Click for next page ( 11


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 10
R&D Consortia in the United States and Japan RATIONALE In both the United States and Japan, policy support for consortia activity was originally prompted by a desire to strengthen domestic companies in industries that needed to catch up technologically with international rivals. Yet the timing and mechanisms of support differ greatly. Because U.S. support for consortia was partly a response to Japan's perceived success, an interesting question is whether the underlying framework for understanding innovation and competition is similar to Japan's. The traditional U.S. view is that, in the relationship between invention and economic growth, proprietary information drives the capitalist engine. The danger, according to some, of cooperation among firms is that it can lead to reduced investment and slower innovation, due to a closing out of multiple paths to invention. In contrast, one of the hallmarks of Japanese industrial policy has been an emphasis on avoiding the perceived disadvan- tages of duplication of effort and "excessive" competition. Here the view is that there can be unnecessary as well as better mousetraps.~ This thread is clear in Japan's antitrust and industrial policies. It is important to note, as shown in Table 1, that in Japan the number of consortia actually formed as research associations was small until the early 1970s. This was a period when many of the formal barriers to foreign a Remarks by Richard J. Samuels, Workshop on R&D Consortia and U.S.-Japan Collaboration. 10

OCR for page 10
11 penetration of Japan's economy through trade and investment were lowered or eliminated. The increased emphasis on consortia as a tool to promote technology development and diffusion coincided with a diminution of MITI's formal powers over Japanese industry, including the authority to force foreign firms to license technology in return for market access. In contrast to MITI, Japanese firms have not always been enthusiastic participants in consortia, at least before the fact. When the VLSI Program was organized in 1975, the companies themselves were nervous about par- ticipating and could not see the rationale for the projector While skeptics point to the gap between the stated goals and the eventual technical out- put of specific projects, others believe that the most important benefits for firms are peripheral to the actual research results achieved in the colla- borative setting. Those with a more positive view of Japanese R&D consortia see other types of benefits coming from development of a shared "technological vi- sion" at a precompetitive stage and the training of young technical person- nel. According to Gerald Hane of MIT, who has studied the management of the International Superconductivity Technology Center (ISTEC), which includes several foreign associate members, the largest group of research- ers at the central lab in terms of age was the 25 to 30-year-old cohort. (See Figure 2.) The consortia experience establishes a jimmyaku, or per- sonal network, that can be used for the informal exchange of information throughout a researcher's career. The seeming specialization of the elec- tronics companies in some low-demand products in semiconductors, a "di- vision of labor" approach that some observers have detected in current Japanese corporate R&D on superconductivity, may also be traced to infor- mal contact between companies at an early stage. The relative lack of mobility in Japan for technical researchers may be another reason why R&D consortia in Japan are seen as mechanisms for technology diffusion, while the movement of individuals from company to company in the United States is more common. Japanese R&D consortia are meant to complement rather than substitute for in-house efforts. Even firms that send the largest numbers of research- ers to work in the ISTEC R&D consortia retain more than half of their researchers at their home laboratories.~3 While participation in consortia is ~ Yoshio Nishi, from his presentation on "The Environment for R&D Consortia in Japan" at the Workshop on R&D Consortia and U.S.-Japan Collaboration. ]2 Gerald Hane, Harvard University Kennedy School of Government, doctoral dissertation, in progress. These young researchers are mostly graduates of Japan's premier research universities. 13 mid.

OCR for page 10
12 4g 35 30 In `, 25 ce a' o at 20 15 10 5 a / / / / / / - \ <25 26-30 31-35 3640 41-45 Age Cohort FIGURE 2 Age cohorts of researchers at ISTEC. Source: Gerald Hane part of a strategy to build "core" technology for the future of the firm's business, technological breakthroughs are expected to occur at the corporate laboratory, not at the R&D consortia. This notion of two tracks is central to the rationale for Japanese consortia.~4 In the United States, where university-based consortia are the norm, a primary motivation has been to monitor peripheral developments in R&D and to recruit new R&D talented In Japan, where industry-based consortia ]4 For a discussion of Japan's successive electronics R&D consortia and their evolution toward more generic research, see National Research Council, Office of Japan Affairs, Learning the R&D System: National Laboratories and Other Non-Academic, Non-Industrial Organizations in Japan and the United States (Washington, D.C.: National Academy Press, 1990), p. 30. Most of the R&D in the early projects was camed out at the separate laboratories of the . . . . part~clpatmg companies. ~5 For a discussion of university-based consortia in the United States and examples of Japanese industry-based consortia, see National Research Council, Office of Japan Affairs, Learning the R&D System: Industrial R&D in Japan and the United States (Washington, D.C.: National Academy Press, 1990), p. 14.

