Conflict and Cooperation in National Competition for High-Technology Industry (1996)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

SUPPLEMENTS 131 Supplements SUPPLEMENT A. HIGH-TECHNOLOGY COMPETITION IN SEMICONDUCTORS The history of the semiconductor industry beginning in the late 1970s illustrates several aspects of international competition in high-technology industries.339 The stakes of this high-technology competition were enor- mous. It is not widely recognized, for example, that the U.S. electronics industry, including semiconductors, is larger than the U.S. steel, automo- bile, and aerospace industries, combined. Employment in the semiconduc- tor industry and related industries is equally significant. U.S. semiconduc- tor makers employ nearly 240,000 people in the United States alone and semiconductors are the enabling technology for the nearly $400 billion U.S. electronics industry, which employs approximately 2.5 million Americans. In addition, the manufacturing capability of the industries mentioned above and, therefore, their global competitiveness depend directly on timely ac- cess to new, electronic manufacturing technologies. Moreover, because the strategic significance of the industry is widely recognized by policymakers around the world, “the semiconductor industry has never been free of the visible hand of government intervention.” 340 A Catch-Up Strategy From its position as a late industrializer, the Japanese government recog- nized the strategic importance of semiconductors to its economy, in terms both of economic growth and national autonomy. In the 1970s Japan made a systematic effort, under MITI guidance, to promote a domestic semicon- ductor industry.341 Building upon their experience in consumer electronics, 339 This section draws heavily from the contributions to Committee deliberations of George M. Scalise, The Nature of High-Technology Competition, 13 December 1996 and William J. Spencer, Technology (Transfer) at SEMATECH. See also the excellent analysis presented by Nortel’s Claudine Simson, Samsung’s Y.S. Kim, and Motorola’s Owen Williams in the discus- sion led by BRIE’s Michael Borrus to the conference Sources of International Friction and Cooperation in High-Technology Development and Trade, 30–31 May 1995. 340 Laura Tyson, Who’s Bashing Whom? p. 85. For a comprehensive review of the exten- sive and ubiquitous government programs designed to develop and support the technologies underpinning the semiconductor industry, see Thomas Howell et al., Creating Advantage, passim. 341 See the partial list of joint research and development projects in microelectronics sponsored by MITI in Laura Tyson, Who’s Bashing Whom? p. 96. Many of these programs continued through the 1980s. In addition to the VSLI manufacturing program, MITI focused on promising technologies such as optical semiconductors, high-speed computing devices, 

132 CONFLICT AND COOPERATION the large Japanese electronics firms invested heavily in the development of semiconductor technology. By the early 1980s Japanese producers had become a major force in the market for dynamic random access memories (DRAMs).342 Escalating Trade Friction The vertically integrated structure of Japanese industry provided major advantages with respect to the capital-intensive investments required for state-of-the-art semiconductor facilities. In this high-tech competition, the Japanese companies benefited from their unique system of corporate gover- nance, which places considerably less emphasis on quarterly or annual prof- its than does the U.S. system. Moreover, the rich revenue stream derived from previous success in penetrating the consumer electronics industry per- mitted cross-product subsidization within companies. At the time, a lower cost of capital encouraged the major Japanese electronics firms to make major counter-cyclical investments. These structural advantages enabled the Japanese industry to undertake a massive capacity build-up in the early 1980s. The Japanese firms then accelerated their gains in market share through highly aggressive price cut- ting. The U.S. industry sustained progressively heavier losses through each generation of DRAM. By 1984–85 with the introduction of the 256K DRAM, the market share of the U.S. industry had gone from roughly ninety percent of the DRAM market in the late 1970s to less than ten percent, with most U.S. synchrotron lithography, and optoelectronic devices, among others. Japan is not alone in providing R&D support in this sector. In the United States, the Defense Advanced Research Projects Agency (DARPA) has extensive technology development programs, though these are generally small scale and focused on “over-the-horizon” technology development, with an emphasis on defense requirements. 342 This strategy was not without its costs. While direct central government expenditure was limited, though important, these programs presumably involved costs in terms of foregone consumer welfare or opportunity costs imposed on other industries, as a 1988 study by Baldwin and Krugman sought to determine. However, this concern for consumer welfare, however laudable, reflects values and assumptions characteristic of American-trained economists, which are not universally shared. See James Fallows, Looking at the Sun, chap. 4, especially pp. 180–190, and the section Producer-Oriented Economies above. As Laura Tyson notes, “it is important to emphasize that the objective of the Japanese industrial policy in the semiconductor industry, as in other industries targeted for development, was not consumer welfare.” (Pp. 86–87, italics added.) Tyson also notes that the Baldwin and Krugman study did conclude that the Japanese policies “succeeded in gaining a share of the 16K DRAM market for Japanese producers, but only at the expense of a net reduction in Japanese economic welfare.” Laura Tyson queries the import of their findings because of their failure to take into account the possible dynamic effects of a Japanese position in the growing DRAM market on profits and technological externalities in future generations of semiconduc- tor products and in systems incorporating semiconductor technology. 

SUPPLEMENTS 133 firms exiting the DRAM market entirely. A similar though less severe trend occurred in erasable programmable read-only memories (EPROM). The dras- tic effects of the Japanese competition led many informed U.S. observers to question the future viability of the U.S. semiconductor industry.343 The Japanese challenge to the U.S. industry was based on three inter- locking elements. Foremost among these was the quality of Japanese pro- duction processes and products. The second, which helped drive the first, was the higher rates of investment made possible by the more integrated, capital-rich structure of the Japanese companies. A third key advantage enjoyed by the Japanese industry in comparison with the U.S. industry was its ability to exclude its global competitors from access to its domestic market. The formidable barriers to investment and the lack of market ac- cess for foreign producers of semiconductors provided a crucial benefit in an otherwise global industry.344 It gave the Japanese industry a sanctuary, from which they could insulate themselves from competitive counter- attacks and preserve a financial cushion for future product development. This enabled Japanese producers, operating from the closed Japanese mar- ket, to dump product on the global marketplace. Because of the shortness of the semiconductor product cycles and the speed with which the foreign markets responded to dumping, the impact of dumping on their U.S. com- petitors was pronounced. The U.S. Industry Response to the Japanese Trade Challenge Leading figures in the U.S. semiconductor industry increasingly recog- nized the unique features of the Japanese competitive challenge. The com- bination of a closed Japanese domestic market, the dumping encouraged by the excess capacity in microelectronics products (encouraged by the capital- rich Japanese keiretsu), and the prospect of supply dependency for key components with the accompanying vulnerability to manipulation of price and availability posed a competitive threat that they were unlikely to sur- mount independently. To meet this threat to the future viability of their industry, several U.S. companies joined in the mid-1980s in filing dumping actions against the Japanese semiconductor producers with respect to 64K DRAMs and EPROMs. In 1985 the Department of Commerce self-initiated a case in reaction to 343 Laura Tyson provides an excellent analysis of the competition for dominance in the semiconductor industry. See Laura Tyson, Who’s Bashing Whom? chap. 4, “Managing Trade and Competition in the Semiconductor Industry,” pp. 85–113, as does the recent study by Kenneth Flamm, Mismanaged Trade? especially chap. 3, 4, and 5. 344 Tyson, p. 87. 

134 CONFLICT AND COOPERATION Japanese dumping of DRAMs of 256K and above. During roughly this same period, the Semiconductor Industry Association brought an action against the Japanese semiconductor producers under Section 301 of the Trade Act of 1974 to address the closed nature of the Japanese semiconductor market. The dumping cases resulted in findings of substantial dumping and rea- sonable indications of injury. The remedy from these cases normally would have been high antidumping duties imposed on the products at issue. How- ever, the U.S. semiconductor industry viewed such a remedy as both inad- equate and flawed. Antidumping duties were considered inadequate because before they could be imposed, most U.S. producers could no longer justify investing in new DRAM capacity and were forced to leave the DRAM market. No compen- sation would be provided to the U.S. industry; the effect of the antidumping duties would be prospective, that is, only to offset future dumping. And antidumping duties on a previous- or current-generation product was little deterrent to the dumping of subsequent products, especially if over-capacity and low marginal costs added to an incentive to capture market share by dumping. Equally important, antidumping duties were a flawed remedy for an in- dustry such as semiconductors. Given the resulting rapid decline in U.S. DRAM production, antidumping duties would impose a burden on the U.S. customers of the semiconductor industry, either by raising the price of DRAMs to them or by forcing them to shift to overseas production of their electronic systems products. To meet this policy challenge, the American producers adopted a novel approach, ultimately accepted by the American government.345 This ap- proach did not seek to protect domestic producers of semiconductors from global competition nor to insulate the U.S. market from technological ad- vance, which would ultimately disadvantage their U.S. customers. Rather than close the U.S. market, the U.S. producers sought to ensure access to the Japanese market for U.S. and foreign suppliers. The U.S. approach sought not only to bring an end to dumping and a return to cost-based competition in the U.S. and other foreign markets, but also to remedy the problem of effective market access in Japan. Previous efforts to obtain equivalent access to the Japanese market had proved elu- 345 While the Semiconductor Agreement and SEMATECH represent innovations in U.S. trade and technology policy, the Reagan administration arguably had a much more pragmatic and innovative approach to economic policy issues than is generally recognized, even when compared with the announced “results-oriented” trade policies of the Clinton administration. See, for example, Alan Tonelson, “Beating Back Predatory Trade,” Foreign Affairs, July– August 1994. Tonelson argues that the efforts to support the semiconductor industry were but one element of a series of successful policies to rescue and restore beleaguered U.S. industries such as steel, machine tools, textiles, and automobiles. 

SUPPLEMENTS 135 sive. No matter how market conditions changed or how well U.S. products competed elsewhere in the world, the U.S. market share in Japan generally stayed around 10 percent,346 well under U.S. shares in other markets. Ef- forts such as the formation in 1982 of an informal group, the U.S.-Japan working group on high-technology industries, did not prove successful. Al- though a series of recommendations by the group was adopted in 1983, by mid-1985 the U.S. share of the Japanese market was lower than it had been when the recommendations were first adopted. Recognizing the failure of this consultative approach and of the Japanese government to fulfill undertakings on market access, the U.S. industry filed a Section 301 action. Sanctions could be authorized against Japan under a Section 301 action for failure to live up to its agreements or for engaging in unjustifiable and restrictive trade practices. For the U.S. industry, however, the objective was to obtain effective market access, not to penalize the Japanese. The Semiconductor Trade Agreement In 1986, just prior to the conclusion of the various legal proceedings, the U.S. industry agreed to suspend the dumping cases, and an agreement was reached on the market access issue. The settlement took the form of a five- year accord called the Semiconductor Trade Agreement (STA). Impor- tantly, the Agreement included a side letter in which the Japanese govern- ment recognized the fact that the U.S. industry expected the foreign share of their market (not only U.S. producers) to grow and to exceed 20 percent by the year 1991.347 The Agreement and the policies associated with its imple- mentation included three principal features: • Provisions designed to lead to concrete market access for non-Japa- nese firms, • A specific target acknowledged as a threshold for foreign market share, and • A mechanism to monitor semiconductor cost so that dumping could be addressed quickly and effectively with the prospect of sanctions should the terms of this agreement not be respected. 346 As noted above, given the enormous R&D costs associated with the semiconductor industry and the resulting importance of competing in all markets to recoup these costs, access to the Japanese market, now the largest in the world, is considered essential for competitive- ness in this global industry. Council on Competitiveness, Roadmap for Results, p. 55. 347 The existence of the side letter was first revealed with the publication of Clyde Prestowitz, Trading Places, Basic Books, New York, 1988. For a discussion of the negotiations, see Kenneth Flamm, Mismanaged Trade? chap. 4. For a discussion of the impact of the side letter agreement specifying 20 percent as a reasonable expectation for foreign market share by 1991, see Kenneth Flamm, Mismanaged Trade? pp. 279–293. 

