Federal S&T Programs and Technology Transfer
Domestic and International Technology Alliances
In recent decades, the speed, complexity, and multidisciplinary
nature of scientific research, coupled with the increased relevance
of science for industrial technology development and the demands
of a globally competitive environment, have increased the importance
of technology linkages for innovation and long-term competitiveness
(Branscomb and Florida 1998).
Although external technology sources, including university research,
have long played a key role in U.S. industry innovation and competitiveness
(Mowery 1983; and Rosenberg
and Nelson 1994), the current environment has encouraged an
innovation system increasingly characterized by networking and feedback
among R&D performers, technology users, and their suppliers
and across industries and national boundaries (Coombs
and Georghiou 2002; and Vonortas
1997). Several Federal S&T policies have also facilitated
private R&D collaboration and Federal technology transfer, as
discussed in more detail throughout this section. (See sidebar,
"Major Federal Legislation Related to Cooperative
R&D and Technology Transfer.")
Available indicators reveal increased cross-sector linkages over
the 1990s. Manufacturing companies increased contract R&D expenditures
at a 4.8 average annual percent rate, in real or inflation-adjusted
terms, between 1993 and 2001, a full annual percentage point higher
than the growth of in-house company-funded R&D expenditures
over the same period. Federal agencies reporting technology transfer
data to DOC increased their invention disclosures, patent activity,
and licensing in FY 2001, reflecting their unique capabilities in
terms of multidisciplinary R&D and specialized facilities. Patents
issued to these Federal agencies topped 1,600 in FY 2001, up 15.6
percent from FY 2000.
The other major intersectoral activity involves cooperative R&D.
U.S. Federal agencies participated in more than 3,600 Cooperative
R&D Agreements (CRADAs) with industrial and nonprofit organizations
in FY 2001, although new CRA-DAs have been stable at about 1,000
annually since FY 1997. In addition, between 1991 and 2001, U.S.
companies participated in more than 4,600 research and technology
alliances worldwide, or about 80 percent of all such alliances involving
U.S., European, Japanese, and emerging-market companies. Activity
was particularly strong in IT and biotechnology.
Outsourcing and collaboration aimed at the acquisition or development
of technologies may reduce costs, expedite projects, or complement
internal R&D capabilities (Howells
and James 2001). Activities linking business, academic, and
government laboratories may take place in special-purpose settings
such as science parks. (See sidebar, "U.S. Science
Parks.") The following sections discuss data on contract R&D,
Federal technology transfer (e.g., patent licensing), and R&D
alliances involving private companies, universities, and government
Many companies have increasingly come to rely on other firms for
a portion of their R&D needs. In fact, the growth rate of contract
R&D, defined as company-funded R&D performed externally,
exceeded that of company-funded R&D performed in-house in recent
years, even after a decline in contract R&D expenditures in
2001. In 2001, more than 1,300 manufacturing companies (8 percent
of all R&D-performing manufacturing companies in the United
States) reported $4.0 billion ($3.6 billion in constant or inflation-adjusted
dollars) in expenditures for contract R&D performed in the United
States, compared with $4.8 billion ($4.5 billion in constant dollars)
in 2000, a decline of 17.5 percent, according to NSF's Survey of
Industrial Research and Development.
In contrast, their in-house company-funded R&D declined only
1.4 percent between 2000 and 2001. Over a longer time span, however,
manufacturing companies increased contract R&D expenditures
at a 4.8 average annual percentage rate in real, or inflation-adjusted,
terms, a full annual percentage point higher than the growth of
in-house company-funded R&D expenditures between 1993 and 2001,
reflecting the importance of outside sources of technology for a
number of corporate technology objectives (appendix
table 4-36 ).
In the manufacturing industry the overall ratio of expenditures
for contract R&D to expenditures for R&D performed in-house
increased from 3.3 percent in 1993 to a peak of 4.7 percent in the
mid-1990s, then moderated somewhat to 3.6 percent in 2001 (figure
In 2001 the proportion was higher for chemicals manufacturing at
11.7 percent (and pharmaceuticals manufacturing at 18.7 percent)
Within nonmanufacturing industries, the contract R&D ratios
for the information sector and the professional, scientific, and
technical services sector were notable at 3.3 and 7.4 percent, respectively.
Within the latter industry, R&D services contracted out $1.3
billion in R&D activities in 2001, which is 12.0 percent of
its $10.9 billion in internal company-funded R&D expenditures.
