in 1983, well after the publication of Axel’s papers
describing cotransformation, and Columbia’s patents eventually covered the
process, the production of specific proteins, and the use of specific markers.
Academic researchers and firms had been using Axel’s cotransformation
process in their labs since the publication of his first paper on the topic in 1977
(Fox, 1983). Accordingly, the research community was not pleased when Co¬
lumbia was granted a patent. Harvard University molecular biologist James
Barbosa noted in 1983:

The [cotransformation] patent’s process has been in use all over the academic world
since ’77 . . . It’s been such a boon in getting mammalian cell gene transfer off the
ground that it has almost become a laboratory reagent. . . that the process has been
patented just doesn’t seem right. And as far as policing industry’s use of it, I don’t see
how Columbia will do it. The way a company makes a product is proprietary infor¬
mation. (“Axel Patent Claims Mammalian Cell Transfer,” 1983, p. 5)

Columbia initially sought to compel firms already using the technology to

negotiate a license with the university. Columbia licensing personnel exam¬
ined the patents, end products, and scientific publications of industrial firms
to identify potential users of the Axel process and informed these firms that if
they were using the cotransformation process to produce proteins, they must
pay royalties to Columbia. In 1984, the assistant director of OSTD at Colum¬
bia announced in a biotechnology trade journal that we understand from the
published literature, and from our own sources, that the process is being used
in research laboratories. Commercial corporations would have to license or
else be in infringement of the patent’’ and further noted that Columbia s po¬
sition is to defend the patent, taking legal action if necessary”(“Axel Patent
Claims Mammalian Cell Transfer,” 1983, p. 5).
Columbia’s technology licensing office communicated the essence of its
views in letters to at least seventy-five companies by early 1984. But the credi¬
bility of Columbia’s threats was undercut by the difficulties inherent in moni¬
toring the use of specific processes by firms to produce proteins. Many firms
did not respond to the letters, either because they were not using the Axel pro¬
cess or did not believe that Columbia could prove that they were doing so.
Ten firms did respond by June 1984 and agreed to a license. Despite the
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

158 uncertainty over Columbia’s ability to enforce the patent, these firms opted to
license the cotransformation technology because of the relatively low royalty
rates offered by Columbia, whose licensing officers stated that the university
pursued a relatively low royalty rate “in order to encourage the use of the tech¬
nology” (“Columbia University rDNA Patent Licensing,” 1984, p. 4).4 Another
series of warning letters produced more licensees, and by December 2000
thirty-four firms had licensed the Axel cotransformation technology.

Commercialization and Development

Cotransformation was an embryonic technology, and its use for the produc¬
tion of specific proteins of commercial quantity and quality required addi¬
tional experimentation and research. These development efforts continued af¬
ter the invention was licensed on a nonexclusive basis. Exclusivity was not
necessary to induce firms to commercialize a product based on the university
technology, perhaps because drugs produced with the cotransformation pro¬
cess could themselves be patented.
Another interesting feature of the exploitation of this technology is the lack
of interaction between the inventor and Columbia’s licensees. Each of the li¬
censee firms employed scientists whose training and research experience en¬
abled them to understand and exploit cotransformation based on the descrip¬
tion of the process in Axel’s published papers. As a result, the inventor did not
have to work with licensees to transfer his know-how or to otherwise convey
important tacit information on the characteristics and applications of co¬
transformation. The inventions technical details were sufficiently codified,
the papers by Axel and colleagues communicated enough of these details,
and the “absorptive capacity” of industrial firms (Cohen and Levinthal, 1990)
was sufficiently well developed for industrial researchers to pursue the com¬
mercial development of cotransformation without direct involvement by Axel
or colleagues.
Since 1984, Columbia’s cotransformation licensees have used the process to
produce pharmaceuticals used in the treatment of a wide range of diseases,
some of which have approached blockbuster status. Columbia has earned over
$370 million (1996 dollars) in gross royalties on the cotransformation patents
since 1983, making cotransformation one of the highest-grossing university in¬
ventions in the post Bayh-Dole era.

Summary

The history of the licensing and commercial exploitation of the cotransfor¬

mation patents suggests that patents and exclusive licensing agreements are
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

not always necessary to achieve commercialization of university inventions. In 159

the case of cotransformation, firms had the capabilities and incentives to use
the process for their own research and drug production in the absence of ex¬
clusive rights to the invention. Indeed, it appears that technology transfer oc¬
curred in spite rather than because of the patents, licenses, and involvement
of the university technology transfer office. The university patent produced
significant income for Columbia, but no evidence suggests that the patent and
associated nonexclusive licenses facilitated commercialization. Columbia’s
nonexclusive licensing agreements for the Axel cotransformation patent, like
the equally renowned (and lucrative) Cohen-Boyer patent jointly licensed by
the University of California and Stanford University, do not appear to have ac¬
celerated or otherwise made feasible the commercial development of this in¬
vention. Instead, these licensing agreements were used by Columbia to levy a
tax on the commercialization of an invention that was published in the scien¬
tific literature and whose commercial development in the absence of licens¬
ing almost certainly would have occurred on the basis of the technical infor¬
mation and demonstration of feasibility provided by the publication.5 The
Axel cotransformation case also suggests that heavy involvement by the uni¬
versity inventor in the commercialization process is less crucial when poten¬
tial users possess sufficient “absorptive capacity” to. exploit the invention.

Gallium Nitride

Gallium nitride (GaN) is a wide bandgap semiconductor; that is, a semicon¬

ductor that emits light across a wide spectrum of colors. GaN allows for col¬
ors, such as bright blue and green, not available from other materials used in
light-emitting diodes (LEDs).6 The technology has a number of other attrac¬
tive features, and its commercial potential has long attracted the interest and
investment of industrial and academic researchers. For example, GaN emits
light at an intensity not available from traditional LED materials such as gal¬
lium arsenide (GaAs) and silicon (Si). Like other LEDs, GaN is more energy
efficient (that is, less energy is lost to heat production in the process of con¬
verting electricity to light) than other lighting technologies. GaN semicon¬
ductors also provide greater transmission power with more efficiency than do
traditional technologies used in mobile phones, military radar, and commu¬
nications satellites.
Because of its unusual characteristics, industrial researchers in large U.S.
firms began to invest in GaN-related R&D in the 1960s. Research on GaN has
throughout its history included efforts by large corporate labs, small start-up
firms, and universities, all of which have made important technical contribu¬
tions. University research on GaN in many cases followed major advances
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

i6o from industrial researchers, and much of the technology transfer in this case
flowed from industry to the research university, rather than vice versa.

Research History

Research on GaN first began in the late 1960s in the laboratory of a prominent
researcher at RCAs Princeton labs, Jacques Pankove. Recognizing consider¬
able potential for GaN, major industrial firms such as IBM, Bell Labs, and
Matsushita had initiated research on GaN and related semiconductors by the
early 1970s, and a number of patents issued by the mid-1970s to these corpora¬
tions. By the early 1980s, however, many of these corporate research efforts had
been abandoned because of limitations in the materials that could be used for
substrates, among other technical problems (Kahaner, 1995). Further basic re¬
search on new materials for substrates and new crystal-growing processes was
necessary to develop GaN technology for commercial applications. These
technical hurdles, combined with reductions by several of these industrial
firms in their central corporate research laboratories during the 1980s, reduced
industrial interest in the family of Class III-V semiconductors to which GaN
belonged.
As GaN projects were being dropped elsewhere, two Japanese researchers,
Isamu Akasaki of Matsushita and Shuji Nakamura of Nichia Chemical, con¬
tinued work (independently of one another) on GaN, receiving over fifty U.S.
patents on GaN and related research through July 2002. Akasaki had worked
for a number of years on related research at Matsushitas research labs. After
meeting Pankove at conferences and visiting Pankoves lab, Akasaki launched
a GaN research program at the Matsushita Research Institute, where he was
the head of the Fundamental Research Laboratory and general manager of
the Semiconductor Department. Akasaki continued his work on GaN after
moving to a faculty position at Nagoya University in 1981. Nakamura, who
was employed by Nichia Chemical, a small Japanese maker of phosphors for
cathode-ray tubes and fluorescent lighting, began research on GaN in the
mid-1980s. He initiated his work on GaN at a time when researchers at large
electronics firms and several universities had begun work on zinc selenide-
based semiconductors, which seemed to have considerable potential for blue-
LED applications and were better understood than GaN.
Although zinc selenide materials were similar in structure to GaN and
were believed to be more feasible to produce, researchers in corporate labs
and U.S. universities failed to produce LEDs based on these materials that
could last longer than a few hundred hours, making them impractical for com¬
mercial use. These failures in zinc selenide R&D meant that Akasakis and
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

Nakamuras successes with GaN renewed the interest of several companies 161

and university scientists in GaN by the early 1990s. But Nakamura had a sub¬
stantial technological lead and produced the first modern prototype GaN-
based blue LED in his lab in 1994.7
Figure 8.1 presents data on the assignees for GaN patents granted between
1972 and 1998 in order to depict the relative importance of industrial, aca¬
demic, and independent inventors during this period in the development of
GaN technology. The most interesting finding from the figure is the impor¬
tant role of industrial, rather than academic, researchers as patentees during
the first decade of this technology’s development. Of 14 GaN-related patents
issuing during 1972-82,11 were assigned to corporations, 2 were assigned to in¬
dividual inventors, and only 1 was assigned to an academic institution.8 Dur¬
ing the later 1980s and 1990s, however, this picture changed somewhat; 29
of the 105 GaN patents issued during 1983-98 were assigned to universities.
Industrial patenting remained dominant during 1983-98, but the share of
GaN patents assigned to academic institutions roughly doubled, from 14 per¬
cent to nearly 28 percent. It is hardly surprising that industry dominated GaN

figure 8.1 Assignees for U.S. GaN-related patents, 1972-98

WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

162 patenting throughout the development of this technology, but the timing of
the expansion in academic patenting is surprising. Rather than dominating
the early stages of GaN patenting, patenting by academic researchers appears
to have followed that of industrial researchers. The flow of knowledge and
technology—in this case, at least —appears to have been as much from in¬
dustry to universities as the reverse.
All but three of the GaN-related patents assigned to universities between
1983 and 1994 covered inventions by Akasaki at Nagoya University.9 Two of the
other three university patents were for inventions by Pankove, who had ac¬
cepted a faculty position at the University of Colorado in 1985.10 The flow of
patentable research relevant to GaN into universities from industry during the
1980s thus was facilitated by the movement of Akasaki and Pankove from in¬
dustry to academic positions.11
One industrial assignee for several of the GaN patents issuing in the early
1990s was Cree Research, a firm founded in 1987 that had licensed patents
from North Carolina State University that covered silicon carbide materials
for wafers and electronic devices. Cree Research, which played an important
role in the exploitation of GaN patents developed at UC Santa Barbara
(UCSB) (see below), had begun a research program during the 1990s to im¬
prove crystal growth methods and manufacturing processes that could be ap¬
plied to GaN.
Nakamuras research and prototype attracted the interest of Steven Den-
Baars and Umesh Mishra of UCSB. DenBaars and Mishra began to investi¬
gate GaN applications in lasers and microwave transistors that had been over¬
looked by other researchers. As Mishra recounted during an interview:

Akasaki and Nakamura were concentrating on a certain aspect, which of course is

the commercially more important area: colored lighting. But then we decided to
make some blue emitters, some lasers, and some microwave transistors [for wireless
applications] . . . The first time we went to a conference, we went without a paper.
We just went to listen. At the next conference, we had a substantial presence giving
papers. (Mishra, 2001)

Established firms provided financial support for DenBaars and Mishra in

their research on new applications for GaN in the mid-1990s. Hughes Elec¬
tronics, which had a large research facility in Santa Barbara, had long main¬
tained a close relationship with UCSB’s engineering school. Although
Hughes had for many years provided philanthropic support for the engineer¬
ing school, the firm had never expressed interest in licensing the results of
UCSB faculty research. Instead, Hughes used its philanthropic support of ba¬
sic research to maintain a “window” on the research activities of UCSB fac¬
ulty and to gain access to promising graduate students.
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

When Mishra and DenBaars decided to establish a laboratory at UCSB for 163
the study of GaN and related research, financial gifts from Hughes provided
“seed funding” for the lab, with no formal agreements governing the disposi¬
tion of intellectual property produced by Mishra, DenBaars, or other re¬
searchers. Additional research funding was provided by grants from the Na¬
tional Science Foundation and the Office of Naval Research, as well as from
a Japanese LED manufacturer, Stanley Electric, that eventually took options,
but no licenses, on inventions from the UCSB lab.

University Patenting and Licensing

The UCSB research facility proved to be quite productive. During the 1995 —
2001 period, Mishra and DenBaars published 223 papers and received twelve
patents on twenty-one invention disclosures based on GaN, GaAs, and related
projects in their laboratories. The dozen patents received by the UCSB re¬
searchers highlight the substantial commercial interest in their results, since
UC licensing officers at that time rarely filed patent applications without
strong signals of commercial interest in the technology, including a willing¬
ness by an industrial optionee or licensee to underwrite the patent prosecu¬
tion costs.
Established firms funded at least a portion of the patent prosecution fees
for several of Mishras and DenBaars’s inventions. In other cases, industrial
firms (including some that had been active in GaN research in the 1970s)
signed “secrecy agreements” with the University of California that allowed
them to conduct technical evaluations of the inventions for which the Uni¬
versity of California had filed patent applications. But the significant indus¬
trial interest in the UCSB research on GaN did not translate into licensing
agreements. In a recent interview, DenBaars (2000) stated that the lack of in¬
terest by established industrial firms in licensing the UCSB GaN patents re¬
flected long-standing R&D practices in the electronics industry: “We were fil¬
ing patents at the university, pretty aggressively at first . . . and the university
was marketing those patents, but our past experience has been with university
patents that large companies generally are not interested in licensing. They’re
interested in doing it themselves.”
In fact, few of Mishras and DenBaars’s inventions were licensed to es¬
tablished firms, although Siemens, Hewlett-Packard, Matsushita, Toshiba,
LumiLEDS (a joint venture between Agilent Technologies and Phillips), and
other large firms all had mounted substantial R&D programs by 2000 to de¬
velop GaN applications. Although twenty-three firms reviewed the UCSB
inventions, only one license (to Cree Research) had been executed as of
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

164 March 2002. Moreover, Cree negotiated its license with UCSB only after it
had acquired a firm founded by Mishra and DenBaars (see below).
Part of the difficulty in marketing these patents was the early stage of de¬
velopment of the inventions. The UCSB patents covered processes that were
small-scale prototypes of those needed for commercial production volumes,
and considerable uncertainty remained about the “scalability” of the manu¬
facturing processes used for their production. The UCSB researchers cited the
importance of personal experience with the technology, rather than formal in¬
tellectual property, as the key source of competitive advantage. As DenBaars
(2000) explained:

Especially in the semiconductor business . . . the basic way big companies get at
your technology, is to get the know-how, and that’s done by hiring your students.
Then [the technology] is in the public domain. The most important thing, at least
in the semiconductor field, is know-how . . . and all of the amassed knowledge dur¬
ing the course of your Ph.D. So, the students are the most important resource we
have. I’d say they’re much more valuable than the patents.

As DenBaars s statement suggests, the lack of industrial interest in licensing

the DenBaars-Mishra patents (including a lack of interest on the part of the in¬
ventors) appears to have been rooted in the characteristics of the “knowledge
base” relevant to industrial innovation in semiconductor materials. In contrast
to many fields of biomedical research, the codified technical information con¬
tained in the patent and licensing documents for the GaN inventions con¬
veyed little of the critical know-how and related information necessary to
translate the invention into commercial applications. At the same time, many
of the firms active in this technical field had substantial related know-how and
knowledge and needed little information beyond the demonstration of the fea¬
sibility of specific technical approaches in order to justify investment in their
own R&D programs aimed at commercial applications. Finally, as was
pointed out in Chapter 2, the strength and economic value of individual pat¬
ents in a field such as semiconductors are widely acknowledged to be lower
than is true of most biomedical patents, and industrial firms therefore may
have seen less need to license relevant academic patents in order to pursue
technical work in the area.

Development and Commercialization

Confronted with limited interest in licensing from large established firms,

Mishra and DenBaars founded a firm to pursue further development of their
inventions. Widegap Technologies, later renamed Nitres, was founded in 1997
by the two UCSB researchers and a former semiconductor industry executive,
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

Fred Blum, and hired several graduate students who had worked in the 165

Mishra-DenBaars labs. After initial negotiations with the university, the in¬
ventors’ new firm did not complete an option or license agreement for the
GaN patents that were based on Mishras and DenBaarss research.
Widegap Technologies/Nitres initially relied on funding from established
firms and government grants. Hughes Electronics and General Electric, re¬
spectively, provided matching funds for a Small Business Innovation Research
(SBIR) grant and a National Institute of Standards and Technology’s Advanced
Technology Program (ATP) grant. Widegap/Nitres also received federal re¬
search grants from the Air Force, the Office of Naval Research, the Defense
Advanced Research Projects Agency, the Ballistic Missile Defense Organiza¬
tion, and the U.S. Army.
The uncertainty associated with the embryonic nature of the new firm’s
GaN technology appears to have limited funding for Nitres from venture cap¬
italists or the equity markets. According to Blum (2000):
We didn’t even approach any VC’s [venture capitalists] initially. We raised angel
funding, but that was through my professional relationships with industry execu¬
tives in the Los Angeles area. We didn’t want VC money initially because they want
to see a prototype too soon, and the technology was still too fragile. If they don’t see
a product coming up soon in the process, VC’s get worried and can put unreason¬
able pressure on the company.

UCSB provided a valuable asset for Nitres’s GaN research during the ini¬
tial years by renting university “cleanroom” space to the firm. According to
DenBaars and Mishra, the firm’s financial constraints and the exorbitant costs
of constructing or acquiring a “cleanroom” meant that Nitres would have
been unable to operate without the UCSB research space.
Cree Research acquired Nitres in March 2000, after Nitres had developed
commercial-level prototypes of some of its technology and had completed the
development projects funded by its initial federal grants. By early 2002, Cree
had released several GaN-based products, including green and blue LEDs for
the commercial market, as well as applications for military and telecommu¬
nications. As of mid-2002, however, the University of California had yet to re¬
ceive any royalty payments from Cree on the GaN licenses.

Summary

The GaN case, like that of the Axel cotransformation patents, is one in which
patents per se appear to have been relatively unimportant for university-
industry technology transfer. In both cases, the existence or absence of a
patent on the key inventions did not affect the development by industry of
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

166 technological innovations based on the university invention. In contrast to the

GaN patents, however, the Axel patents generated substantial income for Co¬
lumbia University. The differences between these inventions and their licens¬
ing history reflect differences in the level of demand for the technologies they
respectively supported, as well as underlying differences in the legal strength
and economic value of patents in the biomedical and electronic fields. An¬
other important contrast with the Axel case was the role of the inventors in
such technology transfer as did occur with the UCSB patents — faced with
limited interest from established industrial firms in their patents, the UCSB
engineering faculty who had developed these technologies started their own
firm. Finally, the history of GaN is characterized by a two-way flow of knowl¬
edge and personnel between industry and academia, rather than a one-way
flow of knowledge and “prototypes” from universities to industry.

Xalatan

Xalatan is an eyedrop solution used to treat glaucoma, an eye disease associ¬

ated with increased fluid pressure in the eye and damage to the optic nerve
that can cause loss of vision or blindness. Xalatan uses a class of compounds
called prostaglandins (PGs) to increase drainage and thereby reduce intraoc¬
ular tension. Before the introduction of Xalatan, glaucoma was typically
treated with beta-blockers, which often had serious side effects in patients with
histories of cardiac problems. Xalatan has fewer side effects and now is the
market leader among glaucoma therapies.
The transfer to industry and the commercial development of Xalatan ap¬
pear to be consistent with the assumptions of the Bayh-Dole Act, in that the
patents and exclusive license covering the drug were essential to its successful
commercialization. In contrast to the Axel cotransformation process, however
(and in a fashion reminiscent of the GaN case), Xalatans academic inventor
was centrally involved in its development and commercialization.

Research History

Laszlo Bito of Columbia University’s College of Physicians and Surgeons be¬

gan research on the effects of PGs on the eye in the early 1970s, funded by a
series of grants from the National Eye Institute of the NIH. In the late 1970s,
he published several papers demonstrating that small doses of PGs could de¬
crease intraocular pressure in cats and rabbits (see, for example, Camras, Bito,
and Eakins, 1977), and in 1981 he published a paper reporting similar results
from experiments on primates (Camras and Bito, 1981).
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

167
University Patenting and Licensing

Bito had originally planned only to publish his research results and gave little
thought to commercialization. But in the course of submitting a grant renewal
application in 1982, he learned that Columbia’s post-Bayh-Dole patent policy
required him to report his research advances to the university’s technology
transfer office. Following discussions with the OSTD at Columbia, Bito hied
an invention report in 1982 on “[t]he use of eicosanoids and their derivatives
for the treatment of ocular hypertension and glaucoma,” a disclosure based on
the research that he had published in 1977 and 1981. Columbia applied for a
patent on the use of PGs for treatment of ocular hypertension and glaucoma
in 1982, and the patent issued in 1986. Columbia subsequently applied for and
received two other patents on the technology.
Even before the first patent had issued, Columbia began to seek licensees
for the invention. Despite a significant marketing effort by Bito and Colum¬
bia, however, there were no takers, perhaps because of the widespread view
among industry experts that PGs were harmful to the eye. In a recent inter¬
view, Bito recalled that prospective industrial licensees were reluctant to con¬
template the use of PGs in the human eye, arguing, “It’s crazy. You can’t put
prostaglandins in the eye,” and a leading eye expert called Bito’s idea “next to
ridiculous” (quoted in Gerth and Stolberg, 2000). Substantial additional re¬
search was required to determine whether and how the technology could be
useful in humans, and few industrial firms saw significant commercial poten¬
tial in Bito’s research advance.
In 1983, however, the Swedish firm Pharmacia agreed to exclusively license
the technology at the urging of a friend of Bito’s who had developed a cataract
treatment for the firm. Carl Camras, a student and colleague of Bito’s who was
instrumental in developing the invention, later noted that if Lazio Bitos
friend hadn’t done wonders for Pharmacia ... no one would have picked it up”
(“Columbia Innovation Enterprise: A Special Report,” 1996, p. 1). The un¬
certainty surrounding the feasibility and commercial potential of Bito’s treat¬
ment meant that the strong patent protection and exclusive licenses negoti¬
ated with Pharmacia were important in the commercialization of Xalatan.

Development and Commercialization

Bito played an important role in the development and commercialization of

Xalatan. Inasmuch as he was one of the few individuals in the world who had
used PGs to treat glaucoma, it is not surprising that Phamacia needed Bito’s as¬
sistance to develop a drug based on his invention that was as effective as exist¬
ing (non-PG) treatments for glaucoma. Some of the collaboration between
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

168 Bito and Pharmacia occurred in the firm’s labs in Uppsala, Sweden, where re¬
searchers worked to understand the operation of PGs in treating glaucoma.
Bito also continued his research at Columbia with funding through a research
agreement between his university and Pharmacia.
By 1991, researchers at both Columbia and Phamacia had developed con¬
siderable confidence in the commercial prospects for the PG-based glaucoma
treatment, and by 1994 Pharmacia began to consider submitting its product to
the U.S. Food and Drug Administration (FDA) for clinical testing and ap¬
proval. But the 1995 merger between Pharmacia and Upjohn, a large U.S.
pharmaceuticals firm, slowed the development of this technology. The new
management team did not assign a high priority to the project, and Bito noted
in a later interview that this “was a very stressful process, because after a good
team had brought the product to fruition, the new people in the management
of the merged company lacked familiarity with it” (“Columbia Innovation
Enterprise: A Special Report,” 1996, p. 4).
By this time, a number of other pharmaceutical firms had approached Bito
and Columbia, expressing interest in obtaining a sublicense from Pharmacia
for the development of the PG-based glaucoma treatment for fields of use
other than those being pursued by the Swedish firm. Faced with evidence that
Pharmacia had scaled down its efforts to develop the glaucoma treatment since
its merger with Upjohn, Bito urged Columbia to require Pharmacia to permit
such a sublicense. The university conveyed Bito’s sentiments to Pharmacia
and encouraged the firm to reinvigorate its commercialization efforts. Perhaps
because of this encouragement from its academic licensor, Pharmacia-Upjohn
began to move forward on commercialization of Xalatan, and the product ob¬
tained FDA approval in 1996. Sales of Xalatan grew rapidly following its com¬
mercial introduction and by FY2001 accounted for more than $740 million
(1996 dollars), making Xalatan the worlds top-selling glaucoma medication. In
the last year (2000) for which data on royalties were available, Columbia
earned almost $20 million in gross royalties from Xalatan (1996 dollars).

Summary

The Xalatan case differs from the Axel cotransformation and GaN cases in that
patents appear to have been important to the transfer and commercialization
of this technology. In part, the importance of patents reflected the fact that this
invention resembled the “prototypes” discussed by Jensen and Thursby
(2001) a lengthy and costly period of development was necessary to bring this
invention to market. And the inventor’s know-how and involvement were in¬
dispensable to this development process, in contrast to the cotransformation
patents. But the Xalatan case illustrates another issue in exclusive licensing
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

agreements for university patents that also appears in the soluble CD4 case dis¬ 169
cussed below. Although a firm may be willing to sign an exclusive licensing
agreement with the university (and although most such agreements include
“due diligence” or “best efforts” clauses that commit a licensee to invest in the
development of an invention), it is difficult for any licensor, let alone an aca¬
demic licensor, to ensure that their licensee will undertake the costly process
of technology development in a timely fashion.

The Ames II Tests

The Ames II Tests are an assay based on the Ames Tests that were developed
in the late 1960s and early 1970s by Bruce Ames at UC Berkeley. The original
Ames Tests were bacteria-based assays used to determine whether a test sub¬
stance has mutagenic potential (the potential to cause mutation in cells). Mu¬
tagenic properties serve as simple proxies for the carcinogenic potency of the
tested substance, and the original Ames Tests are widely used in the pharma¬
ceutical, cosmetics, and food industries in basic product-safety testing. The
Ames II Tests were based on research conducted by Pauline Gee and Dorothy
Maron, two researchers in Ames’s laboratory, two decades after the develop¬
ment of the original Ames Tests.
The commercialization of the Ames II Tests provides an interesting con¬
trast with the commercialization of the original Ames Tests, developed before
Bayh-Dole, which were not patented and were published in the scientific jour¬
nals. As a result, firms and research labs were able to implement the Ames
Tests without any licenses. In contrast, the Ames II Tests were published but
also were covered by two patents and required licenses for their adoption. The
research behind the Ames II Tests was published in the Proceedings of the Na¬
tional Academy of Sciences in 1994, one year after the first patent application
was filed in October 1993 (Gee, Maron, and Ames, 1994) and three years be¬
fore the filing of a second patent application on the invention by the Univer¬
sity of California. The commercialization processes for the Ames and Ames II
Tests also differed from one another in that inventor involvement was crucial
to the development of the Ames II Tests, unlike the original Ames Tests. One
of the Ames II Tests’ inventors participated in the foundation of a firm devoted
to commercialization of the tests and played an active role as an adviser to
other firms seeking to apply this technology.

