National Academies Press: OpenBook

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

Chapter: Annex II: Case Studies in Technology Transfer

« Previous: Technology Transfer by Privately Held, Nonacademic Organizations
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 177
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 178
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 179
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 180
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 181
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 182
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 183
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 184
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 185
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 186
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 187
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 188
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 189
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 190
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 191
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 192
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 193
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 194
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 195
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 196
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 197
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 198
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 199
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 200
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 201
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 202
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 203
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 204
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 205
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 206
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 207
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 208
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 209
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 210
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 211
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 212
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 213
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 214
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 215
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 216
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 217
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 218
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 219
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 220
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 221
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 222
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 223
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 224
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 225
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 226
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 227
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 228
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 229
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 230
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 231
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 232
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 233
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 234
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 235
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 236
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 237
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 238
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 239
Suggested Citation:"Annex II: Case Studies in Technology Transfer." National Academy of Engineering. 1997. Technology Transfer Systems in the United States and Germany: Lessons and Perspectives. Washington, DC: The National Academies Press. doi: 10.17226/5271.
×
Page 240

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.

ANNEX II Case Studies in Technology Transfer BIOTECHNOLOGY Simon Glynn and Arthur E. Humphrey Biotechnology is literally a new technology, enabled by rapid expansion of our understanding of cell biology, especially of DNA, and the development of techniques that use this new understanding to physically change the genetic con- tent of cells. The United States dominates in the biomedical sciences and is the source of the vast majority of basic information in biotechnology. The United States has also dominated early efforts to realize the potential of biotechnology. This paper is intended to review the technology flows that have enabled this success. Defining the Scope of Biotechnology THE TECHNOLOGIES Biotechnology is defined by technologies, not outputs. These technologies, especially the sequencing and decoding of genes on a large scale, have trans- formed our understanding of the function of DNA in cells. These advances also enable researchers to manipulate genetic information in cells. For example, us- ing recombinant DNA technologies, the human gene that codes for insulin (a protein) can be isolated and then inserted in a bacterium. The bacterium can be made to synthesize human insulin, which may then be used to treat diabetes. Genetically engineered cells can produce not only human hormones such as in- sulin or growth hormone, but also blood products like clotting factors, vaccines, and new antibiotics. 177

178 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY These new technologies can also be used to create a class of proteins called monoclonal antibodies that are especially useful in diagnostics. These proteins are not created using recombinant DNA techniques, but by fusing a tumor cell to a white blood cell and then cloning this new cell. The resulting cells produce antibodies that are chemically identical. Monoclonal antibodies are used widely in research to identify the presence of specific types of molecules and to detect the presence of disease. HUMAN THERAPEUTICS AND DIAGNOSTICS Data on biotechnology revenues are inconsistent, but total annual revenues to U.S. companies from products developed using biotechnology appear to be about $10 billion and are projected to increase 15 to 20 percent each year over the next few years. Human therapeutics and diagnostics represent over 90 percent of these revenues (U.S. Department of Commerce, 1993). Table A-1 shows the number of drugs currently in development that use biotechnology techniques. In 1994, there were only 19 biotechnology-based drugs approved for use in the United States. (See Table A-2.) These drugs as a group rely on human hormones that were either understood or thought to be therapeutically useful in the treat- ment of diseases such as diabetes, anemia, and multiple sclerosis. These drugs have about $9 billion in annual global sales, or less than 5 percent of total global sales for pharmaceuticals (Merrill Lynch, 1996). As of February 1992, 640 diag- nostic kits using monoclonal antibodies, DNA probes, and recombinant DNA TABLE A-1 Biotechnology Drugs in Development, 1989–1993 1989 1990 1991 1993 Approved medicines 9 11 14 19 Medicines or vaccines in development Phase I 26 38 48 41 Phase I\II 12 13 16 22 Phase II 23 32 46 53 Phase II/III 8 6 7 6 Phase III 11 15 18 33 Phase not specified 5 3 2 4 Application at FDA for review 10 19 21 11 TOTAL medicines or vaccines in development 95 126 158 170 NOTE: Total medicines or vaccines in development reflects medicines in development for more than one indication. SOURCE: Pharmaceutical Manufacturers Association (1993).

ANNEX II 179 TABLE A-2 Biotechnology Medicines or Vaccines Approved for Use by the Food and Drug Administration as of 1993 Year Product Indication(s) Company Approved Beta interferon Multiple sclerosis Chiron 1993 DNAse Cystic fibrosis Genentech 1993 Factor VIII Hemophilia Genentech, 1993 Genetics Institute IL-2 Renal cell cancer Chiron 1992 Indium-111-labeled antibody Cancer imaging Cytogen 1992 Aglucerase Gaucher’s disease Genzyme 1991 G-CSF Adjunct to chemotherapy Amgen 1991 GM-CSF Bone marrow transplant Immunex 1991 Hyaluronic acid Ophthalmic surgery Genzyme 1991 CMV immune globulin Prevention of rejection in MedImmune 1990 organ transplants Gamma interferon Chronic granulomatous disease Genentech 1990 PEG-adenosine deaminase Immune deficiency Enzon 1990 t-PA Myocardial infarction, pulmonary Genentech 1990 embolism Erythropoietin Anemia associated with renal Amgen 1989 failure, AIDS, cancer Hepatitis B antigens Diagnosis Biogen 1987 Alpha interferon Cancer, genital warts, hepatitis Biogen, Genentech 1986 Hepatitis B vaccine Prevention Biogen, Chiron 1986 Human growth hormone Deficiency Genentech 1985 Human insulin Type I diabetes Genentech 1982 SOURCE: Read and Lee (1994). techniques had been approved by the U.S. Food and Drug Administration (FDA), including screening tests for the AIDS and hepatitis C viruses (U.S. Department of Commerce, 1993). The current generation of biotechnology drugs relies on major advances in biotechnology to identify and decode genes. This large-scale sequencing of genes is being done globally and is coordinated through gene databases on the Internet. Sequencing of the entire human genome may be completed by 2005 (Washington Post, 1996). Two examples of protein drugs based on these techniques are Amgen’s obesity drug Leptin and a protease inhibitor for AIDS that has had a dramatic clearing effect on the HIV virus (Merrill Lynch, 1996). NONMEDICAL USES OF BIOTECHNOLOGY Nonmedical uses of biotechnology are also apparent. Using biotechnology techniques, researchers hope to transfer into plants specific beneficial traits (e.g.,

180 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY resistance to pesticides, tolerance of hostile environmental conditions such as salinity or toxic metals, or higher nutritional content) (National Research Coun- cil, 1987). Bioprocess technologies are also expected to help in diverse sectors of the economy. In the petroleum industry, for example, bioprocessing has potential to degrade wastes or toxic substances (National Research Council, 1992c). Revenues from these nonmedical uses of biotechnology (agriculture, spe- cialties, environmental) are less than 10 percent of total revenues in biotechnol- ogy (U.S. Department of Commerce, 1993). There are several reasons for this. First, the use of biotechnology in areas other than human therapeutics and diag- nostics presents unique research and technical barriers not addressed by biomedi- cal research. Second, the use of biotechnology is constrained by economics. Drugs developed using early biotechnology techniques have tended to be exceed- ingly expensive. But new opportunities will require technologies to synthesize and purify the biological products at sharply lower cost and higher capacity (Na- tional Research Council, 1992c). Finally, commercial development in biotech- nology in the United States (so far) is directed by the size of the opportunity. The most immediate consequence of this is the current focus on human therapeutics and diagnostics, where the returns to investors are expected to be largest; there is relatively less focus on agricultural or industrial applications. For these reasons, many of the nonmedical uses of biotechnology are not expected to be commer- cially available before 2000 (Table A-3). New Biotechnology Companies Two different types of firms are pursuing the commercial potential of bio- technology: new biotechnology firms (NBFs), started specifically to exploit opportunities using biotechnology techniques; and large companies in pharma- ceuticals, chemicals, and other sectors for which biotechnology has important implications. The biotechnology sector included 1,272 biotechnology companies in 1993, of which 235 were public (Read and Lee, 1994). More than 100 of these compa- nies were started in the last 2 years, and 70 percent are less than 10 years old (Read and Lee, 1994). A large proportion of these NBFs, but certainly not all, are developing human therapeutics and diagnostics. Compared with the larger phar- maceutical sector, NBFs as a group are relatively small. According to a survey by Ernst and Young (1993), revenues for biotechnology companies were about $7 billion in 1992, compared with revenues of $114 billion for pharmaceutical companies. The biotechnology sector is nonetheless a very large funder of bio- medical research. According to the survey, NBFs spent nearly $5.7 billion on R&D in 1992, about half the R&D expenditures of the pharmaceutical sector. As is obvious from these levels of R&D spending, the overwhelming majority of biotechnology companies are research organizations with essentially no revenues. Nearly one-third of NBFs have no approved products, and 70 percent had rev-

ANNEX II 181 TABLE A-3 Selected Nonmedical Uses of Biotechnology Animals Vaccines Colibacillosis or scours (1984), pseudorables (1987), feline leukemia (1990) Therapeutic MAbs Canine lymphoma (1991) Diagnostic tests Bacterial and viral infections, pregnancy, presence of antibiotic residues Plants Diagnostic tests Diagnose plant diseases (turfgrass fungi) Biopesticides (killed bacteria) Kills caterpillars, beetles (1991) Bioprocessing Diagnostic tests Diagnose food and feed contaminants (salmonella, aflatoxin, listeria, campylobacter, and Yersinia entercolitica) Chymosin or renin Enzyme used in cheesemaking (1990) Alpha amylase Enzyme used in corn syrup and textile manufacturing (1990) Lipase Enzyme used in detergents (1991) Xylanase Enzyme used in pulp and paper industry (1992) Luciferase Luminescent agent used in diagnostic tests Environment Diagnostic test Detect legionella bacteria in water samples NOTE: MAbs = monoclonal antibodies. SOURCE: U.S. Department of Commerce (1993). enues of less than $5 million in 1992. Moreover, with very few exceptions, development efforts in the majority of these biotechnology companies are several years from approval. Almost all of these small companies will run out of money before their ideas are transferred to clinical practice. This problem becomes critical as the amount of R&D required to move sophisticated medical technologies to commercializa- tion increases. The investment in R&D is also risky. For example, failure to win FDA approval for their sepsis products cost investors in three companies— Synergen, Centocor, and Xoma—about $2.5 billion (Humphrey, 1995). To finance their research and development efforts, the new biotechnology firms have used a variety of funding mechanisms. The most important of these have been investments from venture capital firms, through public financing, and from larger companies. VENTURE CAPITAL The development of the U.S. biotechnology industry has largely been fi- nanced by venture capital firms. Venture capital is available to NBFs because the opportunity to exploit new advances in biotechnology for human therapeutics and diagnostics creates liquidity in public markets (as initial public offerings). Indeed, biotechnology attracted more venture capital financing—$261 million

182 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY invested in 95 companies—than any other sector of the economy in 1992, except software and services (Venture Economics, 1994). Venture capital firms are an important reason for the success of NBFs in the United States. In this country, nearly 75 percent of NBFs started as independent firms, compared with only 5 percent of NBFs in Japan, where venture capital is essentially nonexistent (National Research Council, 1992b). This difference may be an advantage for U.S. firms, since venture capital allows NBFs to form earlier and closer to intellectual capital in universities than would otherwise be possible (Zucker et al., 1994). PUBLIC FINANCING Public markets have also been a valuable source of financing for the higher- quality, larger-capitalization NBFs. In the early 1980s, several start-up biotech firms (Genentech, Cetus) set Wall Street records when they first went public. These firms have also been able to return to the public markets to finance produc- tion scale-ups and clinical trials. It is important to realize that health care reform and regulation impact the availability of venture capital and public financing, since investors focus on the anticipated returns on their investments. For example, regulations that require a certain number of clinical trials to determine the expected time to market of new drugs and therefore the cost of developing them. Health care reform efforts also play an important role by increasing the uncertainty with regard to biotechnol- ogy. Buyers of biotechnology-based drugs are now less often individual physi- cians than health care corporations, and third-party payers are becoming more restrictive, increasing the risk for investors and venture capital. LINKS BETWEEN NBFS AND LARGE COMPANIES Large pharmaceutical companies are an especially important source of fund- ing for new biotechnology firms. Large pharmaceutical companies have been investing in NBFs at an unprecedented pace. These cash infusions are especially important for equity investors in new biotechnology companies, because they reduce the future dilution they face. Linkages to NBFs are important to large pharmaceutical companies for sev- eral reasons. First, major pharmaceutical firms are looking increasingly for new, unique drugs for which there is no analog to treat diseases for which there cur- rently are few or no effective drug therapies. These are diseases such as cancer, Alzheimer’s, and AIDS that account for $500 billion in medical expenses each year in the United States (Merrill Lynch, 1996). Because the technology for developing these drugs is concentrated in the new biotech firms, NBFs have a comparative advantage in developing these new drugs. Second, it is important to recognize that the distinction between pharmaceu- tical companies and biotechnology companies is blurring as pharmaceutical firms

ANNEX II 183 are increasingly using biotechnology techniques to develop new drugs. Accord- ing to a study by the Boston Consulting Group (BCG) (1993), 33 percent of research projects in major pharmaceutical companies in 1993 were based on biotechnology, compared with only 2 percent in 1980. In some larger pharma- ceutical companies, up to 70 percent of the research projects were based on molecular biology techniques. Equity investments have enabled larger compa- nies to access the technology in these NBFs and to develop internal capabilities in biotechnology. Finally, the special strengths of the large pharmaceutical companies continue to be in traditional drug discovery, manufacturing, marketing, and distribution. Large pharmaceutical firms also are experienced in the drug approval process, which is especially difficult for NBFs. Linkages to large pharmaceutical compa- nies thus let NBFs exploit these competencies. For example, Humphrey (1993), at the inaugural meeting of the American Institute of Medical and Biological Engineering in 1992, observed that failure to integrate process design and engi- neering expertise into the development process for biotechnology drugs prior to phase III clinical trials resulted in many nonoptimal bioprocess designs that did not use leading-edge technology. LINKAGES TO FOREIGN FIRMS Technological links are also expanding between new biotechnology firms in the United States and large foreign firms. Foreign pharmaceutical companies understand that a global orientation is required to ensure long-term competitive- ness and financial returns, and they recognize that the United States is the world’s largest health care market. Foreign firms also seek access to advances in biotech- nology developed in the United States (National Research Council, 1992b). From the perspective of NBFs, the need for cash infusions to fund R&D encourages linkages with large, cash-rich foreign firms (National Research Council, 1992b). These linkages so far serve to transfer technology from the United States to foreign countries, although Japan’s strength in enzyme related bioprocessing tech- nologies is a potential opportunity for future technology transfer from Japan to U.S. biotechnology firms. Japan’s Kirin Brewery provided U.S. biotechnology firm Amgen critical robotic bioprocess technologies for the production of Epogen and Neupogen (Box 1). But the implication may be that these transfers represent a future competitive advantage for foreign firms in the U.S. and global markets1 (National Research Council, 1992b). The Importance of Universities The important advances in biotechnology—so far—have been made dispro- portionately by researchers in large U.S. research universities and then have dif- fused to the commercial sector, usually through NBFs. In this sense, U.S. univer- sities perform an incubator role for the biotechnology sector in the United States.