136 CONFLICT AND COOPERATION An End to Dumping The metrics of success for the STA were clear-cut: assured access to the Japanese market for foreign producers and an end to dumping. There would be no offsetting penalties for dumping and no restricted access to the U.S. market. Instead, dumping would cease and an effective deterrent mechanism was devised under which each individual producer would monitor its own cost data to detect dumping. For market access, an objective target for a minimum foreign market share was established at 20 percent of Japan’s annual semicon- ductor consumption. This was a target that focused on access in fact rather than in theory. Because market penetration could be measured, it provided a concrete measure of success. This represented a significant departure from traditional U.S. process-oriented trade policy to a more results-oriented ap- proach to market access; it also provided discipline. Through Government Commitments To be effective, the STA needed government commitment. If there was to be no U.S. government response to noncompliance—the common percep- tion in U.S. industry of the U.S. government’s behavior with respect to past trade agreements with Japan—there would be little prospect of the STA’s meeting its objectives. After widespread evidence of noncompliance with the STA became available in late 1986, President Reagan imposed sanctions in April 1987.348 Following this action dumping in world markets ceased, and the market share in Japan of U.S. and other foreign producers began to climb. By 1990 the increased share of the Japanese market made possible by the agreement translated into additional sales of over $1 billion for American companies alone.349 And Industry Cooperation The success of the STA also depended heavily upon industry commit- ment. The STA not only established a realistic target for market access, but 348 A Council on Competitiveness study concluded that a credible government commitment to sanctions may be necessary to ensure compliance with trade agreements, especially when they affect the interests of powerful foreign companies. The report notes that, “in the case of semiconductors, the U.S. government’s commitment to sanctions was not credible until they were actually applied in 1987. Japanese industry and government were reportedly very sur- prised by the decision.” Council on Competitiveness, Roadmap for Results, p. 56. The report adds that the U.S. government’s commitment to sanctions contributed to the subsequent suc- cess in winning an extension of the Trade Agreement in 1992 and to the dramatic increase in the purchase of U.S. semiconductors by Japanese users. 349 Laura Tyson, Who’s Bashing Whom? p. 113. These additional sales also meant signifi- cant increases in R&D, capital investment, and high-wage employment. 

SUPPLEMENTS 137 was accompanied by major efforts by both U.S. and Japanese industries to pursue the target. These measures were substantial and far-reaching. They included major commitments of marketing activity in Japan by U.S. indus- try.350 Collaborative efforts were undertaken to understand customer re- quirements and product capabilities. These were accompanied by competi- tions at the design-in level. Extensive follow-up and product support were also made available. Supportive Policy Developments A parallel development during this period very much affected the viabil- ity of the U.S. semiconductor industry: the emphasis on intellectual prop- erty protection (including the enactment of a chip protection law in the United States at the end of 1984 and in Japan beginning in 1986). Intellec- tual property protection is essential to commercial development in semicon- ductors and to the large R&D investments necessary to generate new prod- ucts. The chip protection legislation created a new form of intellectual property protection for semiconductor products that was readily adopted internationally. It effectively reaffirmed the importance of general prin- ciples of trademark law and reduced the risks of piracy in semiconductor design. In addition, despite considerable controversy within the U.S. gov- ernment, the decision to establish SEMATECH, a consortium of U.S. semi- conductor manufacturers, backed by the Department of Defense and a sub- stantial annual commitment of funds, sent a powerful signal to foreign producers of semiconductors that the U.S. government was not prepared to see the U.S. semiconductor industry marginalized. Both these developments in- creased the confidence of the U.S. industry and the trade policy community that the cost-based, market access approach of the STA would have an opportunity to work. Success of the Trade Agreement The STA was an unprecedented mechanism to increase competition and reduce friction in a key high-technology industry. Without this effective trade defense, combined with the STA’s innovative market-opening mea- 350 The Council on Competitiveness study emphasized the role of marketing efforts of U.S. semiconductor producers, noting that “as a signal of their commitment to the Japanese market, 125 U.S. semiconductor companies established a presence in the country, and over the 1986– 1992 period, U.S. firms opened an average of one new facility per month in Japan. In general, U.S. industry undertook a wide variety of efforts to market its products in Japan and made great efforts to improve its knowledge of, and exposure to, the Japanese market.” Council on Competitiveness, Roadmap for Results, p. 54. 

138 CONFLICT AND COOPERATION sures, the structure and character of the U.S. industry would be very differ- ent today, and its contribution to the U.S. economy significantly smaller.351 Indeed, some informed participants believe both that the U.S. industry should have forged a collective response earlier and that the government, in turn, should have been more responsive to the needs of this strategic industry. While the policy process ran its course, the U.S. industry suffered signifi- cant damage; thousands of jobs and billions of dollars in revenue were lost before the U.S. responded.352 Although some of the elements of the STA remain controversial, overall the agreement has been very positive.353 Costly and destructive dumping has ceased. U.S. producers of DRAMs and EPROMs have survived and rebounded as world-class competitors. Major new entrants into DRAM production, notably producers in Korea and, increasingly, Taiwan, have substantially increased competitive vitality in the world DRAM market. U.S. and foreign producers have achieved far greater access to the Japanese market. While this progress remains vulnerable to arbitrary reversal, its underpinnings are market based. It has provided both major long-term benefits to global consumers and a competitive environment in which the U.S. industry could recover and prosper. An Agreement Oriented to High-Technology Competition In part, the success of the STA can be attributed to its sensitivity to the peculiar features of high-technology competition. Competition had to be preserved rather than merely offset by government intervention. By requir- ing future competition to be cost based, the STA strengthened market com- petition while minimizing government intervention. However, swift and 351 Ibid., p. 42. 352 Laura Tyson, Who’s Bashing Whom? p. 271, and the Council on Competitiveness study emphasize the special characteristics of high-technology industries and the need for quick decisionmaking processes in responding to trade problems. The report adds that “the U.S. semiconductor industry may have been lucky to survive delay. Other technology-intensive industries may not be.” Roadmap for Results, p. 55. 353 Some observers argue that the agreement was a failure, resulting in significantly higher prices for U.S. consumers of semiconductors and in a windfall gain in profits for Japanese producers. In this view, the agreement’s “bubble profits” for the Japanese producers were plowed back into R&D investment, having the “perverse result of strengthening the Japanese companies for future rounds of competition in new products.” (Laura Tyson, Who’s Bashing Whom? p. 117.) For a critique of the agreement, see David Mowery and Nathan Rosenberg, “New Developments in U.S. Technology Policy: Implications of Competitiveness and Interna- tional Trade Policy.” California Management Review, vol. 32, no. 1, Fall 1989, pp. 107–124. For a positive assessment, see Thomas Howell, et al, Creating Advantage, Laura Tyson also rejects many of these criticisms. See Laura Tyson, Who’s Bashing Whom? p. 87 and p. 132 cited below. 

SUPPLEMENTS 139 predictable government action was a necessary condition of success. The STA provided it in the form of a mechanism that was capable of detecting and combating dumping before the targeted firms were effectively removed as players in global competition. The willingness of the U.S. government to take action, as noted above, was a crucial condition for success.354 Moreover, the policy approach adopted by the American government had the considerable advantage of being genuinely international. The STA ac- complished this by taking a comprehensive approach that addressed, for example, third-country dumping and sought market access for all non-Japa- nese firms, not just U.S. firms. Finally, the arrangement had to be based on market realities. The needs of U.S. and Japanese customers were taken into account, and genuine industry-to-industry interaction, design-ins and tech- nology adaptation were undertaken. This progress was a joint effort. In part, it was facilitated by the ex- panded commitment of U.S. semiconductor firms in Japan. The number of technical personnel was significantly increased, and new design centers were opened at the time of the conclusion of the second Semiconductor Agree- ment. However, MITI played a determining role, urging Japanese semicon- ductor producers to increase procurement of foreign components and to submit reports with purchasing plans for foreign semiconductors as a means of expanding foreign market share.355 The phenomenal growth of the world market for semiconductors has probably contributed to the success of the agreement, not least by reducing the zero-sum nature of its results. This has been especially true in recent years. For example, the world market for semiconductors in 1993 was on the order of $77 billion. By 1995, the market had nearly doubled, to $146 billion.356 Aided by this high-growth environment, progress under the agreement in terms of foreign share of the Japanese market was significant, rising in 1995 354 Thomas Howell argues that the delay in fulfilling the market access commitment made in 1986, until the U.S. retaliated in April 1987, should not be interpreted as proof that the Japanese respond only when confronted with sanctions. He suggests that the explanation is more complicated, noting that Americans tend to see the negotiated solution as final, with implementation flowing naturally, while Japanese tend to see the negotiated solution as one more stage in the negotiations, with implementation a subject of further negotiations. Howell et al., Creating Advantage, p. 89. It might be added that, given the lack of continuity in the upper levels of the U.S. government, the implementation negotiation is likely to be with a new American negotiating team, possibly having no prior experience with the subject at hand. 355 Howell et al., Creating Advantage, p. 93, citing Japanese press reports. See also p. 84. 356 Semiconductor Industry Association, World Semiconductor Forecast, San Jose, Calif., November 1995. Interestingly, U.S. market share has declined from a peak of 43 percent to 41 percent in 1995. Japanese market share is estimated at 39 percent, with the substantial remain- der made up by the rest of the world, notably Korean, Taiwanese, and European production of DRAMs. The DRAM market showed a phenomenal increase, rising from $13 billion in 1993 to $41 billion by 1995. Ibid. 

140 CONFLICT AND COOPERATION to a fourth quarter peak of 30 percent, though some of this increase results from shifts in product mix toward microprocessors, with DRAM-intensive Japanese consumer electronics production moving off-shore. The benefits of this increased market share for U.S. producers were significant.357 Some U.S. and foreign producers, and their customers, were disadvan- taged by price increases and especially by delay in product delivery.358 It is important to recognize, however, that these trends were under way before the trade agreement took effect, and reflected the characteristic “market counter-measures” taken by many Japanese industries in depressed mar- kets.359 The shortage in supply of DRAMs did prove highly profitable for Japanese semiconductor producers. From a long-term perspective, how- ever, U.S. and foreign producers, their equipment suppliers, their employ- ees, and downstream users of semiconductors were advantaged, while glo- bal competition in this enabling technology was strengthened.360 Some critics of the Semiconductor Agreement were less concerned with the agreement’s market effects than with a question of principle, on the 357 While the gains for the U.S. economy from the agreement are difficult to calculate with precision, they are significant. The increased market share translates into R&D and capital investment expenditures as well as high-wage employment. At the end of 1990 the SIA calculated that U.S. firms achieved an incremental $1.16 billion in annual revenues under the STA (and would presumably have achieved an additional $1.16 billion had the 20 percent target been reached). A revenue gain of this magnitude ($1.16 billion) permits increased investment of an estimated $137 million annually in R&D, and of $130 million in annual capital expenditures, which in turn stimulates the sales of U.S.-based semiconductor equipment and materials suppliers. Howell et al., Creating Advantage, p. 89. 358 Howell et. al., Creating Advantage, p. 130. Howell documents cases where European producers reported that Japanese suppliers withheld critical components and machines for competitive purposes during this period. 359 Laura Tyson, Who’s Bashing Whom? p. 117. Howell et al., Creating Advantage, pp. 116–132. Howell argues that “the evidentiary record shows that joint actions by Japanese firms were under way in 1985, well before the [Semiconductor] Arrangement existed; the precipitating event, to the extent one may have been needed to prompt joint market-regulating actions, appears to have been the elimination of most non-Japanese producers of DRAMs in mid-1985.” Ibid., p. 129. Recent analysis supports this view. Kenneth Flamm notes that “Japanese semiconductor analysts in Tokyo stressed both the trend toward more oligopolistic behavior in the chip industry and the [Japanese] government’s role in encouraging it.” See Kenneth Flamm, Mismanaged Trade, p. 206. Flamm also argues that the production cuts and restraint in investment associated with MITI’s implementation of the STA prior to 1988 subse- quently “played some role” in the increase in chip prices by reducing aggregate world supply. He adds, however, that “the brakes applied to Japanese supply apparently greatly exceeded the restraints required” to meet the requirements of the Commerce Department. Ibid., pp. 269– 272. 360 Laura Tyson supports this assessment. She argues that “contrary to popular belief, the agreement was not ineffective at realizing many of its aims, including those of stabilizing the share of U.S. producers in the global DRAM market, reversing the precipitous decline of the U.S. share in the global market for EPROMs and increasing the share of U.S. producers in the 