Research History

Bruce Ames publicly announced his original Ames Tests in 1975 and soon rec¬
ognized the importance of his invention for academic and industrial research.
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

170 Although the original Ames Tests were not patented, the assays were available
from the University of California’s E. Salmonella Mutagenicity Test Resource
Center for a small administrative fee.12 According to the Mutagenicity Cen¬
ters website, more than 3,000 industrial and academic labs have used the
Ames Tests, and many firms and labs manufacture their own strains.13 The
standard version of the Ames Tests exposes a test substance, such as a pharma¬
ceutical molecule, cosmetic, or food, to a strain of Salmonella typhimurium
bacteria that has been chemically modified to no longer manufacture an
amino acid, histidine, that is necessary for growth. Once mixed with the sub¬
stance being tested, some of the strains of bacteria will mutate and grow, just
as a nonmodified, or “wild type,” bacteria would. Since only mutated strains
of bacteria are able to manufacture histidine and grow, a count of the fre¬
quency of growing colonies of bacteria within the tested material provides a
rough index of the material’s.mutagenic properties. If the test substance’s in¬
teraction with the bacteria leads to considerable mutation and growth of the
bacteria, the substance may have strong mutagenic properties and could be
harmful to consumers.
The original Ames Tests provided little information on the characteristics
of the mutagenic properties of the tested substance. During the late 1980s,
however. Gee and Maron began work on an enhanced version of the Ames
Tests that utilized bacteria that had been genetically modified to produce spe¬
cific mutations. By 1992, Gee and Maron had developed the Ames II Tests,
which provided additional information on the mutagenic properties of a test
substance by enabling researchers to determine which genes in the test bacte¬
ria were affected after exposure to a new material or substance.

University Patenting and Licensing

In compliance with UC policy, Gee and Maron submitted their invention,

which included both the genetically engineered bacterial strains and the pro¬
cess through which they were produced, to the UC Office of Technology
Transfer in 1993. The wide adoption of the original Ames Tests seemed to in¬
dicate that industrial interest in licensing Gee and Maron’s follow-on inven¬
tion would be substantial. Surprisingly, however, no established firms ex¬
pressed interest in licensing the Ames II Tests for further development and
commercialization. One biotechnology company expressed interest but did
not pursue the invention with any formal agreement.
Gee stated in a recent interview that at the time of her discovery, she had
no intention of developing and commercializing the invention (Gee, 2001).
Shortly after Gee and Maron’s disclosure of their invention, however. Harvard
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

professor Spencer Farr, a former postdoctoral fellow in the Ames laboratory, 171
began discussions with Gee about commercializing her invention. Financed
by two venture capitalists, Farr had founded a firm to develop and commer¬
cialize his own invention, a gene profiling assay, and he recognized consider¬
able commercial potential for the Ames II Tests as a complementary prod¬
uct.14 Farr invited Gee in 1993 to join his fledgling enterprise, Xenometrix, and
Gee left academia to join the firm. Xenometrix licensed two patents related to
her invention from UC.

Development and Commercialization

According to Gee, Xenometrixs licenses for the Ames II Tests patents aided
the firm’s efforts to raise capital, but the inventors knowledge and experience
with the invention were necessary to commercialize the Ames II Tests. As Gee
(2001) recounted: “The value is not just the patent. . . My experience is in¬
tangible, whereas IP is tangible, and we can wrap legal language around the
IP in our agreements. Clearly what you want is both, but the IP’s protection is
necessary in case someone tries to steal our technology.”
The importance of Gee’s knowledge and experience for development of
the Ames II Tests was underscored by the efforts of Xenometrixs European dis¬
tributor, Xenometrix GmbH, to replicate the Ames II Test strains. By the mid-
1990s, Xenometrix GmbH was concerned about maintaining a sufficient sup¬
ply of the bacterial strains and requested that Xenometrix train the distributor’s
scientists on the techniques used to grow the bacteria. The distributor’s labo¬
ratory scientists spent a week at Xenometrixs U.S. headquarters, undergoing
intensive training in the techniques and equipment needed to grow these bac¬
terial strains. But after nearly a year of effort, the distributor’s European scien¬
tists were unable to produce a supply of strains that met the company’s speci¬
fications and could be sold to customers. In response, Gee spent several days
at the distributors European labs and produced roughly two years’ supply of
the bacteria.
By early 2001, the Ames II Tests had achieved limited commercial success.
Xenometrix continued its development efforts, but Farr had departed to start
another firm, Phase One. In early 2001, Discovery Partners International ac¬
quired Xenometrix to expand sales of the firms two main product lines, Farrs
profiling assay and the Ames II Tests. 1 he University of California has earned
almost $130,000 in royalties on its licenses for the Ames II Tests, although a
substantia] portion of the royalty payments (nearly $50,000) has been delayed
or forgiven to ensure sufficient cash flow for Xenometrix to maintain opera¬
tions (all in 1996 dollars).
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

172
Summary

The commercialization of the Ames II Tests presents some interesting simi¬

larities and contrasts with the GaN and Xalatan cases. Like GaN and Xalatan,
inventor involvement was important and reflected the importance of tacit
know-how for the inventions’ applications. It seems likely that without the par¬
ticipation of the inventor, a license alone would not have sufficed to commer¬
cialize the Ames II Tests. But in contrast to GaN, patent protection for this in¬
vention and the exclusive licensing contract negotiated by its industrial
commercializer proved to be important, just as was the case for Xalatan. In¬
deed, its license for the Ames II Test patents significantly enhanced the avail¬
ability of venture finance for Xenometrix.

Soluble CD4

Soluble CD4 is a prototype for an AIDS drug that inhibits viral entry and in¬
fection. Since CD4 is a receptor for HIV (the virus that causes AIDS), a syn¬
thetic, soluble CD4 can “mop up” HIV before it invades healthy cells. The
idea of a soluble version of CD4 emerged from university and industrial labo¬
ratories in the mid-1980s, a period during which rapid growth in the number
of AIDS cases in the United States created a strong demand for better AIDS
treatments. Soluble CD4 was widely viewed as a product with considerable
commercial potential. The Boston Globe referred to soluble CD4 as “one of
the hottest —though still preliminary —ideas in AIDS treatment since AZT,
the only AIDS drug so far approved” (Foreman, 1988, p. 41), and Anthony
Fauci, a leading AIDS expert, noted at an NIH meeting: “We are all very in¬
terested in this. It works very well in the test tube, [and] there appears to be an
excellent scientific basis to proceed” (Perlman, 1987, p. Ai). Despite this en¬
thusiasm, as of 2001 no soluble CD4-based AIDS treatment had been ap¬
proved for commercial use.

Research History

The discoveries that made soluble CD4-based treatments a possibility were

developed in the mid-1980s in the laboratory of Richard Axel at Columbia
University. Axel was no stranger to commercialization activities, having previ¬
ously invented the cotransformation process described earlier. In 1985, scien¬
tists in the United States, Great Britain, and France provided evidence that the
cellular receptor for the HIV virus was the CD4 molecule, and in 1985 Axel’s
Columbia lab was the first to clone the gene that produces the CD4 receptor
(Foreman, 1988). Together with several graduate students, he showed that the
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

HIV virus could be “decoyed” into attaching itself only to cells into which the 173
CD4 receptor had been inserted.
These experimental results were published in 1985 and led Axel and one of
his graduate students, Paul Maddon, to consider the therapeutic benefits of a
soluble form of genetically engineered CD4. Although their discoveries gen¬
erated enthusiasm among scientists and pharmaceutical firms, uncertainty re¬
mained about the feasibility of CD4-based AIDS treatments. First, the effects
of soluble CD4 on the overall human immune system were unknown. A sec¬
ond concern was the possibility that the body would produce antibodies
against the CD4 decoys, provoking an autoimmune reaction. Third, some sci¬
entists were skeptical that CD4 was the only receptor for HIV, since HIV is ca¬
pable of infecting cells in the brain and immature blood cells in the bone mar¬
row that lack CD4 receptors. By the mid-1980s, soluble CD4 treatments had
not been tested in animals, let alone humans, and the invention was little
more than an idea with significant but uncertain commercial potential.

University Patenting and Licensing

In March 1986, Axel and Maddon submitted two invention disclosures based
on their CD4 research to the OSTD at Columbia. The first disclosure covered
isolation of the gene encoding CD4, and the second disclosure dealt with the
method for producing soluble CD4. The university filed patent applications
on both disclosures in August of that year, and patents issued in 1992. OSTD
began discussions with potential licensees about the CD4 invention in 1987,
and at least six firms expressed interest in licensing the CD4 technologies from
Columbia.
The evolution of Columbia’s licensing strategy for soluble CD4 illustrates
the costs and benefits of exclusive licensing arrangements for early-stage in¬
ventions. Significant additional research was needed to develop a soluble
CD4-based treatment, and any licensee therefore faced considerable risk and
uncertainty about the ultimate commercial feasibility and financial returns
associated with the technology. These high risks, as well as the need for signifi¬
cant additional investments to bring a CD4-based treatment to market, meant
that some form of exclusivity in licensing the CD4 technologies might be nec¬
essary to attract an industrial licensee. But OSTD also faced great uncertainty
about the performance and prospects for commercial success of prospective li¬
censees, especially if any such license was an exclusive one. As a result, OSTD
sought to license several firms and scientific teams that could pursue different
approaches to development of a CD4-based product.
Columbia originally planned to grant a coexclusive license to three firms:
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

174 Biogen, Genentech, and SmithKline Beecham, each of which had been work¬
ing on soluble CD4-based treatments since shortly after Axel’s successful
cloning of the CD4 receptor (Axel had participated in SmithKline Beecham’s
work in this area). But before the final agreements with these three firms were
finalized, Paul Maddon, Axel’s graduate student, founded a firm (Progenies
Pharmaceuticals) to develop a soluble CD4-based treatment and requested
that Columbia add him to the list of coexclusive licensees. Columbia con¬
vinced the other three licensees to modify the draft license agreements, and
Progenies Pharmaceuticals was added to the list of licensees in 1989.

Commercialization and Development

Following the publication by Axel and colleagues of their paper in 1985 con¬
firming that they had cloned the CD4 receptor (Maddon et ah, 1985) and be¬
fore the Columbia license agreements were signed, numerous firms had begun
research on a soluble CD4-based HIV treatment (Foreman, 1988). In De¬
cember 1987, scientists from Genentech published a paper in Science report¬
ing positive results, and scientists from Biogen, SmithKline Beecham, the
Dana Farber Institute, and the Basel Institute of Immunology each published
papers on their respective research on soluble CD4 in Nature later that month.
A contemporary press article observed, “[T]he nearly simultaneous appear¬
ance of five major papers in such a novel area of research is a sign not only
of the intense competition in AIDS research, but also of the tremendous
commercial promise that drugs for AIDS portend”(Foreman, 1988, p. 41). By
the summer of 1988, researchers from several other firms, including Becton-
Dickinson, Genelabs, and Ortho Pharmaceuticals, had published research pa¬
pers on soluble CD4-based treatments.15 In September 1988, seven NIH re¬
searchers reported in Nature that they had developed a poison that, when
combined with CD4, latched on to HIV and killed it; later that year they an¬
nounced plans to license this technology (known as “CD4- PE”) exclusively to
Upjohn Pharmaceuticals (“Licensing Plan for AIDS Drug Draws Fire,” 1988).
By 1989, the first year in which Columbia’s license agreement was in effect,
the race to develop soluble CD4 was well under way and involved numerous
firms and labs other than the Columbia licensees. The presence of several
nonlicensee firms in the race indicates either that the scope of Columbia’s pat¬
ent was narrow and the firms believed that they were not infringing the
Axel/Maddon patents or that the returns to a commercially successful prod¬
uct based on infringement of the patents were so great that industrial devel¬
opers believed that a mutually agreeable and profitable settlement with Co¬
lumbia University could be negotiated in the event of litigation.16
The development activities of these nonlicensee firms also provide evi-
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

dence that licenses for the Columbia/Axel patents were not necessary to sup¬ D5
port development efforts. Nonlicensee firms engaged in research on CD4 ap¬
parently believed that they had other means of appropriating the returns from
their R&D, including patents on any products developed for clinical trials. It
is also noteworthy that only two of the firms in the race to develop soluble
CD4 (SmithKline Beecham and Progenies) enlisted the direct participation
of the university inventors (Axel and Maddon, respectively) in their develop¬
ment efforts. In this case, like that of cotransformation, the university inven¬
tors evidently did not have a monopoly on the tacit knowledge needed to de¬
velop the invention, something that is further corroborated by the large
number of papers on this topic that have been published by scientists in non¬
licensee firms and in other research laboratories.
The enthusiasm surrounding soluble CD4 began to wane by 1990, the year
after Columbia had signed the license agreements for the Axel/Maddon pat¬
ents. Biogens and Genentech’s soluble CD4-based products showed little ef¬
ficacy in Phase I clinical trials, and high development costs, along with pre¬
dictions of high manufacturing costs, led these two firms and SmithKline to
scale back their research on CD4. By 1995, each of these firms’ licenses had
been terminated by Columbia, and only the Columbia start-up licensee, Pro¬
genies, continued efforts to develop a soluble CD4-based treatment.
Before they dropped out of the commercialization race, Biogen, Smith-
Kline, and Genentech had invested substantial funds in research on soluble
CD4, and two of these firms had begun clinical trials. Ultimately, however, no
commercial product was developed under the coexclusive license. In 1995,
Progenies, which by this time was recognized as the leader in soluble CD4-
based treatments, announced plans for clinical trials of its soluble CD4-based
therapy.17 In 1996, the firm renegotiated its license agreement with Columbia,
acquiring exclusive rights for the development of a soluble CD4-based prod¬
uct. In 1996, Progenies disclosed the existence of a second receptor, CC-CKR-
5, for the AIDS virus, a scientific discovery that meant that CD4 was necessary
but not sufficient for HIV to bind to cells. Progenies currently has several drug
candidates based on soluble CD4 and decoy CC-CKR-5 receptors in clinical
trials.
Since the CD4 breakthrough was announced, an alternative approach,
antiretroviral therapy, has been widely adopted for treatment of HIV-infected
individuals. Nevertheless, interest in Progenies’ drug candidates, and in
“entry inhibitors” more generally, remains high.18 As of December 1999, how¬
ever, Columbia had earned less than $10,000 in sales-based royalties (1996 dol¬
lars) from the Axel-Maddon patents, since a CD4-based therapy has yet to be
commercialized.19
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

176 Summary

The history of CD4S development process illustrates the commercial and

technical uncertainties involved in bringing an embryonic invention, even
one that appears to have great commercial potential, from laboratory to mar¬
ketplace. This case also provides some evidence that exclusive licenses may
not be necessary, even for embryonic inventions, if their potential profitabil¬
ity is sufficiently large and downstream innovations can themselves be
patented. Moreover, the case highlights the risks associated with exclusive li¬
censing agreements for such innovations, since it is often difficult for licens¬
ing professionals to determine which of several potential licensees (in the rare
cases in which several firms are interested in pursuing licenses) is most likely
to bring the invention to market successfully. Finally, this case (like the co¬
transformation case) suggests that in contexts where firms have strong links
with the relevant scientific and technological communities, inventor involve¬
ment may be less critical for commercialization.

CONCLUSION

Chapter 5 pointed out that a key premise underpinning the Bayh-Dole Act is
the belief that patenting and licensing are necessary to facilitate the develop¬
ment and commercialization of publicly funded university inventions. Al¬
though the Act does not mandate that universities follow any single specific
policy in patenting and licensing faculty inventions, university administrators
and technology licensing officers frequently assume that the technolog}' trans¬
fer process is essentially similar in different technologies and industries. But
these case studies reveal great heterogeneity within even a small sample of
technologies. There are significant differences among these cases in the role
of intellectual property rights in inducing firms to develop and commercial¬
ize university inventions, in the role of the inventor in postlicense develop¬
ment and commercialization, and in the relationship between academic and
industrial research activities in different technical fields.
Patents and an exclusive license were important to successful commercial¬
ization in one of these five cases (Xalatan), but in at least two cases (cotrans¬
formation and GaN) it seems likely that development and commercialization
would have gone forward without a patent on the university invention. In
these cases, other means of appropriability, such as specialized knowledge or
the prospect of a patent on downstream inventions, were sufficient to induce
firms to invest in development and commercialization. The case of soluble
CD4 also illustrates the difficulties that university licensing officers face in se¬
lecting among prospective licensees when the ultimate commercial prospects
 WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

of the invention and the commercialization capabilities of the licensees are 177
highly uncertain.20
The cases also reveal considerable differences in the extent of inventor in¬
volvement and the role of the inventor in development and commercializa¬
tion. In at least two cases (soluble CD4 and cotransformation), one or more of
the licensee firms had little or no interaction with the inventor, since firms had
sufficient experience and internal expertise in the field of the invention or had
strong relationships with external scientists with such experience. In these
cases, the knowledge and know-how gap between the university inventor and
a would-be industrial commercializer was relatively small, reflecting previous
investments by the industrial firm in internal capabilities and external moni¬
toring of scientific developments.21 But two other inventions discussed in this
chapter (GaN and Ames II) were developed and commercialized by start-up
firms in which inventors played a central role.
The nature of feedback between industrial and academic research differs
among these cases. Bayh-Dole was implicitly based on an assumption of a “lin¬
ear model” of innovation, in which universities perform basic research with
little concern for application and private firms invest in applied research and
commercialization. In this view, patent-based incentives are essential to link
universities, inventors, and industry in the commercialization process. As we
pointed out in Chapter 2, however, this assumption does not accurately de¬
scribe university-industry interactions, before or after Bayh-Dole, in many
technical fields. Indeed, in most of the cases discussed in this chapter, there
was considerable overlap between the scientific and industrial communities
in the nature of research activities (including publication). Consistent with
the work of Zucker and colleagues on biotechnology (Zucker, Darby, and
Armstrong 1994), in these cases technology transfer from universities to firms
took place via a range of channels, including labor mobility and research col¬
laboration. Moreover, in two cases (GaN and CD4) university research itself
built in part on research that originated in industry. There is also little evi¬
dence of significant delays in the disclosure or publication by academic re¬
searchers of their research advances. All of these inventions were the subject
of published papers, and in a majority of the cases the publications appeared
before patent applications were filed.
The research discussed in Chapter 2 highlights significant differences
among industries in the influence of academic research on industrial innova¬
tion as well as in the channels through which these influences operate. This
research also suggests significant interindustry differences in the importance
of patents as vehicles for knowledge transfer among firms or between univer¬
sities and industry and further reveals significant differences among industries
WHAT HAPPENS IN UNIVERSITY-INDUSTRY TECHNOLOGY TRANSFER?

178 in the importance of patents and licenses as channels for the transfer of knowl¬
edge and technology between universities and industry. These case studies do
suggest, however, that patents may be important for start-up firms in their
search for financing. Consistent with previous studies, the evidence from our
small sample of cases suggests that university patenting and licensing were
more important for the biomedical inventions than for the one electronics in¬
vention we studied. But these cases also reveal considerable heterogeneity in
the technology transfer process among biomedical technologies.
These differences among inventions, industries, and technical fields can be
accommodated by flexibility in universities’ patenting and licensing practices.
University licensing offices have at their disposal a number of contractual ar¬
rangements to facilitate technology transfer, including secrecy agreements,
options, licenses, material transfer agreements, and equity investments. To the
extent that universities choose among these instruments carefully, and with
the objective of facilitating use and commercialization rather than maximiz¬
ing royalty income, patents and licenses can advance the mission of university-
industry technology transfer while maintaining the other important missions
of public and private universities in the United States.
 9

Conclusion

In this concluding chapter, we summarize our argument concerning the

Bayh-Dole Act and U.S. universities’ role in economic growth and innovation
and suggest some implications and cautions for U.S. policymakers and uni¬
versity administrators in the held of university patenting and licensing. We be¬
lieve that much of the current discussion of the economic role of U.S. research
universities and the contributions of U.S. universities to the economic boom
of the 1990s exaggerates the role of Bayh-Dole. In fact, U.S. universities have
been important sources of knowledge and other key inputs for industrial in¬
novation throughout the twentieth century, and much of this economic con¬
tribution has relied on channels other than patenting and licensing.
Any assessment of the economic role of universities must recognize the nu¬
merous, diverse channels through which university research influences in¬
dustrial innovation, and vice versa. This recognition is lacking in many of the
current efforts of other nations to emulate Bayh-Dole in initiatives designed to
enhance the contributions of their university systems to innovation and eco¬
nomic growth. U.S. policymakers and university administrators also must keep
in mind these characteristics of university-industry interactions and technol¬
ogy transfer and should base policy and management decisions on a more nu-
anced and realistic view of the relationships between academic research and
commercial innovation. University administrators in particular should heed
the concerns raised by their predecessors during the 1930s and 1940s over the
political risks created by any appearance of university ‘ profiteering from
patenting and licensing.
CONCLUSION

180 U.S. UNIVERSITIES AND TECHNICAL CHANGE IN INDUSTRY

The relationship between U.S. university research and innovation in industry

is a long and close one. Indeed, organized industrial research and the U.S. re¬
search university both first appeared in the late nineteenth century and have
developed a complex, interactive relationship. We pointed out in Chapter 2
that the unusual structure of the U.S. higher education infrastructure, which
blended financial autonomy, public funding from state and local sources, fed¬
eral research support, and substantial scale, provided strong incentives for
university faculty and administrators to focus their efforts on research activi¬
ties with local economic and social benefits. Rather than being exclusively
concerned with fundamental scientific principles, much of U.S. university re¬
search throughout the late nineteenth and twentieth centuries focused on un¬
derstanding and solving problems of agriculture, public health, and industry.
The massive increases in federal financial support for U.S. university re¬
search that began in the 1940s weakened research ties with local industry in
many universities, but this federal funding was dominated during the postwar
period by agencies with two clear missions: national defense and public
health. As a result, much U.S. university research during the postwar period
remained focused on “Pasteur’s Quadrant,” combining a quest for fundamen¬
tal understanding with a focus on the use of that understanding for solutions
to specific problems or missions (Stokes, 1997).
U.S. universities have made important contributions to industrial innova¬
tion throughout the past century, not least through providing both advanced
research and education. The strong links between education and research sus¬
tained a close relationship between the evolving scientific research agenda
and problems of industry or agriculture, while at the same time providing a
powerful and effective channel (in the form of trained students) for the trans¬
fer and application of much of this knowledge to industry and other economic
sectors. In addition, many university researchers in engineering and medical
schools maintained close ties with the users of their research and their gradu¬
ates in industry, medical practice, and agriculture.
The collaboration between university and industrial researchers, com¬
bined with the focus of many U.S. university researchers on scientific prob¬
lems with important industrial, agricultural, or other public applications,
meant that a number of U.S. universities patented faculty inventions through¬
out the twentieth century. Nevertheless, despite the adoption by a growing
number of universities of formal patent policies by the 1950s, many of these
policies, especially those at medical schools, prohibited patenting of inven¬
tions (see Chapter 3), and university patenting was less widespread than was
 CONCLUSION

true of the post-1980 period. Moreover, many universities chose not to manage 181

patenting and licensing themselves. Chapter 4 discussed the role of the Re¬
search Corporation, founded by Frederick Cottrell, a University of California
faculty inventor who wished to use the licensing revenues from his patents to
support scientific research, in managing university patents and licensing dur¬
ing much of the period between 1935 and 1980. But even in these early de¬
cades of patenting and licensing, biomedical technologies accounted for a dis¬
proportionate share of licensing revenues for the Research Corporation and
other early university licensors, such as the Wisconsin Alumni Research
Foundation.
The decade of the 1970s, as much as or more so than the 1980s, represented
a watershed in the growth of U.S. university patenting and licensing. U.S. uni¬
versities expanded their patenting, especially in biomedical fields, and as¬
sumed a more prominent role in managing their patenting and licensing
activities, supplanting the Research Corporation. Agreements between indi¬
vidual government research funding agencies and universities also con¬
tributed to the expansion of patenting during the 1970s. Private universities
in particular began to expand their patenting and licensing rapidly during this
decade.
Expanded interest by U.S. universities in patenting and licensing academic
inventions and growing concern over U.S. economic competitiveness during
the late 1970s combined to produce the Bayh-Dole Act of 1980, which also was
influenced by a broader federal effort during this period to rationalize and
simplify policy governing the disposition of patent rights resulting from feder¬
ally funded research. As we noted in Chapter 5, the Act did not legalize any¬
thing that previously had been prohibited by federal law or regulation, but it
obviated the need for universities to negotiate agreements with individual
agencies or petition for patent rights on a case-by-case basis and reduced un¬
certainty about the direction of government patent policy. Since the trend
toward increased academic patenting and licensing (including patenting of
government-funded research) predates the passage of Bayh-Dole, the Act’s
most important effect arguably was its provision of a congressional endorse¬
ment of patenting and licensing (including exclusive licensing) as appropriate
activities for universities and public laboratories.1
The Bayh-Dole Act is widely credited (or blamed) for the growth in uni¬
versity patenting since 1980. As we have noted, U.S. university patenting had
begun to grow before the passage of the Act, and the patent propensity of
U.S. universities (patents per dollar of academic R&D spending) grew steadily
throughout the post-1945 period with no sharp break in trend after 1980. The
growth of university patenting during the 1970s and 1980s also was affected by
CONCLUSION

182 other developments, including advances in basic science, particularly molec¬

ular biology. Biomedical research advances during the 1970s and 1980s in turn
drew on an academic R&D infrastructure that had benefited from the growth
in federal funding of biomedical research since the 1960s. Expanded univer¬
sity patenting also was affected by judicial and patent office decisions that
clarified and broadened the range of patentable subject matter in the bio¬
medical sciences. No evaluation of the causes and consequences of increased
university patenting since 1980 can overlook the influence of broader U.S. pat¬
ent policy on such growth.
The Bayh-Dole Act did not dramatically affect the patenting and licensing
activities of universities that had long been active in this area, such as Stanford
University and the University of California. Indeed, the biomedical patents
and licenses that dominated these institutions’ licensing revenues during the
1980s and 1990s had begun to grow before the passage of the Bayh-Dole Act. Co¬
lumbia University, an institution with little experience in patenting and li¬
censing before 1980 (and an institution that prohibited the patenting of inven¬
tions by medical faculty until 1975), also had filed for its first blockbuster patent
before the effective date of the Act. Nevertheless, the Act did increase patent¬
ing of faculty inventions at both Stanford and the University of California, al¬
though many of these patents covered inventions of marginal industrial value
and did not yield significant licensing royalties. And the Bayh-Dole Act drew
universities into patenting and licensing that previously had not been active in
these fields. Initially, this wave of entry was associated with the production by
many less experienced universities of relatively low-value patents. Over time,
however, entrants appear to have “learned to patent” and have begun to obtain
patents of greater technological (and potentially, economic) significance.
The analysis of U.S. university patenting and licensing before and after
1980 also highlights the high concentration of these activities in a relatively
narrow range of research fields, primarily the biomedical sciences. As we have
emphasized throughout our discussion of the topic, the biomedical sciences
differ from many fields of engineering and scientific research in the unusual
strength and economic value of patents and licenses for those patents. In ad¬
dition, the advances in molecular biology that revolutionized the field during
and after the 1970s yielded scientific discoveries with significant applications
in the pharmaceutical and other industries, another aspect in which biomed¬
ical research seems to differ from other fields of academic enterprise. The con¬
centration of patenting and licensing in a relatively narrow array of academic
fields means that the effects of increased patenting, before and after Bayh-
Dole, on U.S. universities’ “research culture” cannot be described as perva¬
sive. Instead, any such effects are concentrated in a few areas of research and
 CONCLUSION

in a few academic departments in most universities. The other implication of 183

these important differences among academic fields, of course, is the need to
recognize that patent and licensing policies that may be appropriate for the
biomedical sciences could well discourage university-industry collaboration
in other fields, a point that we discuss in more detail below.
The entry and learning behavior of less experienced university patenters af¬
ter 1980, which we discussed in Chapter 7, illustrate an especially interesting
aspect of the long history of patenting and licensing by U.S. universities — the
perspective that it provides on the operation of markets for intellectual prop¬
erty, a topic of considerable interest to economists and other students of the
“knowledge-based economy.” The mixed experiences of the entrants into
patenting and licensing demonstrate the difficulty of developing institutional
capabilities in managing markets for intellectual property. Successful univer¬
sity licensing officers must combine a deep understanding of a given techni¬
cal field with considerable familiarity with industrial applications. Moreover,
these licensing managers must establish and maintain close working relation¬
ships with faculty inventors whose estimates of both their own talent and the
value of their inventions often are inflated, while negotiating complex con¬
tracts with industrial personnel who may have far more experience in a given
field. In short, licensing officers draw on a set of talents that are not abundant
within most university administrations, and it is hardly surprising that many
universities found it difficult to establish a successful patenting and licensing
organization in the immediate aftermath of the Bayh-Dole Act.
Did the Bayh-Dole Act and the growth in university patenting that charac¬
terized the 1970-2000 period increase the contributions of university research
to the “New Economy” in the United States in the 1990s, as so many of the ac¬
counts cited in Chapter 5 suggest? The contributions of U.S. universities to
economic growth and innovation during the 1980s and 1990s assuredly were
important, but no evidence suggests that these contributions were more im¬
portant than they were during the 1930s or 1950s. Nor does any evidence
“prove” that Bayh-Dole substantially increased these contributions or that any
such expansion would not have occurred in the absence of this Act. The
nature of these contributions and the channels through which they have
been realized before and after the Bayh-Dole Act have been complex and
have included much more than patenting and licensing. But the factors
underpinning these important economic and innovative contributions are
rooted in the structure of the U.S. higher education system, which in many re¬
spects contrasts with those of other industrial economies. U.S. universities
have been closely connected with the external groups that can benefit from
their research and training. They have been relatively quick to expand new
CONCLUSION

184 fields and techniques of science in areas of interest to their constituents in agri¬
culture, medicine, and industry. In some cases, university patenting has facil¬
itated technology transfer. But in many cases, transfer does not depend on the
possession by universities of intellectual property rights.

