The Political Economy of Technological Development: The Case of the Brazilian Informatics Policy

1. The BIP: a capsule chronology

Historical periodization of a policy necessarily involves a degree of arbitrariness because of the cumulative nature of the processes underlying the policy. As an example, it may be claimed that the origins of BIP lay in early fifties, when the ITA (Instituto Tecnológico da Aeronáutica – Air Force Technology Institute) started its electronics engineering course, which provided a considerable part of the technical expertise upon which the Policy was grounded and where a significant number of cadres responsible for the Policy was formed. Because we are dealing with a policy under an evolutionary approach the main criteria adopted here were life-cycle and institutional.

Following such criteria, we may say that the BIP’s life-span covers twenty years: from the creation of the Special Working Group (GTE – Grupo de Trabalho Especial) jointly established by the Navy and the National Economic Development Bank (BNDE) in 1971 to the new Informatics Policy Law of 1991, which drew BIP to a close. Such history may be subdivided into three periods, detailed below:

1.1. Infancy: From the GTE to CAPRE – 1971/75

During the fifties and sixties the Brazilian State (mainly the Federal Government but also the State of São Paulo) fostered the development of technical and scientific capabilities in fields related to BIP by supporting local academic institutions and by providing training abroad. However such support lacked any specific industrial purpose – the use of the resources thus developed was left to the market. The demand for electronic equipment was supplied either by imports or by local production by subsidiaries of multinational companies. As a consequence, employment of skilled personnel was directed mainly to selling equipment and using it, with a minority being absorbed by the academia. State agencies were major users of such equipment.

One of the main souces of funding for graduate education and research was the BNDE, the main industrialization financial agency of the country. Although the Bank’s department for science and technology (FUNTEC) was marginal to the agency, institutionally and resource-wise, it became an important breeding-ground for science and technology policy-makers. In 1968 the department had floated the idea of developing a computer prototype in the local universities. Independently, the Navy had adopted in 1969 a policy of fostering the local production of electronics equipment used for its vessels and airplanes. The purchase of British frigates equipped with computers brought the two agencies together, under a Special Working Group (GTE 111) which led to a project aiming at designing, developing and producing computers for naval use. As a consequence, in 1974 a company (COBRA – Computadores Brasileiros S.A. – Brazilian Computers) was established to produce such equipment – a joint-venture between the Navy supplier of computers to its frigates (Ferranti), a local supplier of other electronic equipment to the Navy (Equipamentos Eletrônicos) and the Bank. A second company was planned, under the same tripartite model, to produce commercial equipment, but never took off. 

In the same period the tripartite model, combining Brazilian State and private capital and foreing enterprises, the latter acting mainly as suppliers of technology, was successfully used for the establishment of the petrochemical industry in the Northeast of the country. The model was conceived in order to ensure simultaneous private and national majority of the capital of the new enterprises, at the same time it provided the enterprises access to foreign technology.

Simultaneously, other electronic data processing equipments were developed by State agencies, mainly for processing fiscal data and for cryptogaphy, and by some universities, mainly for academic purposes. In 1971, the Planning Ministry revamped a small agency dedicated to financing pre-investment studies (FINEP – Financiadora de Estudos e Projetos – Financial Agency for Studies and Projects) as a development bank for science and technology, initially run by bureaucrats from BNDES’s FUNTEC. The support of graduate education and research in computing sciences and the supply of computing equipment to universities soon became a priority of FINEP. 

At the user-end, the Ministry for Planning and Coordination established in 1972 a Commission for Electronics Processing Activities (CAPRE – Comissão de Atividades de Processamento Eletrônico) in order to rationalize the use and purchase of computers by the Federal Government as well as the training of personnel. In 1975, following the severe foreing currency restrictions due to the oil shock, CAPRE was empowered to control imports of computers, parts and components, which had become a main import item, increasing 600% between 1969 and 1974 (Piragibe 1985).

CAPRE was staffed by technical cadres originated from the bureaucracy and from the academia. In turn, CAPRE officers helped to organize scientific societies and meetings related to computing and the associations of electronics data processing professionals (e.g. systems analysts, programmers, etc). The three groups – bureaucrats (civilian and military), academics and professionals – formed BIP’s hard-core, providing its concepts and political support. Given the authoritarian nature of the regime, the former were audible within the State only but the latter two were quite vocal publicly.

Although the First National Development Plan (1972/74) and the First Plan for Science and Technology (1973/74) listed the computer industry among their priorities, singling out minicomputers, and in spite of the initiatives mentioned above, there was not a policy for the sector nor an agency in charge of it. Nonetheless, the main political and institutional building blocks of BIP had been laid down. 

1.2. From youth to maturity under bureaucratic rule: from CAPRE to SEI – 1976/1983

The Brazilian economic history of the second half of the seventies is dominated by the Federal Government Second Development Plan – a major industrialization thrust aiming at, on the one hand, completing the Brazilian industrial structure by large investments in the industries producing intermediary products and capital goods and, on the other hand, changing the energy matrix, by replacing gasoline by alcohol and by complementing the large investments in hydro power with an ambitious nuclear power program, jointly developed with the Federal Republic of Germany.

Underlying the Plan was the idea that Brazil was an “emerging international power”. Altough the Plan was led by the State its aim was to strenghten the nationally-owned enterprises but avoiding as much as possible to antagonize the multinational companies, especially those already established in the country.

The development of a local technological capability, including innovation skills, was an important part of the Plan’s design, since imports were associated with the undevelopment of such capability. Accordingly, FINEP’s resources for developing the research and graduate education system and for granting subsidized loans to technology projects of national enterprises were expanded. State enterprises were instructed to favour locally designed products in their purchases and they greatly expanded their own R&D centres. At the same time, stricter rules for importing technology were adopted, often conditioning imports to local technological efforts.

Parallel to the main thrust of the Plan, which aimed at completing the structure of production of the Second Industrial Revolution, the Government policy-makers established a set of policies aiming at high-tech sectors, the vectors of the Third Industrial Revolution: telecommunications, aeronautics, armaments, nuclear power and, last but not least, informatics. Such sectoral policies has two elements in common: the emphasis on local technological capabilities (including innovation) and the control of the sector by national enterprises.

Within this context, in the begining of 1976 CAPRE was restructured. gaining a Council composed by high officers of several Ministries and empowered to define a national informatics policy. The policy, stated in the middle of the year, had five main objectives:

i) to achieve technological capability to design, develop and produce electronic equipment and software in the country;

ii) to ensure that national firms hold a proeminent position in the national market;

iii) to create opportunities for the development of the informatics parts and components industry;

iv) to create jobs and, especially, more qualified employment for national engineers and technicians;

v) to generate a favourable balance of payments for informatics products and services.

Retaining its power to control imports, which were necessary to manufacture computers in Brazil, CAPRE was also empowered to analyse local manufacturing projects. However, it lacked any positive inducement policy instrument, such as credits or fiscal incentives. 