184 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY BOX 1 AMGEN Amgen, the largest independent biotechnology company in the world, is headquartered in Thousand Oaks, Calif. The firm has research cen- ters in Boulder, Colo., and Toronto, Canada; an international distribution center in Louisville, Kentucky; clinical research centers in Cambridge, England, and Melbourne, Australia; a fill-and-finish facility in Juncos, Puerto Rico; and a European regional headquarters in Lucerne, Switzer- land. Other international operating facilities are located in Australia, Bel- gium, Canada, China, France, Germany, Italy, Japan, the Netherlands, Hong Kong, Portugal, Spain, and the United Kingdom. Founded in 1980 by a group of scientists and venture capitalists, Amgen was able to attract a prestigious scientific advisory board that included several members of the National Academy of Sciences. In the autumn of that year, George B. Rathmann, formerly of Abbott Laborato- ries, was named Amgen’s chairman and chief executive officer—he was the company’s first employee. Amgen commenced operation in early 1981 with a private-equity placement of approximately $19 million, involving venture capital firms and two major corporations. The company chose its Thousand Oaks location to be near such major research centers as the University of Cali- fornia at Los Angeles, the University of California at Santa Barbara, and the California Institute of Technology. The company raised capital through public stock offerings in 1983, 1986, and 1987. Amgen’s stock is traded on NASDAQ’s National Market System under the symbol AMGN. Using techniques of recombinant DNA and molecular biology to cre- ate highly specialized health care products, Amgen scientific achieve- ments have positioned the company at the forefront of the biotechnology industry. As a result of this technology, Amgen has developed several human biopharmaceutical products. Two have been key moneymakers. Its re- PUBLIC FUNDING OF BIOMEDICAL R&D Health R&D now accounts for a rapidly growing share (16 percent in 1995, or $11.4 billion) of the government’s total R&D investment (National Science Foundation, 1994). Health research also received the single largest share–4 per- cent–of federal basic research dollars in 1995, or $6.3 billion. In comparison, general science, which included funding for the National Science Foundation (NSF) and for the research portion of the now-canceled Superconducting Super- collider, accounted for only 20 percent, or $2.9 billion, of estimated federal basic

ANNEX II 185 BOX 1—Continued combinant human erythropoietin, EPOGEN® (Epoetin alfa), stimulates and regulates production of red blood cells; and its recombinant granulo- cyte colony-stimulating factor (rG-CSF), NEUPOGEN® (Filgrastim), se- lectively stimulates the production of a class of infection-fighting white blood cells known as neutrophils. Amgen received its first patent for EPOGEN® on October 27, 1987, and a product license application (PLA) was filed with the FDA 2 days later. On June 1, 1989, EPOGEN® was approved by the FDA for treat- ment of anemia associated with chronic renal failure. Besides the development of key proprietary recombinant DNA meth- ods for the production of EPOGEN and NEUPOGEN, two major process technologies were important to the commercialization of these products. To achieve rapid and successful commercialization of these products, AMGEN made the decision to sharply scale up its roller-bottle technology. To do this, it entered into a joint venture—Kirin Amgen—with Japan’s Kirin Brewery Co., Ltd., in 1984 for the commercial development of re- combinant human erythropoietin. Through this joint venture, robotic bot- tling technology was adapted to roller-bottle manufacturing, A second key factor in the firm’s success was the in-house modifica- tion of the roller-bottle cap, allowing not only easy removal in the robotic process but also an increase in gas-mass transfer across the cap, result- ing in a greater than tenfold improvement in productivity. That this joint technology transfer has been eminently successful can be seen by the financial success of the corporation. Total AMGEN rev- enues for the year ended December 31, 1993, were $1.4 billion, primarily from sales of EPOGEN and NEUPOGEN. Revenues in 1992 were $1.1 billion. Net income for fiscal 1993 was $383.3 million, or $2.67 per share on a primary basis. In just 13 years, starting from scratch, AMGEN has become a Fortune 400 trading corporation. Adaptation of enabling tech- nology was important to achieving this success. SOURCES: Amgen, Inc. (various years). research authorizations (National Science Foundation, 1994). The overwhelming majority of this biomedical funding is directed to U.S. research universities and academic medical centers. To a considerable extent, this support has been concentrated on the emerging genetic engineering techniques in biotechnology, especially for AIDS research. In 1993, the U.S. administration, stating its intention to strengthen the FCCSET (Federal Coordinating Council for Science, Engineering, and Technology) pro- cess, included funding for six presidential initiatives in its initial 1994 budget

186 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY proposal.2 The largest of these initiatives was for biotechnology research, and more than three-quarters of this funding was controlled by the National Institutes of Health (NIH) (National Science Board, 1993). UNIVERSITY-INDUSTRY RELATIONSHIPS In quite a few instances, the mechanism for technology transfer in biomedi- cal R&D has been the establishment of (usually single-product) biotech start-ups, often with individual scientists and their graduate students literally moving from academia to industry. For this reason, universities have been the locus of innova- tion in biotechnology. For biomedical firms, locating near U.S. research univer- sities provides access to state-of-the-art research in fields essential to their contin- ued success. Indeed, about half of all NBFs in the United States are grouped around three major centers of academic biotechnology research: 21 percent of NBFs are close to Stanford, University of California at Berkeley, and UC San Francisco; 18 percent are near MIT and Harvard in Boston; and 12 percent are located near the NIH campus in Bethesda, Md. (Humphrey, 1995). The diffusion of basic information and expertise from U.S. universities to new biotechnology firms is essentially complete. Indeed, these laboratory tech- nologies are now widely disseminated, since virtually all of the research that enables biotechnology was performed in U.S. universities and academic medical centers using public money. There are few valuable strategic positions in these techniques (although separation and purification techniques, and process control are critical, as they create an economic advantage) (Gaden, 1991). Nonetheless, quite a number of interesting case studies seem to indicate that both the number and variety of alliances in biomedical R&D between academia and industry are increasing dramatically. In a recent study, Cohen et al. (1994) identified more than 1,050 research centers at U.S. universities, representing an aggregate budget of $4.12 billion in 1990, exactly half of all federal expenditures on academic R&D that year. Of these university-industry centers, 232, or 22 percent, conducted biotechnology research. Nearly 45 percent of expenditures were for basic research, although this actually represents less of an emphasis on applied research than academia as a whole (Cohen et al., 1995). Data on individual participation suggest that relationships between research- ers in academia and industry are even more pervasive than information on univer- sity-industry alliances indicates. For example, many NBFs have also established scientific advisory boards that include research scientists from U.S. universities and academic medical centers. Blumenthal (1992) found that 47 percent of bio- technology faculty consulted with industry, that 23 percent were involved in for- mal university-industry relations, and that 8 percent had received equity based on their own research. Biotechnology companies encourage these relationships. Genentech, for ex- ample, provides several million dollars of free recombinant materials to academic

ANNEX II 187 researchers every year. As a condition of receiving these materials, Genentech requires that any research findings be reported to Genentech, and Genentech as- serts the first right of refusal on any commercial applications developed (Personal communication from H. Niall, chief scientist, Genentech, to Simon Glynn, re- search associate, National Academy of Engineering, August 10, 1993). These dynamics are important, because federal and industry funding of bio- medical research are not quite the same thing. Industry and universities have increasingly diverging research agendas in biotechnology, and this is reflected in the priorities of academic researchers (Box 2). Of the individuals interviewed for the Harvard biotechnology project, 30 percent of biotechnology faculty with in- dustrial support said that their choice of research topics had been influenced by the likelihood that results would have commercial application. This compared to only 7 percent of faculty without commercial funding who said so. The terms of funding are also different: For extramural grants from the NIH, 92 percent are for 3 years or longer; for industry-funded research in universities, the majority of grants are for 2 years or less, consistent with the shorter time horizon of applied research (Blumenthal et al., 1986b). ECONOMIC INCENTIVES The institutional environment in which academics live is extremely impor- tant for technology linkages. In this respect, the changes in medicine have been faster and more dramatic than in other areas. Few, if any, examples of basic research in academic medical centers attracted commercial interest (unlike phys- ics and chemistry and even music) until the early 1970s and the acceleration of genetic research. Even then, at Stanford, there was significant culture shock (and in some cases even outright hostility) when patents and commercial interest in- truded into these medical departments after the first successful recombination of DNA by Cohen and Boyer in 1973. This is in sharp contrast to the current view of biotechnology at Stanford. Observed a prominent scientist, “The problem [now] is not pushing technology out of the lab; the problem—and this is a prob- lem—is pushing the technology too early. Technology advances too fast from academia to commercialization. I have a staff of nine, and everyone has a pet cure for cancer that they are pushing” (Personal communication from D. Botstein, Stanford University, to Simon Glynn, research associate, National Academy of Engineering, August 9, 1993). A critical element in this culture is the use of programs to provide financial incentives to support and encourage innovation by academics.3 At Stanford, for example, 15 percent is subtracted from total license revenues for the technology licensing office budget (this is usually excessive; the excess then goes to the Dean of Research for a research incentive fund for researchers without sponsor- ship). The net royalties are then divided, one-third to the inventors, one-third to their department, and one-third to the school of medicine (Personal communica-

188 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY BOX 2 The Monsanto-Harvard Agreement In 1982, Monsanto Corp. of St. Louis, Missouri, opened new facilities to expand its fundamental biotechnology research program. These labo- ratories were aimed at developing new products in animal nutrition, agri- culture, and human health, as well as developing and expanding the basic understanding of new biotechnology techniques. Through its program in molecular biology, the company was seeking a window on developing biotechnology through internally conducted research and collaborative research with universities and small start-up companies. The latter in- cluded such firms as Genentech, Genex, Biogen, and Collagen. As part of the program, in 1974 Monsanto entered into a landmark agreement with Harvard University to fund purely basic research for a period of 12 years. The company cosponsored studies on the molecular basis of organ development and tumor angiogenesis. The agreement assigned publication rights to the participating Harvard researchers and commercialization rights to Monsanto. After the initial 12 years, the program was not renewed; rather, it was scaled down and agreements were designed to support the research of individual Harvard scientists whose work was of specific interest to the company. Considerable insight was gained from the research about the carcinogenic behavior of many compounds produced by Monsanto. How- ever, greater commercial advantage was gained through the support of cooperative research with small or emerging biotechnology companies. One such cooperative venture was the joint program with Genentech to produce recombinant bovine and porcine growth hormones. SOURCE: Genetic Engineering News (1982). tion from H. Wiesendanger, Office of Technology Licensing, Stanford Univer- sity, to Simon Glynn, research associate, National Academy of Engineering, Au- gust 10, 1993). A similar policy is in effect at MIT and UC Berkeley. These licensing fees are an important alternative to government funding at the major U.S. research universities and especially in the emerging field of biotechnology, with companies supporting up to 16 percent of university research in this area (Blumenthal, 1992). The patents on DNA recombinant techniques by Boyer and Cohen are an example: The $14.6 million earned by the Cohen-Boyer patents in 1991 represented 58 percent of total income from all patents held by Stanford (Personal communication from H. Wiesendanger, Office of Technology Licens- ing, Stanford University, to Simon Glynn, research associate, National Academy of Engineering, August 10, 1993).

ANNEX II 189 Interaction With NIH and NIH-Funded Investigators NIH is the largest funder of biomedical research in the world. The agency funded $3.5 billion in biotechnology-related R&D in 1992, about 80 percent of all federal spending for biotechnology (National Research Council, 1992a). Rela- tionships between researchers at NIH and university scientists receiving NIH fund- ing are therefore an important dimension of technology transfer in biotechnology. NIH helps industry to develop this research for commercial use in four ways: participating in cooperative R&D agreements (CRADAs), licensing patented materials, training post-doctoral students and research fellows, and publishing. CRADAs involve NIH researchers and facilities in industry-directed re- search. This lets NBFs leverage these NIH resources. Several important prod- ucts have resulted from these collaborations, including the AIDS drugs AZT and DDI, and the HIV antibody tests. Nearly 1,000 CRADAs have been negotiated between the U.S. Department of Health and Human Services and industry since 1987 (U.S. Department of Commerce, Office of Technology Policy, unpublished data, 1996). NIH also facilitates technology links by licensing materials devel- oped by NIH and by training post-doctoral students and research fellows. These technologies and researchers interested in collaboration are listed in an electronic bulletin board funded by NIH.4 NIH researchers also publish about 7,000 techni- cal journal articles per year as well as present research at scientific meetings (National Research Council, 1992b). THE QUESTION OF FOR-PROFIT FUNDING AND RECIPROCITY Funding from for-profit organizations is seen as a potential problem in these relationships, especially if these relationships involve foreign competitors of U.S. firms. The 1993 agreement between the Scripps Research Institute, which re- ceives substantial NIH funding, and Swiss-based Sandoz Pharmaceuticals is an example of this. Under the terms of the agreement, which was scheduled to go into effect in 1997, Sandoz would give Scripps $300 million over 10 years and an option to extend the contract for another 10 years. In return, Scripps, the largest private biomedical research laboratory in the United States, agreed to give all its discoveries to the Swiss firm for the next 20 years (Hilts, 1993a,b). The Scripps-Sandoz agreement was widely attacked by NIH and Congress for giving Sandoz substantial control over the Scripps research laboratories and their findings, and for encouraging the commercial development of federally funded research by non-U.S. companies. Scripps spends about $100 million per year on research and receives $70 million of this funding from NIH. In response, NIH announced that the exclusive rights to patents from biotechnology discover- ies made using these federal funds would be removed. Scripps subsequently agreed to accept a substantially reduced contribution from Sandoz and to modify the terms of the agreement (Hilts, 1993a,b).

190 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY Importance of Fellowships, Conferences, and Specialized Journals An assessment of the factors that facilitate technology transfer must also give attention to the circumstances under which innovation happens. The discov- ery of DNA cloning, for example, derived from basic research. But the discovery occurred in quite exceptional circumstances—an environment conducive to sci- entific discovery and the exchange of information. Inventions almost never happen in isolation. The collaboration by Cohen and Boyer that led to the discovery of DNA cloning was proposed at a U.S./Japan scientific meeting held in Honolulu, Hawaii.5 Technology transfer in biotechnol- ogy depends to a very large extent on the ability of individual scientists or groups of scientists (as opposed to institutions) involved in research to interact with each other. Consequently, the imperative for NBFs is to create close interactions with these academic researchers. According to Hugh Niall, chief scientist for Genentech, this requires estab- lishing a culture in NBFs as similar to universities as possible. Critical to this is the ability of researchers to move back and forth between industry and academia, and to build their academic credentials by doing this. The immediate conse- quence of this is a large network of alumni—senior researchers who leave Genentech are usually retained as consultants, for example. A second conse- quence is that NBFs are able to recruit new scientists to continually renew the research organization. For example, Genentech employed 40 to 50 post-docs (out of 330 researchers) in 1993 in a 2-to-3 year fellowship program funded almost en- tirely by the company. These post-docs do curiosity-driven research, and many of them go on to careers in academia (H. Niall, personal communication, 1993). Conferences, professional organizations, and journals are also very useful for technology transfer. Large organizations, for example the Federation of American Societies for Experimental Biology (FASEB), attract 15,000 to 20,000 people to their meetings (National Research Council, 1992b). Many of these organizations also contribute to the internationalization of biotechnology. For example, more than one-quarter of the members of the American Society for Microbiology (ASM) come from outside the United States. These foreign mem- bers are seen as active contributors. Japanese members are especially active in molecular biology and fermentation technologies, for example. Foreign authors are also significant contributors to the ASM’s many scientific journals (National Research Council, 1992b). Policy Questions INTELLECTUAL PROPERTY RIGHTS Perhaps the most important policy question in university-industry relation- ships relates to intellectual property rights. Universities are the recipients of the majority of federal funding in basic research, but they are not appropriate institu- tions for the development and transfer of these findings to clinical practice. To

ANNEX II 191 address this, the 1980 Bayh-Dole Act gave universities new patent rights for all discoveries resulting from federally sponsored research, thus recognizing a criti- cal element in the transfer of technology: Industry must have reasonable expecta- tions of being able to recover product development costs (which are extremely high in biotechnology) or it will not participate. Patents and licenses on intellec- tual property developed at universities are, consequently, an absolute prerequisite for the transfer of biotechnology to clinical practice. The apparent effect of this initiative on universities is impressive. Before the enactment of Bayh-Dole in 1980, only about 4 percent of the more than 30,000 patents held by the federal government were ever licensed. Now, nearly 50 per- cent of patents are licensed (National Research Council, 1992a). Leading re- search universities also expanded their efforts to transfer technology to industry and to enhance their licensing activities. Indeed, in biotechnology, universities were more efficient in generating patents than private industry. Biotechnology companies in the 1980s were realizing more than four times as many patent appli- cations per dollar invested from university research than from their own labs’ investments (Blumenthal, 1992). The problem with these patents is the requirement that patented inventions be described in enough detail that they can be reproduced without undue experi- mentation. Because microorganisms generally cannot be described in such de- tail, courts have stipulated that this requirement must usually be met by submit- ting a sample of the microorganism to a depository. But this gives competitors direct access to the microorganism, increasing the opportunity for patent infringe- ment. Differences between the U.S. first-to-invent system of patents and the first-to-file system used in Japan and most other countries also create problems (National Research Council, 1992b; Olson, 1986). If the acquisition or enforcement of a patent seems difficult, NBFs may rely instead on trade secrecy laws to protect a product or a process. There are several disadvantages to trade secrecy laws, however. First, they offer no protection against someone who independently discovers the secret. Such a discoverer may then patent the finding and prohibit the original party from using it. Second, trade secrecy prohibits scientists from publishing the results of their own research in the scientific literature. Finally, theft of a trade secret is often difficult to prove (Olson, 1986). REGULATION The use of recombinant DNA has been regulated since the technology’s in- ception. In 1976, NIH issued guidelines for genetic research. These guidelines banned certain types of experiments, reflecting the perceived risks. More typi- cally, experiments had to be performed using various levels of physical and bio- logical containment. As experience with recombinant DNA increased, the NIH guidelines were successively revised. Today, the overwhelming majority of experi- ments using recombinant DNA are exempt from these guidelines (Olson, 1986).