THE BENEFITS AND RISKS OF UNIVERSITY

PATENTING AND LICENSING

The widespread belief held by many policymakers and university adminis¬

trators in the United States and elsewhere that Bayh-Dole has been an un¬
mitigated success and indispensable to the economic contributions by U.S.
universities is based on little evidence. First, data on the growth of U.S. uni¬
versities’ patenting and licensing activities alone provide no basis on which to
conclude that patenting and licensing are essential for technology trans¬
fer, since increased university patenting may cover technologies or inventions
that previously were transferred via other channels. Second, increased aca¬
demic patenting and licensing, as well as growth in other forms of university-
industry collaboration, predate Bayh-Dole. Third, the “evidence” on low rates
of commercialization before Bayh-Dole is weak, as we discussed in Chapter 5.
Any comprehensive assessment of the effects of Bayh-Dole, or the benefits
of increased university patenting and licensing more generally, must consider
the potentially negative effects of these activities on the academic research en¬
terprise, on the access by industry to university research results, and on other
channels of university-industry technology and knowledge transfer. We have
uncovered little evidence that the expanded patenting and licensing activities
of U.S. universities since 1980 have produced significant shifts in the orienta¬
tion of academic researchers away from fundamental research toward more
applied, short-term research activities that might be more easily patented and
licensed. Moreover, as we have repeatedly noted, the concentration of patent¬
ing in a relatively narrow range of academic disciplines means that any such
shift in the orientation of academic researchers would itself be highly con¬
centrated in a few fields. It is also important to keep in mind the relatively
utilitarian research orientation of many U.S. universities and academic re¬
searchers throughout the twentieth century. Moreover, our (limited) evidence
on the characteristics of U.S. university patents before and after 1980 does
not suggest a change in the research orientation of U.S. universities after
Bayh-Dole.
Nor can one mount a convincing argument that patenting of the techno¬
logical products of university research per se is likely to have detrimental con¬
sequences for innovation within the U.S. economy. Patenting, after all, in-
 CONCLUSION

volves the disclosure and publication of the details of a given invention and in 185
some circumstances may reveal as much as a technical paper concerning the
details of an invention.
There are two broad areas for concern over the effects of increased patent-
ing by U.S. universities. The first deals with the extension of patenting since
the 1970s to cover artifacts formerly viewed as “science”: ideas, materials, and
techniques that themselves represent important inputs into the scientific re¬
search process. We emphasize that this extension of the realm of patenting is
only partially attributable to the Bayh-Dole Act; it reflects a set of broader
changes in overall U.S. policy toward intellectual property rights. The second
area for concern deals with the policies adopted by university patentholders
toward the licensing of these patents. These policies can affect both the diffu¬
sion and use of important scientific and technological advances funded by
public resources, and they may also pose significant political risks for univer¬
sities of the sort discussed by university administrators in the debates of the
1930s and 1940s. We discuss each of these risks below.

Patents as Obstacles to Scientific Research

We have noted that much of university patenting comes out of research in

fields in “Pasteurs Quadrant,” where it is difficult to distinguish between a re¬
search result that enhances understanding in the field and points to further
research and one that seems to point to practical future applications. It is
these potential applications, of course, that makes these results potentially
patentable. Nonetheless, the U.S. Patent and Trademark Office and the courts
have in some cases granted broad patents on research results where any appli¬
cations are highly uncertain and far in the future and whose principal near-
term use is as an input to further research. Where this happens, the conse¬
quences for subsequent scientific research may be significant and negative.
Unlike the establishment of property rights over technological inventions,
“privatization” of knowledge inputs that formerly were part of the “scientific
commons” through patenting may impede the progress of research. Increased
academic patenting also may enhance incentives for faculty or universities to
delay publication, restrict sharing of research materials, and/or limit the shar¬
ing by faculty of their research results with the scientific community via con¬
ference presentations or informal communications (Dasgupta and David,
1994; Liebeskind, 2001). A survey of more than 2,000 biomedical faculty con¬
ducted in 1993 by Blumenthal et al. (1997) found that 20 percent of respon¬
dents had delayed publication of research results for at least six months. Nearly
half of these faculty reported that they delayed publication of research results
CONCLUSION

186 in order to protect the patentability of these findings. This survey also found
that 9 percent of respondents had denied other scientists’ requests for access
to their research results. In a follow-up survey by members of this research
team conducted in 2000, 73 percent of respondents working in genetics indi¬
cated that difficulties in obtaining access to data or materials had slowed sci¬
entific progress in this field (Campbell et al., 2002).
The study by Campbell et al. (2002) highlights another potential effect of
academic patenting and licensing —greater restrictions on access by re¬
searchers to inputs for scientific research. If scientific research relies on access
to many patented technologies held by different owners, that is, where there
are numerous potential claimants to particular lines of product development
or research, the transaction costs of obtaining access to these rights may slow
research (Heller and Eisenberg, 1998). One indicator of these increased trans¬
actions costs is the growing reliance by universities and industry on formal li¬
censing contracts and material transfer agreements for the sharing of research
materials (for example, Eisenberg, 2001; Rai and Eisenberg, 2003).2
In these cases, the research outputs are important inputs to a flow of future
research as well as useful inputs for those who are trying to solve practical
problems. Accordingly, the economic benefits from such research advances
are likely to be greater if they can be available to all those working to advance
knowledge and practice in the area. Will such liberal dissemination eliminate
any incentives for researchers to pursue such research, by reducing the ability
of researchers to control or to profit from the use of their findings? We think
not, since the primary reward mechanism for most publicly funded academic
research is based on publication and priority of discovery (Dasgupta and
David, 1994).
Several changes in the U.S. patent system, rather than in the Bayh-Dole
Act, might reduce impediments to scientific research resulting from increased
university patenting. First, the U.S. Patent and Trademark Office could adopt
a more skeptical policy toward patents on discoveries that largely cover natu¬
ral phenomena, for example, by requiring that the applicant make a strong
case that the subject matter of the patent application or patent is “artificial.”
Although these distinctions are difficult to articulate at a general level and
even more difficult to apply on a case-by-case basis, we believe that a more re¬
strictive policy on such patents could reduce any erosion of the “scientific
commons.”
Second, a stricter interpretation of the meaning of “utility” or usefulness
might limit patenting of inputs to science. This issue is particularly important
for patent applications that argue that the research result in question can be
employed to achieve something obviously useful —a case for usefulness once
 CONCLUSION

removed. But the problem here is that the patented “invention” then is useful 187
mainly as an input for future research and represents generic knowledge or
technology that is most useful if kept in the public domain. A stricter inter¬
pretation here would require a more compelling demonstration of significant
progress toward a particular practical solution than appears to be employed at
present.
Third, patent applicants often make claims in their applications that are far
broader than any results they actually have achieved. There are obvious ad¬
vantages to the patentee of being able to control a wide range of possible sub¬
stitutes to what has actually been achieved, but there are also advantages to so¬
ciety as a whole from limiting any obstruction of a field of technological
innovation that may result from broad claims.
More extensive patenting by researchers from industry and academia alike
of inputs to science, as well as the greater efforts of universities to support
patenting and licensing of at least some research discoveries that previously
would have been disseminated through publication, is creating a more com¬
plex landscape of intellectual property rights for scientific research. Indeed,
the risks that academic researchers may knowingly or unknowingly infringe
patents in the course of their fundamental research cannot be dismissed. His¬
torically, academic researchers have been protected from exposure to in¬
fringement litigation or damages by the “research exemption.” There is a long
history of statements by judges that use of a patented invention or advance in
pure research is not a violation of a patent. University researchers and admin¬
istrators have relied on a de facto research exemption in opposing efforts by
patentholders to threaten university researchers with infringement lawsuits,
and U.S. industry historically has granted university researchers a de facto re¬
search exemption in many areas.
The research exemption has never been embodied in law, and the fact that
much of U.S. university research is focused on practical objectives as well as
on the goal of advancing basic understanding, has long complicated univer¬
sity administrators’ arguments for a research exemption. But U.S. universities’
recent efforts to seek patents and licensing agreements for research findings
that previously were placed in the public domain also have changed the situ¬
ation and may increase the reluctance of industry executives to maintain such
a policy. In some fields, such as research materials and tools, university and in¬
dustrial researchers now are competitors as much or more so than collabora¬
tors, and industry researchers are often required to obtain licenses to use
patented university research results. Facing such demands from universities,
industrial research managers see no reason why they should not make similar
demands on universities (Eisenberg, 2001). Many of the obstacles to a research
CONCLUSION

188 exemption thus reflect the changes that have transformed universities from re¬
search collaborators to commercial competitors in some fields. As we note be¬
low, this shift in perceptions of the role of universities reflects their behavior
with respect to licensing and litigation for the patents in their portfolios, rather
than academic patenting per se.
An additional blow to the research exemption came from the 2002 decision
of the Court of Appeals for the Federal Circuit in Madey v. Duke, which es¬
sentially denied the existence or validity of an academic “exemption” from
patent infringement suits. The decision also hints at some change in the per¬
ception of university research by at least some members of the federal judici¬
ary in response to the increased patenting and licensing activities of U.S. uni¬
versities. Ruling on an infringement suit filed by an academic researcher
against Duke University, the court argued that basic and applied research was
part of the “central business” of a university. Since universities benefited fi¬
nancially and otherwise from their research activities, the court argued, it was
reasonable under the law for a patentholder to require that the university take
out a license before using patented material in research.
The Madey v. Duke decision has great potential to impede the progress of
academic research in fields in which patenting has grown in frequency and
has increasingly covered inputs to research.3 If this decision is not overturned
by the U.S. Supreme Court, legislation will be necessary to restore this ex¬
emption. As of mid-2003, however, neither a judicial reversal nor a legislative
solution seems likely. Indeed, if Congress views universities as entities that
profit from their research, both publicly and privately funded, the prospects
for a legislative research exemption could be quite dim. Dreyfuss (2003) pro¬
poses granting a research exemption to university researchers who agree not to
patent research results that rely on patented material. But the Dreyfuss pro¬
posal would require many universities to revise their policies toward patenting
and licensing the results of academic research.
Another solution to the potential problems created by the patenting of sci¬
entific research tools has been proposed by Rai and Eisenberg (2001), who sug¬
gest that the Bayh-Dole Act be amended to strengthen federal agencies’ dis¬
cretion and ability to limit the patenting or restrictive licensing of critically
important scientific discoveries. As the authors note, the National Institutes of
Health (NIH) has on several occasions relied on exhortation to limit re¬
searchers’ patenting of large strands of primary human genomic DNA, and the
NIH has limited the restrictive licensing agreements for embryonic primate
stem cells. Using amendments to the Bayh-Dole Act to create a more trans¬
parent mechanism for such intra-agency deliberations, including advice from
a board of distinguished industry and academic scientific researchers for each
 CONCLUSION

agency, could improve the accountability and predictability of this process, 189
while strengthening the legal basis for something that is presently an ad hoc
procedure. But this proposal also may require considerable change in univer¬
sity licensing policies. Indeed, many universities might well oppose any pro¬
posal granting greater authority to federal research funding agencies to con¬
strain their freedom in the field of licensing.
As we note in our discussion of licensing policies, U.S. universities may
well encounter greater political or judicial hostility if they do not through their
actions (as opposed to their rhetoric) recognize that as public institutions
whose research is largely funded by the public, they have a responsibility for
encouraging their research results to be used as widely as possible. The issues
of “research freedom,” universities’ political standing, and licensing policies
are intertwined, and no politically feasible solution to these problems can fail
to address all of them.

University Licensing Policies

As the previous paragraph suggests, much of the responsibility for managing

patenting and licensing in a fashion that preserves the integrity of higher ed¬
ucation and maintains political support for the large public research budgets
that have been devoted to U.S. universities’ research activities since 1940 rests
with universities. It is important that university administrators recognize that
technology transfer and licensing are components of and subsidiary to their
central institutional missions of education and research. Although there is
abundant rhetorical recognition of this fact, university policies must be con¬
sistent with this understanding as well. A more sophisticated and nuanced ap¬
proach to managing patenting and licensing activities is likely to benefit indi¬
vidual institutions as well as the overall institutional and political strength of
U.S. research universities.
A critical first step in developing an institutional policy for patenting and
licensing is recognition of the many objectives of any such policy and the
trade-offs among these objectives. Among the goals of technology licensing
that university administrators and licensing officers frequently emphasize
are: (1) licensing revenues; (2) maintenance or expansion of industrial re¬
search support; (3) regional economic development; (4) faculty retention; and
(5) technology commercialization. Earlier chapters cited examples of licens¬
ing policies (notably, those pursued by the Research Corporation) that sought
to maximize licensing revenues, in spite of the risks posed by such policies to
the maintenance of cordial relationships with corporate licensees who provide
research support. Similarly, the use of licensing policies to retain faculty may
CONCLUSION

190 involve more generous revenue-sharing with inventors or a licensing policy

that favors new, faculty-founded start-up firms rather than established corpo¬
rations. Regional economic development objectives also raise potential trade¬
offs with licensing revenues.
There is widespread rhetorical recognition of the multiple goals of univer¬
sity licensing policies. But the management of the complex trade-offs among
these goals often is deficient. According to a recent survey of seventy-six major
technology transfer offices, licensing income was the most important criterion
by which technology transfer offices measure their own success (Thursby,
Jensen, and Thursby, 2001). Licensing revenues are a legitimate objective of
licensing policy (although it should not be the only one), but the limited
profitability of technology licensing for most U.S. universities suggests that
an exclusive emphasis on royalties at the expense of these other goals is ill
advised.
A second important premise for institutional licensing and patenting man¬
agement is the recognition that academic research flows to industrial applica¬
tions (and as the case studies in this volume document, industrial research
flows to academic applications) through a variety of channels, including the
movement of faculty, students, and researchers between academia and indus¬
try; the publication of research articles; appearances by faculty and university
researchers at conferences; and faculty consulting. All of these channels are
important and often complement one another. A single-minded focus on
patenting and licensing as the only important or effective channels for tech¬
nology transfer is unrealistic and may produce policies that limit the effec¬
tiveness of other channels that are more important for knowledge transfer and
exchange (Cohen, Nelson, and Walsh, 2002).
Institutional patenting and licensing policies also should recognize the sig¬
nificant differences among academic disciplines or research areas in the im¬
portance of patent-licensing agreements for technology transfer that we dis¬
cussed in Chapters 2 and 8. Patent protection is strong and economically
significant in biomedical research, and the dominance of licensing revenues
by biomedical inventions reflects this fact. But in other areas, such as elec¬
tronics, a commercial device may require access to dozens or hundreds of pat¬
ents, and the average value of a patent often is much lower. Patent licenses typ¬
ically are less important and valuable in these fields, and industrial firms often
collaborate with academic researchers with little expectation of obtaining
rights to key patents. Indeed, in such fields, the insistence by university ad¬
ministrators on extensive agreements covering intellectual property may serve
as a source of friction rather than as a lubricant for research collaborations. It
is important for university research administrators to adjust their intellectual
 CONCLUSION

property policies to accommodate these intersectoral differences, rather than 191

conceptualizing all research collaborations as resembling those common in
biomedical research. Once again, this recognition requires the pursuit of a
broader and more flexible set of objectives through patenting and licensing
policies, rather than focusing solely on licensing revenues.
Where exclusive licenses for patented academic inventions are not essen¬
tial to commercial exploitation, exclusivity may reduce the societal benefits as¬
sociated with an invention. In particular, restrictive licensing may excessively
limit the diversity of further experimentation and development in a context
when multiple, rivalrous development efforts may be more socially desirable
(Merges and Nelson, 1994). Current evidence does not allow us to reach
strong conclusions about the frequency of these cases or to estimate the mag¬
nitude of the associated social costs — they vary across technological fields and
have surely changed over time. Because they involve counterfactuals (for ex¬
ample, “What would have happened without an exclusive license?”), these
costs are difficult to measure systematically.
Nevertheless, the important public mission of universities and university
researchers in the United States and elsewhere, as well as the fact that much
of the research conducted within U.S. universities is supported by public
funds, suggests that university licensing policies should favor relatively broad
dissemination of the patented results of academic research. In our view, U.S.
universities should pursue nonexclusive licensing agreements for the fruits of
publicly funded research whenever possible. That is, the presumption should
be broad dissemination: exclusivity should be employed in licensing agree¬
ments only when it is clear that the technology would not be commercialized
without an exclusive agreement. Nonexclusive licensing agreements need not
result in the sacrifice by universities of significant financial returns. As we have
pointed out, some of the most important sources of licensing revenues for uni¬
versity licensors during the 1980s and 1990s were licensed on a nonexclusive
basis to a large number of firms.
Ultimately, U.S. universities, no less than universities in other nations, will
retain their privileged institutional status as entities that deserve extensive
public financial support and prestige to the extent that they are seen by the
public as serving its broad interests. The single-minded pursuit of revenue
through patent licensing is not always consistent with this lofty position (in¬
deed, this was one reason for the historical reluctance of many U.S. universi¬
ties to assume a direct role in managing institutional patenting and licensing),
and universities must take care to ensure that their core missions are not put
at risk. The interinstitutional competition that has contributed to the remark¬
able dynamism and excellence of many U.S. universities must not be allowed
CONCLUSION

192 to produce a “race to the bottom,” in which institutions focus on revenue-

maximizing strategies to the detriment of education and research.
Our assessment of the Bayh-Dole Act emphasizes the role and responsibil¬
ities of universities, as well as those of public policymakers. U.S. universities
must exercise considerable responsibility and political sensitivity in managing
their intellectual property if the remarkable achievements of the past century
are to be sustained in the new century.
 Notes

CHAPTER 2

1. Portions of this chapter draw on Rosenberg and Nelson (1994).

2. “[Federal] public policy tended to strengthen the competitive market in higher
education by weakening any central authority that could substitute regulations and
standards for competition. It accomplished this by driving decisions downwards and
outwards, by giving more resources and discretion to the consumers of education
and the institutions most responsive to them. It increased the power of the states in re¬
lation to the federal government, as in the defeat of the University of the United States
and [passage of] the first Morrill Act; the power of the institutions in relation to state
governments, as in the Dartmouth College case and the Hatch Act; and the power of
the students in relation to their own institutions, as in the GI Bill and the Higher Ed¬
ucation Act of 1972” (Trow, 1991, p. 274).
3. “[T]he absence of stable and assured support for our [U.S.] colleges and univer¬
sities from the State or an established church has forced them to look for support from
a multiplicity of other sources, notably student tuition and fees, contributions from
alumni and other friends, and especially from wealthy benefactors and institutions. A
constant concern for financial survival and resources for development and growth en¬
gendered a steady sensitivity to the needs and interests of this varied support commu¬
nity, reflected in the enormous diversity of activities and services that our institutions
have provided” (Trow, 1991, p. 159).
4. “Unlike European universities, which often appointed their own star graduates
for faculty posts and rarely recruited faculty from other universities, American univer¬
sities competed in a diverse academic market. European universities featured great in¬
stitutional loyalty and continuity of professional personnel. Senior faculty at the lead¬
ing American universities, however, typically won their posts through participation in
a competitive national academic marketplace’ (Graham and Diamond, 1997* P- 2°)-
5. “Any chemical process, on whatever scale conducted, may be resolved into a
coordinated series of what may be termed unit actions, such as pulverizing, mixing,
NOTES TO CHAPTER 2

194 heating, roasting, absorbing, condensing, lixiviating, precipitating, crystallizing, filter¬

ing, dissolving, electrolyzing and so on. The number of these basic unit operations is
not very large and relatively few of them are involved in any particular process. Chem¬
ical engineering research ... is directed toward the improvement, control and better
coordination of these unit operations and the selection or development of the equip¬
ment in which they are carried out. It is obviously concerned with the testing and the
provision of materials of construction which shall function safely, resist corrosion, and
withstand the indicated conditions of temperature and pressure” (Little, 1933,
pp. 7-8).
6. These paragraphs draw on Mowery (1999).
7. “To say that work like that of Durand and Lesley goes beyond empirical data
gathering does not mean that it should be subsumed under applied science ... It in¬
cludes elements peculiarly important in engineering, and it produces knowledge of a
peculiarly engineering character and intent. Some of the elements of the methodology
appear in scientific activity, but the methodology as a whole does not” (Vincenti, 1990,
p. 166).
8. Vincenti (1990, p. 158) astutely argues that “in formulating the concept of propul¬
sive efficiency, Durand and Lesley were learning how to think about the use of pro¬
peller data in airplane design. This development of ways of thinking is evident through¬
out the Stanford work; for example, in the improvement of data presentation to
facilitate the work of the designer and in the discussion of the solution of design prob¬
lems. Though less tangible than design data, such understanding of how to think about
a problem also constitutes engineering knowledge. This knowledge was communi¬
cated both explicitly and implicitly by the Durand-Lesley reports.”
9. Aiken benefited from the financial and technical support of IBM in developing
the “Mark I” computer, introduced in 1944. Aiken’s device, which had powerful com¬
putational capabilities, nevertheless was still electromechanical, not electronic, in its
basic design.
10. Hence Stokes’s categorization of research with the dual aims of deep under¬
standing and practical value in the physical sciences as well as in the medical sciences
as research in “Pasteur’s Quadrant” (Stokes, 1997).
11. Some of the large R&D programs that were mounted under the exigencies of war
did generate huge societal benefits in the postwar years. A “crash” wartime program
made penicillin, perhaps the greatest medical breakthrough of the twentieth century,
widely available for the treatment of infectious diseases. Another large-scale program
made low-cost synthetic rubber widely available and had lasting effects on the U.S.
chemicals and petrochemicals industries. And wartime research in microelectronics,
directed toward military goals such as improvement of radar systems, provided a rich
legacy of enlarged technological capabilities to the postwar world.
12. This calculation includes reported performance by “universities and colleges”
and “university FFRDCs” in the NSF “National Patterns of R&D Resources” tables.
Since 1995, NSF data (National Science Board, 2002) indicate that the university
share of total national basic research performance has declined to 49 percent in 2000.
Mowery (2002) discusses these trends, which appear to reflect the sharp increase in
industry-funded and -performed basic research in the late 1990s. This growth in
 NOTES TO CHAPTER 2

industry-funded and -performed basic research is likely not to be sustained through the
195
economic downturn of 2000—2001, although no reliable current data are available as of
this writing.
13. The best known of these was the GI Bill, which provided substantial financial
support to all veterans who enrolled in college-level educational programs; others in¬
clude graduate fellowships supported by NSF and AEC funds, training fellowships
from the National Institutes of Health, and the National Defense Education Act
fellowships.
14. See National Research Council (1982); Okimoto and Saxonhouse (1987). Sharp
(1989) argues that the less prominent role played in scientific research by European
universities has contributed to the slower growth of small biotechnology firms: “A re¬
searcher at a CNRS laboratory in France, or at a Max Planck Institute laboratory in
Germany, is the full time employee of that institution. As such his/her prime respon¬
sibility is to public, not private science. Moreover, as a full time employee, he/she will
not find it easy to undertake the ‘mix’ of research frequently undertaken by an Ameri¬
can professor, who combines an academic post with consultancy in the private sector.
Indeed the tradition of funding US academic posts for only nine months of the year,
expecting the academic who wishes to carry out research in the summer to raise re¬
search funds to meet the remaining three months of salary, explicitly encourages the
entrepreneurial academic. In stark contrast, his/her German opposite number at a
Max Planck Institute will find all research costs, including staff and equipment, met as
part of institutional overheads. The opportunity cost of leaving such a research envi¬
ronment for the insecurity of the small firm is all the greater since, once off the aca¬
demic ladder in West Germany, it is more difficult to climb back on again. The same
goes for the opposite number in France, and with the additional disincentive that
French researchers are civil servants and dropping out of the system means both losing
security of tenure/accumulated benefits and difficulty in re-entry should the need
arise. In the circumstances, it is not perhaps so surprising that few spin-offs from pub¬
lic sector research arise, nor, for that matter, that in Europe most such spin-offs are to
be found in the UK, where the organisation of academic science most closely matches
that of the US. In the UK, it is notable that—with the exception of Celltech and the
Agricultural Genetics Company (AGC) —most of the spin-offs from biotechnology
have come from the universities” (pp. 12-13).
15. One might instead argue that the weakening of university-industry research link¬
ages during a significant portion of the postwar period was the real departure from his¬
torical trends. Hounshell and Smith (1988) cite a 1945 memo from Elmer Bolton, di¬
rector of what was to become the Du Pont Company’s central research laboratory, that
made a case for greater self-reliance by the firm in its basic research: “Three things were
necessary: Du Pont had to strengthen its research organizations and house them in
modern research facilities; the company’s existing processes had to be improved and
new processes and products developed; and fundamental research, which will serve as
a background for new advances in applied chemistry, should be expanded not only in
the Chemical Department but should [also] be increased in our industrial research
laboratories and the Engineering Department.’ Bolton stressed that it was no longer
‘possible to rely to the same extent as in the past upon university research to supply this
NOTES TO CHAPTER 3