OCR for page 10
13 are the norm and university research is weaker, there is growing interest to promote cooperation between industry and universities. With regard to industry-based R&D consortia, in industries like electron- ics, U.S. observers say firms are being driven to consortia, not drawn to them.~7 The rationale is that consortia can take risks that individual firms cannot. They can also leverage financial and human resources, provide technology at a lower price than in-house efforts can achieve, and provide firms with access to technology that they would not have had otherwise. Participants determine what is considered "generic" technology, and col- laborai~on in precompetii~ve areas leads to wider access and heightens competition as companies strive for differentiation in their in-house efforts. Participating in an R&D consortium can also facilitate the exchange of information concerning general market needs and other information that is not necessarily technology related. Competition, leveraged buyouts, and the quarterly earnings-per-share men- tality of corporate management are forcing U.S. companies to put more of their R&D resources into distributed research and product development. Grant Dove, chairman and CEO of MCC, notes that U.S. firms are making alarming cuts in R&D with a 4- to 10-year time horizon in order to pay for more distributed research. In a context where the central laboratory is something of an endangered species, consortia can fill a gap.~9 U.S. indus- trial R&D consortia such as MCC, through "strategic partnerships" with members, associates, and academia, can also play a role in taking university research downstream toward the market. A parallel in the Japanese context is the Japan Research and Development Corporation, which was established in the 1960s to commercialize new discoveries at universities. ]6 In its fiscal 1991 budget proposal, the Ministry of Education, Science, and Culture asked for an appropriation of 58.9 billion yen for its Scientific Research Grant program, and expected corporations to contribute 57.9 billion yen to the Industry-University Collaborative Research program. See "Minkan Shikin Izon Tsuyomaru" (Dependence on Private Sector Funding Grows), Nihon Keizai Shimbun, December 29, 1990, p. 11. ]7 Robert Falstad, from his presentation on "The Environment for R&D Consortia in the United States" at the Workshop on R&D Consortia and U.S.-Japan Collaboration. 8 Ibid. ~9 Grant Dove, from his presentation on "The Environment for R&D Consortia in the United States" at the Workshop on R&D Consortia and U.S.-Japan Collaboration.

OCR for page 10
14 ORGANIZATIONAL MODES In Japan, as elsewhere, most industrial research still takes place within individual corporate labs, and most collaborative research is undertaken by two firms to achieve precise technical goals aimed at the commercialization of technology.20 But in R&D consortia Hat include a number of firms, the government of Japan, especially MITI, has taken a lead role in organizing and funding projects.2i The expectation of stable funding of consortia by MITI permits Japanese companies to incorporate consortia activity into their long-term strategies. Government support sometimes takes the form of catalyzing work in a new area, through direct subsidies or, more commonly, conditional loans that are paid back at the commercialization stage. While government leadership often takes place behind the scenes, it has been essential to the success of R&D consortia in Japan.22 (See Table 2.) Several aspects of Japan's R&D system are favorable to the formation and operation of R&D consortia under the auspices of the government. People with technical backgrounds at MITI, the universities, and industry have similar educational backgrounds, and they interact as equals. Univer- sity professors act as neutral brokers, chairing meetings of industry research executives, assisting MITI technocrats in coordination, and serving on re- view boards. Japanese oligopolies have been seen to have a legitimate policy role. The number of major firms in the electronics industry, for example, is limited, and consortia can be organized to include participation by all major players. The unique role of Nippon Telephone and Telegragh (NTT) in collaborative research in the electronics industry also deserves note. NTT traditionally provided a common ground for industry to share precompe- titive results, to develop a common technical focus, and to develop proto- types. In recent years, however, there have been efforts to private NTT. While horizontally structured R&D consortia (among firms competing in 20 The Fair Trade Commission of Japan conducted a survey of collaborative R&D in manufacuturing and found that the overwhelming majority involved collaboration among companies that were not competitors in straight company-to-company deals, while research associations accounted for only about 5.5 percent of the cases. 2\ M. Fransman, op. cit., stresses the role of the government of Japan in reducing the transaction costs of cooperative research (p. 280). The author also notes that while Japan's government is responsible for a smaller share of national R&D expenditure than is the U.S. govemment, in some cases MITI has provided as much as 20 percent of the research expenditures to develop a core technology over a 10-year period. 22 Comments by Michael Mintz, Panel on the Environment for R&D Consortia in Japan.

