Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
4.3. The effects of military technology
The research and development effort for military technology is an important part of US innovative activities and of the government's technological policy. It deserves a specific discussion not only of its remarkable size but also of its quality and nature. The 'positive' effects of military research have been described as civilian 'spin-offs,' while the 'negative' effects can be summarized as a distortion of the research priorities, the pattern of innovation and the orientation of technical change.
The civilian spin-offs of military technology
Large expenditures for military research and development have a contradictory effect on economic and technological development. On the one hand, the US Defense Department research and procurement programmes have created a huge market, financed by public funds, for the emergence of new technologies. It is estimated that about 70 per cent of all Defense Department acquisitions of hardware over the past thirty years have been parts of high-technology systems, including missiles, aircrafts, space, electronics and communication equipment (DeGrasse 1984: 101). In sectors at the technological frontier, such as laser and space, according to Vernon, 'the US military establishment is likely to provide a market whose size cannot be matched in other countries' (Vernon 1982: 159).
On the other hand, the concentration of innovative activities in military projects diverts resources for the development of commercial technologies and increases the competition for limited resources in R&D funds, scientists, laboratories and specialized plants (Tirman 1984: x). Section 4.2 has documented the extent of this military drain of US innovative activities.
These worries are increasingly shared by American industry. According to the American Electronics Association, 'we cannot siphon off a disproportionate share of our skills and technical resources to military application and still stay ahead of Japan in commercial markets' (quoted in Melman 1986a: 65).
The acceleration of the pace of technical change that results from large military R&D cannot be separated from its direction. According to Nelson, the military and space programmes:
surely do not provide us with a model for future policies in support of high technology industries. That US procurement and procurement-related R&D had such a strong effect in building commercial leadership of US firms certainly does not provide a persuasive argument that we should augment our present defense and space programmes to increase 'spillover.' The massive expenditures we mounted then, and are incurring now, surely cannot be justified by the commercial returns.
(Nelson 1984a: 72)
The issue of civilian spin-offs is at the core of many studies on military programmes, from aircrafts to nuclear power, from semiconductors to computers. No general mechanism and pattern of technology transfer from military to civilian applications can be found, although institutional factors, funding for basic research, or procurement contracts at an early stage of development, played a role in the development of some of the new technologies. In the case of semiconductors, early procurement was the most important factor; in the early 1960s the military accounted for half the total sales of semiconductors in the US, a share that fell to 10 per cent in 1981 (Flamm 1984: 36). According to industry sources. half of all the R&D has been paid for by the US Defense Department in this way (ibid.). The lessons of this case, according to Rosenberg, are that
(1) The major innovations were not achieved on projects supported by military R&D. (2) Military R&D on possible alternative routes to miniaturization were largely spent 'betting on the wrong horses'. (3) The procurement needs of the military provided a pervasive and well-understood presence that served as a powerful inducement to innovative activity on the part of private firms spending their own R&D money.
(Rosenberg 1986: 18)
If the US military has been the successful 'midwife' of innovation in semiconductors, this does not mean that it is the legitimate father, nor that no other institutions may have played the same role. The case of nuclear power is an example of a highly unsuccessful technology on commercial terms that was developed from military research on nuclear weapons. The case of supersonic aviation is another example of the attempts to use in civilian areas possible spin-offs from the development of military supersonic aircrafts. This is another case of failure of spin-offs: the SST project by Boeing was never developed and the sixteen Franco-British Concorde airplanes that were produced had development costs of several billion dollars: 'the indiscriminate pursuit of military spillovers thus turned out to be a recipe for commercial disaster when optimal design requirements of the military and civilian sectors were sharply divergent' (ibid.: 24).