Of the manufacturing companies reporting contract R&D in the
NSF survey in 2001, 132 companies (9.7 percent) identified $2.17
billion in R&D expenditures in terms of their R&D contractors
being for-profit companies, universities and colleges, or other
The highest proportion of these identified contract R&D expenditures,
92.0 percent, funded other companies, 5.9 percent funded universities
and colleges, and 2.2 percent funded other non-profit institutions.
For chemical companies, the distribution of contract R&D expenditures
among their R&D contractors was similar (83, 12, and 5 percent,
respectively). However, among companies in the scientific R&D
services sector, the share of identified contract R&D expenditures
performed by universities and colleges was much higher, 35.4 percent,
although still second to the 49.7 percent performed by other for-profit
The relatively higher reliance of U.S. R&D services companies
on universities and colleges as R&D subcontractors may be related
to the broader set of technologies in which these companies work,
complementing their internal capabilities with the wide array of
scientific capabilities of universities.
Federal S&T Programs and Technology Transfer
Concerns over U.S. industrial strength and global competitiveness
in the late 1970s and early 1980s led to a series of legislative
changes that collectively created an environment conducive to industry-government
collaboration in technology development (Link
1999). This section discusses technology transfer and collaborative
activities involving Federal laboratories. Technology transfer
can be defined as the exchange or sharing of technical knowledge,
skills, processes, or products across different organizations.
Technology transfer activities involving Federal laboratories include
patenting, licensing, joint R&D, user-facility agreements, and
Technology transfer functions performed by certain Federal laboratories,
namely, intramural or government-owned-government-operated laboratories,
such as NIH or the Agricultural Research Service, were established
by the Stevenson-Wydler Technology Innovation Act of 1980 (Public
Law 96-480). Later in the decade, the Federal Technology Transfer
Act of 1986 authorized intramural laboratories to enter into CRADAs
with industrial partners, universities, and other organizations,
whereas the FY 1990 DOD Authorization Act (Public Law 101-189) extended
this authority to government-owned-contractor-operated laboratories,
including government-owned FFRDCs
(Schacht 2000). In CRADAs, Federal
laboratories may share or provide personnel, services, equipment,
or facilities (but not funds) with or to a private organization
as part of a joint R&D project with the potential to promote
industrial innovation consistent with the agency's mission. Private
partners may retain ownership rights or acquire exclusive licensing
rights for the developed technologies. More recently, the Technology
Transfer Commercialization Act of 2000 (Public Law 106-404) enhanced
the ability of Federal agencies to license (and monitor) federally
R&D Funding Trends in Federal Laboratories
The share of Federal R&D obligations devoted to intramural
laboratories and FFRDCs declined from 39 percent in the early 1980s
to the low 30s in the late 1990s (NSF forthcoming). Still, the role
of Federal laboratories, either as a source of technology to be
commercialized by private parties or as a research partner, is considerable.
Federal laboratories offer industrial and nonprofit researchers
unique capabilities, such as the ability to perform interdisciplinary
research and to use expensive, specialized equipment (Bozeman
In FY 2001 the Federal Government obligated $27.3 billion, or 34
percent of $79.9 billion in Federal funds earmarked for R&D,
to Federal laboratories (table
compared with $52.6 billion (66 percent of total) in R&D funding
obligated to extramural performers, such as companies and universities
(NSF forthcoming). Within individual agencies, the share devoted
to government laboratories is largest for DOE (71.7 percent) and
smallest for HHS (20.3 percent; 19.6 percent for its NIH component).
Agencies with large amounts or relatively large proportions of their
R&D obligations devoted to intramural and FFRDC performers have
more internal outputs available for patenting and licensing than
agencies that channel their R&D funds to extramural performers.
Federal agencies devoted a higher share of their funds for Federal
laboratories to applied research and development than to basic research.
Of the 34 percent devoted to Federal laboratories in FY 2001, less
than a fourth went to basic research. Individual Federal agencies,
however, varied considerably in the proportion of funds they devoted
to basic research in their laboratories: 52.4 percent of HHS laboratory
R&D funding (59.5 percent for its NIH component), followed by
USDA (49.0 percent) and DOE (35.0 percent). DOD devoted only 5.3
percent to basic research in its laboratories. This profile of character
of work at Federal laboratories, together with the various S&T
emphases of these agencies, suggests that industrial partners are
potentially able to use Federal facilities as a source for a variety
of research outputs.
Federal Technology Transfer Trends
Since FY 1987, 10 Federal agencies have reported data on technology
transfer to the DOC, pursuant to Federal technology transfer statutes
(U.S. DOC 2002).