Within the sector, the priority lay with the segment of small systems, mini and microcomputers and their peripherals, which were not yet produced in the country and to which the efforts to develop national technology should be directed, while for larger systems, locally produced by subsidiaries of multinational firms, especially IBM and Burroughs, the emphasis was laid on rationalizing the investments by using the resources available. 

Under the leadership of BNDE, the State took control of COBRA’s capital and the company was turned around to become a supplier of commercial equipment, mainly for the banking system. A consortium of private banks took an equity interest, followed by the two other main Federal State banks (Banco do Brasil and Caixa Econômica Federal). Although it used licensed technology to produce its banking automation equipment, COBRA continued to invest heavily in its own projects, developing the small-systems concepts originated from the GTE. However, an important by-product of this decision was the reduction of the Navy’s support of the policy.

Simultaneously, IBM anounced its intention to produce locally its minicomputer /32. It was widely recognized that such production would kill the policy. As a consequence, battle-lines were drawn: on the foreground, on one side IBM and other multinational subsidiaries and on the other the academia and professional communities. In the background, the Government was deeply divided and subject to strong internal and external pressures.

Such conflicts, which lasted throughout 1976 and 1977, led to a bid to manufacture minicomputers in Brazil. Projects were assessed by CAPRE according to five criteria: local content of manufacture in terms of components and employment; technological “openness”, giving higher priority to projects which involved greater disclosure of imported technology and greater local technological development; control of the internal market in order to avoid excessive concentration in the hands of a single firm; local equity control and, finally, balance of payments conditions. Implicit, there was a belief that the import of technology would be a once-for-all affair and that next generations of the same type of equipment could rely on locally developed technology. 

Fifteen projects were submitted to CAPRE, out of which three were selected at the end of 1977. All three were based on imported technology with the appropriate clauses of openness. None of the suppliers of technology were industry leaders. In fact, the latter, through their subsidiaries in Brazil (IBM, Burroughs, HP, NCR, Olivetti and TRW) chose to present themselves to the bidding without local partners and with scant attention to the other conditions of the bid. Out of the pressure of segments of the intelligentsia and the bureaucracy and of the rigidity of the multinational leading firms the market reserve was born.

The market reserve for locally-controlled firms was applied to minicomputers and smaller systems only. Combined to import restrictions it implied that the segment of larger systems was de facto reserved to the subsidiaries of multinational firms already established in Brazil – mainly IBM and Burroughs. Because of its product mix the other main foreing supplier, Olivetti, stood more to loose from the policy and eventually withdrew from the sector. Although they were barred from the fastest-growing market segment, the other two, and more clearly IBM, henceforth changed strategy, refraining from explicitly attacking the policy and attempting to circumvent it by playing upon the definition of the borders of the policy, producing computers of “medium” size. However, other U.S. companies, such as Data General, totally excluded from the Brazilian market, started pressing the U.S. Government to take action against the policy.

In 1979, with the change in Government, a new actor came to the fore in the informatics policy: the national security community. Since the mid-seventies the National Security Council and the Foreign Affairs Ministry had been involved with informatics projects related to cryptography, developing equipment and software locally, which led to the creation of a specialized firm. Such experience strenghtened their perception of the weakness of the Brazilian electronics industry. Considering electronics strategic for the objectives of national security, interpreted at the time in wide-embracing terms, they considered CAPRE’s policies weak and limited in terms of the range of electronic activities and they viewed with suspicion its involvement with the academia, a traditional focus of resistance to the military regime. 

Holding very strong powers within the Government, the Security Council, supported by the Foreign Affairs Ministry, sought the control of the informatics policy. Such bid was largely unopposed by the economic ministries, at the time concerned mainly with the foreign adjustment of the economy and with controlling inflation. As a result, CAPRE, where the majority of the Council was held by civilian ministries, was replaced by a Special Informatics Secretary (SEI – Secretaria Especial de Informática) attached to the National Security Council. 

SEI immediately laid down its Directives of the National Informatics Policy. Contrary to many expectations, the objectives of local technological and industrial development under national control were maintained. In fact, the Directives is one of the few official documents in which the market reserve for national firms is explicitly aknowledged. Moreover, the range of activities covered by the policy was broadened so as to include software, national networks for data communication and all components of informatics products and services. 

SEI had three instruments to implement its Directives: the control of imports, the granting of permission to local manufacture and the supervision of purchases by State enterprises and agencies. However, in the latter case it was never able to develop a comprehensive policy for the multifarious Brazilian State. In order to foster local technological development, SEI established at the end of 1982 a research centre (Centro Tecnológico para Informática – Technological Centre for Informatics) with four areas of concentration: computing, process automation, instrumentation and microelectronics.

Over the next three years, following its Directives, SEI defined through its Normative Acts policies and participants in the areas of computers and peripherals, industrial automation equipment, electronics instruments, microelectronic components and software. However, progress in the latter two, the top priority areas, was slow, albeit for different reasons. In microelectronics there were two large industrial groups interested but they demanded fiscal and credit incentives which laid outside the pail of SEI and which the economic ministries were unwilling to give. As a result, investments were procrastinated while, internationally, the minimum scale of plants grew by leaps and bounds. For software, although SEI imposed import controls and conditioned the approval of manufacturing projects for general-purpose microcomputers to the adoption of operating systems locally developed, such measures proved to be ineffective against copying and smuggling of imported software.

Moreover,SEI’s attempt to establish a coherent policy for the electronics complex was foiled at two strategic points – telecommunications and consumer electronics. The first sector was ruled by the Communications Ministry, which followed an independent policy, where products were manufactured by firms which were nationalized subsidiaries of multinational companies using either the technology supplied by the former parent companies or the technology developed by the National Telecomms Research Centre. The second sector was located in a Free Trade Zone in the Amazons region, under the auspices of the Interior Ministry and its firms operated mainly as users of foreign designs and assemblers of imported components, directing their products to the internal market only.

Lacking any positive inducement policy mechanism, the efficacy of SEI’s policies was further undermined by resistance of the economic Ministries (Planning and Treasury), where the objective of local technological development had become a remote priority, as witnessed by decline in the funding of FINEP and the reduction of BNDES’ support to the policy. The same applied to COBRA, the State Enterprise which had led the market technologically in minicomputers, and which was left to languish in terms of mission and resources. The lack of financial incentives to investments affected especially the development of the strategic industry of microelectronics components, the core of the electronics complex, signalling to the firms involved a limited commitment of the Government to the policy.

SEI’s policies were obviously supported by the many industrial firms which had entered the market. However, the policy inevitably brought it into conflict with foreign subsidiaries previously occupying the Brazilian market, which often opted to leave the country altogether or to supply it via local licensors. Users of equipment also often resented the restrictions imposed on the supply and the inevitable higher costs of infant products. Moreover, the military nature of the Secretary estranged the academic community and parts of the civil bureaucracy which had previously supported the policy. Even the main beneficiaries of the policy – the local companies arisen under the umbrella of the reserved market – often resented the restrictions to imports of components and the power wielded by the bureaucracy. Hard-core supporters of the policy also complained about the “laxity” of SEI’s criteria about technology imports, putting a negative premium on the firms which invested more on the local development of technology.

