Technology Transfer Systems in the United States and Germany: Lessons and Perspectives (1997)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

62 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY INTRODUCTION The collective capacity of the United States to deploy technology and techni- cal know-how constitutes the nation’s technology transfer enterprise. The enter- prise involves all of the individuals, public- and private-sector institutions, and other resources (financial and physical capital) involved in the movement of tech- nology within and among organizations operating in the United States. As in other market economies, most of the resources and operational intelligence of this enterprise resides in private companies and is organized and driven by the logic of markets. In 1995, industry performed over 70 percent of all U.S. R&D and employed more than 90 percent of all U.S. scientists and engineers. Simi- larly, the volume of technology transfer that takes place within and between pri- vate firms dwarfs that which takes place between industry and all other sectors of the R&D enterprise combined.1 Indeed, the annual patent royalty income of just one large U.S. high-tech company such as IBM is greater than that of all nonin- dustrial sectors together. Nevertheless, the structure, goals, and performance of the U.S. technology enterprise are profoundly shaped by the contributions of a spectrum of nonindustrial R&D performers that are not themselves directly en- gaged in the commercialization of technology. The specific focus of this report is on the institutions and mechanisms in- volved in the transfer of technology from nonindustrial R&D performers to pri- vate firms, which then use this technology to create new products and services. These institutions include nonindustrial R&D performers: universities and affili- ated institutions, federal laboratories, and an array of public, private, and mixed (public/private) contract R&D institutes and consortia. Also implicated are a diverse group of organizations that perform little, if any, R&D of their own, yet play an important role facilitating technology transfer between the nonindustrial R&D performers and private industry. THE R&D ENTERPRISE A major distinguishing feature of the U.S. R&D enterprise is its colossal size. In 1994, the United States spent roughly $169 billion, or 2.5 percent of its gross domestic product (GDP), on research and development. Calculated in con- stant 1987 dollars, this sum equaled the combined R&D expenditures of Japan, Germany, France, and the United Kingdom (Figure 2.1). As of 1993, there were roughly 963,000 scientists and engineers engaged in R&D work in roughly 41,000 U.S.-based companies, 720 federal laboratories, 875 colleges and universities, and upwards of 2,300 other nonprofit R&D-performing organizations (e.g., re- search institutes, hospitals, consortia, etc.) (National Science Board, 1996; Na- tional Science Foundation, 1996c). As a percentage of GDP, R&D spending in the United States compared fa- vorably with that of most of its major trading partners in 1994 (Figure 2.2). How- 

TECHNOLOGY TRANSFER IN THE UNITED STATES 63 140 United States 120 Japan Billions of constant 1987 dollars Germany 100 France 80 United Kingdom Italy 60 Canada 40 20 0 FIGURE 2.1 International total R&D expenditures, 1994. SOURCE: National Science Foundation (1996b). ever, relative U.S. investments in R&D, estimated at 2.4 percent of GDP in 1995, have been declining since 1991, as have those of Germany and Japan. Moreover, international comparisons of the civilian R&D intensity of national economies (nondefense R&D as a percentage of GDP) reveal a persistent gap between the United States and other major industrialized countries. R&D Funders and Performers For statistical purposes, the U.S. R&D enterprise is divided into four major sectors: (1) government (federal, state, and local), (2) private industry, (3) non- profit colleges and universities, and (4) other private nonprofit R&D funders or performers. GOVERNMENT Prior to 1980, the federal government was the leading source of R&D funds, accounting for as much as 66 percent of the nation’s R&D spending in the early 1960s. During the past decade, however, with the end of the Cold War and declining defense budgets, the federal government’s share has declined rapidly, amounting to only 35.5 percent of the total, about $61 billion, in 1995 (Figure 2.3). Less than one-third of federal R&D funds ($16.7 billion) were used to 

64 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY Total R&D/GDP 4.0 3.5 3.0 Germany United States Japan 2.5 Percent 2.0 United Kingdom France 1.5 Canada Italy 1.0 0.5 0 1981 1983 1985 1987 1989 1991 1993 Nondefense R&D/GDP 4.0 3.5 3.0 Germany Japan 2.5 Percent United States 2.0 United Kingdom 1.5 France Canada 1.0 Italy 0.5 0 1981 1983 1985 1987 1989 1991 1993 FIGURE 2.2 Total and nondefense R&D spending as a percentage of GDP, by country. SOURCE: National Science Board (1996). 

TECHNOLOGY TRANSFER IN THE UNITED STATES 65 Performing sector 175 Billions of current dollars 150 Total R&D 125 Industry 100 75 50 Other nonfederal 25 Federal 0 1970 1975 1980 1985 1990 1995 Source of funds 175 150 Total R&D Billions of current dollars 125 100 Industry 75 Federal 50 25 Other nonfederal 0 1970 1975 1980 1985 1990 1995 FIGURE 2.3 National R&D expenditures, by performing sector and sources of funds. SOURCE: National Science Board (1996). support intramural R&D (i.e., R&D performed by the roughly 750 federal agency- operated research laboratories in 1995). The remaining two-thirds of the federal R&D budget supported R&D performed by private industry ($20.3 billion), uni- versities and colleges ($13 billion), a collection of industry- and university-ad- ministered Federally Funded Research and Development Centers (FFRDCs) ($8 billion),2 and other nonprofit institutions ($2.7 billion) (Table 2.1). The federal government funded roughly 58 percent of all U.S. basic research, 36 percent of all applied research, and 29 percent of all development work in 

66 TABLE 2.1 U.S. Expenditures, by Performing Sector and Source of Funds, 1995 Universities Other Percent Federal and nonprofit distribution, Performing sector Total Industry Government collegesa institutions performers Millions of dollars Total 171,000 101,650 60,700 5,500 3,150 100.0 Industry 119,600 99,300 20,300 – – 69.9 Industry-administered FFRDCsb 1,800 – 1,800 – – 1.1 Federal government 16,700 – 16,700 – – 9.8 Universities and colleges 21,600 1,500 13,000 5,500 1,600 12.6 University-administered FFRDCsb 5,300 – 5,300 – – 3.1 Other nonprofit institutions 5,100 850 2,700 – 1,550 3.0 Nonprofit-administered FFRDCsb 900 – 900 – – 0.5 Percent distribution, sources 100.0 59.4 35.5 3.2 1.8 FFRDC = federally funded research and development center; – = unknown, but assumed to be negligible. NOTE: Data are estimated. aIncludes an estimated $1.6 billion in state and local government funds provided to university and college performers. bFFRDCs conduct R&D almost exclusively for use by the federal government. Expenditures for FFRDCs, therefore, are included in federal R&D support, although some nonfederal R&D support may be included. SOURCE: National Science Board (1996). TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY 