192 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY Several limitations are apparent in the NIH guidelines. First, they apply only to institutions that receive federal funds, and the penalty for violating the guide- lines cannot extend beyond canceling this funding (although several regulatory agencies do require that NIH guidelines be observed). Second, NIH guidelines focus on research, not on commercial development. The scientific review pro- cess used by NIH is inadequate to deal with the volume of commercial develop- ment (Olson, 1986). NIH guidelines have also come in conflict with the rules of several regula- tory agencies. Approval for the release of genetically engineered microorgan- isms into the environment, for example, involves regulatory channels outside the NIH, including those of the Food and Drug Administration (FDA), the Environ- mental Protection Agency (EPA), and the Department of Agriculture (National Research Council, 1992b). There are other regulatory problems, as well. These agencies, in many in- stances, seek several distinct objectives. They have the responsibility to protect human health and the environment from any potential dangers posed by biotech- nology. The FDA, for example, requires pharmaceutical companies to demon- strate through a variety of means, including clinical tests on humans, that a new drug is “safe and effective.” In the instance of drugs developed using recombi- nant DNA, the FDA requires them to undergo the entire approval process irre- spective of identical approved or existing substances manufactured using identi- cal techniques. The reason for this is concern over the possibility of undetected contamination by drugs or chemicals, or the possibility of genetic instability in a recombinant organism. Receiving approval for a new drug developed using recombinant DNA tech- niques is consequently a long and expensive process. Approval of a new drug application (NDA) usually takes 2 years, and the average cost for FDA approval exceeds $200 million (Humphrey, 1995; Olson, 1986). In certain instances, the process can be accelerated. For example, important new drugs for AIDS have recently been approved for use in less than 1 year (Reingold, 1995). But the FDA is also increasingly under pressure to encourage and facilitate the expansion of commercial biotechnology in the United States. By imposing burdensome regulations, winning approval for biotechnology products will take longer. The current lead the United States enjoys in converting the results of biotechnology into commercial products may therefore be lost to biotechnology firms in other countries with less restrictive regulations. THE ORPHAN DRUG ACT AND PRICE REGULATION The viability of the Orphan Drug Act and the pricing of emerging biotech- nology drugs are important questions for regulation. Congress passed the Orphan Drug Act in 1983 to encourage development of drugs that, although clinically useful, had no commercial appeal due to very high development costs or that were intended to treat diseases from which fewer that 200,000 people suffered

ANNEX II 193 (Tregarthen, 1992). The Orphan Drug Act provides two incentives, if the drug is approved by FDA for use. First, a company that wins “orphan” designation for its product may receive tax credits for up to 50 percent of the cost of developing and marketing the drug. Second, the company receives an exclusive 7-year monopoly to market the drug for the specific orphan disease (Mossinghoff, 1992; Tregarthen, 1992). FDA has awarded nearly 500 orphan designations. As of 1993, 60 orphan drugs had been approved for use by the FDA (Mossinghoff, 1992). Seeking FDA approval is risky and expensive. But several important drugs developed under the Orphan Drug Act—for example Taxol, used to treat ovarian cancer, and the AIDS drug AZT (Retrovir)—have demonstrated outstanding com- mercial potential. Both of these drugs were the result of federally funded research (Tregarthen, 1992), and Congress has expressed concern over the high prices charged by drug companies for medicines developed in this way. In 1986, Wellcome priced AZT at $10,000 for a year’s supply but in response to intense public pressure has subsequently reduced the price of the drug to about $2,500 (Tregarthen, 1992; U.S. Centers for Disease Control and Prevention, 1996; Whitehead, 1992). The Future of Biotechnology The current view of the future of biotechnology is of continued U.S. leader- ship in basic research and of technology transfer out of U.S. universities and NBFs into larger pharmaceutical companies in the United States, Japan, and Europe. Large pharmaceutical companies will continue to expand their presence in biotechnology as the new generation of biotechnology drugs now in development by NBFs enters clinical trials. Many of these pharmaceutical companies will be foreign. Swiss-based Sandoz, for example, has acquired an interest in two U.S. NBFs (Genetic Therapy and Systemix) as well as access to advanced technolo- gies through its long-term agreement with the Scripps Research Institute. Indeed, virtually every European country, as well as Canada, Japan, and the Organization for Economic Cooperation and Development, has developed programs to exploit this “new” biotechnology (Gaden, 1991). These programs can be expected to increase international technology transfer in biotechnology, although—with per- haps a few exceptions—the pattern of technology transfer in these programs will continue to represent a net flow of technology out of the United States. THE DEVELOPMENT AND TRANSFER OF MANUFACTURING AND PRODUCTION TECHNOLOGIES TO U.S. COMPANIES Robert K. Carr Introduction Technology for manufacturing and production is difficult to discuss as a ho- mogenous entity. Unlike software, biotechnology, and electronics, the three other sectors studied in this document, manufacturing and production technologies

194 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY overlap a number of traditional academic and industrial categories. Manufactur- ing and production technologies can be product, firm, or industry specific. Broadly defined, they can include parts of associated technologies, such as mate- rials and environmental cleanup, and are an important feature of industries as diverse as biotechnology, automobiles, and microchips. Furthermore, a modern definition of production technology includes not only machinery and other hard- ware, with associated software and computing and communications, but also a range of “soft” technologies, including just-in-time production techniques; lean, flexible, and agile production; total quality management; and a host of other new ways of doing things better. Finally, while the research base for manufacturing and production technologies is centered in engineering, it draws heavily on ad- vances in materials, software, electronics, and other academic disciplines. The following case study examines the infrastructure supporting the devel- opment and transfer/diffusion of production and manufacturing technologies to private industry in the United States. It is important to recognize from the outset that the vast majority of the research and technology transfer of production and manufacturing technologies takes place entirely within the private sector. The research and technology transfer resources at the disposal of private-sector firms will be examined only briefly, however. The focus here will be on the roles of nonmanufacturing research institutions and intermediaries (e.g., universities, fed- eral laboratories, and nonprofit R&D institutions) in the development and trans- fer of production and manufacturing technologies through public-sector technol- ogy transfer networks and other nonmanufacturing technology transfer agents. It may be useful to distinguish between large and/or R&D-intensive manu- facturing firms on the one hand and small and medium-sized firms (SMEs) on the other. The majority of industrial R&D in the United States is performed by large firms. The top 10 R&D-intensive firms perform fully 25 percent of all industrial R&D, and all large firms (with 5,000 employees or more) perform 73 percent (National Science Foundation, 1996c). Furthermore, these large manufacturing firms have their own infrastructure for developing or acquiring manufacturing technologies and for interfacing directly with academic, government, and non- profit R&D entities. Thus, they tend to be less involved in the organizational frameworks that exist to support the transfer of manufacturing technology from the government, nonprofit, and academic sectors. On the other hand, the 375,000 or so SMEs, which constitute fully 98 percent all of U.S. manufacturers, perform little or no R&D. Many of these firms do not produce finished products and are not well known in the marketplace or among U.S. exporters. Nonetheless, they are critical elements in the “food chain” of the manufacturing sector, accounting for up to 60 percent of the cost of manufactured goods (National Research Council, 1993). Therefore, their efficiency and pro- ductivity have a major impact on the overall competitiveness of the manufactur- ing sector. For this reason, the focus in this case study will be on technology transfer programs where small and medium-sized firms are the recipients.

ANNEX II 195 Developing New Manufacturing Technologies: The R&D Base Given the diverse nature of production and manufacturing technologies, the wide range of industries in which they are used, and the very large number of manufacturing firms, it should come as no surprise that gathering precise data on manufacturing R&D poses difficulties. In academia and government, R&D tends to be classified according to academic discipline or government mission, rather than by crosscutting categories such as manufacturing. In industry, data on R&D is collected (primarily by the NSF) according to industry as defined by SIC codes, and by firm size. Nonetheless, some data for production and manufacturing R&D can be identified. PRODUCTION AND MANUFACTURING R&D IN INDUSTRY Industry funds and performs the lion’s share of U.S. research and develop- ment, providing 57 percent of all support for U.S. R&D and accounting for 72 percent of U.S. R&D performance. According to the National Science Founda- tion (1996c), manufacturing firms accounted for 74 percent, or roughly $87 bil- lion, of the R&D performed by industry in 1993. However, process-oriented R&D is not distinguished from new-product R&D in the NSF data. Furthermore, not all process and manufacturing technology R&D is carried out in the manufac- turing sector; service firms that support the manufacturing sector are also active in this area. Compared with other nations, the level of industrial R&D devoted to produc- tion processes in the United States is low. In a 1988 study, Edwin Mansfield stated that American firms “devote about two-thirds of their R&D expenditures to improved product technology and about one-third to improved process tech- nology.” He contrasted the ratio of U.S. process-to-product-oriented R&D with that of Japan, where the proportions are reversed (Mansfield, 1988). Another study (National Science Board, 1992) found that only 19 percent of U.S. indus- trial R&D was devoted to process innovation. These two studies indicate that U.S. industry performed between $23 billion and $40 billion in process-related R&D in 1995 (National Science Foundation, 1996c). Recent data from an ongoing study (Whiteley et al., 1996) by the Industrial Research Institute and Lehigh University’s Center for Innovation and Manage- ment Studies (CIMS) have confirmed these estimates and provided additional survey data on the division between product and process development in several industry areas (Figure A-1). The IRI/CIMS survey divides R&D into basic and applied research, product and process development, and technical services (sup- port for existing products/processes in the field). For the 87 firms surveyed in 1994, 22.5 percent of their R&D efforts were process related and 41.8 percent were product related. Basic and applied research consumed 17.1 percent of their efforts and technical services consumed 18.5 percent. The figures for 1993 were

196 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY quite similar. However, as can be seen from Figure A-1, the ratio of product-to- process-related R&D varies widely by industry group. FEDERAL SUPPORT FOR PRODUCTION AND MANUFACTURING R&D The federal government’s role in production and manufacturing technolo- gies includes both funding research in the university, private, and nonprofit sec- tors, as well as conducting process-oriented R&D within federal facilities. The available data do not permit one to precisely separate federal funding and federal performance of production and manufacturing R&D. However, the nature of funding agency missions provides a general idea. In 1993, the Federal Coordi- nating Council for Science, Engineering and Technology (FCCSET) prepared several S&T initiatives for the FY 1994 budget. One of these initiative areas was Advanced Manufacturing Technology, while another, Advanced Materials and Processing, was supportive of manufacturing technology. These two initiatives no longer exist in their former form, but the budget figures reported for FY 1994 provide a good sense of the level of federal activity in this area. To be sure, much of federal R&D in advanced manufacturing as well as in materials was intended to meet unique federal requirements in defense and space. Nonetheless, there have frequently been significant spin-offs from such activity. The FY 1994 federal total for advanced manufacturing was $1.385 billion and for advanced materials was $2.061 billion (Federal Coordinating Council for Science, Engineering and Technology, 1993). The agencies with the largest bud- gets for advanced manufacturing programs were the Department of Defense (DOD) ($596 million, including the Technology Reinvestment Project, or TRP), the Department of Energy (DOE) ($367 million), the National Institute of Stan- dards and Technology (NIST) ($141 million), and the National Science Founda- tion ($130 million). In the advanced materials initiative, DOE had the largest budget ($946 million), followed by DOD ($422 million), NSF ($328 million), the National Aeronautics and Space Administration (NASA) ($131 million), and the Department of Health and Human Services ($93 million). Almost all of the fig- ures for NSF represent federal funding of extramural research. Most of the other agencies performed the lion’s share of their advanced manufacturing and materials research in their own laboratories, with a small part sent to external performers. The National Science and Technology Council (NSTC), the successor to FCCSET, changed the nature of the FCCSET initiatives, but there are still activi- ties relevant to manufacturing technology. The Manufacturing Infrastructure Ini- tiative is designed to support R&D and other activities that support the entire manufacturing sector. The initiative consists largely of ongoing programs, most of which are described separately in this section. The National Electronics Manu- facturing Initiative (NEMI) was launched in response to industry interest in form- ing a partnership with the federal government to assess the technology needs of electronics manufacturing. NEMI is not a set of specific programs or projects,

ANNEX II 197 FIGURE A-1 Allocation of R&D funds for different industries: product vs. process de- velopment, fiscal year 1994. SOURCE: Whiteley et al. (1996). but rather a way of doing strategic planning and partnering. The Materials Tech- nology Initiative inherited by NSTC has been folded into other NSTC initiatives where its subcommittees provide support to other activities. For the most part, these initiatives include few, if any, new R&D programs among federal agencies or significant new budgets for work in these sectors. They are best thought of as frameworks for budget presentation, coordination, and reporting of manufactur- ing activities in diverse federal R&D programs. Defense Manufacturing Programs The major DOD activities in manufacturing R&D include the TRP and the Manufacturing Science & Technology program (MS&T). TRP funded a number

198 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY of technology extension activities in 1993 and 1994, but has recently been con- verted to a military-specific program and no longer funds industrial extension programs. Although the TRP funded some projects that included the develop- ment of new production technologies, most of its activities in support of manu- facturing were in the deployment arena. The MS&T program, a successor to the ManTech program, consists of manufacturing R&D activities by the services, the Defense Logistics Agency (DLA) and the Defense Advanced Research Projects Agency (DARPA). The MS&T program is managed by the Director for Defense Research and Engineering (DDR&E) and, in 1995, was funded at $329 million. MS&T “matures and validates emerging manufacturing technologies to support low-risk implementation in industry and DOD facilities.” MS&T programs are carried out with all types of organizations, including DOD contractors, suppliers, hardware and software vendors, industry centers of excellence, industrial consor- tia, universities, and research institutes. Cost sharing is part of the MS&T pro- gram, particularly with industrial partners, who may bear a considerable portion of their own costs for an MS&T program. Department of Energy Laboratories Most of the large multiprogram laboratories of the Department of Energy have manufacturing programs. Two DOE laboratories stand out in this regard: Sandia National Laboratories and Oak Ridge National Laboratory. Sandia is primarily an engineering laboratory, and therefore manufacturing programs (par- ticularly those related to weapons manufacture) have been critical to its mission for many years. Oak Ridge, particularly its Y-12 facility, has long been engaged in manufacturing research. The Oak Ridge Center for Manufacturing Technol- ogy has been formed to coordinate manufacturing technology programs at that facility. DOE laboratories tend to use CRADAs to work with industry in manu- facturing as well as in other areas. Although most manufacturing R&D in DOE laboratories supports weapons manufacture, much of it (particularly in areas such as precision machining) is transferable to civilian uses. DOE laboratories also have substantial activities in the area of materials. Some of this work is unique to government (e.g., plutonium), while other aspects of it, like advanced manufac- turing, is transferable. NIST Laboratories NIST is the only federal laboratory that has industry as its principal client. An important part of NIST’s service to industry is the Manufacturing Engineer- ing Laboratory (MEL). With a staff of 300, MEL works with U.S. manufacturers to develop and apply technology, measurements, and standards. MEL operates the National Advanced Manufacturing Testbed, the successor to the Automated Manufacturing Research Facility established over a decade ago. In addition to the MEL, NIST also operates the Materials Science and Engineering Laboratory.

ANNEX II 199 Much of the work of this facility is relevant to production and manufacturing technology. National Science Foundation NSF has a collaborative manufacturing research effort among several NSF directorates that supports manufacturing in several ways. NSF spends about 4 percent of its budget (about $300 million in 1995) on manufacturing-related grants and programs (National Research Council, 1995b). Investigator-initiated R&D projects that support development of the fundamental science and engineering base underlying manufacturing technology are part of NSF’s traditional peer- reviewed grant program. In addition, NSF funds two types of engineering re- search centers that support manufacturing. The centers, described below, are expected to become fully self-supporting. Engineering Research Centers (ERCs) are engaged in cross-disciplinary research and education activities that are important for U.S. competitiveness. ERCs are located at universities and promote links between research and educa- tion. The ERC program was begun in 1985 and currently supports 23 centers involving 100 participating academic institutes and almost 600 nonacademic part- ners. NSF contributed $51 million to the centers’ operation in 1995, while all other sources contributed $96 million. Almost 5,000 people (researchers and students) utilized center facilities in 1995. The 18 ERCs in operation at the end of 1994 were distributed according to their major technology focus as follows (Na- tional Academy of Engineering, 1995a): Design and Manufacturing 5 Materials Processing for Manufacturing 3 Optoelectronics/Microelectronics and Telecommunications 4 Biotechnology/Bioengineering 3 Energy and Resource Recovery 2 Infrastructure 1 Industry/University Cooperative Research Centers (I/UCRCs) and State/Industry University Cooperative Research Centers (S/IUCRCs) are also located in universities and focus on fundamental research areas recommended by industrial advisory boards. S/IUCRCs are similar to I/UCRCs, but are more closely focused on state or regional economic development and are initiated by states with industrial support. I/UCRCs have been in existence since 1973, with the first S/IUCRC added in 1991. In 1995, NSF contributed $8 million to 67 cooperative research centers, while all other sources contributed $79 million. Almost 2,300 people (researchers and students) utilized I/UCRC and S/IUCRC facilities in 1995.