With the end of the military regime on sight, the civilian supporters of the policy, especially the academics, the professionals and the industrialists more deeply committed to local technological development, sought the backing of Congress to pass a law giving the policy appropriate political legitimacy. At the same time, concerned with international pressures and wishing to establish the policy on firmer ground, the military backers of the policy within the Government were also led to submit a Bill to Congress, refraining from using the power of the Executive to pass a decree. 

1.3. Maturity and decay under the law: from 1984 to 1990

In 1984 the Brazilian Congress passed by unanimous vote an Informatics Law. Simbolically, the date was the same in which thirty years before the law creating the oil monopoly had been voted – possibly the single major victory of the country’s nationalist faction.

In essence, the Law was a ratification of CAPRE’s and SEI’s policies. It confirmed the prime objective of industrialization-cum-local technological development and the priority given to locally controlled firms. In this latter aspect it went a step further and defined the “local control” so as to involve technology – a feature latter embedded in the definition of “national enterprise” of the Constitution of 1988. Similarly, the scope of the policy, stemming from the definition of “informatics”, covered the whole electronics complex.

Institutionally, however, the Law led to a break in the previous mold, in which the policy was defined by the Executive alone. Now Congress had to approve the three-years Informatics Plans (PLANIN) and the Council in charge of presenting the Plans to Congress and taking the main decisions regarding the implementation of the policy included a strong representation of the civilian institutions supporting the policy (CONIN – Conselho Nacional de Informática e Automação – National Council for Informatics and Automation). SEI was retained as the executive secretariat of the Council.

The instruments of BIP were also nominally increased, although no mention is made of a market reserve. SEI retained the power to control imports, albeit for a period of eight years and gained the management of some fiscal incentives, especially for microelectronics and software. The creation of a special fund for financing R&D expenditures was vetoed by the President but CTI was confirmed as a policy instrument under the jurisdiction of SEI. COBRA, however, was left under the control of the State banks, the private consortium having left some years before, as the banks which participated in it developed individual informatization strategies and acquired interests in other informatics companies or established subsidiaries to supply banking equipment. Although the importance of State purchasing power as an industrial and technological policy instrument was recognized, no mechanism of coordination was established. In practice, since the fiscal incentives were very limited, the instruments of the policy remained the same as before the Law.

A major effect of the Law was to impart a much stronger legitimacy to the policy, strenghtening its power. However, the unanimity of the last vote in Congress was deceptive. In fact, BIP’s two main attributes – development of a local innovation capability under control of national enterprises – made it intrinsically conflictive. 

The period of preparation of the Bill and the debates in Congress showed that the opposition to the policy was not exclusive to politicians traditionally aligned with foreign interests but came also from a wide range of actors which felt their beliefs and interests threatened. This arc encompassed from objectors to the principle of State intervention to multinational companies active in the electronics complex, with, in between, officers from the economic ministries, local enterprises operating essentially with imported technology and multinationals from other sectors, fearful of the extension of the informatics policy to other areas.

As a result, the final version of the Law contained several compromises, the most important being the limitation of SEI’s power to control imports to eight years. Nothing was said about what would happen the day after. Some interpreted the deadline as the end of the market reserve and others argued that the control could be taken over by another agency (CACEX), which controlled the rest of Brazilian imports. The press, which strongly opposed the policy, supported the first view, which tended to become dominant.

With the new civilian Government, come into power in 1985, the informatics policy was entrusted to the newly-created Ministry of Science and Technology but telecomms and consumer electronics remained under different Ministries. Moreover, the politicians in charge of the three Ministries held diametrally opposed views about the policy, reflected within CONIN. Policy coordination within the electronics complex was rendered all but impossible. Support from the economic ministries was, at best, lukewarm and the retrenchement of the military power implied a weakening of the policy. 

Outside the Government, there was a backlash of the 1984 decision: while the supporters of the Law demobilized their forces, the opposers strenghtened theirs with a mounting press campaign. Since BIP provided a model which could be used for other industrial sectors the trenches gained new participants, especially from other high-tech sectors still undeveloped in the country – biotechnology and fine chemicals. On both sides the participants had the same origins as in informatics, but since the strenght of opposers (e.g. multinational firms from the pharmaceutical industry) was much greater than the supporters’ (e.g. small Brazilian firms producing fine chemicals), the net result of the proposals to expand the scope of the policy to other sectors was to weaken it further. 

Against this background of internal forces came a strong external pressure: in September 1985 the US government announced the start of an investigation into the Brazilian Informatics Policy and the possibility of economic retaliation if discriminatory or unfair trade practices were found against US interests. The possibility of restricting Brazilian exports to the US market of sectors (e.g. orange juice producers) which had nothing to benefit from BIP incresed the private sector opposition to the policy. Within the Brazilian government, the coming of a conflict at a time of balance of payments restrictions and intricate foreign debt renegotiations was anything but welcome at the economic and Foreign Affairs ministries .

In spite of this mounting pressure the Brazilian Executive and Legislative refused to change the Informatics Law and, in order to stem criticisms, SEI was reorganized in 1986 so as to gain greater administrative efficiency. Nonetheless, an incremental process of deterioration of the policy got under way, especially from 1987 onwards, as the more nationalistic faction of the ruling party got weaker and the economic conditions of the country worsened.

Thus, CONIN gave way to US pressures in several instances of import restrictions and Congress passed a Software Bill which was a clear compromise with US demands. A major turning point in the policy, because of the growing importance of software for the sector, the Bill provided very limited protection to locally-developed software, undermining thus the core of BIP. Moreover, the President of the Republic stated to the press that informatics was a “special case”, killing the expansion of BIP to other sectors, and that he would not repeat the vote favourable to the Informatics Law he had given in 1984 as a Senator. Following this, the Presidency took several administrative measures, e.g. regarding the control by SEI of imports for the assembly of electronic products in the Amazon Free Trade Zone and the classification of enterprises as “national”, which gave clear signs that the Executive was withdrawing its support to the policy. 

Less obviously, but not less effectively, the Ministry of Finance put a low ceiling on SEI’s foreign exchange allowance. The imports of electronic products depended on such quota and were restricted accordingly, creating considerable strain between SEI and users of such products, with the former taking the blame. Moreover, a blind eye was turned on smuggling of electronic products, which reduced the market for locally produced goods and services and jeopardized the results of projects aiming at local technological development.

Possibly anticipating the demise of the policy, some of the leading Brazilian informatics firms strenghtened their technological links to foreign suppliers and laid the ground for future joint-ventures, at the same time they used smuggled components massively, undermining the core of BIP and making its end a self-fulfilling prophesy. Such strategies were deeply resented by the enterprises more deeply committed to the technological objective of the policy, causing a rift in the informatics entrepreneurial community.