OCR for page 10
15 TABLE 2 Japanese Government-Sponsored Consortia Name: Pattem Information Processing System Research Association Years: March 1977 to May 1982 (originally projected to last 8 years) Members: Toshiba, NEC, Hitachi, Fujitsu, Mitsubishi Electric Goal: To develop a prototype system that could recognize written symbols, dia grams, shapes of objects, and sounds. Budget: 22.5 billion yen-$81 million at the 1977 rate of 269 yen/dollar over the life of the project. Of the total, 22 billion yen was funded by the govern ment. The remaining 500 million yen was raised through assessments from the member firms. Organization: MITI-managed research, with the Agency for Industrial Science and Technol ogy's Electrotechnical Laboratory responsible for basic research. The mem ber companies were responsible for developing the basic devices, the infor mation management system, and pattern recognition method. The research association itself was responsible for constructing and testing the prototype. Results: The system prototype incorporated many "futuristic technologies" and was awarded Nikkan Kogyo Shinbun's 10th Prime Minister's Prize for Industrial Technology. Name: Intemational Fuzzy Engineering Research Lab Years: Projected to run for 6 years from March 1989 Members: NTT Data, Kao, Kayaba, Kawasaki Steel, Canon, Shimizu Construction, Sony, Takenaka Engineering, Tokyo Electric Power, Toshiba, Thomson Japan, Toyota, Japan IBM, NEC, Japan Electric Computer Company, Hitachi, Fuji Heavy Industries, Fuji Xerox, Fujitsu, Honda, Matsushita, Mitsubishi Electric, Mitsubishi Chemical, Minolta, Yamaichi Securities, Ricoh, and 19 others Goal: Following the maxim that "technology is for the benefit of humanity," the lab will apply "fuzzy theory" to developing "friendly" human-machine interfaces through comprehensive basic and applied research. Possible applications include "home robots" that can understand voice commands, health care, financial services, and systems to aid human decision making. Budget: A total of 5 billion yen 35.7 million at 140 yen/dollar- is projected for 6 years. During 1989, 521 million yen ($3.7 million) was spent, with 216 billion from government and the remainder from member firms. Organization: Twenty-five researchers from member companies are divided into three groups at the lab. Source: Kokagyo Gijutsu Kenkyu Kumiai Sanjunen no Ayumi (Thirty-Year History of the Research System for the Development of Mining and Manufacturing) (Tokyo: Kokogyo Gijutsu Kenkyu Kumiai Kondankai, 1991).