TECHNOLOGY TRANSFER IN THE UNITED STATES 67 1995 (Figure 2.4).3 That year, federal government laboratories performed 9.1 percent of all U.S. basic research ($2.7 billion), 12.3 percent of applied research ($4.9 billion), and 9 percent of development ($9.1 billion). In 1993 the federal government employed over 60,000 scientists and engineers in R&D activity (Na- tional Science Foundation, 1996b). For the most part, national R&D statistics shed little light on the volume and character of R&D funding and performance by U.S. state and local governments. Nonfederal government entities collectively funded roughly 2 percent of all R&D performed in the United States in 1993 (National Science Board, 1996). Most state and local R&D monies are used to support applied research at doctorate- granting state universities. These funds come either directly in the form of re- search grants and contracts or indirectly in the form of general-purpose funds that end up being used for research by the recipient academic institutions. Collec- tively, state and local governments funded between 12 and 17 percent of U.S. academic research in 1995, 7.4 percent (or $1.6 billion) through research con- tracts and grants and an additional 5 to 10 percent ($0.9 to $2.0 billion) through general purpose funds (National Science Board, 1996). INDUSTRY Since 1980, industry has been both the primary source of R&D funds and the largest R&D performer in the United States, financing 59.4 percent ($101.7 bil- lion) and performing roughly 71 percent ($121.4 billion) of all U.S. R&D in 1995 (Figure 2.3). Industry performs the overwhelming majority of the research that it funds, $99.3 billion in 1995, with the remainder, $2.4 billion, going to support research in colleges and universities and other nonprofit research institutions. Industry performed an additional $20.3 billion worth of R&D supported by fed- eral funds in 1995; most of this was defense-related development work financed by the Department of Defense (DOD).4 In addition to R&D performed directly by private firms, federal agencies also funded about $1.8 billion of R&D at indus- try-administered FFRDCs that year. In 1995, industry funded 70.4 percent of all development work, 56.8 percent of all applied research, and 25.3 percent of all basic research performed in the United States. In turn, industry performed about 86 percent ($87.6 billion) of all development work, 67 percent ($26.7 billion) of all applied research, and 24.2 percent of basic research ($7.2 billion) that year. Roughly 764,500 scientists and engineers were engaged in R&D in U.S. industry in 1993. COLLEGES AND UNIVERSITIES Colleges and universities (both private and public) performed 12.6 percent, or $21.6 billion worth, of R&D in 1995. That year, university-administered FFRDCs performed an additional $5.3 billion of R&D. Institutions of higher 

68 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY Performing sector 100 Federal Government Universities and colleges Industry Other nonprofit 80 60 Percent 40 20 0 Development Applied research Basic research Character of work Source of funds 100 Federal Government Universities and colleges 80 Industry Other nonprofit 60 Percent 40 20 0 Development Applied research Basic research Character of work FIGURE 2.4 National R&D expenditures, by performing sector, source of funds, and character of work, 1995. SOURCE: National Science Board (1996). 

TECHNOLOGY TRANSFER IN THE UNITED STATES 69 education are the primary locus of basic research in the United States, accounting for roughly 49 percent of all basic research performed in 1995. Colleges and universities performed about 14 percent of all applied research and only 1.6 per- cent of total U.S. development work that year. Although 875 institutions of higher education reported performing R&D in 1995, the 100 largest of these accounted for 80 percent of all academic research conducted in the United States (National Science Board, 1996). The federal government has long been the primary source of academic R&D dollars (Table 2.2). Although the federal share has fallen significantly since the early 1970s, when it accounted for more than 70 percent of academic research funds, federal agencies still financed over 60 percent of all academic research in 1995. Private universities, which represent an important, highly productive part of the nation’s academic research enterprise, depend much more heavily on fed- eral R&D funds than do public universities, which receive both targeted research funding and general-purpose appropriations from state governments. In 1993, federal agencies funded roughly 56 percent of all research at public universities and 74 percent of research performed at U.S. private universities (National Sci- ence Board, 1996). The second largest source of research funding for colleges and universities is their own institutional funds. This collection of general pur- pose state or local government appropriations, general purpose grants from out- side sources, tuitions and fees, endowment income, and unrestricted gifts totaled roughly $3.9 billion in 1995. The share of academic research supported by insti- tutional funds increased from 13.8 percent in 1980 to 18.1 percent in 1995. In addition to their indirect investment in academic research through general pur- pose appropriations to colleges and universities, state and local government di- rectly funded 7.4 percent of U.S. academic research in 1995. Other nonprofit institutions funded an additional 7.4 percent of the total. Industry accounted for the smallest share (6.9 percent) of academic research support in 1995. However, since the mid-1980s, industry support has increased more rapidly than any other source of academic R&D funding. TABLE 2.2 Support for U.S. Academic R&D, Percent Shares by Sector 1970 1980 1990 1995 (est.) Federal government 70.5 67.5 59.2 60.2 State and local government 9.4 8.2 8.1 7.4 Industry 2.6 3.9 6.9 6.9 Academic institutions 10.4 13.8 18.5 18.1 All other sources 7.1 6.6 7.3 7.4 TOTAL 100.0 100.0 100.0 100.0 SOURCE: National Science Board (1996). 

70 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY 1n 1993, universities and colleges employed nearly 150,000 doctoral scien- tists and engineers, 10,500 individuals with professional degrees, 5,500 scien- tists and engineers with degrees at the master’s and bachelor’s level, and roughly 90,000 graduate students in R&D activity (National Science Board, 1996). OTHER NONPROFITS The least well-documented and well-measured sector of the U.S. R&D enter- prise is that comprising a diverse population of “other nonprofit” R&D funders and performers. Led by private nonprofit foundations, such as the Howard Hughes Medical Institute, “other nonprofit” organizations funded 1.8 percent of total U.S. R&D in 1995 (National Science Board, 1996). That year, these organi- zations funded 5.5 percent of all basic research, 2.6 percent of all applied re- search, and less than 0.5 percent of development work in the United States. In 1995, private nonprofit foundations, independent R&D institutes, private research hospitals, independent medical research centers, consortia, and their affiliates (more than 2,300 institutions altogether) performed about 3.5 percent of all U.S. R&D. In 1995, other nonprofits conducted 7.5 percent of basic research ($2.2 billion), 4.8 percent of applied research ($1.9 billion), and less than 2 percent of development ($1.9 billion). Other nonprofit research institutions are particularly prevalent in medical- and health-related research. In 1994, more than 45 percent of all R&D funded by other nonprofit institutions was in the area of health, as was nearly 42 percent of R&D performed by other nonprofit institutions. Other non- profit institutions employed 10,200 scientists and engineers in R&D activities in 1993 (National Science Foundation, 1996b). Distribution of Publicly Funded R&D Since the 1940s, the federal government has focused its support of the nation’s technology enterprise on mobilizing technical resources to further spe- cific national missions. These missions, championed by various federal agencies, have included national security, the cure of disease, space exploration, food pro- duction, and world leadership in basic science. National economic development and international competitiveness have rarely been explicit objectives of federal technology policies and investments. THE DEFENSE IMPERATIVE A defining feature of the U.S. government’s R&D portfolio has long been its heavy commitment to the needs of national security. In 1955, during the height of the Cold War, defense-related R&D claimed over 85 percent of all federal R&D dollars. During the 1980s, national security accounted for nearly two-thirds of federal R&D spending and one-third of total national (public and private) R&D 