Under such pressures BIP’s policy-making apparatus lost all initiative. SEI, accused of rigidity, ended up by condoning entrepreneurial strategies, such as those mentioned above, which run counter the heart of the policy and CONIN was unable to present Congress with a strong II Plan for the sector, leading to the extension of the First Plan. The absorption in 1989 of the Ministry of Science and Technology by the Ministry of Industry, then run by a notorious oponent of BIP, put an institutional lid on the policy.

Therefore, at the end of eighties, the Brazilian Informatics Policy in its pristine form was agonizing. However, it was not dead yet and, besides the entrepreneurs of the sector, it could still rally support from important segments of society, especially in the intelligentsia and the bureaucracy. Among such supporters there was a consensus, made explicit during the discussions preparatory of the II PLANIN, that the policy should and could be revived, albeit introducing major changes. Such modifications should lead to a greater product selectivity, restricting the range of products to which it should be applied. Local manufacture should embody an increase in the import content so as to better profit from external technological developments and should aim at cost reductions and at exports. It was hoped that if such changes were introduced, the policy could regain internal political legitimacy and retain its main pillars – the commitment to local technological development and the fostering of locally controlled firms.

However, the Presidential elections held at the end of 1989, when the economy was bordering hyperinflation and the State policy-making capability was reduced to a minimum, led to the victory of a candidate explicitly committed to a reduction in State intervention and an opening of economy to imports and foreign capital. 

The supporters of BIP, especially the academics, tried to rally public support to the policy, attempting to revive the Brazilian Informatics Movement (MBI) which had been the mainstay of the 1984 campaign, but with very little success. ABICOMP, the national producers association, preferred instead futile attempts to negotiate with the Executive, ignoring that the past alliance with the bureaucracy was dead. Committed to put an end to the policy, the Executive adopted in 1990 administrative and legal procedures to such purpose, such as withdrawing the II PLANIN from Congress, reducing the number of products subject to import controls and extinguishing SEI. Appropriately, the last battle was fought in Congress and a new Informatics Bill was passed in 1991, after a strong press campaign against the old policy, especially against Congressmen who supported it. 

The new Law changes the concept of “national enterprise”, reducing it to a majority of capital, which allows for joint-ventures in which technological control is in the hands of the foreign partner. CONIN was downgraded and its composition altered, reducing the weight of the groups previously supporting BIP in favour of its opponents. SEI was extinguished and replaced by a Department (DEPIN) of a new Ministry of Science and Technology, holding practically no powers. 

The Bill replaces import administrative controls by tariffs, which decline over time. By 1994 tariffs will range from 35% for finished goods to nil for components not produced in the country. As positive inducement mechanism the Law envisages fiscal incentives for local production and for R&D activities, as well as the use of the State purchasing power. However, none of such mechanisms was yet implemented.

Buffeted by the combination of general recession and policy changes, the industry went into a deep crisis, of which there is no end on sight: as compared to 1989, it is estimated by DEPIN that in 1992 the local industry had reduced net earnings, employment and R&D investment by, respectively, 47%, 60% and 69%. So far the worst affected were producers of components and peripherals. Suppliers of banking automation equipment and of computers have fared relatively better – the former because they have a leading edge on technology combined with a captive and rich market and the latter by virtue of associations with foreign companies. Such partnerships have led to the abandonment of all plans of local technological development and entail no small risks for the survival of the local firms when import liberalization will come into full stride. 

Although it is beyond the scope of this paper to develop more fully conjectures about the future of the electronics industry in Brazil, it is clear that a chapter of the country’s science and technology and industrial policy has come to an end.

2. Objectives, scope, instruments and social support

2.1. Objectives and scope

BIP was an ambitious project, to say the least. It aimed at creating a high-tech sector – first an industry and later the whole electronics complex – in a developing country, endowing such industry with the full range of technological capabilities – from research to marketing – under the control of national entrepreneurs. The two attributes – technological self-reliance and national enterprises – made up its diferentia specifica, carrying it a step further traditional import-substitution policies, which relied almost exclusively upon imports of technology, especially for innovation activies, and upon subsidiaries of international companies. 

The two attributes were organically linked: it was correctly assumed that the subsidiaries of multinational companies would not invest in research and development in Brazil, having the option to do so nearer their parent companies, in locations where they enjoyed strategic and systemic advantages. Therefore, only national companies would be a vehicle for a relative innovation capability. As shown by other sectoral experiences (e.g. petrochemicals), joint-ventures with international companies where the latter held the control of technology did not lead to the transfer of technological innovation capabilities to the local company. Therefore, wholly locally owned companies were necessary. If madness it was, there was method in it.

However, the two attributes were never unqualified. Technological self-reliance was never confused with autarchy: technology imports were always an important part of the strategy, as the starting point for local production, although it was assumed that, over time, they could be replaced by local alternatives. As originally envisaged, it was assumed that having mastered the innovation capability for a family of products (e.g. minicomputers), based on a combination of technology imports and endogeneous investments, the local firms would be able to prescind from further technology imports, being able to follow the international frontier.

This process would be “horizontally” repeated, for new product families (e.g. microcomputers), widening the scope of the policy, but it was understood that this scope was restricted by economic and technlogical factors – e.g. mainframes were excluded. Nonetheless, the technological interdepencies which are characteristic of electronics industries, making them an “industrial complex”, led to a substantial widening of the scope of the policy.

In other words, the objective of the policy, in its most radical version, was to close the technological gap over time for a selected but widening range of families of products. It was a process of import substitution in which local manufacturing and innovation capabilities were sequentially developed . Protection against imports – of products and of the design built-in into such products – was regarded as essential to such purpose. As originally envisaged, such protection should be coupled to positive incentives such as venture capital, credits, tax reductions and State procurement. Thus, to the risk-reduction imparted by import controls should be added other measures, which would reduce risks and costs and accelerate the catching up process. However, as already pointed out, the second part of the policy instruments was never fully set up, retarding the closing up of the technological gap.

The other attribute was qualified too: national enterprises were supposed to take over only the product fields not previously occupied by multinational subsidiaries. This strategy, designed to come to grips with economic and technical realities of the computer sector as well as to reduce the policy conflicts, was applied mainly during CAPRE’s rule. As the policy broadened its scope, the “empty space” approach become more difficult to follow, eventually leading to the exit of some multinational companies as producers in Brazil.

2.2. Managing the technological gap: resources, needs and time

An assumption was critical to BIP’s model: that the national enterprises would invest the necessary amounts to absorb the imported technology and develop their own technology. The amount of investment required was considerable, considering that the firms had no previous electronics experience. Such investment involved high risks too – technical, economic and financial – as well as a long time horizon.

BIP’s strategists clearly underestimated the speed and intensity of technological change in the electronics industry and, consequently, the threshold of resources necessary for catching up with the international technological frontier and remaining there, while, at the same time, absorbing the imported technology in other product ranges. 