SUPPLEMENTS 141 grounds that governments should not directly influence purchasing deci- sions of private enterprises. The 20 percent market share target was singled out as an example of “managed trade.” This criticism, though widespread, presumed that the rest of Japan’s import trade was unrestrained. Industry experts argued that Japan’s high-technology imports were already managed, either directly by MITI or through informal government-industry arrange- ments with industrial groups and distributors, or through private anticom- petitive practices.361 In this view, what was unusual about the Semiconduc- tor Trade Agreement was not that it included a market share target, but that its management was both public and jointly administered (i.e., by the Japa- nese and American governments). Leaving aside ideological objections, the Semiconductor Agreement has the not inconsiderable virtue of having accomplished its goals. It offered an innovative, market-opening solution to a previously intractable competitive problem for the American industry. The negotiation of the STA and the commitment of the U.S. government—and ultimately the Japanese govern- ment—permitted the removal of a potentially debilitating source of friction from a key bilateral relationship. Moreover, by providing an effective means for overcoming the trade problems that threatened the viability of the U.S. industry and for restoring cost-based competition, the agreement established a durable basis for vigorous global competition and allowed the semicon- ductor industry to continue its phenomenal growth from an increasingly dispersed production base. SUPPLEMENT B. GOVERNMENT SUPPORT FOR TECHNOLOGY DEVELOPMENT: THE SEMATECH EXPERIMENT362 Overview The U.S. response to the challenge to its semiconductor industry in the period 1985–1995 involved three inter-related elements. As noted in the Japanese market. Nor did the agreement reduce competition across the board, as is widely believed. Instead the effects on competition also varied by industry segment over time.” Who’s Bashing Whom? p. 87 and p. 132. 361 Howell et al., Creating Advantage, pp. 101–103. MITI’s subsequent success in setting minimum pricing standards for Japanese DRAM manufacturers’ direct export sales lends sup- port to this view. See Kenneth Flamm, Mismanaged Trade? p. 243. 362 This section draws heavily from the contributions to Committee deliberations of George M. Scalise, The Nature of High Technology Competition, and William J. Spencer, SEMATECH, passim. See also William J. Spencer and Peter Grindley, “SEMATECH after Five Years: High-Technology Consortia and U.S. Competitiveness,” California Management Review, vol. 35, Summer 1993. 

142 CONFLICT AND COOPERATION section on the Semiconductor Trade Agreement, the first critical assistance took the form of trade relief. A second element was the creation of the consortium of U.S. chip producers and equipment manufacturers that came to be known as SEMATECH. However, the third indispensable element in the revival of the U.S. industry rested on its willingness and ability to reinvest and innovate. Arguments about which element was most decisive probably miss the point. The recovery of the American industry was achieved by a combina- tion of private effort and public-private interactions which enabled the U.S. firms to profit from shifts in demand, i.e., away from DRAMs (where Japa- nese skill in precision clean manufacturing gave significant advantage) to- ward microprocessor design and production (where U.S. strengths in soft- ware systems and logic design aided their recovery).363 The recovery of the U.S. industry is thus like a three-legged stool. It is unlikely any one factor would have proved sufficient independently.364 Trade policy, no matter how innovative, could not have met the requirement to improve U.S. prod- uct quality. On the other hand, by their long-term nature, even effective industry-government partnerships can be rendered useless in a market un- protected against predatory pricing by foreign rivals. Most importantly, neither trade nor technology policy can succeed in the absence of adaptable, adequately capitalized, effectively managed, technologically innovative com- panies. Because the rest of this section addresses “one leg of the stool,” i.e., the SEMATECH contribution, it is essential to underscore the impor- tance of the other two legs, which are an effective trade policy and a dy- namic private-sector able to attract and retain the capital necessary to com- pete in global markets. The Technology Challenge In addition to the challenge posed by Japanese trade practices (see Supplement A), the U.S. industry faced major technology problems. In the early 1980s, U.S. manufacturing quality had fallen behind the standards of the Japanese industry. The problem was manifest in lower manufacturing yields, higher costs, and inferior product quality. 363 Charles H. Fine and Daniel E. Whitney, “Is the Make-Buy Decision Process a Core Competence?” 364 Michael Porter captures the point, noting that “(g)overnment cannot create competitive industries, firms must do so. Government’s role in competition is inherently partial, because many other characteristics of a nation bear on it. Government can shape or influence the context and institutional structure surrounding firms, however, as well as the inputs they draw upon. Government policies that succeed are those that create an environment in which firms can gain competitive advantage...” Michael Porter, The Competitive Advantage of Nations, p. 620. 

SUPPLEMENTS 143 The Japanese industry had established great strength in manufacturing, whereas the U.S. industry excelled in product design. As the Japanese industry expanded during the early 1980s, particularly in the product cat- egories that involved high-volume manufacturing, their advantage in manu- facturing grew steadily. While the manufacturing technology problem for U.S. semiconductor producers was greatly aggravated by the consequences of the Japanese trade practices, it nonetheless constituted a distinctly differ- ent challenge, requiring its own solution if the U.S. industry was to exploit the opportunities provided by the Semiconductor Trade Agreement. Because the quality technology problem was centered on process tech- nology and the performance of the supporting infrastructure of tools and materials, the manufacturing challenge was common to all U.S. device mak- ers. It was also once removed from direct product or market competition in semiconductors, because it involved technology that was not part of device design or specific device production recipes. The scope of this manufacturing technology challenge was too great for any single company to overcome on its own. As a result, individual compa- nies gradually came to realize that only a consortium that garnered the participation of all those who would benefit from solving the manufacturing challenge was likely to be able to afford the undertaking and make it suc- ceed.365 In 1987 fourteen U.S. semiconductor manufacturing firms, ac- counting for over three-quarters of U.S. production, came together to form a consortium to develop semiconductor manufacturing tools, materials, and technology.366 U.S. Government Interest By the mid-1980s, U.S. officials had become keenly aware of the prob- lems besetting the semiconductor industry and of the implications of the industry’s decline both for economic growth and national security. There was growing concern among senior industrialists and defense officials over the erosion of the U.S. ability to produce key components of critical weap- 365 The Semiconductor Industry Association put forward a proposal in May 1987 for a research consortium supported by shared government-industry funding. 366 Eleven of the original fourteen are still members. Dissatisfaction with the focus of the research program led Micron Technology and LSI Logic to withdraw in 1992. In 1993, Harris Corporation withdrew. For a discussion of the SEMATECH R&D program, see the GAO study SEMATECH’s Technological Progress and Proposed R&D Program, GAO/RCEED/92- 223 BR, Washington, D.C., July 1992. Current SEMATECH membership includes Advanced Micro Devices, Digital Equipment Corporation, Hewlett-Packard Company, Intel Corporation, IBM, Lucent Technologies, Motorola, National Semiconductor, NCR, Rockwell International, and Texas Instruments. 

144 CONFLICT AND COOPERATION ons systems. The national security implications of the growth in U.S. dependency became apparent when Norm Augustine, chairman of the De- fense Science Board and of the Martin Marietta Corporation, informed Secretary of Defense Caspar Weinberger that twenty-one absolutely criti- cal U.S. military systems contained chips available only from foreign pro- ducers. 367 Moreover, defense industry leaders such as Augustine believed that U.S. military suppliers should be able to rely on cost-effective U.S. sources for the manufacture of advanced designs or high-volume, low-cost reliable commercial products. Importantly, the Japanese producers were displacing not only the U.S. semiconductor manufacturers, but also the equipment and materials manu- facturers who supplied them. Because advances in semiconductor technol- ogy are closely linked to equipment and manufacturing capabilities, this meant the United States was losing its capacity in the latest manufacturing technologies—critical determinants of price and quality. This erosion in U.S. technological competency also meant that the United States was well on its way to becoming structurally dependent on Japan for technology critical to the U.S. defense capability.368 In light of these developments, the Defense Department concluded the U.S. industrial base needed the manufacturing technologies that the consor- tium proposed to address. The proposal was closely aligned with DOD’s own plans for the development of semiconductor manufacturing technol- ogy, especially as it was completing a five-year program in the design of integrated circuits for military applications. This technology could allow the Defense Department to deploy advanced integrated circuits in virtually all of its major defense systems in a more effective and cost-efficient man- ner. By participating with the consortium rather than seeking to procure R&D through normal defense contracting procedures, the Defense Depart- ment believed it could achieve the required capability while substantially reducing its costs. To capture these savings, benefit from the anticipated improvements in semiconductor manufacturing technology, and bolster the defense industrial base, the Defense Department supported the consortium with approximately $100 million a year, on the condition that industry match the funding with 367 Clyde Prestowitz, Trading Places, p. 122. 368 Ibid. The National Advisory Committee on Semiconductors confirmed this assessment, stating that “if this vital industry is allowed to wither away, the nation will pay a price measured in millions of jobs across the entire electronics field, technological leadership in many allied industries such as telecommunications and computers, and the technical edge we depend on for national security.” See A Strategic Industry at Risk, November 1989, cited in Council on Competitiveness, Roadmap for Results, p. 39. 

SUPPLEMENTS 145 its own contributions.369 On this matched-funding basis with industry, De- fense Department financial support continued for eight years. In July 1994, the SEMATECH Board voted to terminate government support for its ef- fort, with fiscal year 1996 slated to be the final year of government funding. Despite the end to government funding, SEMATECH members have agreed to continue this innovative program with core funding provided exclusively by corporate contributions. The SEMATECH Experiment SEMATECH represents a significant new experiment for government- industry cooperation in technology development.370 Conceived and funded under the auspices of the Reagan administration, the consortium represented an unusual collaboration for both the U.S. government and the U.S. semi- conductor industry. In many respects, the decision to cooperate represented a profound change for the highly competitive, market-oriented companies that formed its membership. Indeed, the move to form SEMATECH oc- curred only when the top executives from the best U.S. electronics compa- nies came to the sobering realization that they could not prevail alone. As one commentator noted, “the mere formation of SEMATECH required a radically new mind-set at some of America’s leading high-tech corpora- tions.”371 Not only were these independent executives now willing to col- laborate with the government, but these fiercely competitive rivals had to agree to collaborate on research and development that was competitively important to each of them. A Short-Term Focus on Improved Quality The technology focus of the consortium is near-term and of almost im- mediate commercial significance. It seeks to affect existing manufacturing processes, tools, and materials. However, no member company or member company product receives any financial assistance. For member compa- nies, the consortium provides generic manufacturing technology and im- proved tools that allow them to compete more effectively. For equipment and materials suppliers, it provides competitive awards, customer input, and industry-wide demonstration and validation of products. Overall, these measures have had a positive, competitive effect on the U.S. supply base. 369 Membership dues have been fixed at one percent of semiconductor sales, with a mini- mum contribution of $1 million and a maximum of $15 million. 370 For an excellent review of this experiment, see Peter Grindley et al., “SEMATECH and Collaborative Research: Lessons in the Design of High-Technology Consortia.” 371 Hedrick Smith, Rethinking America. Random House, New York, 1995, p. 385. 