A growing divergence between civilian and military needs is evident in many areas, from aircrafts to integrated circuits, from satellites to space (ibid.: 24-5). This is also confirmed by a study commissioned by the ministry for Research and Technology of West Germany; using extensive documentation, the study found examples of transfer of technology from military to civilian uses only in very few cases, when there was a direct compatibility among the products. The possibility of spin-offs falls rapidly as the products develop through the 'life-cycle' and move to a greater differentiation in various applications (Krupp and Kuntze 1986: 27).
A reason of the declining importance of spin-offs is the growing applicative nature of military research; a spokesman for the group Business Executive for National Security' argued that in 1986:
when the Department of Defense will spend more than 32 billion dollars on research, development, evaluation and testing, only 861 million, or about 2% of the total, is to be spent on basic research which might be expected to further commercial technologies. Almost no commercial applications result from the development of particular weapons systems, only from basic research. In addition, the technologies developed for use in military systems are often too costly or sophisticated for commercial application.
(McKenna 1986: 7)
Military R&D is therefore characterized by low levels of productivity of the investment. Melman has reported that the US Commerce Department estimates that a commercial patent requires on average ten man-years of industrial R&D to be developed, and a thousands man-years for the R&D that the Defense Department and NASA contract out or perform in-house. (Melman 1983: 178). A study of the percentage of spin-offs from military R&D has estimated for the US a value between 5 and 10 per cent (ibid.).
Besides the individual interesting cases of civilian spin-offs, there are three basic factors, outlined by Tirman, that reduce their scope and effect: many of the products would have been developed anyway by the industry; the largest benefits have already taken place, as the greatest potential is when a technology is in the early stage of development; finally, any agency that may spend the amount of resources the Defense Department invests in high technology will inevitably lead to some commercial by-products (Tirman 1984: 221).
High military R&D expenditures are therefore a bad technological policy, a view on which there is broad agreement among the major US experts. Edwin Mansfield noted that, in spite of the spin-offs, 'the benefits to civilian technology seem decidedly less than if the funds were spent directly on civilian technology' (Mansfield 1980: 589).
According to Richard Nelson, 'the large spillover from the defense and space programmes of the late 1950s and 1960s was the product of a rather special set of circumstances.... Many analysts have suggested that spillover has diminished markedly since the mid-1960s' (Nelson 1984a: 72).
Nathan Rosenberg added an additional criticism: 'although the beneficial spillovers from the military and space programmes to the civilian sector are often cited, far less attention has been given to their possible deleterious effects in raising the costs of civilian R& D and in reducing the sensitivity of American engineers to cost considerations' (Rosenberg 1982: 284). Such a spin-off can make a lot of engineers unable to produce for competitive cost-conscious markets, such as the civilian ones, with a major effect, this time a negative one, on the pace of innovation and on the direction of technological change.
The distorting effect of military technology
The development of military technologies has an effect on the direction of technological change that goes beyond the simple diversion of resources from civilian innovation. A set of factors - basic principles, technological preferences, performance requirements, nature of the demand - have a strong effect on the kind of technologies developed by the military, in ways that have reduced efficiency, slowed down civilian applications and distorted the overall direction of technical change.
How is it possible to document these effects? The inefficiency of technological systems developed on the basis of military requirements, in the case of numerically controlled machine tools and nuclear reactors, has been shown by detailed reconstructions of their development and by international comparisons with the same technologies developed on other countries in a civilian environment.
David Noble has documented how the development in the 1950s of numerically-controlled machine tools at the Massachusetts Institute of Technology with the funds of the US Air Force has led to machinery that offered a strong centralization of control and wide versatility, while ignoring cost constraints: 'in an effort to meet Air Force specifications therefore, the industry ended up with perhaps the more complex and expensive approach to N/C (numerical control) then available' (Noble 1984: 203).