196 background so that in future years it will be necessary for the Company to provide this
knowledge to a far greater extent through its own efforts.’ To ‘retain its leadership’ Du
Pont had ‘to undertake on a much broader scale fundamental research in order to pro¬
vide more knowledge to serve as a basis for applied research’” (p. 355). Swann (1988,
pp. 170-81) also argues that research links between U.S. universities and the pharma¬
ceuticals industry weakened significantly in the immediate aftermath of World War II,
in part as a result of vastly increased federal funding for academic research in the
health sciences.
16. As we noted, the survey by Cohen, Nelson, and Walsh (2002) included queries
about the importance of different channels for access to research within universities
and government laboratories, issues not covered in the Yale survey. These differences
between the surveys may produce some differences in responses to the Yale and
Carnegie-Mellon surveys, although this issue merits further analysis.

chapter 3

1. McKusick (1948) suggests that “at the beginning of the Depression decade . . .
[t]wo different factors had turned university attention to the patent problem: first, a
steady growth of research sponsored cooperatively by industry demanded a generally
applicable policy toward resulting patents, and, secondly, spectacular inventions on
university campuses demanded immediate concern with patent policy” (p. 212). He ar¬
gues in addition that the frequency of such “spectacular inventions” increased dra¬
matically over the 1930s. Sevringhaus (1932), Gregg (1933), Henderson (1933), and Gray
(1936) all provide contemporary accounts of these debates.
2. The report, titled The Protection by Patents of Scientific Discoveries, deals
with patents on “scientific discoveries” generally, not just discoveries made in uni¬
versities. But most of the discussion is concerned with patenting by universities and
other nonprofit institutions, as opposed to patenting by scientists working in firms
or independently.
3. The committee's cautious language indicates an awareness that, in some in¬
stances, commercial development could proceed apace without patent protection, a
distinction that was largely ignored in the debates that led to the passage of Bayh-Dole.
See Eisenberg (1996) and Chapter 5 for further discussion.
4. The committee quoted Hoskins and Wiles (1921), who argued that “there is at
large a type of engineer commonly called a ‘patent pirate,’ who thrives by monopoliz¬
ing the practical applications of the abstract discoveries of others. The patent pirate is
a menace to industry and a parasite on the community. Nothing would so hamper his
activities as to have the real discoverer take out broad patents in every case” (p. 691).
5. The committee’s report quoted Sevringhaus (1932, pp. 233-34): “The public is
thereby protected against certain ruinous types of exploitation. Assurance can be
gained that technical processes are used in dependable ways. Even the publicity may
be kept on a satisfactorily high plane.”
6. Specifically, the report discussed concerns “that patenting will involve scientists
in commercial pursuits and leave little them little time for research” (p. 9), “that
 NOTES TO CHAPTER 3

patenting leads to secrecy” (p. 11), “that a patent policy will lead to a debasement of re¬ 197
search” (p. 11), and “that the policy of obtaining patents will lead to ill feeling and per¬
sonal jealousies among investigators” (p. 13).
7. The AAAS committee s report characterized the following argument by Zinsser
(1927) as typical: “The invention of an improvement in the mechanism of automobiles,
or of a shoe-buckle, concerns matters of convenience or luxury, and can be dispensed
with easily by those who are forced to do without them. The relief of the sick and the
prevention of unnecessary sorrow by the maintenance of individual and public health
are matters in a different category. As soon as we are in possession of the knowledge of
principles or methods which can contribute to these purposes their free utilization be¬
comes a public necessity; and any procedure which inhibits their most rapid and ef¬
fective application to the needs of the community would seem to us as unjustified as
the cornering of the wheat market or the patenting of the process of making bread”
(p. 154). The widespread reluctance to patent biomedical discoveries is especially par¬
adoxical in view of the fact that biomedical patents proved to be one of the leading
sources of licensing revenues for U.S. universities before and after the Bayh-Dole Act.
8. Citing the example of Frederick Banting and Charles Best of the University of
Toronto, inventors of insulin-production techniques who patented their discovery,
Steenbock noted that they sought patent protection for their inventions to ensure that
“the public is protected against the manufacture of poor preparations and is also pro¬
tected against the extortionate charges” and to “avoid the possibilities of misusing their
discovery which not only would have retarded the further development and use of this
product, but would also have resulted in causing untold suffering among diabetic pa¬
tients” (cited in Apple, 1996, p. 36). Apple (1989) argues that Steenbock also wished to
patent his invention in order to prevent margarine producers from acquiring the pro¬
cess, thus protecting the regions dairy interests (WARF was sued by the federal gov¬
ernment in 1943 over ^ refusal to license margarine producers). Blumenthal, Epstein,
and Maxwell (1986) conjecture that Steenbock was motivated by the experiences of his
colleague Dr. Stephan Babcock, who had developed a method for determining the but-
terfat content of milk (see Chapter 2) and, as was then standard practice, did not file for
patent protection on the invention. But the absence of a patent meant that Babcock
could not prevent low-quality producers from flooding the market with “Babcock
testers,” eventually discrediting the method altogether.
9. Spencer (1939) reports that by 1936, the Steenbock patents had earned more than
$6.7 million (in 1996 dollars) for WARF.
10. “During the last decade, the number of universities which have adopted well de¬
fined policies in regard to inventions made by their employees, or with the aid of uni¬
versity equipment, has steadily increased. Additionally, there are today many other uni¬
versities actively considering the adoption of a ‘Patent Policy’ and it is likely that
another ten years will see such a policy in every school in the country that offers scien¬
tific or engineering courses” (Spencer, 1939, p. 1).
11. Of these 85 policies, 3 were adopted prior to 1930; 6 between 1931 and 1935; 12 be¬
tween 1936 and 1940; 19 between 1941 and 1945; 25 between 1946 and 1950; and 20 be¬
tween 1951 and 1955.
12. Columbia’s policy thus stated that “it is recognized, however, that there may be
NOTES TO CHAPTER 3

198 exceptional circumstances where the taking out of a patent will be advisable in order
to protect the public. These cases must be brought to [the university administration]
for its consideration and approval” (cited in Palmer, 1962, p. 175).
13. The Research Corporations actions reflected its declining financial viability, as
revenues lagged behind operating costs. As we note in Chapter 4, the corporation’s ac¬
tivities in this area appear to have contributed to its eventual demise.
14. Entry is defined as the first year during which AUTM member institutions re¬
ported devoting at least one 0.5 FTE to officially designated “technology transfer” ac¬
tivities (see Chapter 7 for further discussion of this measure).
15. The population of university assignees includes the uilion of all assignees desig¬
nated as “universities” in the Case Western—NBER U.S. Patents database (see Hall,
Jaffe, and Tratjenberg, 2001) and all institutions designated as Research or Doctoral
Universities in the Carnegie Commissions 1973 report. To collect patents assigned to
these universities, we searched the U.S. Patent and Trademark Offices annual reports
for patents assigned to these institutions from 1920 to 1965 (there were none in 1920,
other than patents assigned to the Research Corporation), the DIALOG Corporations
Patents/CLAIMS database for patents from 1963 to 1980, and the Case Western-
NBER database for patents from 1975 to 1980. Searches of multiple sources for the
1963-65 and the 1975-80 periods produced similar results. Unless otherwise specified,
“university patents” refers to this entire sample of patents.
16. It is noteworthy that the increase in university patenting over the 1970s occurred
while overall U.S. patenting was decreasing. Indeed, U.S. universities’ share of domes¬
tically assigned U.S. patents grew from 0.2 percent in 1963 to 0.7 percent in 1979, be¬
fore the passage of the Bayh-Dole Act.
17. Data on total academic R&D were obtained from National Science Board
(2000), Appendix Table 4-4.
18. As we discussed above, two Ivy League institutions—Princeton and Colum¬
bia-signed IAAs with the Research Corporation by 1940. Some patents emanating
from these universities between 1940 and 1945 may have been assigned to the Research
Corporation.
19. Data in Fishman (1996) suggest that MIT generated more patents than any other
private university in 1940 and 1945, although under its IAA the patents were assigned to
the Research Corporation rather than to MIT.
20. Stanford signed an IAA with the Research Corporation in the early 1950s.
21. These calculations exclude Research Corporation patents.
22. To collect patents for these institutions, we classified multicampus universities
by the maximum ranking of any campus in the system, a procedure that is necessary
because patents are typically assigned at the university, not the campus, level. This pro¬
cedure yielded 89 unique Research Universities (49 RUis and 40 RU2S) and 73 Doc¬
toral Universities. Within these groups, 79 of the Research Universities (all 49 of the
RUis, 30 of the RU2S) and 40 of the Doctoral Universities had at least one patent dur¬
ing the 1948-80 period.
23. The basic trends discussed below for “incumbent” and “entrant” university
patenters are not affected by using different threshold values for the number of patents
in defining these groups.
24. Of the 40 universities that had IPAs during the 1970s, 23 had an IPA only with
 NOTES TO CHAPTER 4

HEW, 1 only with NSF, and 16 with both agencies. 199

25. The technological areas we use here are based on a concordance from U.S. pat¬
ent classes to technological fields developed by Adam Jaffe and reported in Jaffe, Fo¬
garty, and Banks (1998).

chapter 4

1. Since Cottrell was a member of this committee, the similarity in views is not
surprising.
2. In a 1911 speech before the American Chemical Society, Cottrell argued that
without an entity charged with advertising, licensing, and developing academic in¬
ventions, such inventions might well fail to find any commercial applications: “[A] cer¬
tain amount of intellectual by-products are going to waste at present in our colleges and
technical laboratories all over the country. There is a great deal of work that is being
developed to a practical or semi-practical standpoint that dies right there because the
men ... do not want to dip into the business side of technology and go out into practi¬
cal fields and the work has not come to the point of economic usefulness that is de¬
sired” (quoted in Cameron, 1993, p. 166).
3. After 1912, Cottrell was no longer active as an inventor in the precipitation field.
Beginning in 1911, he served as chief physical chemist at the U.S. Bureau of Mines, ris¬
ing to become director of the bureau. He subsequently was director of the Fixed Ni¬
trogen Research Laboratory of the U.S. Department of Agriculture during the 1922-27
period, where he oversaw the successful reverse engineering of the Haber-Bosch pro¬
cess that underpinned the synthetic ammonia industry (Mowery and Rosenberg, 1998).
In 1935, Cottrell founded Research Associates to conduct research in other fields, but
this endeavor collapsed in 1938. Cottrell devoted the last decade of his life to research
on nitrogen fixation (Cameron, 1993).
4. From the “Correspondence A-Z: 1918” folder in the Research Corporation col¬
lection at the Smithsonian Institution Archives.
5. The corporation acted “as the intermediary between the inventor and the manu¬
facturer, which can subject alleged discoveries inventions to practical tests and render
a judicial opinion to the manufacturer which can be regarded as trustworthy” (New
York Sun, July 24,1917).
6. Nonetheless, the 1950 Annual Report noted that “although there was always the
expectation that another precipitation business might evolve from these projects, none
did so, and from a financial point of view at least, the operations were distinctly un¬
successful” (p. 39).
7. A 1922 report by a Committee of the Association of Land Grant Colleges on pat¬
ent policies at Engineering Experiment Stations noted that educational and research
institutions are not in a position to do commercial business, such as would be involved
in owning patents, to defend the patent owned, or even to negotiate successfully for the
disposal of rights under patents. Unbusinesslike methods in handling patents by edu¬
cational institutions would hinder rather than promote discoveries or inventions. It will
be necessary to have some outside organization handle the details with reference to the
NOTES TO CHAPTER 4

200 disposition of the patents. The Research Corporation of New York has organized to
handle the patents of those who are not in a position to exploit them” (Committee on
Uniform Patent Policies in Land Grant and Engineering Experiment Stations, 1922,
P- 234)-
8. A September 3, 1937, letter from the Research Corporations president, Howard
Poillon, to the president of MIT, Karl Compton, notes that “our important precipita¬
tion patents are expiring, and we are trying to develop new activities of such type that
will permit us to use all of our staff, when, due to competition, our precipitation busi¬
ness will be decreased” (MIT Archives, “Research Corporation” AC125, Folder 57-12).
9. This section is based on the discussions in Fishman (1996) and Etzkowitz (1994).
10. MIT Patent Committee Statement of Patent Procedure. MIT Archives, AC64,
Box 1.
11. On the advice of Karl Compton, the Research Corporation opted not to handle
patents resulting from National Defense Research Council (NDRC) and OSRD re¬
search. The first reason was fear of litigation: since much of the federal research effort
during World War II funded parallel research efforts in specific technology fields by
several research organizations, Compton predicted that “these particular fields will be
the subject of some exceedingly active patent litigation after the war in order to un¬
tangle the various claims for priority.” Second, because so much of MIT’s wartime re¬
search centered around large Institute laboratories that were staffed by scientists and
engineers from all over the world, at the end of the war “many of the inventors will scat¬
ter over the world and not be readily able to advise in case of Patent Office actions or
other cases when their services are essential. In such cases the Corporation would be
called on to employ others who may be expert in the subject, which will be both costly
and time consuming” (MIT Archives, “Research Corporation” AC125, Folder 57-10).
This latter concern anticipates some of the problems with centralized patent manage¬
ment that we discuss below.
The Research Corporations reduced activities during wartime also reflected the
fact that many of its staff members took leaves to aid in the war effort. An especially se¬
rious loss was the departure of the head of the Boston office, Carroll Wilson, who went
to work at the Office of Scientific Research and Development as Vannevar Bush s chief
of staff.
12. On the other hand, the Research Corporation-MIT relationship had not
yielded significant returns to MIT either by this time. See Fishman (1996).
13. “Its further, and even larger responsibility, was to extend and expand this phase
of the Corporations activities, to further develop the concept of handling the patent
problems of educational institutions, and to do so in a manner that would add funds
available to the Corporation for its program of grants-in-aid, that is, operate at a profit”
(1951 Annual Report, p. 40).
14. Similar concerns led several other Research Corporation client universities to
administer their nonbiomedical patents in-house, rather than submit them to the cor¬
poration, during the 1950s.
15. From its commercial introduction in the mid-1950s until the expiration of the fi-
 NOTES TO CHAPTER 4

nal patent in 1975, nystatin was the Research Corporations biggest single source of cu¬ 201
mulative licensing income.
16. When revenues from hybrid seed corn, which during this period included pay¬
ments of back royalties as a result of a legal settlement, are excluded from this calcula¬
tion for 1970, the share of royalties accounted for by donated inventions declines to
82 percent.
17. Forresters invention resulted from his involvement in the Whirlwind computer
project, originally funded by the U.S. Navy to produce a flight simulator.
18. According to its 1968 Annual Report (p. 37), the Research Corporations policy
“requires that it resort to litigation when patents assigned are believed to be infringed,
and negotiation cannot bring about recognition of its rights.” This commitment to lit¬
igation as a matter of policy (rather than discretion) reflects another cost of large-scale
patent administration: the need to defend all patents (even marginal ones) in order to
avoid developing a reputation of being “soft” and thus inviting future infringement.
19. Another example of such conflicts that involves a clash between university ad¬
ministrators and faculty is the recent litigation between two UC faculty inventors
(Jerome Singer and Lawrence Crooks) and the UC system over the terms of the li¬
censing agreement for these inventors’ MRI patents. Singer and Crooks filed suit
against the University of California, accusing the university of negotiating very low li¬
censing fees with Pfizer Medical Systems in exchange for the firm’s financial support
of UC research activities. The inventors were awarded $2.3 million in back royalties in
a 1996 jury verdict that was upheld on appeal in 1997 (Chiang, 1997).
20. The difficulty of transferring expertise across technology fields is seen in this
statement from the 1968 Annual Report (p. 6): “The evaluation of invention disclosures
is increasingly more complex each year. Not only is the available literature, both tech¬
nical and patent, rapidly increasing, but technology itself is ever more complicated and
requires more concentrated study for proper understanding and evaluation. Only
through considerable use of outside consultants and patent attorneys were we able to
handle this volume. Even so the average time required to complete an evaluation and
the average number of disclosures in process at any given time was [sic] slowly increas¬
ing. The technological areas where these effects were most marked were in the me¬
chanical and the electrical-electronic fields.”
• 21. Legal costs here include expenditures on attorneys for evaluation services and
for patent prosecution, as well as litigation expenses. The latter category also con¬
tributed to the increased costs of patent management over the 1960s and 1970s.
22. This expanded visitation program was motivated by more than simply the cor¬
poration’s desire to maintain close links with academic inventors. During this period,
the Research Corporation also was expanding its efforts to assist universities in adapt¬
ing to changes in government patent policy, as we discuss below.
23. The corporation noted in 1966 that “the result from increased visitation program
and the wider dissemination of information about our patent assistance services has
been a steady increase in the number of inventions submitted for evaluation ... In spite
of the addition of one associate in the past two years, this increased load has required a
substantial increase in the use of technical consultants and patent attorneys. It appears
NOTES TO CHAPTER 4

202 certain that this activity will result in a substantial increase in retained income from in¬
ventions submitted under our agreements with institutions” (1966 Annual Report, p. 3).
These efforts also required more technical experts, and during 1968-73 the technical
staff of the invention administration program nearly tripled (1978 Annual Report, p. 25).
24. See also Jensen and Thursby (2001) on the importance of inventors’ tacit knowl¬
edge for technology transfer. As Lowe (2002) points out, this need for inventor involve¬
ment also is an important reason for the foundation by university inventors of firms to
commercialize their inventions.
25. Lamoreaux and Sokoloff (2002) emphasize the importance of geographic prox¬
imity in the growth of invention intermediaries, who were concentrated in regions of
the United States with relatively high proportions of independent inventors in the late
nineteenth and early twentieth centuries. The analysis of data from the 1970s and 1980s
by Mowery and Ziedonis (2001) found that geographic proximity plays a more promi¬
nent role in the licensing of university patents than in the “knowledge spillovers” rep¬
resented by citations to university patents.
26. The corporation noted that “responsibility for carrying out the procedures must
be assigned unequivocally to an individual or an office in the administrative branch of
the institution” (Marcy, 1978, p. 5).
27. The Research Corporations 1968 Annual Report (p. 3) noted that “royalties re¬
ceived by Research Corporation from the institutional inventions is [sic] only about
one-third the expense of serving these institutions.”
28. Unfortunately, we can say little about cherry-picking by faculty members at
client universities (as opposed to by the universities themselves) without collecting
considerably more data that would enable us to identify the characteristics of all pat¬
ents held by these individual faculty members based on searching for their names in
the records of the U.S. Patent and Trademark Office.
29. Although withholding of valuable inventions by clients is never mentioned as a
possible cause of the lack of profitability of the institutional patent program during the
numerous discussions of this problem in the annual reports of the 1950s and 1960s, the
issue is discussed frequently in the corporations annual reports during the 1970s.
30. As a further check on the degree to which clients were cherry-picking, we com¬
pared the quality of the Research Corporations patents and the independent patents
by client universities by using the number of subsequent citations to a patent as a proxy
for the quality of the patent. For each of the six years discussed above, we regressed the
number of citations received after its issue by a patent on technological field dummies
and a dummy variable indicating whether the patent was assigned to the Research Cor¬
poration or to one of its client universities. In none of the years was the coefficient on
the “assignee” dummy statistically significant at conventional levels. Indeed, the only
year at which the coefficient was significant even at the 15 percent level is 1975; in that
year the sign is negative, suggesting that clients’ patents were of higher “quality” than
those assigned to the Research Corporation. The results of this analysis suggest little ev¬
idence of systematic cherry-picking behavior before the mid-1970s, consistent with the
discussion above.
31. Among the other alternatives considered but not adopted in 1979 was restriction
of the invention administration program to a limited set of technological fields. This
 NOTES TO CHAPTER 5

proposal, which reflected a loss of confidence in the feasibility of cost reductions 203
through a technologically diversified patent licensing program, was not adopted be¬
cause “it would mean losing contact with other disciplines and narrowing our interface
within the scientific community” (Research Corporation, 1979, p. 83).
32. The Corporations 1964 Annual Report (p. 46) noted that “the program of more
frequent visits to institutions with which Research Corporation has patent agreements,
begun in 1963, was continued throughout 1964 and at an expanded pace in the latter
months of the year as the result of institutions accepting the foundations offer of assis¬
tance in matters relating to the new Government patent policies.”
33. “Our staff members have been making themselves freely available to these ad¬
ministrators during the year, and have undertaken to establish mutually helpful rela¬
tionships with as many of these key research administrators as possible” (1967 Annual
Report, p. 9).
34. Entry into patenting during the 1970s was dominated by institutions that were
not active Research Corporation clients or that were not Research Corporation clients,
that is, those with little or no history of patenting. This characteristic of entrant insti¬
tutions once again suggests that increased interinstitutional dispersion of funding,
rather than cherry-picking, inefficiencies at the corporation, or increased enthusiasm
for patenting, drove institutional entry into this activity during the 1970s.
35. With this point in mind, it is interesting to note that the survey of university li¬
censing officers in Jensen and Thursby (2001) reported that “revenue” was rated as “ex¬
tremely important” by nearly 75 percent of the respondents, ranking as the most im¬
portant objective of university licensing offices in their survey.

CHAPTER 5

1. See, for example, Bush (1943) and Bush’s testimony on the Kilgore bill, portions
of which are reprinted in U.S. Department of Justice (1947)- Most of the debate over
this issue occurred during congressional hearings on Kilgore s proposal for postwar sci¬
ence policy. Although Bush actively opposed the provisions in the Kilgore bill that
would give title to the government, his own proposal for postwar science policy (out¬
lined in Science: The Endless Frontier) avoided specifics, recommending only that “the
public interest will normally be adequately protected if the Government receives a
royalty-free license,” that is, if contractors retain title (“Letter of Transmittal” in Bush,
r945). Although the disagreements during the Bush-Kilgore debates centered on the ef¬
fects of government patent policy on private contractors, Bush also supported patent¬
ing by university researchers, as we noted in Chapter 2.
2. NASA administrator T. Keith Glennan, commenting on the differences between
NASA and DOD policies, observed that “two such contradictory patent policies, fol¬
lowed by government agencies working in closely related fields of research and devel¬
opment, can be detrimental to the kind of cooperation that we must have from indus¬
try” (“Glennan Asks Review of NASA Patent Policy,” 1959, p. 33)-
3. Memorandum and Statement of Government Patent Policy, 28 Federal Register
10,943-46 (1963); Memorandum and Statement of Government Patent Policy, 36 Fed¬

eral Register 16,886 (1971).