TECHNOLOGY TRANSFER IN THE UNITED STATES 71 expenditures. In spite of a significant decline in defense spending during the past 5 years, defense-related R&D still accounted for 55 percent of federal R&D spending, or roughly one-quarter of all R&D spending in 1995 (Table 2.3).5 Al- though federally funded R&D as a share of total industrial R&D has declined rapidly since the late 1980s, from 33 percent in 1988 to 17 percent in 1995, DOD remains the source of over 80 percent of all federal R&D dollars spent by private industry. During the past 4 decades, defense-related R&D and procurement have fos- tered the development of important “dual-use” technologies (technologies having both civilian and defense applications) and provided a powerful stimulus to inno- vation in a select number of high-tech civilian industries such as microelectron- ics, software, and aerospace.6 As of 1994, federal R&D dollars (predominantly DOD funds) still accounted for 61 percent of industrial R&D in the aerospace sector (Table 2.3). Nevertheless, the overwhelming majority of this defense- related R&D (an estimated 90 percent as of the early 1990s) has been for the TABLE 2.3 U.S. Defense-Related R&D, Various Comparisons 1955 1960 1970 1980 1990 1995 Share of federal R&D that is defense related 85 80 58 51 63 55 Share of total U.S. R&D that is defense related 48 52 33 24 26 23 Share of federal support of academic engineering research that is defense related * * 45a 55 44 45b Share of all government-funded R&D in U.S. industry that is defense relatedc * 81 68 63 83 80 Federal share of total R&D funds in aerospace industry 88d 89 77 72 76 61e Federal share of total R&D funds in electrical machinery and communications 66d 65 52 41 38 14e *Data not available. a1971 data. b1993–1995 average federal academic research obligations. cDepartment of Defense only, data for 1962, 1981, and 1989. d1957 data. e1994 data. SOURCES: National Science Foundation (1990; 1991; 1992a,b; 1994; 1996b). 

72 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY “development, testing, and evaluation” of weapons and other systems that have no markets other than the military7 (Alic et al., 1992). National security has also long been the focus of government support for engineering R&D in U.S. universities and government laboratories. Although DOD accounted for only 12.2 percent of federal funding for all fields of academic R&D in 1995, the agency remains a major funder of university-based engineer- ing research. As of 1994, DOD accounted for over 49 percent of all federal obligations for academic research in math, computer sciences, and all fields of engineering combined. This included 60 percent of federal funds for academic electronics and electrical engineering research, 54 percent for metallurgy and materials research, 52 percent for aerospace engineering research, 41 percent for mechanical, 47 percent for civil, and 4 percent for chemical engineering research8 (National Science Foundation, 1997). Finally, the demands of national defense have largely determined the struc- ture and objectives of the government’s system of federal laboratories, particu- larly in the physical sciences and engineering research. In 1995, DOD accounted for nearly half of all obligated expenditures of federal laboratories and, as of 1993, employed more than half of all federal laboratory R&D scientists and engi- neers9 (National Science Board, 1996; National Science Foundation, 1995a). GOVERNMENT CIVILIAN R&D PRIORITIES Between 1987 and 1994, the share of federal R&D funds dedicated to civil- ian or nondefense-related agency missions increased from 31 percent to 45 per- cent (Table 2.4). In 1994, over 60 percent of the federal civilian R&D portfolio was allocated to the missions of health and civilian space exploration. The shares of federal civilian R&D funds dedicated to the missions of health, energy, the “advancement of research,” and agriculture all declined slightly between 1987 and 1994. These declines were offset by increases in the shares allocated for research related to civilian space, infrastructure, environmental protection, and industrial development. Of these four mission areas, industrial development R&D has grown most rapidly since the late 1980s, albeit from a very small base. More than two-thirds of federal civilian R&D funds went for basic and ap- plied research in 1995. In contrast, 90 percent of federal defense-related R&D went for exploratory development (Figure 2.5). The vast majority of federal support for basic research flows from a few civilian agencies. The Department of Health and Human Services (DHHS), more specifically its National Institutes of Health, is overwhelmingly the largest funder of basic research—DHHS obliga- tions in 1995 were $6.3 billion, three or more times those of the National Science Foundation (NSF) ($2.0 billion), NASA ($1.8 billion), and the Department of Energy ($1.7 billion). By way of comparison, DOD’s obligations for basic re- search were $1.2 billion in 1995 (National Science Board, 1996). Likewise, that year, civilian agencies accounted for over 78 percent of all federal obligations for 

TECHNOLOGY TRANSFER IN THE UNITED STATES 73 TABLE 2.4 Distribution of Government R&D Appropriations by Socioeconomic Objective in the United States, 1987 and 1994 Percent of Public R&D Funds Total Funds Civilian Funds Percent change in share Objective 1987 1994 1987 1994 1987–1994 Agriculture 2.3 2.5 7.3 5.6 –23 Industrial development 0.2 0.6 .6 1.3 117 Energy 3.6 4.2 11.5 9.4 –18 Infrastructure 1.8 2.9 5.7 6.5 14 Environmental protection 0.5 0.8 1.6 1.8 13 Health 11.9 16.5 37.9 36.9 –3 Civilian space 6.0 10.9 19.1 24.4 28 Defense 68.6 55.3 — — — Advancement of research 3.6 4.0 11.5 8.9 –23 General university fundsa — — — — — Not elsewhere classified — 2.3 5.1 SOURCES: National Science Board (1989, 1996). applied research. Nearly half of all federal obligations for basic research and a third of those for applied research went to support research in the life sciences (biological, agricultural and medical sciences) in 1994 (Figure 2.6). Until recently, direct support by states of applied research projects (predomi- nantly at academic research institutions) appears to have been concentrated in a relatively small number of fields or mission areas, including health, agriculture, and transportation. Although data regarding the distribution of state and local government R&D funds are fragmentary, it is estimated that between 60 and 75 percent of all of research supported with nonfederal government dollars in 1994 was health related. During the past decade, however, some states have broadened their R&D portfolios to support industrial and economic development more ex- plicitly and aggressively (Coburn, 1995). NEW FEDERAL INDUSTRIAL R&D INITIATIVES The small claim of industrial development on the federal R&D budget testi- fies to the weak commitment of the federal government to economic develop- ment as an explicit mission of public R&D and technology policy. Indeed, the federal government and the private sector have long maintained a stark division of roles with regard to the funding of research versus the funding of development and deployment of technology for most sectors of the nation’s economy.10 Basic research and the development and application of technology relevant to accepted 