146 CONFLICT AND COOPERATION Industry Driven From the outset, leadership for the consortium was provided by industry. The CEO of National Semiconductor, Charles Sporck, provided crucial im- petus to the concept of the consortium, playing an instrumental role in garnering support among the original fourteen members of SEMATECH. The consortium also benefited from the prestige and leadership of Bob Noyce, who strengthened the credibility of the endeavor within the industry and in Washington. Reflecting these origins, the operations of SEMATECH have remained industry driven. While it continues to benefit from strong leadership, no single entity dominates the consortium or determines its direction. Mem- bers, including Department of Defense officials, reach consensus on techni- cal direction; the consortium management is then left to implement that direction. The industry interaction within the consortium changed the dy- namics between device makers and suppliers, with the collaboration gener- ating new technical perspectives for the participants and encouraging the give-and-take between manufacturer and supplier so as to enrich, improve, and expedite the technology development process. Because SEMATECH programs are focused on SEmiconductor MAnufacturing TECHnology (hence SEMATECH’s name and its mission), the largest por- tion of its budget is spent on programs which relate to equipment improve- ment. Quite often these programs are carried out at equipment supplier sites or at member company sites. Teams of SEMATECH engineers are involved in these programs, as well as the on-site engineers. Technology diffusion is thus achieved through first-hand transfer. Member companies provide technical personnel to work at the consortium and regularly rotate them back to the member company.372 By having a member’s output from SEMATECH depend largely on the quality of the people assigned to the consortium, technology transfer is enhanced and there is a strong incentive to provide high-quality people to the consortium.373 Closed Membership—Open Output Membership in SEMATECH and in the group of participating supplier firms known as SEMI-SEMATECH is limited to U.S.-based companies. This 372 “In 1992, some 225 of the 722 personnel employed at SEMATECH were assignees from member firms. Most of these individuals were members of the consortium’s 300 member professional research staff.” See Grindley et al., “SEMATECH and Collaborative Research,” p. 729. 373 SEMATECH also produces reports and holds meetings to provide technology informa- tion to member companies. For example, in 1994 SEMATECH held over 600 meetings and hosted over 25,000 visitors. See Spencer, Technology (Transfer) at SEMATECH, 14 December 1995 contribution to Steering Committee deliberations, p. 1. 

SUPPLEMENTS 147 preference was based on the corporate restrictions imposed by the U.S. indus- try participants, although the policy had the tacit support of U.S. Defense Department officials. This limitation was also probably necessary in order to maintain political support for continued government funding.374 Originally, the results of SEMATECH research were licensed to member firms only on an exclusive basis, although that requirement was changed in 1991. At present, the output of the consortium is not restricted to U.S. entities, although mem- bers of the consortium do receive some priority in ordering and receiving equipment derived from innovations funded by SEMATECH.375 Currently, members and supplier firms alike are encouraged to deploy SEMATECH tech- nology throughout the world. Moreover, new projects, like the I300I initiative to develop an industry standard for the move to the 300mm wafer, involve extensive international participation.376 The Semiconductor Roadmap The scope and success of this initiative ultimately led to a systematic review of manufacturing needs across the entire technology spectrum. The Semiconductor Roadmap exercise brought together, in a systematic fashion, expert representatives of industry, universities, and government to examine the technological challenges facing the industry. The Semiconductor Roadmap represents an unprecedented collective effort to identify opportunities and obstacles to the continued exponential growth in performance of the industry’s products. SEMATECH’s efforts became the basis for the development of 374 Andrew Procassini, the former president of the Semiconductor Industry Association, confirms this point in Competitors in Alliance, p. 74. See also Grindley et al., “SEMATECH and Collaborative Research,” p. 747. The authors note that SEMATECH’s May 1993 an- nouncement that it would continue to work with a U.S. equipment manufacturer despite an impending alliance with a major foreign competitor was criticized by the House Armed Ser- vices Committee. The Committee questioned “the rationale of Congressionally initiated in- vestments if these investments do not result in jobs and industry in the U.S.” See also J. Robertson and J. Doresch, “GCA Revives as U.S. Drafts Corning, AT&T.” Electronics News, 16 August 1993 cited in Grindley et al., “SEMATECH and Collaborative Research.” 375 Grindley et al., “SEMATECH and Collaborative Research,” p. 731. The authors note that the change in disclosure policies both benefited U.S. equipment manufacturers and ad- dressed criticism from some semiconductor manufacturers about the use of general tax rev- enues to benefit member companies at the expense of nonmembers. Ibid. This policy has also enhanced the ability of nonmembers to obtain a “free-ride,” that is benefit from SEMATECH advances without contributing to the program. P. 745. 376 As noted above, Japanese producers did not join the international effort organized by SEMATECH to develop the 300mm wafer, preferring a separate national effort, though some information will be exchanged between the two projects. Kenneth Flamm, Mismanaged Trade? p. 441. The author also suggests that SEMATECH’s sponsorship of the I300I program is a move toward greater international cooperation made possible by the absence of the constraints associated with government funding. 

148 CONFLICT AND COOPERATION aggregate technology roadmaps that could identify technical opportunities and gaps that had not previously been brought to light. Was SEMATECH A Success? An assessment of an R&D program such as SEMATECH is a difficult analytical undertaking, not least because of the “counter-factual challenge” of knowing what would have happened to the industry without SEMATECH— keeping in mind that the consortium’s technological accomplishments were widely disseminated to both participants and nonparticipants. Moreover, the success of R&D efforts are difficult to measure with precision.377 The criteria are elusive, the output can rarely be quantified, and know-how or technology is only one—and not necessarily the most decisive—element in the product development and commercialization process.378 Despite these inherent uncertainties, the assessments summarized below suggest a record of substantial accomplishment.379 Industry Goals Met From the industry perspective, SEMATECH’s goals have been largely achieved.380 The manufacturing quality and efficiency of the U.S. industry have at least reestablished parity with Japan. With a few notable excep- tions, at least one world-class U.S. supplier exists in all the primary seg- ments of the semiconductor manufacturing process.381 The worldwide mar- 377 Assessing technological success, and a consortium’s contribution to technological ad- vance, can be complex. Success has many fathers. Moreover, the contribution of technology to commercial success is difficult to assess, especially for government-funded consortia. As Grindley, Mowery, and Silverman note “the politically relevant timetable for such evaluation is often far shorter than the period of time needed for the realization of the economic effects of consortia research.” Grindley et al., “SEMATECH and Collaborative Research,” p. 724. 378 Ibid. In reviewing the difficulties SEMATECH encountered in maintaining a viable U.S. producer of lithographic equipment, the authors underscore the argument that “technology alone is rarely sufficient to restore the competitiveness of firms or industries that lack critical, complementary personnel, marketing, managerial, and financial resources.” P. 743 and p. 751. 379 Ibid., p. 736 and passim. The authors discuss the challenges of evaluating SEMATECH and provide a critical, but generally positive, assessment of the consortium’s accomplishments. They also note that the consortium’s goals changed over time, reflecting the changing percep- tions of its members’ needs. This operational flexibility is a strength and probably essential in an industry evolving as rapidly as the semiconductor industry. 380 For example, SEMATECH has met its objectives in the development of process tech- nology, the supply of manufacturing equipment, and collaboration between manufacturers, suppliers, and research centers. Ibid., p. 743. 381 Relative to both Europe and Japan, the United States lags in terms of company rankings in sectors such as packaging equipment, mircrolithography equipment, and semiconductor materials. Presentation by VLSI Research Inc., President’s Conference, 11 April 1995, Wash- ington, D.C. 

SUPPLEMENTS 149 ket share decline of both U.S. semiconductor firms and suppliers has been reversed and worldwide market share leadership regained. While other factors were also in play, many in the industry believe SEMATECH made a major contribution to these turnarounds. The consortium certainly achieved its initial objective to match and then overtake its competitors in semiconductor manufacturing technology. U.S. manufacturers were able to construct integrated circuits with features as small as 0.8 micron on five-inch silicon wafers by 1989, 0.5 micron features on a six-inch wafer in 1990, and 0.35 micron features by the end of 1992. Keeping in mind that a 0.35 micron particle is approximately 1/200th the width of a human hair, this is a remarkable technical achievement. As noted in the section on international cooperation above, SEMATECH is now playing an instrumental role in the industry move to the 300mm wafer standard. The consortium itself has taken the view that an important part of the programs is the measurement of results. Earlier in its history, SEMATECH established a return on investment measurement for each program. The average return tended to be between 3.5 and 4 times for the member compa- nies, with a range of return from about 8 to about 2 times investment.382 In 1994 a new measurement system was established measuring user satisfac- tion and user support for each program. Plotted in four quadrants with the quadrants divided at the 50 percent line for customer support and customer satisfaction, the goals for SEMATECH programs are to have 70 percent in the third quadrant, where customer support and customer satisfaction are both over 50 percent. Those programs that fall in the first quadrant are candidates for termination. The creation of SEMATECH encouraged technology-based competition among equipment and materials suppliers. Rather than alter market forces, SEMATECH built on them. Consequently, government-industry collabora- tion on semiconductor technologies did not fundamentally undermine mar- ket forces. Inefficient companies were not propped up by government funds, and assured commercial success was not bestowed on any consortium mem- ber. Instead, this innovative collaboration offered a source of technology that would then have to be effectively commercialized before a firm could extract any economic benefit for itself.383 SEMATECH’s role in supporting equipment development and qualification does mean that it has substantial influence on technological choices in an industry faced by great uncer- 382 Spencer, Technology (Transfer) at SEMATECH, p. 1. 383 The impact of government support for technology is generally believed to be minimally distortive of the market because it tends to occur at the beginning of the product development cycle rather than at the end. This is the rationale for the current exemptions in the GATT for government R&D support, as noted in the body of the report. 

150 CONFLICT AND COOPERATION tainty. However, SEMATECH’s private sector management has sought to maintain a diversified portfolio of technological alternatives and has hedged against technological risk by pursuing alternative technologies with more than one supplier and more than one engineering approach.384 Perhaps the best measure of SEMATECH’s success for the industry is the continued willingness of members to invest in the consortium even without government participation. Moreover, the improved technology and more capable suppliers that became available through SEMATECH’s ef- forts supported not only U.S. semiconductor manufacturers but the emerg- ing industries in Korea and Taiwan as well. European and Japanese indus- tries were also able to draw on the results of consortium research. The technical accomplishments of SEMATECH resulted in higher productivity, better product, and a strengthened U.S. supply base while contributing to the diversity of equipment supply which is essential for the maintenance of vigorous global competition. DOD Goals Met The Defense Department’s return from SEMATECH can be similarly assessed. The Defense Department benefited from cost and quality im- provements in the microelectronics that increasingly account for the perfor- mance advantage of U.S. defense systems. In addition, the vitality of the supporting U.S. infrastructure was preserved, and concentrated dependency on foreign sources of semiconductor tools and materials was avoided. Like other defense R&D programs, SEMATECH delivered technology that en- abled the Defense Department to do its mission better. University Spillovers Even though government participation in the consortium was justified and funded as a defense program, broader national interests also benefited. SEMATECH provided a focus and exchange of technical ideas that, as it did for industry, strengthened ongoing R&D in both universities and gov- ernment agencies and laboratories. For example, SEMATECH helped cre- ate, in cooperation with the Semiconductor Research Corporation, the uni- versity centers of excellence for semiconductor research. In addition, the Semiconductor Roadmap exercise mentioned above contributed a great deal to the identification of key research areas. 384 Grindley et al., “SEMATECH and Collaborative Research,” pp. 745–745. 

SUPPLEMENTS 151 National Economic Benefits Perhaps most important, SEMATECH helped to anchor the future devel- opment of a strategic technology in the United States. This makes it likely that the United States will be the hub for subsequent deployment of the technology throughout the world. The United States is thereby assured of capturing a significant proportion of the economic growth, high-wage em- ployment, and technological competency to be gained from the phenomenal growth that characterizes this industry. More broadly, the United States will continue to maintain and benefit from the economic leadership, inde- pendence, and prosperity that are uniquely associated with the mastery of high-technology manufacturing. BOX J. LESSONS IN TECHNOLOGY TRANSFER SEMATECH management has identified several principles of operations which underpin successful technology transfer activities. They include: 1. User Involvement: It is important to have the users involved in the determi- nation of the programs, and the continual evaluation of the programs. These users have the responsibility for returning that technology to their member companies. 2. Metrics: Specific metrics for each program are essential for successful technol- ogy development. For example, programs are evaluated quarterly at SEMATECH. 3. Direct Involvement of Senior Management: This is essential to stop what does not work. Engineers quite often believe that if a program can continue for only a short time longer, the problems can be solved and the program will turn out to be of great value. Usually senior management in the member companies of SEMATECH is more inclined to stop programs. Consequently, the involvement of senior management is essential to close out programs with inadequate returns. 4. High Return: Focus on those programs where there is a high return. The users of technology are the best judges of rates of return. Support and satisfaction are one measure; SEMATECH has also experimented with return on investment. There are certainly other metrics that can be used. What is essential is that there be an accepted metric to determine which programs should be continued and which should be stopped. 5. Effective Selection: Over a period of nearly eight years, SEMATECH has developed a series of processes for choosing technology development programs which are relevant to the core of interests of the member companies. These processes are continually evaluated and updated. The consortium’s goal is to find the best processes from anywhere in the world and apply them to its programs. 