OCR for page 10
16 the same area) attract much of the attention given to Japan's collaborative research, it is important to remember that 80 percent of collaborative R&D in Japan is vertically structured, involving firms at different stages of the production process. Contrary to what might be expected, the keiretsu sys- tem of corporate linkages does not play a large part in determining which firms enter vertical consortia. The predominant style of R&D collaboration in Japan involves unrelated firms, but firms that are not competing directly in the same market segment.23 Some features of Japanese consortia have evolved over time. The elec- tronics industry consortia are good examples. The VLSI Program had a central laboratory, but most of the work was done by individual companies. Foreign visitors were not allowed, let alone invited to corporate member- ship. In the Future Electron Device Program, the follow-on to the VLSI project, the companies were busy with in-house semiconductor activity and the project was based on coordinated in-house research.24 While foreigners were invited to review sessions for the Future Electron Device project, ISTEC represents the "new wave" of Japanese consortia. There is a central ISTEC facility and two levels of membership. In contrast to projects of the past, where the "industry" consisted of the five or six largest firms, there is now a much wider range of industries. Banks and other service industries are included as well as manufacturers. ISTEC also has several foreign associate members. In the United States, MCC and SEMATECH are both structured honzon- tally but also take account of vertical relationships. MCC does some con- tract research for the government, while government support constitutes half of SEMATECH's budget. In contrast to Japanese consortia organized by the government, most of which operate for a specified time period and are then replaced by a successor project or transformed into an industry association, many U.S. consortia are designed to work in the specified tech- nical area indefinitely. In contrast to most Japanese consortia, which normally utilize research- ers from the member companies, MCC employs its own researchers and staff in addition to accepting researchers from member companies. At the 23 See Jonah D. Levy and Richard J. Samuels, "Institutions and Innovation: Research Collabora- tion as Technology Strategy in Japan," MIT Japan Program Working Paper 89-02, 1989, p. 10. In ISTEC, two-firm patent applications usually involve firms from the same keiretsu, those involving three or more firms usually include firms from different keiretsu. 24 Yoshio Nishi, from his presentation on "The Environment for R&D Consortia in Japan" at the Workshop on R&D Consortia and U.S.-Japan Collaboration. M. Fransman, op.cit., documents the pattern of continued "coordinated in-house" research in Japan and the relative absence of "joint research facilities" in Japan.

OCR for page 10
17 outset there were four large MCC R&D programs open only to sharehold- ers. (Membership is drawn from a number of industry groups, including computers, telecommunications, aerospace, and semiconductors.) Initially, members were required to pay the fee for the entire program even if they were interested in only a few specific projects. Several years ago MCC "unbundled" its programs and projects and began distinguishing between core efforts and satellite projects. Members can now pick and choose the projects they want to fund, thereby utilizing resources more efficiently. Participation has also been opened to nonshareholders. MCC has launched an associates' program, recognizing the importance of technology trans- fer to smaller firms for establishing and maintaining a vertical infrastructure in the computer and semiconductor industries. The evolution in MCC's structure is related to a concurrent evolution in its technical goals. The consortium was launched as a response to Japan's Fifth Generation Computer Project, whose research was conducted by the Institute for New Generation Computer Technology (ICOT). ICOT had as its goal the development of a "thinking computer," and initially MCC adopt- ed a similar technology targeting strategy.25 Before long, however, it became clear that the research focus on technological breakthroughs at MCC did not match well with the technical agendas of the member compa- nies. Over time the emphasis of MCC's work has shifted from "revolution- ary" to "evolutionary" advances. Though it still seeks to make significant breakthroughs and appears to be on the verge of several, it is understood that revolutionary goals may be achieved in a series of incremental steps.26 The incremental approach, some hope, may lead to faster overall innova- tion, because breakthroughs can be difficult to incorporate into products. This is an important consideration for a consortium that must transfer its technology to member companies, because investment in technology trans- fer can sometimes be as large as the investment in research itself. SEMATECH, a consortium of 14 semiconductor companies and the De- partment of Defense, has as its mission the strengthening of U.S. semicon- ductor manufacturing technology. Small U.S. firms in the semiconductor equipment manufacturing industry have been losing ground, for some time, to Japanese firms with close ties to the large integrated electronics compa- nies. There is a fear expressed by the U.S. industry that, if Japanese firms 25 Institute for New Generation Computer Technology, Fifth Generation Computer: Outline of Fifth Generation Computer Project (Tokyo: Institute for New Generation Computer Technology, 1987, p. 9.) 26 Grant Dove, from his presentation on "The Environment for R&D Consortia in the United States" at the Workshop on R&D Consortia and U.S.-Japan Collaboration.