74 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY Other 5% NSF 3% R&D plant R&D plant NASA 12% Development Commerce 2% Development Applied research HHS 17% Basic Agriculture 2% research DOD 50% Applied research DOE 9% Basic research Total R&D and R&D plant ($69.4 billion) FIGURE 2.5 Federal obligations, by agency and type of activity, 1995. SOURCE: National Science Board (1996). federal agency missions (though conducted principally by private-sector actors) have been regarded as legitimate activities for funding by the public sector. The identification, development, and adoption of technology for commercial products and services not directly associated with public missions has been seen as the preserve of the private sector. Until relatively recently, the only notable excep- tion to this division of labor was the technical support (standards, testing, and evaluation) provided by the Department of Commerce’s National Institute of Stan- dards and Technology (NIST, formerly the National Bureau of Standards). During the 1970s and 1980s, growing concerns regarding the health and com- petitive performance of the U.S. commercial technology enterprise prepared the way for a number of new initiatives by the federal government that would engage it explicitly, however tentatively, in support of civilian technology for national economic development. A series of laws were passed to promote government- industry partnerships and to foster technology transfer and collaborative R&D between and within sectors of the nation’s technology enterprise. The 1980 Patent and Trademark Amendments (P.L. 96-517), known as the Bayh-Dole Act, per- mitted recipients of federal grants and contracts to retain title to inventions devel- oped with government funds. Bayh-Dole provided a major impetus for universi- ties and colleges in particular to get into the business of patenting and licensing technologies developed on their campuses.11 Similarly, the Stevenson-Wydler Technology Innovation Act of 1980 (P.L. 96-480) and subsequent amendments to it during the ensuing decade were directed at engaging federal laboratories more extensively in the transfer of technologies to private firms as well as foster- ing cooperative research among federal laboratories, state and local governments, universities, and private firms. The Federal Technology Transfer Act of 1986 

TECHNOLOGY TRANSFER IN THE UNITED STATES 75 Basic research 6 Life sciences 5 Billions of constant 1987 dollars 4 3 Physical sciences 2 Engineering 1 Environmental sciences Mathematics and computer sciences 0 1980 1985 1990 1995 Applied research 6 5 Billions of constant 1987 dollars 4 Life sciences Engineering 3 2 Physical sciences 1 Environmental sciences Mathematics and computer sciences 0 1980 1985 1990 1995 FIGURE 2.6 Federal obligations for basic and applied research, by field. SOURCE: National Science Board (1996). 

76 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY (P.L. 99-502) amended Stevenson-Wydler to authorize cooperative research and development agreements (CRADAs) between federal laboratories and other enti- ties, including state governments.12 In 1984, Congress passed the National Co- operative Research Act (P.L.98-462), which fostered the proliferation of indus- trial R&D consortia and joint ventures by removing the threat of treble damages under U.S. antitrust law for firms that filed with the Department of Justice infor- mation concerning their involvement in such activities.13 In addition to these changes in law, Congress funded a range of programs designed to foster industrial technology development and technology transfer during the 1980s and early 1990s. Coburn (1995) identifies five basic types of federal and state cooperative technology programs that were established during the past 10 to 15 years. These include programs directed at technology develop- ment (i.e., those that support the development and application of new or enhanced industrial products or processes); programs focused on industrial problem solv- ing, particularly for small business, through the diffusion of technology and best- practice applications; technology financing programs involving public capital or facilitated access to private capital; start-up assistance, primarily through public support of technology incubators and research parks; and teaming, or assistance in forming partnerships. Major federal technology development programs include the National Sci- ence Foundation–sponsored portfolio of university-industry research centers— Industry/University Cooperative Research Centers (begun in 1973), Engineering Research Centers (1985), Science and Technology Centers (1987), Materials Re- search Science and Engineering Centers (1993), Supercomputer Centers (1986)— each designed to serve different objectives yet sharing a commitment to facilitate university-industry research cooperation and technology transfer.14 Another tech- nology development initiative is NIST’s Advanced Technology Program (ATP), established in 1989 to fund businesses, especially SMEs, in the research and de- velopment of generic, precompetitive technologies to foster high-risk, high-po- tential products, processes, and technologies. ATP’s budget reached a high of $340.5 million in 1995, but has since declined to $225 million in fiscal 1997. With the advent of a Republican-controlled Congress following the election of 1994, ATP has been under constant threat of elimination. Other technology de- velopment programs include DOD’s Manufacturing Technology Program and SEMATECH (now totally privately funded), the Department of Transportation’s Intelligent Vehicle Highway Systems and Maglev programs, and multiagency initiatives such as the Technology Reinvestment Project and the Small Business Technology Transfer Program.15 Primary federal industrial problem-solving initiatives include the Department of Agriculture’s (USDA) long-standing agricultural extension service, set up in the 1914 to diffuse results of USDA research and modern farming technology and methods, and NIST’s Manufacturing Extension Partnership (MEP). MEP was conceived initially in 1988/89 as a system of manufacturing technology cen- 

TECHNOLOGY TRANSFER IN THE UNITED STATES 77 ters designed to transfer advanced production technology from NIST’s research facilities and other federal laboratories to small and medium-sized enterprises (SMEs). By the early 1990s, however, the focus of the program shifted to the provision of technical extension/industrial modernization services for SMEs through what would become a nationwide network of extension centers and agents. MEP centers assist SMEs with adoption of improved manufacturing tech- nologies, training, management, and networking. MEP’s budget in fiscal 1997 was $95 million. (For further information on MEP, see Annex II, pp. 207–209.) Although no federal programs provide direct general-purpose financing of technology-based companies, there are several federal grant programs that help finance the development and commercial application of technologies relevant to federal agency missions. Most notably, in 1982, the Small Business Innovation Research (SBIR) program was created to direct a small share (initially not less than 1.25 percent, now 2.5 percent) of each major mission agency’s total annual R&D budget to fund R&D at small and medium-sized firms and to stimulate the commercialization of new products and services (National Science Board, 1996). Other programs that help finance the commercialization of technology by private companies include NASA’s Aerospace Industry Technology Program, and the Environmental Protection Agency’s Environmental Technology Initiative. Four federal agencies, the Department of Commerce, DOD, the Department of Labor, and NASA, sponsor “teaming or network-building” programs that pro- vide assistance to industry through information dissemination, networking, and databases. Start-up assistance in the form of technology incubators and research parks remains the exclusive preserve of state and local governments. Coburn (1995) estimates that the federal investment for all cooperative tech- nology programs grew from $1.7 billion in fiscal 1992 to $2.7 billion in fiscal 1994. Forty percent of federal spending on cooperative technology programs in fiscal 1994 was for technology development initiatives, 28 percent for technol- ogy financing, and 25 percent for industrial problem solving (Table 2.5). Although the proliferation of federal cooperative technology programs dur- ing the past decade has been impressive, collectively these programs amounted to less than 4 percent of the fiscal 1994 federal R&D budget. Furthermore, since the Congressional elections of 1994, several major federal cooperative technol- ogy programs have ended (e.g., TRP and DOD funding for SEMATECH16 ), or are on the verge of being eliminated (ATP) by a more skeptical Republican con- trolled Congress. In other words, at the federal level at least the role of govern- ment in direct financial support of industrially relevant civilian R&D and tech- nology transfer is still seeking potential legitimacy. STATE INDUSTRIAL TECHNOLOGY PROGRAMS In contrast with the federal government, state governments traditionally have had few political reservations about using public funds to actively promote indus- 