152 CONFLICT AND COOPERATION SUPPLEMENT C. IMPLICATIONS OF THE U.S. DUAL-USE STRATEGY The rapid advance of commercial technologies has led U.S. officials to push for a new defense acquisition strategy which emphasizes the use of new technologies, developed by commercial industry, to meet defense needs. While announced with some fanfare in the United States, in the view of some observers dual-use policies have in fact been practiced by others in the Cold War period, perhaps most effectively by Japan, and is now the policy of many European countries.385 It is U.S. policy which, in the post–Cold War world, is evolving toward the practice of the other industrial powers. The dual-use strategy adopted by the United States early in the Clinton administration has two key elements. First, the Department of Defense is to make use, wherever possible, of components, technologies, and subsystems developed by commercial industry. The Department is to develop defense- unique products only where necessary.386 Second, for this acquisition strat- egy to be successful, defense R&D efforts must nurture technologies and capabilities that will continue to be advanced through industries’ efforts to remain competitive in commercial markets. The goal of the strategy is to leverage commercial technological advance in order to create military ad- vantage, while ensuring the resulting equipment remains both affordable and the most advanced in the world.387 In the words of one senior defense official, the objective of this policy is “to marry the momentum of a rigorous, productive, and competitive com- mercial industrial infrastructure with the unique technologies and system integration capabilities provided by our defense-industrial base.”388 This policy represents a sharp departure from the deeply ingrained practice of defense funded R&D, with procurement and production consigned to spe- cialized defense suppliers. In the past, when defense requirements consti- tuted a significant portion of high-technology markets, this policy was often effective. Indeed, government-military procurement was instrumental in accelerating innovation in aerospace and even in generating new industries 385 Presentation by Jacques Gansler to the conference Sources of International Friction and Cooperation in High-Technology Development and Trade, 30–31 May 1995. For a discus- sion of the need for the U.S. military-industrial complex to adjust to the end of the Cold War and maintain technological superiority with a reduced defense budget that nonetheless en- hances rather than reduces U.S. economic competitiveness, see Jacques Gansler, Defense Con- version Transforming the Arsenal of Democracy, MIT Press, Cambridge, Mass., 1995. 386 Flat Panel Display Task Force, Building U.S. Capabilities in Flat Panel Displays, Final Report, U.S. Department of Defense, Washington, D.C., October 1994, p. 1.1. 387 Presentation by Under Secretary of Defense Paul Kaminski, in Sources of International Friction and Cooperation in High-Technology Development and Trade, 30–31 May 1995 and the presentation by Stephen A. Cooney to the conference Towards A New Global Framework for High-Technology Competition, 30–31 August 1995, Kiel, Germany. 388 Kaminski, op. cit. 

SUPPLEMENTS 153 such as microelectronics. However, the military emphasis on “the extremes of performance,” with mission capability given first priority and cost being a secondary consideration, led military manufacturers to emphasize high- cost custom production rather than the low-cost, high-volume production that characterizes commercial markets.389 Some suggest this led military contractors “to lose the capability to de- sign, manufacture, and sell in a competitive marketplace.”390 This less competitive, performance-only orientation had been identified by the 1986 Packard Commission, which stated that commercial industry could provide cost savings and quality improvements for military programs through the application of civilian technology.391 Because the commercial sector relies on advanced technology both to lower costs and to enhance performance, through their interaction, cost and performance have improved faster in the commercial sector than for military applications. Moreover, the high devel- opment costs for new technologies are best supported by high-volume sales of consumer products. In addition, the acceleration of product life-cycles in competitive markets means that commercial producers have become signifi- cantly more responsive and flexible than military producers.392 As a result of these trends, the Defense Department no longer leads in critical technolo- gies such as information systems, telecommunications systems, microelec- tronics, and a variety of fields associated with advanced design tools.393 Senior defense officials now assert that declining defense budgets, demand levels, and resources make reliance on defense-only suppliers both unaffordable and ineffective.394 To accommodate these trends, current U.S. policy is to rely on commer- cial producers for leading-edge capabilities, at reduced cost, in a shorter timeframe. Insertion time is increasingly important, because in the global markets which characterize many new technologies, access to the same commercial technology base by both allies and adversaries will become the norm. In this environment, military advantage will accrue to the nation with the best “cycle time,” that is, the ability to acquire leading-edge tech- nologies at low cost, insert them in weapons systems, and field soldiers trained to exploit them effectively.395 If successful, this approach will 389 Richard Samuels, Rich Nation, Strong Army, pp. 22–24. 390 Ibid. Samuels also cites Seymour Melman (ed.), The War Economy in the United States: Readings on Military Industry and Economy, Saint Martin’s Press, New York, 1971, p. 72. 391 Richard Samuels, Rich Nation, Strong Army, p. 27. 392 Ibid., p. 28. 393 Paul Kaminski in Sources of International Friction and Cooperation in High-Technol- ogy Development and Trade. 394 Flat Panel Display Task Force, Building U.S. Capabilities in Flat Panel Displays. 395 Paul Kaminski in Sources of International Friction and Cooperation in High-Technol- ogy Development and Trade. 

154 CONFLICT AND COOPERATION mean a sharp departure from the current fifteen-year acquisition cycle for major U.S. defense systems, and enable U.S. forces to benefit from the commercial turnover cycle of three to four years. For many systems, it will mean that the acquisition of defense-only technologies becomes the excep- tion rather than the rule. The adoption of a dual-use strategy by the U.S. also undercuts, at least in part, the widespread perception that U.S. defense programs and procure- ment served as “hidden subsidies” to the U.S. commercial sector. In part this view is widespread because U.S. defense and space agency officials sought to justify their budget requests with the argument that their programs had powerful and positive commercial benefits.396 Further legitimacy was accorded this argument by countries directly subsidizing commercial tech- nologies who saw U.S. defense budgets as a justification for their own commercially oriented efforts.397 This claim was reinforced by companies responsible for defense work, which justify high defense budgets by assert- ing the benefits of defense acquisition. In some cases, there clearly have been benefits to U.S. industries such as aerospace and semiconductors— though the benefits of such support, especially in aerospace, seem to be somewhat unevenly distributed.398 The validity of a direct comparison of the two types of support, the one commercially oriented from the outset (i.e., direct subsidies such as those for Airbus) and the other derivative (i.e., learning through military programs), is often overstated, particularly with respect to today’s technologies and markets. 396 Richard Samuels, Rich Nation, Strong Army, p. 21. See also Steven Rosen, Winning the Next War: Innovation and the Modern Military, Cornell University Press, Ithaca, N.Y., 1991. 397 If the military procurement approach were best, presumably countries in Europe would have increased military budgets rather than direct support for commercial technology. Kende points out, for example, that “while many of the actions taken by the European governments in support of their computer industry were similar to those taken by the U.S. government, the goals were different, as were the results.” Kende, “Government Support of the European Information Technology Industry,” p. 4. In his analysis, Kende argues that the underlying concept of the national champion approach was flawed. For example, French, German, and British efforts in the 1960s to develop national champions to compete with IBM focused on company size per se, rather than focusing on the new technologies and quality products that characterized the American competition. He contends that the merger of several disparate national firms with incompatible product lines did not result in a dynamic global competitor— not least because governments removed the discipline of competition from the favored firms, which then benefited from protected national procurement markets. 398 Some analysts argue that the unwillingness of the American government to intervene in the aerospace industry for commercial objectives has actually weakened U.S. producers, nota- bly McDonnell Douglas. Laura Tyson, Who’s Bashing Whom? p. 192. Tyson argues further that the scale, scope, and learning economies mean new entrants face much higher production costs than do incumbents. Without “the visible hand” of government support, the particular conditions which characterize the aircraft industry pose insurmountable barriers to new com- petitors. Ibid, p. 157. 

SUPPLEMENTS 155 Even in the heyday of defense led R&D, the assistance to U.S. industry derived from the U.S. defense budget was often constrained by the defense mission. Some analysts argue that the U.S. defense acquisition system, far from being a guise for the support of commercially relevant industry, has in fact created disincentives and barriers to the operation of market forces. These include “the unique government oversight requirements, the unique procurement requirements, (and) the unique military specifications” associ- ated with military procurement.399 Indeed, these restrictions are seen as so onerous that some world-class U.S. companies, e.g., Hewlett-Packard, refuse to accept defense R&D contracts.400 Some analysts suggest that U.S. de- fense acquisition practices lowered overall return on U.S. R&D spending.401 This is because the normal output of defense R&D “spun away” rather than “on” to commercial applications. Normal technology diffusion could not take place as a result of defense secrecy requirements.402 Some analysts take the case a step further, not only outlining the shortcomings of the “spin-off paradigm,” but arguing that massive defense spending did great harm to the [U.S.] economy.403 This is because the defense industry has 399 Presentation by Jacques Gansler in Sources of International Friction and Cooperation in High-Technology Development and Trade. 400 Ibid. 401 See Robert A. Solo, “Gearing Military R&D to Economic Growth,” Harvard Business Review, November–December 1962, pp. 49–60, cited in Richard Samuels, Rich Nation, Strong Army. Samuels adds that “it is far from clear that the military acts primarily as an agent of technological innovation, that a linear relationship exists between defense and civilian prod- ucts, or that military spending has had a positive impact on the civilian economy.” Ibid., p. 22. 402 “Governments have, of course, struggled to prevent diffusion of military technologies... This enforced isolation has taken a great many forms, including separate accounting and audit systems, different bidding and marketing procedures, export controls (extending to dual-use technologies), separate divisional structures, classification schemes, and nationality require- ments, all of which aim to prevent (or at least to impede) technology diffusion.” Ibid., p. 22. 403 Exponents of this view cited by Samuels are Jay Stowsky, Beating our Plowshares into Double-Edged Swords: The Impact of Pentagon Policies on the Commercialization of Ad- vanced Technologies, Berkeley Roundtable on International Economics, Berkeley, Calif., 1986; Robert Solo, “Gearing Military R&D to Economic Growth,” Harvard Business Review, No- vember–December 1962; Amitai Etzoini, The Moon-Doggle, Doubleday, New York, 1964; Seymour Melman et al., The War Economy of the United States: Readings on Military Industry and Economy, St. Martin’s Press, New York, 1971; Mary Kaldor, The Baroque Arsenal, Hill and Wang, New York, 1981; James Fallows, National Defense, Random House, New York, 1981; Lloyd Dumas, The Overburdened Economy, University of California Press, Berkeley, Calif., 1986; Nathan Rosenberg, “Civilian ‘Spillovers’ from Military R&D Spending: The U.S. Experience Since World War II,” in S. Lakoff and R. Willoughby (eds.), Strategic Defense and the Western Alliance, Lexington Books, Lexington, Mass., 1987; Jacques Gansler, “Integrating Civilian and Military Industry,” Issues in Science and Technology, Fall 1988; John Alic et al. (eds.), Beyond Spinoff, Harvard Business School Press, Boston, Mass., 1992; and Jay Stowsky and Burgess Laird, “Conversion to Competitiveness: Making the Most of the National Labs,” American Prospect, Fall 1992. 