Cost constraints have made the application of these machines extremely limited in civilian sectors. In 1978, twenty years after their introduction in 1958, only 2 per cent of metal-working machinery in US manufacturing industry was numerically controlled. Only in the aerospace industry, largely financed by the Defense Department, the ratio was 6 per cent. According to Melman, the military nature of their development 'probably had the long-term effect of severely retarding the adoption of advanced technology in the metalworking industry' (Melman 1983: 107). This delay in the diffusion of numerical-control systems allowed Europe and Japan to reduce the advantage of the US in this field, and made easier their successful effort to export their machine tools in the US market.
A second example is that of nuclear power. US nuclear reactors were developed after a strong R&D effort by the US Navy, without exploring alternative designs and without competitive mechanisms, that have led, according to various studies in this field (G. Thompson 1984) to the failure of the industry. In the US, 'the design fostered by the US Navy was heavily promoted by the Atomic Energy Commission, a design flawed in many respects. The consequence, as we have seen in the 1980s, is a wholesale economic disaster compared with the widely held expectations for the technology' (Tirman 1984: 217).
The main current example of the distorting effects of military technology is provided by manufacturing automation. The same tendency towards eliminating the control and presence of production workers that had so deeply marked the development of numerically controlled machine tools has led to a major involvement of the US military in a variety of industrial automation programmes, from the Air Force Integrated Computer-Assisted Manufacturing (ICAM) aiming at a 'workerless factory' (Melman 1983: 236), to the Navy's Rapid Acquisition of Manufactured Parts (RAMP) (Business Week, 20 April 1987, p.60; see section 3.2 above).
The effects of such a form of industrial automation can be seen in the project sponsored by the US Air Force for producing the B1-B bomber in a 'factory without workers.' Vought Aero Products, a division of LTV Aerospace & Defense Co. has built a $10.1-million factory with eight Flexible Manufacturing Systems that can build 564 parts of the bomber. The system has been in operation since July 1984 and it has been considered 'an absolute success' (Business Week, 3 March 1986). But the bombers so far completed have shown major defects in the electronic systems, fuel leaks and an early need of parts replacement; repairing them is expected to take two years and $7 billion, on top of the $27 billion already spent for the first 100 aircrafts (Business Week, 24 November 1986, p.47). From being the gem of US military technology, the B1-B bomber has become another major scandal of US military procurement.
This time, however, on the road to militarized high technology, the US is not alone. In West Germany, Messerschmitt-Bolkow-Blohm is producing parts for the European MRCA-Tornado aircraft in a factory with twenty-eight computer-controlled machining centres and milling machines, and automated transport systems for tools and components (Schneider 1984: 160).
The defects, cost overruns, and the frequence of breakdowns of military high-technology systems provide additional evidence of the distortion of military technology. Many studies have listed the innumerable cases of malfunctioning, inefficiency, long down time and waste in weapons systems of all kind (Kaldor 1982a; Melman 1983; Adams 1982), proving the systematic nature of the problems of military technology.
Two recent examples are the US fighter aircraft F-15 and the Aegis anti-aircraft system for the destroyers of the US Navy. The former, with 127 individual electronic units, is not available, due to maintenance and other work, for 45 per cent of the time. The latter is available only for 42 per cent of the time, and its software works only for two and a half hours between failures (Melman 1986a: 66). The seriousness of these problems has been recognized also by the official report of the US Commission on defense management headed by David Packard, former deputy secretary of defense and president of Hewlett Packard corporation. The report admitted that 'weapons systems take too long and cost too much to produce' and that 'too often they do not perform as promised' (The Economist, 8 March 1986, p.34). Furthermore, the performances required are designed for imaginary combats; in 1980 the US Defense Science Board noted that the 'Defense Department creates requirements to meet threat projections that often do not materialize' (ibid.).
Extreme design sophistication and strong centralization of control over production are key characteristics of military-oriented technological development. Its very high costs, however, continue to limit the applicability of its results to civilian industry, which pays the price of the diversion of resources, loss of efficiency and distortion of innovation. In the factories, this results in workers losing their skills and control over the production process. In the educational system and the labour market for qualified engineers new distortions emerge and, according to Rosenberg, this
may have had a significant effect in slowing the rate of productivity growth and in contributing to declining international competitiveness that has plagued the American economy in recent years. The possibility needs to be seriously entertained that large-scale military R&D programmes do not offer a solution to these intractable economic problems; rather, they may have constituted a serious part of the problem itself.
