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 63
II FINDINGS AND RECOMMENDATIONS
OCR for page 64
This page in the original is blank.
OCR for page 65
Findings The federal government has played a significant role in supporting the growth of the semiconductor industry since its inception.1 The industry has benefited from close cooperation with government, both through generous procurement contracts such as those related to defense and space exploration, and through research consortia. This support for industry research and development is fully justified. The semiconductor industry’s technological progress has enabled major advances in technologies directly relevant to core government missions including those in national security, communications, health, weather, the environment, and education. In addition, there is growing recognition of the importance of the industry’s contributions to the productivity growth of the U.S. economy.2 The contribution of new technologies to growth, especially information technologies, is now recognized at the highest levels of U.S. policy making. Notably, 1 As Laura Tyson observed in 1992: “The semiconductor industry has never been free of the visible hand of government intervention. Competitive advantage in production and trade has been heavily influenced by policy choices, particularly in the United States and Japan. Some of these choices, such as the provision of public support for basic science, R&D, and education in the United States, have had general, not industry-specific objectives. But other choices, such as the provision of secured demand for industry output through military procurement in the United States and through preferential procurement of computers and telecommunications equipment in Japan, have been industry specific in intent and implementation.” Laura D’Andrea Tyson, Who’s Bashing Whom?: Trade Conflict in High Technology Industries, Washington, D.C.: Institute for International Economics, 1992, p. 85. 2 Dale Jorgenson and Kevin Stiroh, “Raising the Speed Limit: U.S. Economic Growth in the Information Age,” in Brookings Papers on Economic Activity, (2), 2000, p. 125-212.
OCR for page 66
Federal Reserve Chairman Alan Greenspan has affirmed the contribution of new technologies to the low inflation, low unemployment, and the continued high growth rates that characterized the U.S. economy in the latter half of the 1990s.3 Much of the technological advance that has made these productivity gains possible is dependent on the unprecedented decrease in cost of increasingly more powerful semiconductors.4 I.THE SEMICONDUCTOR INDUSTRY: A HISTORY OF COMPETITION AND COOPERATION Firms in the U.S. semiconductor industry have a deserved reputation as fierce competitors in both American and foreign markets. Yet, at key points in the history of the American semiconductor industry, particularly in the decade of the 1980s, the industry launched cooperative efforts through organizations such as the Semiconductor Research Corporation (SRC—1982) and SEMATECH (1987).5 This cooperative research has pooled expertise, lowered costs, and encouraged the dissemination of knowledge across the industry.6 After two decades of relative declines, the decade of the 1990s witnessed a major resurgence in the competitive position of American industry in many sectors.7 As 3 Alan Greenspan, Technological Innovation and the Economy, Remarks Before the White House Conference on the New Economy, Washington, D.C. April 5, 2000, Federal Reserve Board. 4 Ibid. See also National Research Council, Measuring and Sustaining the New Economy, Washington, D.C.: National Academy Press, 2002. 5 The Semiconductor Research Corporation, founded in 1982, is based in Research Triangle Park, North Carolina, and has an office in San Jose, California. Its stated goal is to operate globally in order to provide competitive advantage to its member companies as the world’s premier university research management consortium delivering relevantly educated technical talent and early research results. In the SRC’s words; “The goal in 1982, as it is today, was to define common industry needs, invest in and manage the research that would expand the industry knowledge base and attract premier students to study semiconductor technology.” In addition, the SRC also trains and produces graduates who are highly and relevantly skilled to perform at the frontier of semiconductor research. The SEMATECH (SEmiconductor MAnufacturing TECHnology) consortium was a public-private industry partnership formed in 1987 in order to reinvigorate the semiconductor industry in the United States, which had lost significant market share to Japanese firms. The consortium eventually focused on encouraging cooperation among firms to establish standards and helped to develop roadmaps for the evolution of the industry. The consortium stopped receiving federal support in 1996 and has further evolved to include foreign firms. It is now known as International SEMATECH. Currently, partnerships are under way with members, equipment and materials suppliers, national laboratories, and other consortia. 6 Kenneth Flamm and Qifei Wang, “Sematech Revisited: Assessing Consortium Impacts on Semiconductor Industry R&D,” in this report. 7 See National Research Council, U.S. Industry in 2000: Studies in Competitive Performance, Washington, D.C.: National Academy Press, 2000.
