Seven
Esprit
Opening the Door

By 1980 the dimensions of Europe's crisis in telematics were becoming clear. A decade and a half of national-champion strategies had failed to close any gaps between Europe and the telematics leaders, Japan and the United States. Europe faced the prospect of a neck-and-neck race between American and Japanese industries that would leave Europe eating dust. This was the challenge facing Europe's policy-makers at the beginning of the new decade.

My argument is that the crisis led European elites to question their unilateral, national strategies for promoting telematics industries. Policy-makers in each of the major countries entered what I have labeled the adaptive mode: They were all searching for new approaches. The evidence consists of high-level, high-profile, official studies of telematics policies in the United Kingdom, France, and the Federal Republic of Germany. In each case blue-ribbon commissions produced new telematics plans. Simultaneously, the same governments were agreeing to collaborate in the Commission's ESPRIT program. The evidence is clear: Across Europe, governments were in a state of crisis-induced adaptation in the early 1980s.

The previous chapter also spelled out the technological changes that underpinned the intensified interest of private enterprises in interfirm alliances. In this chapter, I show how a transnational coalition of the major telematics companies combined with an entrepreneurial Commission of the European Communities to produce

a collaborative response to the crisis. To be sure, collaboration was by no means the necessary outcome of the policy adaptation going on in the various countries. Collaboration emerged because the Commission and the industry coalition (the Twelve Roundtable companies) were able to gain the necessary political support from governments looking for new directions in telematics policy.

Crisis-Induced Adaptation

Although collaboration eventually emerged in the form of ESPRIT, it did not replace national programs to build up domestic IT industries. In fact, national programs to support IT flourished with renewed vigor in the early 1980s. In this first section I describe the adaptive process in France, Germany, and the United Kingdom, focusing on changes in national policies. It bears remembering that although each country adopted new national strategies, the ESPRIT program was being created during the same period (1982–1984). In other words, collaboration did not replace national policies but took its place beside them.

Several themes link the national cases. First, each government believed more firmly than ever in the importance of IT in driving economic development and growth. Even those governments ideologically opposed to public intervention in industry (the Helmut Kohl government in Germany and the Thatcher government in Great Britain) provided massive new supports to the IT sectors in the early 1980s. Second, government officials, company executives, and Brussels technocrats all spoke in the early 1980s of the urgent necessity of meeting the Japanese and American challenge.

Third, as Malerba points out, government and industry had come around to the view that strength in telematics required a presence in the design and manufacturing of advanced semiconductors. Both business and government had missed the boat the first time around, and Europe by the late 1970s was absent from world microprocessor and memory-chip markets. Only then did governments conclude that "a domestic productive capability in LSI devices was of strategic value for reasons of national security and for sustaining the electronic and the manufacturing industries as a whole ."^[1] Although it is not his analytic concern, Malerba thus provides evidence

[1] Malerba, Semiconductor Business , 162, my emphasis.

of the kind of policy adaptation I expect. Further confirmation of this adaptation comes from a well-placed observer, Pasquale Pistorio, president of the Italian firm SGS: "European governments are beginning to see that they've got to protect this very important industry [semiconductors] if they're going to preserve the quality of the European electronics industry."^[2] The programs described in this section show that governments were looking for new ways to support their domestic firms' reentry into advanced digital microelectronics.

The semiconductor companies themselves were also trying to return to advanced digital ICs; by the early 1980s Siemens, Philips, Thomson, SGS, and Inmos were all starting again to produce highly integrated memory devices.^[3] A Financial Times survey of the European electronics industry in 1984 summarized the by-then general consensus that the production of standard ICs (especially memories) was essential as it drove product, design, and production innovations.^[4] Furthermore, as final electronics systems were increasingly built into the chips, the competitiveness of European systems makers increasingly depended on mastery of chip technologies. Executives from Europe's major semiconductor firms confirmed this observation. Klaus Ziegler, vice-president of the Components Group at Siemens, explained his company's efforts starting in the early 1980s to rejoin the competition in advanced semiconductors: "First, for a company like us that integrates microelectronics into its finished products, it's essential to avoid becoming dependent on our competitors. Secondly, increasing system know-how is going into components all the time, and we're justifiably reluctant to have our expertise siphoned off by others."^[5] Cees Krijgsman, managing director for ICs at Philips, speaking of Europe's position in electronics and advanced manufacturing more generally, declared, "To be competitive, we simply have to be strong in semiconductors."^[6]

Fourth, and last, promoting collaboration among companies and between industry and academia became a feature of European IT

[2] Quoted in Thane Peterson, with Amy Borrus, "Europe's Chipmakers Pull Out of a Long Losing Streak," 22.

[3] Paul Betts, "Thomson Has Another Try," Financial Times , 25 October 1985, p. 18.

[4] Guy de Jonquieres, "Electronics in Europe," Financial Times , 28 March 1984, Survey, p. 1.

[5] Siemens, "Japan's Challenge—Europe's Response: Interview with Klaus Ziegler," 18.

[6] Quoted in Peterson, "Europe's Chipmakers," 22.

support programs. Such links were a central part not only of ESPRIT but also of national programs in Germany and the United Kingdom. Collaborative R&D was perceived as a crucial part of American and especially Japanese successes in promoting high-technology industries. In a sense, then, collaborative R&D in Europe was an adaptive response based on imitation of the acknowledged leaders. Table 7.1 presents a chronology of the major events in IT policy.

France

French electronics and data-processing programs in the first years of the Mitterrand government were even more nationalist than previous ones, difficult as that may be to imagine. The French trade deficit in electronics goods doubled from 1981 to 1982, to FFr 12 billion.^[7] President Mitterrand was determined to reverse that trend. The Giscard government, prior to leaving office, had commissioned a series of studies on French high-tech capacities. The new government expanded these studies and used them as a basis for the colloque national on research in January 1982, designed as a forum in which those interested (unions, industries, banks, researchers, civil servants) could provide input for a new science and technology policy. The new policy emerged in July 1982 with the Loi d'orientation et programmation (LOP). The LOP laid out ambitious goals to make France "the world's third scientific power" by the 1990s.^[8] The overall goal was to spend 2.5 percent of GNP on domestic R&D by 1985, from a base of 2.01 percent in 1981, with the effort focused on seven high-priority programmes mobilisateurs .^[9] One of the priority programmes was for the electronics sector, the filière électronique .

Simultaneous with the LOP process the government set up within the Ministry of Research and Technology a special Mission filière électronique to study that sector in particular. Under Abel Farnoux (who lent his name to the report) the Mission submitted its conclusions in March 1982. The Farnoux Report declared, "If France is to maintain its independence, master the new communications systems and put the crisis behind it, it is imperative for the country

[7] David Marsh, "Bid to Narrow the Gap," Financial Times , 21 March 1983, Survey, p. 6.

[8] OECD, Innovation Policy , 67.

[9] Ibid., 71–75.

to master the key activities of the electronics sector."^[10] The report became the basis for the Programme d'action pour la filière électronique (PAFE). The PAFE envisioned a wholesale revamping of French telematics policy, with the different sectors seen as parts of a single whole extending from ICs to final systems in communications, consumer electronics, data-processing, and office and factory automation. Again, ambitious national goals were the order of the day. For the period 1983–87 France would shoot for a production increase of 9 percent (versus the actual trend of a 3 percent increase), 50,000 new jobs in the sector (trend: 10,000 lost), and a trade surplus of FFr 14 billion (trend: FFr 20 billion deficit by 1987).^[11] This program would be accomplished via several major actions:

Structuring the program through national projects resembled the Japanese model for mobilizing whole industries.^[13]

In semiconductors French goals were ambitious: The microelectronics segment of the filière électronique aimed at French independence in semiconductor technology by 1986. The government was convinced that "without integrated circuits, the entire French electronics industry would crumble."^[14]

The report also mentioned the possibility of eventually including firms from other European countries in the national projects and

[10] Quoted in ibid., 218.

[11] Ibid., 219.

[12] Such an obligation was apparently quite common in the early 1980s in France. Users of both ICs and computers were pressured to buy French (Kenneth Dreyfack, "France Wants Bigger Piece of Pie," 98; John Morris, "France Plans DP Sell-Off," 66).

[13] Robert T. Gallagher, "French Want National as Partner," 104.

[14] Quoted in Thomas R. Howell et al., The Microelectronics Race , 170.

possibly even American or Japanese firms in the long term.^[15] But, as the OECD study of innovation policy in France points out, the "PAFE was not set in an international context from the outset. It was not until over a year later after the programme was adopted in September 1983 that France made the first attempts at achieving some complementarity between the French vision of the electronics sector and European action."^[16] Indeed, President Mitterrand announced early in his term that his objective was "reconquering the domestic market" and making France the third IT power behind the United States and Japan.^[17]

The government reiterated in September 1983 its intention to invest FFr 140 billion over five years to achieve what Mitterrand called the "great electronic leap forward." Minister for Industry and Research Laurent Fabius declared, "Meeting the challenge of electronics and information technologies is the number one priority in industrial policy for the country."^[18] The total investment included FFr 60 billion from the state, the rest to come from industry.^[19] The state contribution to R&D support was to total FFr 30 billion, approximately double what it would have contributed otherwise. The state contributions would come from several ministries (largest shares from defense and the DGT) and take the form of contracts, grants, and equity funding. In practice the funding channels were extremely complex, so it was hard to tell whether the spending targets were attained.^[20] They almost certainly were not.

France's inability to meet its targets was behind its sudden about-face in 1983 and 1984. When it became clear that France could not possibly afford such an ambitious and nationalist strategy, the French government became a vigorous advocate of European collaboration. How this turnabout occurred is instructive, for it demonstrates the recognition on the part of a government of the failure of a national-champion approach and the subsequent turn toward cooperation.

[15] Robert T. Gallagher, "France Urged to Reorganize Electronics," 105.

[16] OECD, Innovation Policy , 219.

[17] David Marsh, "French Computer Strategy Under Fire," Financial Times , 15 April 1983, p. 32; and David Marsh, "Stronger Emphasis on Joint Ventures," Financial Times , 11 July 1984, Survey, p. 7.

[18] "M. Mitterrand: la France proposera un plan européen de haute technologie en 1984," Le Monde , 27 September 1983, p. 44.

[19] Paul Betts, "France Campaigns for Electronics Collaboration," Financial Times , 28 September 1983, p. 1.

[20] OECD, Innovation Policy , 221.

What is surprising is how quickly the shift occurred: The Farnoux Report was issued in March 1982; by September 1983 the French government was beginning to speak out for European high-technology cooperation.

As the OECD report on France points out, economic growth rates were far lower in the early 1980s than the government had predicted. The economy was not producing at the level needed to support the huge investments planned for the filière électronique . Evidence of trouble came in July 1983, when a large chunk of the PAFE program was placed under the DGT. This switch meant that DGT revenues from its profitable network activities (which were independent of the national budget) would be used to underwrite the electronics plan, including huge subsidies to loss-making firms like Bull and Thomson and in addition to the usual funding for telecoms suppliers like CGE and CGCT. Not surprisingly, the reshuffling quickly ruined the DGT's former extremely strong financial position.^[21] The following year the Direction des Industries Electroniques et de l'Informatique, which had been in charge of the filière électronique , was quoted as saying, "Unfortunately, in the present context, we cannot afford to make the effort required. Certainly for the new [microelectronics] plan which will start in 1987, a European effort will have to be considered."^[22] In short, France was rapidly discovering the limits to what one European nation, even one of the largest, could hope to achieve autonomously in high technology.

Thus, by the fall of 1983 the French were already beginning to sing a cooperative tune. The Mitterrand government in September 1983 circulated a memorandum among its fellow EC members advocating "un espace industriel européen ." President Mitterrand announced at the end of the month that France was preparing to propose a European plan for high technology, especially IT. The French president declared in a speech: "We should show the same ambition in high technology; our independence as Europeans and our identity are the stakes. Otherwise we will be submerged by competitors like the United States and Japan."^[23]

[21] Ibrahim Warde, "French Telecommunications," 109–10.

[22] Howell et al., Microelectronics Race , 171.

[23] "M. Mitterrand: la France proposera un plan européen de haute technologie en 1984," Le Monde , 27 September 1983, p. 44.

By this time France also had a new minister for industry and research, Fabius, who was a strong advocate of increased European R&D collaboration. Fabius had replaced Jean-Pierre Chevènement, who was in office during the preparation of the LOP and the PAFE and who was a vigorous proponent of French technological independence. Indeed, Chevènement had issued instructions that French scientists publish their research in French and defend French as a scientific language in international meetings.^[24] Chevènement resigned amid controversy over his interventionist approach to the nationalized industries in March 1983 and was replaced by Fabius.

As Mitterrand began talking up European cooperation on high technologies, Fabius also took up the theme. The minister for industry and research spoke out in favor of European collaboration in electronics and telecommunications, including the gradual opening of public procurement markets. The government also expressed the view that alliances among European industrial groups should be favored over those with non-European firms.^[25] The French even criticized an AT&T-Philips deal as a Trojan horse for the American company.^[26] At the same time, however, French electronics firms were striking alliances with U.S. companies—Thomson with National Semiconductor, Matra with Harris. Nevertheless, from that point on the French became consistent supporters of the Commission's efforts leading up to ESPRIT.^[27]

Just over a year later, Fabius ascended to the post of prime minister. Mitterrand's selection of Fabius reflected both his hopes for the technological modernization of French industry and his growing commitment to European cooperation. As chairman of the Council of Research Ministers of the EC during the first half of 1984 Fabius had been a staunch supporter of the Commission's technology plans and had been a consistent backer of ESPRIT, whose first phase was approved during Fabius's tenure. To fill his former cabinet position, Fabius chose Hubert Curien, though the industry brief was moved

[24] John Walsh, "France Readies New Research Law," 712.

[25] Paul Betts, "France Campaigns for Electronics Collaboration," Financial Times , 28 September 1983, p. 1.

[26] Guy de Jonquieres, "Government Aims to Float Off Slice of BT in U.S.," Financial Times , 20 September 1983, p. 6.