The threshold was further raised and the nature of necessary technical resources revolutionized by the emergence of the microcomputer and the changes in microelectronics. At the begining of BIP, its policy-makers correctly identified the minicomputer as a window of opportunity and emphasized product-design skills. However, the coming of the microprocessor and of the microcomputer with open architecture transformed the fastest-growing product of the industry into a commodity, placing the emphasis on production and marketing skills. Moreover, such skills changed too. Production moved from batch-assembly to automated mass-production and the main users changed – from large and medium enterprises where purchases were made putting a premium on technical characteristics of the product, to small enterprises and households, where price and user-friendliness were the main considerations.

In other words, the international frontier had not only moved forward but it had also changed its shape and the gap had changed accordingly. However, BIP reacted slowly to this momentous transformation, retaining the emphasis on product-design and not paying enough attention to production techniques and economics and even less to marketing. This was one of the important causes of BIP’s downfall and, below, we return to some possible reasons for this slowness of reaction.

The fact that electronics is a fan-shaped industrial complex, composed by several industries catering for different markets (e.g. data-processing, telecomms, entertainement) but made interdependent by a common technological basis, provided by microelectronic componentes and software, raised the threshold above mentioned as well as the gap between resources and needs. 

SEI explicitly embedded the concept of the industrial complex into BIP but failed to gain control of the technology and industrial strategy of two industries which are critical for the electronics complex – telecomms and durable consumer goods. As a consequence, economies of scope across industries and economies of scale for components and software were reduced. As previously discussed, the insufficiency of BIP’s instruments (mainly import controls) contributed to stymie the development of software and components. As a result of this negative sinergy within the electronics complex, the industries under BIP’s aegis had their costs and risks augmented, especially when compared to other national industries, raising therefore the threshold of resources and the risk-propensity required for closing the gap.

There was very limited evidence of this Schumpeterian behaviour in Brazilian industry, where import substitution had established reliance on imported technology as a norm. The main exceptions to this norm came from State enterprises (e.g. in the oil, telecomms and aeronautics sectors), where there was a politically-driven motivation to technological autonomy and, more recently, from export-oriented private firms. It was no accident that COBRA, the State enterprise, was a leader in technology investments in the computer industry, but it was an outstanding achievement of BIP that the private informatics enterprises at the mid-eighties were investing between 8 and 10% of their sales on technological activities, in a country where, on average, manufacturing industries spent less than 1% of their sales on such activities. It is also noteworthy the fact that national informatics firms allotted about a fourth of their graduate employees to R&D activities, in contrast with their multinational competitiors, which allocated less than 6% of their graduate employees to such activities (Hewitt 1992). 

In order to understand the actual behaviour of the privately-owned electronics enterprises it is worth examining in some more detail the effects of BIP and other policies upon their investment rationale.

The undevelopment of the Brazilian science and technology system meant that local firms had to internalize costs and run risks that, elsewhere, were not privately borne, raising the threshold and risk levels of investment. The timing and intensity of policy instruments was critical. During the first half of the eighties, when the scope of BIP was broadened by SEI, public funding of science and technology declined sharply, cutting short the process of structuring the S&T electronics system started during the seventies. Although this was partially reversed by the Ministry of Science and Technology during its short life-span, when electronics was a priority area, it was not possible to create a critical mass of human and institutional resources. In fact, even top priority projects, such as CTI, were not properly funded. This lack of resources reflected not only budgetary constraints but also the lack of commitment to the policy from the economic ministries¹.

The dearth of externalities was not restricted to the top end of technical skills. Although subsidiaries of multinational firms had been operating in the country for decades, staffed mainly with local personnel, the skills they developed was concentrated on marketing of mainframes, by and large inapproriate to the problems faced by local enterprises striving to desing, produce and market different types of products. 

The same limitation applied to the network of suppliers of parts and components, which was very rudimentary, since the subsidiaries were supplied mainly by imports. As already mentioned, the durable consumer goods industry, which in other countries provided a strong market for locally produced parts and components, was supplied mainly by imports thoroughout the duration of BIP. As a consequence, although imports allowances of CAPRE and SEI privileged parts and components, local firms were obliged to verticalize production, spreading their investments and incresing their costs. A side-effect was to increase the profitability of smuggling parts and components, enhanced by the lack of effective repression. In fact, the main deterrent to smuggling was the moral suasion of the informatics community, which, as previously mentioned, tended to decline over time. 

The same propensity to import, smuggling if necessary, apllied to software, where import controls were, at the same time, cumbersome and ineffective and incentives to local production were scarce. For both microcomponents and software the international standardization of products under way during the eighties led to an increasing pressure to import – legally or otherwise. As a consequence, the attractiveness of local investment in such activities, the core of the electronics complex was reduced.

In other words, differently from other countries, the development of the Brazilian electronics complex was plagued by a vicious circle of a cumulative nature, in which the industries which made up the shafts of the fan (e.g. computers, industrial automation) did not benefit from the economies of scale and scope arising from the hub of the fan (components and software). Such negative sinergy was internalized by the enterprises which, following the industrial complex approach, entered several industries of the complex and which were obliged to follow different and contradictory strategies in their product-divisions or subsidiaries.

To such external factors it should be added the fact that the local enterprises entering the electronics area had no previous experience of the field, suffering thus all the pains and errors of learning. Although most of them had technology licensing agreements, the knowledge imparted by the licensors was limited, restricted mainly to the provision of product designs and specs. In most cases, investing, manufacturing and selling was an on-the-job learning process with all its inherent mistakes. 

To counter the combined effect of such internal and external factors BIP had practically only one instrument – the control of imports. Fiscal incentives were almost negligible and the funding from public development banks was very limited, especially during the start-up phase of the industry, when it was most necessary. Moreover such funding was provided under the form of loans and not as risk capital. The undercapitalization of COBRA, where the technology ethos was strongest, was often remarked but never corrected by its owners, the State banks. State purchasing power, always present in BIP’s documents, as a main instrument for fostering the industry, was used haphazardly and discontinuously, depending on whims and political orientation of the managers of State agencies. Such weakness in terms of instruments and time-horizon was not accidental – it reflected the opposition BIP suffered within the State, which, in turn, reflected the conflicts it evoked in the Brazilian society

The paucity of instruments used by BIP stands in stark contrast with the wide array of policy instruments used in other countries, which ranged from State intensive and extensive funding of R&D and production to State procurement and support in the internal market and abroad. In such countries, the development of the electronics complex was stimulated by measures which, at the same time, reduced risks and costs. Moreover, in other countries the policies for the segments of the electronics complex were convergent and had a long, normally indeterminate, time horizon, reflecting a wider and more permanent social support². 

It is true that the control of imports sheltered the local companies from price and perfomance competition and allowed them to reap considerable margins on the products they sold. If, on the one hand such margins could be justified on the grounds that, in the absence of alternative sources of funding, they were necessary to allow the firms to make their “primitive accumulation” to develop local technological and industrial capability, on the other hand, they attracted rent-seekers, aiming at making the highest possible short-term profits with very limited commitment to the development of local resources. The possibility of smuggling and the limited capability of SEI to control the implementation of the projects it approved (again because of “budgetary constraints”) facilitated the latter behaviour. 