OCR for page 10
18 come to dominate the semiconductor equipment market, access to cutting- edge process technology by the U.S. semiconductor industry could be de- layed or denied. Therefore, SEMATECH, along with the companies partici- pating in the affiliated group of equipment manufacturers, has a technical mission with a distinctly vertical cast. In structuring R&D consortia, one important question is whether all members have the same opportunities. SEMATECH has only one level of membership, and members may send researchers to the central laboratory commensurate with their financial contributions. According to those in- volved, the substantial financial commitment of the corporate members helps to forge cooperation among them. Questions concerning intellec- tual property rights and tax treatment have been, in the case of SEMATECH, not too difficult to resolve once the overall structure was agreed. While the environments for R&D consortia are changing in both the United States and Japan, the basic fact is that consortia are not easy to organize or manage in either country. In both countries the major initial impetus for R&D consortia is increasing concern over foreign competition and a desire to band together to promote indigenous technological capabil- ities. It will be interesting to see whether Japan will be successful in its efforts to transform R&D consortia into organizations that foster more fundamental, even basic, research. Only time will tell whether the United States can use consortia to deepen manufacturing capabilities and com- mercialization of new ideas. Of course, in both countries one can find examples of R&D consortia that span the range of R&D activities from basic to applied research to manufacturing. Evaluations of future success will depend, to some extent at least, on expectations and new goals. RESULTS AND SIGNIFICANCE Comparing the rationales for R&D consortia with the actual results sug- gests some broader themes in U.S.-Japan science and technology rela- tions and competition in high-technology industries. Americans often place more emphasis on the technical goals and outputs than do the Japanese. It is true that Japanese R&D consortia are often launched with a great deal of fanfare and ambitious technical goals, but after several years it is often difficult to say what the output has been. Many have questioned, for ex- ample, whether the Fifth Generation Computer Program achieved its techni- cal goals. In contrast, Americans often emphasize the value or lack of value of consortia primarily in terms of technical output. It was the per- ceived technical success of the VLSI Program that helped to spark enthusi- asm for consortia in the United States, and the performance of domestic collaborative efforts such as MCC and SEMATECH are and will continue to be evaluated primarily on technical grounds.

OCR for page 10
19 TABLE 3 Japanese Private Sector Evaluation of Government Technology Policy Policy Area No. of Measures Score (0-100) Research associations 1 72 Tax incentives 10 71 Information dissemination 3 70 Commission research mediation 2 68 Commission research fees 9 68 Organizational measures 11 68 Standards 4 68 Patent system 4 67 Patent filing fees 8 66 "Success conditional" loans 8 65 Basic guidance 9 64 International aspects 9 62 Industry -govemment-university activities 7 61 Research facilities 13 59 Human resource development 3 57 Budget 1 56 Source: Results of a survey conducted by Keidanren's Industrial Technology Committee in 1989. Results appeared in Kokogyo Gijuisu Kenkyu Kumiai Sanjunen no Ayami (Thirty-Year History of the Research System for the Development of Mining and Manufacturing) (Tokyo: Kokogyo Gijutsu Kenkyu Kumiai Kondankai, 1991). Despite uneven reviews from foreign observers of Japanese R&D con- sortia, all of the major industrial technology policy initiatives launched by MITI in the 1980s and thus far in the 1990s provide incentives for some form of collaborative research. There must be perceived benefits from the points of view of MITI and the firms. A survey conducted by the Industrial Structure Commitee of Keidaaren (Federation of Economic Organizations) in October 1989 indicates that members chose "research cooperation" as the greatest benefit of all policy instruments used by the Japanese goverment to promote technology policy.27 (See Table 3.) Not only MITI but also other agencies of the Japanese government have adopted R&D consortia in recent years. Compei~i~on among ministries for turf and budgets is apparent. In the past year, five ministries have launched programs that will promote research on the role of carbohydrates in cell 27 See Kogyo Gijutsu Kenkyu Kumiai Sodankai (Industrial Technology Research Assoca- tion Advisory Committee), Kogyo Gijutsu Kenkyu Kumiai 30 Hen no Ayumi (The 30 Year History of Industrial Research Associations) (Tokyo: Nihon Kogyo Gijutsu Shinko Kyokai, 1991), p. 38 for a summary of the survey results.