OCR for page 67
previous analysis by the National Research Council suggests, an important part of the improvement in the competitive position of American industry can be attributed to the growth in the application of information technologies, particularly after 1995.8 A key challenge of the new century is to sustain the high rate of technological advance that has characterized the semiconductor industry, which underpins the information technologies which have in turn contributed to the growth of the American economy as a whole.9 A.A Steady Increase in U.S. R&D Investments In aggregate terms, the outlook for R&D investments in the United States appears favorable. On December 20, 2001, Congress approved a record federal R&D budget for FY 2002 of $103.7 billion—a 13.5 percent increase over FY 2001. Total R&D funding reached a preliminary $264.6 billion in 2000, or 2.68 percent of total GDP. This amount reflects an increased R&D share of GDP from 2.63 percent in 1999.10 Total U.S. R&D expenditures show a steady increase. For example, between 1995 and 2000, R&D expenditures increased at an average rate of 7.74 percent.11 B.Increases Mask Substantial Shifts Overall, increases in R&D investments are widely recognized as a good thing in that the social returns on such investments (that is, the gains for society as a whole) are very high, on the order of 40 to 50 percent.12 However, the composition of R&D investments also matters and, in this regard, current trends are a cause for concern. Federal support for R&D has not kept pace with private-sector investments, which have risen dramatically (see Figure 1). In 1980, the federal share of R&D was roughly 48 percent. In 1999, it had fallen to 28 percent of the 8 Ibid. See also: Jorgenson and Stiroh, op. cit.; and National Research Council, Measuring and Sustaining the New Economy. 9 Ibid. 10 These figures use a preliminary estimate for 2000 from the National Science Foundation. The data are derived from the National Science Foundation’s, National Patterns of R&D Resources: 2000 Data Update, “National expenditures for R&D, from funding sectors to performing sectors: 1993-2000.” 11 Ibid. 12 There is a substantial literature on the social benefits or “spillovers” attributed to R&D investments, from Edwin Mansfield’s early work in 1977 to more recent analysis by Zvi Griliches. See, for example, Martin N. Baily and A. Chakrabarti, Innovation and the Productivity Crisis, Washington, D.C.: The Brookings Institution, 1998; and Zvi Griliches, The Search for R&D Spillovers, Cambridge, MA: Harvard University Press, 1990.
OCR for page 68
FIGURE 1. Total Real R&D Expenditures By Source of Funds 1960–2000. SOURCE: National Science Foundation, National Patterns of R&D Resources. total. In part, this reflects U.S. industry’s commitment to developing new products and processes, and, in part, the declines reflect the budgetary constraints and uncertainties of the mid-1990s. Whatever the cause, the current differential trends are a source of concern because, as noted below, the government and industry focus on different phases of the innovation system. The Nature of Industry R&D The contribution of industry to the R&D budget has focused more on product development than on basic research. Of the $43.5 billion of federal R&D devoted solely to research in FY2001, $22 billion, or 50.7 percent, was channeled into basic research.13 The semiconductor industry devotes $14 billion to R&D, or 14 percent of sales per year (as of 2000).14 The industry also devotes a significant portion of its R&D effort to university-based research through the Semiconductor 13 See American Association for the Advancement of Science; Congressional Action on Research and Development in the FY 2001 Budget. 14 See <http://semichips.org/ind_facts.cfm>
OCR for page 69
BOX A THE MICROELECTRONICS ADVANCED RESEARCH CORPORATION (MARCO)a MARCO, a cooperative program organized under the auspices of SRC, funds and operates a number of university-based research centers in microelectronics as part of its Focus Center Research Program (FCRP). The Focus Centers concentrate on those areas of microelectronics research that must be addressed to maintain the historic productivity growth curve of the industry. Focus Programs involve multiple universities and place strong emphasis on cross-fertilization of ideas during the basic research stage. Focus Program research is typically longer term—normally eight years away from commercialization. Advances made under the Focus Program can become proposals to the SRC to address long-term needs identified in the Industry Roadmap (ITRS).b a MARCO was established in 1998. For additional information on MARCO, see the SRC website. <http://src.org/member/about/fcrp.asp>. b See International Technology Roadmap for Semiconductors, <http://public.itrs.net/Files/2001ITRS/Home.htm>. Research Corporation, the jointly funded Focus Center Research Programs (see Box A), and industry R&D collaboration through SEMATECH.15 In general, however, industry has been less inclined to fund the basic research on which the future growth of the economy ultimately depends.16 Yet, much of the current technological progress the United States and, indeed, the rest 15 The Focus Center Research Program is a national research network formed in 1998. It is jointly funded by the U.S. semiconductor industry and the federal government. Its purpose is to address core issues in technology development for the semiconductor industry. The program supports long-range, broad-based research that seeks to establish new perspectives and approaches to technological challenges facing the industry. In March 2001, the Semiconductor Industry Association announced that it would double the size of the FCRP. Subsequent economic conditions have made the timing of this decision uncertain. 16 National Research Council, Allocating Federal Funds for R&D, Washington, D.C.: National Academy Press, 1995. See also National Research Council, Capitalizing on New Needs and New Opportunities: Government-Industry Partnerships in Biotechnology and Information Technologies: New Needs and New Opportunities, C. Wessner, ed., Washington, D.C.: National Academy Press, 2002.