(Rosenberg 1986: 29)
The effects of military technology on innovation and on the economy have been summarized by Mary Kaldor with the notion of the 'baroque arsenal':
Modern military technology is not advanced; it is decadent. Over the years, more and more resources have been spent on perfecting the military technology of a previous era. As a consequence, modern armaments have become increasingly remote from military and economic reality. They are immensely sophisticated and elaborate; they are feats of tremendous ingenuity, talent and organization; and they can inflict unimaginable destruction. But they are incapable of achieving limited military objectives and they have successively eroded the economy of the United States and the economies of those countries that have followed in her wake.
(Kaldor 1982a: 1)
This has a major impact on the economy: 'baroque military technology artificially expands industries that would otherwise have contracted. It absorbs resources that might otherwise have been used for investment and innovation in newer, more dynamic industries. And it distorts concepts of what constitutes technical advance' (ibid.: 3).
These characteristics of military technology have affected the development and commercial success of many technologies of the past, and they are now influencing the direction of progress at the technological frontier, in areas such as microelectronics, computers, optics, telecommunications, industrial automation, space and materials research. Together with market relations, the logic of military technology can be traced in the technological strategies of corporations and governments, which are examined in the following sections.
4.4. The technological strategies of corporations
The technological performances of the US, Europe and Japan are the result of the activity of corporations and of government policy. Firms are the place where most innovative activities take place and where new products, processes and forms of organization are introduced, but the strategy of corporations, for R&D as well as for production and marketing, is increasingly international. Innovation takes place more and more within a network of relations with research institutions, universities, government laboratories and other firms. Two apparently contradictory processes have developed here. On the one hand, there is a growing internationalization of the technological strategies of corporations, with greater and more rapid flows of technology and greater integration of innovative activities at a world scale. On the other hand, there is strong pressure towards national co-operation in R&D between firms, government agencies and research institutions, in a strategy to advance a country's technological level.
The internationalization of technology
The development of new technologies has been associated with a growth of the international flows of technology. A study by Mansfield et al. on a large number of US firms found that US-based multinational firms are transferring their technology to their foreign subsidiaries much more quickly than in the past. In 1969-78, about 75% of the technologies in our sample that were transferred to subsidiaries in developed countries were less than five years old; in 1960-68 the proportion was about 27%' (Mansfield et al. 1982: 209).
These greater flows of technology facilitate the imitation by other countries. The study found also that
in about one-fourth of the cases, the technology transfer seemed to hasten foreigners' access to these technologies by at least two and a half years. And in about one-third of the cases, the technology transfer resulted in at least a two-and-a-half-year reduction in the length of time elapsing before a foreign competitor imitated the innovation.
(Ibid.: 213)
The degree of internationalization of firms is greater in the high-technology sectors. Analysing a sample of fifty-seven US multinational corporations, Vernon found that the innovations are transferred more rapidly than in the past, particularly in the firms with higher R&D expenditures (Vernon 1982: 151). The cost of the technology transfer within a multinational firm has been estimated at 20 per cent of the cost of establishing a plant abroad (Mansfield et al. 1982: 215).
These greater technology flows have led to a different distribution of R&D activities within multinational corporations. The study by Mansfield found also that by the mid-1970s 'about one-half of the firms we studied felt that worldwide integration of overseas and domestic R&D had been achieved' (ibid.: 210). This evidence puts the relative decline of innovative activities within the US in a different perspective. In the early 1970s US multinational corporations controlled, through their foreign subsidiaries, one-seventh of all industrial R&D in West Germany and Britain, and a half in Canada (ibid.).