NOTES TO CHAPTER 5

204 4. See also National Science Board (2002), Table 4-3. Universities and colleges (ex¬
cluding federally funded R&D Centers) performed 6 percent of federally funded R&D
in 1953, the first year for which the National Science Foundation has published data.
This share had grown to more than 33 percent of federal R&D spending by 2000.
5. President Kennedy’s 1963 “Memorandum on Government Patent Policy” had
charged the FCST with analyzing the effects of different patent policies on utilization
and commercialization of government funded research.
6. The firms screened these compounds for potential biological activity.
7. HEW had instituted an IPA program in 1953, and eighteen universities had ne¬
gotiated IPAs with the agency by 1958. But after 1958, no additional requests for IPAs
were approved by HEW because “opinions of responsible agency officials differed con¬
cerning the value of such agreements” (GAO, rg68, p. 24). Pharmaceutical companies
also complained that these IPAs were ambiguous about the scope of exclusive rights
that licensees could retain.
8. The purpose of the HEW review was “to make sure that assignment of patent
rights to universities and research institutes did not stifle competition in the private sec¬
tor in those cases where competition could bring the fruits of research to the public
faster and more economically” according to the testimony of Comptroller General
Elmer Staats during the Bayh-Dole hearings (U.S. Senate Committee on the Judiciary,
1979a, p. 37)•
9. A patent attorney formerly employed by Purdue University, Norman Latker, was
also the chief architect of the changes in HEW’s patent policies in 1968 and was fired
from HEW after denouncing the agency’s subsequent review of these policies. He re¬
turned to the HEW’s patent office in 1978, after his dismissal was overturned by a civil
service review board on procedural grounds. Reporting on these events in Science,
Broad (1979a, p. 476) noted: “The reinstatement is timely. Support is now building for
the Bayh-Dole patent bill, and Latkers return to the HEW is seen by many university
researchers and patent transfer fans, to whom Latker is something of a hero, as a shot
in the arm for their cause.”
10. Identical legislation (H.R. 2414) was introduced in the House of Representatives
by Rep. Peter Rodino (D-N.J.) in 1979.
11. Another IPA restriction dropped in the Dole-Bayh bill was the requirement that
grantees and contractors try first to offer nonexclusive licenses. According to an
anonymous aide to Senator Bayli who was quoted by Henig (1979, p. 281), “‘It’s too
hard and inefficient a process. Universities don’t have the financial capability to beat
the bushes and try to find someone who is willing to accept a license on a non¬
exclusive basis.’”
12. A contemporary account noted that limiting the bill to universities and small
businesses was “a tactical exclusion taken to ensure liberal support” (Henig, 1979,
p. 282). A Senate aide commented, “We’d like to extend [the policy] to everybody
but if we did the bill would never have a chance of passing” (Broad, 1979b, p. 474). The
original bill also included several provisions designed to defuse criticism that it would
lead to “profiteering” at the expense of the public interest, including a recoupment pro¬
vision requiring that institutions pay back a share of licensing income or sales to fund¬
ing agencies. The final version of the Bayh-Dole Act eliminated this provision “because
 NOTES TO CHAPTER 5

there was no agreement on whether the funds would be returned to the agencies or to 205
general revenue, or how the collection and auditing functions would be conducted”
and “fears that the costs of the infrastructure required to administer such a program
would exceed the amounts collected.” See http://www.nih.gov/news/oyoioiwyden.htm.
13. For example. Dole opened the Senate Judiciary Committees hearings on the
bill by observing that “the damaging impact of the Federal patent policy on the econ¬
omy is dramatic. That we have lost our leadership role to Japan in the field of elec¬
tronics and shipbuilding is no accident” (U.S. Senate Committee on the Judiciary,
x979a, P- z8)-
14. See Harold Bremer, “Public Patents, Public Benefits?” submitted with his testi¬
mony during the Senate Judiciary Committees Bayh-Dole Hearings (U.S. Senate
Committee on the Judiciary, 1979a).
r5. See, for example, the opening statements by Senators Bayh, Dole, and Orrin
Hatch (R-Ut.), as well as the testimony by Elmer Staats, Walter Syniuta, and Betsy
Ancker-Johnson (U.S. Senate Committee on the Judiciary, 1979a).
16. See, for example, the testimony by Hector Deluca, Frederick Andrews, and
Harold Bremer during the Senate Judiciary Committees Bayh-Dole Hearings (U.S.
Senate Committee on the Judiciary, 1979a).
17. The Senate Judiciary Committees report on S. 414 noted: “A number of wit¬
nesses also pointed out to the committee that when government agencies retain title to
inventions made by nonprofit organizations or small business contractors there is ab¬
solutely no incentive for the inventor to remain involved in the possible development
of the patentable discovery. Virtually all experts in the innovation process stress very
strongly that such involvement by the inventor is absolutely essential, especially when
the invention was made under basic research where it is invariably in the embryonic
stage of development” (U.S. Senate Committee on the Judiciary, i9798> P- 22)-
r8. Much of the floor debate in both chambers focused on whether large businesses
should be allowed to retain rights to patents resulting from federal contracts. Rep. Rob¬
ert Kastenmeier (D-Wis.), a supporter of extending rights to large business contractors,
noted that “rather than hold hostage [the] non-controversial areas [that is, allowing uni¬
versities and small businesses to retain title], I think we have no real option but to move
forward with this [the bill as amended] and send it to the White House” (U.S. House
of Representatives, 1980, p. 30560.) The original bill passed by the House (H.R. 6933)
contained the same provisions as the Senate bill introduced by Senators Bayh and Dole
but also allowed large firms to obtain “field of use” exclusive licenses for inventions de¬
veloped with government funds. Senate amendments to this bill eliminated that pro¬
vision, and the House subsequently passed a bill identical to that passed by the Senate.
19. A journalist covering the hearings observed that although the Dole-Bayh bill is
receiving nearly unprecedented support, some congressional aides point out that it still
leaves unanswered fundamental questions about patents in general and patents on uni¬
versity campuses in particular” (Henig, 1979, p. 284).
20. For details, see the Council on Government Relations (1999).
21. Memorandum to the Heads of Executive Departments and Agencies: Govern¬
ment Patent Policy, Pub. Papers 252 (February 18, 1983); Executive Order 12591
(4/10/87); codified at 3 C.F.R. 221.
NOTES TO CHAPTER 6

206 22. Trademark Clarification Act of ^84, PL 98-620. Among other things, this
amendment also allowed contractors at Government Owned, Contractor Operated
Labs (GOCOs) to retain title to federally funded inventions.
23. According to Katz and Ordover (1990), at least fourteen congressional bills
passed during the 1980s focused on strengthening domestic and international protec¬
tion for intellectual property rights, and the Court of Appeals for the Federal Circuit,
created in 1982, has upheld patent rights in roughly 80 percent of the cases argued be¬
fore it, a considerable increase from the pre-1982 rate of 30 percent for the federal
bench.
24. “Regulatory reform in the United States in the early 1980s, such as the Bayh-
Dole Act, have [sic] significantly increased the contribution of scientific institutions to
innovation. There is evidence that this is one of the factors contributing to the pick-up
of US growth performance” (OECD, 2000, p. 77).
25. “In 1980, the enactment of the Bayh-Dole Act (Public Law 98-620) culminated
years of work to develop incentives for laboratory discoveries to make their way to the
marketplace promptly, with all the attendant benefits for public welfare and economic
growth that result from those innovations. Before Bayh-Dole, the federal government
had accumulated 30,000 patents, of which only 5% had been licensed and even fewer
had found their way into commercial products. Today under Bayh-Dole more than 200
universities are engaged in technology transfer, adding more than $21 billion each year
to the economy” (Hasselmo, 1999, p. 3).
26. “In the 1970s, the government discovered the inventions that resulted from pub¬
lic funding were not reaching the marketplace because no one would make the addi¬
tional investment to turn basic research into marketable products. That finding re¬
sulted in the Bayh-Dole Act, passed in 1980. It enabled universities, small companies,
and nonprofit organizations to commercialize the results of federally funded research.
The results of Bayh-Dole have been significant. Before 1981, fewer than 250 patents
were issued to universities each year. A decade later universities were averaging ap¬
proximately 1,000 patents a year” (Dickinson, 2000, p. 2).
27. The Bayh-Dole Act turned out to be the Viagra for campus innovation. Uni¬
versities that would previously have let their intellectual property lie fallow began fil¬
ing for and getting patents at unprecedented rates. Coupled with other legal economic
and political developments that also spurred patenting and licensing, the results seems
nothing less than a major boon to national economic growth” (Zacks, 2000).

chapter 6

1. According to a recent ranking of academic institutions by licensing revenues by

the Association of University Technology Managers (AUTM), Columbia University,
the UC system, and Stanford University were the first, second, and fifth leading insti¬
tutions, earning $90.5 million, $76.4 million, and $37.9 million, respectively. Florida
State University and Yale University were ranked third and fourth, receiving $54.1 and
$38.5 million, respectively (AUTM, 2000, p. 97; all figures in 1996 dollars).
2. According to a March 10,1975, letter from UC president Charles). Hitch to Gov¬
ernor Edmund G. Brown Jr. (University of California Office of the President, Office of
 NOTES TO CHAPTER 6

Technology Transfer Archives), “The possibility of developing a formal patent policy 207
and program was first considered in the University in the Thirties. But the idea did not
achieve full impetus until the war years when the Federal Government began to spon¬
sor research in the University on a large scale and inventions began to be made under
research contracts.”
3. The “Patent Board” was a committee of UC faculty and administrators charged
with oversight of the Patent Office. As revised in 1973, the “University Policy Regarding
Patents” states that “an agreement to assign inventions and patents to The Regents of
the University of California, except those resulting from permissible consulting activi¬
ties without use of University facilities, shall be mandatory for all employees, academic
and nonacademic.” The policy statement goes on to emphasize that “the Regents is
[sic] adverse to seeking protective patents and will not seek such patents unless the dis¬
coverer or inventor can demonstrate that the securing of the patent is important to the
University.” This latter sentiment notwithstanding, UC administrators were actively
seeking patent protection for faculty inventions by the mid-1970s, as the historical data
of the Office of Technology Transfer show.
4. These “independent” licensing offices, which continue to pay a portion of their
revenues to the state government, are in charge of invention disclosures (along with
any revenues or expenses associated with these disclosures) occurring after their foun¬
dation. The UC Office of Technology Transfer, however, continues to collect and re¬
port data on disclosures, patents, and licenses from all UC campuses.
5. Almost simultaneously with this shift in university patent policy, an internal study
by the OTL Advisory Board in 1993 recommended that “OTL need not be constrained
by the principle of‘preference for non-exclusive licensing’” (Stanford University Of¬
fice of Technology Licensing, 1994b, p. 2).
6. Reflecting faculty sensitivity over assignment to the university of all ownership of
all copyrighted material produced under university sponsorship, Stanfords OTL ex¬
empted ownership of books, articles, popular nonfiction, novels, poems, musical
compositions, or other works of artistic imagination which are not institutional works”
from the policy governing software (Stanford University Office of Technology Licens¬
ing, 1994b, p. 1).
7. Figure 6.7 excludes licenses for the Cohen-Boyer patents, which were managed
by Stanfords OTL on behalf of the UC system and Stanford University. Strictly speak¬
ing, since the revenues from these licenses are split between the UC system and Stan¬
ford University, the licenses also should be allocated between the two institutions. Ex¬
clusion of this heavily licensed invention thus understates the growth in the biomedical
share of Stanford and UC licensing agreements during the 1980s in Figure 6.7.
8. Some indication of the relative magnitudes of licensing revenues from these site
licenses," which for some years were administered by the OTL Software Distribution
Center, is given by the following data cited in the 1988-89 report of the OTL, which
separated software licensing revenues into those derived from direct software distri¬
bution through OTL’s Software Distribution Center ($453,581 from 515 use licenses)
and from royalties paid by commercial distributors ($420,000 from 40 distribution
licenses to software firms, computer companies, and publishers)” (Stanford Uni¬
versity Office of Technology Licensing, 1990, p. 4). Unfortunately, we have been
NOTES TO CHAPTER 6

208 unable thus far to consistently separate software licenses between these two channels
of distribution.
9. Portions of this section are based on Crow et al. (1998).
10. Biotechnology inventions accounted for 45 percent of the biomedical inventions
that resulted in patents and nearly 70 percent of the biomedical inventions that were
licensed.
11. Columbia also experienced growth in the number of licenses for software in¬
ventions during the 1980s (a measure that heavily weights inventions licensed on a
nonexclusive basis). Software licenses account for well over 50 percent of Columbia li¬
censing agreements after 1988; the majority of these licenses (420 of a total of 648) were
associated with one software invention. In addition, and similar to the situation at Stan¬
ford University (see below), more than 300 of the 420 licensees for this software inven¬
tion were academic institutions.
12. In order to deal with the problems of “truncation bias” while accommodating
the fact that our data end in 1997, we have imposed a six-year “trailing window” on our
invention disclosures. In other words, the analysis includes issued patents or licenses
only if these events occur within six years after the date of disclosure of the invention.
This convention is used to avoid unfairly biasing the indicators of “productivity” in fa¬
vor of older disclosures, which have much longer time periods during which to pro¬
duce patents or licenses.
13. As we noted earlier, a large percentage of Stanfords software licenses cover low-
royalty “site licenses” at other academic institutions, which may well raise the shares of
Stanford disclosures that yield licensing income without necessarily having a signifi¬
cant effect on overall licensing income. In addition, the Stanford invention disclosures
and licensing data contain a large number of agreements covering “clones” of various
pieces of genetic material —such agreements are less common in the Columbia or UC
data for the period covered by this analysis. These licenses are somewhat more formal
than Materials Transfer Agreements and often involve the payment of modest licens¬
ing fees. But as with the Stanford software licenses, including these agreements in our
data increases the share of disclosures that are licensed or that yield licensing income
without having much effect on overall licensing income.
14. See the case study of Cohen-Boyer in the National Research Council workshop,
Intellectual Property Rights and Research Tools in Molecular Biology (National Re¬
search Council, 1997).
15. Data on Stanford patent applications are unavailable, and we therefore are able
to compute measures using patent applications only for the University of California.
Because the analysis reported in Table 6.3 examines two earlier time periods than
Table 6.2, we used a longer, eight-year “trailing window” (see Note 12; we include only
issued patents or license agreements within eight years following the invention disclo¬
sures) to compare the yield and productivity of Stanford and UC patenting and li¬
censing before and after Bayh-Dole.
16. Since the UC disclosure and licensing data largely exclude software inventions,
the most nearly comparable Stanford data are those excluding software inventions and
licenses.
17. The data on Stanford licenses are incomplete, as was noted earlier. Moreover,
 NOTES TO CHAPTER 6

the licensing income data are reported on a fiscal year basis and license agreements on 209
a calendar year basis, making it difficult to reconcile these data. We have attempted to
adjust the licensing income and agreements data to address this incompatibility, but
the data for Stanford in the last line of Table 6.3 should be treated with caution.
r8. This share is essentially stable, however, when software inventions are added to
the sample, reflecting the inclusion within the Stanford software licensing data of nu¬
merous academic “site licenses,” each of which produced a small but positive licens¬
ing income.
19. For example, average income per license may have increased in the second pe¬
riod, although the skewed distribution of the licensing income of both the Stanford
and UC technology transfer offices means that any such changes are likely to be small.
See Sampat and Ziedonis (2003) for an analysis of the relationship between the level of
licensing revenues at the University of California and Columbia and the characteris¬
tics of these universities’ licensed patents.
20. This third subset of our data is intended to separate those faculty disclosures that
occurred before Bayh-Dole could affect the content of academic research but whose
commercial promotion by universities (that is, through filing patent applications and
seeking licensees) occurred after the effective date of the Act. Unfortunately, the small
size of this sample of patents reduces the power and significance of our tests of statisti¬
cal significance for any differences among the characteristics of the three subsets of
patents.
21. Our control samples are constructed differently from those of Henderson, Jaffe,
and Trajtenberg (1995, 1998a). They used as their control sample a 1 percent random
sample of all U.S. patents granted during the time period covered by their university
patent sample. In contrast, our control sample matches each university patent with a
nonuniversity patent from the same patent class with an application date at or near the
date of application for the university patent.
22. Our definition of biomedical patent classes is based on Technology Assessment
and Forecast reports published by the U.S. Patent and Trademark Office that identify
three-digit patent classes and subclasses related to medical and biomedical technolo¬
gies. For a more detailed description of this taxonomy, see Ziedonis (2001).
23. In another analysis not reported because of space limitations, we subdivided the
post-1981 period into three subperiods of equal length in order to test for more signifi¬
cant reductions in the importance of Stanford and UC patents during the late 1980s.
Henderson, Jaffe, and Trajtenberg (1998a) find that the greatest reduction in impor¬
tance in their sample of academic patents, especially those assigned to the leading aca¬
demic patenters, occurs late in the 1980s. Our tests for differences in the means of the
university and control patent samples, however, reveal no significant differences.
24. Interestingly, the corrected results reported in the errata appendix to Trajten¬
berg, Henderson, and Jaffe (1997) find that the importance (measured in terms of the
number of forward citations) of academic patents, relative to those from corporate in¬
ventors, declines between 1975 and 1980, before the Bayh-Dole Act. 1 Ins finding is
broadly consistent with our conclusion that Bayh-Dole per se had little measurable ef¬
fect on the relative importance of academic and nonacademic patents. ;
" 25. Any such effect was significant during only the early years of Columbia s
NOTES TO CHAPTER 7

210 patenting and licensing activities, since by 1986-90 the share of disclosures resulting
in issued patents and the share of disclosures that result in licenses yielding positive roy¬
alty income are fairly similar at Columbia, the University of California, and Stanford,
as we noted earlier in this chapter.
26. We also examined the number of UC and Stanford patents from all three sub¬
periods that yielded no citations during the six years following their issue in an effort
to replicate the analysis in Henderson, Jaffe, and Trajtenberg (1998a) of “losers.” Here,
too, we find no consistent trend. “Zero-citation patents” account for 22 percent and
18 percent, respectively, of the UC and Stanford patents for period 1 in Tables 6.5 and
6.6; 16 percent and o percent of the UC and Stanford patents in period 2; and 11 per¬
cent and 8 percent of the UC and Stanford patents in period 3.
27. We also tested, but do not report results, for the significance of year-specific ef¬
fects on the relative importance of these academic and nonacademic patents, using a
negative binomial specification. We find no significant decline in the importance of
these three universities’ patents relative to nonacademic patents during the post-1980
period.
28. Stanford s middle period biomedical patents also display a lower mean gener¬
ality score than do their control patents, and UC nonbiomedical patents applied for
and issued before 1981 also exhibit slightly lower mean “generality scores” than do their
respective control samples.
29. Sampat, Mowery, and Ziedonis (2003) empirically analyze the effects of trun¬
cation bias on measures of change in university patent quality after Bayh-Dole.

chapter 7

1. As we pointed out in Note 21 in the previous chapter, the patent control samples
used in our analysis are constructed differently from those of Henderson, Jaffe, and
Trajtenberg (1995,1998a).
2. Our construction of the academic patent dataset yields a small number of obser¬
vations for 1992, and we accordingly omit this year from the results reported in
Tables 7.2-4.
3. Specifically, we estimate a “double” tobit model with an lower limit of zero and
an upper limit of 1.
4. We also ran probit regressions that examined the number of patents assigned to
universities in these three groups that received no citations in the first six years follow¬
ing their issue. Henderson, Jaffe, and I rajtenberg (1998a,b) use this “zero-citation” mea¬
sure as another measure of change in the characteristics of university patents. The re¬
sults of these probit regressions were broadly similar to those reported in Tables 7.2-4.
5. Our findings of differences in significance in importance and generality among
the “high-intensity” incumbent, “low-intensity” incumbent, and “entrant” subsamples
and their respective control groups could be affected by differences in standard errors
due to the various sizes of the three subsamples. To check the robustness of our results,
we conducted but do not report similar regressions combining all three subsamples
and obtained broadly similar results to those reported in Tables 7.2-4.
 NOTES TO CHAPTER 7

6. Our matching algorithm used to generate the control sample generated the same 211
control patent for 307 university patents. We thus have 10,574 different control patents
matched with our 10,881 university patents.
7. Lanjouw and Schankerman (1998) suggest that five-year citation windows are suf¬
ficient to construct meaningful measures of a patents “importance.”
8. We also conducted Poisson regressions but tested for and found evidence of
overdispersion, suggesting that the negative binomial model is more appropriate.
9. We also conducted but do not report probit regressions comparing the probabil¬
ity that an academic and nonacademic patent is ever cited within five years of issue us¬
ing the same university and control group samples. Point estimates of the coefficients
for ENT8l83 and ENT8486 in this analysis imply a marginal effect of about -0.05 (that is,
entrants’ patents applied for during 1981-83 and 1984-86 were approximately 5 per¬
cent less likely to be cited at all than were incumbents’ patents).
10. Unreported probit regression results for biomedical patents indicate that en¬
trants’ patents applied for during 1984- 86 were significantly less likely (at the 5 percent
level) to be cited than were incumbents’ patents. This difference was not statistically
significant, however, for entrant-university and incumbent-university patents applied
for during 1981-83.
11. In unreported probit regressions for nonbiomedical patents, entrants’ patents ap¬
plied for during 1981—83 are significantly less likely to be cited at all than are the simi¬
lar patents of incumbent universities (at the 5 percent level), but this differential be¬
comes smaller and statistically insignificant for patents applied for during 1984-86.
12. Improvement in entrants’ patent importance could reflect changes in funding
patterns. For example, if the average importance of patents in an institution’s patent
portfolio is correlated with the size of its research budget, increases in their research
budgets could lead to an improvement in the importance of entrant universities’ pat¬
ents in the absence of “learning.” There is little empirical support for this hypothesis,
and we do not test it here. Nevertheless, Geiger and Feller (1995) found evidence of in¬
creased interinstitutional dispersion in federal funding of university research during
the 1980s, and higher dispersion may well have benefited universities with less pre-1980
experience in patenting. Geiger and Feller (1995) found that growth during the 1980s
in funding dispersion was greatest in biomedicine, a field in which we find little evi¬
dence of improvement in the importance of entrants’ patents. Although it merits addi¬
tional analysis, this evidence provides little support for a “dispersion hypothesis in ex¬
plaining the increased measured importance of entrant universities’ patents during the

late 1980s. ,
13 In separate regressions (not reported), we analyze the average importance ot an
institution’s patents at time t as a function of cumulative patenting experience for the
entire 1975-92 period. These regressions did not reveal a relationship between cumu¬
lative patenting and patent importance.
14 Michael Crow, personal communication, September 28, 2000.
15. We also conducted but do not report the results of regression specifications us¬
ing interaction terms for individual application years and patent technology classes.
These regressions produced results similar to those from regressions where application
NOTES TO CHAPTER 8

212 years were aggregated into three-year intervals. The first application year dummy and
first patent class dummy were omitted to avoid perfect multicollinearity.

chapter 8

1. Similar results are reported in Thursby and Thursby (2002), based on a survey of
112 firms that licensed university inventions between 1993 and 1997.
2. Two important publications describing the process were Wigler et al. (1977) and
Wigler et al. (1979).
3. Since Columbia did not have a technology transfer office in early 1980, the
General Counsel’s office handled all potentially patentable inventions disclosed to the
university.
4. Columbia’s low royalty rates for a key “research process” patent resemble the li¬
censing policies followed by Stanford in its licensing of the Cohen-Boyer patents de¬
veloped by researchers at the University of California and Stanford University.
5. Niels Reimers, the first head of Stanford’s Office of Technology Transfer and
manager ofthe licensure of Cohen-Boyer, subsequently noted, “[W]hether we licensed
it or not, commercialization of recombinant DNA was going forward. As I mentioned,
a nonexclusive licensing program, at its heart, is really a tax . . . [b]ut it’s always nice to
say ‘technology transfer’” (Reimers, 1998).
6. Materials similar to GaN that are based on aluminum and indium have similar
properties and are now used in many of the same applications listed above as a result
of the research effort discussed below. Although we focus on GaN, research in this area
included research on a class of related semiconductors that included these cousins,
namely Class III-V.
7. Blue GaN LEDs had been produced previously —Pankove himself had devel¬
oped a prototype earlier. But Nakamura’s prototype is widely credited as the first GaN
LED with significant commercial potential.
8. “GaN-related patents” were selected through a text-based search for “gallium ni¬
tride or GaN in the title or abstract of all U.S. patents issuing during 1972-98.
9. Akasakis patents issued in 1983 and 1984 are largely based on his work at Mat¬
sushita, since the patents were applied for during the same year Akasaki accepted his
university position.
10. The other university patent issuing during 1983-94 was assigned to Harvard
University and covered a thin film process that could be used for GaN as well as other
Class III-V compounds.
11. Another significant instance of industrial knowledge flowing into academia oc¬
curred in 2000, when Nakamura accepted a position at UC Santa Barbara.
12. I11 2002, responsibility for the distribution of the Ames Tests assays was shifted to
Xenometrix, the start-up firm that commercialized the Ames II Tests. Xenometrix has
maintained the original UC distribution policies for these Ames Tests, policies that in
many respects resemble the Materials lransfer Agreements widely used by academic
and industrial research laboratories.
13. http://ist-socrates.berkeley.edu/mutagen/salmut_facil_core.html.
14. Gene profiling is the process of collecting information on gene activity after ex-
 NOTES TO CHAPTER 9

posure to another organism, a chemical, or an environmental change. 213

15. The Wall Street Journal announced after the publication of the paper by Ortho
scientists that “a sixth scientific group has entered the already crowded field vying to
develop [CD4]” (Chase, 1988). As was the case in the GaN study, the large number of
published scientific papers from industrial scientists in this field illustrates the impor¬
tance of knowledge flows from industrial research to academic institutions.
r6. Referring to Upjohn’s licensure from the NIH of the competing CD4 treatment
(CD4-PE), the Antiviral Agents Bulletin observed that “it is interesting to note that
Upjohn has licensed this patent application for CD4, although a patent claiming the
entire CD4 gene and active portions has been issued to Columbia . . . [i]t appears that
Columbia could claim infringement or otherwise try to induce Upjohn to take a li¬
cense on its CD4 patent, should its CD4-PE show sufficient clinical and commercial
potential” (“CD4-PE40 Exotoxin Conjugate in Trials; Other CD4 Conjugates in De¬
velopment,” 1992, p. 69).
17. A r995 article observed that “although other, much larger companies have pur¬
sued development of CD4-based therapeutics, none appear to have active develop¬
ment efforts in this area, and it is somewhat surprising that so few organizations are pur¬
suing development of CD4-based therapeutics ’ (‘ Progenies Developing CD4-IgG2 for
HIV-Infection,” 1995, p. 166).
18. Interest within the research and industrial communities in soluble CD4-based
treatments remained high, because antiretroviral drugs have side effects that many pa¬
tients cannot tolerate, and a growing share of HIV patients have developed resistance
to such drugs.
19. Two of the licensees sold their soluble CD4 preparations to other researchers,
accounting for the sales-based royalties. In addition to these sales-based royalties, Co¬
lumbia collected at least $2 million in up-front payments, milestone payments, and le¬
gal reimbursements from the licensees.
20. Nonetheless, interviews with licensing officers suggest that very few university
inventions face such strong demand from prospective licensees that the officer can se¬
lect from among several “applicants” for a license in a given field of use.
21. The substantial flow of scientific papers from industrial scientists in AIDS re¬
search that was noted earlier also supports this characterization of the fiinrs engaged in
commercial development of the soluble CD4 invention.

chapter 9

1. Indeed, Howard Bremer, former patent counsel for the Wisconsin Alumni Re¬
search Foundation (WARF) and president of the Association of University Technology
Managers (AUTM) during 1978-79, acknowledged in a statement celebrating the
“technology transfer harvest” following Bayh-Dole that the growth of patenting of fed¬
erally funded academic research began not with Bayh-Dole but with the NSF and
HEW Institutional Patent Agreements (IPAs) in the 1960s and 1970s, and he notes that
“in fact, that law is often looked upon as a codification of the terms and provisions of
the IPAs” (Bremer, 2001).
2. In June 1998, a report of the NIH Working Group on Research Tools raised con-
NOTES TO CHAPTER 9

214 cerns about difficulties scientists and institutions faced in getting access to patented
biomedical research tools developed with federal funds. In response to these concerns,
in the “Technology Transfer Commercialization Act of 2000,” Congress added a “tech¬
nical amendment” to Bayh-Dole, noting that federally funded inventions should be li¬
censed “without unduly encumbering future research and development” (Roumel,
2003). The effects of this change in the language of the Bayh-Dole Act on academic
patenting and licensing practices remain unclear.
3. Walsh, Arora, and Cohen (2003) argue that the reliance by academic researchers
and research administrators on a de facto research exemption from infringement suits
has been an important “working solution” to problems arising from the potential as¬
sertion by patentholders of rights over research materials, tools, and similar inputs for
the scientific research process. The authors conclude that the Madey v. Duke decision,
if not reversed judicially or legislatively, “could well chill some of the ‘offending’ bio¬
medical research that is conducted in university settings” (p. 56).
 References