In this context, the imposition of a time-limit to SEI’s power to control imports by the Informatics Law of 1984 probably played a contradictory role: it stimulated both learning and predatory behaviours. However, since the time period was established arbitrarily, as a political compromise between the supporters and opposers of the policy in Congress, bearing no relationship to learning curves, its net effect was probably tilted to making the best of the policy while it lasted.

The behaviour of entrepreneurs is obviously dependent on the development of the market they serve. BIP was predicated upon the assumption that the Brazilian market would provide a basis wide enough to develop the industry. The increasing role played by static scale economies in the more standardized industries was clearly underestimated by policy-makers. As suggested above, this may be partly explained by the technological emphasis on product design. However, political economy factors may have acted too, albeit sometimes unconsciously: if scale economies were a prime criterion for guiding SEI’s decisions to allow the local manufacture of a product, the agency would be led to restrict the number of local suppliers. As a consequence, it would have to meet an inner-circle hostility, from the frustrated newcomers, and more strident accusations from opposers of the policy that it was nurturing an exclusive set of enterprises. As it was, SEI could truthfully contend that it was fostering competition in the industry. 

In fact, a trait that distinguishes BIP from other countries’ informatics policies is the absence of “national champions”. From the outset of the policy, at the time of CAPRE’s market reserve decision, policy-makers were unwilling to commit resources to one or a few enterprises, which explains the fate of COBRA, an obvious candidate to be a national champion. As a consequence, the policy tended to follow a “product design” approach insted of focusing on enterprises and their economics. Such bias was reflected in the staffing of SEI, where there were very few economists and in the procedures of analysis, which priorized technical aspects.

Finally, it is also probable that the estrangement between SEI and the academic community, brought about by the “national security” origin of the former and the ruthless way in which it replaced CAPRE (where the relationship between bureaucracy and academia was very close) contributed to reducing SEI’s capability to properly assess the technical changes under way at the international level³.

Moreover, because of the levels of investment which were required and of the structure of funding of such investment, based mainly on retained earnings and very expensive short-term credit, the internal market had to grow fast – the enterprises being in the position of the cyclist which must keep pedalling in order to stand upright. Indeed, until 1987 the Brazilian market grew at spetacular rates, impervious to the crisis which affected the rest of the economy during the eighties. According to data from SEI, between 1980 and 1986 the earnings of national informatics firms grew 7.4 times – circa 40% yearly average. In 1987 the growth was 14% – a healthy rate for any industry. Nonetheless, as an indication of the financial fragility of the industry, such decline in growth provoked a serious crisis.

An export oriented strategy, eventually developed along selective lines, could have provided some solutions. However, pressures to export were applied only on the foreign subsidiaries, which complied in exchange for grater import flexibility. Although local firms did achieve some export successes in specific market niches (e.g. services automation) the strategy was very inward-looking. Only at the end, in the late eighties, BIP begun to consider exports as a priority, but without any policy instrument to back it up. 

The inwardness of the policy was the result of several, convergent factors. First, the very success of internal sales reduced the incentives to export. Second, the policy itself, by emphasizing local design of products and software reduced their tradability and contained no incentives to export. The scant attention paid to production technology and economics, especially to scale economies, as mentioned above, had important negative consequences for the international competitiveness of the products covered by BIP, especially after the microcomputer revolution made standardization and scale economies prime requisites to international trade. Finally, BIP was deprived of the export basis of other developing countries, especially of Southeast Asia: the durable consumer goods industry. While in the other countries such industry provided the experience of mass-production and international trading for exporting other electronic products, in Brazil the durable consumer goods industry produced limited quantities oriented to the internal market too and used mainly imported components. In this aspect too the sinergy of the Brazilian electronics complex was largely negative.

1987 was a watershed for BIP . In that year coincided a market slackening with the withdrawal of Government support, as shown above. The two events were closely connected. The diminished demand growth of 1987 was largely a result of the failure of the stabilization attempt of the year before (the Cruzado Plan), which led the Brazilian Government to declare a moratorium on external debt payments and to the weakening of the groups which were the strongest supporters of the policy within the Governement coalition. The stepping up of US pressures against BIP, with threats of export retaliations, could not have come at a worst time for BIP. Although BNDES stepped in, providing relief credit for some of the largest firms in dire straits, the support of BIP in the economic ministries and the Foreign Affairs Ministry – never enthusiastic – was then substantially reduced, as discussed above. The combination of macroeconomic foreign exchange restrictions with the low priority attached to the success of BIP led to the already mentioned greater import restrictions of parts and components, stimulating smuggling and undermining the competitiveness of enterprises which adhered to the policy. At the end of the day, macroeconomic interests reimposed their hegemony over sectoral rationale.

As the policy weakened under internal and external pressures, another cumulative process set in: reading a writing on the wall many firms reduced their commitment to BIP’s objectives and rules and started to negotiate partnerships with foreign companies (using as their main asset their hold on the Brazilian market), increased (legally and illegally) imported content, contained investments in R&D, etc. Such actions undermined BIP further and thus stimulated other firms to follow a similar course, straining the relationships between such enterprises and those which still adhered to the policy. As a result, when the tide of liberalization mounted in 1990 BIP was bereft of support even within the informatics entrepreneurial community, with the exception of a few stalwarts which still upheld the banner of technological autonomy. The latter presently have adapted to the new norm or are exploring small market niches.

The speed by which a substantial part of the informatics entrepreneurial community changed gears, throwing their lot with a reversal of BIP, may be viewed as a sign of entrepreneurial capability for interpreting market signals and adapting strategies accordingly. Another interpretation is that many of the entrepreneurs were never really committed to BIP’s objectives, having supported them as an ideological shield under which they were able to gain rents, which were good as they lasted. 

There is probably some truth in both interpretations. Undoubtedly there was a strong component of rent-seeking in the support given by entrepreneurs to the policy. Many paid only lip-service to BIP’s technological objectives, using them as beachead to enter the market. However, it must be recognized that the limited array of instruments BIP was able to muster and its fixed time-horizon were factors, internal to the policy, which stimulated such behaviour or, at least, led the entrepreneurs to keep in mind a policy reversal as a possible alternative. An exclusive commitment to BIP required a Faustian drive to local technological independence for which Brazilian entrepreneurs were never noted. Indeed, one of BIP’s achievements was to reveal that, even under unfavourable conditions, there were entrepreneurs capable of such commitment.

A policy of management of the technological gap using import controls is based not only on assumptions about the behaviour of entrepreneurs: it relies also on assumptions, explicit or implicit about consumers’s reactions. At the very least it must assume that the latter will bear, with grace or by force, the costs of the policy: the restriction of the range of products offered and the higher price/performance ratio of the goods actually supplied.