The 10 agencies reporting data were DOC, DOD, DOE, DOI, the Department
of Transportation, the Environmental Protection Agency, HHS, NASA,
USDA, and the Department of Veterans Affairs. In general, available
metrics indicate an increased level of Federal technology transfer
activities since the late 1980s. Data include inventions disclosed,
federally owned patents, licenses, licensing income, and the number
In FY 2001, Federal agencies reporting data on technology transfer
activities logged more than 3,900 invention disclosures (table
Invention disclosures increased 9.7 percent from FY 2000, close
to the 4,000 mark reached in the early and mid-1990s (figure
Patent applications increased to a peak of 2,172 in FY 2001, up
4.3 percent from FY 2000, after remaining at or just below 2,000
for most of the 1990s. Patents issued to these Federal agencies
reached 1,608 in FY 2001, up 15.6 percent from FY 2000. Between
FY 1997 (the first fiscal year for which these data were available
from DOC) and FY 2001, a total of 7,178 patents were issued to these
10 Federal agencies.
At the agency level, DOD and DOE had the largest shares of inventions
disclosed, patent applications, and patents issued in FY 2001. These
two agencies accounted for 6575 percent of those Federal technology
transfer indicators among the reporting agencies. Differences in
R&D funding structures and character of work may drive some
of these results at the agency level. Furthermore, Federal agencies
are engaged in other technology-related activities (e.g., technology
procurement, safety or material standards, and technology assistance
to businesses), offering other venues for technology diffusion not
covered in this section.
Federal Laboratories in Collaborative Research Agreements
Two indicators of Federal laboratories' participation in research
alliances show selected features of these activities: the first
identifies their industrial focus, and the second describes Federal
agency participation in CRADAs.
Ninety-nine R&D agreements registered from 1985 to 2001 in
the Federal Register (11.5 percent of 861 R&D agreements)
had at least one Federal laboratory partner.
Thirty-seven of these industry-government R&D alliances were
classified in electronic and other electrical equipment and components
Ten alliances were classified in chemicals manufacturing (which
includes pharmaceuticals), another 10 in industrial machinery and
computer equipment manufacturing, and eight in transportation equipment
manufacturing. Leyden and Link (1999) report that registered alliances
with Federal laboratory partners tend to have more participants
than do alliances without government partners. Federal laboratories
in large alliances not only increase economies of technological
scope but also reduce monitoring costs, increasing potential benefits
to all members (Leyden and Link 1999).
The 10 Federal agencies reporting technology transfer activities
to DOC executed 926 new CRADAs with industrial and university partners
in FY 2001, up 5.9 percent from FY 2000, but little changed from
the 1,000 mark since first reported in FY 1997. The 2001 increase
brought the number of active CRADAs to 3,603 (figure
Three agencies accounted for more than 80 percent of active CRADAs
in FY 2001: DOD, which participated in 1,965 CRADAs, or 54.5 percent
of all CRADAs; DOE, which participated in 558, or 15.4 percent;
and HHS, which participated in 490, or 13.6 percent.
The FY 2001 increase in active CRADAs was driven by increases in
DOD and HHS CRADAs (44 and 12 percent, respectively) compared with
a 19 percent decline in DOE CRADAs.
DOE had the largest share of CRADAs through the mid-1990s, driving
the overall agency count to its FY 1996 peak, when CRADAs began
their declining trend. Smaller increases in DOD CRADAs sustained
the overall trend from further declines to FY 2000. Compared with
other forms of technology transfer activities, cooperative research
activities, both CRADAs and non-CRADA joint R&D projects, involve
a number of additional managerial and organizational requirements
for both agency and company participants. For agencies, an additional
factor is the R&D or administrative budget devoted to technology
transfer planning and management (U.S.
Small Business S&T Programs
The Small Business Innovation Research (SBIR) program, created
in 1982 (Public Law 97-219), leverages existing Federal R&D
funding toward small companies (those with 500 or fewer employees).
Although larger firms dominate R&D performance in the United
States, as discussed earlier in this chapter, small firms may have
capabilities or incentives to innovate, which may or may not come
to fruition due to a number of constraints, including financing.
SBIR's sister program, the Small Business Technology Transfer Program
(STTR), was created in 1992 to stimulate cooperative R&D and
technology transfer involving small businesses and nonprofit organizations,
including universities and FFRDCs. Both programs leverage existing
Federal R&D funding to small-company and nonprofit performers
to stimulate innovation, technology transfer, and R&D commercialization.
SBIR and STTR are administered by participating agencies and coordinated
by the Small Business Administration.
