ABSTRACT
The first major initiative in 1971 was directly concerned with the creation of a national computer industry in Brazil. Naval technical personnel, led by Commander José Luis Guaranys Rego of the Electronics Division of the Navy, were worried about the growing technological dependence of the Navy and in particular the inability of Brazilians even to service the British-made minicomputers aboard six recently purchased British frigates.4 Civilian technical personnel concentrated in various agencies of the Ministry of Planning, especially the National Economic Development Bank (BNDE), had similar concerns about technological dependence based on their own experiences with the state of research and development in Brazil as well as influential anti-dependency writings by Brazilian scientists beginning in the early 1960s. Economists at the BNDE had also been strongly influenced by anti-dependency ideas associated with the Economic Commission for Latin America, with whom the BNDE had worked closely since the early 1950s. The BNDE group had already identified the minicomputer (less sophisticated technologically than large mainframes and not being produced by any multinational company inside Brazil) as a promising start-up opportunity for a possible Brazilian industry. The Navy created a special working group called the GTE/FUNTEC 111, a joint (and exclusive) operation of the Navy and the BNDE with BNDE’s Scientific and Technical Development Fund (FUNTEC) providing 60 percent of the necessary funds and the Navy 40 percent.5 The key members of the group were Commander Guaranys from the Navy and a BNDE engineer with a Ph.D. from Stanford, Ricardo Saur.6 The Brazilian computer policy would eventually embrace the principle of reserving a major segment of the computer market for companies in which Brazilians controlled 100 percent of the ownership, management, and technology (including technology purchased from abroad). In the beginning, however, the special working group pursued a different model, a three-way joint venture (known as the tri-pé) involving the Brazilian state, private national capital, and private foreign capital. This model was already proving successful in the development of an eventually quite sophisticated petrochemical industry in Brazil.7 A joint statement by the Ministry of Planning and the Navy as late as 1975 clearly foresees the government as a temporary partner and describes the foreign role as simply a matter of technology transfer, allowing for the participation of foreign firms “provided that they accept minority participation, possess valid technology, agree to transfer it at a fair price, contribute capital and create the possibility of exports.”8 However, the effort to implement the tri-pé model was short-lived in the computer sector after the creation of COBRA in July 1974. No major market leader in the developed economies could be found willing to enter as a minority partner into a joint venture and transfer full control of its technology over time to its Brazilian partners. IBM and Burroughs declined even to make proposals (IBM later in the decade would pull out of India completely rather than accede to demands for a joint venture with Indian firms9), while Digital, Data General, and
Hewlett Packard all made demands for majority status or control of technology that were unacceptable to the Brazilians, intent on building a national industry.10 The special working group then selected Ferranti, the small British computer company that had supplied the minicomputers for the Brazilian Navy’s frigates. Ferranti, however, suffering from financial difficulties in its home market and frustrated by conflicts with the other members of the tri-pé, gradually reduced its involvement in the enterprise and by 1977, only three years after COBRA’s founding, controlled less than 5 percent of the capital.11 Moreover, from the beginning, the special working group had pursued a parallel project to produce a minicomputer designed in Brazil by Brazilians. Although COBRA initially marketed minicomputers using Ferranti’s technology and later shifted to a model based on technology purchased outright from Sycor, a small Amer ican company, by 1979 the company was producing and selling its own line of minicomputers that were a direct product of the special working group’s commitment to developing Brazilian technology. The development work was begun at Brazilian universities and later completed via COBRA’s own research and development operation.12 With the development of this flagship line of products based on Brazilian designs, the company no longer needed a foreign source, let alone a foreign partner, to provide minicomputer technology. So a model that began with the presumption that a foreign partner to provide technology was essential soon found itself with only Brazilian participants. The original effort to involve private Brazilian capital in the new industry ran into difficulties even more rapidly. The special working group selected E.E. Equipamentos Electrônicos (EEE), a company that had some experience manufacturing electronic equipment for the Navy, as the national partner in the tri-pé, but by April 1975, less than a year after the founding of COBRA, EEE had effectively withdrawn from the venture, and no other substantial Brazilian company showed any interest13 despite approaches being made to some 60 different Brazilian companies.14 The still infant company not only became an almost purely national operation but also became essentially an enterprise of the Brazilian state with private capital participating only in a distinctly secondary role. This outcome had little in common with the tri-pé that the special working group had been originally commissioned to create. Foreign capital had been largely excluded, despite Brazilian eagerness to include a foreign partner, because of the reluctance to participate in technology transfer of the sort contemplated. Private Brazilian capital was also on the sidelines as a result of its general indifference to the project and secondarily because of the weakness of firms with any relevant experience in the electronics area. Once the prospects for a genuine tri-pé collapsed, there was a constant battle within the Brazilian state over COBRA’s status. After it established a measure of technological competence and commercial success, there were a number of campaigns to simply sell the company to one of several private banks or electronics firms that eventually became interested in the company. There was also tremendous confusion within the Brazilian state over who was to control the selection of COBRA’s management team and shape the policies that would
determine COBRA’s role in the national informatics industry. Sometimes due to keen interest and other times by virtue of almost inadvertently becoming the owners of a major portion of COBRA’s stock through a reshuffling of administrative agencies, all of the following state agencies at one time or another participated in running COBRA: CAPRE and its successor, SEI; the Ministry of Planning; the Ministry of Finance; the Ministry of Science and Technology; the Bank of Brazil; the Caixa Econômica Federal (a government savings bank); the BNDE (later the BNDES); DIGIBRÁS; SERPRO (the government dataprocessing agency); the Navy; and the SNI (the military intelligence service). With such pulling and hauling by very disparate forces within the state (controlled by a considerable variety of ideological orientations), it proved difficult ever to fully integrate the increasingly significant COBRA into the overall informatics policy that came under the control first of CAPRE and later of SEI. In many respects, COBRA maintained a relationship to the policymaking agencies as if it were a private (albeit somewhat privileged) enterprise, pursuing its own entrepreneurial agenda that sometimes reinforced and sometimes undermined the larger policy. The difficulty in coordinating the actions of the state’s informatics agency and the state’s own informatics company illustrates the fragmented character of the Brazilian state. The existence of multiple, sometimes competing agencies in control of the variety of resources related to informatics within the Brazilian state practically guaranteed that all of these resources could never be fully integrated into a coherent and effective policy. Their integration was a political challenge never completely met.