200 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY PRODUCTION AND MANUFACTURING R&D IN ACADEMIA: UNIVERSITY INDUSTRY RESEARCH CENTERS (UIRCS) Although research relationships between industry and universities date back to the late 19th century, the wholesale formation of centers involving industry- university collaboration is a relatively recent phenomenon, dating from the 1970s and 1980s. As of 1991, there were over 1,000 UIRCs with a total estimated budget of $4.12 billion, of which $2.53 billion was devoted to R&D (most of the balance was for educational activities) (Cohen et al., 1994). A relatively small number of centers receive support from the NSF. Overall, government provided 46 percent of UIRC funding (federal sources accounted for 34.2 percent; state sources for 12.1 percent) while industrial participants provided 30 percent of the centers’ financial backing (representing over 70 percent of industry’s financial support for academia) as well as additional noncash support. UIRCs are quite diverse in their size, organization, relationship to industrial needs, and research activities. Some have a traditional academic orientation, pursuing research for its own sake, while others (about 25 percent) are focused on the goals of industry. According to Cohen et al. (1995), “More than one-quarter of UIRCs conduct R&D relevant to the manufacturing sector exclusively, while more than two-thirds conduct R&D that is relevant to both the manufacturing and non-manufacturing sectors.” The technology focus of UIRCs includes significant emphasis on manufac- turing (Cohen et al., 1995). Almost 20 percent of UIRCs carried out research in manufacturing technologies, while 30 percent of centers were involved in envi- ronmental technology and waste management and 27 percent in advanced materi- als, the latter two technologies being of considerable interest and concern to manu- facturers. Another study (Dickens, 1995) identified 1,030 university-based engineering research units (including UIRCs) at 154 universities. A survey of the directors of these units revealed that 45 percent were working in materials, 42 percent in energy and environmental technologies, 29 percent in manufacturing, 27 percent in information and communications, 17 percent in transportation, and 13 percent in biotechnologies and life sciences. Although UIRCs are the most visible type of university-industry research cooperation, the establishment of a UIRC is not essential for such interaction to occur. Many universities and their departments (particularly engineering) have relationships with industry, receiving support for research activities and contrib- uting knowledge to industrial sponsors. OTHER CENTERS OF MANUFACTURING R&D In addition to government, universities, and industry, independent nonprofit institutions also engage in R&D related to manufacturing. They make up by far the smallest of the four groups in terms of their numbers and R&D budgets. Their manufacturing R&D activities are proportionately smaller, in part since over half

ANNEX II 201 of nonprofit independent R&D institutes conduct work primarily in biomedical and other areas not directly relevant to manufacturing. Nonetheless, there are some significant centers of excellence in manufacturing research in organizations such as the Battelle Memorial Institute, SRI International, and Southwest Re- search Institute. These types of institutions are frequently called upon to solve industrial problems relating to production and manufacturing R&D. Transferring Manufacturing Technology to Industry There is substantial need for and substantial barriers to the acquisition of modern technology by small firms. At the end of the 1980s, over three-fifths of the machine tools used by U.S. manufacturers were over 10 years old, and more than one-quarter were over 20 years old (Shapira, 1990). Furthermore, more recent data indicate that while larger firms are modernizing, smaller firms con- tinue to lag behind. Table A-4 shows the percentage of large and small U.S. firms that have adopted nine key types of modern production technology. Small U.S. firms lag larger firms substantially in this regard. Both large and small U.S. firms lag their Japanese counterparts in all but one category. As well as lacking mod- ern equipment, U.S. small manufacturers tend to have neither highly trained staff nor modern operating methods. They are often content with this arrangement because it is similar to that of their nearby competitors, and it is often still per- ceived as sufficient for corporate survival. Transferring Manufacturing Technology from Federal Laboratories and Universities Federal laboratories transfer technology through a number of mechanisms, but three are particularly relevant to manufacturing: licensing, cooperative R&D, and technical assistance. These three mechanisms are used to different degrees by different laboratories. NASA centers have tended to focus on technical assis- tance, while DOE laboratories have preferred cooperative R&D. Licensing is the least-used mechanism, particularly for manufacturing technologies. The DOE reported in late 1994 that manufacturing technologies accounted for 18 percent of all DOE CRADAs. Two closely related areas, advanced materials and instru- mentation, and pollution minimization and remediation, accounted for 18 percent and 12 percent, respectively, of CRADAs (U.S. Department of Energy, 1994). Many federal laboratories have active technical assistance programs for manufacturers. While some labs have been offering technical assistance to firms in their immediate vicinity for many years, more and more federal laboratories are becoming technology sources in state-run industrial extension services, and their technical assistance activities will begin to be reflected in the data collected by these state extension services. The National Technology Transfer Center (NTTC) refers callers (mostly from the business community) to sources of technology in federal laboratories. In one

TABLE A-4 Use of New Technology in Manufacturing, Japan and the United States, 1988 202 Technology Users as a Percentage of Small and Large Manufacturing Large/Small Japan/U.S. Enterprisesa Ratio Ratio Japan U.S. Japan U.S. Small Large Type of New Manufacturing Technology Japanese Definition (closest U.S. (a) (b) (c) (d) (e) (f) (g) (h) definition in parentheses) Small Large Small Large Numerically controlled and customized numerically controlled machine tools (NC.CNC machine tools) 57.4 79.4 39.6 69.8 1.4 1.8 2.0 1.1 Machining centers (FMS cells or systems) 39.4 67.4 9.1 35.9 1.7 4.0 7.4 1.9 Computer-aided design (and computer-aided engineering) 39.1 75.2 36.3 82.6 1.9 2.3 2.1 0.9 Handling robots (pick-and-place robots) 22.6 62.2 5.5 43.3 2.8 7.8 11.2 1.4 Automatic warehouse equipment (automatic storage and retrieval) 10.9 44.9 1.9 24.4 4.1 13.1 24.1 1.8 Assembly robots (other robots) 8.3 41.4 3.9 35.0 5.0 8.9 10.6 1.2 NOTES: a The comparisons between Japan and the U.S. are approximate since differences exist in technology definitions and employment size categories. Addi- tionally, the Japanese data are enterprise-based, while the U.S. data are establishment-based. The Japanese define a small manufacturing enterprise as having less than 300 employees and a large enterprise as having 300 employees or more. The U.S. government defines small enterprises as firms with 50 to 499 employees and large enterprises as firms with 500 employees or more. (e) = (b)/(a). (f) = (d)/(c). (g) = (a)/(c). (h) = (b)/(d). SOURCES: After Shapira et al. (1992, p. 5). (a),(b) Ministry of International Trade and Industry (1989). TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY (c),(d) U.S. Department of Commerce (1989).

ANNEX II 203 13-month period (from June 1, 1994, to June 30, 1995), requests for assistance in the area of manufacturing technology ranked second, accounting for 9 percent of all inquiries, while materials sciences and environmental pollution and control accounted for 16 percent and 7 percent, respectively. Of the firms making re- quests during that period, over 70 percent were small businesses, most with fewer than 100 employees. Small Business Development Centers (SBDCs), operated by the Small Busi- ness Administration (SBA), provide educational and research resources to small businesses. There are over 900 SBDCs in operation, providing direct counseling to small business owners and managers. SBDCs are initiated at the state level and funded by state governments as well as the SBA. SBA funds go to a state university or economic development agency, which serves as the “lead center” in the state, with subcenters established at other educational institutions and cham- bers of commerce. In 1995, federal funding for the SBDC program was $73.5 million, while matching state funds amounted to $81.6 million. Individual SBDCs vary according to geographic area and in terms of clients and services offered, but many actively support small manufacturers. SBA has a collaborative working relationship with the Manufacturing Extension Partnership (MEP) at NIST. Many, if not most, SBDCs are integrated into state technology extension pro- grams and are part of the network of service providers available to small busi- nesses. In states where active industrial extension networks resolve manufactur- ers’ technology problems, SBDCs tend to focus on managerial issues such as finance. Universities transfer technology primarily through licensing, the formation of spin-off companies, faculty consulting, cooperative R&D (particularly in UIRCs), and the flow of graduates to private firms. There is little information on the technology areas of university licensing and faculty consulting, and therefore it is difficult to know what percentage of these activities is related to manufactur- ing. It is likely that most of the manufacturing technology flows from engineer- ing programs through faculty consulting and graduates as well as from coopera- tive R&D in UIRCs. It is probable that much of this technology is at the high end, useful primarily to large or technologically sophisticated firms. As is the case with federal labs, some university centers and engineering departments have be- come resources for state extension networks. Community colleges are even more frequently involved in state industrial modernization and extension systems. UIRCs are active in technology transfer to their industrial sponsors. Accord- ing to the Cohen et al. (1995) study, almost two-thirds of the UIRCs indicated that transferring technology to industry was “important,” even though only a small percentage of the effort of centers was specifically devoted to technology trans- fer. The centers reported that collaborative R&D, exchange of research person- nel, delivery of prototypes or designs, and informal contacts were the most effec- tive technology transfer mechanisms. Although not surveyed as a technology transfer mechanism, the flow of students to industry is clearly another important

204 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY avenue of transfer. The 1995 study indicates that industrial sponsors of UIRCs hire a substantial number of center graduates students. Transferring Manufacturing Technology: Industrial Extension Programs THE AGRICULTURE MODEL Development of the agricultural research and extension system began in 1862, when the federal government established agricultural colleges, run by the states, to offer practical instruction in agriculture. Fifteen years later, the federal government established a system of state agricultural experiment stations, again under the state auspices. Finally, the Cooperative Agricultural Extension Service (AES), a partnership among federal, state, and county governments, was created by the Smith-Lever Act in 1914. The AES has grown into a nationwide system with more than 9,600 county agents and 4,600 university researchers. Funding is spread among federal, state, and local governments. In an era of farming by large agribusiness, some believe the extension ser- vice no longer plays a critical role. Still, when the AES first came into existence, farmers were small businessmen who ran into many of the same barriers to tech- nological advancement as do small manufacturers today. Thus, when manufac- turing extension was first considered, the Agricultural Extension Service was an obvious model. However, in spite of the parallels, there are some critical differ- ences between the model and the realities of modern manufacturing. For one thing, farmers in a local area tend to have the same problems, while manufactur- ers’ problems are often very different. BOTTOM-UP APPROACHES: STATE AND LOCAL PROGRAMS Some states have long recognized that small manufacturers, like farmers, had much to gain from technical assistance. While most state efforts in the tech- nology area initially focused on research and development, a few states created technical assistance programs. North Carolina began such a program for manu- facturers in 1955, and Georgia followed in 1960. Both of these programs are still flourishing. North Carolina’s cooperative technology program is the country’s largest, with an annual budget of $37 million, while Georgia’s now ranks fourth at $30 million (Coburn, 1995). Pennsylvania followed suit in the mid-1960s with the Pennsylvania Technical Assistance Program (PENNTAP). By the end of the 1970s, Maryland, Massachusetts, Michigan, New York, Ohio, and Virginia had begun their own extension programs. By 1994, 40 states had technology extension programs (Coburn, 1995) in addition to other efforts to assist manufacturers. Approximately half of the state programs were operated by educational institutions, with the balance managed by nonprofit organizations or state agencies. These programs offer different types of

ANNEX II 205 services, including supply of technical information, seminars and workshops, demonstrations, referral of consultants and other experts, and in-plant consulta- tion. However, intensive field assistance (generally agreed to be the most effec- tive technique) was provided by only a few programs (Shapira et al., 1995). In Modernizing Manufacturing, Philip Shapira (1990) groups state industrial extension programs into four categories: Technology broker programs focus on providing technical information and referrals for client firms. Typically, these programs have large numbers of re- quests, each of which receives modest attention from program staff (generally less than a day). In 1993, PENNTAP assisted 490 client firms (the majority of them small firms) and handled 700 requests with a staff of eight at no cost to the requesting firms. University-based field office programs generally make engineers available to assist firms in a wide range of problem areas. By virtue of being university based, these programs can easily access engineering faculty or R&D centers for assistance. Service is normally provided for free. One of the oldest and largest such programs, the Georgia Industrial Extension Service, is operated by the Georgia Institute of Technology. Founded in 1961, it has 13 regional offices (being expanded to 17), which served over 1,000 companies and communities in 1994 using $1.55 million in state funds. Technology centers and state-sponsored consulting services are not part of a university, although they may have links to one. Their focus is generally on technology modernization (i.e., technology assessments, upgrade recommenda- tions, implementation, training, etc.). Assistance is often provided by private consultants subsidized or paid for by the state program. Pennsylvania’s Indus- trial Resource Centers (IRCs) are an example. Established in 1988, the IRCs are private nonprofit corporations operated by private-sector boards. Seven of the IRCs serve the traditional manufacturing sector, and one is devoted to the sup- port of the biotechnology industry. They provide assistance with their staff and through private consultants. Initial assistance is generally free, but in-depth assess- ments and services by outside consultants require some company payment. Manufacturing networks are regional networks of firms that cooperate in tech- nology diffusion, training, design, finance, and marketing. To a certain extent, these networks have been influenced by the successful small-firm networks in Northern Italy. In this country, the Southern Technology Council (STC) has established networks in North Carolina and Arkansas, which involve commu- nity colleges and economic development authorities along with local firms. The Arkansas Industrial Networking Project was created by the Arkansas Science and Technology Authority with a $90,000 grant from the STC. The project’s goal is to improve the competitiveness of small manufacturers by facilitating cost sharing for R&D, purchasing, training, and expensive technology, and en- couraging cooperation on contract bidding. The Arkansas networks, which are focused on the wood-products and metal-working industries, involve about 100 companies.

206 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY Shapira (1990) surveyed state industrial extension programs for information about the types of assistance they provided to client firms. The most frequently offered services, in descending order of popularity, were: (1) improve/solve prob- lem with existing production technology; (2) identify vendor of new technology/ software; (3) specify new production/process technology; (4) refer client to train- ing source; (5) improve quality control/statistical process control; (6) improve existing plant/layout operations; (7) identify new markets; (8) address waste man- agement/environmental problems; and (9) improve/debug and existing product. Service type (1) was by far the most frequently sought, while types (2), (3), and (4) were roughly equal in importance. In addition to industrial extension programs, the states engage in a number of other activities in support of their technology base. Coburn (1995) identifies a number of different models, including the following: University-industry technology centers (UITCs), which feature interdiscipli- nary research in areas relevant to industry. These centers are either university based or are operated by a nonprofit in close association with a university. They are often supported by federal agencies, as in the case of the NSF engineering centers. Kansas, New Jersey, New York, and Ohio have strong center programs. Government-industry consortia are groups of firms that, with other R&D in- stitutions, such as universities, focus on research in a given area. Although some states sponsor such consortia, they are more typical at the federal level. University-industry research partnerships are similar to consortia but are project centered and have a start and end date. Arkansas, Connecticut, Kansas, and Maryland sponsor such programs. Equipment and facility access programs provide state firms access to expen- sive and sophisticated equipment and facilities and associated staff expertise. Colorado, Maryland, North Carolina, and Pennsylvania, among other states, have such programs. Technology financing programs provide capital to firms and to specific projects under a broad spectrum of arrangements including grants, low-cost loans, guar- antees to third-party lenders, and investment in exchange for equity. Start-up assistance includes state-supported incubators for new businesses as well as research parks, where high-tech companies can obtain both research space and (usually) access to a nearby source of technology such as a research university. TOP-DOWN APPROACHES: FEDERAL MANUFACTURING TECHNOLOGY PROGRAMS From the 1960s to the 1980s, industrial modernization programs were found only at the state and local level. Federal manufacturing programs provided only limited and uncoordinated support for these efforts and were focused primarily

ANNEX II 207 on basic research and the defense sector (Shapira et al., 1995). In 1988, Congress enacted a mandate for the federal government (through NIST) to assist state in- dustrial development efforts, and, in 1992, the Clinton administration expanded substantially the manufacturing technology efforts that NIST had undertaken up to that date. Manufacturing Extension Partnership The 1988 Trade and Competitiveness Act established at NIST a new MEP program. By 1992, seven Manufacturing Technology Centers (MTCs) had been established under MEP, initially with the goal of transferring advanced technol- ogy from NIST’s Advanced Manufacturing Research Facility and other federal labs. However, it quickly became obvious that what most manufacturing firms needed was proven, off-the-shelf technologies. Therefore, the MTCs shifted their emphasis to helping small and medium-sized firms adopt less-advanced tech- nologies, including “soft” technologies such as training, management, and net- working. As of October 1995, 42 states and Puerto Rico had established or were plan- ning to establish manufacturing extension centers, and over 60 individual centers are currently affiliated with MEP. These centers employ 2,500 agents in over 250 field offices. MEP also operates the State Technology Extension Program (STEP), which provides grants to states to plan and begin manufacturing exten- sion services, although by the end of 1995, only a few states were still without industrial extension programs. To assist states to evaluate and improve their centers, MEP is developing a uniform system of program evaluation. Since the program’s inception, the resources devoted to MEP have increased dramatically. Direct appropriations to the MEP program can be seen in Figure A-2. However, the total funding for extension activities, including MEP and TRP as well as state, local, and private matching funds was over $250 million in 1994, a threefold increase in industrialization funding from just 2 years earlier (Shapira et al., 1995). MEP support for individual manufacturing technology centers is sup- posed to end after 5 years, presumably after the centers have become self-suffi- cient. However, recent evaluations have called into question whether the centers can continue to operate for a longer (perhaps indefinite) period without federal funding. While each center tailors its services to meet the needs dictated by its loca- tion and manufacturing clients, some services are common to most extension centers. These include: • assessment of technology and business needs • definition of needed changes, and • implementation of improvements. Many centers also assist companies with quality programs, employee training, workplace organization, business systems, marketing, and financial issues.

208 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY MEP has created links with a number of affiliated organizations that can provide assistance to manufacturing extension programs or directly to small busi- nesses. For example, MEP and EPA have launched a program to assist smaller manufacturers solve environmental problems before they become regulatory con- cerns. MEP also conducts research to better understand the barriers to modern- ization faced by small manufacturers and to discover additional ways to over- come these barriers. The Technology Reinvestment Project The Technology Reinvestment Project (TRP), created by the 1992 Defense Authorization Act, made a substantial down payment on the U.S. industrial mod- ernization system. TRP had three areas of focus: technology development, tech- nology deployment and diffusion, and manufacturing education and training. In its first 2 years, TRP funded the majority of the manufacturing extension centers that make up the MEP. Since 1994, TRP has ceased providing grants for tech- nology deployment (i.e., extension) activities, but during FY 1993 and FY 1994, TRP awarded $223 million to 95 separate projects. This funding was in addition to MEP direct appropriations (Figure A-2). Required matching funds totaled somewhat more than that figure, meaning that nearly half a billion dollars worth of new technology extension activities were funded in a 2-year period as the result of TRP grants (U.S. Department of Defense, 1995). TRP also funded projects to increase manufacturers’ access to federal tech- nologies, particularly those in federal laboratories. TRP no longer exists, and its remaining activities and reduced funding were recently transferred to the Dual- Use Technology Office of the Office of the Secretary of Defense. The new pro- gram has been redirected toward more military-specific projects. FIGURE A-2 MEP appropriations, including 1995 recision and 1996 continuing resolu- tion. NOTE: 1997 figure is from President’s budget. SOURCE: Unpublished data from National Institute of Standards and Technology.