The growing divorce between the performance of the US national economy and that of US-based multinationals, already discussed in section 3.3, is developing also in the field of technology. The presence of large US firms in the foreign markets that are emerging as new sources of innovation puts them in the position to benefit from the new, reversed flows of technology. This has led to a growing number of agreements among firms of different countries for R&D projects and investment in new laboratories and plants. Japanese and European firms are expanding their research activities in the US, and US firms are doing the same in Japan and Europe. This pattern will be investigated in detail in the case of two major industries, semiconductors and telecommunications, in the next sections.
The scale of the new R&D co-operation among firms is unprecedented; according to a OECD report, companies 'have not previously cooperated so directly, on such a scale, in planning, financing and carrying out joint R& D' (Brainard and Madden 1985: 64). This leads to new business strategies, besides the traditional pattern of takeovers, mergers and joint ventures. Joint development agreements, sharing of research and design resources are increasingly frequent, together with cases of large firms investing as minority shareholders in small, dynamic companies, which are allowed to maintain their independence and innovative capacity, while offering the large company access to their results (ibid.: 66).
These transformations, according to Nelson, are leading to more integration and closer inter-firm relations at an international level (Nelson 1984a: 76). Vernon has argued that this will not necessarily result in a growing degree of concentration, as the current trend to mergers among national firms in high-technology industries is developing at a slower pace than the internationalization, with the entry of new foreign firms in national markets (Vernon 1982: 150).
The national co-operation in R&D
In contrast to the growing internationalization of innovative activities, within the national boundaries of all advanced industrial countries, there is evidence of an increased co-operation in the research and development of new technologies between firms, governments and research institutions. A network of relationships between industry, universities and government agencies has been a permanent characteristics of the R&D and innovation system in advanced capitalist countries (Noble 1977). In many cases, what used to be an informal network with a clear separation of roles, is transformed into close co-operation. The risk, the size of the R&D investment and in some cases the strategic importance of technological projects in key areas, have drawn together formerly competing firms, government funds and research institutions.
In Europe and Japan the intensification of co-operative R&D projects does not represent a qualitative difference with the past, as technological strategies have always been characterized by a strong role of governments in coordinating and funding private research. It is in the US that the emergence of co-operative strategies has marked a turning point, breaking with a competitive tradition that used to be strengthened by the anti-trust legislation. Corporate strategies are in this way becoming increasingly close and intertwined to the technological strategies of governments.
Striking similarities in the research fields, and deep institutional differences have emerged in a comparative study of US, European and Japanese high technology programmes (Pianta 1988).
The research on computers of a new generation illustrates this trend. In Japan the project of the Ministry for International Trade and Industry (MITI) for the 'Fifth-generation computers' has been a classical example of the co-operation between government and corporations that is common in Japan. Launched in 1981, it involves an expenditure of $450 million over ten years (The Economist, 24 August 1986, p.13). Europe and the US have made efforts in the same direction. The EEC countries have developed the 'Esprit' programme (European Strategic Programme of Research in Information Technology), which covers a wide area of research, and is funded by $800 million of EEC funds, which must be matched by an equal amount from the firms involved in its projects. Started in 1984, 'Esprit' has led in its first year to 104 co-operative projects that have involved 270 European firms, including affiliates of US multinationals. Its focus is pre-competitive research, but there are pressures to move towards demonstrative projects for commercial products (Brainard and Madden 1985: 65).
In the United States, the challenge for the new-generation computers has been taken up first of all by the Defense Department, with the 'Strategic Computing Programme' of DARPA, the agency specialized in directing high-technology defence projects (ibid.: 59). DARPA has spent $300 million over the past five years on military Very High Speed Integrated Circuits (VHSIC) and plans to spend $1 billion on research on supercomputers (The Economist, 23 August 1986, p.13). This response is typical of the US technological strategy that is discussed in section 4.7.