TABLE 2.5 Federal and State Government Investment in Cooperative Technology Activities, by Type of Program, 78 FY 1994 Industrial Technology Problem Technology Start-Up Federal Agency Development Solving Financinga Assistanceb Teaming $ millions (number of programs) Department of Agriculture 0.0 (0) 434.6 (1) 16.1 (2) 0.0 (0) 0.0 (0) Department of Commerce 199.5 (1) 30.2 (1) 3.7 (4) 0.0 (0) 0.0 (1)c Department of Defense (FY93) 374.5 (6) 205.7 (2) 346.0 (1) 0.0 (0) 48.2 (2) Department of Energy 1.7 (1) 9.0 (3) 61.3 (3) 0.0 (0) 0.0 (0) Department of Health and Human Services 4.1 (1) 0.0 (0) 129.0 (1) 0.0 (0) 0.0 (0) Department of Labor 0.0 (0) 1.4 (4) 0.0 (0) 0.0 (0) 0.2 (1) Department of Transportation 345.4 (4) 0.0 (0) 4.2 (1) 0.0 (0) 0.0 (0) Environmental Protection Agency 0.0 (0) 0.0 (0) 56.8 (4) 0.0 (0) 0.0 (0) National Aeronautics and Space Administration 3.6 (1) 8.1 (1) 149.2 (3) 0.0 (0) 7.7 (2) National Science Foundation 161.2 (6) 0.0 (0) 29.8 (2) 0.0 (0) 0.0 (0) Total federal support by type $1,090.0 (20) $689.0 (12) $761.1 (20) $0.0 (0) $56.1 (6) Total state support by type 127.5 (69) 59.5 (97) 101.8 (108) 7.2 (28) 5.9 (34) Grand Total, federal and state support by type $1,217.5 $748.5 $862.9 $7.2 $62.0 aDollaramounts for the ATP are included in the Technology Development totals. bWhilethere are no formal federal programs in this category, there is ad hoc activity in some agencies. cThe Department of Commerce teaming program is the Clearinghouse for State and Local Initiatives, which has received no funding so far. TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY SOURCE: Coburn (1995). 

TECHNOLOGY TRANSFER IN THE UNITED STATES 79 trial and economic growth within their borders. Most of the 50 states sponsor economic development programs, the earliest of which date from the late 18th century (Coburn, 1995). The use of internal technology resources to foster eco- nomic development was first promoted by North Carolina in the early 1960s, an effort that led to the creation of the Research Triangle Park complex. Other states have followed suit, and now all 50 states have technology-based development programs of one sort or another. Coburn (1995) estimates that state governments spent just over $384 million in fiscal 1994 on cooperative technology programs. Of this, approximately one- third was used for technology development, mostly matching support for univer- sity-industry technology centers funded primarily by federal initiatives (Box 1). One-quarter went to support technology financing, about 60 percent of which went to projects and 30 percent to companies. Fifteen percent of state coopera- tive technology funds were used for industrial problem solving, predominantly for technology extension and deployment programs (state initiatives as well as MEP matching funding). Twenty-one percent of state funds went to educational programs at institutions of higher education that sponsor the development, diffu- sion, and use of technology and improved practices to benefit specific companies. As of 1994, 42 states had some form of industrial problem solving17 and technol- ogy-financing program in place; 31 had technology development programs; 18 supported start-up assistance incubators or industrial technology parks; 21 funded teaming (Coburn, 1995). North Carolina, one of the pioneers, invests the most of any state in coopera- tive technology programs ($37 million in 1994), followed by Pennsylvania ($34 million), Texas ($30 million), Georgia ($30 million), Connecticut ($27 million), Ohio ($27 million), New York ($23 million), New Jersey ($20 million), Michi- gan ($14 million), and Maryland ($13 million). The highest per-capita invest- ments are made by Arkansas, Connecticut, Nebraska, North Carolina, and South Dakota (Coburn, 1995). The Industrial R&D Enterprise The U.S. industrial R&D enterprise is distinguished by its large size, both in terms of R&D volume and the number of firms involved; its dynamism as re- flected in the changing sectoral distribution of R&D activity over time; and its capacity for spawning new technology-based products and industries. In 1995, over 18,000 manufacturing and 23,000 nonmanufacturing compa- nies reported performing a total of $102 billion of R&D in the United States.18 Collectively, these firms employed 764,500 scientists and engineers in R&D ac- tivity in 1993 (National Science Foundation, 1996b). The vast majority of indus- trial R&D spending is concentrated in a small number of firms. In 1993, for example, the 20 largest R&D spending companies accounted for one-third of all industrial R&D expenditures; the 200 largest firms accounted for 71 percent 

80 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY BOX 1 State Spending, by Category, on Cooperative Technology Programs, 1994 ($000s) Technology Development (34 percent) University-Industry Technology Centers $104,606 University-Industry Research Partnerships 12,118 Government-Industry Consortia 4,810 Equipment and Facility Access Programs 5,965 Technology Financing (26 percent) Project Financing $62,172 Company Financing 30,861 Small Business Innovation Research 3,185 Technology Reinvestment Program/ Advanced Technology Program 5,593 Related Educational Initiatives (21 percent) and Other $82,635 Industry Problem Solving (15 percent) Technology Enterprise Divisions $54,851 Federal Technology Application Programs 3,805 IMPs 850 Start-Ups (2 percent) Incubators $7,238 Research Parks not available Teaming (2 percent) Networks $4,376 Databases 1,531 TOTAL $385,000 SOURCE: Coburn (1995). (National Science Board, 1996). In 1994, 19 companies each spent more than $1 billion on R&D, and another 49 spent more than $200 million. SHIFTING SECTORAL DISTRIBUTION OF R&D ACTIVITY There has been significant change in the sectoral distribution of U.S. indus- trial R&D in recent decades, reflecting changes in the composition of the nation’s 