ANNEX II 209 Federally Sponsored Consortia Federally sponsored consortia include groups such as SEMATECH, a con- sortium of microchip manufacturers that has received funding for about half its operating costs from DARPA, as well as the National Center for Manufacturing Sciences (NCMS), which receives funding from the Defense Department’s MS&T program and elsewhere. Both of these consortia are working to develop new production technologies in different industry sectors. Although federal labo- ratories play an active role in the research of these consortia, the primary federal input is money. Technologies developed by these consortia are transferred to their member firms. Although the members are generally active in planning con- sortia R&D activities and are presumably interested in the results, effectively transferring the results of those activities has proved to be a difficult management issue. Most federally funded consortia are registered under the National Coop- erative Research Act of 1984 (P.L. 98-462), although they represent a small per- centage of all the ventures so registered. The National Cooperative Research Act The National Cooperative Research Act of 1984 grants special antitrust treat- ment to joint research and development ventures and consortia that conduct re- search, analysis, experimentation, or testing. Industrial participants can protect themselves from possible treble damages imposed under antitrust laws by regis- tering with the Department of Justice. In 1993, the act was amended (P.L. 103-42) to add the same antitrust protection to industrial participants in joint production ventures. This legislation has facilitated the formation of several R&D consortia, including the federally funded consortia described above, that are engaged in research related to manufacturing. The NCMS is one example of a consortium whose formation was made possible by the act. TRANSFER OF MANUFACTURING TECHNOLOGY WITHIN THE PRIVATE SECTOR Several mechanisms exist for the transfer of production and manufacturing technologies within the private sector. The most commonly used are: • Vendors and suppliers of manufacturing technology, which are often an excellent source of information and assistance for their clients. Obvi- ously, their interests are not always identical to those of their clients, since they are primarily interested in selling their equipment. Many smaller manufacturers do not know how to evaluate proposals from vendors and have difficulty making an informed choice among many competing sup- pliers. However, if a choice can be made, suppliers are often the source of considerable assistance in plant reorganization, training, and other changes that will let customers take full advantage of new equipment, even though their support tends to diminish with time.

210 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY • Professional and trade associations, which are a potential source of tech- nical information for some firms. Professional associations (particularly engineering associations) have considerable information about advanced manufacturing technologies. For example, the Society of Manufacturing Engineers produces and markets a newspaper, a refereed academic jour- nal, two trade magazines, seven special-interest technical newsletters, hundreds of video-based training programs and reference books, and thou- sands of technical papers. The society also features a manufacturing- oriented library and an on-line electronic information service. However, as noted above, many manufacturing engineers do not belong to profes- sional and trade associations. • Consultants and service firms, which are a major source of new process technology for U.S. manufacturers. Such firms tend to be hired by larger manufacturing companies, and thus their approach to problem solving will tend to reflect solutions appropriate to larger manufacturers. Firms at the smaller end of the spectrum are often not able to afford such consultants, and those that can are often not able to implement the consultant’s recom- mendations without continuing assistance. As noted above, some state- sponsored manufacturing extension programs fund or subsidize private consultants to deliver services. The state of New York spends several million dollars annually on private consultants for small manufacturers (National Research Council, 1993). • Supplier development programs, which have been used by a number of large manufacturers to improve the quality and efficiency of their suppli- ers. Since as much as 80 percent of the cost of products such as airplanes, automobiles, and computers may be purchased from outside suppliers, large manufacturers have a strong interest in the performance of the sup- plier community. Studies indicate that close relationships between sup- pliers and their customers may induce suppliers to adopt more modern technology (National Research Council, 1993). Although supplier devel- opment programs can be an excellent source of assistance in adopting new technologies, they generally do not reach below the first tier of the supplier chain. The nature of these customer-led programs varies widely, from general reviews of supplier progress and advice to comprehensive customer-mandated programs to help suppliers meet mandated quality and other standards. • The American Supplier Institute (ASI) was formed in 1981 to provide training to suppliers to the automotive industry and is now chartered as a nonprofit educational institute in Michigan. Its board of directors includes representatives of major automotive companies and American universi- ties. ASI focuses on quality and management improvement programs, using the ideas of W. Edwards Deming, Genichi Taguchi, and others. It does not provide technical assistance, per se, for manufacturers, but the

ANNEX II 211 quality and management programs generally lead to improvements in all areas of firm activity. Analysis and Trends The organization of manufacturing R&D and extension services in the United States reflects the decentralized nature of America’s governmental structure and of the nation itself. The diverse state and federal programs designed to assist industry have frequently been criticized as uncoordinated and confusing to ac- cess, although states are beginning to establish single points of contact for indus- trial modernization programs. Such actions may improve coordination from the point of view of the company. However, on the federal level, programs are au- thorized, funded, and operated in different areas of the government, and coordi- nation is likely to continue on an informal level at best. Although a model for public-sector manufacturing extension programs ex- isted for nearly a century in the agricultural sector, no pressing need was felt to extend it to manufacturing until global competition began to challenge U.S. eco- nomic performance. Until that time, federal policymakers were not generally concerned by the state of the U.S. manufacturing base, and public-sector indus- trial modernization programs of the time were efforts of individual states to pro- tect their firms against competition from other states. Even now, one might ask why government should not leave the task of modernizing small firms entirely to the private sector and the marketplace. However, as shown in Table A-4, the market has not brought the same rate of utilization of new production technologies to SMEs as it has to large companies. The slow pace of small-firm modernization has an impact on the economy far beyond the community of small manufacturing firms. Lack of modern equip- ment, techniques, and management practices reduces the productivity and quality of small manufacturers and, since these firms account for such a large portion of the value of final products, of the entire manufacturing sector, a critical part of the U.S. economy. Why is the rate of modernization among SMEs slower than among large firms? Primarily it is because there are structural barriers to small-firm modernization that large firms do not face. At the firm level, these barriers include lack of financing; lack of awareness of available proven technologies; fear of change; insufficient time to study and implement changes; lack of skill and training of technical personnel; paradigm shift in new equipment (numerically controlled versus mechanical); inability to per- form comprehensive cost analysis; prior bad experience with new technologies; and inability to select the correct product and vendor (Shapira, 1990). These typical characteristics of small firms are not shared by larger manufacturing companies. Furthermore, the short-term nature of most supplier relationships in the United States and the absence of effective supplier networks is another barrier to modernization faced by small firms, particularly in comparison with similar firms in Germany and Japan (Shapira, 1995). Trade associations are sometimes a source

212 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY of modernization assistance, but in the United States, associations generally fo- cus less on technical information and assistance and more on influencing govern- ment policies. Therefore, in the absence of public-sector industrial modernization activities, equipment vendors are often the sole source of modernization informa- tion for small firms. However, the vendor’s interest (closing the sale) is often quite different from that of their clients. Small firms lack the wherewithal of large firms to recover from an investment mistake, and, particularly for firms that have already made a bad investment, suspicion of the impartiality of vendors and dealers is a critical barrier. Many clients of industrial extension programs, when surveyed, have said that impartial advice about modernization options and avail- able equipment is one of the most valuable services provided by those programs. Do industrial extension services work? Are they a cost-effective way of boosting the competence and competitiveness of small manufacturers? Do they provide what their clients need? There is a large body of anecdotal evidence suggesting the answer to all three questions is yes. The fact that states have begun and continued such modernization services in the face of their own finan- cial problems supports the notion that such programs are successful. However, the amount of systematically collected data and analysis currently is inadequate to prove this. NSF has devoted considerable time and resources to the process of evaluating the university-based centers it supports (including ERCs and I/UCRCs), and a number of centers have been defunded following unsatisfactory evalua- tions. For its part, the MEP has an active program to develop evaluation method- ologies that can be used by state and local extension programs to evaluate their efforts. It includes activities to define, measure, analyze, and report on short- and long-term impacts of MEP centers on the operations of client firms. In 1995, the General Accounting Office (GAO) issued a report based on a survey of clients of 57 extension programs in 34 states that had received at least 40 hours of assistance in 1993. The 551 completed questionnaires covered the most common types of MEP center services. Most manufacturers responding (73 percent) believed that MPE assistance had affected positively their overall busi- ness performance. A minority (15 percent) said the assistance had not affected their business performance, and the rest said it was too early to tell or they had no way to estimate. The majority of respondents said that the impacts of MEP center assistance had a positive effect on their use of technology (63 percent); improved quality of their product (61 percent); and improved productivity of workers (56 percent). About half of the respondents indicated that MEP assistance had a positive impact on their customers’ satisfaction, their profits, and their ability to meet production schedules. In a related survey of small firms that did not use the MEP, most (82 percent) told the GAO they were unaware of MEP services, while another 10 percent were aware of but did not need MEP services. Although federal technology programs (particularly the Advanced Technol- ogy Program and the TRP) have come under fire since the election of the Repub- lican majority in the 104th Congress, the MEP itself has relatively few critics.

ANNEX II 213 The MEP has succeeded in developing strong bipartisan support in states and localities, which have in turn been effective in persuading their federal legislators that the MEP should continue as an active program. While the growth of MEP funding has slowed in the past year, the program has nonetheless enjoyed modest increases in the face of cuts nearly everywhere else, and it has grown substan- tially since its inception. In spite of the apparent success of U.S. industrial extension programs, both in the field and in the political arena, the goal of government-funded manufactur- ing extension programs can and will probably remain modest compared with the absolute numbers of firms in the manufacturing business. Seven years ago, Shapira (1990) asserted that to have a significant effect, state and federal pro- grams should move far beyond assisting a few thousand firms per year (the rate at that time). In fact, MEP-affiliated centers are now providing some level of ser- vice to about 15,000 new manufacturing firms per year. However, even at this rate, another 25 years would be required to reach all the small manufacturing firms in the United States. The National Academy of Engineering (1993) sug- gests that the goal should be to “catalyze the development of a dense national network of public and private providers of industrial modernization services that is capable of meeting the diverse technical, managerial, training, and related needs of 20-25 percent of the nation’s small and medium-sized manufacturing compa- nies by the year 2000.” What rate of industrial extension activity is sufficient to have a significant impact on the technological status of small U.S. manufacturers? As implied in the previous paragraph, it is probably not necessary to reach every small and medium-sized firm. Many firms do not need or do not want any assistance from government. Many others are far from the economic mainstream, in niche or local- ized markets; their competitive position is not threatened by the world economy; or their technological limitations detract little from the competitiveness of the United States. Still other firms belong to private-sector supplier networks or have found private assistance, such as consulting, on their own and are receiving the critical assistance they need. It would be difficult to gauge the sizes of these groups, but together they probably make up a significant part of the U.S. manufacturing sec- tor. The key is to identify the most important sectors, geographical areas, and firms that can benefit most from a manufacturing extension program. MICROELECTRONICS Simon Glynn and William J. Spencer Microelectronics are a vital enabling technology, one that is critical to the U.S. economy. Sales of microelectronics represent nearly 11 percent of U.S. GDP, and the microelectronics industry is one of the fastest growing sectors of the U.S. economy, increasing at a CAGR of 9.3 percent between 1987 and 1994

214 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY (Council on Competitiveness, 1996). Most of the innovations in microelectron- ics were first developed in the United States. The United States has nonetheless faced considerable challenges in microelectronics compared with the software or biotechnology industry, especially from Japan and the newly industrialized econo- mies (NIEs) of Asia. This case study discusses the reasons for the United States’ success in early innovation and current technology transfer in microelectronics intended to counter competitive challenges. Defining Microelectronics This paper focuses on technology transfer in two areas: semiconductors and flat panel displays (FPDs). These two technologies are critical competencies in microelectronics. Semiconductor content in personal computers, for example, increased from $750 to $1,500 between 1989 and 1993. Semiconductor content is also approaching 50 percent of the total cost of new weapons systems for the DOD. The world market for semiconductors increased 32 percent in 1994, to $102 billion (Council on Competitiveness, 1996). The United States has faced considerable global competition in semiconduc- tors. The U.S. global semiconductor market share in 1992 was 43.7 percent, nearly equal to Japan’s 43.4 percent. This lead has widened in recent years, to 46 percent for U.S. semiconductor companies versus 41 percent for Japan in 1995 (Council on Competitiveness, 1996). Flat panel displays are also key enablers of electronic systems and represent an increasing share of the total cost of these systems. For example, FPDs represent 25 to 30 percent of the cost of PCs and over 50 percent of the cost of personal digital assistants (PDAs). The world market for FPDs was $11.5 billion in 1995 and is expected to approach $22 billion by 2000 (Council on Competitiveness, 1996). Japanese firms control 95 percent of the world’s FPD market (Council on Com- petitiveness, 1996). The U.S. FPD industry is very small and fragmented by com- parison. Many large U.S. firms exited the FPD market in the 1980s. In Japan, on the other hand, several large firms invested aggressively in today’s dominant technol- ogy of active-matrix liquid crystal displays (AM-LCDs). In the United States, there are some dozen small and medium-sized firms pursuing a variety of FPD technolo- gies. A few U.S. companies have successfully positioned themselves as materials or equipment suppliers, but there is almost no U.S. presence in the multibillion dollar LCD market (Council on Competitiveness, 1996; Saccocio, 1996). Research and Development SEMICONDUCTORS The U.S. semiconductor industry spent $3.7 billion on R&D in 1994, or 13 percent of revenues. R&D spending in the underlying semiconductor equipment and materials industry is also estimated at about 12 to 15 percent of revenues (Council on Competitiveness, 1996).

ANNEX II 215 Progress in semiconductors also depends—critically—on the experience of implementing advanced technologies in new semiconductor fabrication. This feedback influences the next set of technical goals in semiconductor R&D. Con- sequently, advanced R&D in semiconductors also requires continuous invest- ment in new fabrication (Borrus, 1988). But the capital-intensive nature and ever-increasing complexity of semicon- ductor manufacturing make large investments in R&D quite difficult. For ex- ample, the cost of building a new world-class 16MB DRAM chip fabrication facility, or “fab,” is now about $1 billion (Council on Competitiveness, 1996). As many as 125 new fabrication facilities are now planned or under construction around the world (San Jose Mercury News, 1996). These costs create enormous pressure on levels of R&D spending. “Fab- less” semiconductor R&D companies emerged in the mid-1980s in response to these high capital costs. By going fab-less, these companies can increase their R&D spending to focus on design and testing. Several companies also use smaller, specialized “mini-fabs.” But fab-less companies are increasingly vul- nerable as global semiconductor fabrication capacity tightens. This dynamic is complicated by declining profit margins as each generation of semiconductor technology becomes commoditized. This, in turn, reflects the rapid improvement in price/performance ratio of semiconductors. (“Moore’s Law” predicts that semiconductor performance will double every 18 months, without any increase in price.) In response to these dynamics, leading U.S. semiconductor manufacturers, including Intel, Motorola, and Texas Instruments, are participating in consortia and other cooperative mechanisms to leverage R&D, especially in generic (precompetitive) technologies (Rosenbloom and Spencer, 1996). A generic tech- nology (e.g., superlattice or heterostructures) typically has broad applications. Cooperation to leverage R&D in these new technologies reduces the financial risk for individual competitors. Cooperation also eliminates duplication of R&D (Rosenbloom and Spencer, 1996). This approach to generic R&D largely mimics the Japanese approach to semiconductor R&D (Borrus, 1988). The U.S. and Japanese semiconductor industries are relying increasingly on the equipment industries to help support the R&D and capital costs for new de- vices. Today, semiconductor manufacturing equipment is typically developed through cooperation between individual semiconductor component firms and their equipment suppliers. As a result, several different technologies are used in mate- rials and equipment, and no single industry-wide specification exists. This is a significant problem for the industry (Council on Competitiveness, 1996). FLAT PANEL DISPLAYS R&D spending for flat panels is more difficult to estimate than for semicon- ductors. There are currently no U.S. companies that compete in AM-LCDs and

216 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY other high-volume FPD markets. The U.S. FPD sector consists entirely of a group of small companies pursuing “technology-push” strategies. A large por- tion of the funding for these small companies is derived from federal funding and contracts, especially from the Advanced Research Projects Agency (ARPA) (Saccocio, 1996). FPDs are perhaps the highest-potential, strategically significant competency in microelectronics. Analysts predict the integration of semiconductors directly onto FPDs, for example. In the next decade, new applications will drive the demand for FPDs, including PDAs, virtual reality, and portable computing. As yet, there is no agreement on what FPD technologies will dominate these applications. Future FPD market growth is almost entirely dependent on techni- cal advances in FPDs. For example, FPDs with good contrast and resolution that meet the cost and power requirements for all applications do not yet exist. Con- sequently, different FPD technologies are being developed to address the needs of different applications. U.S. FPD companies lead in developing many of these technologies (Saccocio, 1996). Early Innovation in Semiconductors6 As recently as the 1970s, the United States dominated semiconductor tech- nology and the manufacture of semiconductors. To a considerable extent, this dominance reflected the success of the U.S. economy is exploiting the earliest innovations in semiconductors. This success was shaped by several factors. First, AT&T/Bell Laboratories was extremely important for early innovation in semi- conductors. Bell Laboratories received nearly 350 patents in semiconductors, or more than one-quarter of all semiconductor patents between the time the transis- tor was invented at Bell Labs in 1948 to the time the integrated circuit (IC) was developed at Texas Instruments and Fairchild in 1958. Bell Labs’ rapid dissemi- nation of these results helped develop the semiconductor industry. For example, as early as 1952, AT&T provided licenses to 35 companies under its transistor patents, even as AT&T started to fabricate germanium transistor semiconductors for internal use. Technical symposia were also held by Bell Labs to transfer technology and recent R&D developments—including silicon oxide diffusion and oxide masking, which enabled large-scale semiconductor fabrication—to these licensees. Second, the transfer of individuals also encouraged early innovation in semi- conductors. For example, Shockley, one of the inventors of the transistor at Bell Labs, left to form start-up Shockley Transistor Corporation at Stanford in 1955; researchers recruited by Shockley then left to form Fairchild Semiconductor in 1957. Fairchild itself became a source of new spin-outs, including Intel and AMD. Texas Instruments as well as Motorola recruited Bell Labs researchers. Third, early innovation in semiconductors was directly encouraged by de- fense policy—especially military and aerospace demand for the new technology.