OCR for page 70
of world enjoys today rests on inventions and investments made 30 and 40 years ago.17 In addition, many of the large industry laboratories that once supported major technological advances, such as the transistor, no longer exist or have seen their research strategies substantially modified.18 C.The Expansion of Foreign National R&D Programs As noted above, governments around the world have played an active role in the development of the semiconductor industry. In its early years, the U.S. industry received substantial support for research and development from the federal government, particularly to achieve national missions in defense (e.g., the Minuteman program) and in space exploration (e.g., the Apollo program). Governments around the world have also taken an active approach in supporting the entry of their national firms into the global semiconductor market.19 Japan’s early VSLI program, for example, helped bring its producers to the forefront of the industry in only a few years.20 In 1987, SEMATECH was founded to aid a beleaguered U.S. industry. Korea followed in the late 1980s and 1990s with generous state-supported financing for DRAM production by its chaebols. Taiwan’s innovative policy mix of equity finance, technical support, favorable tax treatment, and the development of the Hsinchu Science and Technology Park Complex helped propel its industry forward in the 1990s. Enhanced R&D support and other programs are not confined to new entrants. Several European countries, operating in conjunction with the European Union, have put in 17 David Tennenhouse, vice-president and Director of Research and Development at Intel, emphasized this point in his presentation at The Global Computer Industry Beyond Moore’s Law: A Technical, Economic, and National Security Perspective, a Joint Strategic Assessments Group (SAG) and Defense Advanced Research Projects Agency (DARPA) Conference, January 14-15, 2002. 18 Richard Rosenbloom and Bill Spencer, Engines of Innovation: U.S. Industrial Research at the End of an Era, Boston: Harvard Business School Press, 1996. For a condensed version, see Rosenbloom and Spencer, “The Transformation of Industrial Research” in Issues in Science and Technology, 8(3):68-74. 19 See the proceedings below for a discussion of the wide range of programs under way in Japan, Taiwan, and Europe (the latter at the EU, regional, and national levels). The paper by Thomas Howell in this report, “Competing Programs: Government Support for Microelectronics,” provides original documentation concerning the focus and funding of many of these programs. The steady growth of these programs and the levels of public support reflect both the perceived importance of the industry and the perceived success of the American model. See for example the chart summarizing program goals and funding in Thomas Howell, op cit., in this report. Howell identifies about 17 programs currently under way in the microelectronics industry outside the United States. 20 For a description of these programs and a prescient prediction of the recovery of European firms, see Thomas, Howell, Brent Bartlett, and Warren Davis, Creating Advantage: Semiconductors and Government Industrial Policy in the 1990s, Santa Clara, CA: SIA, 1992.
OCR for page 71
place programs that have contributed to a strengthened competitive position for European producers. A wide variety of policy instruments—ranging from substantial government funding for national R&D programs to favorable tax treatment (e.g., short depreciation allowances) and, in the past, trade measures such as tariffs and private restraints of trade (i.e., restrictive internal market arrangements)—have been used to promote domestic semiconductor firms. Given the perceived contributions of SEMATECH, countries and regions interested in supporting the semiconductor industry have adopted the consortium model as a means of encouraging cooperation among firms within a national industry and as a vehicle for providing government support. The combination of technical challenges facing the semiconductor industry and the perceived success of cooperative programs in the United States have led policy makers in several countries to increase government funding in support of their national semiconductor industries.21 Box B describes current trends in national programs to support national semiconductor industries. II.THE SEMICONDUCTOR INDUSTRY FACES SIGNIFICANT CHALLENGES The substantial increases in semiconductor power—predicted by Moore’s Law—are becoming more challenging to continue. To do so, the industry must overcome a series of technical hurdles, including the need for both new materials and designs. It must also address the need for skilled labor required to overcome these hurdles amidst emerging changes in the structure of the industry. Given the economic importance of the industry, there is very limited research on the impact of SEMATECH on R&D in the semiconductor industry, its role in the resurgence of the U.S. industry, and its potential lessons for other U.S. consortia.22 With regard to the industry as a whole, there is limited economic research as to the sources of the industry’s pronounced cyclical swings, its contributions to productivity, and its subsequent impact on the economy at large. Scant public policy attention has been focused, as well, on the research requirements needed to keep this industry on its positive course, and on the skilled labor and advanced training needed to sustain this trajectory. 21 Ibid. 22 See Flamm and Wang, op. cit., in this report.
