The Design of High-Technology Consortia: Lessons from SEMATECH

Abstract

Introduction Paper attempts to derive lessons from Sematech for design of consortia in other industries. SEMATECH has shifted from horizontal research cooperation to vertical collaboration between its members, major US users of semiconductor process equipment and materials; and US suppliers of these goods. Technology diffusion is a key role of SEMATECH. However, the vertical orientation has not satisfied all of the original members, some of whom have left. Also, some firms have failed. Design and Policy Issues for R&D Consortia In theory, research consortia produce several economic benefits. However, in order to exploit externally performed research, participant firms must invest in the creation of in-house expertise, so some duplication necessarily results. Also, reduction in duplication reduces diversity, which may act as a hedge. Moreover, there is no reason why all forms of industrial R&D exhibit economies of scale. Finally, concerns with appropriability and spillovers may hamstring the consortium’s operations. Three challenges arise in consortium design and management: definition of research agenda, transfer of research results to participants and adaptation to change in technological and economic environment. Agenda design suffers as the result of appropriability concerns and divergent member firm objectives. One solution is a decentralized structure and research agenda. Time horizon may be another problem because focus may be short-term. Technology transfer is aided by in-house investments and rotation of member firms’ personnel through collaborative projects. Adaptation is facilitated by keeping a strong, flexible administrative structure. The Foundation and Structure of Sematech Sematech founded by 14 companies in response to Japanese firms’ growing domination of global semiconductor market. Member firms make contributions and supply personnel. SEMI/SEMATECH facilitates linkages between US ME suppliers and Sematech. Original objective was process improvement, but this raised competitive concerns among members, so objective was changed to improvement of technological capabilities and strengthening of vertical cooperation between US suppliers and users of semiconductor process equipment. Knowledge diffusion and technology transfer became central objectives. SEMATECH developed “generic” technologies alike . 0.35 micron line width process technology. Intellectual property policy was liberalized in accordance with new strategy. Continued interest of members despite looser intellectual property policy suggests that some excludability still characterizes SEMATECH research results. SEMATECH also provides cheaper testing services. Still some conflict among members. SEMATECH was distinctive in concentrating research in one facility, having involved managers, and having received impetus for its creation from industry rather than government. Commitment of firms together with centralized management made the organization flexible. Member firms also stronger than West European counterparts. Evaluating SEMATECH Difficult to evaluate because agreement on goals and evaluation criteria are lacking, plus long time horizon. Also, many original goals no longer relevant to evaluation. SEMATECH has been credited with improving competitiveness of US semiconductor producers and equipment firms, but there is little evidence that it is directly responsible for the gains. Moreover, in lithographic “steppers,” SEMATECH programs have not prevented erosion of competitiveness. Assessment of technical performance suggests a more positive verdict. SEMATECH has met most of its revised objectives in the development of process technology and equipment, and most members have reported positive returns on their investments in SEMATECH Although SEMATECH’s emphasis on infrastructure has reduced appropriability problems among its members, some of these problems have shifted upstream to the ME industry. Also, SEMATECH can influence selection of a technology or individual firm. Also, its policies favor firms with more resources. Other concerns involve tension from contracts awarded to non-member firms and questions about efficiency of testing services. Perhaps greatest challenge in regard to vertical agenda, however, it the relative weakness of the financial and managerial resources of ME firms. Exclusion of foreign firms may be neither feasible nor desirable as other international joint ventures arise. Lessons of SEMATECH Numerous lessons. Near-term, supplier infrastructure focus means that SEMATECH R&D cannot substitute for government R&D. Also, where an industry’s problems reflect broader weaknesses such as lack of talent, incremental programs like SEMATECH may be insufficient.. Centralized structures may not be well suited to development of proprietary technology. Industry initiative and involvement are key. Can public funding be justified? Helped credibility, provided continuity, policy cooperation. However, little “dual-use” or “spin-on” value. Finally, some activities that involve little risk may not warrant government support but could be done by trade associations. Also, other funding mechanisms might be more desirable. Can “vertical collaboration” model of SEMATECH strengthen US suppliers and sectoral competitiveness in these and other US industries? Mayer: Sematech has moved from a horizontal research cooperation within the semiconductor industry to vertical collaboration between its members and US sources of semiconductor manufacturing equipment. It is also concerned with technology diffusion. Traditionally hypothesized advantages to collaborative research: (1) abilities of participating firms to lower costs and spread risks, (2) reduced duplication of R&D investment, and (3) exploitation of economies of scale in R&D. However, these hypothesized advantages have been empirically tested. A potential drawback to cooperative R&D is that it could reduce the diversity of R&D. Also, appropriability issues do not disappear with cooperative research--how quickly to spread benefits to non-members. Three keys to consortium design and management: (1) definition of the research agenda and choice of projects (most successful consortia have a short-term time horizon), (2) transfer of research results to participants, and (3) adaptation to change in the economic or technological environment. Consortia in Europe and Japan are compared to Sematech. European consortia tend to have small projects performed by a few firms, but are very resistant to change--the consortia is just an administrative body with no real power. Sematech conducts research in a central facility, while European and Japanese firms tend to work at the labs of member firms. Also industry is in charge of Sematech, while gov't is often in charge in Europe. Sematech also has a more limited focus, which is helpful. Another issue is that cooperation is harder once a country hits the technological frontier, because it is no longer clear in what direction to proceed. Sematech's research agenda shifted from development of state of the art processes to knowledge diffusion and technology transfer--which led to loosening of its intellectual property policies. Participation with Sematech does provide firms with an advantage in adopting Sematech-related technology because they are more familiar with it--kind of an absorptive capacity argument. There have been issues of alignment between the goals of industry participants and government. There have been tremendous gains by the US in the semiconductor and semiconductor manufacturing equipment industries since 1987, but many other factors have contributed to this--there is no direct evidence that Sematech is responsible for such gains. In fact, holding many factors constant, there is some evidence that Japanese suppliers of semiconductor manufacturing equipment have outperformed their US rivals. Experience in the semiconductor manufacturing equipment (SME) industry suggests that "...technology alone is rarely sufficient to restore the competitiveness of firms or industries that lack critical, complementary personnel, marketing, managerial, and financial resources." (pg. 15) Appropriability issues still exist, also with SME firms now that they are involved. Conflicting objectives for Sematech are again important--allow competitive advantage for semiconductor firms or spread SME technology. The authors proceed to argue that excluding firms from the consortia--either foreign firms or small firms by virtue of prohibitively high entry fees--is counterproductive. Lessons from Sematech: (1) Take a short-term view in which cooperative R&D is a complement to individual firm R&D, (2) may need to split up Sematech into smaller cooperatives that can pursue a broad array of projects with only minimal top level coordination , (3) maintain broad involvement and direction by member firms, (4) the authors justify public funding for Sematech (for reasons of credibility, continuity, catalytic effect, and existence of high uncertainty, long time horizons, substantial spillovers and external effects)--the authors do believe however that Sematech should have to compete for funding in order to stay competitive and focused. To maintain operations after 1996 (when they give up federal support), Sematech will have to allow more members of either smaller or foreign firms. Sematech also needs to maintain a short-term focus. Three keys to success of a consortia: (1) structure of the consortia, (2) political or economic expectations imposed on the consortia, and (3) 'fit' between the activities of the consortia and the competitive problems of an industry. Consortia are a complement to managerial and technical competence--not a substitute for these vital resources. INTRODUCTION Semiconductor Manufacturing Technology consortium (SEMATECH) Contrary to the predictions of Cohen and Noll (1992) SEMATECH has been followed by a number of other U