ANNEX II 217 This policy had several aspects. Defense R&D programs in semiconductors in the 1950s and 1960s served as important technology transfer mechanisms linking the vast number of DOD semiconductor programs in the commercial sector. Defense funding of academic research at U.S. universities also contributed di- rectly to early innovation. Start-ups in semiconductors tended to concentrate around these academic programs—notably in Boston (MIT, Harvard), and San Francisco (Berkeley, Stanford). In this way, learning in these federally spon- sored R&D programs was transferred to commercial use. DOD and NASA programs also created demand for advanced semiconductor technologies. For example, the two agencies were responsible for nearly 50 per- cent of revenues from transistor sales in 1960. This early demand was met at very high unit costs. As innovations in processes reduced unit costs, transistor tech- nology extended into commercial uses. For example, from 1963 to 1965, DOD and NASA funded 14 programs that called for the use of ICs, notably the Minute- man II missile and Apollo spacecraft guidance systems. In 1963, these programs represented 94 percent of the market for ICs, at a unit price of $31. In 1965, DOD and NASA procurement represented only 72 percent of demand for ICs as com- mercial use expanded, and unit prices for ICs dropped to less than $9. Fourth, the development of computers, and especially IBM’s development of transistorized computers, was critical to the successful U.S. exploitation of inno- vation in semiconductors. IBM’s enormously successful System 360 was the first computer not based on discrete semiconductor design, but on integrated cir- cuits as well as magnetic tape drives and flexible software architectures that were all developed under government funding and adapted almost immediately for commercial use. IBM and its competitors created enormous demand for new semiconductor technologies, driving up profits and encouraging innovation. Technology Flows Technology flows in the U.S. microelectronics sector depend to an unusual extent on formal relationships between companies, equipment suppliers, univer- sities, and the federal government. CONSORTIA Consortia play an especially important role in semiconductors. Collabora- tion between U.S. semiconductor manufacturers and equipment suppliers has helped the United States compete against the Japanese vertically integrated keiretsu (relationships between Japanese semiconductor manufacturers and original equipment manufacturers). The most important of these consortia are SEMATECH (for Semiconductor Manufacturing Technology) and the Semicon- ductor Research Corporation (SRC). Consortia in FPDs, notably the Microelec- tronics and Computer Technology Corporation, have been largely ineffective.

218 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY SEMATECH SEMATECH is a consortium of U.S. semiconductor manufacturers, govern- ment, and academia. It was formed in 1987 as a cooperative effort between DOD and the semiconductor industry in response to the perceived targeting of the U.S. semiconductor industry by international competitors. Its purpose is to sponsor and conduct research in semiconductor manufacturing technology. SEMATECH’s members include Advanced Micro Devices, AT&T, DEC, Hewlett Packard, IBM, Intel, Motorola, National Semiconductor, and Texas Instruments. SEMATECH’s focus is on developing and diffusing precompetitive manufac- turing technologies and processes. (See, for example, Box 1.) Early SEMATECH programs succeeded so well that U.S. semiconductor manufacturing capabilities and equipment can now be met domestically, except in the critical field of photolithogra- phy. SEMATECH is now developing the next generation of semiconductor manu- facturing technologies needed to create 300-millimeter semiconductor wafers. SEMATECH has invited domestic and foreign firms that have fabrication facilities in the United States to participate in the so-called 300 Millimeter Initiative. Results will then be transferred to consortium members. SEMATECH expects four or five U.S. companies and five or six companies from Europe and Asia to participate (Council on Competitiveness, 1996; Spencer, 1996). SEMATECH annually received $100 million in federal funds during the Reagan and Bush administrations, which it matched with an equivalent amount for a total yearly budget of $200 million. Federal funding was reduced by about $10 million in each of 1994 and 1995 and is expected to be phased out entirely by 1997. The 300 Millimeter Initiative will be funded entirely by consortium mem- bers (Council on Competitiveness, 1996). Semiconductor Research Corporation (SRC) SRC was established in 1982 by the Semiconductor Industry Association (SIA) to plan and execute a program of R&D at U.S. universities in areas of interest to the U.S. semiconductor industry. SRC participants include industry and government agencies. The SRC research program spends about $37 million annually and supports more than half of all silicon-related generic research in U.S. universities. The effect has been to dramatically increase research into silicon-based technology. In 1982, only 20 to 30 graduate students were pursuing silicon-based projects. Now, SRC funds over 350 faculty and 900 graduate students at more than 60 universities. Coordination with SRC member companies has helped faculty focus on the areas of highest commercial potential (Council on Competitiveness, 1996). Microelectronics and Computer Technology Corporation (MCC) MCC was created in 1982 by 10 major U.S. computer and semiconductor manufacturers with the goal of maintaining the U.S. lead in computer technolo-

ANNEX II 219 gies. Members now include large companies (3M, AMD, Andersen Consulting, AT&T, Cadence, Ceridian, DEC, Eastman Kodak, GE, Harris, HP, Honeywell, Lockheed Martin, Motorola, National Semiconductor, Nortel, Rockwell, Westing- house), associates (various companies, government agencies, and academic insti- tutions), small business associates, and university affiliates. Unlike SEMATECH, MCC was privately created. MCC has widely been seen as a failure compared with SEMATECH. More recently, MCC has abandoned the goal of maintaining the U.S. lead in computer technologies and has focused primarily on two areas: a high-volume electronics division that develops packaging, interconnect, and display technology; and an enterprise-integration division dedicated to building a global data highway and networking and database technologies. These initiatives have (so far) had no impact on the U.S. FPD industry. GOVERNMENT-INDUSTRY RELATIONSHIPS SEMATECH is not the only example of government-industry relationships in U.S. microelectronics. Indeed, the federal government has always played an active role in this sector, as noted earlier for semiconductors. Federal spending for FDP R&D over the last 5 years was about $650 million. The DOD has been the largest funder of R&D in the areas of microelectronics seen to be critical for national security. For example, DOD spending represents nearly 90 percent of federal funding for FPD technology. Key FPD initiatives funded by this spend- ing include ARPA’s High Definition Systems and Head Mounted Display pro- grams. In 1994, ARPA awarded a 3-year, $21.4 million grant to Xerox for con- tinued development of its AM-LCD technology (Council on Competitiveness, 1996; Saccocio, 1996). DOD funds have also gone beyond FPD research and development. In 1994, the DOD announced a 5-year, $500 million program to support future domestic FPD manufacturing. In 1993, ARPA and Optical Image Systems (OIS) an- nounced they would build a $100 million LCD plant in the United States to pro- vide displays for military and commercial use. ARPA funding represents about half of the costs for the fabrication facility. Also in 1993, ARPA awarded a multiyear infrastructure development grant to the USDC, including $12 million in the first year (Council on Competitiveness, 1996; Saccocio, 1996). Other federal agencies are also involved in government-industry relation- ships in microelectronics. For example, NIST has been funding research in semi- conductor measurement technology. Many segments of the microelectronics in- dustry are also collaborating with DOE laboratories through CRADAs. For example, Sandia National Laboratories has worked with SEMATECH starting in 1993 under a CRADA that covers research and development in critical areas of contamination-free manufacturing (Council on Competitiveness, 1996).

220 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY BOX 1 Cost of Ownership Technology Transfer The Cost of Ownership (CoO) concept can be traced to the nuclear power industry in the 1960s. It was then that the notion of total system cost was conceived as a way to estimate operating expenses over the life of a nuclear power plant. Unfortunately, the implementation was not successful, in part due to the number of variables and the lack of comput- ing power needed to manage the large number of calculations. CoO, in the form known today, originated at Intel Corp., where it was used to examine the total cost of acquiring, maintaining, and operating purchased equipment. Dean Toombs, an Intel assignee, introduced the concept to SEMATECH. At SEMATECH, the method was first used to evaluate different lithography technologies. Lithography is used to im- age the semiconductor patterns onto a light-sensitive emulsion to pro- duce the patterns that define integrated-circuit performance. The smaller the line widths, the more critical the imaging process. Each layer in the manufacturing of the semiconductor involves lithographic imaging. Today’s semiconductors involve multiple layers and have requirements for a high degree of alignment with very fine resolution. At the time, the two competing lithography technologies were projec- tion imaging and near-contact imaging (stepper). The projection imaging system was relatively low-cost when compared with the stepper process. The apparent throughput rate was also higher with the projection pro- cess. However, the engineers, the operators, and the manufacturing personnel knew that the projection process had a lower yield and re- quired more maintenance. There was a need for a tool that could be employed to permit an accurate evaluation of the process and provide a verifiable method of analyzing the equipment. Decisions to buy equip- ment are often based on purchase price and the cost of installation. These costs do not consider the effect of equipment reliability, production utilization, or product yield. Over the life of a system, these factors may have a greater impact on CoO than the initial purchase and installation costs. CoO was applied to this project to provide an accurate analysis of the life-cycle costs. SEMATECH incorporated and expanded CoO as part of quality and equipment improvement programs. This led to the development of a method for estimating in some detail the total life-cycle cost of owning and operating equipment for a single semiconductor process step. This work was transferred from Toombs to Ross Carnes, a Motorola assignee who continued to refine the method. Joann Trego, a SEMATECH director, was responsible for training users in the correct application of the software. SEMATECH implemented this methodology in a spreadsheet program and distributed it to member and supplier companies. CoO measures the life-cycle costs of equipment improvement or purchase for both sup-

ANNEX II 221 BOX 1—Continued pliers and users. It became possible for suppliers to measure them- selves against their competition. Users could employ CoO to evaluate various supplier equipment in order to determine the best selection for their facility. CoO quickly became widely used by SEMATECH member companies in their purchase decisions. As with any product with signifi- cant market impact, copies began to appear. While there were a number of them, the quality varied and the values could be manipulated by the user. The original CoO locked critical elements of the program and pre- vented values from being changed. In 1991, Daren Dance acquired re- sponsibility for the CoO effort. He guided it through numerous minor revisions and one major revision. The last version that SEMATECH pro- duced was release B. The last release incorporates over 150 param- eters that cover aspects of the equipment from tool usage through con- sumables and maintenance parts and support. The SEMATECH software became a de facto industry standard, but the support diverted resources from other SEMATECH activities and the quality of competing software was suspect; so, SEMATECH guided an effort, with Semiconductor Equipment and Materials International (SEMI), to develop standard definitions and equations for CoO. SEMATECH, through Dance, worked to develop a consensus on the definitions and formulation for a generic CoO standard. The resulting SEMI E35 guide- lines were accepted by worldwide balloting. This helped the situation but did not solve the problem. Due to the popularity of the program, SEMATECH had over 1,200 copies of CoO in use and was spending a significant amount to keep the software updated and member companies trained. There was a need for a commercial supplier to be given the responsibility for maintaining the software and providing customer support. Dance led the SEMATECH effort to find a commercial supplier. A state- ment of work was developed and an open bidding process commenced. This was completed with the selection of a commercial supplier, Wright, Williams, and Kelly (WWK), which incorporated the SEMATECH code into its existing interface and marketed it as TWO COOL. The story does not end with the transfer of support responsibility. WWK developed a marketing strategy that was based on providing com- panies with site licenses that bundled the support costs. SEMATECH member companies, which had been receiving the software and support as part of their return on investment from their dues, now had to agree to pay an acquisition cost of less than $10,000 in order to receive the latest software and the associated support. The transition was not easy. It took WWK almost 6 months to make its first sale; after 1 year, the firm had five of the SEMATECH member companies on board. SOURCE: W. J. Spencer, Chairman and Chief Executive Officer, SEMATECH.

222 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY UNIVERSITY-INDUSTRY RELATIONSHIPS Industry consortia in semiconductors continue to create relationships with leading U.S. research universities. As noted earlier, SRC is an important link between the semiconductor industry and U.S. universities. SRC also administers SEMATECH’s Center of Excellence program. This program directs 5 percent, or $9 million, of SEMATECH’s $180 million budget to R&D in specific areas of interest to semiconductor manufacturers. All but one of these centers is located at U.S. universities. However, there are concerns at the universities that SEMATECH support for academic research may disappear when federal funding is phased out in 1997 (Council on Competitiveness, 1996). In FPDs, ARPA has started several centers of excellence, including the Dis- play Phosphor Center of Excellence at the Georgia Institute of Technology and the National Center of Advanced Information Components Manufacturing lo- cated at Sandia National Laboratory. Other FPD research programs at U.S. uni- versities include the NSF Science and Technology Center on Advanced Liquid Crystalline Optical Materials at Kent State University and research programs at Temple University and the University of Michigan (Saccocio, 1996). Interestingly, most of the important innovations in microelectronics—unlike biotechnology—have derived from industry R&D, not academic research. Uni- versity programs are nonetheless exceedingly important in microelectronics, be- cause the speed and extent of R&D in both semiconductors and FPDs depend on a relatively small number of students trained in these technologies. Demand for experienced scientists and engineers is forcing U.S. semiconductor manufactur- ers to compete for the people necessary to perform desired levels of R&D. U.S. FPD manufacturers are also constrained by relatively few talented scientists and engineers educated in FPD technologies in U.S. research universities. TECHNOLOGY MAPPING IN SEMICONDUCTORS The Semiconductor Industry Association introduced the concept of technol- ogy mapping in the 1980s to guide investment in R&D. This map has evolved into the National Semiconductor Technology Roadmap, which maps technology goals for semiconductors over the next 10 to 20 years. Technology mapping has helped to coordinate research across industry, universities, and federal laborato- ries, although the actual fit between the technology maps and the marketplace is quite weak (Council on Competitiveness, 1996). SMALL FPD COMPANIES AS A SOURCE OF TECHNOLOGY Recently, several large U.S. customers of electronic displays have formed alliances with small U.S. FPD companies to commercialize new FPD technolo- gies. For example, Motorola has invested $20 million in a joint venture with In

ANNEX II 223 Focus Systems to develop a new LCD technology. Rockwell International has formed a strategic relationship with Kopin to develop AM-LCD panels based on Kopin’s Smart Slide technology (itself a result of development with Standish, another small FPD company, and the Sarnoff Research Center). Kopin has also signed a broad product-development effort with Philips North America Corp. Standish is also working with Xerox to manufacture a very-high-resolution AM- LCD (Saccocio, 1996). International Technology Transfer International alliances are increasingly important in microelectronics. The technical challenges of developing the next generation of semiconductors and the enormous costs of fabrication facilities are driving U.S. semiconductor manufac- turers to form technology transfer agreements with foreign competitors. Several notable examples are described below. • Texas Instruments and Japan’s Hitachi formed an alliance in 1988 to share technology related to the production of 16MB DRAMs. The alliance has subsequently evolved into an arrangement to develop 64MB and 256MB DRAMs and to establish a joint-manufacturing arrangement (Council on Competitiveness, 1996). • IBM, Japan’s Toshiba, and Germany’s Siemens have established two joint development agreements (one for 256MB DRAMs in 1992 and another for second-generation 64MB DRAMS in 1994). IBM and Toshiba also collaborate in other areas (Council on Competitiveness, 1996). • In 1992, Intel and Japan’s Sharp joined forces to develop, manufacture, and sell 8/18MB flash memory devices. Each company shares the R&D, and each expects to expand their flash memory market presence (Council on Competitiveness, 1996). • Major alliances in semiconductors are also forming between Japanese companies and the newly industrialized economies in Asia. For example, Japan’s Hitachi and South Korea’s Goldstar are collaborating on the pro- duction of 16MB and 256MB DRAMs (Council on Competitiveness, 1996). These alliances are also forming in the FPD market. Numerous large U.S. defense contractors have devoted considerable effort to develop displays for the DOD. Hughes, for example, successfully produced a large color FPD using LCD technology for the command and control units on Navy ships. Declining defense spending is now forcing these companies to explore commercial uses for this technology. For example, Hughes is reconsidering the commercial FPD market for its liquid crystal light-valve technology developed for the U.S. Navy. In 1992, Hughes formed a joint venture with Japan’s JVC to develop, produce, and market liquid crystal light-valve projectors (Saccocio, 1996).

224 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY Outlook for the Future Demand for microelectronics is expected to increase at a compound annual growth rate (CAGR) of 6 to 8 percent through the 1990s. Moore’s Law is also expected to continue to drive the price-performance ratio in microelectronics, demanding radical advances in semiconductor design and integration of micro- electronics, especially semiconductors and FPDs. In semiconductors, the direc- tion of technology advance is relatively clear (even mapped). Unless leapfrog technologies are developed, the extensive technology flows forced by technologi- cal challenges and very high fabrication costs make it unlikely that the competi- tive advantage from these advances will accrue exclusively to any one company or national economy. In FPDs, the future is less clear. Japan controls virtually all of the FPD market. U.S. FPD manufacturers may continue to lead in developing new FPD technologies, but the direction of technology development in FPDs is not at all clear. Finally, developing the infrastructure, manufacturing, and applications needed to exploit these new FPD technologies appears exceedingly difficult with- out Japanese involvement. SOFTWARE Simon Glynn The United States has excelled in software and computers because of an exceptional ability to develop ideas. A unique convergence of institutions and relationships has created an environment that encourages not only new ideas, but also their development and application. This paper is intended to introduce the various relationships between the public and private sectors that have enabled this. Overview of Software Research and Development Software is now critical to many commercial and defense technologies. In- deed, any product or service enabled by computers depends on “embedded” soft- ware. One informal estimate is that perhaps 70 percent of Hewlett Packard’s development engineering is concerned with software engineering (Mowery, 1996).7 For this reason, it is difficult to communicate the scale of computer and information technologies research and development in the U.S. economy. Spending by the federal agencies for basic and applied research in mathemat- ics and computer science exceeded $900 million in 1991. Of this spending, nearly half was for basic research. Indeed, the rate of increase in funding for basic research in the 1980s has been faster for mathematics and computer science than for any other field—although in absolute terms the base is only 4 percent of federal spending for basic research in 1991 (National Science Board, 1993).