OCR for page 10
20 processes a theme that is relevant to commercial biotechnology. Three of the projects are directed toward commercial applications, and two encom- pass consortia with private sector participations There are several ways to explain a divergence between initial stated goals and results in Japanese R&D consortia. First, many consortia receive "success conditional" loans to be repaid only if the project is a success. Consortia receiving these loans therefore have an incentive to downplay their results. In addition, no one in Japan expects Me breakthroughs in proprietary technology connected with a firm's core businesses to occur in R&D consortia. It would be a mistake to conclude that Japanese R&D consortia have no impact on technology development. Participants can normally identify con- crete technology-based results. The patent battles that raged among the participants for years after the conclusion of the VLSI Program illustrate that significant technology is developed in the collaborative setting. In addition to technology-related effects, evaluating success must take into account the intensity and complexity of consortium interactions with the members and other organizations. In evaluating the Future Electron De- vice project, for example, there were important effects in stimulating government-indus~y-university collaboration as well as technical outputs like resonant tunneling devices.29 The Japanese R&D consortium provides members with a forum for the exchange of precompetitive technology and a mechanism for promoting a "culture of exchange" in which there is more interfirm interaction at various levels of the hierarchy than would otherwise occur. The "career path" benefits of Japanese R&D consortia relate to the seniority-based hierarchy and lifetime employment system common in large Japanese companies. Typically, those involved in corporate R&D enter a company in their mid- twenties after receiving master's degrees. Japanese electronics companies tend to send younger researchers to R&D consortia. The rotation of person- nel through a consortium to the proprietary in-house group working on the same technology is analogous to the standard internal rotation of techni- cal personnel from basic research to applied research and then to manufac- turing and line management within the business division. In addition to training and relationship building among younger research- ers, Japanese R&D consortia are seen to encourage interaction among labo 28 "San Shocho de Shin Purojokuto" (Three Agencies Launch New Projects), Nihon Keizai Shinbun, September 3, 1990, p. 17. 29 Yoshio Nishi, from his presentation on "The Environment for R&D Consortia in Japan" at the Workshop on R&D Consortia and U.S.-Japan Collaboration.

OCR for page 10
21 ratory directors and others at a more senior level. Those who reach the laboratory director level have 30 to 60 patents to their credit, have written papers and books, and have substantial experience doing research them- selves. Top management gives the broad technical direction, but the R&D managers implement and evaluate from the bottom up. Indeed, the implications of the "culture of exchange" are most clear when comparing the normal levels of interaction that occur between consor- tia members in Japan and in the United States. In Japan the interaction between the consortia and the company is more frequent among senior people.30 Japanese companies participating in consortia send senior people to participate in steering committees. Japanese R&D consortia are also organized to enhance interactions between senior and junior people. These factors combine to enhance the transfer of technology from the R&D con- sortia to the participating firms. In Japan, interaction among companies often comes at an earlier stage in industries like the semiconductor industry. For example, "technical mile- stones" are set early. There are incentives to show evidence that work is going smoothly, to promote interactions between R&D and systems people early, and to show and benchmark technical output at the International Solid State Circuit Conference (ISSCC) earlier than is usually the case for U.S. companies.3i Japanese companies also tend to apply for patents earlier. (See Figures 3 and 4.) One implication of these different patterns may be that the technical community in the Japanese semiconductor indus- try has given a false impression to U.S. industry both the circuit design community and top management that Japan is a generation ahead in process technology. The unfortunate result is to discourage investment in the U.S. industry, push the American systems houses to go to Japan for components, and encourage an aggressive approach among U.S. process suppliers. In contrast to the Japanese experience, where observers point to training and a "culture of exchange" as the major benefits of consortia, there seems to be a greater need for U.S. consortia to define and justify themselves in terms of their actual technical results. MCC reports success in computer- aided design, packaging, and expert systems and claims to be on the verge 30 Yoshio Nishi, from his presentation on "The Environment for R&D Consortia in Japan" at the Workshop on R&D Consortia and U.S.-Japan Collaboration. 31 Hitachi's announcement that it had successfully developed the world's first 64M DRAM in 1990 was criticized by competitors as being premature. See IEEE Spectrum, "The Main Event," January 1991, p. 30.