TECHNOLOGY TRANSFER IN THE UNITED STATES 81 industrial base, changes in the relative R&D intensity of different industries over time, and changes in the way U.S. industrial R&D activity is measured (Figure 2.7). The most notable change over the past 10 years has been the rapid increase in the share of R&D claimed by nonmanufacturing (predominantly service) in- dustries. Until fairly recently, nonmanufacturing industries were believed to ac- count for less than 5 percent of all industrial R&D spending. Since the early 1980s, however, their share of the total has increased rapidly, from 5.1 percent in 1983 to 26.7 percent in 1993. Much of the increase in nonmanufacturing R&D over this period can be attributed to changes in NSF’s survey of industrial R&D in 1991, which changed and greatly expanded the sample of companies surveyed, thereby incorporating more accurate information on the R&D performance of smaller firms and firms classified in the nonmanufacturing sector. According to NSF, these changes resulted in an upward revision of total nonmanufacturing R&D in 1991 from roughly $10 billion previously reported to $21 billion. That year, an additional $7 billion of R&D was reclassified from manufacturing to nonmanufacturing categories. Much of this latter shift is believed to accurately Industrial Chemicals Pharmaceuticals 1993 Fabricated Metal Products 1983 Nonelectrical 1973 Machinery Motor Vehicles Aerospace Electrical Machinery and Apparatus Office Machinery and Computers Electronic and Communi- cation Equipment Instruments Other Manufacturing Nonmanufacturing 0 5 10 15 20 25 30 Percent FIGURE 2.7 U.S. industrial R&D spending, by sector, 1973, 1983, and 1993. SOURCE: Organization for Economic Cooperation and Development (1996a). 

82 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY reflect changes in the output mix of companies formerly classified in manufactur- ing industries (National Science Foundation, 1995b). Three industries account for the majority of nonmanufacturing R&D: com- puter programming and related services, including software (8.1 percent of total R&D in 1993); communications services (4.4 percent); and research, develop- ment, and testing services (1.5 percent) (National Science Board, 1996).19 Two manufacturing industries—pharmaceuticals and professional and scien- tific instruments—have also significantly increased their share of total industrial R&D expenditures during the past 2 decades. Notably, the growth of these indus- tries’ share of industrial R&D tracks the growth in the share of federal R&D dedicated to health-related research as well as the associated growth of the nonin- dustrial research base in the life and medical sciences during the period. (See Figure 2.6, p. 75.) Four manufacturing industries—aerospace, electronics and communications equipment, office machinery and computers, and electrical machinery—have seen their shares of total industrial R&D contract dramatically during the past 10 to 20 years. Historically, these industries, particularly aerospace, have been the benefi- ciaries of DOD R&D and procurement, which has declined dramatically during the past decade. As of 1988, the aerospace industry absorbed more than 60 per- cent of all federal R&D funds for industry. However, by 1994, industrial aero- space R&D amounted to about 39 percent of the total (National Science Founda- tion, 1996a). Moreover, data from the Aerospace Industry Association (1994) also indicate a 25-percent decrease in revenues from sales of military-related hard- ware from 1990 and 1993. CHANGES IN THE COMPOSITION AND ORGANIZATION OF INDUSTRIAL R&D Changes in the sectoral distribution of industrial R&D spending over the past 2 decades have been accompanied by compositional and organizational shifts. Relative Decline in Industrial Basic and Applied Research First, there has been a change in the character of industrial R&D (i.e., basic research, applied research, and development) since 1991. While the inflation- adjusted industrial R&D expenditures overall declined 5.9 percent between 1991 and 1995, industrial performance of basic and applied research declined more than did industrial exploratory development. Since 1991, industrial basic re- search as a share of total industrial research has declined from 6.7 percent to 5.9 percent,20 that of industrial applied research declined from 23.5 to 22.0 percent, while that of industrial development increased from 69.8 to 72.2 percent. These shifts are explained, in part, by the dismantling of several companies’ large cen- tral research facilities and a general movement in several industries away from long-term fundamental research toward more short-term applied research and 

TECHNOLOGY TRANSFER IN THE UNITED STATES 83 development in order to meet intensifying international competition (National Science Board, 1996). Accompanying this latter trend have been an increased emphasis on R&D as a tool for scanning for and exploiting knowledge generated or applied beyond national boundaries, as well as closer integration of R&D with activities farther downstream in the value-added process (i.e., changes designed to leverage scarce R&D dollars and speed commercialization of new technology) (National Academy of Engineering, 1993, 1996b). Increased Cooperative R&D and R&D Outsourcing Second, there has been an increase in both cooperative R&D and R&D outsourcing among firms as well as between firms and nonindustrial R&D per- formers during the past decade. This is in part explained by the rapid growth in the number of R&D consortia, joint ventures, and other forms of strategic alli- ances in R&D at the hands of U.S.-based companies during the past decade.21 Though by no means a measure of all U.S. R&D consortia and joint venture activity, the number of “joint research ventures” (JRVs) registered each year with the Department of Justice (DOJ) has grown significantly since passage of the 1984 National Cooperative Research Act (P.L. 98-462). As of 1995, more than 565 JRVs had been registered with the DOJ (Vonortas, 1996). Likewise, Hagedoorn (1995) has documented a marked increase in the level of U.S.-firm participation in international strategic technology alliances since the early 1980s (Figure 2.8). 500 1990–94 1985–89 Europe–United States 400 Japan–United States 1980–84 Europe–Japan 300 Number 200 100 0 Information technology Biotechnology New materials FIGURE 2.8 Number of new strategic technology alliances, by industry and region. SOURCE: National Science Board (1996). 

84 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY Between 1985 and 1995, industry funding of R&D at universities and col- leges (in inflation-adjusted dollars) nearly doubled, and industry research fund- ing at other nonprofit organizations grew nearly 65 percent. During the same period, company-financed R&D performed within industry grew less than 27 percent in constant 1987 dollars. Consistent with these trends has been a signifi- cant increase in the level of industrial involvement in collaborative research with academic researchers via university-industry research centers as well as rapid growth (albeit from a very small base) in the volume of technology licensed by industry from academic research institutions during the past decade (Association of University Technology Managers, 1996; Cohen et al., 1994).22 Similarly, there has been rapid growth in the number of CRADAs between companies and federal laboratories since the mid-1980s.23 Yet another indicator of the growth of research collaboration between indus- try and nonindustrial research institutions is the rapid increase in the share of scientific and technical articles that are coauthored by individuals in industry and researchers based at nonindustrial research institutions. Between 1981 and 1993, the share of scientific and technical articles that had industry-based authors grew from 27.3 percent to 47 percent (National Science Board, 1996). Most of this increase was accounted for by growth in the volume of academic-industry coau- thored literature. Internationalization of U.S. Industrial R&D Third, the past 2 decades have witnessed a growing internationalization of U.S. industrial R&D activity, predominantly at the hand of foreign direct invest- ment (multinational companies) and international strategic alliances (National Academy of Engineering, 1996b). Between 1985 and 1993, U.S.-owned compa- nies increased their investment in overseas R&D three times faster than their investment in U.S.-based R&D activity. As of 1994, these investments amounted to roughly 10 percent of all company-financed R&D in the United States. Even more pronounced has been the growth of foreign participation in the U.S. indus- trial R&D enterprise since the early 1980s (Figure 2.9). From 1984 to 1994, R&D spending by the U.S. affiliates of foreign-owned companies24 increased as a share of all company-financed U.S. R&D from 9 percent to nearly 16 percent. As of 1994, foreign-owned companies financed roughly 2 percent of all research conducted at U.S. universities and federal laboratories (National Academy of Engineering, 1996b). THE SPECIAL ROLE OF START-UP COMPANIES A unique feature of the U.S. industrial technology enterprise is the critical role start-up companies play in the transfer and commercialization of fast-mov- ing, science-based technologies. This happens generally via movement, or “spin- out,” of researchers and technology from universities, large established compa- 