ANNEX II 225 Most of this federal spending has been for academic research. Total expen- ditures for university research in mathematics and computer science were nearly $775 million in 1991 (National Science Board, 1993). Of this, federal spending represented more than $500 million, or about 70 percent of all spending for aca- demic research in mathematics and computer science in 1991. The difference, $240 million, or 30 percent of academic spending in mathematics and computer science, came from nonfederal sources, including industry, state and local pro- grams, and academic institutions (National Science Board, 1993). This apparent division of intellectual effort between universities and industry is illuminated by data on Ph.D.’s employed in computer research and develop- ment. Industry employs a higher percent of Ph.D.’s in computer science (58 percent) than it does Ph.D.’s from any other science field. Nearly 90 percent of these Ph.D.’s are employed in applied research and development. In academia, on the other hand, almost 60 percent of Ph.D.’s are employed in basic computer science research (National Science Board, 1993). In this sense, academia contin- ues to be the locus of basic scientific and technological learning. Two observations deserve special attention with respect to this learning in software. First, the development of the software sector in the United States has been shaped by the very large contribution of federal funding to R&D in comput- ers and software. The United States enjoys a “first-mover” advantage in soft- ware, because it is very difficult to displace a successful first-mover in software and because demand for new computer and software technologies developed first in the United States (Mowery, 1996). These first-mover advantages were created not only by commercial activity, but also by federal funding for research and development and the early development of computer science in U.S. universities (Steinmueller, 1996). Second, the development of the software sector has involved the transfer of learning and technology beyond institutional boundaries. Innovation in comput- ers and in software has depended on the opportunity for individuals to move among academia, the federal labs, and technology-intensive companies, for ex- ample IBM. In this respect, new, technologically innovative software compa- nies—and the environment that encourages them—represent an increasingly im- portant way for individuals to transfer technology. Size and Scope of Software Sectors Spending for software may be divided into two types: prepackaged software (SIC code 7372) and customized software and services (including computer pro- gramming services, SIC 7371; and computer systems integration, SIC 7373). Data on these sectors are presented in Tables A-5 and A-6. Global revenues for customized software and services by U.S. companies in 1993 were estimated to be $38.7 billion and are expected to exceed $40 billion in 1994. More than 40 percent of these revenues are from markets outside the United

226 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY TABLE A-5 Revenue Trends and Forecasts, Customized Software and Services (dollars in billions), 1991–1997 CAGRc CAGRc 1991 1992a 1993b 1994c 1991–1994 1994–1997 Systems integration 16.2 17.7 19.3 20.9 9% 8% Programming services 15.6 17.6 19.4 21.2 11% 9% TOTAL 31.8 35.3 38.7 42.1 10% 9% aRevised. bEstimated. cForecast. NOTE: Totals and percent changes are based on unrounded revenue data. CAGR = compound annual growth rate. . SOURCE: U.S. Department of Commerce (1994). States, where experience in the technologically advanced U.S. market provides a competitive advantage. Several large U.S. players compete in these markets, including Electronic Data Systems (1992 revenues of $7 billion) and consulting firms such as Andersen Consulting and SHL-Systemhouse (Ferne and Quintas, 1991; U.S. Department of Commerce, 1994). More recently, large defense con- tractors, including Boeing and McDonnell Douglas, have entered these markets because of their expertise in developing large systems (National Research Coun- cil, 1992b). Larger computer makers such as IBM are also focusing on systems integration services as their customers migrate from high-end hardware to a com- plex environment of mainframe and distributed desktop systems. Global spending for prepackaged software (including operating systems) was estimated to be $71.9 billion in 1993 (U.S. Department of Commerce, 1994). The United States is by far the largest geographic market for prepackaged soft- ware, representing 45 percent of spending ($32 billion). Japan is second, repre- senting 9.6 percent ($7 billion). The individual markets of western Europe as a group invested $25.7 billion, or 36 percent of global spending in 1993 (U.S. De- partment of Commerce, 1994). U.S. software companies dominate this sector. According to International Data Corporation (IDC), revenues to U.S. companies were nearly $50 billion in 1992, or more than 70 percent of global spending (U.S. Department of Commerce, 1994). Accurate estimates of revenues for PC-based applications software are not widely available; however, they may be estimated from 6-month data published by the Software Publishers Association (U.S. De- partment of Commerce, 1994).8 Using these estimates, spending for PC-based applications software in the United States and Canada totaled more than $6.6 billion in 1993. It is important to understand that internal development is not included in

ANNEX II 227 TABLE A-6 Global Spending for Prepackaged Software Markets, 1991–1997 (dollars in millions) CAGRa CAGRb 1991 1992 1993a 1991–1993 1993–1997 United States 25,330 28,460 32,040 13% 13% Western Europec 21,091 23,850 25,699 11% 10% Japan 5,270 5,967 6,938 15% 19% Canada 1,078 1,188 1,374 13% 10% Australia 941 980 1,094 8% 13% Latin Americad 1,054 1,242 1,471 18% 18% Asiae 584 780 974 29% 21% Other 1,674 1,846 2,094 12% 15% WORLD 57,022 64,313 71,864 12% 13% aEstimated. bForecast. cIncludes Austria, Belgium, Denmark, Finland, France, Germany, Italy, Netherlands, Norway, Spain, Sweden, Switzerland, and the United Kingdom. dIncludes Argentina, Brazil, Chile, Mexico, Venezuela. eIncludes China, Hong Kong, India, Malaysia, Singapore, South Korea, Taiwan, and Thailand. NOTE: CAGR = compound annual growth rate. SOURCE: U.S. Department of Commerce (1994). these estimates, and the majority of software continues to be developed inter- nally. Spending on internally developed software code exceeded $200 billion in 1990, compared with nearly $75 billion in spending for commercially available software (National Research Council, 1990). The preponderance of these efforts is for incremental improvement to existing software, not the development of new systems (Organization for Economic Cooperation and Development, 1985). Factors Shaping Academic Computer and Software Research Research performed in the leading U.S. research universities has effectively defined software as an academic and engineering discipline. The creation of a new academic discipline that is exceedingly instrument-dependent has been shaped by large public-sector investments. Software as a concept did not exist before the development by John von Neumann of the conceptual architecture for computers in 1945. Indeed, even in 1959, there were virtually no formal programs in computer science. Yet, com- puter science is now an academic discipline of substantial intellectual depth: In 1989, U.S. universities produced 531 Ph.D.’s in computer science and 3,860 doc- toral-level researchers. There were 5,239 people with doctorates teaching com- puter science in 1989 (National Research Council, 1992a).

228 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY TABLE A-7 Federal Funding for Computer Science and Engineering Research and All Science and Engineering Research, Fiscal Year 1991 Computer All Science and Science Engineering Research Percentage Research Agency ($ millions) of Total ($ millions) Defense 418.7 62 3,805 National Science Foundation 122.7 18 1,847 National Aeronautics and Space Administration 52.2 8 3,463 Energy 33.3 5 2,963 Commerce 18.4 3 444 Interior 11.4 2 549 Environmental Protection Agency 8.3 1 343 Transportation 6.1 0.9 146 Agency for International Development 3.6 0.5 290 Treasury 1.7 0.3 22 Health and Human Services 1.5 0.2 8,201 Agriculture 1.5 0.2 1,177 Education 0.9 0.1 157 Housing and Urban Development 0.2 0.03 11 Federal Communications Commission 0.1 0.01 2 Other Agenciesa — — 631 TOTAL 680.6 — 24,051 aOther agencies that support some type of basic or applied research but not in computer science. NOTE: All funding in fiscal year 1991 dollars. SOURCE: National Research Council (1992a). Funding for academic computer science is from NSF and DOD’s ARPA, as well as NASA and DOE. These four agencies accounted for 92 percent of fund- ing for computer science research in 1991, including basic and applied research (Table A-7) (National Research Council, 1992a). NSF is the second-largest funder of computer science research, spending $122.7 million in 1991. Almost all of this funding went to universities. Indeed, NSF is largest funder of individuals in academic research in computer science (as opposed to departments or universities). These funds tend to emphasize basic research (National Research Council, 1992a). That said, NSF-funded research has contributed enormously to software development. The development of the BASIC and PASCAL programming languages was funded by NSF, for example, as well as software engineering and early object-oriented languages like CLU. NSF also funds a large computing infrastructure. The most important compo- nents of this infrastructure are the four NSF supercomputing centers and the NSFNET, which supports the Internet.

ANNEX II 229 Academic computer science is also funded by ARPA. Among federal agen- cies, the DOD continues to be the largest funder of computer science research, spending $418.7 million in 1991; typically, about one-third of this is for aca- demic computer science (National Research Council, 1992a). In this context, ARPA has had an extraordinary influence in defining the research agenda for academic computer science. In contrast to NSF support, which is mainly in the form of grants to individuals, ARPA support for academic programs tends to be concentrated among the leading U.S. research universities: Carnegie Mellon, MIT, Stanford, and UC Berkeley (Mowery and Langlois, 1996). The objective of ARPA funding has been to develop a basic research infrastructure in computer science that may be exploited by defense agencies. This infrastructure-building goal incorporates support for education as well as research: In 1990 one-quarter of faculty in the 40 leading U.S. departments of computer science had received their computer science Ph.D. from one of the three major universities supported by ARPA (Carnegie Mellon, MIT, and Stanford) (Mowery and Langlois, 1996). The High Performance Computing and Communications (HPCC) program is currently a large component of this funding for academic computer science. Started in 1992, HPCC is coordinated across all federal agencies, receiving sig- nificant funding from NSF and ARPA. HPCC is defined in the context of spe- cific applications of computing—a series of “grand challenges” in science and engineering, for example modeling global climate change and weather—that can only be solved using powerful computing. The HPCC program areas are high- performance computing systems; advanced software technology and algorithms; networking; and human resources and basic research. Funding for the individual programs is included in Table A-8. If fully funded over the proposed 5 years, the HPCC program will represent about a $2 billion investment, not including baseline spending for 1991 (National Research Council, 1992a). Spin-On Effects from Defense-Related Research and Development As well as funding basic research in U.S. universities, many of the initiatives funded by ARPA to develop new technologies have had surprising “spin-on” effects for commercial use. What is surprising is that these research projects, selected to complement the defense community, have had an enormous impact on the commercial sector. Examples of this research include timesharing, parallel processing, computer-enabled graphics modeling, and artificial intelligence. Perhaps the best example of this is the Internet. The original concept for the Internet may be traced to an ARPA research project on internetworking in the early 1970s. The concept used by the Internet, of distributed computing and communication by a technology called packet-switching, was proposed in the 1960s and developed using ARPA funding. The concept was deployed as ARPANET, a secure communications network for military and university com- puters. Protocols were also developed to enable different networks to connect,

230 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY TABLE A-8 Agency Budgets by HPCC Program Components, FY 1994 Agency HPCS NREN ASTA IITA BRHR TOTAL ARPA 151.8 60.8 58.7 — 71.7 343.0 NSF 34.2 57.6 140.0 36.0 73.2 341.0 DOE 10.9 16.8 75.1 — 21.0 123.8 NASA 20.1 13.2 74.2 12.0 3.5 123.0 NIH 6.5 6.1 26.2 24.0 8.3 71.1 NSA 22.7 11.2 7.6 — 0.2 41.7 NIST 0.3 1.2 0.6 24.0 — 26.1 NOAA — 1.6 10.5 — 0.3 12.4 EPA — 0.7 9.6 — 1.6 11.9 ED — 2.0 — — — 2.0 TOTAL 246.5 171.2 402.5 96.0 179.8 1,096.0 KEY: HPCS = High-Performance Computing Systems; NREN = National Research and Educational Network; ASTA = Advanced Software Technology and Algorithms; IITA = Information Infrastruc- ture Technology and Applications; BRHR = Basic Research and Human Resources. SOURCE: Office of Science and Technology Policy (1994). based on a now widely used protocol, TCP/IP. By the early 1980s, the success of ARPANET caused NSF to fund its own NSFNET using the same technologies. As demand for advanced computing power has accelerated, other local networks have quickly developed, linked by the NSFNET and connected to other sites and networks around the world. These technologies, as well as the protocols and standards, are collectively referred to as the Internet (National Academy of Engi- neering, 1995b). These ARPA initiatives continue to demonstrate marked spin-on effects. In networking, to extend this example, current ARPA initiatives to develop new technologies are concentrated in the HPCC. One of these initiatives, the National Research and Educational Network (NREN), is expected to advance networking technology in two phases. The first phase of NREN is to increase the communi- cation speed of NSFNET from 1.5 million bits per second to 45 million bits per second. The second phase involves research and development on “gigabit test- beds” to develop networking technology that will enable computer networks that can communicate at speeds of 1 billion bits per second (one gigabit). Most of this R&D is expected to be done in close collaboration with larger telecommunica- tions and computer companies to encourage the transfer of these technologies to commercial high-speed data communications networks (Office of Science and Technology Policy, 1994; U.S. Congress, Office of Technology Assessment, 1993). These surprising effects of the ARPA research initiatives illuminate a related point: Research and development are relatively “closer” in computer science

ANNEX II 231 than in other disciplines. U.S. universities in this sense have provided important channels for the dissemination and diffusion of these innovations in software between academia and the defense and civilian research efforts in software. Digi- tal Equipment Corporation (DEC) is an example of the importance of these tech- nology flows. DEC’s founder, Ken Olsen, developed many of his ground break- ing ideas for the minicomputer while working as a research assistant at MIT on Project Whirlwind, a DOD-funded project that was the precursor of a massive programming effort to develop the Semi-Automatic Ground Environment (SAGE) air defense system (Lampe and Rosegrant, 1992). Indeed, some believe that a lack of interchange between military and civilian researchers and engineers weakened British efforts in computers. The Colossus machine built at Bletchley Park during World War II for code breaking, for ex- ample, was never further developed, and some aspects of it are still classified by the British government (Grindley, 1996). The very different situation in the United States enhanced the competitiveness of the U.S. computer and software efforts (Mowery and Langlois, 1996). Mechanisms to Encourage Technology Transfer in Academic Computer Science Other formal mechanisms have also been important in the transfer of tech- nology from the military to the commercial sphere. For example, the federal government influenced the development of early automatic programming tech- niques through its support for information dissemination. The Office of Naval Research (ONR) organized seminars on automatic programming in 1951, 1954, and 1956. These conferences circulated ideas within a developing community of practitioners who did not yet have journals or other formal channels of communi- cation. The ONR also established the Institute for Numerical Analysis at UCLA, which made important contributions to the overall field of computer science (Mowery and Langlois, 1996). Yet another formal mechanism for technology transfer is the Software Engi- neering Institute (SEI) at Carnegie Mellon University. SEI was started by ARPA in 1984. In contrast to the applications-focus of many ARPA initiatives, SEI is intended to encourage the development and dissemination of generic tools and techniques for use in software engineering for defense applications. (For infor- mation on the SEI program, see their Internet home page at http://sei.cmu.org.) Several professional societies have also influenced the development of com- puter science, especially the Association for Computing Machinery (ACM) and the IEEE Computer Society. The publications and conferences of the ACM and IEEE Computer Society are the major channels for dissemination of research and conceptual advances in computer science. The ACM has also shaped the devel- opment of the undergraduate curriculum in computer science (National Research Council, 1992a).