156 CONFLICT AND COOPERATION become progressively isolated from commercial industry, and because mili- tary and civilian laboratories compete (at least to some extent) for the same research personnel and research funding.404 Indeed, the evolution of U.S. policy toward a dual-use strategy essen- tially supports these claims. Instead of the commercial sector benefiting from technological spin-off, the military now requires commercial “spin- on.” This strongly suggests that the previous U.S. approach, with the cur- rent barriers to the integration of the civil and military industrial bases, has proven less efficient than more commercially oriented strategies for a wide variety of defense components and equipment. Consequences for Competition— The more widespread adoption of dual-use policies is also likely to have important consequences for high-technology competition and cooperation. From the perspective of global competition, efforts by the United States and other nations to ensure an adequate commercial-industrial base to meet de- fense needs raise concerns that the defense rationale may be applied to a broad range of leading-edge industries, with defense requirements becom- ing an important source of financial support.405 For example, government efforts to develop military weapons systems, such as the current Japanese production of the FSX, components for the Boeing 777, or the recently announced plan to produce an antisubmarine patrol plane, can have direct impact on commercially oriented aerospace products, much as U.S. Cold War military programs sometimes encouraged U.S. commercial aircraft de- velopment.406 Because reliance on defense-unique industries is no longer cost effective, or even possible in some cases, the Defense Department appears prepared to foster, or at least encourage, a viable domestic industry that is globally competitive in terms of both cost and technological capabil- ity, as a means of meeting defense needs. The most striking example of this approach is the U.S. dual-use program on flat panel displays, an unusual effort to encourage a U.S.-based commercial industry—where none exists— to meet military needs from lower-cost, high-volume, commercially ori- ented production facilities.407 404 Richard Samuels, Rich Nation, Strong Army, pp. 22–23. 405 Sylvia Ostry, “Technology Issues,” paragraphs 21–24. Ostry cites in particular the evolution of U.S. policy toward dual-use technology. Richard Samuels, Rich Nation, Strong Army, argues that Japan has consciously pursued a dual-use policy for many years. 406 As noted, aerospace spin-off can be overstated. See the section above, and the presen- tation by Sally Bath, director of aerospace, U.S. Department of Commerce in Sources of International Friction and Cooperation in High-Technology Development and Trade. 407 For a description of this innovative program, see Flat Panel Display Task Force, Build- ing U.S. Capabilities in Flat Panel Displays. 

SUPPLEMENTS 157 —and Cooperation At the same time, the dual-use strategy includes an important coopera- tive element. In order to obtain affordable weapons systems, the U.S. de- fense strategy places increased emphasis on cooperation with U.S. allies in the acquisition of defense equipment. The renewed emphasis on coopera- tion rests on a combination of political, military, and economic rationales. Politically, the cooperative programs “help strengthen the connective tis- sues, the military-industrial relationships, that bind...nations in a strong se- curity relationship.” 408 The military rationale rests on the likelihood of combined operations in a coalition environment where performance can be considerably enhanced with interoperable equipment and rationalized logis- tics. And the economic rationale is simply that the budgets of the U.S. and allied militaries are shrinking. Consequently, what may not be affordable individually may be possible collectively.409 As with commercial alliances, international cooperation for military sys- tems is not without its problems. These include the very long decision times often associated with cooperative projects, in which complex national decision processes have to be combined to produce a collective decision. This often results in “very poor cycle times and very long schedules.”410 Cooperative programs can also result in the absence of effective competi- tion once an international team is assembled, which can lead to high cost and inefficient performance for weapons systems. Commercial considerations can also pose significant challenges to coop- erative programs in militarily relevant technologies. For example, the U.S. practice of arms-length relations with private producers, often through the intermediary of a national laboratory, is not mirrored by the practices of some U.S. allies. These governments often work closely with leading na- tional companies, in which the government may also have a direct financial stake, and which have commercial interests in competition with those of U.S. firms.411 Similarly, U.S.-Japan cooperation in defense and dual-use technology has been asymmetrical. A recent study suggests that, in part, the asymmetry is due to the unwillingness of Japanese industry and govern- ment to cooperate technologically on reciprocal terms. The study also sug- gests that efforts by the U.S. government have lacked consistency and coor- dination in the pursuit of more balanced technology flows in defense technology 408 Paul Kaminski in Sources of International Friction and Cooperation in High-Technol- ogy Development and Trade. 409 Ibid. 410 Ibid. 411 Presentation by Rear Admiral Marc Y.E. Pelaez, Office of Naval Research, in Sources of International Friction and Cooperation in High-Technology Development and Trade. 

158 CONFLICT AND COOPERATION collaboration, while U.S. industry has tended to pursue increased sales rather than the acquisition of new technology.412 This lack of sustained attention to technology may reflect the U.S. preoccupation with the defense of terri- tory during the Cold War, and the tendency—from a position of great eco- nomic and technological strength—to subordinate economic interests to diplomatic and military aspects of national security. Japanese policy tends not to make this distinction between national defense objectives and economic interests. Japanese policymakers reject such arguments, seeing technology and pro- duction, as well as territory, as national interests that can and must be defended.413 These differences in perspective can generate friction with respect both to trade in high-technology products and to cooperative programs to de- velop new technologies. In some cases, these differences can be outweighed by the substantial benefits offered by equitable international cooperation. However, they underscore the need for care in the establishment and man- agement of cooperative programs. More broadly, as dual-use programs continue or expand in the U.S., Japan, and Europe, increased international friction is likely to occur over program goals and the resources devoted to their achievement. SUPPLEMENT D. GLOBAL POSITIONING SYSTEM: GOVERNMENT MISSIONS, COMMERCIAL APPLICATIONS, AND POLICY EVOLUTION Recent developments in U.S. policy on the Global Positioning System illustrate the powerful synergies made possible by government-funded re- search and infrastructure investments and by the private sector’s develop- ment of innovative commercial applications. For national security pur- poses, the U.S. government invested approximately $10 billion over two decades in a constellation of twenty-four satellites orbiting 11,000 miles above the earth. Each of the satellites has four atomic clocks to provide accurate time. In addition to the satellites, which produce the GPS signals, the “basic” GPS system consists of ground stations, data links, and associ- ated command control facilities. This satellite system is operated and main- tained by the U.S. Department of Defense. A System Developed for Military Advantage— The initial development of the Global Positioning System was under- taken under the auspices of the U.S. military in the 1970s to provide ex- 412 Maximizing U.S. Interests in Science and Technology Relations with Japan. National Academy Press, Washington, D.C., 1995, pp. 4–5. 413 Richard Samuels, Rich Nation, Strong Army, p. 4. 

SUPPLEMENTS 159 tremely accurate guidance for U.S. aircraft, missiles, and ships anywhere on earth. In combination with other U.S. information-gathering satellites, GPS provides precise targeting data for U.S. weapons systems and accurate posi- tioning information for allied troops. The precision afforded by GPS played a valuable role in U.S. air and ground operations in the course of the Gulf War. Partly on the basis of this experience, the substantial military advan- tage provided by GPS is now widely recognized, and this innovative system is now being integrated into a wide range of military applications. —With Growing Civilian Applications Over the past ten years, however, GPS has evolved far beyond its mili- tary origins, becoming “an information resource which supports a wide range of civil, scientific, and commercial functions—from air traffic control to the Internet—through its ability to provide precision location and timing information.”414 Moreover, the rapid advances in microelectronics have contributed to the development of equipment with lower costs, augmented capability, and increased portability.415 For example, the cost of a receiver has dropped precipitously, from $150,000 per GPS survey receiver in the mid-1980s to approximately $5,000 for an equivalent survey receiver in 1996. GPS handheld receivers for consumers originally sold for approxi- mately $3,000 in 1989, and are now selling at under $200. The GPS en- gine, or board receiver for the car navigation system, is currently selling, in moderate volume, in the consumer car navigation market at under $100 per unit. By picking up transmissions from the GPS signals and comparing the distances to different GPS satellites, commercial GPS receivers can identify their position on the globe with unprecedented accuracy from 300 meters down to—in some cases—a few centimeters. While applications are growing rapidly in both the military and commer- cial sectors, the market for civilian applications now exceeds that for mili- tary applications “by roughly three-to-one.” This is expected to widen to eight-to-one over the next few years, with projected sales of commercial GPS equipment expected to soar from the current $2 billion to $8.5 billion in the year 2000.416 The rapid evolution of GPS represents a significant, if 414 Critical Technologies Institute, A Policy Direction for the Global Positioning System: Balancing National Security and Commercial Interests, Research Brief, Washington, D.C., December 1995. 415 A recent Academy report points out that the development of concurrent technologies including satellites and microelectronics was essential to the creation of GPS, as was the assignment of responsibility to the Air Force for the development of a navigation system for all military services. See Allocating Federal Funds for Science and Technology, National Research Council, p. 49. 416 Scott Pace et al., The Global Positioning System: Assessing National Policies, Critical Technologies Institute, RAND, Santa Monica, Calif., 1995, p. 201. 

160 CONFLICT AND COOPERATION specialized, spin-off from national investment in high-technology military systems.417 It also underscores the unpredictable consequences of techno- logical advance and the powerful synergies which can occur as a result of public investments in new technologies, especially in high-technology in- frastructure, and the private sector’s development of commercial applica- tions utilizing the innovative capabilities of these systems.418 Interestingly, the evolution of the GPS system has benefited from the rapid advance in cost and performance driven by commercial imperatives.419 While the initial GPS market was derived from government policies de- signed to achieve defense objectives, the subsequent development of com- mercial markets has contributed to the ability of the government to accom- plish the national security mission of GPS. A recent report notes that “the demand by civilian commercial users of GPS for smaller, better, and cheaper receivers has directly benefited systems designed specifically for military use.” The report cites the case of the precision lightweight GPS receivers used by U.S. military forces, which were built at a low cost and delivered on time, in large part because of the technical benefits derived from re- search and development carried out by private firms for commercial appli- cations.420 Issues for U.S. Policy The rapid growth in commercial applications for the Global Positioning System has required an innovative policy approach to balance U.S. national security objectives with the interests of an increasingly international user community. A recent study noted that the success of the GPS system has raised “complex policy questions for U.S. decision-makers on a variety of issues affecting national defense, commerce, and foreign policy.”421 The report identified four policy areas: • Dual-Use issues: the need to balance the different requirements of national security and commercial interests; • Funding regime: whether the U.S. government should continue to carry the cost of maintaining the Global Positioning System or seek to collect user fees; • GPS governance: how should the Global Positioning System be gov- erned in the future? 417 See Supplement C for a broader discussion of dual-use technologies and commercial spin-offs. 418 Allocating Federal Funds for Science and Technology, passim, especially pp. 31–32. 419 This process is described in the supplement on dual-use (Supplement C). 420 Scott Pace et. al., The Global Positioning System, p. 251. 421 Critical Technologies Institute, A Policy Direction for the Global Positioning System, p. 1. 

SUPPLEMENTS 161 • Concerns of foreign users: how should the U.S. government guaran- tee domestic and foreign users—both public and private—that they will have assured, reliable access to GPS signals in a stable policy environment? These policy areas represent a broad consensus identified in a 1993 re- port requested by Congress.422 Asked to assess the future of GPS, the report found that “most aspects of GPS technology, governance and man- agement, and funding are remarkably sound. Among the aspects of the system that are working well and should, therefore, be retained are opera- tional control and funding of the basic GPS satellites by DOD; the aggres- sive application of GPS technology to public safety and public service needs by civil government agencies; and a dual-use policy that allows room for innovation and entrepreneurship in the GPS industry both at home and abroad.” The study emphasized the need for the United States to maintain its leader- ship position in satellite navigation technology.423 Because GPS originated as a military system, national security concerns dominated its initial evolution. As a recent report noted, “the military is necessarily risk-averse in its approach to technology,” reflecting its objec- tive of unilateral advantage on the battlefield.424 These security concerns have, however, been largely overcome by a combination of factors includ- ing new technological solutions which should continue to provide a mili- tary-competitive edge, the growth in the capabilities of foreign users, and the risks which would be posed by the emergence of competing systems.425 As the U.S. security perspectives became less dominant, U.S. commer- cial interests sought to win acceptance of the view that the successful ex- ploitation of GPS required, above all, a stable policy framework guarantee- ing access for all users to the most accurate signals the system could offer. For GPS applications to continue to develop (see below), it is essential the 422 Legislative Provisions, Global Positioning System. “In view of the pressures on the defense budget, the necessity for increased civil-military cooperation, the importance of dual- use technology for economic competitiveness and conversion, and the President’s interest in effective infrastructure investments, the committee believes that the time has come to examine carefully a number of GPS technical and management issues. It is clear that GPS offers the potential to revolutionize the movement of people and goods the world over. Civil and commercial exploitation of GPS could soon dwarf that of the Department of Defense and lead to large productivity gains and increased safety in all trans- portation sectors.” National Defense Authorization Act for Fiscal Year 1994, Conference Report to Accompany H.R. 2401, Report 103-357, U.S. House of Representatives, Washington, D.C., 10 November 1993, p. 63. 423 The Global Positioning System: Charting the Future, National Academy of Public Administration, Washington, D.C., 1995, p. vii. 424 Scott Pace et al., The Global Positioning System. 425 Ibid., pp. 186–201. For security-based criticisms of the current U.S. policy, see John Mintz, “U.S. Opens Satellites to Civilians,” The Washington Post, 30 March 1996. 