Agrawal, A., and R. Henderson. 2002. “Putting Patents in Context: Exploring Knowl¬
edge Transfer from MIT.” Management Science 48: 44-60.
American Association for the Advancement of Science (AAAS). 1934. The Protection by
Patents of Scientific Discoveries: Report of the Committee on Patents, Copyrights, and
Trademarks. New York: Science Press.
Apple, R. D. 1989. “Patenting University Research.” Isis 80: 375-94-
_1996. Vitamania: Vitamins in American Culture. New Brunswick, N.J.: Rutgers
University Press.
Arrow, K. 1962. “Economic Welfare and the Allocation of Resources for Invention.” In
R. R. Nelson, ed., The Rate and Direction of Inventive Activity. Princeton, N.J.:
Princeton University Press.
Association of University Technology Managers (AUTM). 1994. The AUTM Licensing
Survey: Executive Summary and Selected Data, Fiscal Years 1993, 1992, and 1991.
Norwalk, Conn.: AUTM.
_ 1996. AUTM Licensing Survey 1996, Survey Summary. Norwalk, Conn.: AUTM.
_ j998. AUTM Licensing Survey 1998, Survey Summary. Norwalk, Conn.:

AUTM.
_2000. The AUTM Licensing Survey: FY1999. Norwalk, Conn.: AU TM.
“Axel Patent Claims Mammalian Cell Transfer.” McGraw Hills Biotechnology
Newswatch 3 (1983): 5.
Barrett, P. 1980. “Harvard Fears Congress May Not Pass Patent Bill. Harvard Crimson,
October 7.
Ben-David, J., 1968. Fundamental Research and the Universities. Paris: OECD.
Blum, F. 2000. Telephone interview.
Blumenthal, D., E. Campbell, M. Anderson, N. Causino, and K. Louis. 1997. With¬
holding Research Results in Academic Life Science: Evidence from a National Sur¬
vey of Faculty.” Journal of the American Medical Association 277: 1224-29.
REFERENCES

216 Blumenthal, D., S. Epstein, and J. Maxwell. 1986. “Commercializing University Re¬
search: Lessons from the History of the Wisconsin Alumni Research Foundation.”
New England Journal of Medicine 314:1621-26.
Bok, D. Beyond the Ivory Tower. 1982. Cambridge: Harvard University Press.
Bremer, H. 2001. “The First Two Decades of the Bayh-Dole Act as Public Policy.” Pre¬
sentation to the National Association of State Universities and Land Grant Col¬
leges. Available from World Wide Web: http://www.nasulgc.org/COTT/Bayh
-Dohl/Bremer_speech.htm.
Broad, W. 1979a. “Patent Bill Returns Bright Idea to Inventor.” Science 205: 473-76.
-1979b. “Whistle Blower Reinstated at HEW.” Science 205: 476.
Bryson, E. 1984. “Frederick E. Terman: Educator and Mentor.” IEEE Spectrum
(March): 71—73.
Burn, B. B., P. G. Altbach, C. Kerr, and J. A. Perkins. 1971. Higher Education in Nine
Countries. New York: McGraw-Hill.
Bush, V. 1943. “The Kilgore Bill.” Science 98: 571-77.
- 1945- Science: The Endless Frontier. Washington, D.C.: U.S. Government Print¬
ing Office.
Cameron, F. 1993. Cottrell: Samaritan of Science. Tucson, Az.: Research Corporation.
Campbell, E. G., B. R. Clarridge, M. Gokhale, L. Birenbaum, S. Hilgartner, N. A.
Holtzman, and D. Blumenthal. 2002. “Data Withholding in Academic Genetics:
Evidence from a National Survey.” Journal of the American Medical Association
287: 473-80.
Camras, C. B., and L. Z. Bito. 1981. “Reduction of Intraocular Pressure in Normal and
Glaucomatous Primate (Aotus Trivirgatus) Eyes by Topically Applied Prostaglandin
F2 Alpha.” Current Eye Research 1: 205-9.
Camras, C. B., L. Z. Bito, and K. E. Eakins. 1977. “Reduction of Intraocular Pressure
by Prostaglandins Applied Topically to the Eyes Of Conscious Rabbits.” Investiga¬
tive Ophthalmology and Visual Science 16:1125-34.
Carnegie Commission on Higher Education. 1973. ^ Classification of Institutions of
Higher Education: A Technical Report. Berkeley, Calif.: Carnegie Commission on
Higher Education.
- !993- A Classification of Institutions of Higher Education; A Technical Report.
Berkeley, Calif.: Carnegie Commission on Higher Education.
Caves, R., H. Crookell, and P. Killing. 1983. “The Imperfect Market for Technology
Licenses.” Oxford Bulletin of Economics and Statistics 45: 249-67.
“CD4-PE40 Exotoxin Conjugate in Trials; Other CD4 Conjugates in Development.”
Antiviral Agents Bulletin (March 1992): 69.
Chase, M. 1988. “Group Enters Field Studying CD4 for AIDS.” Wall Street Journal,
June 3.
Chiang, H. 1997. Court Reinstates Royalty Award to UC Professors.” San Francisco
Chronicle, December 2.
Cohen, W. M., R. Morida, and R. Goe. 1994. “University-Industry Research Centers in
the United States. Pittsburgh: Center for Economic Development, Carnegie-
Mellon University.
Cohen, W. M., R. Florida, L. Randazzese, and J. Walsh. 1998. “Industry and the Acad-
 REFERENCES

emy: Uneasy Partners in the Cause of Technological Advance.” In R. Noll, ed., 217
Challenges to the Research University. Washington, D.C.: Brookings Institution.
Cohen, W. M., and D. A. Levinthal. 1990. “Absorptive Capacity: A New Perspective on
Learning and Innovation.” Administrative Science Quarterly 35: 128-52.
Cohen, W. M., R. R. Nelson, and J. P. Walsh. 2002. “Links and Impacts: The Influence
of Public Research on Industrial R&D.” Management Science 48: 1—23.
“Columbia Innovation Enterprise: A Special Report.” Reporter 7 (1996).
“Columbia University rDNA Patent Licensing.” Blue Sheet 27 (1984): 3-4.
Colyvas, J., M. Crow, A. Gelijns, R. Mazzoleni, R. R. Nelson, N. Rosenberg, and B. N.
Sampat. 2002. “How Do University Inventions Get into Practice?” Management
Science 48: 61—72.
Committee on Uniform Patent Policies in Land Grant and Engineering Experiment
Stations. 1922. “Report of the Committee on Uniform Patent Policies.” Proceedings
of the Association of Land Grant Colleges and Universities: 283-84.
Cottrell, F. 1911. “The Electrical Precipitation of Suspended Particles.” Journal of In¬
dustrial and Engineering Chemistry 3: 542-50.
-1912. “The Research Corporation, an Experiment in the Public Administration
of Patent Rights.” Journal of Industrial and Engineering Chemistry 4: 864-67.
- 1932. “Patent Experience of the Research Corporation.” Transactions of the
American Institute of Chemical Engineers 26: 222-25.
- 1937- “The Social Responsibility of the Engineer.” Journal of the Western Soci¬
ety of Engineers 42: 65-77.
Council on Government Relations. 1999. The Bayh-Dole Act: A Guide to the Law and
Implementing Regulations. Available from World Wide Web: http://www.cogr.edu/
bayh-dole.htm.
Crow, M. M., A. C. Gelijns, R. R. Nelson, H. J. Raider, and B. N. Sampat. 1998. “Re¬
cent Changes in University-Industry Research Interactions: A Preliminary Analysis
of Causes and Effects.” Unpublished working paper, School of International and
Public Affairs, Columbia University, New York.
Dasgupta, P., and P. David. 1994. “Towards a New Economics of Science.” Research
Policy 23: 487-521.
David, P. A., and D. Foray. 1995. “Accessing and Expanding the Science and Technol¬
ogy Knowledge Base.” STI Review 16:13-68.
David, P. A., D. C. Mowery, and W. E. Steinmueller. 1992. “Analyzing the Economic
Payoffs from Basic Research.” Economics of Innovation and New Technology 2:
73-90.
Davis, E. W. 1964. Pioneering with Taconite. St. Paul: Minnesota Historical Society.
DenBaars, S. 2000. Personal interview.
Department of Health and Human Services. 2001. NIH Response to the Conference Re¬
port Request for a Plan to Ensure Taxpayers’ Interests Are Protected. Bethesda, Md.:
National Institutes of Health. Available from World Wide Web: http://www.nih
.gov/news/oyoioiwyden.htm.
Dickinson, Q. Todd. 2000. “Reconciling Research and the Patent System.” Issues in
Science and Technology 16: 27-31.
“Dr. Cottrell and the Research Corporation.” Scientific Monthlv 22 (1926): 181-86.
REFERENCES

2i8 Dreyfuss, R. 2003. “Varying the Course in Patenting Genetic Material: A Counter-
Proposal to Richard Epstein’s Steady Course. " New York University School of Law,
Public Law and Legal Theory Research Paper Series, Research Paper No. 59.
Eisenberg, R. S. 1996. “Public Research and Private Development: Patents and Tech¬
nology Transfer in Government-Sponsored Research.” Virginia Law Review 82:
1663-1727.
-2001. “Bargaining over the Transfer of Proprietary Research Tools: Is This Mar¬
ket Emerging or Lading?” In D. L. Zimmerman, R. C. Dreyfuss, and H. First, eds.,
Expanding the Bounds of Intellectual Property: Innovation Policy for the Knowledge
Society. New York: Oxford University Press.
Eskridge, N. 1978. “Dole Blasts HEW for ‘Stonewalling’ Patent Applications.” Bio¬
science 28: 605-6.
Etzkowitz, H. 1994. “Knowledge as Property: The Massachusetts Institute of Technol¬
ogy and the Debate of Academic Patent Policy.” Minerva 32: 383-421.
Evenson, R. E. 1982. “Agriculture.” In R. R. Nelson, ed., Government and Technical
Progress: A Cross-Industry Analysis. New York: Pergamon.
Executive Office of the President. 1983. “Memorandum to the Heads of Executive De¬
partments and Agencies: Government Patent Policy, February 18,1983.”
Federal Council on Science and Technology (FCST). 1978. Report on Government
Patent Policy, 1973-1976. Washington, D.C.: U.S. Government Printing Office.
Feldman, M., I. Feller, J. E. L. Bercovitz, and R. Burton. 2002. “Equity and the Tech¬
nology Transfer Strategies of American Research Universities.” Management
Science 48:105—21.
Fishman, E. A. 1996. “MIT Patent Policy 1932-1946: Historical Precedents in
University-Industry Technology Transfer.” Ph.D. diss., University of Pennsylvania.
Flexner, A. 1968 [1930]. Universities: American, English, German. London: Oxford Uni¬
versity Press.
Foray, D., and A. Kazancigil. 1999. “Science, Economics and Democracy.” Man¬
agement of Social Transformation (MOST) Discussion Paper 42. Prepared for
UNESCO World Conference on Science 1999.
Foreman, ]. 1988. “Scientists Race to Create AIDS Virus ‘Decoy’: Crucial Issues Re¬
main, but Interest Is Intense in Using Fake Receptors.” Boston Globe, January 11.
Fox, J. 1983. “Columbia Awarded Biotechnology Patent.” Science 221: 933.
Furter, W. F., ed. 1980. History of Chemical Engineering. Washington, D.C.: American
Chemical Society.
Gee, P. 2001. Personal interview.
Gee, P., D. M. Maron, and B. N. Ames. 1994. “Detection and Classification of Muta¬
gens: A Set of Base-Specific Salmonella Tester Strains.” Proceedings of the National
Academy of Sciences 91: 11606-10.
Geiger, R. 1986. To Advance Knowledge: The Growth of American Research Universities,
1900-1940. New York: Oxford University Press.
- 1993- Research and Relevant Knowledge: American Research Universities Since
World War II. New York: Oxford University Press.
Geiger, R., and I. Feller. 1995. “The Dispersion of Academic Research in the 1980s.”
Journal of Higher Education 65: 336-60.
Gelijns, A., and N. Rosenberg. 1999. “Diagnostic Devices: An Analysis of Comparative
 REFERENCES

Advantages.’ In D. C. Mowery and R. R. Nelson, eds., Sources of Industrial Lead¬ 219

ership. New York: Cambridge University Press.
Gerth, J., and S. Stolberg. 2000. “Drug Companies Profit from Research Funded by
Taxpayers.” New York Times, April 23.
Gilles, J. r99r. “Research Corporation Technologies Offers Universities ‘Critical Mass’
of Diverse Expertise.” Technology Access Report, August: 1+.
“Glennan Asks Review of NASA Patent Policy.” 1959. Aviation Week, March 30, 33.
Government University Industry Research Roundtable (GUIRR). 1991. Industrial Per¬
spectives on Innovation and Interactions with Universities. Washington, D.C.: Na¬
tional Academy Press.
Graham, H. D., and N. Diamond. 1997. The Rise of American Research Universities.
Baltimore: Johns Hopkins University Press.
Gray, G. 1936. “Science and Profits.” Harpers 172: 539-49.
Grayson, L. 1977. “A Brief History of Engineering Education in the United States.” En¬
gineering Education (December): 246-64.
Gregg, A. 1933. “University Patents.” Science 77: 257-60.
Hall, B. H. 2000. “A Note on the Bias in the Herfindahl Based on Count Data.” In
A. B. Jaffe and M. Trajtenberg, eds., Patents, Citations, and Innovations: A Window
on the Knowledge Economy. Cambridge: MIT Press.
Hall, B. H., A. B. Jaffe, and M. Tratjenberg. 2oor. “The NBER Patent Citation Data
File: Lessons, Insights and Methodological Tools.” National Bureau of Economic
Research Working Paper 8498.
Harbridge House Inc. 1968a. “Effects of Patent Policy on Government R&D Pro¬
grams.” Government Patent Policy Study, Final Report. Vol. 2. Washington, D.C.:
Federal Council for Science and Technology.
-r968b. “Effects of Government Policy on Commercial Utilization and Business
Competition.” Government Patent Policy Study, Final Report. Vol. 4. Washington,
D.C.: Federal Council for Science and Technology.
Hasselmo, N. 1999. Priorities for Federal Innovation Reform. Washington, D.C.: Asso¬
ciation of American Universities.
Heaton, George R., Jr., Christopher T. Hill, and Patrick Windham. 2000. “Policy In¬
novation: The Initiation and Formation of New Science and Technology Policies
in the U.S. During the 1980s.” A Report to Japan External Trade Organization New
York office and Japan New Energy and Industrial Technology Development Orga¬
nization, Washington, D.C.
Heilbron, J., and R. Seidel. 1989. Lawrence and His Laboratory: A History of the Law¬
rence Berkeley Laboratory. Berkeley: University of California Press.
Heller, M. A., and R. S. Eisenberg. 1998. “Can Patents Deter Innovation? The Anti¬
commons in Biomedical Research.” Science 280: 298.
Henderson, R., A. B. Jaffe, and M. Trajtenberg. 1994. “Numbers Up, Quality Down?
Trends in University Patenting, 1965-1992.” Presented at the CEPR Conference on
University Goals, Institutional Mechanisms, and the “Industrial Transferability” of
Research, Stanford University, Stanford, California.
_1995. “Universities as a Source of Commercial Technology: A Detailed Analysis
of University Patenting, 1965-1988.” National Bureau of Economic Research Work¬
ing Paper 5068.
REFERENCES

220 -1998a. “Universities as a Source of Commercial Technology: A Detailed Analy¬

sis of University Patenting, 1965—88.” Review of Economics and Statistics 80:
n9~27-
- 1998b. “University Patenting Amid Changing Incentives for Commercializa¬
tion.” In G. Barba Navaretti, P. Dasgupta, K. G. Maler, and D. Siniscalco, eds., Cre¬
ation and Transfer of Knowledge. New York: Springer.
Henderson, R., L. Orsenigo, and G. Pisano. 1999. “The Pharmaceutical Industry and
the Revolution in Molecular Biology: Interactions Among Scientific, Institutional
and Organizational Change.” In D. C. Mowery and R. R. Nelson, eds., Sources of
Industrial Leadership. New York: Cambridge University Press.
Henderson, Y. 1933. “Patents Are Ethical.” Science 77: 324-25.
Henig, R. 1979. “New Patent Policy Bill Gathers Congressional Support.” Bioscience
29: 281- 84.
Hoskins, W., and R. Wiles. 1921. “Promotion of Scientific Research." Chemical and
Metallurgical Engineering 24: 689-91.
Hounshell, D. A., and J. K. Smith. 1988. Science and Corporate Strategy. New York:
Cambridge University Press.
“Innovations Golden Goose.” 2002. Economist 365: T3.
Jaffe, A. B., M. S. Fogarty, and B. A. Banks. 1998. “Evidence from Patents and Patent
Citations on the Impact of NASA and Other Federal Labs on Commercial Innova¬
tion.” Journal of Industrial Economics 46:183-205.
Jensen, R., and M. Thursby. 2001. “Proofs and Prototypes for Sale: The Licensing of
University Inventions.” American Economic Review 91: 240-58.
r Kahaner, D. 1995. “Blue LEDs: Breakthroughs and Implications.” Available from
World Wide Web: http://www.atip.or.ip/public/atp.reports.95/atip95.59r.html.
Katz, M. L., and J. A. Ordover. 1990. “R&D Competition and Cooperation.” Brookings
Papers on Economic Activity: Microeconomics: 137-92.
Kevles, D. J. 1977. “The National Science Foundation and the Debate over Postwar Re¬
search Policy, 1942-45.” Isis 68: 5-26.
-i978. The Physicists. New York: Norton.
-1990. Principles and Politics in Federal R(PD Policy, 1945-1990: An Appreciation
of the Bush Report. Washington, D.C.: National Science Foundation.
Kitch, E. W. 1977. “The Nature and Function of the Patent System.” Journal of Law
and Economics 20: 265-90.
Kortum, S., and J. Lerner. 1999. “What Is Behind the Recent Surge in Patenting?” Re¬
search Policy 28: 1-22.
Lamoreaux, N., and K. Sokoloff. 2002. “Intermediaries in the Market for Technology
in the United States, 1870-1920.” National Bureau of Economic Research Working
Paper 9017.
Lanjouw, J. O., and M. Schankerman. 1998. “The Quality of Ideas: Measuring Inno¬
vation with Multiple Indicators.” National Bureau of Economic Research Working
Paper 7375.
Levin, R. C., A. Klevorick, R. R. Nelson, and S. Winter. 1987. “Appropriating the Re¬
turns from Industrial Research and Development.” Brookings Papers on Economic
Activity 3: 783-820.
 REFERENCES

Levine, D. O. 1986. The American College and the Culture of Aspiration, 1915-1940. 221
Ithaca, N.Y.: Cornell University Press.
“Licensing Plan For AIDS Drug Draws Fire.” 1988. Seattle Times, October 25.
Liebeskind,}. 2001. “Risky Business: Universities and Intellectual Property.” Academe
87. Available from World Wide Web: http://www.aaup.org/publications/Academe/
oiSO/sooilie.htm.
Little, A. D. 1933. Twenty-five Years of Chemical Engineering Progress. New York: Van
Nostrand.
Louis, K., L. Jones, M. Anderson, D. Blumenthal, and E. Campbell. 2001. “Entrepre¬
neurship, Secrecy, and Productivity: A Comparison of Clinical and Non-Clinical
Life Sciences Faculty.” Journal of Technology Transfer 26: 233-45.
Lowe, Robert A. 2001. “The Role and Experience of Start-ups in Commercializing
University Inventions: Start-up Licensees at the University of California.” In
G. Libecap, ed., Entrepreneurial Inputs and Outcomes. Amsterdam: JAI Press.
- 2002. “Entrepreneurship and Information Asymmetry: Theory and Evidence
from the University of California.” Unpublished working paper, Haas School of
Business, University of California-Berkeley.
Maddison, A. 1987. “Growth and Slowdown in Advanced Capitalist Economies: Tech¬
niques of Quantitative Assessment.” Journal of Economic Literature 25: 649-98.
Maddon, P. J., D. R. Littman, M. Godfrey, D. E. Maddon, L. Chess, and R. Axel. 1985.
“The Isolation and Nucleotide Sequence of a cDNA Encoding the T Cell Surface
Protein T4: A New Member of the Immunoglobulin Gene Family.” Cell 42: 93-104.
Mansfield, E. r99r. “Academic Research and Industrial Innovations.” Research Policy
20: 1-12.
Marcy, W. 1978. “Patent Policy at Educational and Nonprofit Scientific Institutions.”
In W. Marcy, ed., Patent Policy: Government, Academic, and Industry Concepts.
Washington, D.C.: American Chemical Society.
Marx, J. r98o. “Gene Transfer Moves Ahead.” Science 210:1334-36.
Mazzoleni, R., and R. Nelson. 1998. “The Benefits and Costs of Strong Patent Protec¬
tion: A Contribution to the Current Debate.” Research Policy 27: 274-84.
McKusick, V. 1948. “A Study of Patent Policies in Educational Institutions, Giving Spe¬
cific Attention to the Massachusetts Institute of Technology.” Journal of the Franklin
Institute 245: 193-225.
Merges, R., and R. Nelson. 1994. “On Limiting or Encouraging Rivalry in Technical
Progress: The Effect of Patent Scope Decisions.” Journal of Economic Behavior and
Organization 25: 1-24.
Merton, R. K. 1973. The Sociology of Science : Theoretical and Empirical Investigations.
Chicago: University of Chicago Press.
Mishra, U. 2001. Personal interview.
Mowery, D. C. 1981. “The Emergence and Growth of Industrial Research in American
Manufacturing, 1899-1945.” Ph.D. diss., Stanford University.
_1983. “The Relationship Between Intrafirm and Contractual Forms of Industrial
Research in American Manufacturing, 1900-1940.” Explorations in Economic His¬
tory 20: 351-74-
_ 1999. The Evolving Structure of University-Industry Collaboration in the United
REFERENCES

States: Three Cases, in Research Teams and Partnerships: Trends in the Chemical Sci¬
ences. Washington, D.C.: National Academy Press.
-2002. “The Changing Role of Universities in the 21st Century U.S. R&D Sys¬
tem.” In A. H. Teich, S. D. Nelson, S. J. Lita, eds., AAAS Science and Technology
Policy Handbook. Washington, D.C.: American Association for the Advancement of
Science.
Mowery, D. C., R. R. Nelson, B. N. Sampat, and A. A. Ziedonis. 1999. “The Effects of
the Bayh-Dole Act on U.S. University Research and Technology Transfer: An Analy¬
sis of Data from Columbia University, the University of California, and Stanford
University.” In L. Branscomb and R. Florida, eds., Industrializing Knowledge. Cam¬
bridge: MIT Press.
- 2001. “The Growth of Patenting and Licensing by U.S. Universities: An
Assessment of the Effects of the Bayh-Dole Act of 1980.” Research Policy 30:
"-ng.
Mowery, D. C., and N. Rosenberg. 1993. “The U.S. National Innovation System.” In
R. R. Nelson, ed., National Innovation Systems: A Comparative Analysis. New York:
Oxford University Press.
-1998. Paths of Innovation: Technological Change in 20th-Century America. New
York: Cambridge University Press.
Mowery, D. C., and B. N. Sampat. 2001a. “Patenting and Licensing University Inven¬
tions: Lessons from the History of the Research Corporation.” Industrial and Cor¬
porate Change 10: 317-55.
- 2001b. “University Patents, Patent Policies, and Patent Policy Debates, 1925-
1980.” Industrial and Corporate Change 10: 781-814.
Mowery, D. C., B. N. Sampat, and A. A. Ziedonis. 2002. “Learning to Patent: Institu¬
tional Experience and the Quality of University Patents.” Management Science 48:
73-89.
Mowery, D. C., and T. Simcoe. 2002. “The Origins and Evolution of the Internet.” In
D. Victor, B. Steil, and R. R. Nelson, eds., Technological Innovation and National
Economic Performance. Princeton, N.J.: Princeton University Press.
Mowery, D. C., and A. A. Ziedonis. 1998. “Market Failure or Market Magic? Structural
Change in the U.S. National Innovation System.” STI Review 22: 101-36.
- 2001. “The Geographic Reach of Market and Nonmarket Channels of Tech¬
nology Transfer: Comparing Citations and Licenses of University Patents.” Na¬
tional Bureau of Economic Research Working Paper 8568.
Mowery, D. C., and R. R. Nelson, eds. 1999. Sources Of Industrial Leadership: Studies
of Seven Industries. New York: Cambridge University Press.
National Research Council. 1982. “Research in Europe and the United States.” In
Outlook for Science and Technology: The Next Five Years. San Francisco: W. H.
Freeman.
- 1997. Intellectual Property Rights and Research Tools in Molecular Biology.
Washington, D.C.: National Academy Press.
National Resources Committee. 1938. Research—A National Resource. Vol. 1. Wash¬
ington, D.C.: U.S. Government Printing Office.
National Science Board. 1996. Science and Engineering Indicators: 1996. Washington,
D.C.: U.S. Government Printing Office.
 REFERENCES

-2000. Science and Engineering Indicators: 2000. Washington, D.C.: U.S. Gov¬ 223
ernment Printing Office.
- 2002. Science and Engineering Indicators: 2002. Washington, D.C.: U.S. Gov¬
ernment Printing Office.
National Science Foundation. 1958. Funds for Scientific Activities in the Federal Gov¬
ernment: Fiscal Years 1953 and 1954. Washington, D.C.: U.S. Government Printing
Office.
-1980. National Patterns of Science and Technology Resources: 1980. Washington,
D.C.: U.S. Government Printing Office.
- 1994- National Patterns ofR&D Resources: 1994. Washington, D.C.: U.S. Gov¬
ernment Printing Office.
-1996. National Patterns ofR&D Resources: 1996. Washington, D.C.: U.S. Gov¬
ernment Printing Office.
- 1998. National Patterns ofR&D Resources: 1998. Washington, D.C.: U.S. Gov¬
ernment Printing Office.
-2000. National Patterns ofR&D Resources 2000. Washington, D.C.: U.S. Gov¬
ernment Printing Office.
-- 2001. National Patterns ofR&D Resources 2001. Washington, D.C.: U.S. Gov¬
ernment Printing Office.
Nelson, R. R. 1959. “The Simple Economics of Basic Scientific Research.” Journal of
Political Economy 67: 297-306.
- 1984. High-Technology Policies: A Five-Nation Comparison. Washington, D.C.:
American Enterprise Institute.
-1992. “What Is ‘Commercial’ and What Is ‘Public’ About Technology, and What
Should Be?” In N. Rosenberg, R. Landau, and D. C. Mowery, eds., Technology and
the Wealth of Nations. Stanford, Calif.: Stanford University Press.
-, ed. 1982. Government and Technical Progress: A Cross-Industry Analysis. New
York: Pergamon.
- 093 • National Innovation Systems: A Comparative Analysis. New York: Oxford
University Press.
Nelson, R. R., and P. Romer. 1997. “Science, Economic Growth, and Public Policy.”
In B. L. R. Smith and C. Barfield, eds., Technology, R&D, and the Economy. Wash¬
ington, D.C.: Brookings Institution.
Noble, D. F. 1977. American by Design. New York: Knopf.
Office of Technology Transfer. 1997. Annual Report: University of California Technol-
ogy Transfer Program. Oakland: University of California Office of the President.
Okimoto, D., and G. Saxonhouse. 1987. “Technology and the Future of the Economy.”
In K. Yamamura and Y. Yasuba, eds., The Political Economy of Japan. Vol. 1, The
Domestic Transformation. Stanford, Calif.: Stanford University Press.
Organization for Economic Cooperation and Development (OECD). 2000. A New
Economy? Paris: OECD.
-2002. Benchmarking Science-Industry Relationships. Paris: OECD.
Palmer, A. M. 1934. “University Patent Policies and Procedures.” Journal of the Patent
Office Society r6: 96-13^
_ J947- “Patents and University Research.” Law and Contemporary Problems 15:
680-94.
REFERENCES