[27] Paul Betts, "Government Encourages Closer European Collaboration," Financial Times , 11 July 1984, Survey, p. 2; and David Marsh, "Stronger Emphasis on Joint Ventures," FT , 11 July 1984, Survey, p. 7.

elsewhere and the ministry once again covered research and technology. The appointment of Curien was also significant. Curien was at the time president of the European Science Foundation and was a former chairman of the council of the ESA. In other words, Curien was a committed and experienced Europeanist on technology matters.^[28] Thus, two key positions in the Mitterrand cabinet after July 1984 were filled by advocates of European collaboration. Curien would later be one of Mitterrand's point men in selling EUREKA to France's partners. In barely three years the French government had evolved from a vigorous promoter of national champions and independence into an enthusiastic advocate of European collaboration.^[29]

Federal Republic of Germany

Germany, like the United Kingdom, frequently finds its desire to advance its IT industries at odds with its noninterventionist instincts. The Ministry of Economics tends to uphold market solutions to industrial problems, while the Ministry of Research and Technology (BMFT) has been willing to formulate programs to stimulate high-tech sectors. The CDU/CSU-FDP coalition that has governed since 1982 is ideologically uncomfortable with state intervention. Still, the Kohl government, like its SDP predecessors, has instituted programs to stimulate the IT industry.

In the early 1980s the BMFT's Microelectronics Program, begun in 1979, was supporting R&D for chip design, production technology, materials, and applications. Financial support shifted emphasis from data-processing equipment to microelectronics and applications. In 1980 state funding for the computer sector was more than twice that for microelectronics; in 1983 the ratio was 5:2 in favor of microelectronics.^[30] In addition the government agreed in 1984 to help underwrite the Mega project involving Siemens and Philips, whose object was the production of one-megabit static random-access memories (SRAMs) and four-megabit DRAMs. The German government would contribute DM 300 million (Siemens

[28] David Dickson, "France's New Technocrats," 486–87.

[29] See Guy de Jonquieres and Paul Betts, "The Euphoria Is Over," Financial Times , 7 February 1985, p. 14.

[30] Erik Arnold and Ken Guy, Parallel Convergence: National Strategies in Information Technology , 142.

claimed to have allocated DM 2.2 billion for the project, including investment in a new plant).^[31] The German government saw microelectronics as one sector whose fate could not be left entirely to the operations of the market.

The BMFT announced in 1984 a new Informationstechnik program.^[32] The plan had been long awaited, held up in part by objections from the Ministry of Economics.^[33] The new program allotted DM 3 billion ($1.14 billion in 1984) over five years, from 1984 to 1988. The government declared that its IT program was a necessary response to technological competition from the United States and Japan.^[34] The minister for research and technology, Heinz Riesenhuber, declared:

Our share of the world market in high technology goods is stagnating, while Japan's share in the last decade has doubled. We cannot overlook this, if we want to maintain our leading position in the 1990s. . . . To this end, the government has approved the comprehensive plan to support microelectronics, information and communications technology. . . . In this plan, we are documenting our determination to accept the challenges of information technology and to improve our competitiveness.^[35]

The BMFT report noted that IT production (components, consumer electronics, communications, computers, industrial automation) was growing far more rapidly in the United States and Japan than in Germany. The position of the German semiconductor industry was unsatisfactory: Whereas the United States and Japan were exporting ICs to the world, German industry supplied only 60 percent of the country's demand for ICs.^[36]

The Informationstechnik program was the first German IT support scheme in which industry was consulted in the planning stages; fifteen companies, including both large and small ones, helped prepare the program.^[37] Though the plan specified assistance only for

[31] Guy de Jonquieres, "Philips Joins Siemens in Microchip Project," Financial Times , 11 October 1984, p. 1.

[32] This announcement came at about the time the first year of the full ESPRIT program was getting underway.

[33] Arnold and Guy, Parallel Convergence , 140.

[34] Jonathan Carr, "Bonn Wakes Up to the Challenge of the Chip," Financial Times , 28 March 1984, Survey, p. 6; Interview 4.

[35] Quoted in Howell et al., Microelectronics Race , 174.

[36] Rupert Cornwell, "Bonn Bid to Push Ahead in High-Technology Field," Financial Times , 15 March 1984, p. 3.

[37] John Gosch, "Germany Spends for Its High-Tech Future," 101.

R&D, and only up to 50 percent of project costs, the government indicated that state procurement, both civil and military, would provide a further stimulus to industry. The Bundespost alone was to spend a predicted $750 million annually on broadband telecommunications. The largest chunk of the DM 3 billion would go for components R&D (DM 1.41 billion), including a DM 608 million project on submicron-chip technologies. Other major areas were data-processing, industrial automation (CAD and CAM), and telecommunications.^[38] A final noteworthy aspect of the program was that it would emphasize cooperation among companies and between industry, universities, and research institutes. Indeed, collaboration was becoming a keyword in German technology policy.^[39] Two executives from major German electronics firms told me that interfirm collaboration had been stressed by the BMFT in its programs since 1982.^[40]

However, nothing like the enthusiasm of the French for European collaboration evolved in Germany. Rather, the government was divided internally. Foreign Minister Hans-Dietrich Genscher (of the FDP) emerged as the most avid and consistent proponent of European collaboration. As one official in the BMFT put it, "Since 1984, Genscher has mentioned technology in nearly every speech. He always speaks of the necessity of European cooperation in space and information technologies." This official noted a high level of concern for technology questions ever since Genscher appointed Konrad Seitz as director of his planning staff at the Foreign Ministry.^[41] The Finance Ministry, in contrast, consistently questioned the value of European collaborative programs (including the EC's Framework Programme and EUREKA, for example) and sought to restrict their budgets. The BMFT favored collaboration for basic research, as in fusion, space, and synchrotron radiation. For more commercially relevant technologies, the BMFT argued for selectivity, so that European programs would complement national ones. Thus, the BMFT saw the major components of the Framework Programme (including ESPRIT) as essential but would have preferred to reduce the budget for other programs.^[42]

[38] Ibid., 101–2.

[39] Rupert Cornwell, "Bonn to Spend DM3bn on High Technology Boost," Financial Times , 9 March 1984, p. 1; Arnold and Guy, Parallel Convergence , 141.

[40] Interviews 45 and 46.

[41] Interview 4.

[42] Ibid.

United Kingdom

Great Britain during the Thatcher years had been torn between an anti-interventionist ideology and an increasing perceived need to support the IT industries. These contrary impulses led to what Erik Arnold and Ken Guy call "stop-go policies."^[43] Thus, the Thatcher government began with measures to privatize that part of the industry that had come under state tutelage. The National Enterprise Board was transformed into the British Technology Group and told to sell of its shares in ICL and Inmos. The government announced that a £35 million cash infusion for Inmos in January 1984 would be the last.^[44] After offers for Inmos from AT&T and a Dutch consortium had been turned down in 1984, Thorn-EMI bid £144 million, raised later to £192 million, for the state's 75 percent interest in the company. This proposal was approved. The same month (August 1984) the electronics and telecommunications conglomerate STC purchased ICL for £653 million; ICL had required government loans and loan guarantees in 1980–81, though it had achieved a profit in 1981–82. Some political opposition arose to the deal because 35 percent of STC was owned by the American firm ITT; but the government approved the acquisition after ITT announced a reduction of its interest to 25 percent.^[45]

In late 1980, however, the government asked the ACARD to report on whether the state should promote the IT industries in the United Kingdom. The government decided that it should, and in early 1981 Thatcher appointed a minister of state for information technology, Kenneth Baker. He immediately announced a "Year of Information Technology" in 1982. The DTI began that year to fund R&D through its Support for Innovation program (SFI). The importance attached by the DTI to IT is witnessed by the large share of its SFI spending that went to IT: In 1983–84, out of £322.53 million, £206.4 million supported IT projects.^[46] However, after a change of minister in September 1984, government support for IT moved from funding specific projects to more indirect activities. The

[43] Arnold and Guy, Parallel Convergence , 120.

[44] In all, Inmos had received £211 million in government funds, about twice the amount initially foreseen, and had not turned a profit. Still, the company was generally regarded as technologically advanced if limited by a restricted product range (Kevin Smith, "Inmos Forced to Get Off the Dole," 106).

[45] Hart, "British Industrial Policy," 154.

[46] Ibid., 121.

total DTI budget for industrial R&D in 1985–86 was £176.6 million, of which £97.7 million was earmarked for IT.^[47]

Britain's major new support plan for IT in the 1980s was a reaction to Japan's 5G program.^[48] British delegates in attendance at the conference in which Japan made its announcement returned gravely pessimistic about British prospects in the face of such an effort. The government quickly ordered a study of the United Kingdom's IT industries, to be carried out by an independent committee chaired by John Alvey, technical director of British Telecom. Reporting within months, the committee proposed a major program of cooperative R&D to be sponsored by the government. In addition the R&D would focus on areas that were of industrial interest but still some distance from production for the market; this vague category of research became known, especially within the ESPRIT program, as precompetitive.

Whitehall fretted about industrial intervention but in the end approved the Alvey Program after cutting back some of the recommendations. The Alvey Committee had requested 60 percent funding for industrial R&D; the government reduced the level to 50 percent, apparently a personal decision by Thatcher.^[49] The government also slimmed the proposed Alvey Directorate from thirty persons to about six, and the total government contribution from about £250 million to £200 million.^[50] The funds came from the Ministry of Defense (£40 million), the Department of Education (£50 million via the Science and Engineering Research Council), and the DTI (£110 million).^[51] Industry would contribute £150 million, bringing the total budget to £350 million (about $550 million at the time). This budget dwarfed that of any previous civil R&D program. As Patrick Jenkin, secretary of state for industry, told the House of Commons, "This is the first time in our history that we shall be embarking on a collaborative research project on anything like this scale."^[52]

[47] Ibid., 123.

[48] See Freeman, Technology Policy , 126.

[49] Alan Cane, "Britain Enters the Great Race," Financial Times , 9 May 1983, p. 12.

[50] "UK Fifth-Generation Computer Program Passes Final Hurdle," 76; Jason Crisp, "State Joins Industry in £350m Research on Super Computers," Financial Times , 29 April 1983, p. 1.

[51] Jason Crisp, "State Joins Industry . . . ," Financial Times , 29 April 1983, p. 1.

[52] Alan Cane, "Britain Enters the Great Race," Financial Times , 9 May 1983, p. 12.

Alvey followed the Japanese model, being designed to encourage collaborative research among U.K. companies, universities, and laboratories.^[53] As Jenkin observed, "Industry, academic researchers, and government will be coming together to achieve major advances which none could achieve on their own."^[54]

The technical aims of Alvey fell into four major areas: VLSI technology and design tools, software engineering, "intelligent knowledge-based systems" (AI), and person/machine interfaces (display technology, speech- and image-processing). Note again the prominence of advanced microelectronics: Alvey had asserted that the United Kingdom needed access to world state-of-the-art VLSI technology.^[55]

The British Alvey planners were able to coordinate the program with ESPRIT, which was emerging at the same time, so as to avoid excessive overlap between the technical objectives of the two programs. Such planning was possible in large part because the head of the Alvey Directorate, Brian Oakley, was a U.K. member of the ESPRIT Management Committee at the earliest stages.^[56]

Other Countries

The big three were not the only European countries with a growing interest in IT. Italy in the mid-1980s began a program for R&D in VLSI technologies that focused on developing prototype ICs, CAD for chips, and process technologies. SGS, the major Italian semiconductor firm (before its merger with the commercial semiconductor activities of Thomson), was slated to receive 13 billion lire in subsidies and 13 billion lire in soft loans.^[57] The Dutch government agreed to chip in DM 200 million (compared with West Germany's DM 300 million) for the Mega project involving Philips and Siemens; Philips took charge of the part of the project aiming at producing one-megabit SRAMs.

Even Belgium was getting into the act. The Belgians began construction of a state-of-the-art microelectronics research laboratory at Katholieke Universiteit Leuven in 1984. Called IMEC for Interuniversity MicroElectronic Center, the facility was designed to concentrate and optimize the electronics research going on at Belgium's

[53] Kevin Smith, "UK Pursues Fifth-Generation Computer," 101.

[54] Ibid.

[55] Ibid., 102.

[56] Interview 47.

[57] Howell et al., Microelectronics Race , 175.

three Flemish-speaking universities. IMEC's research would cover advanced processing, packaging, VLSI design methods, and optoelectronics. Sixty percent of the funding would come from the Belgian government and 40 percent from industrial research contracts. The government already had in place a program to train personnel from Belgian companies in advanced microelectronics.^[58] Two years later Belgium's first native IC maker came into being. The new enterprise, Mietec, was a joint venture between a development-capital company owned by Flanders state and Bell Telephone Manufacturing, ITT's Belgian subsidiary. The national government did not participate, as it had in IMEC. Mietec was given the best available fabrication equipment and the capacity to increase production rapidly.^[59]

Third, in each case the national program was a response to an acute (Japanese and American) threat to national capabilities in the IT sectors. Fourth, promoting collaboration among industry, universities, and research laboratories was a feature of national programs, most notably in Britain's Alvey but also in Germany and Italy. Finally, at the same time they were devising new nationalist responses to the high-tech challenge, each country also agreed to unprecedented degrees of international collaboration in IT. France became the most vigorous supporter among the big three of European

[58] Robert T. Gallagher, "Belgians Set Up Advanced IC Lab," 18–19.

[59] Robert T. Gallagher, "Now Belgium Has Its Own Player in the IC Game," 20–22.

IT collaboration. Even the reticent governments eventually agreed: Europe faced a crisis in the information technologies that were crucial to continued economic growth, and European collaboration could be one dimension of the response.

The Genesis of Esprit

Officials within the Commission were by the second half of the 1970s already convinced that Europe had to cooperate in IT in order not to lag ever further in that economically crucial area. But national governments during that decade were completely unreceptive to Commission appeals. At the end of the 1970s the mounting disillusionment with past policies coincided with the appointment of a new commissioner of industry in Brussels, Davignon. Davignon was the entrepreneurial IO official par excellence. He was able to take the arguments for cooperation developed within the Commission and sell them first to industry and then, with the help of the powerful business coalition, to national governments. Davignon was energetic, passionately committed to European collaboration in telematics, and politically savvy. But I do not want to overstress his role, important as it was.

The Commission initiative for ESPRIT bore fruit because of several factors: Purely national policies had failed; national policy-makers were in adaptive mode; the major industrial actors were for their own reasons interested in cooperation; and an activist Commission mobilized a political coalition behind collaboration.