162 CONFLICT AND COOPERATION United States provide a continuous, stable signal, with a widely accepted commitment to avoid the imposition of arbitrary policy changes such as encryption or user fees. A stable policy framework is also seen as the basis for maintaining the current GPS system as a de facto standard. Ultimately, congressional action may be desirable to secure policy stability, an under- taking which should prove possible given the history of broad bipartisan support and common agreement in this area. The question of standards has enormous stakes, not only in terms of market investments and product development, but also as the system be- comes more pervasive, internationally accepted standards become crucial to a wide range of civil, commercial and military applications involving issues such as encryption, safety certification, and international spectrum specifi- cations. A formal, national policy on GPS, providing a predictable environ- ment for both public service and private business decisions, is essential. A Stable Policy Framework for Investment From 1983, when President Reagan first offered access to GPS, interna- tional reliance on GPS has expanded exponentially. To meet the concerns of increasing reliance on GPS, and the resulting need for a clearly articu- lated and stable policy framework, in March 1996 the Clinton administra- tion issued a presidential decision directive approving a comprehensive na- tional policy on the future management and use of the U.S. Global Positioning System.426 The presidential directive was designed to address the broad range of military, civil, commercial, and scientific interests, both national and international, concerned with GPS operations. Its fundamental objec- tive was to establish a clear, high-level commitment to a stable policy envi- ronment for the development of international standards facilitating both private and public sector investments. This policy stability is of special interest to public authorities which are increasingly dependent on GPS for the provision of critical public functions. To this end, the policy guidelines adopted for the operation and manage- ment of GPS under this directive included decisions to • provide the GPS Standard Positioning Service for peaceful, civil, com- mercial, and scientific use on a continuous, worldwide basis; • make the service available free of direct user fees; • discontinue the Selective Availability of GPS signals “within a de- cade,” thereby making available for commercial use the most precise posi- tioning information, formerly reserved only for military applications; 426 See U.S. Global Positioning System Policy Fact Sheet, The White House, OSTP/NSC 29 March 1996. 

SUPPLEMENTS 163 • seek the acceptance of the U.S. GPS as a standard for international use, while ensuring an appropriate balance between international security interests and the requirements of international, civil, commercial, and scien- tific users; • establish a permanent GPS Executive Board chaired by the Depart- ments of Defense and Transportation, charged with the management of GPS in consultation with interested parties. Lessons from GPS The development of the Global Positioning System captures many of the special features of high-technology development and competition. As noted, the industry grew as a result of U.S. defense expenditure on a military system, although it does not represent the traditional spin-off of defense technologies, insofar as it is the additional capacity of the defense system itself which provides the infrastructure for the industry. A separate GPS satellite signal was made available for commercial use and, as a result, industry investments in R&D created an array of productivity-enhancing information products. The initial government role was critical to the development of both the technology and the infrastructure. At its origins, the start-up of such a system was most unlikely, not to say impossible, on the part of purely private investors, in terms of both the cost and the untried concept. How- ever, because the government eventually adopted a dual-use approach to the GPS technology, commercial and non-defense public use of the system has expanded far beyond the initial military applications. At the same time, the rapid technical development driven by commercial applications has contrib- uted to increased technical capability and availability for military systems. The military-commercial synergies thus obtained illustrate the power of a dual-use approach to new technologies. The Growth in Applications The breadth of benefits made possible by the GPS system were certainly unforeseen when the initial U.S. investment was made. Applications now extend to key public infrastructures such as the Global Information Infra- structure (GII), the Internet, mobile communications and other global wire- less applications, and transportation. Public safety applications include civil aviation, disaster management and response, and emergency fire, police, and medical response. The potential applications for private automobiles, and the transport industry generally, are immense. The increasingly wide- spread use of differential GPS for augmenting basic GPS signals, yields great accuracy, i.e., within centimeters. This capability is translating into 

164 CONFLICT AND COOPERATION “an incredible array of applications,” such as demonstrating new systems for landing aircraft in bad weather; robotic plowing, planting, and fertiliz- ing of fields, with tremendous economies and productivity gains; monitor- ing train and ship locations and even tracking and contributing to the clean- up of oil spills; carrying out twenty-four-hour remote-controlled mining operations; and optimizing the positioning of emergency vehicles to provide the critical response-time margin that saves lives.427 The Central Role of International Cooperation For an information industry to continue to grow globally, however, it requires that national policy evolve into a stable international policy frame- work. The role of foreign public and private users, and their perceptions and needs, have to be taken into account in U.S. policymaking. As the system evolves toward a de facto global information utility, the need to ensure that the requirements of all stakeholders are accommodated becomes paramount. By seeking regional agreements with Japan and Europe on security and economic issues, the adoption of GPS as a global standard for position location, navigation, and timing can be assured. By pursuing a cooperative approach, the rapid expansion of the industry may continue to be unhindered by potentially debilitating conflict over standards or by the artificial segmentation of global markets. Finally, notwithstanding these policy challenges, it is important to keep in mind the unprecedented opportunities offered by the Global Positioning System. The growth in GPS applications offers major benefits in terms of consumer welfare, national productivity, and the ability of national, state and local governments to better perform their missions. The evolution of this system highlights the unforeseen benefits to be derived from sustained public investment in new technologies, the benefits of the dual-use ap- proach for military acquisition, and the central role of international coop- eration. As information technology becomes increasingly important to in- ternational security and commerce, the experience of GPS provides an instructive model for cooperative efforts to promote commercial competition and col- lective security. SUPPLEMENT E. DISCRIMINATORY PUBLIC PROCUREMENT: PROSPECTS FOR PROGRESS Of the many tools governments use to pursue their industrial policy goals, one of the oldest and most direct is targeted government procurement. To 427 See Allocating Federal Funds for Science and Technology, p. 49. 

SUPPLEMENTS 165 support noncompetitive domestic companies or to assist in the development of what are expected to be “sunrise” industries, governments often reserve attractive contracts for local suppliers. This can be done through explicit restrictions on who may bid on a contract, or by stated preferences for local suppliers. More often, the government’s policy goes unstated, and the mechanism through which it supports local companies is shielded by opaque procure- ment procedures. The Government Rationale for Discrimination In recent years, high-technology industries have been a favorite target of directed government procurement. An experience in Norway in the late 1980s offers a particularly clear example of how some governments have sought to use government procurement to support domestic industry. At that time, the city of Oslo planned to procure an automated toll collection system. After following a competitive bidding procedure, the Oslo authorities decided to sign a contract with a U.S. supplier whose system was cost-competitive and based on proven technology. At that point the Norwegian government— which was providing R&D support to a fledgling Norwegian producer of auto- mated toll collection systems and was also funding the Oslo toll road project— entered the picture and directed Oslo to buy the Norwegian product. Oslo resisted and the matter became the focus of intense media attention. In the process, the Norwegian Minister of Transport stated publicly that if the com- petition had been between the U.S. firm and another foreign firm, the U.S. firm would have won the contract. However, he continued, in this instance the Norwegian government was obliged to support its domestic supplier because the Oslo contract would give the winner a tremendous boost toward winning similar contracts throughout Europe. It would provide a field test in a Euro- pean environment, drive down unit costs for the Norwegian producer through economies of scale, create customer acceptance, and influence the develop- ment of European standards for toll collection equipment. Ultimately, the Norwegian firm was awarded the Oslo contract and a subsequent contract in the city of Trondheim.428 The Power of Government Purchases The power represented by government purchases is staggering. Accord- ing to U.S. government estimates, total procurement by central govern- 428 Norway is a member of the Government Procurement Agreement (GPA), and the United States challenged both of these Norwegian procurements in that forum. The U.S. company received compensation for the costs it incurred in bidding on the Oslo project, but was unable to overturn the contract awards. 

166 CONFLICT AND COOPERATION ments, sub-central governments, and electrical utilities in the European Union is approximately $115 billion per year. Central government procurement in Japan approaches $30 billion. The figures are even more dramatic when the purchasing activities of state-owned enterprises are taken into account. Ac- cording to a World Bank study, state-owned enterprises accounted for 10.7 percent of economic activity in developing countries and 4.9 percent of economic activity in developed countries from 1978 to 1991.429 Govern- ment procurement policies and decisions are thus one of the most important potential influences on international trade flows.430 As a result, the elimination of restrictive procurement practices is a high priority in international trade negotiations. The United States has been especially active. With Japan alone, the United States has negotiated sectoral procurement agreements covering supercomputers, computers, cellular tele- phones, medical equipment, and telecommunications equipment. As with other trade and investment issues—discussed in the body of this Report— these bilateral initiatives are designed to address discriminatory practices for which effective multilateral arrangements are not in place. The Exclusion of Government Procurement from Multilateral Disciplines Although the importance of these purchases was widely recognized, when the international community adopted the General Agreement on Tariffs and Trade in 1948, liberalization of government procurement procedures were not included. In fact, Article III.8 of the GATT specifically excepted public procurement from the principle of national treatment. This exclusion was a 429 Within these statistics, there is considerable variation among countries. During the period 1978 to 1991, state-run enterprises accounted for 10.5 percent of economic activity in France, and 7.1 percent of economic activity in Germany, but only 1.2 percent of economic activity in the United States. Bureaucrats in Business, Oxford University Press, New York, 1995, pp. 268–271. 430 It should be noted, however, that the efficacy of using government procurement to direct business toward domestic high-technology ventures in some sectors is of uncertain value. In the United States, the federal government has succeeded in developing technologies, such as space launch vehicles, where: (1) there was a clear government need and (2) the lack of commercial demand and the high cost of entry made private sector investment uneconomical. The Norwegian toll road case, in which the government sought to provide critical support to launch a new technology, illustrates a third category in which some would argue for the efficacy of government intervention. As a means of supporting new generations of existing commercial technologies, however, the benefits of selling to the government are far less clear. In the semiconductor industry, for instance, government specifications tend to lag behind the state of the art and can impose additional requirements that have no commercial benefit. Companies that become dependent on government contracts may survive, but they are likely to see a steady erosion of their competitive position in the far-larger commercial market. 

SUPPLEMENTS 167 foregone conclusion from as early as 1946 when the GATT Preparatory Committee recommended in its first session that government procurement not be covered by the most-favored-nation (MFN) clause. The participating governments saw their procurement decisions as inherently political and were unwilling to give up the use of one of their most flexible industrial policy tools. This outright exclusion of such a large portion of economic activity was recognized by many as a serious gap in GATT coverage, but for thirty years the gap remained unfilled. Only in 1979, with the conclusion of the Government Procurement Agreement as one of the Tokyo Round codes, did the GATT contracting parties begin to introduce a measure of discipline and international obligation to govern- ment procurement practices. For the purchases it covers, the GPA requires signatories to accord MFN and national treatment to other signatories. It also mandates certain procurement procedures designed to ensure that its objectives are not undercut through the introduction of bid requirements or procedures biased toward a local bidder. The framers of the GPA sought to create an incentive for additional participation by making the benefits of the GPA available only to other signatories. The GPA, however, also has a number of serious shortcomings: • First, participation in the GPA is voluntary, and the membership from the beginning did not extend essentially beyond OECD countries.431 As a result, the GPA has done nothing to change procurement practices in the rapidly growing countries which offer great opportunities for increased exports from the industrial countries.432 The expectation that additional countries would join the GPA in order to obtain the benefits of access to government procurement in the developed world has proven to be unfounded. Most developing countries simply do not expect that their suppliers would be net beneficiaries of the agreement, and under those circumstances they are unwilling to circumscribe their own flexibility in procurement deci- sions. Limited participation has left procurements in most developing countries open to arbitrary and nontransparent purchasing procedures. Moreover, because transparent procedures offer one of the best defenses against cor- 431 The following countries currently apply the GPA: Canada; the European Union (in- cluding Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxem- bourg, the Netherlands, Portugal, Spain, Sweden, and the United Kingdom); Israel; Japan; Norway; Switzerland; and the United States. Korea will begin to apply the GPA on January 1, 1997. Singapore is in the process of acceding to the agreement. 432 U.S. exports to the ten “Big Emerging Markets” (BEMs) identified by the U.S. Com- merce Department already exceed U.S. exports to either Japan or the European Union, and by the year 2000 they are expected to exceed U.S. exports to Japan and the EU combined. None of these eighteen countries now apply the GPA, and only Korea and Singapore are likely to begin applying the agreement in the near future. 