-1948. Survey of University Patent Policies. Washington, D.C.: National Research

Council.
-1949. University Research and Patent Problems. Washington, D.C.: National Re¬
search Council.
—— 1952. University Patent Policies and Practices. Washington, D.C.: National Acad¬
emy of Sciences-National Research Council.
- 1955. Nonprofit Research and Patent Management Organization. Washington,
D.C.: National Academy of Sciences-National Research Council.
-1957. Patents and Nonprofit Research: A Study Prepared for the Subcommittee on
Patents, Trademarks, and Copyrights of the U.S. Senate Committee on the Judiciary.
Washington D.C.: U.S. Government Printing Office.
- 1962. University Research and Patent Policies, Practices, and Procedures. Vol. 1.
Washington, D.C.: National Academy of Sciences-National Research Council.
Perlman, D. 1987. “Biotech ‘Decoy’ May Fool AIDS Virus.” San Francisco Chronicle,
December 18.
Potter, A. 1940. “Research and Invention in Engineering Colleges.” Science
91: !-7-
“Progenies Developing CD4-IgG2 for HIV-Infection.” 1995. Antiviral Agents Bulletin
(June): 166.
Pugh, E. W. 1984. Memories That Shaped an Industry: Decisions Leading to IBM Sys-
tem/360. Cambridge: MIT Press.
Pursell, C. 1979. “Science Agencies in World War II: The OSRD and Its Challengers.”
In N. Reingold, ed., The Sciences in the American Context: New Perspectives. Wash¬
ington, D.C.: Smithsonian Institution Press.
Rai, A. T., and R. S. Eisenberg. 2001. “The Public and the Private in Biopharmaceutical
Research.” Presented at the Conference on the Public Domain, Duke University.
-2003. “Bayh-Dole Reform and the Progress of Biomedicine.” American Scientist
91: 52-59-
Reich, L. S. 1985. The Making of American Industrial Research. New York: Cambridge
University Press.
Reimers, Niels. 1998. “Stanfords Office of Technology Licensing and the Cohen/
Boyer Cloning Patents.” An oral history conducted in 1997 by Sally Smith Hughes,
Regional Oral History Office, The Bancroft Library, University of California,
Berkeley. Available from the Online Archive of California, http://ark.cdlib.org/
ark:/i303o/kt4b69n6sc.
Research Corporation. 1947-87. Annual Report. New York and Tucson, Az.: Research
Corporation.
- 1972. Science, Invention, and Society: The Story of a Unique American Institu¬
tion. New York: Research Corporation.
- 1979. “Report of the Committee on Goals and Objectives.” Unpublished MS,
Research Corporation Archives, Tucson, Az.
Research Corporation Technologies. 1996. Annual Report. Tucson, Az: Research Cor¬
poration Technologies.
Rosenberg, N. 1992. “Scientific Instrumentation and University Research.” Research
Policy 21: 381-90.
 REFERENCES

- 1998. “Technological Change in Chemicals: The Role of University-Industry 225

Relations.” In A. Arora, R. Landau and N. Rosenberg, eds., Chemicals and Long-
Term Economic Growth. New York: John Wiley.
Rosenberg, N., and R. R. Nelson. 1994. “American Universities and Technical Advance
in Industry.” Research Policy 23: 323-48.
Roumel, T. 2003. “Development of a Policy to Ensure the Sharing of Unique Bio¬
medical Research Resources in the Biomedical Community.” In OECD, Turning
L Science Into Business: Patenting and Licensing by Public Research Organizations.
Paris: OECD.
Roush, W., E. Marshall, and G. Vogel. 1997. “Publishing Sensitive Data: Who’s Call¬
ing the Shots?” Science 276: 523-26.
Sampat, B., D. C. Mowery, and A. Ziedonis. 2003. “Changes in University Patent
Quality After the Bayh-Dole Act: A Re-Examination.” International Journal of In¬
dustrial Organization 21: 1371-90.
Sampat, B. N., and R. R. Nelson. 2002. “The Emergence and Standardization of Uni-
versity Technology Transfer Offices: A Case Study of Institutional Change.” Ad¬
vances in Strategic Management 19.
Sampat, B. N., and A. A. Ziedonis. 2003. “Cite-Seeing: Patent Citations and Economic
Value.” Presented at the Conference on Empirical Economics of Innovation
and Patenting, Centre for European Economic Research, Mannheim Germany,
March.
Servos, J. W. 1980. “The Industrial Relations of Science: Chemical Engineering at
MIT, 1900-1939.” Isis 71: 531-49.
Sevringhaus, E. L. 1932. “Should Scientific Discoveries Be Patented?” Science 75:

233-34-
Shane, S. 2002. “Selling University Technology.” Management Science 48: 61-72.
Sharp, M. 1989. “European Countries in Science-Based Competition: The Case of
Biotechnology.” Designated Research Center Discussion Paper 72, Science Policy
Research Unit, University of Sussex, Brighton, U.K.
Simon, H. A. 1969. The Sciences of the Artificial. Cambridge: MIT Press.
The State of Academic Science: The Universities
Smith, B. L. R., and J. J. Karlesky. 1977.
in the Nation’s Research Effort. New York: Change Magazine Press.
Spencer, R. 1939. University Patent Policies. Chicago: Northwestern University Law

School.
Stanford University Office of Technology Licensing. 1983- Annual Report. Stanford,

Calif.: Stanford University.

_1990. Annual Report. Stanford, Calif.: Stanford University.
_ 1992. 1991-92 Annual Report: Office of Technology Licensing. Stanford, Calif.:

Stanford University.
_1994a. Copyrightable Works and Licensing at Stanford. Stanford, Calif.: Stanford

University.
1994b. Office ofTechnology Licensing Guidelines for Software Distribution. Stan¬

ford, Calif.: Stanford University.

Stata Corporation. 1999. Stata Statistical Software: Release 6.0. College Station, Tex.:

Stata Corporation.
REFERENCES

226 Stephan, P. E., S. Gurmu, A. J. Sumell, and G. Black. 2002. “Patenting and Publish¬
ing: Substitutes or Complements for University Faculty?” Unpublished working pa¬
per, Georgia State University, Atlanta.
Stern, N. 1981. From ENIAC to Univac. Bedford, Mass.: Digital Press.
Stokes, D. E. 1997. Pasteur’s Quadrant: Basic Science and Technological Innovation.
Washington, D.C.: Brookings Institution.
Swann, J. 1988.Academic Scientists and the Pharmaceutical Industry: Cooperative Re¬
search in Twentieth-Century America. Baltimore: Johns Hopkins University Press.
Thursby, J., R. Jensen, and M. Thursby. 2001. “Objectives, Characteristics and Out¬
comes of University Licensing: A Survey of Major U.S. Universities.” Journal of
Technology Transfer 26: 59-72.
Thursby, J., and M. Thursby. 2002. “Who Is Selling the Ivory Tower? Sources of
Growth in University Licensing.” Management Science 48: 90-104.
Tocqueville, A. de. 1990. Democracy in America, trans. by P. Bradley. New York:
Vintage.
Trajtenberg, M., R. Henderson, and A. B. Jaffe. 1997. “University Versus Corporate Pat¬
ents: A Window on the Basicness of Inventions.” Economics of Innovation and New
Technology 5:19-50.
Trow, M. 1979. “Aspects of Diversity in American Higher Education.” In H. Cans, ed.,
On the Making of Americans. Philadelphia: University of Pennsylvania Press.
- 1991. “American Higher Education: ‘Exceptional’ or Just Different.” In B. E.
Shafer, ed., Is America Different? A New Look at American Exceptionalism. New
York: Oxford University Press.
Trune, D., and L. Goslin. 1998. “University Technology Transfer Programs: A Profit/
Loss Analysis.” Technological Forecasting and Social Change 57:197-204.
Trustees of Columbia University in the City of New York. 1944. “Statement of Research
Policy and Patent Procedures.” Columbiana Library, “Patents” folder, Columbia
University, New York.
U.S. Congress Joint Economic Committee. 1999. Entrepreneurial Ehmamism and the
Success of U.S. High-Tech: Joint Economic Committee Staff Report. Washington,
D.C.: U.S. Government Printing Office.
U.S. Department of Health Education and Welfare. 1974. DHEW Obligations to In¬
stitutions of Higher Education and Selected Nonprofit Organizations, FY1965-1972.
Washington, D.C.: U.S. Department of Health, Education, and Welfare.
Investigation of Government Patent Practices and
U.S. Department of Justice. 1947.
Policies: Report and Recommendations of the Attorney General to the President.
Washington, D.C.: U.S. Government Printing Office.
U.S. General Accounting Office (GAO). 1968. Problem Areas Affecting Usefulness of
Results of Government-Sponsored Research in Medicinal Chemistry: A Report to the
Congress. Washington, D.C.: U.S. Government Printing Office.
-1995. University Research: Effects of Indirect Cost Revisions and Options for Fu¬
ture Changes. Washington, D.C.: U.S. Government Printing Office.
U.S. House of Representatives. 1980. Floor Debate on H.R. 6933. Congressional
Record, 96th Congress, Second Session, November 21: 30556-60.
 REFERENCES

U.S. Office of Management and Budget. 1963. “Memorandum and Statement of Gov¬ 227
ernment Patent Policy.” Federal Register 28: 10943-46.

-1971. “Memorandum and Statement of Government Patent Policy.” Federal Reg¬

ister 36: 16886.

- 1995- The Budget of the United States Government for Fiscal 1996. Washington,

D.C.: U.S. Government Printing Office.

U.S. Patent and Trademark Office. 1998. U.S. Colleges and Universities —Utility Patent
Grants, 1969-1998. Washington, D.C.: U.S. Government Printing Office.

U.S. Senate Committee on the Judiciary. 1979a. S. Rpt. 96-480 Accompanying S. 414,
the University and Small Business Patent Procedures Act. Washington, D.C.: U.S.
Government Printing Office.

- 1979b. The University and Small Business Patent Procedures Act: Flearings on
S. 414, May 16 and ]une 6. Washington, D.C.: U.S. Government Printing Office.
U.S. Senate Subcommittee of the Senate Select Committee on Small Business. 1963.

Economic Aspects of Government Patent Policies. Washington, D.C.: U.S. Govern¬

ment Printing Office.

Veblen, T. 1918. The Higher Learning in America: A Memorandum on the Conduct of

Universities by Business Men. New York: W. B. Huesch.
Vincenti, W. 1990. What Engineers Know and How They Know It. Baltimore: John

Hopkins University Press.

Walsh, J. P, A. Arora, and W. M. Cohen. 2003. “Research Tool Patenting and Licens¬

ing and Biomedical Innovation.” In W. M. Cohen and S. Merrill, eds., The Patent
System in the Knowledge-Based Economy. Washington, D.C.: National Academies

Press.

Weiner, C. 1986. “Universities, Professors, and Patents: A Continuing Controversy.”

Technology Review 83: 33-43.

Weissman, R. 1989. “Public Finance, Private Gain: The Emerging University-

Business-Government Alliance and the New U.S. Technological Order.” Under¬

graduate thesis, Harvard University.

White, H. 1963. Industrial Electrostatic Precipitation. Reading, Mass.: Addison-Wesley.

Wigler, M., S. Silverstein, L. S. Lee, A. Pellicer, Y. Cheng, and R. Axel. 1977. “Transfer

of Purified Herpes Virus Thymidine Kinase Gene to Cultured Mouse Cells. Cell
11: 223-32.

Wigler, M., R. Sweet, G. K. Sim, B. Wold, A. Pellicer, E. Lacy, T. Maniatis, S. Silver¬

stein, and R. Axel. 1979. “Transformation of Mammalian Cells with Genes from

Procaryotes and Eucaryotes.” Cell 16: 777—85.

Wildes, K. L., and N. A. Lindgren. 1985. A Century of Electrical Engineering and Com¬
puter Science at MIT, 1882—1982. Cambridge: MIT Press.

Williamson, O. E. 1979. “Transaction Cost Economics: The Governance of Contrac¬

tual Relation.” Journal of Law and Economics 22: 233-62.

Zacks, R. 2000. “The TR University Research Scorecard 2000.” Technology Review

(July/August): Available from World Wide Web: http://www.technologyreview

.com /articles /scorecardoyoo. asp.

Ziedonis, A. A. 2001. “The Commercialization of University Research: Implications for

REFERENCES

228 Firm Strategy and Public Policy.” Ph.D. diss., Haas School of Business, University

of California-Berkeley.

Zinsser, A. 1927. “Problems of the Bacteriologist in His Relation to Medicine and the

Public Health.” journal of Bacteriology 13: 147-62.

Zucker, L., M. Darby, and J. Armstrong. 1994. “Inter-Institutional Spillover Effects in

the Commercialization of Bioscience.” 1SSR Working Papers in Social Science 6.3.

 Index

academic patents: citations, 138-39; effect Ames II Tests: commercialization of, 172;

of, on basic research, 117-18; entrants vs. development of, 169,170,172; research

incumbents, 139-42; importance of, history of, 169-70; university patenting

138-39 and licensing, 170-71

academic research: Bayh-Dole Act effects Ames Tests: description of, 170; University

on, 85, 92,117,148,184, 2091220; biomed¬ of California Berkeley, 169

ical, 43; in chemistry, 29, 54; disclosure Anderson, M., 185

norms, and university patent policies, Apple, R. D., 39,19708

applied and engineering sciences in U.S.

95,127; focus of, 117; growth in, 23-25,

41, 62; and industrial research, 34; in lin¬ universities, 19-21

applied research: in agriculture, 38; in

ear model of innovation, 34,177; new

Bayh-Dole deliberations, 177; in biomed¬

structure of, 27; in pharmaceuticals, 28,

icine, 95; classification of, 19; in indus¬

34; in physics, 29—30; research exemp¬

tion for, 187-88. See also federal funding

try, 38; of industry, 195-96015; local ap¬

plications of, 46; in mining, 13; National

of academic research

Institute of Health (NIH) support for, 23;

aeronautical engineering, 17-18,194117,

role of, in universities, 188

19408
Armstrong, J., 117,177
Agrawal, A., 118

Arora, A., 21403

agricultural inventions, 104

ASET (consortium), 96
agriculture: applied research in, 38; univer¬

Association of American Universities, 94

sity research effect on, 29

Association of University Technology Man¬

Aiken, Howard, 19,19409

aircraft (DC series), 18

agers (AUTM): AUTM Licensing Survey
(1996), 82; major universities patent
air pollution control, 58

management, 45; patent data by, 146; re¬

Akasaki, Isamu, 160,162

port of licensing revenues by, 20601

American Association for the Advancement

Atanasoff, John, 19
of Science (AAAS) patent policy report,

Atomic Energy Commission (AEC), 23, 25

38> 59
Axel, Richard, 155,156,172,173
Ames, Bruce, 169
INDEX

23O Axel cotransformation patent: before Bayh- patents, 209024; on technology transfer,
Dole Act, 109; Columbia University pat¬ 117, 206025, 206026, 206027; on univer¬
ents covering, 157; income for Columbia sity patenting, 36,46, 85,126; on univer¬
University, 166; licensing structure of, sity research culture, 128
113; as major advance in gene transfer, Bayh-Dole bill, 88-92; debate over, 20301;
156; nonexclusive licensing of, 159 issues ignored during arguments, 205018,
Axel-Maddon patents, royalties from, 175 205019; issues ignored during debate on,
AZT, 172 19603; non-exclusive licenses, 204011
Ben-David, J., 12
Babcock, Stephan, 19708 Bercovitz, J. E. L., 147
Babcock test, 11,19708 Best, Charles, 19708
Banks, B. A., 199025 Biogen, 174,175
Banting, Frederick, 19708 biomedical discoveries, reluctance to pat¬
Barbosa, James, 157 ent, 19707
Basel Institute of Immunology, 174 biomedical inventions: biotechnology
basic research, 10, 30; academic patents inventions as portion of, 208010; growth
effect on, 117—18; vs. applied research, in number of, 102—3; increase in patent¬
117; agency funding of, 26; classification ing of, 50; licensing income from, 55,
of, 19-20, 26; driving patenting and li¬ 127; reports of, at Columbia University,
censing, 95; federal funds for, 43; as fo¬ 109-10
cus of Bayh-Dole Act, 206026; growth in, biomedical patents: class identification of,
194-95012; by industry, 195-96015; in 209022; of entrant universities, 51, 53;
linear model of innovation, 177; in medi¬ generality analysis of, 132; growth in
cal sciences, 21,128; as source of eco¬ number of, 35-36,103; home-runs, 69;
nomic growth, 22-23, Hi *n U.S. univer¬ of incumbent universities, 51; MRI pat¬
sities, 20 ents, 201019;vs- nonbiomedical patents,
Bayh, Birch, 89, 90 patent importance of, 139
Bayh-Dole Act: amendments to, 188, 21402; biomedical research, federal funding of, 25
assumptions of, 27,153,177; becomes ef¬ biomedical sciences, 43,182
fective, 92-93; before and after data, 115; biotechnology inventions: broad patents in,
as catalyst for university patenting, 89, 93; feasibility of licensing, 126; as portion
93-94; emulation of, 95-97,179, of biomedical inventions, 208010; uni¬
206024; encouragement of entrant uni¬ versity research effect on, 28
versities, 137; growth in university patent¬ Birenbaum, L., 186
ing attributed to, 36, 46,181; impact on Bito, Laszlo, 166,167,168
Research Corporation, 84; incentives for Black, G., 118
inventor involvement, 153; issues ignored Blum, Fred, 165
during arguments, 92; justifications for, Blumenthal, D., 55,185,186,19708
92; limitations of, removed, 93; origins Bolton, Elmer, 195-96015
of, 85; passage of, 91; policy shifts in, 112; Boston Globe, 172
political history of, 86-93; provisions of, Bremer, Howard, 21301
91; role of, 179 Britain, 96
Bayh-Dole Act effects, 93-95; on academic Broad, W., 89
research, 85, 92,117,148,184, 209020; on Brown, Edmund G. Jr., 206-702
disclosure numbers, 105; on patenting Brown, Rachel, 71
and licensing, 52,129; on relative impor¬ Burton, R., 147
tance of academic and nonacademic Bush, Vannevar, 19, 22-23, 64, 86
 INDEX

California Institute of Technology, 18 Cohen-Boyer patents: on DNA slicing 231

Campbell, E. G., 185,186 technique, 101; licensing revenues from,
Camras, Carl, 167 20707; licensing structure of, 113; nonex¬
Canada, 96 clusive licensing of, 159
Carnegie Commission on Higher Educa¬ Columbia University: Axel cotransforma¬
tion, 50 tion patent, 109,158,175; biomedical in¬
Carnegie Doctoral Universities, 51 ventions reports of, 109-10; biomedical
Carnegie-Mellon survey, 28, 29, 30-32 technology license share, K>7(fig); case
Carnegie Research Universities: classifi¬ studies of, 152,154; College of Physicians
cation of, 198022; patenting by entrants and Surgeons, 155,156,166; Columbia
and incumbents, 52(fig); patenting by Innovation Enterprise, 109; comparative
IPA status, 53(fig); patent policy and disclosure and licensing of, ii3(fig); com¬
practices of, 42; with Research Corpora¬ puter science inventions, 110; effects of
tion IAAs, 42(tab); RUis, 50-51, 52; Bayh-Dole on, 111; electrical engineering
RU2S, 50-51 inventions, 110; engineering curricula es¬
Carter, Jimmy, 91 tablished, 14; as entrant university, 119;
case studies: of Ames II Tests, 169-72; of Faculty Flandbook, 109; forward citations
Columbia University, 152; of cotransfor¬ of, i22(fig); generality of patents of,
mation, 155-59; of gallium nitride, 159— i24(fig), 125-26; invention disclosures of,
66; of soluble CD4,169,172-76; of Uni¬ no(fig); licensing income, i02(tab), 112,
versity of California, 152; of Xalatan, 158,168, 206m; number of patents of,
166-69 i2o(fig); Office of Science and Technol¬
case studies methodology, 154 ogy Development (OSTD), 109,123,167,
Causino, N., 185 173; Office of the General Counsel, 156;
CD4-based treatments. See soluble CD4 patenting and licensing in Bayh-Dole
Center for Telecommunications Research, Act, 129; patenting of, 1987118; patent liti¬
110 gation by, 157-58; patent policies of, 42-
chemical engineering: development in, 43,108-12,147; patents, by year of issue,
193-94715; industrial research in, 195- lii(fig); patents, importance of, 117-26;
967115; unit operations, 16 ramping up of invention reports, 109-10;
chemical patents, 53, 54 royalties from Axel-Maddon patents, 175;
chemistry, academic research in, 29 School of Mines, 14; Science and Tech¬
cherry-picking, 77, 80,141, 2021128, 202030, nology Ventures, 109; software inven¬

203034 tions at, 2087111; “Statement of Research

citation measures vs. yield, 127-28 Policy and Patent Procedure,” 108;
citations: academic patents, 138—39; and trends and categories of patents, 121-23;
correlation with licensing revenues, 128; and Xalatan patent, 153
dispersion of, 125; forward, 119,131; for¬ commercialization and development:
ward, of Stanford University, i22(fig); for¬ Ames II Tests, 171-72; cotransformation,

ward, of University of California, 158; gallium nitride, 164-65; soluble

i22(fig); importance based on, 138; as CD4,174-75; Xalatan, 164-65
measure of patent importance, 128; right commercialization of inventions, 85

truncated, 119; self-citations, 138 Compton, Karl, 64, 20071718,11

computer science, inventions of, 110
Clarridge, B. R., 186
coexclusive licensing, 173-74,175 computer science and engineering, 18-19

Cohen, W. M., 196016, 21403 conflict of interest, 84

Cohen-Boyer DNA slicing technique, 101 copyrights for software inventions, 111
INDEX

232 Cornell University, 15,40 Eisenberg, R. S., 91,188

cortical hormones, 71 electrical engineering, 15-16,110
cotransformation patents: commercializa¬ electrostatic generator, 64
tion and development, 158; licensing in¬ electrostatic precipitators: invention of,
come from, 155-56; research history of, 59; market for, 63; patents, 61; profitabil¬
156; summary of, 158-59; university ity of, 61; Research Corporation and,
patenting and licensing, 156-58 60
Cottrell, Frederick Gardner, 58, 59,181; on embryonic inventions: development and
founding of Research Corporation, 60; commercialization of, 91; uncertainty in
life of, 199113; on prospect argument, 82; bringing from lab to market, 91
role in Research Corporation, 61; on embryonic technologies: cotransformation
technology transfer, 199712; on university as, 158; inventor cooperation in develop¬
conflict of interest, 83 ment of, 153; uncertainty associated
Court of Appeals for the Federal Circuit with, 165; U.S. advantage in, 26
(CAFC), 93 engineering education: aeronautical engi¬
Cree Research, 162,163-64,165 neering, 17; chemical engineering, 16;
Crooks, Lawrence, 2011119 computer science and engineering, 18-
19; curricula established in U.S. universi¬
Dana Farber Institute, 174 ties, 14; development in, 12,14; electrical
Darby, M., 117,177 engineering, 15; in European universi¬
DC series (aircraft), 18 ties, 14
DenBaars, Steven, 162,163,164-65 ENLAC (Electronic Numerical Integrator
Department of Defense (DOD): in FCST and Computer), 18,19
report, 91; guidelines for patenting by, entrant learning: learning curve, 131,137;
79; IPAs signed, 88; patent policies of, from patent volume, 145; in post-Bayh-
45, 203112; R&D expenditures of, 23, 24 Dole era, 130; from Research Corpora¬
Department of Energy (DOE), 23, 25 tion relationship, 145; from Technology
Department of Health, Education, and Licensing Offices (TLO), 146
Welfare (HEW), 43, 45, 52; funding in¬ entrant universities: biomedical patents of,
crease by, 56; guidelines for patenting by, 51; characteristics of, 138; Columbia Uni¬
79; IPAs signed, 88; IPAs with, 204177, versity as, 119; definition of, 132; effects
213m; patent policies of, 87-88, 20409; of, 129; vs. incumbents, 139-44; interin-
stalled on patent assignments, 89 stitutional funding dispersion, 203034;
de Tocqueville, Alexis, 9-10,11,27 learning sources for, 144-47; patent im¬
Diamond, N., 11 portance of, 130-31,142-44; patenting
Diamond v. Chakrabarty, 93,126-27 by- 52(%)> 198^23
disclosure norms, 95 entry, defined, 198014
Discovery Partners International, 171 Epstein, S., 55,19708
Dole, Robert, 89, 90 European universities: control of, 10; engi¬
Douglas Aircraft, 18 neering education in, 14; faculty mobil¬
Dreyfuss, R., 188 ity vs. U.S. universities, 195014; scientific
Duke University, 188 research at, 195014
Dupont Company, 195-96015 exclusive licensing: costs and benefits of,
Durand, W. F., 17,19407, 19408 173; defined, 112; disadvantages of, 191;
for early-stage inventions, 173; under
Eckert, Presper, 18 IPAs, 57; issues in, 168-69; need for, 153,
Economist, 94 166,167,172,175; risks associated with,
Edison, Thomas A., 15 176; for Xalatan, 167
 INDEX

faculty inventors. See inventors assistance Forrester, Jay, 72 233

faculty mobility: university to industry, 34; Forrester patent, 71, 84, 201017
in U.S. universities, 13; of U.S. vs. Euro¬
pean universities, 193114,1951114 gallium nitride (GaN): applications for,
Farr, Spencer, 171 162; commercialization and develop¬
Fauci, Anthony, 172 ment, 164-65; in lasers and microwave
Federal Company, 15 transistors, 162; patent assignees for,
Federal Council for Science and Technol¬ i6i(tab); patents for, 166; R&D projects
ogy (FCST), 87; patent policy analysis, for, 159-60; research history of, 160-63;
20405; report data, 91; reports by, 90 summary of, 158-59,165-66; technical
federal funding of academic research, merits of, 159; university patenting and
24(tab), 25, 25(tab); importance of, to in¬ licensing, 163—64; z*nc selenide as alter¬
dustrial R&D, 33(tab); increases in, 50, nate to, 160-61
180; interinstitutional concentration of, Gee, Pauline, 169,170,171
43; interinstitutional dispersion of, Geiger, R., 11,40, 211012
211012; policy changes in, 79-82; post¬ Geiger “16”, 48
war, 41; postwar structure, 23-27; rise in Gelijns, A., 25
support for, 43-44; transfer and com¬ Genentech, 174,175
mercialization of, 85; and university pat¬ General Accounting Office (GAO), 87-88
ent policies, 206-702; World War II, 21- General Electric, 165
generality: analysis of, 132; of biomedical
23
federal funding of research and develop¬ patents, 132; calculations of, 125; of Co¬
ment: academic share of, 20404; postwar, lumbia University patents, i24(fig); de¬
scribed, 123,125; of entrants’ patents, 136;
87
federal funds: for basic research, 43; for of high and low intensity incumbents’
biomedical sciences, 43 patents, 133; of nonbiomedical patents,
federally funded inventions: patent rights 132; of post-Bayh-Dole era patents, 129; of
debate over, 86—87 University of California patents, i24(fig)
Federally Funded Research and Develop¬ generality measure, 125-26
ment Centers (FFRDCs), 23, 26, 43, 87 gene transfer, 155,156
federal patent policy. See government pat¬ geographic proximity, 75, 83, 202025

ent policy Germany, 96

federal public education policy and U.S. GI Bill, 19302,195013
university structure, 19302. See also Mor¬ Gilles, J., 82
glaucoma treatment, 153,166,167
rill Act
federal research and development: pluralis¬ Glennan, T. Keith, 20302
tic nature of, 26; post-wartime expendi¬ Gokhale, M., 186
tures, 23, 26; wartime expenditures, 22 Goslin, L., 84
government-funded patents, 90
Feldman, M., 147
government-operated laboratories, 24
Feller, I., 147, 211012
financial control and support of U.S. uni¬ Government Owned, Contractor Operated

versities, 11; vs. European universities, 12 Labs (GOCOs), 206022

government patent policy: changes in, 45;
Fishman, E. A., 69,198019
effects of on private contractors, 20301;
Flexner, Abraham, 10, 21
Memorandum on Governmental Patent
Florida State University, licensing revenues
Policy, 20405; of National Aeronautics
at, 20601
and Space Administration (NASA),
Fogarty, M. S., 199025
20302; by National Institute of Health
forestry, 29
INDEX