Esprit Prehistory

In 1978, on the basis of the "Europe + 30" report, the Commission created a working group, Forecasting and Assessment in the Field of Science and Technology (FAST). Under FAST, research institutes in the EC competed for grants to forecast developments in three main areas: work and employment, the information society, and the biosociety. The resulting thirty-six reports were synthesized by the twelve-person FAST staff at Science, Research and Development.^[60] The FAST synthesis concluded that autonomous socioeconomic development in the EC would require "a coordinated approach

[60] Ros Herman, The European Scientific Community , 153.

in order to derive maximum value from the scientific, technological and industrial potential of the Community countries."^[61]

Regarding IT the group also noted the impact of technological change in making collaboration appear attractive to firms and governments: Because of rapid innovation and market change, it was difficult to foresee which technological options would prove essential in the long run. The report states:

No country in Europe has the capacity to cover such a spectrum of technological options, which is the only way to achieve a satisfactory degree of technological flexibility. . . . Hence there is a two-fold task for the Community: first, to achieve a certain collaboration and division of labour in order pursue jointly a wide range of technological options; and second, to carry out those specific programmes where there are evident advantages of scale.^[62]

The report singled out semiconductors as an area in which co-operative action was necessary to meet the strategies of the United States and Japan. The threat from Japan and the United States would become one of the Commission's primary refrains in pushing for ESPRIT. Indeed, one EC official was quoted in late 1981 as saying that the Japanese 5G initiative was an acute challenge for European firms, "many of whom are managing to stay profitable only by technical linkups with the Japanese. But this is dangerous for the future of European industry."^[63] In planning for an EC IT program, Commission officials also had in mind the Japanese model of precompetitive R&D carried out by industrial consortia.^[64]

Commission preparation for ESPRIT began among those officials working on FAST. According to the FAST report, Roland Hüber of the FAST team "initiated the conception and coordination" of a research project on EC needs and activities in "long lead-time R&D in information technology" (LLTRD-IT).^[65] This preparatory study, begun in 1980, lasted about eighteen months and led to the formation of the Joint European Planning Exercise in IT. The aim of this project was to "create the conditions for European collaboration in long lead time R&D . . . through the European Strategic Programme for R&D in Information Technologies (ESPRIT) and

[61] FAST, Eurofutures: the Challenges of Innovation , xi.

[62] Ibid., 74.

[63] Tom Manuel, "West Wary of Japan's Computer Plan," 104.

[64] Interview 22.

[65] FAST, Eurofutures , xii.

[contribute] to the definition of the necessary complementary measures to promote a competitive European IT industry in the 1990s."^[66]

The LLTRD-IT project commissioned studies that became ammunition for the Commission in persuading the IT companies that cooperation was necessary. For instance, the Battelle Memorial Institute and other consulting firms worked on a three-year LLTRD-IT study to identify long-term R&D needs in data-processing.^[67] The array of studies marshaled by the Commission made it knowledge-able enough about the technologies and the industries to be credible.

Davignon came to the industry post at the Commission in 1979 already possessed of a desire not to see repeated in IT the same "tragedy" that had overtaken the European steel industry, a sector with which he was familiar.^[68] Thus Commissioner Davignon became the patron of the small group working on IT at the Commission. Davignon protected them from "the bureaucrats who wanted to squash anything new or original."^[69] He converted the small group within FAST into the Information Technologies Task Force (ITTF), an independent body under the Commission not subject to any of the existing directorates. Davignon argued that industry had to make a major contribution to any EC program.^[70] In September 1979 he proposed that industry, governments, and the EC together work out a telematics strategy.^[71] His ability to rally industry would prove to be the key to getting ESPRIT started.

In the meantime the Commission was slowly educating the Council on the need for EC action in IT. By the fall of 1979 the Commission (that is, the ITTF) had drafted a document on telematics strategy to submit to the European Council at its meeting in Dublin at the end of November.^[72] The document was based on an "exhaustive

[66] Ibid., 74.

[67] Tom Manuel, "West Wary . . . ," 104.

[68] According to a senior scientific counselor to the Commission during the entire period who also participated in the founding of ESPRIT. Interview 15.

[69] Interview 22.

[70] Agence Europe , 24 March 1979, p. 10.

[71] Agence Europe , 29 September 1979, p. 12.

[72] The European Council consists of the heads of state or government of the member countries. Their meetings are also referred to as EC summits. The European Council should not be confused with the Council of Ministers (usually referred to simply as the Council), which in principle consists of the foreign ministers. The specialized councils have the same legal status as and are technically equivalent to the Council of Ministers; these group ministers with specific portfolios, such as research, agriculture, finance.

general survey" by the Commission of national ministries of industry, posts, and research, which indicated a desire for the Commission to outline an approach. The paper submitted to the Council was titled "La société européenne face aux technologies de l'information: pour une réponse communautaire." It outlined six points for an EC telematics strategy:

The Commission noted that telematics was a rapidly growing sector (over 15 percent per annum) and was "strategically important," having a "direct effect on the competitive position of numerous other branches of activity."^[73]

The Council responded by asking for specific proposals in the areas of microelectronics and telecommunications. The Commission proposed a microelectronics program the following July.^[74] The Commission had begun working in late 1978 with industry and research institutes to specify technical and performance goals.^[75] Davignon asserted in announcing the document that major national programs in France, Italy, Germany, and the United Kingdom would not be enough to narrow the gap between Europe and Japan or the United States.^[76] The Commission envisioned a budget of 100 million UAs over four years.^[77]

Time dragged on. Though the Council neglected the Commission proposals, they won approval in other EC bodies. Both the Economic

[73] Agence Europe , 7 November 1979, p. 8; 26 November 1979, p. 9; 28 November 1979, p. 5.

[74] Agence Europe , 18 July 1980, p. 6.

[75] Agence Europe , 29 November 1979, p. 16.

[76] Agence Europe , 18 July 1980, p. 6.

[77] Agence Europe , 11 September 1980, p. 5.

and Social Committee and the European Parliament endorsed the objectives and expressed concern that the financial means being discussed were too limited.^[78] The research ministers finally approved a program in November 1981, though at only 40 MECU. The Microelectronics Program would fund 30 to 50 percent of project costs. The areas chosen for funding all related to IC design and production equipment: step-and-repeat equipment, electron beam for direct writing on wafers, plasma etching and deposition, test equipment, and CAD for VLSI circuits. The program was to promote cooperation among makers and users of the technologies, and between industry and the universities.^[79] Over two years had elapsed since the Council responded to Commission prodding by requesting proposals for a microelectronics program.

Davignon and the Knights of the Roundtable

Within months of submitting his first proposals for an EC telematics strategy to the Council in late 1979, Davignon invited senior officials from the ten largest computer and telecommunications manufacturers in Europe to meet with him. The agenda for the February 1980 meeting consisted of discussing the possibility of a European telematics strategy.^[80] Those executives in attendance^[81] agreed in general with Davignon's assessment and accorded his notion of a European strategy a highly positive reception. Some of the industrialists felt that the Commission had underestimated the likely impact of vigorous Japanese development efforts on European and world markets by 1990. They expressed strong agreement that there was a need to support the development of basic microelectronics technologies and to develop common standards for data-processing and telecommunications. The companies agreed to submit specific observations on the Commission's proposals and to meet again in March.^[82]

[78] Agence Europe , 23 February 1981, p. 15; 1 May 1981, p. 13; 7 May 1981, p. 10; 9 May 1981, p. 7.

[79] Agence Europe , 10 October 1981, p. 14; 9 November 1981, p. 5; 14 November 1981, p. 13; the Council regulation is in the Official Journal , no. L/376, 30 December 1981.

[80] Agence Europe , 8 February 1980, p. 10.

[81] Including representatives of Siemens, Nixdorf, CGE, Thomson, CII-HB, ICL, Plessey, Olivetti, and Philips.

[82] Agence Europe , 11 February 1980, p. 10.

Davignon stepped up his campaign at the time the proposed Microelectronics Program was submitted to the Council (July 1980). In a briefing for industry executives, he argued that the choice for Europe was whether to be assistants to the United States and Japan in their strategies or to take an active role. The pointed out that Japan had caught up to the United States in IC technology and production in only four years, and did it with a lower government expenditure than existed in Europe as a whole. Davignon held out the goal of increasing Europe's share of semiconductor production from 6 percent to 12 percent by 1985. He also advocated support for the semiconductor-equipment industry, the opening of markets for telecommunications equipment, and common standards.^[83]

In late 1981 Davignon invited directors of the twelve largest European IT companies (the Twelve) to "roundtable" discussions on the future of IT in Europe.^[84] Davignon told the assembled executives that Europe faced formidable challenges from the United States and Japan, especially given the major new programs emerging in both countries. He argued that national programs were not the right answer, as Europe would end up reinventing the wheel many times. One participant (an independent senior advisor to the Commission on IT) paraphrased Davignon's message as follows: "If you remain alone, you will be subordinate to the Americans and the Japanese." Davignon told the Twelve that if they perceived a problem and if they wanted to cooperate among themselves, then he would find the funds for a major program. The company directors agreed with Davignon's analysis, but the general response was that they could not immediately commit themselves to a major cooperative program.^[85]

After discussions within firms and some initial contacts between them, the Twelve met in a second Roundtable. Davignon reaffirmed his promise and told the companies to propose a program. A steering committee of the Twelve was composed to do so. The steering committee heard reports from outside consultants and, after reaching

[83] Agence Europe , 11 September 1980, p. 5.

[84] The companies were Bull, Thomson, and CGE from France; Siemens, Nixdorf, and Allgemeine Elektrizitaets-Gesellschaft (AEG) from Germany; General Electric Company (GEC), ICL, and Plessey from the United Kingdom; Olivetti and Societa Finanziaria Telefonica (STET) from Italy; and Philips from the Netherlands. These companies became known, and I will refer to them hereafter, as the Twelve or the Roundtable companies.

[85] Interviews 15, 46.

agreement on the broad outlines of a strategy, constituted a number of expert panels to work out the specifics. The general strategy, defined by the steering committee working with Commission officials, included five areas as being appropriate for a European program: advanced microelectronics, advanced computing, software technologies, office automation, and integrated computer systems for industry. The expert panels were in place by the spring of 1982. Already the Commission was aiming at an initial phase of pilot projects to begin in January 1983, with EC funding of about 10–12 MECU.^[86]

At this stage, in fact, the Commission made its first formal submission to the Council, on 25 May 1982. The proposed program was titled ESPRIT: European Strategic Programme for Research and Development in Information Technology. The Commission document had the purpose of starting the process of discussion.^[87] It argued for a set of pilot projects, proposals for which would be submitted in the fall of 1982, with actual work to begin in January 1983. The proposed program met with generally favorable reactions from the national delegations. At the expert level the Research Council delegations agreed with the Commission analysis of the crisis facing European industry and on the need for swift action.^[88]

The Research Council gave a favorable reception to the ESPRIT proposal at the end of June 1982. The president of that meeting, in summarizing the conclusions of the debate, noted that IT was "crucial for the future of European industry." The Research Council also concluded that an EC program was urgent given "the state of competition in world markets" and Europe's having fallen behind.^[89]

Throughout this period the Twelve continued to meet, at the executive level in the steering committee (meeting about once a month) and at the technical level in the five working panels.^[90] Through these efforts the overall work plan and the pilot phase took shape. One participant, a senior executive from a Roundtable company, described

[86] Interview 15; Agence Europe , 26 May 1982, p. 13.

[87] CEC, Towards a European Strategic Programme for Research and Development in Information Technologies .

[88] Agence Europe , 28 June 1982, p. 7.

[89] CEC, Communication from the Commission to the Council: On Laying the Foundations for a European Strategic Programme of Research and Development in Information Technology: The Pilot Phase .

[90] Interview 53.

how at least one of the meetings worked. He was in charge of a one-day conference held at his corporate offices. Each company set up shop in a room, and the executives circulated among the rooms, working in pairs to specify projects on which they could cooperate. At the end of the day they had agreements on potential projects, which later became the pilot phase.^[91]

For the technical panels the Twelve sent engineers and technologists. Initially, about 100 people were involved.^[92] As companies other than the Twelve, as well as universities and research laboratories, became involved, this number rose. André Danzin, advisor to the Commission on IT, later said that creation of the work plan involved 400 technologists from industry, plus 150 from government and private laboratories and universities.^[93] Their discussions quickly became open and lively, leading some industry executives to worry about what they might be "giving away."^[94] Danzin said that being authorized to speak openly, the scientists "told all," like an "intellectual striptease."^[95]

The Pilot Phase

All this activity led by August 1982 to a proposal for the ESPRIT pilot phase. The Commission document declared that the overall objective of ESPRIT was "to provide the basic technologies which European industry needs to be competitive with that of Japan and the USA." The paper then outlined the fundamental features of the program. The most important was that ESPRIT had to be "aimed at pre-competitive technology."^[96] The term precompetitive is a vague one. Research can take place anywhere on a continuum from basic scientific research to product development and manufacture. The precompetitive condition was meant to keep ESPRIT away from work on actual products nearing the market. The program had to be precompetitive for several reasons. Some of the Commission's

[91] Interview 1.

[92] CEC, Communication from the Commission to the Council: On Laying the Foundations, 7.

[93] Speech at a seminar sponsored by the Centre d'Etudes Prospectives et d'Informations Internationales (CEPII), 26 June 1987, at the Commissariat Général du Plan, Paris.

[94] Interview 5.

[95] Danzin, CEPII speech.

[96] CEC, Communication from the Commission to the Council: On Laying the Foundations, 6.

reasons are somewhat abstract, dealing with larger questions of optimal research funding; other reasons are more pragmatic.

The Commission argued that in Europe long-term industrial research had been neglected for the sake of catching up with developments elsewhere. Thus, Europe had been unable to focus enough resources on long-term research to reach the critical mass of effort. "The result is a lack of punch in any major attempt to develop new technologies, which no one company or country can tackle alone." ESPRIT would remedy that defect. It would "reverse the current emphasis and focus it on longer-term technological research, establishing co-operation between industries across the European frontiers as well as co-operation between universities, research institutions and industry."^[97]

A more pragmatic reason for the emphasis on precompetitive R&D was that the Treaty of Rome did not provide for the Commission to run industrial policies. Anything that looked like the promotion of specific firms and industries in the marketplace would raise thorny political and constitutional questions. Another good reason for limiting the R&D to long-term areas was that companies would be reluctant to share technologies that they were preparing to introduce on the market. Precompetitive R&D did not directly threaten any commercial interests.