OCR for page 72
Box B Significant Global Trends in the Semiconductor Industry as a Result of National Programs Substantial Support for Microelectronics in Japan. The Japanese government has initiated a series of R&D initiatives and has provided substantial support in cooperation with the Japanese industry. These initiatives are intended to contribute to a “national revival” in the competitive position of the Japanese semiconductor industry. Strong Support for Microelectronics in the European Union. European Union and national government supported R&D projects— such as JESSI (the Joint European Submicron Silicon Initiative), MEDEA+ (the Micro-Electronics Development for European Applications), and IMEC (Inter-university Micro-Electronics Centre)—have helped to reverse declines in the European microelectronics industry, considerably improving its global competitive standing. The comparative advantage of other nations in leading-edge mobile communications and digital technologies will be leveraged. Japan and the EU have designed comprehensive strategies to challenge the U.S. leadership in microelectronics by leveraging their present and expected future advantages in mobile communications and digital home appliances. The Challenge of the foundry model. Taiwan’s pioneering of a new business model—the dedicated foundry—has the potential to revolutionize the industry. Taiwan has emerged as a major production base, in part through government capital to launch the industry and through a supportive environment such as the Hsinchu Park Complex. China: A future competitor. China is making a concerted effort to become a significant competitor in microelectronics. Its government has made substantial efforts to attract foreign investment and promote the diffusion of more advanced foreign technology to the Chinese mainland, while promoting indigenous producers. The evolution of the Chinese microelectronics industry will continue to be augmented by the ongoing movement of Taiwanese manufacturing information technology to mainland China.
OCR for page 73
A.Need for Highly Skilled Human Capital 1.Continued Progress Depends on the Supply of Talented and Skilled Labor. In order for the semiconductor industry to maintain high growth rates and respond to the growing challenges within the industry, the United States faces a long-term need to bolster support for highly skilled workers. While, at this writing, the industry is showing the effects of a historically sharp downturn, this cyclical feature should not mask the growing long-run demand for the skilled workers needed to keep the industry on its current growth path. Sustaining the industry’s remarkable rate of technological advance requires persistent creativity and ingenuity. Such an innovative environment is sustained by a trained workforce well grounded in the disciplines—such as physics, mathematics, and engineering—that underpin the semiconductor industry. This long-term growth in the demand for skilled labor has emerged against the decline in U.S. federal funding for these disciplines (see Figure 1). The United States is also competing globally to generate and attract the human capital necessary to the long-term health and development of the semiconductor industry.23 2.Generating a Skilled and Qualified Workforce in the Microelectronics Industries. Despite ongoing initiatives to address the skilled manpower needs of the industry by organizations such as the Semiconductor Research Cooperation (SRC), there is concern within the industry as to whether there will be enough skilled graduate students to meet future demand—a problem some believe is worsening. The SRC has documented a significant drop in the graduation rate of electrical engineering students in the United States from 1988 to the present and projects no recovery from these low levels in coming years.24 To compound this challenge, competition for the limited pool of talented workers in engineering fields is global, and U.S. industry will face increased difficulty in attracting the young, skilled workers it needs to continue growing. The United States exhibits one of the lowest yields, 5.3 percent, in producing engineers when comparing the number of bachelor’s degrees in engineering to all bachelor’s degrees. In the sheer production of engineers on a yearly basis, the U.S. is surpassed by Asian nations as a group by almost six times.25 Other na 23 See Howell, op. cit. 24 For a discussion of the decrease in engineering students and its implications for the microelectronics industry, see Michael Polcari’s statements in the Proceedings of this report, “Current Challenges: A U.S. and Global Perspective,” pp. 115-116. For a recent review of these challenges, see National Research Council, Building a Workforce for the Information Economy, Washington, D.C.: National Academy Press, 2001. 25 This group of Asian nations consists of China, India, Japan, South Korea, and Taiwan.