TECHNOLOGY TRANSFER IN THE UNITED STATES 85 20 15.7 15 15 14.4 13.8 12.6 12.8 11.5 10.4 Percent 10 9.2 9.2 9.5 9.1 9.0 5 0 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 Year FIGURE 2.9 R&D spending by U.S. affiliates of foreign-owned firms as a percentage of all privately funded U.S. R&D, 1982–1994. SOURCES: National Science Board (1996) and U.S. Department of Commerce (1996a). nies, and government laboratories. U.S.-based high-tech start-ups are credited with commercializing the technologies that launched the new biotechnology and computer software industries. Although growth in the number of new U.S. high- tech companies established during the past decade is considerably slower than that from the mid-1970s to the mid-1980s, nearly half of all U.S. high-tech com- panies operating in 1994 were established during the past 15 years (Table 2.6) (National Science Board, 1996). More than one-quarter of all new businesses started since 1980 (and operating in 1994) were software companies, and soft- ware continues to create more new start-ups than any other technology field.25 Similarly, the rate of formation of new firms dedicated to the exploitation of one or another aspect of recent advances in biotechnology has been phenomenal: 800 new enterprises were founded in the 1980s, and the industry currently includes more than 1,200 firms. A few of these firms have become large, successful oper- ating companies (e.g., Amgen), however, the vast majority are still small, inves- tor-funded ventures. From 1980 to 1994, the shares of start-ups in computer hardware, advanced materials, photonics, optics, and telecommunications also increased.26 High-tech start-ups have played important roles in the U.S. technology enter- 

86 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY TABLE 2.6 High-Tech Companies Formed in the United States, 1960–1994 All High- Auto- Biotech- Computer Advanced Period Formed Tech Fields mation nology Hardware Materials Number of Companies 1960–1994 29,358 1,939 735 2,845 1,045 1980–94 16,660 917 546 1,907 487 1980–84 7,727 483 213 842 212 1985–89 6,510 331 225 756 194 1990–94 2,423 103 108 309 81 Percentage of all high-tech companies formed during each period 1960–1994 100.0 6.6 2.5 9.7 3.6 1980–94 100.0 5.5 3.3 11.4 2.9 1980–84 100.0 6.3 2.8 10.9 2.7 1985–89 100.0 5.1 3.5 11.6 3.0 1990–94 100.0 4.3 4.5 12.8 3.3 Percentage of all U.S. high-tech companies 1960–1994 100.0 100.0 100.0 100.0 100.0 1980–94 56.7 47.3 74.3 67.0 46.6 1980–84 26.3 24.9 29.0 29.6 20.3 1985–89 22.2 17.1 30.6 26.6 18.6 1990–94 8.3 5.3 14.7 10.9 7.8 aOther fields are chemicals, defense related, energy, environmental, manufacturing equipment, medical, pharmaceuticals, test and measurement, and transportation. SOURCE: National Science Board (1996). prise because they can accept a level and type of risk that larger companies usu- ally cannot. Able to serve highly dynamic niche markets, start-ups often serve as a “test-bed” for new products and services, a few of which might develop into large-volume businesses (National Academy of Engineering, 1995c). Further- more, start-ups are considered particularly adept at drawing effectively upon new product ideas of customers, suppliers, universities, research laboratories and oth- ers as well as at rapidly commercializing innovations.27 Many factors have enabled high-tech start-up companies to perform their unique roles in the U.S. innovation system.28 The following are among the most important: • the existence of sophisticated financial markets, particularly access to a large volume of venture capital and highly developed public equity markets; 

TECHNOLOGY TRANSFER IN THE UNITED STATES 87 Photonics Electronic Telecom- Other and Optics Software Components munications Fieldsa 977 7,661 2,923 1,556 9,677 507 5,196 1,293 933 4,874 221 2,467 629 408 2,252 191 1,962 508 370 1,973 95 767 156 155 649 3.3 26.1 10.0 5.3 33.0 3.0 31.2 7.8 5.6 29.3 2.9 31.9 8.1 5.3 29.1 2.9 30.1 7.8 5.7 30.3 3.9 31.7 6.4 6.4 26.8 100.0 100.0 100.0 100.0 100.0 51.9 67.8 44.2 60.0 50.4 22.6 32.2 21.5 26.2 23.3 19.5 25.6 17.4 23.8 20.4 9.7 10.0 5.3 10.0 6.7 • the large scale and technological intensity of relatively homogeneous seg- ments of the U.S. domestic market; • the large size, high mobility, accessibility, and entrepreneurial orientation of the U.S. technical workforce; • the sheer scale and accessibility of U.S. publicly funded nonproprietary research, particularly university-based research; • the scale of federal procurement combined with explicit preferences or set-asides for small and medium-sized vendors and suppliers; • a history of regulatory and other public policy commitments conducive to high-tech start-up companies, including the competition-oriented or tech- nology diffusion–oriented enforcement of intellectual property rights and antitrust law (competition policy), as well as relatively risk-friendly sys- tem of company law, particularly bankruptcy law; and • a highly individualistic, entrepreneurial culture nurtured in industry and many U.S. research universities by private practices, public policies, and various institutional mechanisms such as technology business incubators and venture capital firms that encourage risk taking. 