168 CONFLICT AND COOPERATION rupt behavior in government procurement, failure to sign up more members to the GPA has meant that one of the great potential values of the GPA has gone unrealized. • Second, the GPA disciplines apply only to procuring entities “sched- uled” by the signatories. Although the United States scheduled most of its central government procurement in 1979, “entity coverage” by other coun- tries was more limited, resulting in significant imbalances. For instance, prior to the Uruguay Round agreement, the European Union and most other signatories refused to schedule their telecommunications, electrical, trans- portation, and water utilities. In fact, very few countries scheduled their state-owned enterprises, whatever their field of business.433 U.S. compa- nies operating in these sectors thus obtained no new market access under the GPA. The imbalance was compounded by the fact that shortly after the GPA went into effect the United States introduced competition—including competition from international suppliers—into its own telecommunications and electrical utility markets. U.S. dissatisfaction with this imbalance, combined with European desires to obtain more U.S. coverage of state and local enti- ties, resulted in a serious U.S.-EU trade dispute. This dispute led to agree- ments in 1993 and 1994 to undertake coverage of electrical utilities and subcentral entities,434 but telecommunications procurement issues were not resolved, resulting in retaliation by both sides. • Third, the GPA procedures are difficult to understand. While they are laudable in their intentions, the excessive detail of the GPA procedures, particularly by comparison with other elements of the GATT, tends to wrap the GPA in a cloak of mystery. It is only a select few, the high priests of the agreement as it were, who understand the commitments and undertake to interpret them for the rest of the world. Even if they were to feel that GPA participation were otherwise warranted, this factor deters nonsignatories 433 State trading companies are covered by Article XVII of the GATT, which requires non- discrimination in their purchasing practices, but exempts purchases that are ultimately intended for governmental purposes. 434 Coverage of electrical utilities and U.S. subcentral entities was achieved first through a bilateral agreement, and, after 1 January 1995, through coverage of most of these entities in the GPA. The bilateral agreement did not, however, prove satisfactory. For example, U.S. suppli- ers of heavy electrical equipment claim that they continued to encounter discriminatory pro- curement procedures and inadequate remedies in the German market. As a result, on 30 April 1996 the U.S. Trade Representative identified Germany (under Title VII of the 1988 Omnibus Trade and Competitiveness Act) for “a significant and persistent pattern” of discrimination and failure to adequately implement (its government procurement) obligations.” This identifica- tion of Germany triggered a sixty-day period of consultations including, at this writing, the possibility of retaliatory action by the U.S. government if the consultations do not prove successful. Whatever the eventual resolution of this dispute, it highlights both the importance of high-technology markets to even the most industrialized countries and the persistence of government procurement of high-technology products as a source of friction within the inter- national system. 

SUPPLEMENTS 169 from participation, to the extent their authorities believe they are dependent on foreign expert interpretations of the agreement and do not wish to make themselves vulnerable to arcane dispute settlement procedures. • Fourth, the GPA is based on an all or nothing principle that differs from the rest of the GATT. If an entity is scheduled and the GPA is applied for that entity to another GPA signatory, the entity is subject to all GPA disciplines. If not, it is subject to no disciplines. This all-or-nothing approach leaves no room for the type of steady improvement in market access that has been so successful with tariff reductions in the GATT. Since its adoption, the Government Procurement Agreement has been modified through almost continuous negotiations, but GPA membership has expanded only marginally.435 In an attempt to plug procedural holes that permitted signatories to follow the GPA rules and still steer contracts to local suppliers, most modifications to the GPA during the 1980s added new procedural requirements, further complicating the agreement. Progress in the Uruguay Round During the Uruguay Round, substantial progress was made on several gov- ernment procurement fronts.436 For the first time, countries were permitted to take derogations from their national treatment obligations, thus moving a step away from the all-or-nothing principle. The GPA was also expanded to cover the procurement of services (including construction services) and subcentral entities and public utilities. When combined with additional entity coverage at the central government level, these changes resulted in a tenfold increase in the value of procurements covered by GPA procedures.437 Another major addition to the GPA was a requirement that signatories provide procedures for bid challenges as well as judicial remedies in disputes over the award of a contract. The GPA thus serves a valuable purpose. It has shown that govern- ments can subject themselves to international obligations in the field of gov- ernment procurement, and it has opened hundreds of billions of dollars of contracts to international competition. None of these changes, however, adequately addresses the fundamental flaws with the GPA as identified above. The prospects for any significant expansion in membership remain limited;438 entity coverage among existing 435 Most of the growth in GPA membership has occurred through enlargement of the European Union. Greece, Spain, and Portugal all joined the GPA after joining the EU. 436 Technically, the government procurement negotiations, which were ongoing at the time the Uruguay Round began, were not part of the Round. In practice, however, revisions to the GPA made up an important part of the final Uruguay Round package. 437 “Creating open competition in government procurement,” WTO Internet Home Page, 1996. 438 Unlike most of the other Tokyo Round codes, the GPA was not incorporated as an integral element of the Uruguay Round single act, and membership remains voluntary. 

170 CONFLICT AND COOPERATION signatories is still incomplete; procurement procedures remain unnecessar- ily intimidating, and the structure of the agreement provides slim hopes for achieving improvements outside of the existing membership. As a result, some analysts believe the time has come for a new approach; an approach that will recognize and lock in the progress made to date by existing GPA members, demystify the agreement, and encourage a steady process of lib- eralization rather than expecting new GPA members to leap directly to full liberalization.439 In doing so, an expanded GPA would expose procurement practices to public scrutiny, thereby contributing to the worldwide effort to reduce corruption in government. A broader, more effective agreement would also offer a means to reduce a major source of friction in high- technology trade while providing the benefits of transparent competition to government acquisitions of high-technology products. A New GPA A new Government Procurement Agreement would look to the GATT itself, and in particular the success of the GATT in eliminating quotas and reducing tariffs, as a model for steady improvements in trade liberaliza- tion.440 With this model in mind, a reconstituted agreement might embody the following six elements: 1. Membership: all WTO members would be required to join the GPA. 2. Entity coverage: GPA members would be required to make binding commitments covering purchases by all of their government entities, includ- ing all state-owned enterprises. Under a reconstituted GPA, unlike the existing one, however, these commitments could involve exceptions from national treatment, such as a specified preference level for domestic suppli- ers. Existing GPA members should, at a minimum, bind their existing commitments with respect to other current GPA signatories (including the 439 See the statement by R. Michael Gadbaw in Sources of International Friction and Cooperation in High-Technology Development and Trade. 440 At the first GATT/ITO Preparatory Committee meeting in London in 1946, the United States set out five objectives for the new organization, many of which could be applied today to the GPA. Professor John Jackson describes these objectives as follows: “(1) Existing barriers to international trade should be substantially reduced; (2) International trade should be multilateral rather than bilateral; (3) International trade should be non-discriminatory; (4) Stabilization policies for industry and agriculture...[are] so intimately related to international trade policies that the two must be consistent and coordinated; and (5) The rules for interna- tional commerce should be drafted so that they would apply with equal fairness and equal force to the external trade of all nations regardless of whether their internal economies were organized upon the basis of individualism, collectivism or some combination of the two.” John H. Jackson, World Trade and the Law of the GATT, Bobbs-Merrill Company, New York, p. 54. 

SUPPLEMENTS 171 national treatment element inherent in those commitments). Transparency (item 4 below) and remedy (item 5 below) procedures should apply to all entities in all WTO members. 3. Liberalization method: GPA members would engage in regular ne- gotiations to improve their degree of procurement liberalization. In these negotiations, as in tariff negotiations, countries could trade off reductions in their discriminatory practices in exchange for similar reductions by others. 4. Procedures: The GPA would be amended to focus on the key com- mitments necessary to achieve improved procurement opportunities. These include requirements for: (a) adequate notice; (b) neutral standards (i.e., standards should not be set in a way that gives an advantage to any one bidder); (c) objective and pre-specified bid criteria; (d) public bid opening; and (e) award of contracts to the lowest compliant bidder or demonstrably best overall value on the basis of the objective criteria. The essential ele- ments of these commitments should be distilled and inserted into the GPA so that the principles are readily apparent and readily applied. 5. Remedies: The current procedures on bid challenge, access to an independent appeal body, and dispute settlement should be retained. In addition, once the GPA becomes a requirement for all WTO signatories, cross-retaliation441 should be permitted. In the event that a WTO member were found by a panel to have violated the GPA but did not implement the panel ruling, cross-retaliation would increase the options for retaliation available to an aggrieved party. This should provide an improved deterrent to viola- tions of the agreement. 6. National Treatment/MFN: To accord with the shift from an all-or- nothing approach to an approach that seeks broad and steady liberalization, the national treatment requirement of the GPA would have to be eliminated, although the MFN requirement should be retained as a core principle. Current Prospects Leading WTO members have begun to consider the need for changes to the GPA. At the meeting in Kobe, Japan, on 1921 April 1996, the trade ministers of Japan, Canada, the European Union, and the United States (the Quad ministers) addressed the need to bring additional members into the GPA and endorsed an approach other than the all-or-nothing approach to procurement liberalization. The chairman’s statement from that meeting 441 Cross-retaliation would permit retaliation outside of the government procurement area if a GPA member were to violate its GPA commitments and then fail to implement a panel ruling against it. Full cross-retaliation would also permit retaliation in the government pro- curement area for violations related to other WTO agreements. Currently, the GPA does not permit cross-retaliation. 

172 CONFLICT AND COOPERATION reports, “We agreed to renew our efforts to expand membership in the WTO Agreement on Government Procurement (GPA) and to improve its disci- plines through reducing barriers to government procurement. As a first step, we agreed to initiate work on an interim arrangement on transparency, openness and due process in government procurement, which would help to reduce corruption as an impediment to trade.”442 While the Quad ministers did not provide any further detail on what an interim agreement on transparency, openness, and due process might be, the market-opening potential of such an agreement is enormous. If an interim agreement included the procedures described in points (4) and (5) above, it would offer three principal benefits. First, such an agreement would serve as a means by which significant numbers of additional countries could sign on to government procurement disciplines. Second, even without a national treatment commitment, transparency provisions alone would greatly improve the environment for foreign companies seeking to do business in signatories to the interim agreement. Third, as the Quad ministers’ statement suggests, a truly transparent procurement system is incompatible with illicit payments. For all of these reasons, the transparency initiative represents a welcome step. In the short term, however, the road to adoption of an interim agreement with provisions on transparency, openness, and due process is likely to prove difficult. A recent statement by the ASEAN economic and trade ministers opposed discussion of corruption issues (including the issue of transparency in procurement) at the Singapore Ministerial meeting. This is of particular concern because Singapore, the host of the Ministerial meet- ing, is an ASEAN member. The position taken at the meeting of the ASEAN economic and trade ministers highlights the different perspectives of countries at different lev- els of development. As described in the first section of the report, it also illustrates the different policy objectives of competitors for high-technology industry in the global economy. And at a more fundamental level, it may be the ASEAN position reflects differing values and a determination on the part of newly industrialized countries to assert their independence from Western precepts. Responding to this ASEAN challenge will require a clear presentation of the benefits that can accrue to those countries participating in efforts to increase transparency. In this regard, the Quad ministers’ conclusion that transparency, broader participation in the Government Procurement Agree- ment, and improvement of its disciplines is a positive step. For further progress, it is likely to be necessary to undertake the fundamental reform of the Government Procurement Agreement outlined above. 442 Chairman’s Statement, Kobe Quadrilateral Trade Ministers’ Meeting 1921 April 1996, p. 2.