OCR for page 80
1.Private partnerships and government-industry partnerships represent an integrated national approach to develop semiconductor technology. a. Japan’s Pre-competitive R&D Programs. In Japan, Selete (SEmiconductor Leading Edge Technology Corporation), a joint-venture company established in 1996, conducts R&D on behalf of the Japanese semiconductor industry. Selete, which is not directly funded by government, has been successful in the promotion and evaluation of technologies, developing advanced technologies, and carrying out special projects. By comparison, ASET (Association of Super-Advanced Electronics Technologies) is completely funded by the government and focuses on equipment and chip R&D for 70- to 100-nm technology. b. Europe’s Multinational, Multi-firm Partnerships. MEDEA (Microelectronics Development for European Applications) is a multinational, multifirm partnership. It is similar to SEMATECH in that it was jointly financed by government and industry. MEDEA has helped develop a better understanding between semiconductor suppliers and system houses, helped develop a better idea of where to focus R&D resources, and fostered closer cooperation among companies in different European countries—both vertically and horizontally. MEDEA officials report that these efforts demonstrate that collaboration in the semiconductor industry can have positive effects for society (employment in the industry increased) and that it can be a productive use of public funds. As a result of MEDEA’s success, MEDEA-Plus was initiated in 2001 to address the challenges facing the semiconductor industry noted above. c. Government-industry Partnering in Taiwan. The semiconductor industry in Taiwan was born out of government support and partnerships in the mid-1970s. Today, the major semiconductor companies in Taiwan are world leaders in their specialties. One of the most successful joint ventures between industry and government is Taiwan Semiconductor Corporation (TSMC). TSMC is a positive example of a government-industry equity partnership in terms of return to society on public investments.37 The dynamic effects for the Taiwanese economy associated with the establishment of a rapidly growing, highly competitive industry are substantial.38 37 The most evident return is in the investment itself. Originally, the government had invested about $100 million in TSMC; the stock was later sold for $400 million. 38 The highest return is the long-run economic impact of having a leading semiconductor producer in Taiwan. Taiwan’s science-park approach (notably Hsinchu Park) to creating partnerships between government and industry has shown substantial success. Though in the beginning there were many foreign companies in the park, by the late 1990s roughly 80 percent of the companies were either local or domestic. Direct equity investment by the government has proved effective. Most of the financial capital provided to the park originated from the government. Currently only about 4 percent of financial capital comes from the government.
OCR for page 81
IV.THE IMPACT OF SEMATECH: A GOVERNMENT-INDUSTRY PARTNERSHIP SEMATECH is widely perceived as effective in accomplishing its goals. The consortium’s members believe that participation in the consortium has been worthwhile, as evidenced by their continued participation and contributions. This positive assessment is further reflected in the industry’s willingness to discontinue public funding while continuing to support the consortium. The foreign competitors of the U.S. industry share the perception that SEMATECH contributed to the resurgence of the American semiconductor industry and have established a variety of similar programs. These programs are often on a significantly larger scale and have greater underlying political support. Furthermore, a significant number of foreign producers have affirmed their belief in the program’s effectiveness by joining SEMATECH since it became an international consortium in 1999.39 These trends underscore the importance of public-private cooperation to support research and technology development in the semiconductor industry. In light of the growing significance of R&D collaboration in both the equipment and device industries, providing policy and financial incentives to encourage such cooperation is an increasingly important way to sustain the investments needed to transition to successive generations of new technologies.40 39 This “market-based” judgment of the consortium’s utility is, in the end, the most compelling. Economic analysis of SEMATECH’s impact is extremely challenging. As Kenneth Flamm and Qifei Wang observe, “Finally, the underlying models of R&D cooperation which ultimately must be the basis of a scientific effort to untangle the chains of causality are simply too simplified at this point to capture the complexity of the real world of SEMATECH: a real world in which companies committed to R&D carried out within a joint venture while at the same time competing through internal R&D efforts which also may have spilled over to competitors, a real world in which the menu of consortium activities changes over time with experimentation and learning. At the end of the day, the only absolutely certain thing about SEMATECH is that a substantial portion of its member companies must have found it to be of net value—having actually run the experiment of ending public subsidy, and finding that its consumers continued to buy its output.” See Flamm and Wang, op. cit. 40 The operation of an effective consortium entails an agreement on achievable goals in accordance with a sense of shared interests. In addition, a consortium needs an effective management structure tightly linked to member interests, as well as a long-term commitment from its participants to contribute highly trained and qualified personnel and to provide financial support. For a further discussion of characteristics of successful national and international consortia, see National Research Council, Conflict and Cooperation in National Competition for High-Technology Industry, Washington, D.C.: National Academy Press, 1996, pp. 48-51.