OCR for page 10
22 ~ ~1 system concept ~yslern design unit process RED unit process RED ll | Internat~analElectron | I I DevicsMeetrg I bgiciciraJit prdo~p~ ch p design l design swy proces ~prac - e integration SPICE paramelem final design rub final chip design prdo-labrication reiabBily . . . . ....... I International Sdid state1 | htenabonal Sdid State | | Circuile Canierence | | Circuits Conferees | vdun~ production yield improvement FIGURE 3 Typical VLSI R&D pattern in the United States. Source: Yoshio Nishi of breakthroughs in holographic storage and knowledge representation.32 As proof of success, those involved cite growing membership and a request from members for a 10-year plan. It would be difficult for U.S. firms and the U.S. government to use the same yardstick Japan uses to measure the effectiveness of R&D consortia. First, Japanese projects usually run for a limited time, so consortia do not have to justify their existence as permanent entities. Further, the technical output of particular projects, or consortia in general, does not seem to factor as strongly in the Japanese funding and policy environments. There is little evidence that "failures" have an adverse impact on the ability of govern- ment agencies to organize new projects. In the United States, in contrast, I_ system design urn' process RID hgicicircuit design prototype chp dew early process ir~tegration at-risk design rule international Electror, Device M`3stirg nalanl red ongiriserrig sample dew process irtegration final design rule hiterr~aliorial Sdid State Circuits Canbrsnos patent customer visit (RSD) design revision sample for approval process "rov~t reliability design rule trttrnirg t~ndogy transfer patent | volume production 1 I Abed improvement I cost down customer | I c~storner~(sal95) | - make process available | t~ ~ - Ilk| Ird~aled Circus with | lull library l FIGURE 4 Typical VLSI R&D pattern in Japan. Source: Yoshio Nishi 32 Grant Dove, from his presentation on "The Environment for R&D Consortia in the United States" at the Workshop on R&D Consortia and U.S.-Japan Collaboration.

OCR for page 10
23 failure to show technical results would have an adverse impact on future funding. Committing resources to R&D consortia in the United States must be justified in terms of the high probability of concrete results in the near term. Even if the training benefits gained by sending researchers to work at consortia are substantial and quantifiable, there is less emphasis on these benefits in the United States where labor mobility is high. Perhaps underlying conceptual differences between the two cultures have an impact on disparities in the organization and management of R&D con- sortia as well. Independently staffed, permanent R&D consortia mesh with a stress on technology development at one location, followed by a transfer to the corporate setting where it will be applied. Likewise, temporary consortia characterized by more intensive communication and exchange be- tween organizations fit with the Japanese view of technology as, fundamen- tally, "skills" that reside in people. What of the "shared technical vision" that some observers point to as a benefit that accrues to members of Japanese consortia? Indeed, it has been as- serted that American companies have a greater potential for benefiting from consortia precisely because they do not have the "shared vision" of the future development of their industry and the economy that MITI formulates in Japan. The positive evaluation in the United States of innovation that results from individual entrepreneurial behavior is somewhat in tension with an interest in understanding Japanese methods in order to compete with Japan in critical industries. Some will say that developing "shared technological vision" crosses the line into an area that has been considered collusive behavior under U.S. law since the early years of this century. Fears that R&D collaboration is another name for a subsidy to a noncom- petitive industry and that small firms will be disadvantaged have been voiced by opponents in the United States.33 There is, however, another school of thought about R&D consortia and collaboration more generally.34 Some believe that the standardization of components, the integration of information processing and distribution in- dustries, and digitalization are creating an environment in which the small venture capital-backed start-up, which has long seemed to be the most ef- fective means of quickly bringing technological innovation to the market, will not compete effectively. In this view the future belongs to networks of firms that are linked by equity and other mechanisms, large-scale groups that have the wherewithal to make significant long-term investments in 33 Rich Karlgaard, "An Interview with T. J. Rodgers," Upside, December 1990, p. 67. 34 Charles H. Ferguson, "Computers and the Coming of the U.S. Keiretsu," Harvard Business Review, July 1990, p. 55.

OCR for page 10
24 R&D and capital equipment. Vertical R&D collaboration and horizontal R&D collaboration at early stages are seen as ways to ensure that U.S. semiconductor and systems companies have the same critical mass of re- sources that the Japanese industrial groups command in order to compete. There is no way to resolve these differences, either between Japan and the United States or among U.S. observers, here. The issues raised by comparing R&D consortia in the United States and Japan are related to larger questions relevant to U.S. international competitiveness, including impacts of industri- al structure, foreign investment, and government policies on science and tech- nology. In considering prospects for U.S.-Japan collaboration in R&D, it is important to consider the potential gains and risks to the U.S. economy and technology base as well as to the U.S. organizations participating.