234 government patent policy: (continued) Princeton University, i98m8; of Re¬

(NIH), 188; by National Science Foun¬ search Corporation and MIT, 63-65,
dation (NSF), 79; Report on Government 1981119; of Stanford University, 198020
Patent Policy (1978), 90; and university IBM: Forrester patent, 71-72; Mark I com¬
research, 87-88 puter, 194119; relationship with MIT, 72
Government-University-Industry Research importance and generality: declines in, of
Roundtable (GUIRR), 27, 28-29 academic patents, 129; of Stanford Uni¬
Graham, H. D., 11 versity patents, 137; of University of Cali¬
Guggenheim Aeronautical Laboratory, 18 fornia patents, 137
Guggenheim Foundation, 18 importance of patents: academic, 138-39;
Gurrnu, S., 118 based on citations, 138; of biomedical vs.
nonbiomedical, 138-42; entrants vs. in¬
Harbridge House: report data, 91; reports cumbents, 139-44; *n post-Bayh-Dole
by, 87-88, 90 era, 123,129
Hazel, Elizabeth, 71 incumbent universities: Bayh-Dole Act ef¬
Henderson, R., 25,118,119,121,125,126,127, fects on, 129; biomedical patents of, 51;
129, 2097121, 2091124, 210114,21on26 definition of, 132,198023; vs. entrants,
Henig, R., 204ml 139-44; patenting by, 52(fig); technology
Hewlett-Packard, 163 transfer at, 117
higher education: competitive market in, industrial firms: absorptive capacity of,
193112; structural characteristics of, 9-13 158-59; and academic research, 34; in¬
Higher Education Act of 1972,193m novation and university research, 27—33;
high-intensity incumbents: definition of, knowledge flows of, to academic re¬
131; importance and generality among, search, 213015; public R&D value to,
210115; regression coefficients for, i34(tab) 33(tab); in technology transfer, 161; uni¬
Hilgartner, S., 186 versity science relevance, 3i(tab)
Hitch, Charles J., 206-7112 institutional experience in university
HIV virus, soluble CD4 as receptor for, 172 patenting, 131—37
Holtzman, N. A., 186 Institutional Patent Agreements (IPAs),45,
home-run patents: arrival rate of, 58; in bio¬ 5i-52, 56, 57, 88
medical applications, 62, 69; expiration intellectual property rights: Diamond v.
of, 71, 76; gross income to Research Cor¬ Chakrabarty, 93,126-27; domestic and
poration from, 7o(fig); and licensing in¬ international protection for, 206023;
come, 84; limitations of, 76; nystatin, ownership of, 17 96; role of, 59,176; U.S.
200-2011115; problems with reliance on, policy of, 185
84; shortage of, in Research Corpora¬ interinstitutional funding dispersion,
tion, 76-77; and university attention, 203034
196m Internal Revenue Service, 81
Hoskins, W., 196114 interuniversity personnel transfer, 147
Hounshell, D. A., i95ni5 invention disclosures and licenses, 112-15
Hughes Electronics, 162-63, *65 inventors assistance: vs. absorptive capac¬
hybrid-seed growing process, 71 ity, 158-59; in patent development,
hydrogen bomb, 19 205017; with patent process, 75; Pauline
Gee, 171,172; in technology transfer, 152,
lAAs (Invention Administration Agree¬ 166,167,168,176,177, 202024
ments), 40, 63; of Columbia University, inventor startup firms, 153,174
i98m8; cost sharing under, 145; of IPAs, exclusive licensing under, 57
 INDEX

Jaffe, A. B., 118,119,121,125,126,127,129, at Yale University, 20601. See also cherry- 235
1997725, 209021, 209024, 21004, 210026 picking
Japan, 96 linear model of innovation, 34,177
Japanese universities, 96 Little, Arthur D., 16
Jensen, R., 91,152,168, 203035 Lodge, Sir Oliver, 59
Jones-Mangelsdorf patent, 71 Louis, K., 185
low-intensity incumbents: importance and
Kastenmeier, Robert, 205018 generality among, 21005; regression
Katz, M. L., 206023 coefficients for, i35(tab)
Kendall, Edward, 71 LumiLEDS, 163
Kennedy, John F., 87, 20405
Kilgore, Harley, 86 Maddon, Paul, 173,174
Kilgore bill, 20301 Madey v. Duke, 188, 21403
Kitch, E. W., 59 magnetic core memories, 71
klystron, 15 mainframe computers, 24
knowledge-based economy, 183 Manhattan Project, 22
knowledge spillovers, 202025, 202027 Mansfield, E., 27,28-29
Koch, Robert, 21 march-in rights, 93
Maron, Dorothy, 169,170
Lamont-Doherty Earth Observatory, 110 Matsushita, 160,163
Lamoreaux, N., 202025 Mauchly, John, 18,19
Lanjouw, J. O., 21107 Maxwell, J., 55,19708
large corporations vs. small business, 86, McKusick, V., 40,19601
204-5012, 205018 Medical Education in the United States
lasers and microwave transistors, 162 and Canada (Flexner, Abraham), 21
Latker, Norman, 20409 medical patents, 19707; opposition to, 38;
learning and learning curve: by entrant as revenue source, 56. See also biomed¬
universities, 130,131,137,143,145,147; by ical patents
networking, 147; and research budget, medical sciences, 21
211012 Memorandum on Governmental Patent Pol¬

Lesley, E. R, 17,19407,19408 icy, 20405

Lewis, Warren, 16 Milas, Nicholas, 55, 64
licensing agreements, limits of, 83 Mines Experiment Station, 13
licensing income: from biomedical inven¬ Mishra, Umesh, 162,163,164-65
tions, 55; from Cohen-Boyer patents, MIT: chemical engineering established,
20707; at Columbia University, io2(tab), 16; Committee on Patent Policy, 64;
112, 20601; from cotransformation pat¬ electrical engineering established in, 15;

ents, 155-56; decline in yield of, 116; dur¬ engineering curricula established, 14;

ing depression era, 63; distribution of, and IBM relationship, 72; patent policy

65; at Florida State University, 20601; of, 64; relationships with industry, 17; Re¬

and home-run patents, 84; and patent ci¬ search Corporation and Invention Ad¬
ministration Agreement (IAA), 63-65;
tations, 128; shift from focus on, 84; from
site licenses, 207—808, 208013; from sol¬ wartime R&D expenditures, 22

uble CD4 treatments, 213019; at Stanford MIT IAA with Research Corporation, 63-
65; MIT IAA cancellation, 71-72; MIT’s
University, io2(tab), 106,112, 20601; at
University of California, io2(tab), 104, Forrester patent, 84; origins of, 20008

112, 20601; from university patenting, 127; molecular biology, development of, 44
INDEX

236 Moore School of Electrical Engineering, open science, 37, 85

18-19 Ordover, J. A., 206023
Morrill Act, 9,10,11,193112 organic compounds, 54
Mowery, D. C., 2021125, 2101129 Organization for Economic Cooperation
MRI patents, 201019 and Development (OECD), 93-94
Orsenigo, L., 25
Nagoya University, 160,162 ownership rights for researchers, 96
Nakamura, Shuji, 160,161,162, 21207,
212011 Palmer, A. M., 42
National Aeronautics and Space Adminis¬ Pankove, Jacques, 160, 21207
tration (NASA), 24-25, 20302 Pasteur, Louis, 21
National Institute of Health (NIH), 25,43, Pasteurs Quadrant, 20, 26,180,185,194010
1.56 —57; limits on patenting by, 188; med¬ Patent, Trademark, and Copyright Office
ical chemistry program, 87; National (PTCO), University of California, 100
Eye Institute, 166 Patent Awareness Program, of Research
National Institute of Standards and Tech¬ Corporation, 75
nology, 165 patent citations. See citations
National Research Council, 27 patent importance. See importance of
National Science Foundation (NSF), 24, patents
45, 52, 75,163; guidelines for patenting patent litigation, 157, 200011,201018,
by, 79; IPAs signed, 88; IPAs with, 213111; 201021, 213016
survey of, 118 patent management: academic, difficulties
Nature, 174 of, 71; across technical fields, 69, 72-73,
Nelson, R. R., 196016 201020; growing costs of, 72—73; high
Nichia Chemical, 160 costs of, 58; Research Corporation trains
Nitres, 164-65,165 universities in, 75-76; unit costs of, 72;
Nixon, Richard M., 87 university involvement in, 43-46
nonbiomedical patents: generality analysis Patent Management Division, of Research
of, 132; regression coefficients for, lzp(tab) Corporation, 58, 65-79; closed, 81; clo¬
nonexclusive licenses: dropped provision of sure of, 79; costs of, 67; invention portfo¬
Bayh-Dole bill, 204011; revenue stream lio of, 69-71; responsibilities of, 66; sup¬
from, 113; university patent policies on, port for, 73; university clients of, 67
20705 patent pirates, 36, 37,196114
nonexclusive licensing: of Axel patents, patent policies and practices: of Carnegie
157,158,159; of Cohen-Boyer patents, 159 Research Universities, 42; of Columbia
nonlicensee firms, development activities University, 42-43; effects of different,
of, 174-75 204115; university changes to, 44-46. See
North Carolina State University, 20-21,162 also government patent policies
nystatin, 71, 200-201015 patent propensity of universities, 181
patents and patenting: ambivalence to¬
Office of Naval Research, 163 ward, 35; dispersion in, 50; donations, 62;
Office of Science and Technology Devel¬ extension of, 185; importance of univer¬
opment (OSTD), 123,157 sity patents, 138-44; need for, 165,168,
Office of Scientific Research and Develop¬ x72> U5; need for, by startups, 177; need
ment (OSRD), 22 for, unproven, 184; number of, by year of
Office of Technology Licensing (OTL), 45, issue, i33(tab); as obstacles for scientific
104-5 research, 185-89; utility of, 186; values
Office of Technology Transfer (OTT), 100 of, 190
 INDEX

penicillin, 194ml research and development: in wartime, 237

Pfizer Medical Systems, 2011119 18-19
pharmaceuticals, 28 research collaboration, 95
Pharmacia, 167,168 Research Corporation, 39,44, 48, 55; after
physics, 29 1980, 81-82; Annual Reports of, 65, 67,
Pisano, G., 25 73, 77, 80; assistance to Columbia’s fac¬
Poillon, Howard, 64, 200118 ulty, 108; Bayh-Dole Act, support of by,
political risks, 92 81; Boston office for MIT, 75; as chari¬
postwar structure of federal funding of table foundation, 81; clients of, and en¬
academic research, 23-27 trant learning, 144,145; Committee on
precipitation patents, 61 Goals and Objectives of the Corpora¬
Princeton University, 160,1987118 tion, 77; cost increases of, 74; current
Proceedings of the National Academy of operations of, 82; decline of, 79- 82; do¬
Sciences, 169 nated inventions, 71; early activities, 61-
productivity, indicators of, 208012 63; evolution of, 61-65; expansion of,
Progenies, 174,175 62, 74; expenditures of, 73(fig); financial
proofs of concept, 152 condition of, 76, 2021127; Forrester patent
propeller tests, 17,19408 management by, 71; founding of, 60;
prospect argument, for university patent¬ gross income from home-run inventions,
ing, 82 7o(fig); gross royalties of, 68(fig); history
prostaglandins (PGs), 166,167 of, 58; home runs, shortage of, 76-77;
The Protection of Patents by Scientific IAA data, 42; IAAs negotiated, 66(fig),
Discoveries (AAAS), 19602 66-67; IAA termination with MIT, 72;
public R&D. See federal funding of aca¬ Internal Revenue Service ruling, 81; in¬
demic research vention disclosures to, 67(fig); legacy of,
public universities, local funding focus of, 83; and limits in scale economies, 77-
79; litigation of, 61; negotiations with
57 -

public vs. private universities: growth pat¬ IBM, 72; net royalty income, 79(fig); op¬
ent management role of, 52; university erating costs growth, 77; origins of, 59-

patent policies of, 46,49 60; patent management, 40, 61; patent
Purdue University: research foundation at, management costs growth, 72-73; patent
management scale of, 65; patent reve¬
40
nues source, 56; patents issued to,
68(fig); patents of, 53; patent withhold¬
Quaker Oats, 39
ing, 77, 2027129; purpose of, 60; relation¬

R&D projects: external information ship with entrants, 138; reliance on bio¬
medical patents, 69; role of, 181; share of
sources in, 32(tab)
U.S. university patents, 8o(fig); Stanford
Rai, A. T., 188
income from, 45; support to Research
Rajchman, Jan, 71, 72
Corporation Technologies (RCT), 81-
RCA, 71
82; total costs per IAA, 78(fig); total oper¬
Reagan, Ronald, 93
ating costs, 78(fig); universities trained in
recombinant DNA method, 156, 21205
patent management by, 75-76; visitation
Reimers, Niels, 147, 21205
and marketing efforts of, 73-75; World
Rensselaer Polytechnic Institute (RPI), 14
War II effects on, 65. See also cherry-
Report on Government Patent Policy (1978),
picking; electrostatic precipitators; MIT
90 r ... IAA with Research Corporation; Patent
research: commercialization of, 15; publi¬
Management Division
cation of, 17
INDEX

238 Research Corporation Technologies Sokoloff, K., 202025

(RCT), 81-82 soluble CD4: as AIDS drug prototype,
research exemption for academic research, 172; case study of, 169; clinical trials
187-88 for, 175; commercialization and devel¬
research foundations, 40 opment, 174-75; developments in, 172,
research history: of Ames II Tests, 169-70; 174; licensing revenues from, 213019; re¬
of cotransformation, 156; of gallium ni¬ search history of, 172-73; summary of,
tride, 160-63; °f soluble CD4,172-73; 176; university patenting and licensing,
of Xalatan, 166-69 273-74
research process patent, 212714 Spencer, R., 19709
reserpine, 71 Sputnik crisis, 23
Roosevelt, Franklin D., 22 Staats, Elmer, 20408
Rosenberg, N., 25 Standard Oil of New Jersey, 17
RUs. See Carnegie Research Universities Stanford Physics Department, 15
Stanford University: before and after data
Salmonella typhimurium bacteria, 170 at, 116; before and after patent analysis,
Sampat, B. N., 2101129 118; biomedical inventions, 105-6; bio¬
Schankerman, M., 211117 medical technology license share,
Science, 90,174 i07(fig); comparative disclosure and li¬
science-based industries, 95 censing of, ii3(fig); disclosure and licens¬
Science: The Endless Frontier (V. Bush), 22, ing of, ii4(tab); disclosures, patenting
203m and licensing of, 115; disclosure share of,
scientific commons, 185,186 116; disclosures of, io6(fig), 112; Federal
scientific research: and clinical practice Company, 15; forward citations of,
combined, 21, 25; patents as obstacles to, i22(fig); generality of patents of, i24(fig),
185-89; restrictions to, by patenting, 186. 125-26; importance and generality of
See also academic research patents of, 137; licensing income,
secrecy agreements, 163 i02(tab), 106,112, 206m; marketing effort
SELETE (consortium), 96 of, 116; Office of Technology Licensing
SEMATECH (Semiconductor Manufac¬ (OTL), 45,104-5; patenting and licens¬
turing Technology), 96 ing at, 115-17,129; patent policies of,
Senate Judiciary Committee, 90, 2051117 104-8, 207715, 20706; patents, by year of
Sevringhaus, E. L., 19605 issue, io6(tab); patents, importance of,
Sharp, M., 195014 117-26; patents, number of, i2o(fig); pro¬
Siemens, 163 peller tests, 17-18; site licenses from, 108;
Simon, Herbert, 19, 20 Software Distribution Center, 108; soft¬
Singer, Jerome, 201019 ware inventions, 108,116; university pat¬
site licenses, 108, 207-808, 208013, 209018 ent policies exemptions, 20706; yield of
small businesses, 86, 204-5012, 204012, licensing at, 116-17
205018 statistics, 20
Smith, J. K., 195015 Steenbock, Harry, 39,19708
SmithKline Beecham, 174 Steenbock patents, 39, 55,19709
Smithsonian Institution, 60 Stephan, P. E., 118
software inventions: at Columbia Univer¬ Stevens Institute of Technology, 14
sity, 208011; copyrights for, 111; licensing Stokes, Donald, 20,194010
revenues from, 127; at Stanford Univer¬ Sumell, A. J., 118
sity, 108; university patent policies ex¬ Swann, J., 195-96015
emptions, 20706 Sweden, 96
 INDEX

tacit knowledge, 175 of Patents, 100; case studies, 152,154-55; 239

Taft, William Howard, 60 Cree Research royalties, 165; data on dis¬
taxpayer-benefit arguments for university closures, patenting, and licensing at,
patenting, 92 ioo(fig), ii4(tab), 115; forward citations
technological innovations, effect of univer¬ of, i22(fig); General Counsel’s Office,
sity research on, 27-28 100; generality measure of patents of,
technology licensing offices (TLOs): en¬ 125-26; generality of patents of, i24(fig);
trant learning from, 146; and entrants’ importance and generality of patents of,
patent importance, 144-45 137; licensing income, io2(tab), 104,112,
Technology Review, 94 206m; marketing effort of, 116; MRI pat¬
technology transfer: and academic re¬ ent litigation, 201/119; number of patents
search, 85; diverse channels of, 179; ef¬ of, i2o(fig); Office of Technology Trans¬
fects of Bayh-Dole on, 117; inventors as¬ fer (OTT), 100,170; Patent, Trademark,
sistance in, 152,166,167,168,176,177, and Copyright Office (PTCO), 100; pat¬
202024; patents needed for, 36; process ent importance and generality of, 118;
of, 152; in science-based industries, 95; as patenting and licensing at, 115-17;
a tax, 212/15; university developments in, patenting and licensing in Bayh-Dole
80; by university patenting, 200014; uni¬ Act, 129; patents, by year of issue,
versity to industry, 34,152-78; by U.S. ioi(fig); patents, importance of, 117-26;
universities, 44,45 patents issued to, 101; Salmonella Muta¬
third-party management agents, 40 genicity Test Resource Center, 170; se¬
Thursby, J., 117 crecy agreements with, 163; trends and
Thursby, M., 91,117,152,168, 203035 categories of patents, 121-23; UC Patent
Toshiba, 163 Office, 100; University Patent Fund, 100;
Trajtenberg, M., 118,119,121,125,126,127, university patent policies, 207/13; yield of
129, 209021, 209024, 21004, 210026 licensing at, 116-17
Trow, M., 11 University of California, Berkeley, 54“55>
truncation bias, 126, 208012, 210029 169
Trune, D., 84 University of California, Santa Barbara,

Turing, Alan, 18 162,163

University of Colorado, 162
unit operations, 16,193-9405 University of Minnesota, 13,40
universities: autonomy of, 10; focus of, 10 University of Pennsylvania, 18-19
University and Small Business Patent Act, University of Toronto, 197/18
University of Wisconsin, 12, 39. See also
89
university funding and community focus, WARF
university patenting and licensing, 198/118;
19303
university-industry interaction, 95 Ames II Tests, 170-71; arguments for, 35;

university-industry research links: charac¬ and basic research, 38; before the Bayh-

teristics of, 179; growth of, 184; postwar, Dole act, 35-57; benefits and risks of,
184-92; characteristics of, 137-47; co¬
195015; prewar, 41
transformation, 156-58; difficulties with,
University of Akron, 12
199-200/17; direct involvement of uni¬
University of California: Ames II Tests, roy¬
versities in, 46-47; dispersion and entry
alty payments for, 171; before and after
in, 50-53; entry, importance and gener¬
patent analysis, 118; biomedical disclo¬
ality in, 131-37; entry and experience, ef¬
sures, io3(fig); biomedical inventions of,
fects of on, 129-48; gallium nitride, 163-
103; biomedical patents, 103; biomedical
64; growth of, 47-48, 50, 56,182,198/116;
technology license share, io7(fig); Board
INDEX

24O university patenting and licensing, pharmaceuticals, 28; on R&D projects,

(continued) 31; on taconite extraction and refine¬
licensing revenues from, 127; of MIT, ment, 13
198019; motivation for, 38; negative ef¬ Upjohn Pharmaceuticals, 168,174
fects of, 37; political criticism for, 56-57; U.S. academic research, federal funding
potential risks created by, 92, 97,186; for. See federal funding of academic
pre-Bayh-Dole era, 35; prewar, 196m; research
of Princeton University, 198018; propo¬ U.S. Army, 19
nents of, 37; Research Corporation, 63; U.S. Department of Agriculture (USDA),4i
soluble CD4,173-74; sources of data on, U.S. Food and Drug Administration
198015; technology class distribution, (FDA), 168
54(fig), 55(hg); Xalatan, 167. See also Co¬ U.S. Military Academy, 14
lumbia University; MIT; Research Cor¬ U.S. Patent and Trademark Office, 77, 94,
poration; Stanford University; University 119,185,186
of California U.S. patent system, 186
university patent policies: during 1960s and U.S. Supreme Court, 188; Diamond v.
1970s, 44; before the Bayh-Dole act, 35- Chakrabarty, 93,126-27
57; for biomedical patents, 40,43; U.S. universities: basic research, 20; central
changes to, 189; of Columbia University, business of, 188; conflict of mission with
147,197—98012; conflicts between uni¬ patenting, 189,191-92; enrollment of, 11;
versity and faculty, 201019; conflict with faculty mobility in, 13; financial control
scientific research, 196-9706; develop¬ and support of, 11; financing research at,
ment of, 197010; disclosure norms of aca¬ 22; focus of, 180; patent propensity of,
demic research, 95; effect of on univer¬ 181; and pharmaceutical industry, 195-
sity mission, 189,191-92; effects of, on 96015; relationships with industry, 13,16,
scientific research, 59; and federal fund¬ 17; research culture vs. patenting, 182; re¬
ing of academic research, 206-702; goals search orientation of, after Bayh-Dole,
of, 189-90; history of, 36-46; impact of 184; structural differences from Euro¬
World War II on, 41-43; by institution, pean universities, 33; structure of, 21-22,
4i(tab); for medical patents, 43; on 183; and technical changes in industry,
nonexclusive licenses, 20705; political 180-84
risks of, 179; postwar, 42; prewar, 38-41; utility requirement for patentability, 186
public vs. private universities, 46,49; Re¬
search Corporation assistance in, 79-80; Van de Graaf, Robert, 64
on software inventions, 20706; and Varian, Hal, 15
WARF, 39-40 Varian, Sigurd, 15
university patents: 1925-80, 47(fig); basic Veblen, T., 10
trends in, 47-49; by institution, 49(tab); Vincenti, W., 18
per R&D dollar, 48(fig); by technologi¬ visitation program of Research Corpora¬
cal field, 53-56; by university category, tion: development of, 74, 203032; failure
51(fig) of, 76; and licensing income, 201-2023;
university research: in agriculture, 29; in motivation of, 201022
biotechnology, 28; contributions of to vitamin B„ 54, 62, 71
technological innovations, 27-28; in vitamin D, 38
forestry, 29; and government patent pol¬ vitamin patents, 64
icy, 87-88; and industrial innovation, vitamin research, 54-55
27-33; industry rating of, 3o(tab); on Von Neumann, John, 19
 INDEX

Walcott, Charles, 60 Xalatan: commercialization and develop¬ 241

Wall Street Journal, 2137115 ment, 164-65,168; for glaucoma treat¬
Walsh, J. P., 1967716, 214773 ment, 166; research history of, 166-69;
WARF (Wisconsin Alumni Research Foun¬ summary of, 168-69; university patent¬
dation), 49, 63,181; as model for others, ing and licensing, 153,167
39-40; patent revenue source, 56; Steen- Xenometrix: Ames II licenses, 171; Ames
bock patents, 39, 55,19777778,9 Tests assays, 212012; development of, 171
wartime R&D expenditures, 194ml, 2007711
Waterman, Robert E., 62 Yale survey, 28, 29-30,196016
Western Electric, 22 Yale University, 14, 20601
Widegap Technologies, 164 yield: vs. citation measures, 127-28; decline
Wiles, R., 19604 in, 116,128; and intensity, 115,116; of mar¬
Williams, Robert R., 62 keting efforts, 115; at Stanford University,
Williams-Waterman method, 54, 62 116-17; at University of California, 116-
Wilson, Carroll, 2007711 U
Windham, Patrick, 89
Wisconsin Alumni Research Foundation Ziedonis, A. A., 202025,210029
(WARF). See WARF Zinsser, A., 19707
Woodward, Robert, 71 Zucker, L., 117,177
World War II and federal funding of aca¬ Zuse, Konrad, 18
demic research, 21-23
WYLBUR operating system, 108
 .

'
.
 *