For the reasons above outlined, consumers, inequivocally and literally, footed the policy’s bill. Opposers of BIP inevitably pointed out consumers’ satisfaction as a main reason to scrape it and, therefore, it is worth considering the argument in more detail. Schmitz and Hewitt (1992) point out that “the diffusion of computers made by national firms has been rapid by any standards. The average annual growth rate of the Brazilian microcomputer market between 1984 and 1987 was the highest in the capitalist worls at 74 per cent” (p.31), showing that the policy, at the very least did not prevent diffusion. The same authors provide a good survey of the evidence on price differentials between Brazil and other markets at the end of the eighties, when the attack on the policy was at its peak: they show that the former were double the US prices and not higher than a fourth of the European prices. Over the eighties, according to the same source, the technological lag in terms of product-vintage between Brazilian and international supply had been reduced too, although it was still significant – at least two years. 

No matter how satisfactory such diffusion and learning performances were from a long-term industrial policy point of view, there was a shift from an attitude of cooperation and tolerance with BIP’s products to hostility and civil resistance, as expressed by the smuggling of finished goods or by purchases of products assembled with aknowledgedly smuggled in parts and components. Curiously, consumer disaffection increased with the shift from mini to microcomputers, although the purchasers of the former were probably better informed than the latter about price-performance characteristics of the international frontier. This is probably attributable to the greater commitment of the technical profession to the policy.

It is also important to notice that there was no noticeable difference in the performance of national firms and subsidiaries of multinational companies. Moreover, similar differentials of price and product vintages were observable in the segments of the electronics complex where BIP was not applicable, such as the consumer goods assembled in the Amazons Free Trade Zone, as well as in other industries structured by traditional import-substitution policies and dominated by multinational firms, such as the automobile and pharmaceuticals industries. Finally, a detailed study of the cost structure of informatics products showed that sourcing was a major cause of the Brazilian higher costs and that import-liberalization would not necessarily bring about a reduction of such costs (ibid.). 

In other words, although electronics did cost more in Brazil, it was not significantly different in this respect from the rest of the industry established in the country and its higher costs could not be ascribed to the specificities of BIP. Nonetheless, this was what precisely happened. The press played a major role in this process by carefully reporting all the inefficiencies of the industry and by, equally carefully, abstaining to report its achievements or to compare its performance to other industries and policies. To give an example, it is widely recognized that the automation of the Brazilian banking system is highly sucessful (Cassiolato 1992). Although millions use it, such success was never related to BIP .

In this way, consumer dissatisfaction, especially of householders, was nurtured, increasing the opposition to the policy. Such opposition, spreading by way of mouth, was probably crucial for the overthrow of the policy. The standards may have been double but they were inequivocally sucessful in depriving the policy of social support. After all, a well-known apothegm of Brazilian politics is that “facts do not count but versions of them”.

The same may be said about State intervention. In order to pay the external debt the other Brazilian industries were subject to import controls similar to those imposed on electronics, managed by another Government agency (CACEX) and it was never explained to the public that the import quota of SEI was not decided by it but by the Treasury. Other industries of the electronics complex, not subject to BIP, depended even more strongly than BIP’s on Government incentives, such as the fiscal exemptions of the Amazons Free Trade Zone or the procurement of telecomms equipment, which was coupled to transfers of technology from the State sectoral research centre. To other industries, dominated by multinational companies, such as the electric power equipment the role played by State intervention was not less important.

We have argued above that BIP’s strategists underestimated the problems facing the policy. Such problems would have been reduced by greater selectivity. For instance, by distinguishing between families of products that would have to be imported for the perceivable future, products that could be manufactured locally but with imported technology and, finally, products to which the policy of technology import-substitution could be applied⁴.

As shown above, BIP started selectively, by concentrating on small and medium computers but soon, driven by the logic of the industrial complex, the scope of the policy was broadened. However, since the policy did not gain the control of the complex, the losses were doubled: sinergy was negative and the available resources were spread too thinly. By the end of the eighties the need for greater selectivity was getting wider recognition – it was a recurrent theme during the preparation of the Second PLANIN. But it was not unanimous: there were those which considered introducing greater selectivity as caving it to the pressures of the US Government and internal opposers of BIP⁵. In fact, one of the results of the pressure under which BIP, in general, and SEI, in particular, were placed during the late eighties was to rigidify the stand of part of the policy supporters, preventing them from making major strategic changes. 

Probably, it was too late, anyway. As discussed in more detail en the following section, more was at stake with BIP than supply and consumption of electronic products and delenda BIP had become a symbol of modernity. Under the prevailing circumstances, no greater selectivity could have saved it.

2.3. Values, interests and social support 

A crucial aspect of BIP was its value content. The history of the policy may be seen as the apogee of a weltanschauung and its decline. BIP’s paradigmatic force, as a model applicable to other industries did not escape anyone – supporters and opposers alike, inside the country and abroad. Such attention was obviously reinforced by the strategic role played by electronics in the modern world. 

BIP strategists assumed that the notions of national autonomy and local creativity would be strong enough to warrant it social support, giving the industry enough time to proceed along learning curves and reach close to an international technical and economic efficiency frontier. Such support involved the legitimacy of the policy instruments – State intervention discriminating against imports and in favour of nationally owned enterprises and of local technological efforts and the social actors (scientists, engineers, etc) responsible for such activities. 

Although a considerable part of the debate surrounding the policy was conducted using “objective” arguments such as the performance and price of the products of BIP, the objectiveness was more apparent than real, since what was at stake in the economic and technical efficiency debate was an income distribution question – how costs and benefits of the policy would be distributed along different social groups over an imprecise time duration – a question which cannot be answered without resort to value judgements. Moreover, underneath such arguments laid unquantifiable values about creativity and power – of local versus foreign agents, of producers versus consumers and of the State versus the market. The fierceness of the debate, which often turned vicious, and the polarization of opinions is indicative of the indivisibility of such values, which rendered impossible a meaningful compromise. 

As shown above, the balance of forces shifted along time and it is worthwhile to try to identify the factors conditioning such movement. Let us begin with the contextual factors, since BIP was a sectoral expression of a pattern of development. As mentioned, at the inception of BIP a similar pattern was apllied to other sectors too. Expansive macroeconomic conditions, a strong State which led the economy and had an assertive international positioning, confidence about the future, supported the assumptions of BIP.

Although the macroeconomic conditions worsened considerably during the first years of the eighties, by 1984 an export-led recovery was under way, lending credence to the view that the crisis had been overcome and that the Brazilian economy was bound to return to its “natural” high-growth course. Optimism was greatly reinforced by the iminent end of the military regime. The approval of the Informatics Bill by Congress fits into this context.

In turn, the debate about the Bill contributed to make informatics a national issue and a symbol of State intervention, focusing the opposition to the latter on BIP. If at its inception the informatics policy was a kind of guerrilla warfare, as aptly described in the literature (Adler, 1986; Dantas, 1988), its institutionalization by Congress turned the conflict into a war of position, with the occupants of the trenches well defined in ideological terms. 