Still, the definition of the boundary between precompetitive and competitive R&D was intentionally left vague. One Commission official told me the best definition of precompetitive that he had heard: "Anything two companies are willing to do together."^[98] The Commission proposal made it clear that "nothing in the programme will prevent companies continuing to compete in the market with products and systems derived from the work."^[99] Indeed, one early member of the ITTF informed me that they had in mind the Japanese model of interfirm cooperation for precompetitive R&D, then competition on the market.

Still, the Commission has made efforts to give the term precompetitive a firm juridical basis, even though these developments followed the pilot-phase period I am discussing in this section. The legislation finally establishing ESPRIT defines precompetitive R&D

[97] Ibid., 6, 16, 17.

[98] Interview 26.

[99] CEC, Communication from the Commission to the Council: On Laying the Foundations , 17.

as that for which commercial possibilities remain five to ten years in the future, placing ESPRIT somewhere between fundamental research and applied industrial R&D.^[100]

Additionally, in late 1984 the Commission adopted Regulation 418/85, which exempted from EC competition (antitrust) rules certain kinds of R&D collaboration. The regulation is worded so as to apply clearly to R&D agreements under the Commission's programs like ESPRIT. For instance, the exemption applies if the R&D is performed within a defined program. The exemption covers joint R&D "on products or processes and joint exploitation of the results of that R&D." The rule seeks to preserve market competition by limiting the duration of the exemption in cases where a joint production and marketing agreement involves partners whose combined share of the EC market exceeds 20 percent. Even in the new regulation, however, concepts and wording remain fuzzy.^[101] In practice, the competition rules have been relaxed for firms developing products on the basis of Commission-sponsored research.^[102]

The proposal for the pilot phase set out several key themes that undergirded Commission efforts throughout the evolution of ESPRIT and even through RACE. First, the Commission highlighted the need to fortify Europe's ability to compete with the United States and Japan. In the section on microelectronics the Commission noted: "Both Japan and the United States have a significant lead in VLSI technology. Submicron work is required to stay in competition with them." One of the applications areas chosen by the Roundtable planners was office automation. Regarding that sector the Commission wrote: "Office automation is expected to become the largest single Information Technology market. In the USA both IBM and Xerox each spend more on this subject than European industry and academia combined, and the Japanese fifth generation computer concept is also aimed at these applications."^[103] This acute awareness of Japanese and American competition was shared by the national governments, as I showed in the first section of this chapter.

[100] Official Journal , no. L/365, 31 December 1985, 4.

[101] Alexis Jacquemin and Bernard Spinoit, Economic and Legal Aspects of Cooperative Research: A European View , 37–44.

[102] See, for example, William Dawkins, "Brussels Moves on 'Know-How' Accords," Financial Times , 3 September 1987, p. 4.

[103] CEC, Communication from the Commission to the Council: On Laying the Foundations , 10, 13; Annex II, 4.

A second key theme laid out in the pilot-phase proposal was, not surprisingly, that national programs would not suffice. "Given that there is a shortage of qualified manpower and other resources, which means that companies or governments individually cannot address all topics on a sufficient scale, the concentration of ESPRIT on precompetitive research will enable the necessary critical mass to be reached in key areas." At stake were crucial new technologies, "which no one company or country can tackle alone."^[104] As shown in the first section of this chapter, France at least was rapidly becoming convinced that national programs were insufficient, and the other major countries were also dissatisfied with the national-champion programs of the past.

Finally, a third important theme of the proposal was the need to overcome market fragmentation through common European standards. In fact, the Commission identified standardization as one of the two most important reasons for EC action, referring to "the relationship between R&D, standards and markets." In other words, the Commission linked R&D to standardization and market unification. "The importance of standards for the creation of a more homogeneous European home market in the IT sector is capital." The Commission also made specific mention of standards in the sections discussing the microelectronics, software, and office-automation components of the program.^[105] Standards would remain a consistent part of Commission telematics initiatives.

I have highlighted these three themes (international competition, the inadequacy of national programs alone, and the need for European standards) at some length because they dominated Commission efforts to sell the governments on collaboration.

In its proposal to the Council the Commission was careful to clarify that approving the one-year pilot phase "will involve no commitment to any further phases."^[106] Such a limited commitment was seen as unlikely to alarm member states.^[107] In practical managerial terms, a pilot phase, to begin in January 1983, would give the Commission and the participants experience in the "complex and difficult issues" of running such a program. The Twelve had outlined a total of fifteen pilot projects, which were described in

[104] Ibid., 6, 16.

[105] Ibid., 10, 11, 13, 15.

[106] Ibid., 19.

[107] Agence Europe , 10 September 1982, p. 11.

some detail in the proposal. The Commission added one more, for the development of a system that would allow ESPRIT participants to communicate quickly and efficiently. All sixteen pilot projects shared three main characteristics: They could be started immediately; they were useful in their own right; and, though lasting several years, they could be evaluated at the end of the pilot phase on the basis of "milestones."^[108] EEC funding for the initial projects would be 11.5 MECU, to be matched by the companies involved.^[109] The pilot phase was also meant to encourage the participation of small companies.^[110]

The pilot projects fell under the five main headings chosen by the Twelve. The first three were basic technologies, the other two were major applications areas: (1) microelectronics, (2) advanced information processing (AIP), (3) software, (4) office automation, and (5) CIM.

In Annex I to the proposal, the Commission laid out in some detail each of the pilot projects, including technological objectives, first-year review milestones, and subsequent review milestones. For instance, the first microelectronics project, "Advanced Interconnect for VLSI," specified the goal of developing a four-layer, metal interconnection technology at a level of one-micron features, capable of production utilizing CAD. The project would last four years, with two review milestones to have been achieved at the end of the first year. The first software project, called the "Portable Common Tool Environment," aimed at developing the hardware, software, and standard interfaces for software design. This project would allow European software designers to work with a common set of methods and tools. The project had nine first-year milestones and was the largest of the pilot-phase projects at 3.1 MECU.

Annex III of the proposal outlined the technical objectives of the full ESPRIT program. Each of the five technical areas was subdivided into a number of subprograms, the microelectronics area having the most specific subprogram objectives. The bulk of the microelectronics program was devoted to silicon VLSI technologies, but it also addressed compound semiconductor devices (like gallium arsenide chips), sensors, optical devices, flat-panel displays, and chip

[108] CEC, Communication from the Commission to the Council: On Laying the Foundations , 21.

[109] Ibid., 25.

[110] Ibid., 19.

packaging. The AIP and software areas had less precise objectives, while the office automation and CIM applications areas indicated rather general themes in their subprograms.

The Commission felt confident enough of a favorable reaction to ESPRIT by member governments that it published an advance notice for participation in the pilot phase in the Official Journal of 18 October 1982.^[111] The European Parliament voted in favor of ESPRIT and urged haste.^[112] In early November the Council passed a "common orientation,"^[113] with formal approval coming on 21 December 1982 at the requested EC funding level of 11.5 MECU.^[114]

The Commission published an official invitation for tenders in February 1983, drawing over 200 proposals from 600 companies and institutes. The total value of all the proposals reached 50 MECU.^[115] The Commission, with the help of a senior executive committee that included national IT officials, chose thirty-eight projects to receive funding. Not surprisingly, about 70 percent of the pilot-phase funds ended up going to the Twelve Roundtable companies that had prepared the program and were thus already in a position to submit relevant proposals.^[116]

Preparing the Main Phase

The success of the pilot phase eased the way for the main program. Both companies and governments responded positively to the trial run.^[117] The thirty-eight contracts had barely been signed when the Commission introduced a proposal for the ESPRIT main phase. The Commission submitted this proposal on 2 June 1983.^[118]

The document reiterated two of the Commission's main arguments. First, major programs in the United States and Japan threatened Europe's future in IT.^[119] The stakes were depicted as incalculably

[111] Agence Europe , 21 October 1982, p. 11.

[112] Agence Europe , 30 October 1982, p. 10.

[113] Agence Europe , 8 November 1982, p. 12.

[114] CEC, Proposal for a Council Decision Adopting the First ESPRIT , 6. The pilot phase was passed in the form of a Decision, which was published in the Official Journal , no. L/369, 29 December 1982, 37.

[115] Agence Europe , 5 April 1983, p. 7.

[116] Sherry Buchanan, "Small Firms Find a Niche in EC's ESPRIT Program," International Herald Tribune , 21 March 1984, p. 9.

[117] Paul Cheeseright, "How ESPRIT Will Help to Bridge the Gap," Financial Times , 28 March 1984, Survey, p. 8; David Fishlock, "A Shared Approach to Scientific Study," Financial Times , 12 December 1984, p. 12; Interviews.

[118] CEC, Proposal for a Council Decision Adopting the First ESPRIT .

[119] Ibid., 1–2.

high: "IT will be the driving force of economic growth for at least the rest of the century, and is therefore a major factor in economic affairs ."^[120] IT was the key to new products, processes, and services, and thus to increased exports and employment.^[121] Second, the Commission naturally repeated its argument that fragmented national programs were too small and overlapped too much.^[122] In addition, the Commission highlighted the support of industry:

Community industry has acknowledged that, in order to reverse the trend of increasing reliance on importing technology, only joint strategic long-term research planning and the concentration of resources through the definition and funding of technology goals of common interest on a Community scale, can have a good chance of redressing the situation .^[123]

Thus the Commission proposed the launching of the first five-year phase of an envisioned ten-year ESPRIT program.

The total budget would be 1,500 MECU, with half that amount coming from the EC. The sum proposed could appear almost negligible, argued the Commission, compared with the total EC industry investment in R&D in the sector, $5 billion per year. Some American companies by themselves poured $2 billion annually into R&D for IT. The Commission asserted that the ESPRIT program would bring the amount of resources devoted to long-term, precompetitive R&D into the same range as that found in the United States and Japan. Europe's main competitors were allocating between 5 and 10 percent of their R&D investment to precompetitive research; the 1,500 MECU from ESPRIT would comprise 6 percent of the total EC R&D effort in IT, raising the share of precompetitive R&D in Europe so that it was on a par with that in Japan and the United States.^[124]

In addition to the budget the Commission proposed a staff of 150 and a 7.7 percent share of the total budget for staff and administration.^[125] The proposal promised that the work program for ESPRIT's second five-year phase would be ready before the end of the fourth year of Phase I. The Commission also provided for an overall

[120] Ibid., Appendix, 4, emphasis in the original.

[121] Ibid., 7.

[122] Ibid., Appendix, 20.

[123] Ibid., 9, emphasis in the original.

[124] Ibid., 4; Appendix, 44–45.

[125] Ibid., Appendix, 62 and Financial Statement.

evaluation of the program at its mid-point, two and a half years into the work. The document projected a January 1984 launching.

Selling ESPRIT

The key to Commission success in winning approval for ESPRIT was that it did not try to do the job itself. It got industry to sell the program to the national governments. Quite simply, the Commission recruited industry, industry became committed to the cause, and industry took the cause to the national capitals. The launching of ESPRIT stems directly from the alliance struck by the Commission with a powerful transnational coalition, namely, the Twelve Roundtable companies. Of course, as I argued in the first section of this chapter, the governments were already in an adaptive mode, the failure of national-champion policies having led to a search for new solutions to the telematics crisis. This, in fact, was a necessary precondition. That the adaptive process resulted in collaboration is explicable only by the actions of an entrepreneurial IO allied with a potent industrial coalition.

Why were the Roundtable companies responsive to Davignon's invitation to put together a cooperative R&D program? I outlined in the previous chapter the major motives behind the rapid increase in interfirm alliances involving European companies in the early 1980s. These were the increasing costs of R&D, the blurring of boundaries between sectors (convergence), and the uncertainty induced by rapid technological and market change. The studies I cited in that chapter were based on extensive databases that recorded the motives expressed by businesses entering into partnerships. In other words, the motives I attribute to European firms that established technological alliances were not derived from some deductive schema but rather reflect the actual, empirical needs expressed by European telematics companies in the early 1980s. The Commission invitation presented a new way of meeting those needs.

In fact, there had been scattered talk about cooperation in European industry before Davignon's invitation.^[126] The Commission acted as a seed around which European collaboration could crystallize. The Roundtable companies agreed to talk, and their discussions

[126] According to an executive with a French Roundtable company, a British Roundtable executive, and a Commission official involved since the origin of the ITTF. Interviews 18, 28 and 60.

entailed a learning process. According to all the participants I spoke to, it was through the meetings of the steering committee and later the technical panels (as described previously) that the Twelve discovered strategic areas in which they could fruitfully collaborate. Those areas became the basis of the work programs. The loyalty of the Twelve was ensured by having them prepare the pilot projects and the work programs. In a real sense ESPRIT was their program.

Later, the companies were instrumental in selling the program to the governments. The numerous studies the Commission had paid for in working up to ESPRIT had to vanish. The Twelve had to present ESPRIT to their governments as something built by their efforts. If they had relied on Commission studies the Twelve would have carried no credibility in their respective capitals. "That was the key to selling the program to the governments."^[127] The substantial work invested by the Twelve in the steering committee and the technical panels thus gave ESPRIT an indispensable industrial legitimacy. In its proposal to the Council the Commission quoted extensively from a letter written jointly by the Twelve to Davignon. The letter noted Europe's miserable market position and argued that national programs even in the larger states would be insufficient. The prescription of this transnational industrial coalition was equally clear: "Unless a cooperative industrial programme of a sufficient magnitude can be mounted, most if not all of the current IT industry could disappear in a few years time. "^[128]

Industrialists and government officials in France, West Germany, and the United Kingdom confirmed to me that the Roundtable companies played an essential role in persuading their governments that ESPRIT was a good idea. One French executive said that Davignon used the industrialists to convince the governments. The same executive noted that the three French Roundtable firms (CGE, Thomson, and Bull) presented a unified, clear view to the Ministry of Industry, whose minister was persuaded and took up the cause of ESPRIT.^[129] Another French Roundtable industrialist, from a different company, said that he personally visited the Ministry of Research and Higher Education many times and lobbied the finance minister as well. His opinion was that the Twelve had helped immensely

[127] Interviews 23 and 28.