In SBIR, Federal agencies with extramural R&D obligations exceeding
$100 million must set aside a fixed percentage of such obligations
for SBIR projects. This set-aside has been 2.5 percent since FY
1997. To obtain this Federal funding, a small company applies for
a Phase I SBIR grant of up to $100,000 for up to 6 months to assess
the scientific and technical feasibility of ideas with commercial
potential. If the concept shows further potential, the company can
receive a Phase II grant of up to $750,000 over a period of up to
2 years for further development. In Phase III, the innovation must
be brought to market with private-sector investment and support;
no SBIR funds may be used for Phase III activities.
SBIR awarded about $12 billion to 64,300 projects through FY 2001.
Projects included research and commercialization activities in the
areas of computers, information processing and electronics, materials,
energy, environmental protection, and life sciences. In FY 2001
the program awarded $1.29 billion in R&D funding ($1.18 billion
in 1996 dollars) to 4,748 projects (figure
In FY 2001, DOD led the 10 participating agencies in SBIR funding,
obligating $576 million (45 percent of total SBIR funding), followed
by HHS at $412 million (32 percent) in FY 2001 (appendix
table 4-39 ).
STTR involves cooperative R&D performed jointly by a small
business and a research organization and is also structured in three
phases. The participating research organization must be a nonprofit
institution, as defined by the Stevenson-Wydler Technology Innovation
Act of 1980, or an FFRDC. Five Federal agencies with extramural
R&D budgets exceeding $1 billion participate in the program:
DOD, NSF, DOE, NASA, and HHS. The required set-aside has been 0.15
percent from FY 1996 to FY 2003, compared with 2.5 percent for SBIR.
STTR awarded about $460 million to more than 2,400 projects from
FY 1994 to FY 2001, including $71.3 million ($65.1 million in 1996
dollars) to 337 projects in 2001. DOD and HHS are the largest agency
table 4-40 ).
The Advanced Technology Program
The Advanced Technology Program (ATP), sponsored by DOC's National
Institute of Standards and Technology (NIST), was established by
the Omnibus Trade and Competitiveness Act of 1988 (Public Law 100-418;
15 USC, Section 278n) to promote the development and commercialization
of generic or broad-based technologies. The program provides funding
for high-risk R&D projects through a competitive process on
a cost-share basis with private-company participants.
From ATP's inception through FY 2002 more than 1,300 companies,
nonprofit institutions, and universities participated in 642 projects
costing $3.8 billion, which were funded about equally by ATP and
table 4-41 ).
Over the same period, 447 projects (70 percent) were single-company
projects and 195 (30 percent) were joint ventures; two-thirds of
participants were members of joint ventures. Participants pursued
projects in five technology areas: biotechnology, electronics, IT,
advanced materials and chemistry, and manufacturing.
In FY 2002, 61 R&D projects costing $289 million were initiated,
with about 54 percent funded by ATP and the balance funded by participants.
Public Law 108-7 appropriated $180 million for the program for FY
2003, a decline of 2.4 percent from FY 2002 (Schacht
2003). At the time of this writing, the Bush administration's
FY 2004 budget calls for the suspension of new awards and requests
funding only for administrative and close-out expenses (U.S.
Domestic and International Technology Alliances
Over the past 2 decades, U.S. firms have not only turned to technology
outsourcing but also increased their participation in technology
alliances domestically and globally. Technology alliances can
be defined as collaborative relationships or partnerships among
legally distinct parties that involve joint R&D or technology
Technology alliances allow firms to share R&D costs, pool technical
and market risks, and complement and further develop internal capabilities
(Vonortas 1997). Collaborative
networks are not without risks, however. Unintended transfer of
proprietary technology is always a concern for businesses. Cultural
differences among different industries, public partners (government
or academic), or international partners present additional difficulties
in managing alliances. Lastly, public-private collaboration presents
challenges for intellectual property policy and concerns for the
free flow of basic scientific knowledge.
Types of Technology Alliances
Technology alliances can be classified and analyzed according to
several criteria (Hagedoorn, Link,
and Vonortas 2000). In terms of their organizational structure,
they can be classified as equity alliances, or research joint
ventures (RJVs), in which two or more partners form a separate business
entity with long-term objectives. In contrast, nonequity alliances
are mostly contractual agreements governing short-term projects.
By membership profile, they may be private-private alliances (involving
only business partners such as suppliers, customers, or competitors)
or public-private alliances (involving government laboratories and
Technology alliances may focus on a number of innovation-related
activities, ranging from industry-wide issues such as basic or precompetitive
research, standards settings, or regulatory issues (Tassey
1997) to firm-specific projects. They can also range from longer
term learning and capabilities-building activities to shorter term
development projects closer to commercialization goals. These varied
goals, together with firm-specific characteristics (e.g., size,
age, internal organization, and R&D capabilities) and the underlying
technology and market characteristics, affect the choice of partners
and the organizational structure of these alliances.