The debacle of the 1986 Cruzado stabilization plan put an end to optimism and the next year ushered in a period in which the Brazilian unatended structural problems claimed their price as an economic crisis still under way. The credibility of the Executive, wich was already weakened by successive and unsuccessful attempts to reduce public deficits, resulting only in a deterioration of the State apparatuses, was finally swept away as the economy bordered hyperinflation. This could not but affect the legitimacy of a policy, such as BIP, which hinged on State intervention. 

Nonetheless, the legitimacy of the values underlying BIP was still strong enough to lead Congress to enshrine in the 1988 Constitution a definition of “national enterprise” closely patterned upon BIP, a characterization of the Brazilian market as a “national asset” and to reserve to the State strategic sectors such as oil and telecomms. However, the same Congress established, against the strenuous efforts of BIP supporters, the maintenance of the privileges of the Amazon Free Trade Zone. In fact, the process of writing up the Constitution led to the consolidation of the conservative forces within Congress and the articles which supported BIP were a compromise with the waning nationalist faction. To all practical purposes, they were the latter’s swan song.

The disenchantment of public opinion with State intervention due to internal reasons was further enhanced by the international tidal wave of neoliberalism which swept the world during the eighties. The media, which had been strenuosly campaigning against State intervention since the late seventies rode high on the wave and singled out BIP as a prime target, inclusive for the conflicts with the U.S.

Much has been made of the US pressure against BIP as a cause of its demise, but its role and strength can be gauged only against the background above described, in which loomed the huge Brazilian external debt. As shown by the case of the nuclear policy during the second half of the seventies, different economic and political conditions made even stronger pressures sustainable by Brazilian Governments. Nonetheless, under the conditions prevailing at the end of the eighties, the US pressure played an important role in the reduction of the Government support to BIP, previously examined.

In short, during the second half of the eighties the macroeconomic and political conditions became highly unfavourable to BIP’s values. The latter were further undermined by the policy’s specific problems of managing the technological gap, previously commented. The behaviour of entrepreneurs and consumers shows that the mores of the Brazilian society were mismatched with BIP’s in at least three critical aspects. First, on the valuation of independence, technological and otherwise. Public opinion, as expressed by consumers, wanted Miami, not the Silicon Valley. Entrepreneurs were quite willing to parlay autonomy of decisions for a safer existence. Second, the time-horizon: although industries take as long as human beings to reach maturity, as shown by other international experiences (e.g. the Japanese car industry), in Brazil there was little patience with BIP. Third, a cavalier approach to the law, as witnessed by smuggling and other corner-cutting procedures. The latter points to a deeper problem, which lies at the heart of BIP – the lack of legitimacy of the State. 

As shown, throughout BIP’s life-span its supportive coalition remained essentially the same: academics, technicians and bureaucrats – social actors who had professions which led them to place high value on independence, especially technological independence. To such coalition were later added the entrepreneurs of the sector, with the misgivings already mentioned. Such alliance, which had to fight against a growing number of oppositors as BIP encroached on a broadening range of interests, could never convince public opinion that it was fighting for the common good. The media said and people believed that it was only a particularist policy – “for some colonels and a few inefficient entrepreneurs”⁶. It was not hard to believe, given the record of particularist policies of the Brazilian State, strongly reinforced during two decades of military regime. In fact, BIP was often branded as an exclusive product of the military, in spite of being one of the very few sectoral policies submitted to Congress. 

The fact that policies alternative to BIP served other particular interests was conveniently disregarded by its opposers. However, the important point is that BIP failed to capture the hearts and minds of the majority of those involved. Therefore, as a result of the general conditions and of its specific problems, at the end BIP was hollowed of most social support. Accordingly, it ended with a whimper, not with a bang. 

3. What is left

The metaphor of the half-full glass, which is also half-empty depending on who looks at it is well applicable to BIP as an industrial and technological policy. It produced a sizable industry with considerable technical capabilities, although far short of its ambitions. Its products, in terms of price, performance and updateness, are similar to those of industries spawned by the traditional import substitution model. It provided jobs which involved the use and development of technical skills wich other industries in the country did not foster but which they now can use. Moreover, it showed that under appropriate circumstances there are some Schumpeterian entrepreneurs in Brazil. 

However, as an ideological construct – as the proof that a more independent pattern of development was both desirable and feasible – in pragmatic terms BIP failed. At the end it had very little social support, no matter if justifiedly or not. As pointed out above, several lessons about the objectives and mores of the Brazilian society can be drawn from such experience, as well as about the relative weight of interests and how they operate, within and outside the State. As for the latter, it is a reminder of its limits and of the penalties attached to hubris.

At the same time, looking again at the glass, BIP was an example of audacity, hope, hard work and honesty – qualities that are not amiss in a society such as the Brazilian. Moreover, since the values which inspired it are not dead, they may find a different form of expression, eventually as a new policy. Opposers of the policy may read this as a threat and think of Dracula but its supporters may find relief and hope in the image of the Phoenix.

(*) Professor of the Faculdade de Economia e Administração of the Universidade Federal do Rio de Janeiro and Director of the Banco Nacional do Desenvolvimento Econômico e Social. This paper was prepared for the project on The Politics of Technology Policy Institutions in Latin America organized by the United Nations University Institute for New Technologies (UNU/INTECH) under the coordination of Dr. Maria-Inês Bastos. I thank Arthur Pereira Nunes, Ivan da Costa Marques and José Guaranys for the time they spent discussing and recollecting facts and ideas about the policy. The final version of the paper benefited from the careful reading by Wilson Suzigan of its first version, as well as from the comments of the participants of the seminar on The Politics of Science and Technology Institutions in Latin America, held at INTECH in April 1993. The usual disclaimer that the article represents strictly personal views obviously applies, but, since the policy was ladden with values, as argued in more detail in the text, it is convenient, from the begining, to “declare an interest”: in my double capacity, as a bureaucrat and academic, I was an active supporter of BIP. Since “objectivity” is not to be confused with “neutrality” I hope I have provided a balanced account of the policy but corrections are all the more welcome.

• F.Erber, Professor of the Faculdade de Economia e Administração of the Universidade Federal do Rio de Janeiro and Director of the Banco Nacional do Desenvolvimento Econômico e Social. This paper was prepared for the project on The Politics of Technology Policy Institutions in Latin America organized by the United Nations University Institute for New Technologies (UNU/INTECH) under the coordination of Dr. Maria-Inês Bastos. I thank Arthur Pereira Nunes, Ivan da Costa Marques and José Guaranys for the time they spent discussing and recollecting facts and ideas about the policy. The  final version of the paper benefited from the careful reading by Wilson Suzigan of its first version, as well as from the comments of the participants of the seminar on The Politics of Science and Technology Institutions in Latin America, held at INTECH in April 1993. The usual disclaimer that the article represents strictly personal views obviously applies, but, since the policy was ladden with values, as argued in more detail in the text, it is convenient, from the begining, to “declare an interest”: in my double capacity, as a bureaucrat and academic, I was an active supporter of BIP. Since “objectivity” is not to be confused with “neutrality” I hope I have provided a balanced account of the policy but corrections are all the more welcome.