More unusual has been the initiative of twenty-one US electronics firms - all the major ones, with the exception of IBM, for antitrust fears - that for the first time pooled together resources and in 1983 formed the Microelectronics and Computer Technology Corporation (MCC). Each company contributes to the $75 million annual budget, financing the research performed by 250 scientists in the new MCC laboratories in Austin, Texas. The four areas of research are systems architecture, software engineering, manufacturing of microprocessors and computer-aided design of Very Large Scale Integrated Circuits (VLSI). The ten-year systems architecture project has the same objectives as the Japanese programme (ibid.).
As US legislation prevents full co-operation among firms in the development of a product, MCC focuses on pre-competitive research; MCC will own the patents on the results and will license them to the individual companies of the pool, who will then carry out the development and production (Brainard and Madden 1985: 64; The Financial Times, 27 January 1986). The degree of success of this experiment is still uncertain, and the founder of MCC, Admiral Bobby Inman, resigned as chairman at the end of 1986 (see also Pianta 1988), but the same model of the MCC is now being proposed for the semiconductor industry, as we will see in the next section.
Behind the apparent paradox of growing internationalization and growing national co-operation, there is a common tendency to greater inter-firm agreements. Both processes search for the size and scale of research efforts that is needed, with government help, to remain at the technological frontier. Two classic cases of these apparently opposing tendencies will be presented in the next two sections: the semiconductor industry, that has been dominated by the internationalization of R&D and production; and the telecommunications industry, characterized by close government control and co-operative national research.
4.5 The case of semiconductors
The semiconductor industry offers a major example of corporate strategy in a new technology that has been deeply characterized by the internationalization of R&D and production. A recent OECD study (Ypsilanti 1985) has examined the forms of technology transfer, from direct investment to joint ventures, licences and technology agreements. The summary of the processes is that
firstly there was a period when American firms dominated the industry and were involved in direct overseas investment and some licensing agreements. Secondly, there was a period marked by the emergence of Japanese competition and the increasing drive of European firms to obtain technology. Finally, in more recent years, there has been an increasing emphasis on direct investment and on international cross-licensing between firms which are at a similar technological level.
(Ypsilanti 1985: 48)
By documenting the many foreign investments, joint ventures and inter-firm agreements, the study draws a picture of highly dynamic corporate strategies, large international technology transfers and rapid changes in the relative position of individual firms, with regard to their specialization, R&D activities, production and marketing strategies.
The picture of direct investments shows a slow down of those by US firms in East Asian countries, a growing presence of European and Japanese firms in the US market and of Japanese companies in Europe. The agreements that are reviewed between 1979 and 1983 are largely for licensing and joint ventures, aiming to get access to new markets and to improve the technological capacity of each company, with greater integration upstream or downstream of production. These agreements have become more frequent after 1983, marking a change in the industry structure (ibid.: 51).
The web of agreements around Intel, a major US semiconductor company, is indicative of what is happening in the industry. Around the ownership by IBM of a share of Intel, there is an extensive network of inter-firm relations. Eleven electronics companies have licenses from Intel (Sanyo, Oky, Nec, Signetics, Mitsubishi, Matra-Harris, Toshiba, Memorex, Siemens, Amd, Harris). Three others (Fujitsu, the former Burroughs and Texas Instruments) have developed agreements with either IBM or Intel. Furthermore, there is extensive cross-licensing between IBM, AT&T and NTT (ibid.: 55). This network is also constantly changing, both in terms of the technology agreements and in terms of equity ownership.
In 1986 these international links have increased further with twenty-seven agreements between Japanese and Western companies (Business Week, 20 April 1987, p.62). The main one took place between Motorola, the US company leader in logic chips, and Toshiba, the Japanese leader in memory chips; they plan to exchange technology and open a joint factory in Japan. Other links were established between American Micro Devices and Sony; LSI Logic and Toshiba; Boeing and NEC; Siemens, Toshiba and General Electric (ibid.).