[128] CEC, Proposal for a Council Decision Adopting the First ESPRIT, 3, emphasis in original.

[129] Interview 18.

in winning approval for ESPRIT by lobbying their governments.^[130] An official concerned with ESPRIT in the French Ministry of Research and Higher Education informed me that the industrialists did help convince the French government of the program's merits. As he put it, "When the three largest firms in electronics and information technologies are all saying it's a good idea, the officials listen."^[131]

A similar phenomenon took place in Britain. Roughly the same people represented the British Roundtable companies (GEC, Plessey, and ICL) in ESPRIT as worked with the Alvey Committee. They therefore had a direct input into British policy-making circles. The United Kingdom's big three companies were definitely in favor of ESPRIT.^[132] In addition, the senior civil servant placed in charge of the Alvey Directorate, Brian Oakley, also sat on the ESPRIT Management Committee from the very early stages. This arrangement permitted close coordination of the work of the EC program with that of Alvey. Oakley also became an advocate of ESPRIT within the British government. On the industry side, Derek Roberts, technical director of GEC, became persuaded at an early date that European collaboration was important.^[133] As early as December 1982 Roberts was quoted as saying, apropos of ESPRIT, "There is simply no alternative to co-operation between previous rivals, and between industry itself and the academic world."^[134] British industrialists became even more enthusiastic as the pilot phase progressed.

Finally, the German Roundtable companies also weighed in favorably with their government on behalf of ESPRIT. The BMFT held discussions with industry and industry associations regarding ESPRIT. German companies were in favor of the program. One BMFT official speculated that part of the industrial support for ESPRIT in Germany might have been due to reductions in BMFT industrial research funds at that time.^[135]

In short, the picture painted by Commission officials, Roundtable executives, and national officials confirms the central role played by the Roundtable companies in winning national approval for ESPRIT.

[130] Interview 23.

[131] Interview 48.

[132] Interview 47.

[133] Interview 60.

[134] Richard Brooks, "ESPRIT Puts d'accord into Europe," Sunday Times, 19 December 1982, p. 44.

[135] Interview 4.

Of course, final agreement did not come without political battles and compromises. In the rest of this section I describe the principal struggles. Note, however, that the disagreements centered on details of the structure and administration of ESPRIT, not on the fundamental merits of the program itself. The sales pitch of the Roundtable companies had reached national governments searching for new approaches to IT policy.

By summer 1983 it appeared that all ten member governments were in complete agreement on the importance and the content of the program. But signs of trouble were already appearing: Germany wanted to wait and see the results of the Stuttgart Mandate on the reform of EC finances. The Germans declared themselves willing to approve the new expenditures for ESPRIT only if they could be financed within the current budget allocations.^[136] This would prove to be the main sticking point.

The Research Council meeting of 26 October 1983 brought forth a surprise however. French Energy Minister Jean Auroux, sitting in for Minister of Industry and Research Fabius, blocked any decision on ESPRIT. He rejected the funding level proposed by the Commission (750 MECU) and suggested a limit of 400 MECU. Auroux also asked for a provision limiting EC funding to genuinely European companies. His stand provoked puzzlement and confusion among the other delegations. For one reason, France had just the month before circulated its memorandum on increased technological cooperation in Europe. Furthermore, at the same time Auroux was applying the brakes to ESPRIT, French Minister of European Affairs Andr%e Chandernagor was advocating in the Groupe Unique de Pr%eparation the development of an EC industrial policy to improve the competitiveness of European enterprises.

At the Research Council meeting seven delegations expressed agreement on the proposed level of finance. West Germany and the United Kingdom did not oppose the 750 MECU budget but withheld their approval pending the results of the Athens Summit scheduled for December, at which the overall EC budget problems would be addressed. Some of the smaller countries, especially Belgium, expressed the fear that they would not benefit from the results of the program because they had no giant firms like the Roundtable companies.^[137]

[136] Agence Europe, 24 October 1983, p. 7.

The ministers scheduled a special session for early November to resolve these differences.

At the Research Council of 5 November, Fabius at the outset lifted all the French objections. What explains the rapid turnaround? The problem stemmed from a French bureaucratic glitch. Apparently, when Auroux was detailed to substitute for Fabius, Auroux was briefed by functionaries who were not up-to-date on the government's thinking about ESPRIT and were unenthusiastic about the EC plan. Remember that the fall of 1983 was the period when the Mitterrand government was shifting to its pro-European stance. Auroux therefore thought he was representing the French position, but that position was in the process of being changed.^[138] In the same meeting the Ten agreed on a budget of 700 MECU, though Germany and Britain continued to reserve final assent until after the Athens Summit.

Other organizational details were hammered out. First, the annual work program would have to be submitted to the Council, where it would be approved by qualified majority.^[139] Second, Type A projects would receive 75 percent of the total budget, and Type B projects were guaranteed 25 percent of the total. Type A projects were defined as those with budgets of 10 MECU or more, later reduced to 5 MECU;^[140] Type B projects would have no minimum. This was a concession to the small countries, led by Belgium, who wanted some assurances that small projects would not be overlooked.

Third, in special cases (namely, for universities and small- and medium-sized companies), the limit of 50 percent EC funding could be exceeded if agreed by qualified majority in the Management

[137] Europolitique, 29 October 1983; 9 November 1983.

[138] Interview 15.

[139] Qualified majority is a voting procedure in the Council in which national votes are weighted and passage of a measure requires more than a simple majority of the weighted votes. Prior to 1986 the voting of the ten member states was weighted as follows: ten votes each for France, Germany, Italy, and the United Kingdom; five each for Belgium, Greece, and the Netherlands; three each for Denmark and Ireland; and a generous two for Luxembourg. Out of the total of sixty-three votes, forty-three were needed to pass a measure that required a qualified majority. When Spain and Portugal joined the EC, they received eight and five votes respectively; the qualified majority changed to fifty-four out of seventy-six votes. Qualified-majority voting in bodies other than the Council (like the ESPRIT Management Committee and the RACE Management Committee) follows the same format.

[140] ESPRIT Review Board, Mid-Term Review of ESPRIT, A2.

Committee. Fourth, Type A projects had to be approved by qualified majority in the Management Committee. If the qualified majority could not be achieved but there was a clear majority opinion, the Commission could make the final decision itself. The Germans expressed a reservation on this point, fearing too great latitude was being given the Commission. Fifth, the share for personnel and administration in the ESPRIT budget was fixed at 4.5 percent, which involved a concession from the Germans, who favored a 1 percent limit.

Finally, the French relaxed their demand that ESPRIT funds go only to genuinely European firms. They agreed to a statement in one of the "considering" clauses that "those enterprises best placed to achieve the strategic objectives of the ESPRIT program will be recipients of assistance." Because the main ESPRIT objective was to promote European industry, the clause indicated that European firms would receive the grants.

The November meeting concluded with a political agreement on ESPRIT, with final approval awaiting the outcome of the Athens Summit. Commissioner Davignon expressed the view that if the Athens meeting failed to achieve reforms that would free funds for ESPRIT from other areas, the Commission would take money for ESPRIT from its other R&D programs.^[141] In the meantime, both the Economic and Social Committee (unanimously, with one abstention) and the European Parliament approved the ESPRIT program.^[142]

The Athens Summit failed to resolve the EC budget crisis. ESPRIT thus remained blocked, even though all the members approved of the technical content and the management provisions.^[143] The Germans indicated an area of possible compromise: They would be satisfied with guarantees that the money for ESPRIT could be found in existing research appropriations.^[144] The British did not express any firm position. In fact, according to a Financial Times story the British were in a holding pattern behind the Germans because of divisions within Whitehall.^[145] This speculation was confirmed to me

[141] Europolitique, 9 November 1983; Agence Europe, 7 November 1983, p. 5.

[142] Agence Europe, 1 October 1983, p. 15; 15 October 1983, p. 11.

[143] Europolitique, 14 December 1983.

[144] Agence Europe, 14 December 1983, p. 10.

[145] Paul Cheeseright, "EEC Nears Information Technology Agreement," Financial Times, 27 February 1984.

by a well-placed British civil servant. The cabinet was divided, with different ministries expressing contradictory positions in public.^[146] Clearly, Thatcher herself had not yet decided in favor of ESPRIT.

At this point serious wooing of the Germans and British began. The French, who had the presidency of the Council for the first six months of the new year, initiated a series of high-level contacts with Bonn in order to reach an agreement. The Commission also pursued discussions with the holdout governments. Davignon himself made an appeal to the Council of Ministers meeting in late January 1984. He argued that the program could not await the next European Council, which was slated for March 19.^[147] He warned the British and West Germans that their footdragging jeopardized Europe's efforts to catch up with the United States and Japan.^[148] Eventually, Davignon traveled to London, where he met with Thatcher, and to Bonn.^[149]

The Commission proceeded by publishing on 30 December 1983 an advance notice for participation in the ESPRIT main phase, so that interested parties would be informed about application procedures once the program was finally launched. Indeed, the work program had already been unanimously approved.^[150]

At the Council meeting of 23 January the ministers agreed that a final decision on ESPRIT should be reached in the 28 February meeting of the Research Council. Contacts between the Commission and Bonn had produced a compromise solution. Germany would agree to the funding level of 700 MECU agreed on in December, provided that Commission R&D spending not exceed 600 MECU per year for the next two years. In addition, for that fixed amount, ESPRIT, the Joint European Torus, and the Joint Research Centers should have priority. This compromise would leave just enough money to keep other existing research programs going but limit some new projects. The smaller states objected to such a move because they derived benefit from the other programs but considered themselves less able to capture benefits from the three large programs.^[151]

[146] Interview 47; see Agence Europe, 10 February 1984, p. 7.

[147] Agence Europe, 19 January 1984, p. 13.

[148] Giles Merritt, "ESPRIT under Budget Cloud," Sunday Times, 22 January 1984, p. 25.

[149] Jon Turney, "Cash Agreed for Europe's New Venture," Times Higher Education Supplement, 2 March 1984, p. 1.

[150] Agence Europe, 19 January 1984, p. 13; 5 January 1984, p. 7; 25 January 1984, p. 7.

[151] Agence Europe, 25 January 1984, p. 7.

Even so, the smaller states never argued against the ESPRIT program or its objectives. They did seek measures to assure that their organizations would be allowed full participation, as described previously regarding the Type A—Type B formula and the possibility of greater than 50 percent funding for some Type B projects.

By that point impatience was starting to mount. The European Parliament Energy and Research Committee passed an emergency resolution urging the Council to pass ESPRIT by 28 February at the latest.^[152] There were vague rumors that the large IT companies themselves were getting fed up with the delays and had talked of setting up a program of their own, outside the EC. A program run by the major firms would naturally have reduced the role of small and medium enterprises (SMEs). In fact, the large firms were at least a little wary of including the SMEs, fearing that they would gain access to technology from the Twelve without contributing much in return.^[153]

The Commission assented to the German conditions, promising to manage R&D expenditures so that ESPRIT would be run within existing appropriations in 1984 and 1985.^[154] After that the Commission's own resources were scheduled to increase. This assurance satisfied the Germans and the British, and ESPRIT was finally released from its budgetary deadlock by the research ministers on 28 February 1984. Interestingly, the Council approved an ESPRIT budget of 750 MECU, the level originally requested by the Commission and higher than the compromise level of 700 MECU agreed to the previous December.^[155]

Having prepared for this moment, the ITTF immediately published a call for proposals. By the deadline, it was inundated with 441 submissions. ESPRIT was taking off.

Esprit in Action

This section examines ESPRIT from two different perspectives. First, how does it work, in organizational and procedural terms? Second, has ESPRIT been "successful"? The struggle surrounding the

[152] Agence Europe, 4 February 1984, p. 12.

[153] Sherry Buchanan, "Small Firms Find a Niche in EC's ESPRIT Program," International Herald Tribune, 21 March, 1984, p. 9.

[154] Agence Europe, 25 February 1984, p. 7.

[155] Europolitique, 29 February 1984; Agence Europe, 1 March 1984, p. 7.

launching of the second phase reveals much about the perceptions of ESPRIT on the part of participants, research ministries, and national governments.

The Mechanics of ESPRIT

The gears and levers to make ESPRIT work were already agreed on before the program received definitive funding. The work program projected that the resources (measured in man-years) would be allocated among the sectors as shown in Table 7.2. The R&D work would be precompetitive, and projects would as a rule have to include companies, universities, or laboratories from at least two different EC countries.

The Commission, through the ITTF, had overall responsibility for managing the program. Assisting the ITTF were ESPRIT's three other administrative bodies, the ESPRIT Management Committee (EMC), the ESPRIT Advisory Board (EAB), and the Steering Committee. The EMC represents the member governments, each state placing two members on it. As the states did not want to leave project selection entirely to the Commission, the EMC must give final approval to all projects and must specifically approve (by qualified majority) each project worth 5 MECU or more (Type A projects). The EMC must also approve any departures from official ESPRIT

rules by simple majority—as in exceeding the 50 percent funding limit, for example. Voting in the EMC is weighted.^[156]

The EAB is the participants' body. Its role is purely consultative, advising the Commission on technical questions. Its membership includes sixteen persons not on a representative basis but serving à titre personnel . About half the members come from Roundtable companies, the other half from small companies, universities, and laboratories. As one British Roundtable executive explained it, expanding the EAB beyond the Twelve was politically necessary; in order to win approval for the program, it could not appear to be merely a subvention for Europe's largest telematics houses.^[157] The Twelve retain a central role through the ESPRIT Steering Committee (separate from both the EMC and the EAB), whose function is to outline the work plans every year. The work plans are then fleshed out by technical panels involving the Twelve, smaller companies, academics, and other researchers.

After the work program is approved and a call for proposals is published in the Official Journal, the ITTF begins the job of evaluating the proposals and making selections.^[158] For this purpose the Commission retains technologists from industry, universities, and laboratories for a period of about three to six weeks. They serve personally, not as representatives of their organizations. Committees of experts evaluate the first part of the proposals, which at this stage are anonymous; the proposals lay out the technical objectives, show how they fit the ESPRIT program, describe the resources needed, and specify milestones (for checking progress) and deliverables. To fail at this stage, a proposal must be rejected by a majority of the experts on the panel. The committees then submit short lists with justifications of their choices.