232 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY The Importance of Defense-System Acquisitions The federal government is also a prodigious consumer of information tech- nology. DOD programs to develop new, very complex computer systems had a tremendous influence on the development of U.S. software and computer compe- tencies. Perhaps the most conspicuous example of this is the development of the SAGE air defense system in the 1950s, which far exceeded previous program- ming efforts. SAGE was developed from the Whirlwind project at MIT to coor- dinate the control of radar installations into a national air-defense system. Devel- opment of SAGE was directed by a division of the RAND Corporation, the System Development Corporation (Mowery and Langlois, 1996). By 1955, RAND already employed 25 people, perhaps 10 percent of the programmers in the United States. By 1960, SDC had spun out of RAND and hired more than 800 programmers for developing SAGE. By 1963, SDC and SAGE had seeded the emerging software and computer industry with more than 6,000 individuals from the SAGE development effort (Mowery and Langlois, 1996). Indeed, by 1967, the Air Force had started divestiture proceedings to spin out SDC from the federally funded research and development centers, as compe- tition in software made this status unnecessary. SAGE’s legacy also includes IBM’s development of transistorized comput- ers. In 1955, IBM delivered the XD-1 (patterned after Whirlwind, which also inspired Digital Equipment Corporation and the minicomputer) to serve as the “brain” of SAGE. Critical to its performance was a new memory architecture, called magnetic core memory, that later would appear in IBM’s enormously suc- cessful System 360 computer. Also key were magnetic tape drives and flexible software architectures, all developed under government funding and adapted al- most immediately for commercial use. IBM also recruited Emanuel Piore, head of the Office of Naval Research, as chief scientist, and increased research spend- ing to 35 percent in the 1950s, and to 50 percent by the 1960s and 1970s. By the 1960s, IBM’s computer R&D budget was bigger than the federal government’s. Even as late as 1960, defense spending represented 35 percent of IBM’s research budget (Ferguson and Morris, 1993). More recently, the Software Productivity Consortium (SPC) has performed an analogous (if less dramatic) role in technology transfer to improve U.S. soft- ware and computer competencies. SPC was established by its member compa- nies in 1985, uniting more than half of U.S. aerospace and defense firms in a for- profit consortium. The goal is to develop processes and methods that improve the design and implementation of complex software systems. This includes develop- ing prototypes and technical reports, but not commercial products (Software Pro- ductivity Consortium, 1996). Until a few years ago, this goal was pursued with something of an ivory tower mentality by SPC, without involving consortium members and usually resulting in products that were off the mark. More recently, the SPC approach has emphasized intensive collaboration with members to de-

ANNEX II 233 velop a technical program more closely aligned to members’ needs (Robert K. Carr, consultant in technology transfer, measurement and evaluation, and inter- national technology, unpublished notes, 1993). The resulting program is seen as a useful resource for U.S. organizations to leverage investments in software de- velopment and to evaluate methods and processes. Software Depends Critically on Innovation in Computer Technologies Opportunities in software also depend on innovation in computer technolo- gies. In this sense, the development of networked personal computers and work- stations marks the transition to a profoundly different environment for software development. IBM is currently the world’s largest supplier of software, despite its current difficulties; IBM’s revenues from software (including operating sys- tems) in 1992 were $11.1 billion. Several trends have affected IBM’s thinking about the software side of their business (and by extension, the thinking of other large computer makers). First, software has developed as an opportunity that is quite distinct from computers. Operating systems and enterprise-scale applica- tions software have become very expensive and complex to develop. For ex- ample, IBM’s OS/2 operating system is estimated to have required at least 5 years and 400 programmers and cost as much as $1 billion. In addition, in recent years, independent software companies have pushed advances in several areas of oper- ating systems and applications software. Second, as computer hardware is increasingly commoditized, differentiation is less on the physical performance of the electronics than on the performance of the systems software and the collection of applications software and services available to users. For example, the success of Apple’s Macintosh computer (whose development was mainly in sophisticated operating systems software) depended on the commercial availability of software designed and marketed by start-ups and smaller software companies. Consequently, computer makers have learned to encourage independent software companies to develop applications based on their architectures. These dynamics contribute to a first-mover advantage for U.S. software com- panies. In contrast to customized software and systems integration, the personal computer and, more recently, networked computing, are radically changing the demand for software by creating very high-volume markets. Indeed, by 1984, the installed base of PCs was 23 million machines, compared with less than 200,000 for large- and medium-sized systems (Steinmueller, 1996). These high-volume opportunities easily absorb the fixed costs of software development. Also, stan- dardization of personal computer architectures in the United States has enabled software companies to create software and operating systems that can be incorpo- rated by different computer makers. This is in marked contrast to Japan, for example, where 6 of the top 10 software companies are tied through industrial groups to different computer makers (the top four are NEC, IBM Japan, Hitachi,

234 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY and Fujitsu), each using a proprietary operating system. In the United States, by comparison, none of the top 10 software companies is tied to a computer maker (Friedland, 1993). These trends reflect the divergence within the overall U.S. software industry between commercial and military applications. The DOD focus on systems de- velopment and on embedded systems doubtless limits the spin-on effects of these technologies for commercial use. The concept of software engineering continues to be relevant to the creation of large-scale, complex defense software systems, especially for embedded applications. But many of these systems are irrelevant to the commercial sector. Small Companies Exploit These New Opportunities In sharp contrast to the computer makers, most of the new independent soft- ware firms are relatively small, entrepreneurial companies. In Utah’s “software valley,” for example, three-quarters of the more than 1,120 technology-intensive companies have fewer than 25 employees, and 50 percent have revenues of $200,000 or less (The Economist, 1994). Several characteristics shape the oppor- tunities for these new software companies. First, the initial capital requirements to start in software are extremely low (with the exception, of course, of intellec- tual capital). This is likely to be as true in the future as it has been in the past. These extremely low barriers to entry, especially in the decentralized, software- intensive low end of the hardware spectrum, limit the amount of risk a software entrepreneur must accept. Second, these emerging software companies often exploit technologies or markets deemed too small or too risky for established players. Thus, new mar- kets and narrow, niche markets that sometimes lead to considerably larger mar- kets let new software companies develop the revenue stream, product, and core competencies of valuable new businesses. On the other hand, once such compa- nies are established in a market niche, they in turn become vulnerable to new players with a better idea. Since the development cycle of sophisticated software is lengthy and requires highly focused skills, reacting to a competitive threat is usually not an easy task. As a result, software companies tend to be divided into three groups. The first group, quite rare, consists of the few that become large and develop the internal resources to have long-term staying power and to stay on the advancing technology curve. The overwhelming majority of start-ups in soft- ware are in the second group, which develops niche-market products, with com- pany revenues in the $5 million to $15 million range. The life cycle of these companies is also quite short. Typically, they will either fail when their product life cycle has run its course or be acquired by or merged with other players to reach sustaining capabilities. The third group includes those new software com- panies that, for a variety of reasons, are not successful and fail.

ANNEX II 235 Economic and Technological Risk is Encouraged The dynamics of these opportunities for new companies in software are very appealing to venture capital. In 1992, software and related services attracted more venture capital financing than any other sector of the economy, including biotechnology. Some 214 software and services companies received 22 percent, or $562 million, of venture capital invested in 1992. As new software companies demonstrate the viability of new technologies or markets, the risk is less and these opportunities then become valuable to larger companies, creating liquidity by acquisition. Compared with other sectors, the valuations are also relatively high, encouraging the formation of new companies. For example, Microsoft’s bid to acquire Intuit for $1.5 billion represented a breathtaking 40 percent pre- mium over Intel’s (then current) market value. The Importance of Large, Technology-Intensive Companies These innovative new software companies tend to be distributed according to a specific geography. That is, technology-intensive communities, for example Boston or San Francisco, that have reached critical mass in software tend to be self-perpetuating. This is because relationships with other, larger technology companies are very important to small software companies. First, the software industry is marked by a large number of spin-off companies or by entrepreneurs leaving larger software companies (or hardware companies) to create their own companies. These new software companies tend to concentrate in areas that in- clude larger, technology-based companies where such spin-offs are common. DEC, for example, has spawned numerous spin-offs. The spin-offs are less well documented in software than in other high-tech fields but occur equally frequent (if not more so). Lotus Development, for example, spun off at least three new firms during its first 3 years, including Iris Associates (which developed the very successful Notes program using venture capital from Lotus). As well as providing a source of entrepreneurs, large, technology-based com- panies also provide a critical base of new technology. There is a broad consensus that concepts are best transferred by the individuals who understand the new technology. To this end, small start-up firms have been responsible in software for an overwhelmingly large share of new commercial applications, often ex- ploiting research and ideas developed elsewhere—usually in universities or in large, technology-based companies. The laboratories of IBM and AT&T Bell Labs especially, and also Xerox PARC, have developed software technologies that have been successfully commercialized by new software companies. Unresolved Policy Questions Several unresolved policy questions shape opportunities for companies in software. Concern about the domination of IBM’s extraordinarily successful

236 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY System/360 architecture led the U.S. Justice Department to assert that this suc- cess represented an illegal monopolization. In response, IBM decided to “un- bundle” the pricing of its systems instead of including the software in the pricing of the computer, essentially creating the opportunity for Microsoft and other in- dependent software companies to sell competing software. Recently, similar questions have been raised in connection to the extraordinary successes of Microsoft Corp. in personal computer software. Uncertain intellectual property rights (IPRs) are a second problem. Existing mechanisms for securing IPRs assume that something is either an expression of ideas (in which case, the expression of these ideas may be protected by copyright law, but not the ideas themselves), or a patentable process (in which case it may be protected by patent law). But software is both an expression of ideas, as lines of code, and the process that the algorithm describes—and that process is valu- able. For this reason, IPRs are an imperfect mechanism (at best) for protecting innovations in software. Intellectual property rights may disadvantage start-ups and smaller software companies. Patents, especially, present special problems in software. Many soft- ware companies are using patents to compensate for recent legal decisions deny- ing them the copyright protection they feel they need. But patents are costly to obtain and difficult to enforce and defend. Large companies are, consequently, more likely to be able to threaten litigation and to defend against litigation. There is also ambiguity about what is and is not patentable. These problems have con- sequences for innovation, because small companies and start-ups are disadvan- taged by the costs and uncertainties of litigation. Also, because larger companies and universities are usually the sources of the technology for spin-offs and smaller companies in software, stronger IPRs for software may actually impede innova- tion as patent portfolios grow but their value remains ill-defined. Remarks on the Future The United States will continue to lead in developing new technologies and markets in software. Most of this innovation will be centered in small software companies. (For large companies, nurturing creativity and innovation has often proved difficult, and the risk-reward equation dictating product development is typically very demanding [Hooper, 1993].) As a result, successful software start- ups will continue to spin out of larger companies led by entrepreneurs with a riskier agenda. These dynamics create an advantage for the United States in software. But even as new technologies present opportunity for new entrants, many of these smaller companies may not have the resources to adopt these innovations and remain competitive. As software programs (including prepackaged software) have become larger and more complex, software developers have started to run into problems of quality and reliability, referred to in the literature as the “soft-

ANNEX II 237 ware bottleneck.” Major delays in product releases, including 1-2-3 (Lotus), dBase IV (Ashton-Tate), OS/2 (IBM), and Windows NT (Microsoft) are examples of this trend (Brandt, 1991). Attempts to address these problems include efforts to replace the current approach to software development with a more rigorous one, using code re-use and object-oriented designs (Brandt, 1991; Ferguson and Morris, 1993). Tech- nology transfer in these new software technologies (if real) may present an op- portunity for other countries to compete with the United States in software. Japan and Europe, especially, have put a premium on developing process innovations in software design and automation, although they have not yet realized any com- mercial advantage from these initiatives. ELECTRIC POWER RESEARCH INSTITUTE: THE BOILER TUBE FAILURE REDUCTION PROGRAM Jim Oggerino Background The Electric Power Research Institute (EPRI) has been the centralized R&D arm of the U.S. electricity industry since 1972. Its members include over 600 utilities that together provide about 70 percent of the electricity generated in the United States. EPRI manages research projects performed by contractors, in- cluding universities and large and small companies. Typically, there are over 1,000 projects in process at any given time, supported with an annual budget in excess of $400 million. Over its 23-year history, EPRI has developed many technology transfer methods and processes. This case study focuses on boiler tube failures (BTFs) in fossil-fueled power plants and the technology transfer process used to ensure that solutions to the BTF problem reach the electric-power industry. The Issue Roughly 40 percent of the energy consumed in the United States today is in the form of electricity, and in the next 50 years that value could grow to 60 percent. Given the extent of public reliance on electric power, any technical problem or flaw that affects the availability of the boilers that make steam to drive steam-turbine generators is serious. Historically, BTFs represented the largest single source of lost generation in fossil-fueled power plants in the na- tion. According to the North American Electric Reliability Council Generating Availability Data System (NERC-GADS), coal-fired units 200 megawatts (MW) or larger experienced more than 15,000 boiler tube failures during the 6-year period from 1983 to 1989. These failures represented a loss of 81 million mega-

238 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY watt hours (MWh) per year. At an estimated replacement power cost of $10 per MWh, this represented an annual loss of $810 million (McNaughton and Dooley, 1995). Thus, the objective of EPRI’s BTF-reduction project was to improve boiler availability at more than 800 EPRI member generating facilities, comprising about 2,000 generating units. At the time, there were 22 known mechanisms of BTF. Solutions for some of those mechanisms were available in other countries, but most had not yet been adopted in the United States. The challenge was to per- form R&D on the as-yet-unsolved failure mechanisms, and to transfer knowledge about BTF solutions to EPRI members. Previous experience indicated that one could not effect technology transfer only by granting users access to the technical information. Face-to-face assistance, organizational and operational changes, and management commitment were required to succeed with a technology transfer challenge of this magnitude. EPRI contracted with General Physics Corp. (GPC) to carry out the research on the unresolved failure mechanisms and to assist with the technology transfer program. Barry Dooley, the EPRI project manager, had come to EPRI from Ontario Hydro (a Canadian utility with a large research department), where he had done considerable work in this area. Dr. Dooley is considered a world-class expert which was, and is, an important technology transfer factor. The contract with General Physics Corp. consisted of cost-plus remuneration and was typical for EPRI at the time. In addition to EPRI and General Physics, the utilities that provided their power plants for technology demonstration were part of the solu- tion teams that were formed. At the time EPRI contracted with GPC, the provisions for intellectual prop- erty were that all rights were to be retained by EPRI. However, in a large number of cases, the GPC investigators were given exclusive, or sometimes nonexclu- sive, licenses to sell the resulting research products in any market except the U.S. utility market. Utilities receive EPRI results free through their membership. Depending on the circumstances, technology licensing for EPRI members may be cost free or require the payment of a considerable fee. EPRI is funded by its member utilities to perform collaborative R&D and be involved in the transfer of research results to members. EPRI generally funds its research projects at the laboratory investigation level. However, at the full-scale demonstration phase, it is often necessary to use a member’s generating station. It takes the highest level of trust on the part of a member utility to use an operat- ing plant as a test bed, because any plant unavailability can result in costs of hundreds of thousands of dollars per day. One technology transfer mechanism used to obtain sponsors for demonstra- tion or shared R&D projects is a one-page document called a “host utility.” This document is distributed through EPRI’s Technical Interest Profile (TIP) system, which member utility staff join by submitting a TIP interest sheet. The interest sheet has roughly 100 technical areas to choose from. Most TIP users check three

ANNEX II 239 or four items and receive routine mailings on the topics of their choice. Mailing lists in each technical area contain between 2,000 and 7,000 names. To initiate the BTF demonstration program, EPRI distributed a host utility document throughout the industry. Based on the responses it received, EPRI selected 10 utilities, representing about 40,000 MW of capacity, to begin the program. About a year after the project started, EPRI held workshops and semi- nars to announce interim progress to the rest of the industry. This resulted in many additional utilities volunteering to become part of the program, and so EPRI issued another host utility invitation. Two years after the first set of 10, 6 addi- tional utilities (another 20,000 MW of capacity) were added to the project. Prior to inviting utilities to participate, EPRI established criteria for partici- pation. Each utility had to assign a BTF program coordinator for the project; issue a corporate BTF program mandate or philosophy statement; include in the program all fossil-generating units operated by the utility; and form cross cutting BTF program teams for which training attendance was mandatory. After EPRI selected the first 10 utilities, it convened a meeting of BTF coor- dinators. All were asked to obtain senior management signatures on the corpo- rate mandates prior to beginning activities at their utilities. EPRI then held train- ing sessions at each utility for each BTF team, consisting of staff from the utility’s engineering, operations, maintenance, and management units. The senior man- ager who signed the mandate had to attend the course for at least one hour. These meetings were held at home offices or at various power stations, with the selec- tion left up to the sponsoring utility. The training material included descriptions of what actions and organizational and operational changes were required to ad- dress each of the 22 failure mechanisms. Six-month follow-up meetings were held by the EPRI team to determine if corrections or additional changes were required. It should be noted that GPC carried out essentially all of the training sessions for EPRI. GPC played a major role in the technology transfer process, freeing the EPRI project manager to focus on the R&D portion of the project. From the outset in 1985, the target for the project was to transfer technology to achieve an average equivalent availability loss (EAL) of 1.45 percent from BTFs. This represents a nearly 60-percent reduction from the national EAL of 3.4 percent in 1985. Figure A-3 shows the EAL reductions achieved from 1985 through 1991. By 1987, the first group of 10 utilities had reduced their EAL from 2.5 percent to 1.8 percent. By 1991, that same group had further reduced their EAL to 1.5 percent. The 10 utilities predict savings of at least $41 million annu- ally for the next 10 years. The second group, which had started 2 years later and at a much higher EAL (3.4 percent) had reduced their average to 2.0 percent. This project demonstrated to the electric-power industry that successful tech- nology transfer does not consist solely of being exposed to research results or technical fixes. Management commitment to support tech transfer programs and, in most cases, organizational rearrangements, also are necessary. In addition, operational changes and training programs are often required.

240 TECHNOLOGY TRANSFER SYSTEMS IN THE UNITED STATES AND GERMANY 3.8 Percent Equivalent Availability Loss BTFRP (6 utilities) 3.4 3.0 National Average 2.6 2.2 BTFRP (10 utilities) 1.8 1.4 Target (1.45 percent) 1.0 1985 1986 1987 1988 1989 1990 1991 1992 BTFRP = boiler tube failure reduction program FIGURE A-3 Equivalent availability loss due to boiler tube failure, 1985–1992. SOURCE: McNaughton and Dooley (1995). The demonstration program and subsequent adoption of BTF solutions by EPRI member utilities led to annual savings in the hundreds of millions of dol- lars. The technology transfer process was so successful it is being used for two other major EPRI programs: Plant Life Extension and Cycle Chemistry. Perhaps more important than resolving the 22 BTF mechanisms (since ex- panded to more then 30), the demonstration project has resulted in utility man- agement recognizing more fully the need to support internal product champions with funding and organizational clout. Thus, three elements—research results, a demonstration host site, and senior management support—were all required to achieve success.

Next: Part III: Technology Transfer in Germany »
Technology Transfer Systems in the United States and Germany: Lessons and Perspectives Get This Book
×
Buy Paperback | $80.00 Buy Ebook | $64.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

This book explores major similarities and differences in the structure, conduct, and performance of the national technology transfer systems of Germany and the United States. It maps the technology transfer landscape in each country in detail, uses case studies to examine the dynamics of technology transfer in four major technology areas, and identifies areas and opportunities for further mutual learning between the two national systems.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!