OCR for page 82
A.Sources of SEMATECH Contributions 1.Flexible Objectives and Industry Leadership By definition, an R&D consortium’s contributions (Box D) are due in part to its ability to adapt its goals to the conditions of a rapidly evolving industry.41 The cost sharing arrangement with the government and industry management of the research agenda has contributed to this flexibility. Indeed, while it benefited from strong leadership, no single entity dominated the consortium or determined its direction. Members, including Department of Defense officials, reached a broad consensus on technical goals and then left the consortium management to implement the program. The industry interaction within the consortium, and between the consortium members and the suppliers, improved the dynamics between the device makers and the equipment industry, with collaboration generating new technical perspectives for the participants and encouraging the give-and-take between manufacturer and supplier necessary to expedite the technology development process. 2.Analyzing SEMATECH Measuring the contributions of research consortia is a difficult task (Box E). As noted in this report, there have been relatively few empirical analyses of the impacts of R&D cooperation on industrial R&D.42 For the semiconductor industry, some empirical analysis suggests that the consortium has boosted the “effec 41 While many believe that SEMATECH contributed to the resurgence of the U.S. semiconductor industry in the early 1990s, it was by no means the only element in this unprecedented recovery. For example, time for the industry to reposition itself was provided by the 1986 Semiconductor Trade Agreement. The U.S. industry also repositioned itself, profiting from shifts in demand, i.e., away from DRAMS (where Japanese skill in precision clean manufacturing gave significant advantage) towards microprocessor design and production (where U.S. strengths in software systems and logic design aided in their recovery.) Arguments about which of these elements were most decisive probably miss the point. The recovery of the U.S. industry is thus like a three-legged stool. It is unlikely that any one factor would have proved sufficient independently. Trade policy, no matter how innovative, could not have met the requirement to improve U.S. product quality. On the other hand, by their long-term nature, even effective industry-government partnerships can be rendered useless in a market unprotected against dumping by foreign rivals. Most important, neither trade nor technology policy can succeed in the absence of adaptable, adequately capitalized, effectively managed, technologically innovative companies. In the end, it was the American companies that restored U.S. market share. 42 Few researchers have empirically assessed the effects of joining SEMATECH on its member firms’ expenditure on private R&D. As noted by Flamm and Wang, op. cit., in this report, even
OCR for page 83
Box D Contributions of the Consortium SEMATECH has made a variety of important contributions to the health of the semiconductor industry in the United States. For example, the consortium has: Played an integral role in promoting the development of effective manufacturing technology in the semiconductor industry. Developed industry-wide standards for manufacturing tools, notably through collaboration with the equipment industry and through industry-wide technology roadmaps. Fostered a shared perspective on the technological development required to maintain the industry’s high growth rate through the semiconductor roadmap process. Aided companies in: developing reliable manufacturing tools creating an effective quality control process understanding the needs of the industry and advancing the sophistication of the manufacturing process. though SEMATECH is the highest-profile R&D consortium in the United States, it has been the focus of study for only three statistically rigorous papers. One study (Douglas A. Irwin and Peter J. Klenow, “High-Tech R&D Subsidies: Estimating the Effects of SEMATECH,” Journal of International Eco-nomics 40(3-4):323-44, May 1996) found that SEMATECH firms reduced their individual expendi-tures on R&D by about $300 million dollars. They further concluded that the reduction in firm-level R&D of member firms does not justify public support for the consortium, since firms are essentially free-riding on federal funds and would have expended the equivalent federal funds out of their own budgets had there not been a consortium. Irwin and Klenow argue further that firms joined SEMATECH to “share” information but not to necessarily “commit” funding for high spillover R&D, which, if true, would have resulted in an increase in R&D. There are both conceptual and economet-ric flaws with this argument, as pointed out in Flamm and Wang, op.cit. Irwin and Klenow interpret the data as member firms reducing R&D expenditure that would have been conducted in the absence of the consortium, rather than reducing poorly appropriable R&D, which would not have occurred at all had it not been for the consortium’s formation. Further, even a “commitment” approach to R&D among consortium members, with relatively low spillovers, would actually lower R&D, a conclusion the authors do not draw.
OCR for page 84
Box E Organizing Successful Consortia Because of its contributions, SEMATECH is sometimes considered a model for future public-private partnerships.a Some of its lessons for organizing a successful consortium are:b Understand the Need for Cooperation—the great range of R&D needs, from basic science to manufacturing infrastructure to whole new industries, is, arguably, best understood in terms of a process where industry works in close cooperation with universities and government research laboratories. To achieve the full benefits of cooperation, it is important to: Ensure Quality Leadership, including key leaders of the major participating industries Convey your Message publicly to leaders in the government and private sectors Focus the Program on key sectors and build on this developed strength, rather than approach the entire industry Set Measurable Objectives for advancing generic or pre-competitive knowledge Set Uniform Requirements of participation so that support is not fragmented Plan first—Spend later: Roadmaps are needed before consortia can be properly launched Develop an Industry-driven process—recent collaborative work, such as that on extreme ultra-violet lithography, shows that successful consortia are industry driven. a National Research Council, “Government support for technology development: The SEMATECH experiment,” Conflict and Cooperation in National Competition for High-Technology Industry,” pp. 141-51. b See remarks by William Spencer in the Proceedings section of National Research Council, Partnerships for Solid-State Lighting, Report on a Workshop, Washington, D.C.: National Academy Press, 2002. Toshiaki Masuhara of Hitachi also suggests a number of criteria for organizing a successful consortium. See his presentation in Panel VI of the Proceedings of this volume.