In this international restructuring of semiconductor R&D and production, the crisis of the US industry should not be overlooked. While in 1980 fifteen US firms were profitably producing a very large share of world chips, in 1987 only three were left; they had total losses of about $800 million in 1986 and since 1981 65,000 workers lost their jobs in the industry (Reich 1987: 65).
The situation in Europe is equally difficult. In 1985 European semiconductors represented only 13 per cent of logic and memory chips and 6 per cent of microprocessors (The Economist, 22 February 1986). Half of the European market is controlled by US and Japanese producers, and the European capacity in specialized productions is declining (ibid.).
The confusion on the European scene is stressed by Ypsilanti, who pointed out that 'the requirements of European firms do not necessarily coincide with an industrial strategy for a "unified" European semiconductor industry. This is true not only in the case of semiconductors, but also in other branches of the electronic industry (computers, telecommunications)' (Ypsilanti 1985: 59).
The strategy of European firms has focused on establishing new direct links with US companies, in a variety of ways, from direct investment, acquisitions of firms and joint ventures, to licensing agreements, second sourcing and co-operative research. Among the major European companies, Philips bought Signetics, Siemens acquired five small US firms and Thomson invested in Mostek (ibid.: 59-60; Business Week, 20 April 1987, p.63). Growing links are also developed within Europe, with the $1 billion project of Siemens and Philips for the next generation of memory chips, and the merger between Thomson and Italy's SGS.
The free-wheeling internationalization of the semiconductor industry has now halted. In July 1986 the US and Japanese governments reached an agreement on market shares and prices that led, for the first time, to a 10 per cent increase in semiconductor prices (The Economist, 6 September 1986). In February 1987 a report by the US Defense Science Board criticized the 'unacceptable' reliance of US military on foreign semiconductors and proposed a $2 billion programme to support American producers. Estimates by Reich suggest that about 40 per cent of the value of advanced electronic equipment in US military systems now comes from Japan, and at the present trend, the share will reach 55 per cent in 1991 (Reich 1987: 64).
Then, in March 1987, the US government prevented the takeover of Fairchild by Fujitsu that was announced in October 1986. Fairchild, a US company owned by a French family, was almost bankrupt and in search of a buyer, but a large share of its production consisted of military chips for the Defense Department. Its sale to the Japanese was considered a threat to 'national security', and the US government intervened, sending a clear signal on the limits that the internationalization of the industry should not trespass. Furthermore, citing the Japanese failure to comply with the market-sharing agreement, the US government imposed unprecedented trade sanctions against Japanese electronic goods, opening a 'trade war' across the Pacific. Few months later, the Fairchild case ended with its acquisition by the US company National Semiconductor (Time, 14 September 1987, p.37).
The intervention of the US government in support of the domestic chip producers gave new weight to the project of the Semiconductor Industry Association, the US trade group, for creating a Semiconductor Manufacturing Technology Corporation (Sematech) along the lines of the Microelectronics and Computer Technology Corporation (MCC) described in the previous section. All large US manufacturers of semiconductors would join in a major programme for developing the manufacturing system of the next generation of chips. The $1 billion needed for the project may come from the new funds the US Defense Department is prepared to spend in support of American semiconductor firms, and an initial funding of $100 million a year was approved by Congress early in 1987 (Business Week, 20 April 1987, p.63).
This project would dramatically expand inter-firm co-operation in the US, that is presently limited to the Semiconductor Research Corporation, a non-profit consortium of thirteen companies funding, with $35 million a year, university research on advanced integrated circuits (The Economist, 23 August 1986, p.13; see also Pianta 1988).
The lessons are that a new phase is now opening for the semiconductor industry; a much greater role will be played by inter-firm co-operation and by the state, not only in Japan and Europe, but especially in the United States, in controlling and subsidizing the industry, negotiating market shares and directing technological strategies, in an industry where the corporations' strategies had led to rapid internationalization.