88 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY COMPARATIVE STRENGTHS AND WEAKNESSES OF THE INDUSTRIAL R&D ENTERPRISE The U.S. technology enterprise excels in the development and exploitation of new commercial technologies. Major shifts in the sectoral composition of U.S. industrial R&D and industrial production during the past 2 decades, as well as the large and rapidly expanding population of U.S. high-tech start-up compa- nies, attest to this fact. Further evidence of the dynamism and future growth orientation of the U.S. industrial technology base is offered by the U.S. patent and export statistics. The patent activity of U.S. companies encompasses a broad spectrum of tech- nologies and new product areas. However, recent patenting by U.S. companies demonstrates a strong emphasis on technologies or fields—medical and surgical devices, telecommunications, aeronautics, electricity transmission, advanced materials, biotechnology—that are expected to serve as engines of future eco- nomic growth29 as well as technologies associated with the extraction and use of the nation’s abundant natural resources (Table 2.7). Not surprisingly, these areas of patent emphasis reflect the competitive strength of U.S. industry in global high-technology product markets. In 1994, 25 percent of U.S. manufacturing exports were high-tech manufactured goods, and 3 of the 10 classifications of high-technology products accounted for nearly 85 percent of these technology exports: information technology (computers, software, and communications) (35.5 percent), aerospace (29.0 percent), and electronics (21.3 percent) (National Science Board, 1996). By way of comparison, U.S. patent activity by German companies in 1993 indicates an emphasis on technology areas associated with heavy manufacturing industries (motor vehicles, printing, power generation, and new chemistry and materials) that have long been a source of German compara- tive industrial strength in world markets.30 In contrast to the relative strength of the U.S. industrial R&D enterprise and its supporting nonindustrial R&D infrastructure in opening up new technological frontiers and launching new industries, the U.S. enterprise appears to be less ef- fective than some of its trading partners at serving the R&D and technology trans- fer/diffusion needs of technologically mature industries. In particular, U.S. companies in many technologically mature manufacturing industries appear to operate increasingly on the periphery of the nation’s nonin- dustrial R&D system. The R&D portfolios of U.S. research universities, federal laboratories, and most nonprofit research institutes have not overlapped much with the process and product R&D needs of firms, particularly small and me- dium-sized firms, in these industries. Many observers have noted gaps in the R&D portfolios of major technologically mature industries (Competitiveness Policy Council, 1993; National Academy of Engineering, 1993). Of particular concern have been perceived emerging gaps in these industries’ “infrastructural” R&D portfolios—R&D directed at the discovery and development of low-techni- cal-risk, difficult-to-appropriate technologies that have the potential to enhance 

TECHNOLOGY TRANSFER IN THE UNITED STATES 89 TABLE 2.7 Top 20 Most-Emphasized U.S. Patent Classes for Inventors from the United States and Germany, 1993 Ranking of class United States Germany 1 Wells Fluid-pressure brake and analogous systems 2 Mineral oils; processes and products Plant protecting and regulating compositions 3 Surgery, patent class 604 Printing 4 Surgery, patent class 606 Internal combustion engines 5 Chemistry, hydrocarbons Organic compoundsa 6 Special receptacle or package Synthetic resins or natural rubbersb 7 Surgery: light, thermal, and electrical Organic compoundsa applications 8 Chemistry: analytical and immunological Conveyors: power-driven testing 9 Fluid handling Organic compoundsb 10 Liquid purification or separation Winding and reeling 11 Error detection/correction and fault Organic compoundsa detection 12 Illumination Land vehicles 13 Chemistry: natural resins or derivatives Plastic articles 14 Receptacles Organic compoundsa 15 Amusement devices: games Synthetic resins or natural rubbersb 16 Communications: directive radio wave Organic compoundsa systems and devices 17 Information processing system Fluid sprinkling, spraying, and diffusing organization 18 Surgery Organic compoundsa 19 Hydraulic and earth engineering Compositions: coating or plastic 20 Supports Material or article handling aPart of the class 532–570 series. bPart of the class 520 series. SOURCE: National Science Board (1996). the performance of a broad spectrum of firms within an industry or related indus- tries. Also of concern are gaps in these industries’ “pathbreaking” R&D base— R&D aimed at discovering and developing high-technical-risk technologies with the potential for transforming existing industries (Alic et al., 1992). Factors that have helped weaken the connection between firms in many in- dustries and the nation’s nonindustrial research enterprise include the highly con- centrated (by industry and technology field) and mission-driven nature of federal R&D funding; the fragmented structure and low levels of industrial self-organi- zation of many technologically mature U.S. industries; and changes in the indus- trial composition of the U.S. economy (i.e., the increasing shares of total U.S. output accounted for by service and high-tech manufacturing industries). Nu- merous federal industrial technology initiatives of the past decade have sought to 

90 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY strengthen government-university-industry R&D cooperation as well as foster industrial consortia in selected industries (e.g., semiconductors, automotive). However, the volume of federal R&D dollars devoted to these initiatives has been small, and it is not yet clear whether these programs have been effective at forg- ing tighter linkages between industrial and nonindustrial R&D performers in es- tablished technologically mature industries. Another relative weakness of the U.S. industrial R&D/technology transfer enterprise is its limited capacity for diffusing new technology and know-how, particularly manufacturing or production technology, within technologically ma- ture industries and SMEs in particular (National Academy of Engineering, 1993). In recent years, there has been a concerted effort at both the federal and state levels to develop a more far-reaching network of private- and public-sector pro- viders of technical extension/industrial modernization services to SMEs. Ex- amples of this are NIST’s manufacturing extension partnership and related state initiatives. (See Part II, pp.76–79, and Annex II, pp. 205–209.) There are indica- tions that a growing percentage of U.S.-based manufacturers are adopting ad- vanced manufacturing technologies more rapidly (National Science Board, 1996). Nevertheless, compared with its German counterpart, the U.S. infrastructure for diffusion of production technology and other technologies to established indus- tries is much more uneven and fragmented. Technology Transfer to U.S. Industry in Context In order to begin to place technology transfer from nonindustrial R&D per- formers to U.S. industry in context, it is important to recognize that the volume of technology transfer that takes place internally among divisions of large private firms and externally between firms is by far the largest segment of U.S. technol- ogy transfer. This activity occurs through formal measures (such as mergers and acquisitions, and licensing of patents, software, and trade secrets) as well as through less formal mechanisms (such as sharing technical know-how, exchanges of personnel, and technical and marketing assistance). Data collected by the U.S. Internal Revenue Service show that in 1992, corporate royalty income in the U.S. manufacturing sector alone was almost $33 billion, roughly 100 times the royalty income of all of U.S. universities and federal laboratories combined. Indeed, that year several large technology-intensive firms reported royalty incomes of over $1 billion (e.g., IBM, Texas Instruments, and Bellcore). Several recent surveys of R&D-intensive companies shed light on the per- ceived relative importance of industrial and nonindustrial sources of commercial- izable ideas and technology. A 1992 survey by Roessner (1993) of member companies of the Industrial Research Institute (mostly large, research-intensive firms) found that respondents considered other companies (U.S. and foreign) to be the most significant sources of external technology, with universities second, private databases third, and federal laboratories fourth. Similarly, a 1994 pilot