OCR for page 85
tive” R&D level of its members.43 The work of Flamm and Wang suggests that SEMATECH reduced the R&D expenditures of its membership somewhat, in part by eliminating duplication.44 In essence, if the number of dollars spent on similar R&D projects across firms is reduced, and the yield of overall industry R&D is unchanged, then this is a better outcome from both a social standpoint, for society, and for industry since resources are freed which can be put to productive use elsewhere. This is a positive result both for the firms and for society as a whole. Moreover, this outcome lends credence to the idea that a consortium can add to the dynamic efficiency of both its member firms and the industry as a whole. While the precise measurement of contributions is difficult, SEMATECH is widely believed, within the industry, both in the United States and abroad, to have made a positive contribution to the resurgence of the U.S. semiconductor industry. More indirectly, the consortium’s activities have contributed to greater cooperation among producers, suppliers, and the government. For example, the current promising cooperation on next-generation lithography tools, (i.e. the EUV Consortium) illustrates this enhanced willingness to collaborate in innovative ways. This positive perception of SEMATECH has contributed to its emulation, notably in foreign programs to support national or regional semiconductor industries and among other U.S. industries, (e.g. in optoelectronics and nanotechnologies).45 More broadly, SEMATECH helped sustain the rapid technological progress of the industry as projected by Moore’s Law. This technical progress was facilitated by the collaborative research encouraged by the consortium, including the development of the Semiconductor Industry Roadmap. For its part, the government partner achieved many of its objectives. The Department of Defense achieved its goal of maintaining a robust, technologically advanced manufacturing capability within the United States. SEMATECH thus helped the government achieve a key objective, namely, sustaining a U.S.-based industry able to provide cutting-edge, low-cost devices to support defense requirements46 and thereby avoiding the risk of dependency on foreign suppliers for U.S. defense systems.47 Throughout the decade of the 1990s, the Defense Department was able to acquire higher-performance, lower-cost components from 43 See Flamm and Wang, op. cit. 44 Ibid. 45 See Box B in the Introduction in this volume. See also National Research Council, Small Wonders, Endless Frontiers: A Review of the National Nanotechnology Initiative, Washington, D.C.: National Academy Press, 2002. 46 A healthy U.S. industry also ensures a surge capacity for the defense industrial base, should it be required. 47 The erosion of the U.S. semiconductor industry’s position was a source of growing concern to federal defense officials and was reflected in the creation of the National Advisory Commission on Semiconductors (NACS). One of the NACS’ missions concerned the dependency of modern weap
OCR for page 86
commercial suppliers than would have been available from a dedicated defense production facility. This trend contributed to dual-use defense acquisition designed to benefit from the rapid evolution of commercially available semiconductors characterized by rapidly increasing performance and falling costs.48 The combination of rapid gains in semiconductor capabilities and sharply falling costs has contributed to the government’s capacity to carry out many other non-defense missions more efficiently. These contributions are reflected in the economy as a whole. Also, as noted above, the U.S. economy recorded substantial gains in productivity growth between 1995 and 1999, with productivity growth more than double that of the 1973-1995 period. The Council of Economic Advisers attributed “these extraordinary economic gains” to three factors, namely, technological innovation, organizational changes in businesses, and public policy.49 Two of these factors concern information technology, in particular the simultaneous advances in information technologies—computers, hardware, software, and telecommunications—which combine these new technologies in ways that sharply increase their economic potential. Progress in semiconductor capabilities enabled advances in information technologies, driving innovation in each of these product areas. In short, the government and the economy as a whole have benefited from the contributions of a robust U.S. industry.50 ons systems on state-of-the-art semiconductor devices. Specifically, under the legislation creating the Commission, Congress notes in its findings that: “Modern weapons systems are highly dependent on leading-edge semiconductor devices, and it is counter to the national security interest to be heavily dependent upon foreign sources for this technology.” The charter further states that this Committee shall “identify new or emerging semiconductor technologies that will impact the national defense or United States competitiveness or both.” For the objectives set forth for NACS, see <http://www4.law.cornell.edu/uscode/15/4632.html>. 