Dedicated databases tracking these developments and sponsored in
part by NSF include the Cooperative Research (CORE) database, housed
at the University of North Carolina at Greensboro, and the Cooperative
Agreements and Technology Indicators database, compiled by the Maastricht
Economic Research Institute on Innovation and Technology (CATI-MERIT).
The CORE database covers U.S.-based alliances and RJVs recorded
in the Federal Register, pursuant to the provisions of the
National Cooperative Research Act, as amended.
Trends in the CORE database are illustrative only, because the registry
is not intended to be a comprehensive count of cooperative activity
by U.S.-based firms. The CATI-MERIT database covers international
technology agreements and is based on announcements of alliances
and tabulated according to the country of ownership of the parent
Domestic Research Partnerships
A total of 861 technology alliances were registered in the CORE
database from 1985 to 2001. The database shows the following trends:
- In 2001 there were 26 new technology alliances, compared with
45 in 2000. New filings increased between 1986 and 1995, when
they peaked at 115 (figure
Brod and Link (2001) developed
a statistical model to explain the trends in RJV filings, including
the decline after 1995. They found that filings are likely to
be countercyclical. In particular, they argue that "[w]hen the
economy is strong and...R&D is growing, firms may rely less
on cooperative research arrangements...than when the economy is
weak and internal resources are more constrained" (p. 109).
- About half of the technology alliances in 19852001 involved
activities classified in three industrial areas: electronic and
electrical equipment (18 percent), communication services (16
percent), and transportation equipment (15 percent).
- Fifteen percent (125 of 861) of these alliances involved a
U.S. university as a research member, whereas about 12 percent
(99 of 861) included a Federal laboratory.
International Technology Alliances
The data from the CATI-MERIT database are annual counts of new
technology alliances formed by domestic and multinational corporations
and their subsidiaries or affiliates worldwide. Most of the alliances
recorded in the database were owned by, and/or had R&D partners
located in, the United States, Western Europe, and Japan, the so-called
From 1991 to 2001, there were 5,892 new technology alliances formed
worldwide in six major sectors: information technology (IT), biotechnology,
advanced materials, aerospace and defense, automotive, and (nonbiotech)
chemicals. This total includes 602 alliances formed in 2001, a 25
percent increase from 483 in 2000 (figure
This is the first increase since a 19.5 percent increase in 1995
to its all-time high of 674 technology alliances.
The majority of these alliances were organized as nonequity, or
contractual, agreements (figure
In particular, the share of nonequity alliances increased from 61
percent in 198090 to 86 percent in 19912001. The more
flexible and project-based organization of nonequity agreements
favors activities in highly dynamic high-technology sectors such
as IT and biotechnology research and product development, as opposed
to more mature technology sectors (Hagedoorn
2001). Indeed, these two sectors are the top technology sectors
of these alliances.
The participation by U.S.-owned companies and their subsidiaries
is considerable. About 80 percent (4,646 of 5,892) of the 19912001
technology alliances worldwide involved at least one U.S.-owned
company (table 4-16 ),
up from two-thirds between 1980 and 1990. About half of the U.S.
alliances between 1991 and 2001 (or 39 percent of the all countries
total) were alliances exclusively among U.S.-owned companies. Thirty-four
percent of the U.S. alliances (27 percent of the total) were formed
between U.S.- and European-owned companies. European companies participated
in 2,604 (44 percent of 5,892) technology alliances during the period
19912001, up from 1,989 alliances in 19801990. However,
contrary to the pattern for U.S. companies, the majority of European
technology alliances were between U.S.- and European-owned companies,
as opposed to alliances exclusively among European-owned companies
and subsidiaries. Japanese companies participated in 779 technology
alliances worldwide between 1991 and 2001, down from 1,013 alliances
between 1980 and 1990, according to the CATI-MERIT database.
IT was the major focus among most ownership categories shown in
during 19902001. Notably, 46 percent of the alliances owned
exclusively by U.S. companies in 19912001 were focused on
IT activities. In contrast, the most frequent technology activity
of U.S.-European alliances was biotechnology at 33 percent (table
(The IT share for U.S.-European alliances was the second largest
at 21 percent.) Indeed, biotechnology alliances began to outpace
IT alliances in 2000 (figure
driven by intense activity in this sector by U.S. and European companies
(van Beuzekom 2001). In 1995
a new breed of alliance combining IT and biotechnology activities
emerged in the database. From 1995 to 2001, a total of 46 alliances
performed activities in areas such as bioinformatics applications.
U.S. companies participated in 37 (80 percent) of these alliances,
including 19 with European firms.