The next stage is a study of the short lists by review committees each headed by a division head from the ITTF and including ITTF personnel and technical experts. At this point the proposals are no longer anonymous, as the committees must assess the experience and technical capabilities of each partner. The review also judges the statements of products and markets that might be exploited on

[156] Arthur F. P. Wassenberg, Strategies and Tactics of European Industrial Policy-Making, 11.

[157] Interview 60.

[158] In the following paragraphs I rely on interviews with Commission officials and on ESPRIT documents, especially CEC, ESPRIT: 1987 Information Package .

the basis of the research. Each division head submits a short list. The final choice of projects is made by the head of the ITTF in conjunction with the division heads. Their choices then go before the EMC.

In practice the percentage of projects approved by the ITTF and rejected by the EMC is very low. The EMC does have the prerogative to suggest changes in projects. It is in the EMC that countries press for a good share of the work. Naturally, sometimes countries strike bargains to include one more partner on a project or to combine two projects. Nevertheless, according to one member of the EMC project selection has not been highly political and juste retour concerns have not posed a problem.^[159] There is no formula to assure juste retour, and the Commission does not publish any figures on national shares. Significantly, organizations that have been involved in ESPRIT have not expressed any complaints about political meddling for juste retour . In a survey of participants conducted for the ESPRIT mid-term review, respondents did not complain of political interference in the selection process. Some did, however, complain that merging projects caused serious problems. But merging projects is not necessarily the outcome of political bargaining; it is also due to the desire to fund as many participants as possible on a limited budget.^[160]

Once projects receive approval, the Commission signs a contract with each consortium partner. When the contract is signed, the consortium receives an advance of 30 percent of the Commission contribution. The partners are jointly and individually liable for fulfillment of the terms of the contract. One of the partners serves as the consortium leader (or manager) and acts as liaison for the project with the Commission. Within the ITTF a number of project managers keep tabs on an assigned set of projects. Each project manager must submit a monthly report on activities and progress. A larger evaluation involving outside experts retained by the Commission takes place every six months at the Commission, not at the work sites. Further disbursal of EC funds depends on satisfactory evaluations.

Regarding intellectual and industrial property rights, contractors in a project have equal ownership and equal rights to exploit the

[159] Interview 47.

[160] ESPRIT Review Board, Mid-Term Review, 17–19.

results, with specifics to be agreed on among themselves. Participants in different ESPRIT projects have privileged access to the results of another project if those results would enhance their ESPRIT work. Other EC companies can acquire the rights to knowledge generated in ESPRIT on a regular commercial basis. The Commission retains the right to require an ESPRIT participant to grant licenses if it does not wish to exploit results itself.^[161]

Beginning with the pilot phase EC organizations responded enthusiastically to these arrangements. Out of 145 proposals submitted for the pilot phase, 36 were selected. As noted, the first call for proposals for Phase I attracted 441 submissions, including pilot-phase projects applying for continuation under the full program. A total of 110 projects received funding, including 23 carried over from the pilot phase. The total funding commitment for the 1984 call for proposals, after some changes in projects, settled at 372 MECU.^[162] Of this first batch, exactly half were Type A projects and half Type B. The May 1985 call for proposals drew 389 responses, of which 95 were approved for a total of 279.7 MECU. Thus, by June 1986, after some settling out among the projects, 201 were underway with a total allocation of 677.7 MECU, or about 90 percent of the total 750 MECU budgeted for Phase I.^[163]

A limited call for proposals followed in the fall of 1985, restricted to the software area and totaling 18.9 MECU. (The Commission had noted a less-than-expected level of participation in the software segment.) Out of over forty proposals, eleven received the stamp of approval.^[164] Finally, the Commission published in April 1986 a final call for proposals for Phase I, with only 62 MECU left in the coffer. The office-automation area was excluded from this final round, the Commission judging that the program was already meeting its objectives in that sector.^[165] Twenty projects were retained from this final round.^[166] After the usual settling down and some attrition, the total number of projects in ESPRIT Phase I reached

[161] CEC, Proposal for a Council Decision Adopting the First ESPRIT .

[162] ESPRIT Review Board, Mid-Term Review, A2, n. 1.

[163] CEC, ESPRIT: The First Phase: Progress and Results, 7.

[164] Europolitique, 26 February 1986.

[165] Europolitique, 19 April 1986. The totals from all the bidding rounds appear to total over 750 MECU because the termination or scaling down of some projects released a small amount of funds that could be added to the 1986 call.

[166] Europolitique, 19 October 1986.

226. The number of projects in which each country was represented was as follows; there are no surprises here:^[167]

A few numbers will sketch in the nature of the participation in ESPRIT I. By type of organization, participation was as follows: 62 small companies (fewer than 50 employees), 84 medium companies (50–500 employees), 181 large companies (over 500 employees), and 199 universities and research institutions.^[168] The average number of partners per consortium at the end of 1986 was 5.1, though the average number of industrial partners per project was 3.5. This difference was due to the broad participation of nonindustrial organizations (universities and research institutes): Out of 201 projects as of December 1986, 150 included nonindustrial partners. SMEs (those with fewer than 500 employees) participated in 50.6 percent of Type A projects and 61.4 percent of Type B projects. They did over one-quarter of the work on 60 percent of the projects in which they participated.^[169] Overall, SMEs had a role in 65 percent of the projects and received 14 percent of the funding. The Roundtable companies participated in 70 percent of the projects and received

[167] CEC, Directorate General XIII, ESPRIT: The Project Synopses, vols. 1–8.

[168] CEC, ESPRIT: 1988 Annual Report, 3.

[169] CEC, ESPRIT: The First Phase, 78–85 and Tables 7–9.

50 percent of the funding, though their share has been falling over time.^[170]

The average number of partners per project was inflated by a few extremely large projects whose primary aim was standardization. When standards are at issue, not being included can be costly. And, indeed, for effective standards, the more the industrial participants the better. The largest project (in number of partners) was "AMICE: A European Computer Integrated Manufacturing Architecture," which had twenty partners. A number of projects had nine or ten members, like the office-automation project "Standardisation of Integrated LAN Service and Service Access Protocols."

Assessing the Impact of ESPRIT

An evaluation of the technological and industrial results of the total ESPRIT program is impossible at this point. In any case, the technical impact of ESPRIT is peripheral to the arguments of this study, which focuses on the politics of international cooperation. Nevertheless, we can look at the three evaluations of the ESPRIT ^I program undertaken by the Commission: one at the mid-point as promised in its proposal, one near the end of Phase ^I , and one after its completion. I shall summarize the main conclusions of the reports as succinctly as possible, then analyze the politics surrounding ESPRIT Phase II. As noted, the launching of ESPRIT II reveals much about the perceptions of the program on the part of the Commission, industry, and governments.

Judging ESPRIT I

Some observers have claimed that the stated goals of ESPRIT have changed over time. For example, Arnold and Guy argue that technological advance has been replaced as the chief objective, with European collaboration becoming an end in itself.^[171] This claim is only partially true. Certainly the Commission has shifted its emphasis to highlight ESPRIT's success in expanding IT collaboration in Europe and to tout that as a primary end. To be fair, however, the Commission could not possibly point to ESPRIT's success in narrowing technology gaps simply because it is far too soon to detect such changes. Both technological advance and the promotion of cooperation now appear as the prime goals of ESPRIT,

[170] ESPRIT Review Board, The Review of ESPRIT, 1984–1988, 17.

[171] See Arnold and Guy, Parallel Convergence, chap. 5.

though sometimes one appears first and sometimes the other. The mid-term review, for instance, claimed that its charge was to examine how well the program was meeting its objectives; it listed the first objective as "the promotion of European industrial cooperation" and the second as "the provision of basic technologies." The third objective was to "pave the way for standards."^[172] However, when the progress and results report reiterated the three main objectives of ESPRIT, it reversed the order of the first two and placed technology first.^[173] The report of the Third ESPRIT Conference incorporated both orderings: In the introduction it placed "basic technologies" first; but in his speech the Commission official in charge of ESPRIT, Jean-Marie Cadiou, mentioned industrial cooperation first.^[174]

The Review Board appointed by the Commission for its midterm evaluation was composed of three outside experts: A. E. Pannenborg, chairman, formerly of Philips; André Danzin, formerly of Thomson and an advisor to the Commission on IT; and H. J. Warnecke of the Fraunhofer Institut/IPA. The secretariat to the Review Board also consisted of three outsiders: V. C. Grandis of Tecnetra, P. A. Monseu of Comase, and P. A. Walker of Mackintosh International. Their evaluation was based on personal interviews or group meetings with representatives of 131 organizations (participants and governments) and a questionnaire sent to 477 participants, of which 238 responded.^[175]

The first conclusion of the review was that "there is now unanimous agreement as to the initial success of ESPRIT, particularly with respect to the promotion of trans-European cooperation." Respondents noted that cooperation added overhead costs (travel, coordination) of about 10 to 20 percent of project costs, but that nevertheless cooperative R&D was more beneficial than isolated efforts. There was "unanimous agreement" that ESPRIT had permitted the size, scope, and goals of research projects to be increased. The resources applied to projects, both manpower and diversity of skills, were greater than they could otherwise be. ESPRIT was also leading to increased confidence and cohesion within the European IT industry. On the whole, respondents thought that the

[172] ESPRIT Review Board, Mid-Term Review, 5.

[173] CEC, ESPRIT: The First Phase, 7.

[174] CEC, 3rd ESPRIT Conference: Building Momentum, 3, 37.

[175] ESPRIT Review Board, Mid-Term Review, Executive Summary, i.

ITTF management of the program was exceptionally proficient. Several corporate officers mentioned that ESPRIT was at least as efficiently run as national programs with which they had experience. With regard to national programs, participants felt that they were complementary to ESPRIT, with no severe problems of competition.^[176]

The main complaints regarding the program were that the application process was costly; there were sometimes delays in receiving contracts and payments; the merging of projects caused serious problems; the monthly reports were an onerous task; and the ESPRIT Electronic Information Exchange System was not at all adequate. National governments, then, would have some reason to accuse ESPRIT of being excessively bureaucratic, a rationale on the part of some states for supporting EUREKA (as will be seen).^[177]

The questionnaire elicited a number of illuminating responses. I will highlight some of them in summary form:

[176] Ibid., 33.

[177] Ibid., 43–49.

[178] Ibid., Annex C.

The Review Board also made recommendations concerning the future of the program. Many of the suggestions were distilled from the comments of participants. These included:

Most of these suggestions were incorporated in the plans for ESPRIT II.

The progress and results report was based on 200 reports from the technical evaluators who conducted the six-month progress evaluations. It attempted to present to the Council "concrete technology results." The report noted that almost 90 percent of the projects were on schedule as of mid-1986, with fourteen projects three to six months behind schedule and only eight projects lagging more than six months. It summarized the positive role of ESPRIT as concentrating scarce resources, both personnel and finance; allowing the pursuit of more options; accelerating research; and taking advantage of other EC policies such as standardization and trade policy.^[180] The bulk of the document consisted of numerous specific examples of technical milestones achieved in each of the five technical areas. I will summarize some of them for purposes of illustration.

In microelectronics a project produced a design for a 10K-gate array with two-micron line widths and 200 picosecond delay. Siemens

[179] Ibid., 50–57.

[180] CEC, ESPRIT: The First Phase, 2–8.

was building a 100 MECU production line for the arrays. Another project developed methods for multilayer metal interconnections in VLSI circuits, a technique exploited by Plessey on a 200,000-element complementary-metal-oxide-on-silicon (CMOS) chip. One major development in the software area was the prototype for a computerized software production system, the Portable Common Tool Environment, that could be run on computers of different makes. One company started marketing a commercial application of the system.^[181]

In AIP, ESPRIT projects produced enhanced versions of the logic-programing languages (computer languages that run on logical rules rather than mathematical operations) used in AI work. A Belgian institute released a compiler for the Prolog language that produces faster executing code than any other in the world. Work in the office-automation area led to a technical standard for multimedia (voice, text, and image) office documents. The Herode project developed what became the European and world (International Standards Organization, or ISO) standards for office documents. The CIM projects also produced the bases for standards in manufacturing systems, such as factory communications and robot integration. Numerous large and small companies developed prototypes and marketable products from ESPRIT CIM projects.^[182]

The 1988 annual report of ESPRIT reviewed further technical achievements obtained from projects. Without going into details, the report claimed that by the end of 1988, 130 of the 226 projects had produced 166 "major results." This output included forty-two direct contributions to products or services already commercially available, and forty-eight to products or services under development but not yet on the market. Thirty technical results had contributed to international standards.^[183]

Like the mid-term review, the final review of ESPRIT I was carried out by a team of outside experts, again headed by Pannenborg. The Review Board based its conclusions on 210 interviews with participants, 949 questionnaires returned by participants, interviews with external consultants, and ESPRIT documents. Their report praised ESPRIT for substantially achieving its major goals but

[181] Ibid., 25.

[182] Ibid., 68–74.

[183] CEC, ESPRIT: 1988 Annual Report, 3.

also frankly discussed the program's shortcomings. The Review Board concluded that "in the vast majority of projects trans-European cooperation has been a success and resulted in significant benefits for the participants." The report lauded ESPRIT for improving Europe's technology base, in both material and human resources, and for improving links between industry and the universities. On the whole, program management was "satisfactory and smooth."

Responses to the survey questionnaire showed that over 60 percent of participants thought that ESPRIT I objectives had been "well" or "adequately met" in AIP and CIM, but only about 40 percent responded similarly for the other areas (software, microelectronics, office systems). An additional 25–45 percent answered that ESPRIT objectives had been "partially" met, leaving a small portion (5–20 percent) who thought they had been "poorly" met. Ninety percent of respondents thought that collaboration with partners from other countries had worked "well" or "adequately." The most commonly reported benefit from ESPRIT work was "enhanced know-how" (70 percent of respondents), followed by "more ambitious R&D goals" (60 percent), "new products" (47 percent), "improved development techniques" (45 percent), and "improved products" (35 percent).

The major criticisms of the program were as follows:

The Review Board noted that many of the criticisms had been taken into account while preparing ESPRIT II, and the second phase would therefore include a number of improvements.^[184]

In my own interviews, British, French, and German executives from Roundtable companies all expressed general satisfaction with ESPRIT. The major benefit they cited was that the program gave them access to more results that would have normally been possible with their limited R&D personnel. Many company officials cited a new sense of self-confidence and community in the IT industries across Europe.^[185] As I will argue later in this chapter and in succeeding chapters, the momentum built by ESPRIT Phase I propelled the EC forward into ESPRIT II and RACE and in other directions outside the EC framework. The new sense of "Euro-optimism" in IT thus preceded and foreshadowed the mounting enthusiasm tied to the 1992 movement. ESPRIT showed that Europe could act coherently to promote joint gains in crucial areas.

Perceptions of ESPRIT I and the Battle over Phase II

The growing reservoir of support for ESPRIT in industry and among research ministries did not mean automatic approval for the second phase. It was not sufficient because ESPRIT II became tied up in the debate over the Commission's proposed Framework Programme for R&D, of which ESPRIT was the largest element. The Framework Programme embraces all the R&D activities run by the Commission, including nuclear fusion, energy, environmental protection, and medicine. The governments of two states in particular—Germany and the United Kingdom—sought reductions in the proposed budget for, and increased control over, the content of the Framework Programme.

When it became clear that ESPRIT I funds would be fully allocated by early 1986, the Commission announced its intention to press for an accelerated launching of ESPRIT II. Officials in Brussels declared that in order to preserve the momentum achieved under the first phase, the second should begin in 1987, two years earlier than originally planned.^[186] Meetings with the Roundtable companies

[184] ESPRIT Review Board, Review of ESPRIT.

[185] Interviews 1, 5, 18, 37, 45, 46, 53, 54 and 60.

[186] Peter Marsh, "EEC Bid to Boost Spending on ESPRIT," Financial Times , 18 November 1985, p. 3.

began in order to devise a plan for Phase II. The industry group recommended that the level of work be increased to 30,000 man-years, about triple the amount that the first phase would reach. This expansion would also require about triple the level of funding, or about 2.2 BECU from the EC.^[187]

A first communication to the Council regarding ESPRIT II was completed in May 1986. The document stressed familiar themes. First, the economic importance of IT: "IT is not only a major industry in its own right but contributes significantly to the competitive status of most economic activities." Second, the Commission pointed out Europe's declining position in IT. It cited studies projecting a decline in the share of European producers in world IT markets from 23 percent in 1980 to 21 percent in 1990, by which time Europe would provide 30 percent of world demand, the largest single market.^[188] Third, the document repeatedly highlighted Japanese and American programs, including SDI, that threatened to increase the technological lead of these countries over Europe.^[189]

The Commission proposed a few major changes for ESPRIT II, though the basic attributes of the program would remain unchanged (precompetitive research, the system of proposals and contracts). I will describe these as quickly as possible. One big change was that ESPRIT II would be proposed within the context of the Framework Programme on R&D. Pursuant to the Single European Act (Luxembourg, December 1985) the Commission was to encourage industrial R&D in the EC. Unanimous approval of the Council was needed only for the Framework Programme as a whole; the specific R&D programs constituting it would be decided by majority vote. Thus, the Commission in 1986 was preparing the Framework Programme to cover the five years from 1987 to 1991. One highly placed Commission official within the ITTF told me that inclusion of ESPRIT II within the Framework Programme was a tactical decision. Technically, it could have been submitted individually. The Commission thought that attaching the big programs with broad support, like ESPRIT, to the Framework Programme

[187] Europolitique, 7 December 1985. My conversations with industrialists confirmed that the enlarged goals for Phase II did indeed emerge from industry.

[188] CEC, Communication from the Commission to the Council: The Second Phase of ESPRIT, 14–15.

[189] Ibid., 4, 8, 10–11, 16.

would pull along the smaller programs that would not be approved if left on their own.^[190] I will explain momentarily the difficulties to which this decision led.

Another significant change was that the Commission recommended that organizations from countries belonging to the European Free Trade Association (EFTA) be allowed to participate not just as subcontractors (as under Phase I) but as contracting partners. Non-EC partners would have to pay their own costs plus, when appropriate, a share of the operational expenses.^[191] In addition the five technical areas would be collapsed into three: microelectronics, information-processing systems (including the old software and AIP sectors), and applications systems (including the old office-automation and CIM categories). More importantly, the Commission responded to Council desires for a more practical emphasis than that of ESPRIT I by proposing the inclusion among ESPRIT II projects of a small number of large-scale, focused technology-integration projects (TIPs). The TIPs would require the integration of ESPRIT systems and components results in applications relevant to EC users. The Commission also suggested creating European "centers of excellence" in basic IT research. Finally, the Commission proposed a set of measures to encourage technology transfer—that is, the diffusion of results throughout EC industry and users and especially to SMEs.

The total proposed by the Commission for the new, five-year (1987–91) Framework Programme (the overall plan for EC spending on R&D) was 10.3 BECU, which compared to the current Framework Programme of 3.7 BECU for the four years 1984–87. Of course, the new plan included major new R&D programs, notably ESPRIT and RACE. The Commission's figure was soundly rejected by the three large states, with the other nine members supporting the Commission. France, Germany, and the United Kingdom all argued that the program needed to set priorities, to be more selective, to take greater consideration of what was being done in national programs, and to include more rigorous evaluation procedures for present and future work. Britain's minister for IT, Geoffrey

[190] Interview 67.

[191] CEC, Communication from the Commission to the Council: The Second Phase of ESPRIT, 5, 41, 42. The EFTA countries are Austria, Finland, Norway, Sweden, Switzerland, and Iceland.

Pattie, said the 10.3 BECU budget was about twice as high as it ought to be.^[192]

During the torturous struggle that extended over the following year, the value and importance of ESPRIT Phase II were never questioned by those states that resisted the Framework Programme. In attaching ESPRIT II to the Framework, Commission officials had run the risk of delaying their flagship program in the fight over small ones. A handful of programs, including ESPRIT, RACE, and BRITE (Basic Research in Industrial Technologies for Europe), were those that the French, Germans, and British saw as essential. They wanted other parts of the Framework Programme pared. Indeed, because specific programs needed only majority approval once the Framework passed unanimously, the large states used the initial battle over the Framework budget to try to control the list of programs that would later be included.^[193]

Thus, it was an irony that ESPRIT was one program that the member governments vigorously approved of. In fact, the Council had passed a Resolution in April 1986 expressing a strong commitment to ESPRIT and urging the Commission to proceed with the program while implementing many of the changes suggested by the Review Board.^[194] National officials in France, Germany, and Great Britain all told me either that their governments strongly supported ESPRIT II or that it (and RACE) were their top priorities in the Framework Programme.^[195] The programs that were in doubt included the Community's Joint Research Centers, the Joint European Torus, and a number of smaller programs.

The battle focused on the budget, with Germany and Britain holding out for cuts. By spring 1987 bargaining had reduced the Framework Programme budget to 6.48 BECU, to which all states except Britain agreed. The British came under increasing pressure and even scorn from the Commission, the European Parliament, and other governments. In mid-April the European Parliament voted 141 to 4 to have the Commission withdraw the Framework Programme if the United Kingdom did not approve within three weeks.

[192] Paul Cheeseright, "EEC Bid to Triple Research Spending Fails," Financial Times, 11 June 1986, p. 2.

[193] Several of my sources made this point.

[194] Official Journal, no. C/102, 29 April 1986, 1.

[195] Interviews 4, 30, 47, 48 and 51.

Some delegates discussed the possibility of setting up a separate technology program among the other eleven countries.^[196] Interestingly, Britain would be a net beneficiary of the research programs (receiving in work more than it contributed).

As national elections approached in the United Kingdom, it became clear that no decision would be taken until after the vote. The Commission was claiming that further delays for ESPRIT would cause irreparable harm to the IT sector.^[197] Pressure was also mounting from industry. The heads of the Roundtable companies met with Commissioner Karl-Heinz Narjes in late June to express their concern. They argued that the delays already meant that 600 researchers would have to be released or reassigned temporarily until the fresh funds became available to continue ESPRIT projects. Such an outcome would compel the dismantling of some research teams, which the industrialists argued would alone constitute a serious technological and industrial setback.^[198]

Indeed, industry was enthusiastically behind the second phase of ESPRIT. It was at the initiative of the Roundtable that the work program for Phase II was tripled over that of Phase I, to 30,000 man-years. Executives from IT companies in France, Germany, and Great Britain expressed to me strong approval for the second phase. No one spoke against it.^[199] It would be difficult to argue that the prospect of subsidies inevitably disposed industry to look favorably on the program; the companies put up substantial earnest money for the program—namely, one-half the cost of each of their projects. German executives strongly favored an expanded ESPRIT II, one of them noting that it was much better to spend the money on R&D than on wasting tomatoes or piling up butter.

A French Roundtable executive declared that French industry presented a united front to the government in lobbying hard for ESPRIT II, and that his company would likely increase its ESPRIT involvement from 130 to about 200–220 researchers. A British executive told me that his firm was committed to supporting the increased funding for ESPRIT II with its own financial contributions.

[196] William Dawkins, "UK Attacked on Research Funds," Financial Times, 10 April 1987, p. 2.

[197] William Dawkins, "EC Research Ministers Scrap Talks on Cash Plan," Financial Times, 10 June 1987, p. 2.

[198] Europolitique, 24 June 1987.

[199] Interviews 1, 5, 18, 37, 40, 45, 46, 53 and 60.

Another, from a separate company, declared, "We have been doing our best to persuade the U.K. government to support the Framework Programme, especially including ESPRIT and RACE as essential."

The Electronic Engineering Association, Britain's largest trade association for the electronics industry, called publicly for British approval of the Framework Programme. The Association stated that further delays "must be detrimental to UK interests" and that 60 percent of its members participated in EC-sponsored R&D programs. The statement was times for political leverage, coming out the week before an EC summit at which the Programme would be discussed.^[200]

At the end of June, after the British elections that returned Thatcher, the new cabinet had two new ministers with responsibility for the Framework Programme, Kenneth Clarke and Lord Young. In contrast to Pattie, neither opposed it.^[201] Finally, unanimous political agreement on a 5.2 BECU Framework Programme came from the budget ministers in late July 1987. An additional 417 MECU was held out at British insistence until there was evidence of progress in EC budget control. This compromise was formally ratified by the research ministers in late September 1987.^[202]

As the Council was finally unblocking the Framework Programme, the Commission was submitting its formal proposal for ESPRIT Phase II. It had created a new organizational basis for doing so: In April 1987 the ITTF had been converted into Directorate General (DG) XIII for Telecommunications, Information Industries, and Innovation. DG XIII would administer the ESPRIT and RACE programs as the old ITTF had done. In the ESPRIT II proposal the original budget of 2.2 BECU for the Community contribution had been cut to 1.6 BECU, as even some of the national officials who were sold on ESPRIT thought the first figure too high.^[203] The Commission once again cited the "substantial increase in public support

[200] Terry Dodsworth, "Electronics Body Calls for End to EC Budget Delay," Financial Times, 26 June 1987, p. 7.

[201] William Dawkins, "UK Close to Decision on EC Research Proposals," Financial Times, 24 June 1987, p. 2; William Dawkins, "Belgium Pushes for UK Backing on Research," Financial Times, 25 June 1987, p. 3.

[202] William Dawkins, "Commission Approves Eight Research Projects," Financial Times, 23 July 1987, p. 2; William Dawkins, "EC Unblocks Long-Delayed Research Funds," 29 September 1987, p. 2.

[203] Interview 47.

of funding of R&D in [the] USA and Japan."^[204] The proposal outlined the research areas as follows:

Phase II also included a basic-research component, with a 50 MECU budget.^[206] Topics for research would include molecular electronics, AI and cognitive science, solid-state physics for IT, and advanced system design.

At the same time, the Commission presented its draft work program for the new phase. The plan had been prepared by representatives from 150 EC companies working in the technical panels.^[207] Each of the three major areas constitutes a chapter, divided into the subheadings listed above. Each subheading is further divided into specific projects. Type A projects are outlined in some detail, including technological objectives, approaches, and time scales. Most subheadings contain one and sometimes two TIPs intended to combine technologies into usable systems slightly downstream from the precompetitive realm. The TIPs "aim at meeting ambitious, well-defined industrial targets."^[208] The TIP projects specify milestones

[204] CEC, Proposal for a Council Regulation concerning the European Strategic Programme for Research and Development in Information Technologies (ESPRIT), 4.

[205] Ibid., 16–23.

[206] Ibid., 23–24; Europolitique, 24 June 1987.

[207] CEC, Proposal for a Council Regulation concerning ESPRIT, 32.

[208] CEC, Draft ESPRIT Workprogramme, 0–3.

and deliverables. For instance, the TIP for high-density ICs aimed at a complete integrated system for the design, production, and testing of application-specific ICs. A TIP in advanced systems architectures had as the ultimate goal a prototype parallel-processing high-performance computer. IT applications support systems included a TIP on a multimedia integrated workstation that would employ emerging technologies for workstations capable of handling voice, text, and image data.

Type B projects are not specified; rather, the plan lists within each subheading research topics considered appropriate for the small projects. The Type B themes consist of a general descriptive phrase—for example, under the CIM subheading, "strategies and tactics for advanced flexible automated assembly" and "advanced grippers and artificial skins," and fourteen others.^[209]

After the adoption of the Framework Programme, each of its constituent programs had to be approved by majority vote in the Council. This approval came for ESPRIT II in April 1988. But the Commission had already, the previous December, published a call for proposals. The deadline for submissions was 11 April 1988, and by that date the Commission had received well over 1,000 applications. The following day the Research Council gave final approval to ESPRIT II at 1.6 BECU. Project selection had to be particularly severe: Only 156 of the proposals, involving 585 organizations, were granted contracts. The new basic-research program in ESPRIT also released its first call for proposals in spring 1988. The Commission received 300 proposals. Late in 1988, sixty-two proposals (285 organizations) were selected and eleven working groups were constituted, with a total budget of 63 MECU for thirty months. As noted, the basic-research program focuses on fundamental research on topics relevant to ESPRIT goals, including microelectronics, computing, and AI.^[210] Table 7.3 shows national levels of representation in ESPRIT II and basic-research projects.

In short, the drawn-out struggle to win Council approval for ESPRIT II revealed solid, enthusiastic backing for the program among Europe's major IT firms. The Twelve proposed a tripling of resources

[209] CEC, Proposal for a Council Regulation concerning ESPRIT, III-17.

[210] CEC, ESPRIT: 1988 Annual Report, 3, 65–67; Terry Dodsworth, "Joint Bid for EC Computer Project," Financial Times, 11 April 1988, p. 34; William Dawkins, "Ministers Give Go-Ahead to High-Tech Research Projects," Financial Times, 12 April 1988, p. 2.

devoted to ESPRIT—in the end they were doubled—and were ready to match the EC contribution with funds of their own. The avalanche of proposals for Phase II also demonstrated an impressive degree of interest in the program on the part of IT organizations. The research ministries in the three large countries all favored the continuation and expansion of ESPRIT. Yet, in spite of

such impressive backing, the program was held hostage for over a year to the opposition of a pair of states to a separate—though connected—issue, the Framework Programme.

Secondary Effects of Esprit

Will ESPRIT alter the structures of the IT industries in Europe? Will it make Europe a formidable competitor in world markets? No one can yet say. However, already the IT business in Europe is different. ESPRIT has produced ripple effects, or changes that have taken place outside the program but that are traceable to its influence. I will briefly describe two sets of phenomena growing from the program. The first includes concrete instances of technology cooperation that spun off from ESPRIT; the second is a less tangible but nevertheless widely recognized sense of community in European telematics industries.

Spinoffs

One indication that ESPRIT is having a durable impact on telematics in Europe is that a number of additional projects have begun as a direct result of contacts and cooperation established in ESPRIT. Indeed, RACE is a follow-on to ESPRIT. But ESPRIT also spun off other ventures that did not belong to any EEC or intergovernmental program.

In the spring of 1983 Bull, ICL, and Siemens began discussing the possibility of creating a joint laboratory for R&D on advanced computing. The idea was initially launched by Jacques Stern, president of Bull.^[211] By September the three companies had agreed on a joint research center to focus on precompetitive R&D with full sharing of the results but no joint product development.^[212] The European Computer Research Center (ECRC) officially came into being in January 1984. The ECRC has its own facilities outside of Munich and a staff of fifty full-time researchers. Its projects divide into four groups: computer languages, knowledge-based systems (expert systems), person/machine interaction, and symbolic computer architectures

[211] Guy de Jonquieres and Paul Betts, "European Computer Makers Plan Joint Research," Financial Times, 22 March 1983, p. 44.

[212] Guy de Jonquieres, "European Alliance in Computer Research," Financial Times, 2 September 1983, p. 34.

(as opposed to math-based architectures).^[213] Executives at both Bull and Siemens traced the joint research center directly to ESPRIT and the increased contacts among the Twelve that it brought about.^[214]

Another example of an ESPRIT spinoff is the merger of SGS Microelettronica (the semiconductor division of the state telecommunications company, Societa Finanziaria Telefonica, or STET) and the commercial semiconductor division of Thomson. The new SGS-Thomson, announced in April 1987, would be Europe's second largest producer of ICs, after Philips.^[215] Thomson had been one of France's national champions in electronics, and its marriage to SGS signaled how far France had retreated from its ambitions for national autonomy in the filière électronique . An executive at Thomson told me before the merger was publicly finalized that he was absolutely certain that it came about as a result of ESPRIT interactions.^[216]

Of greater significance than the industrial spinoffs of ESPRIT has been the acceleration of European standardization. Of course, a large share of ESPRIT projects addresses IT standards, and common specifications for future telecommunications systems have been a primary theme of RACE. But a handful of efforts to promote European standards in data-processing and computer networking have sprung up outside the EC programs. The first of these was the Standards Promotion and Application Group (SPAG). The group included the same twelve telematics giants that had begun meeting in the Commission's Roundtables in early 1982. The mobilizing initiative came from France's Bull. Stern, the new president of Bull, appointed by the Mitterrand government in 1982, was personally committed to European standardization for the computer industry. His goal was European adoption of the Open Systems Interconnection (OSI) standards for allowing computers of different makes to communicate and share programs and files. The OSI standards were (and are) being progressively defined by the ISO. Thus, Bull officials organized the first meeting of what became the SPAG group

[213] Paul Tate, "Picking Up Speed," 64.

[214] Interviews 18 and 45.

[215] John Tagliabue, "Europe Joins Semiconductor Battle," International Herald Tribune, 12 May 1987, p. 19.

[216] Interview 37.

in November 1982. One executive closely associated with the process at Bull said that SPAG was inspired in its origins by ESPRIT.^[217] SPAG officially came into being in March 1983, with encouragement from the Commission. The group called on national governments to require conformance to OSI standards in public procurement.

At bottom SPAG was an effort to combat the domination of IBM. IBM had its own proprietary standard—Systems Network Architecture (SNA)—for linking computers. OSI would, its backers hoped, open the market by assuring customers that they could buy from any maker equipment compatible with any other machines. European firms hypothesized that many users bought IBM equipment so as to ensure that it would run on SNA with their previous IBM purchases. Thus, SPAG made specific proposals for European standards (based on the ISO models) in such areas as packet-switched data networks and message handling to European standardization bodies (Comité Européen de Normalisation, Comité Européen de Normalisation Électronique, and CEPT). SPAG submissions have been accepted as the basis for future standards work in Europe.^[218]

SPAG efforts have generated other results. The British government announced in mid-1984 that it would start requiring that computers purchased by the government conform to OSI standards. In September 1984 IBM Europe announced that it would start developing products based on OSI.^[219] After these initial victories part of the SPAG group created a joint company, SPAG Services S.A., in October 1986.^[220] Based in Brussels, with an initial budget of 2.4 MECU per year, SPAG Services offers test facilities for verifying compliance with European standards. It therefore supports and demonstrates the interconnectability of European IT products.^[221]

[217] Robert T. Gallagher, "Stern Spells Europe's Future O-S-I, Not I-B-M," 43; Interview 18.

[218] Eric Le Boucher, "L'Europe Informatique," Le Monde, 16 March 1984, p. 1; Herbert Donner, "The OSI World, Seen from SPAG Europe." Donner was chairman of SPAG Services S.A.

[219] Richard L. Hudson, "IBM Europe Backs a Computer Language Pushed by Its Rivals," Wall Street Journal, 2 May 1986, p. 1.

[220] The initial shareholders were Bull, ICL, Nixdorf, Olivetti, Philips, Siemens, STET, and Thomson.

[221] Jean-Jacques Chiquelin, "L'Informatique Européenne Rentre dans la Norme," Liberation, 3 October 1986, p. 12; Donner, "The OSI World."

The second major standards initiative in IT addressed the operating system for computers. The operating system is the set of internal instructions in a computer that manages the flow of information within the machine and the execution of programs. This time the initiative came in 1984 from Robb Wilmot, then of Britain's ICL. The initial discussion involved five of Europe's computer makers (all of them Roundtable companies). An ICL executive informed me that the process of proposing the group to other partners was greatly eased by the technical contacts created in ESPRIT.^[222] In February 1985 six firms (Bull, ICL, Nixdorf, Olivetti, Philips, and Siemens) announced the formation of the Open Group for Unix Systems. The objective of the group was to encourage the use of the Unix operating system developed by AT&T's Bell Labs, and it later became known as the X/Open Group.^[223]

Again, the idea was to create an alternative to SNA. The thinking was that if a large group of computer makers committed to Unix, it would open the market for their products. First, software writers would be more willing to write programs that could be used on a variety of makes of computer. Second, buyers would be more willing to buy the computers knowing that a large body of software was available to run on them. In addition any programs written to conform to Unix would be operable on any machine built to the standard. Unix would provide operating systems for the fast-growing minicomputer and personal-computer markets.

In time, other makers joined X/Open, including Ericsson of Sweden and the European subsidiaries of DEC and Sperry of the United States. The X/Open Group has produced a manual that provides software producers with guidelines for writing programs according to the Unix applications environment. Initial victories came as some governments (France, the Netherlands, and Sweden) established Unix as a national standard, requiring it for public procurement orders. Significantly, IBM announced that it would offer a version of the Unix operating system on equipment ranging from personal computers to mainframes. A battle over which version of Unix would become the standard has since emerged in the United States, but X/Open agreed on the System V version. The Commission later

[222] Interview 5.

[223] "Six Constructeurs Européens Choisissent un Logiciel d'ATT," Le Monde, 19 February 1985; Europolitique, 20 February 1985; Robert T. Gallagher, "Europeans Are Counting on Unix to Fight IBM," 121.

granted an exemption to the X/Open Group from the EC ban on agreements among firms because any standards produced by the group would be published and would therefore not distort competition.^[224]

These spinoffs provide tentative indications that ESPRIT is changing the way the telematics industries are organized and do business in Europe. At a minimum the collaborative programs have accelerated developments that would have occurred anyway but probably only after further delays and footdragging. ESPRIT provided an organizational structure in which the proper contacts could be made and a consensus could be fashioned. Prior to ESPRIT European firms sought out American companies for technology partnerships. Because of ESPRIT European companies now seek out European partners. In fact, Cadiou, the Commission's director of the ESPRIT program, pointed out that in 1983 (the year before ESPRIT was launched) there were thirty-two alliances linking European to American firms compared with only six between European firms. In 1986 forty-six intra-European linkups almost matched the forty-nine created with U.S. companies.^[225] In addition, although European standards would likely have emerged in time, there is no doubt that ESPRIT hastened the process.

Less Tangible Effects

ESPRIT triggered a burst of collaborative activity in Western Europe and set off changes that are potentially more profound than those in organizations and procedures. The program has altered fundamental perceptions about European telematics. The best way to summarize those shifts in perceptions is to speak of the new-found sense of a European telematics community. Most of the industrial officials interviewed for this book spoke enthusiastically of a new appreciation of Europe's technological strengths and a fresh set of informal ties linking companies and technologists across the

[224] Gallagher, "Europeans Are Counting," 121–22; Agence Europe, 20 December 1986, p. 11; Don Clark, "Computer Giants Gang Up on AT&T-Sun," San Francisco Chronicle, 17 May 1988, pp. C1, C4.

[225] Louise Kehoe, "Seven Take a United Stand against AT&T," Financial Times, 18 May 1988, p. 27. The number of linkups with Japanese firms remained constant at eight.

continent. The new awareness and associations could alter the way European firms conduct technology and product planning.

Officials at ICL in the United Kingdom, for example, told me that the benefits of cooperation were far more important than the funding received: "We have new contacts and new confidence." One official noted that there was, after ESPRIT, more of a European IT community. Another executive declared that the main success of ESPRIT had been the creation of an IT network in Europe: "I can call my friends at Siemens or Philips and ask what's going on. This is hard to quantify but very important." A research director at GEC remarked that although his firm identified real technical benefits in software standards and semiconductors, there was "almost more benefit from the companies knowing each other better, to take advantage of the potential that was spread around Europe."^[226]

A French Roundtable executive expressed a similar notion, noting that ESPRIT brings together some "2,000 or so European scientists and technologists, creating a European community of researchers." Another French industrialist declared that one of the spinoffs of ESPRIT had been extensive visiting of other laboratories, which allowed people to know better what was going on technologically elsewhere and so make better decisions about alliances.^[227] More recently, Stern, chairman of Bull, argued that European cooperation in IT should go well beyond precompetitive R&D. He said that European firms are now more willing to collaborate than they were previously because of the success of ESPRIT and progress on common standards.^[228]

Officials at Siemens recognized that ESPRIT allowed them to take advantage of the "synergies" of working with the other partners, from whom they had learned a great deal. An executive at Nixdorf noted a new sense of confidence in European IT firms that resulted from ESPRIT.^[229]

In short, regardless of whether Europe closes any technology gaps because of ESPRIT and its follow-ons, both telematics policy and business practice have changed. Cooperative industrial ventures, new

[226] Interviews 5, 53, and 60.

[227] Interviews 37 and 68.

[228] Alan Cane, "High-Tech Warning to Europe," Financial Times, 8 June 1989, p. 2.

[229] Interviews 45 and 54.

standards initiatives, and the fresh sense of a telematics community imply that some of the changes will be durable.

Conclusion

ESPRIT confirms the utility of the theoretical framework outlined in Chapter 2. First, it addresses the question of demand for cooperation. Theory must account for the cognitive process by which state leaders come to choose cooperation, especially in an area as dominated by nationalist policies as IT had been. The cognitive-process variables were clearly at work in ESPRIT: The national-champion strategies of the past decade had failed to close any technology gaps. Moreover, European national policy-makers recognized the failure and were in an adaptive mode, searching for new means of promoting their IT industries. The evidence for this adaptative mode consists of major studies of the IT sector in France, Germany, and the United Kingdom, all leading to new and enlarged national support programs.

Second, analysis must account for the political leadership necessary to organize cooperation. ESPRIT shows that IOs can exercise that leadership. Indeed, the Commission was indispensable in ESPRIT. It both mobilized and shaped the European consensus behind collaboration. The Commission was technically well prepared, was headed by an entrepreneurial leader in Davignon, and faced an opportunity for leadership because of the policy crisis in member governments. The most important political factor in ESPRIT was the alliance struck by the Commission with the Roundtable companies. The Twelve designed the program and sold it to their governments. No political explanation of ESPRIT could suffice without assigning importance to EC institutions and to the transnational industrial alliance.

ESPRIT also incorporated organizational features that enhanced the prospects for successful collaboration. For one, the program was large. ESPRIT I was the largest EC R&D program to date, and ESPRIT II accounted for fully one quarter of the 1987–91 Framework Programme. The total budget for both phases was over $5 billion. The program was substantial enough in the first phase to include 420 organizations from across the EC. With that many pieces of work, chances were far better that each state could receive an

acceptable share than had there been only a few dozen pieces to fight over. Furthermore, in the EMC states could bargain somewhat for the inclusion of national organizations in ESPRIT projects. Still, no participants pointed to EMC give-and-take as a problem for ESPRIT. Participation is voluntary, left up to the individual companies, universities, and institutes. Though each member country contributes to ESPRIT through the general budget, no country can complain of being forced to participate in projects it finds useless. Thus, ESPRIT satisfies juste retour concerns via a modified form of à la carte.

ESPRIT might not propel Europe to the same level as the United States and Japan in IT. But it has gone far in creating a European IT community. At a minimum ESPRIT was a bold political initiative, one that ended two decades of exclusively unilateral, nationalist IT policies in Europe.