48 See Jacques Gansler, Defense Conversion: Transforming the Arsenal of Democracy, Cambridge, MA: MIT Press, 1995. See also the presentation by Paul Kaminski, then Under Secretary of Defense for Technology and Acquisition, in National Research Council, International Friction and Cooperation in High-Technology Development, Washington, D.C.: National Academy Press, 1997, pp. 132-133. Dr. Kaminski points out that tighter linkages with commercial markets shorten cycle time for weapons-systems development and reduce the cost of inserting technological improvements into DoD weapons systems. By placing greater reliance on commercial sources, DoD can field technologically superior weapons at a more affordable cost. 49 Council of Economic Advisers. Economic Report of the President, Washington, D.C.: Government Printing Office, 2001. 50 These broader contributions may be relevant with respect to proposals for recoupment of federal contributions to government-industry partnerships. Some analysts suggest that the best means of recoupment is, in fact, the tax system. For example, many of the companies that have thrived following SEMATECH’s inception have returned the government’s original investment to the consortium many times over in the form of tax revenue. See testimony by Christopher T. Hill, Vice Provost for Research and Professor of Public Policy and Technology, George Mason University, before the Subcommittee on Environment, Technology and Standards, Committee on Science, U.S. House of Representatives, March 14, 2002.
OCR for page 87
As described in greater detail in the Introduction, the SEMATECH consortium’s contributions to the resurgence of the U.S. industry were significant but are best understood as one element of a series of public policy initiatives that collectively provided a positive policy framework for U.S. semiconductor producers. Overall, SEMATECH’s record of accomplishment was achieved in no small part through the flexibility granted its management and the sustained support provided by DARPA, the public partner, complemented by the close engagement of its members’ senior management and leading researchers.
OCR for page 88
Recommendations The Committee’s recommendations outline a series of modest steps that nonetheless may prove important to the long-term welfare, economic growth, and security of the United States. RESOURCES FOR UNIVERSITY-BASED SEMICONDUCTOR RESEARCH To better address the technical challenges faced by the semiconductor industry and to better ensure the foundation for continued progress, more resources for university-based research are required. The Committee believes that universities have an important role in maintaining a balance between applied science and fundamental research. This balance is key in generating ideas for future research. The Committee suggests consideration of the development of three-way partnerships among industry, academia, and government to catalyze progress in the high-cost area of future process and design. These partnerships would: Sponsor more initiatives that encourage collaboration between universities and industry, especially through student training programs, in order to generate research interest in solutions to impending and current industry problems.
OCR for page 89
Increase funding for current programs.51 Research programs that are already operational, such as the Focus Center Research Program developed by the SRC, could usefully be augmented through substantially increased direct government funding. These centers also represent opportunities for collaborative research with other federal research programs, such as those supported by the National Science Foundation. Create Incentives for students. A key role for universities is to ensure the flow of technical innovation and skills that originate with students. In order to address the undersupply of talented workers and graduate students in the industry, more incentive programs should be established. Since professors typically respond to appropriate research incentives, augmented federal support for programs that encourage research in semiconductors would attract professors and graduate students.52 In addition, specific incentive programs could be established to attract and retain talented graduate students. 51 The president’s FY 2003 budget makes important steps in this direction. It calls for a 3 percent increase, to $1.9 billion, in the Networking and Information Technology Research and Development Program (NITRD). This particular program could play a key role in funding the basic research that confronts the technical challenges in the semiconductor industry. The NITRD coordinates key advanced information technology research across multiple agencies to make broad advances in computing and networking. These advances manifest themselves in the development of new technologies such as computing platforms and software, which can support advances research in physics, materials science and engineering as well as biomedical and earth and space science. The 2003 budget envisions emphasizing critical areas of research such as networks security issues; high-assurance software and systems; micro- and embedded-sensor technologies; and revolutionary architectures to reduce cost, size, and power requirements of high-end computing. The budget emphasizes research on the social and economic impacts of developments in the fields of information technology. For the text of the president’s proposed initiatives, see Fiscal Year 2003, Analytical Perspectives, Budget of the United States Government, Washington, D.C.: U.S. Government Printing Office, 2002, p. 164. 52 See Paula Stephan and Grant Black, “Bioinformatics: Emerging Opportunities and Emerging Gaps,” in National Research Council, Capitalizing on New Needs and New Opportunities: Government-Industry Partnerships in Biotechnology and Information Technologies, p. 244.
OCR for page 90
This page in the original is blank.
Representative terms from entire chapter: