New Departures for Technology Policy in Brazil

Jörg Meyer-Stamer, German Development Institute, E-mail jms@cs.tu-berlin.de

Published in: Science and Public Policy, Vol. 22, 1995, No. 5, pp. 295-304


This article(1) starts with the notion that there is no national system of industrial innovation in Brazil. What has emerged in the era of import substitution was hardly system and limited innovation: Apart from a few isolated cases, there was little interaction between universities and technology institutes on the one hand and industrial firms on the other; and the firms themselves could prosper on the basis of limited innovative efforts. This, I will argue, has to do with the incentives that both sides were facing in the import substitution era.

I will then outline what role technology policy can play in the period of transition to a more competitiveness- oriented industrial development model. I will discuss different approaches to and instruments of technology policy. I will then discuss a number of questionable assumptions that are widely held in the Brazilian technology discussion and that make the transition to a more effective technology policy difficult. Finally, I will elaborate a number of policy proposals as well as address the question of an appropriate governance structure (between government and societal actors, but also between different levels of government).

1 Why there is no national system of innovation in Brazil

In the past, national development policies as well as development cooperation have been based on the notion that one key ingredient for industrial development was the existence of a science and technology infrastructure, and that science and technology institutes and industrial firms would somehow automatically join their hands in the development of technology. This notion was based on the experience of advanced industrial countries, and the experience of countries such as the East Asian NICs has underlined the fact that there is a certain truth about this notion. However, in most developing countries policy makers failed to notice two points. First, institutes that try to conduct basic research often fail to contribute to technological development. Second, there is no automatism at all that leads to the establishment of close relationships between research institutes and industrial firms. Whether or not this happens is largely dependent on the incentives that actors on both sides face (2). The fact that some actors in some places have deliberately ignored the prevailing incentive structure (i.e. have put a lot of effort into moving close to the technological frontier without pressure that forced them to do so) has sometimes confused the observers as to what these incentives really were.

In the particular case of Brazil, the prevailing incentive structure first of all did not stimulate innovative behaviour inside firms, at least not in the way it is normally understood in industrialized countries. This is not to say that Brazil was a stationary economy with no technical change at all; the contrary has been documented (eg Sercovitch 1984, Teubal 1984, Dahlman and Fonseca 1987). However, the larger part of technical change followed the typical pattern of inward-oriented latecomer industrialization where firms try to master technologies that have been developed elsewhere, try to adapt them to their needs and possibly try to improve them so that they better fit into the local environment; like in other Latin American countries 'typical R&D efforts would be determined by the need to use different raw materials, scale-down (to smaller) plant size, diversify the product mix, change the product design, use simpler, more universal, less automated and lower capacity machinery, stretch out the capacity of existing equipment, etc.' (Teitel 1987, p. 109). This kind of activity does not particularly require intra- company R&D but rather process engineering activities.

In the closed market environment, firms felt little pressure to extend their technological effort beyond incrementalism. As the competitive pressure was low, they had little incentive to introduce truly innovative products or to look systematically for radical improvements in the production process. There were only few exceptions, like the aircraft industry where a Brazilian firm tried to compete at the leading edge in the world market for commuter planes (Frischtak 1992), or in oil exploration where the internationally available know-how on deep water exploration apparently was limited (Hawrylyshyn 1991), or in the production of alcohol fuel that simply did not exist anywhere else. In these cases, firms have set up a systematic R&D effort and have also cooperated with universities and research institutes. Thus, some sectoral innovation systems have emerged. However, the claim that there is a national system of innovation (3) is misleading, at least in the sense of term introducted by Freeman ("The network of institutions in the public and private sectors whose activities and interactions initiate, import, modify and diffuse new technologies may be described as 'the national system of innovation'", 1987, p 1). In fact, Freeman's considerations are illuminating for an evaluation of the Brazilian experience:

"The rate of technical change in any country and the effectiveness of companies in world competition in international trade in goods and services, does not depend simply on the scale of their Research and Development and other technical activities. It depends upon the way in which the available resources are managed and organised, both at the enterprise and at the national level. The national level of innovation may enable a country with rather limited resources, nevertheless, to make a very rapid progress through appropriate combinations of imported technology and local adaptation and development. On the other hand, weaknesses in the national system of innovation may lead to more abundant resources being squandered by the pursuit of inappropriate objectives or the use of ineffective methods." (Freeman 1987, p 3)

In the case of Brazil the numbers point to the fact that it was the misallocation rather than the shortage of resources that restricted technological dynamism. In 1976/77, there were 1,050 firms that included R&D expenses in the tax declarations. This number shrank to 780 until the end of the recession of 1981/83 and then again grew to 1,090 in 1985. The average R&D/turnover ratio of these firms was only 0.4% (Dahlman and Frischtak 1993, p. 425). No less than 62.5% of the R&D expenses came from state companies, eight of which accounted for more than half of the total. The low R&D effort on the side of private firms is being confirmed by look at the data on research personnel. In 1986, the clear majority of the 52,863 persons worked for the state 62% in universities, 20% in technology institutes, 3.4% in state firms and 6.1% in other government bodies. Further 6.5% worked for private universites, which leaves 1.9% for private firms. This points to the fact that the number of firms who are seriously and systematically investing into R&D is far lower than 1,000; Brazilian observers have offered estimates that range from more than 200 to 366 (4). The true R&D/turnover ratio may have been as low as 0.16% in the early 1990s (5).

Most Brazilian firms did not have R&D departments. They are an important feature in the organization of R&D cooperation with external agents since they form the major recipient structure for the internalization of research results from external sources. In fact, as a rule Brazilian firms did not confront problems, challenges, and opportunities that would have made it necessary or advisable to look for research results or external research support. Therefore, researchers in universities and research institutes had to have the impression that there was little potential demand for their results in industry. Yet, this was only one reason for the clear separation between research and industry. Other reasons were (6)

the easy availability, particularly in the 1970s, of research funds from public financing institutions which gave the emerging scientific community a lot of space to define research priorities according to their personal interests;

the research ideals that PhDs returning from abroad brought with them, leading them to strive for academic reputation rather than application of their researach results; they often also oriented their research so that it fit into the line of activity of their foreign alma mater rather than domestic priorities;

a certain degree of introspection that emerged inside the science system since the early 1970s, where communication mainly took place inside a circle made up by people from research institutes, universities and public funding agencies; reputation and career perspectives were dependent on the peer evaluation inside this circle, not on things like successful cooperative links with firms.

By and large, the environment for research activities during the 1980s deteriorated not as much as some observers have stated (eg Bastos 1993). In fact, after decreasing in the crisis of the early 1980s, the government R&D allocations grew by 50% between 1984 and 1987 and remained high until 1990 when it started to drop again (Figure 1). In the second half of the 1980s, the core funds for S&T development (CNPq, CAPES, FINEP, PADCT, FNDCT) (7) remained reasonably stable. Nevertheless, there can be little doubt that the framework conditions for research became less friendly. Even if R&D funds did not decrease, they became unpredictable as the federal government tried to reduce its deficit by delaying the allocation of all sorts of funds (Franco 1993). Moreover, as budget allocations did not keep up with the growth of the science system, the competition between different research groups became intense; apparently, it moved from healthy rivalry to a situation that was marked by mutual mistrust and open hostility (Castro 1989). This implies that there is not only no national system of innovation with a dense tissue of relationships between science and technology institutes and the productive sector. Possibly even the science system does not really deserve this name since the term 'system' suggests the existence of horizontal links and interaction.

On the whole, the case of Brazil supports the notion developed elsewhere that what has been labelled the national system of innovation should rather be called national system of institutions since it is the broader framework conditions that shape the way innovations are, or are not, taking place in a given economy (8). Due to the institutional framework of import substitution it was perfectly rational for Brazilian firms and reseachers not to strive for the construction of innovation networks.

2 Technology policy in Brazil: Old approaches to a new situation

In 1990, the government has radically changed the framework conditions for industrial development. Acknowledging that the import substitution model had run into a dead end, it opted for a policy of gradually opening the market to foreign competitors, thus creating an environment that requires international competitiveness and thereby forces companies to attain international levels of quality and efficiency. This has been accompanied by a number of technology and industrial policy programmes (9). However, they were either not implemented, or only after long delays, or have had little impact so far because the recession inhibited private sector investments. Even the Quality and Productivity Program (Programa Brasileira de Qualidade e Produtividade, PBQP) that has pursued an innovative approach (mainly trying to build a consciousness for quality issues inside firms) and got a lot of publicity in Brazil apparantly has only had a limited impact (PBQP 1992).

Brazilian policy makers and researchers tend to explain the limited effect of technology policy initiatives with the economic crisis and the low investment propensity of companies. I will argue that some widely favored policy initiatives have limited prospects to succeed even under improved economic conditions. This is due to certain structural features of the Brazilian economy and science system; I will discuss the cases of fiscal incentives and, again, university-company-links. I will point at some implicit technology policy measures that undermine the effects of the explicit technology policy.

2.1 Fiscal incentives

Brazilian policy makers perceive fiscal incentives as a major, if not the central technology policy instrument. Throughout the 1980s, the proponents of the informatics policy have unsuccessfully tried to establish a system of fiscal incentives; and the industry and technology policy projects of the Collor government put heavy emphasis on fiscal incentives. Policy makers are pinning great hopes on a law, passed in 1993, that introduces fiscal incentives for firm R&D (10). However, there are some major problems with the use of fiscal incentives in technology policy.

1) From a strictly economic point of view, the justification for fiscal incentives for R&D is much more linked to R than to D. In industrialized countries, innovative behavior, i.e. a high development effort, is simply a basic feature of any competitive strategy, and this needs not be stimulated by fiscal incentives (even though they exist in some countries); rather, companies simply will drop out of the market if they do not undertake this effort. In Brazilian companies, the larger part of R&D outlays is in development. It reflects the specific characteristics of the Brazilian industrialization experience that policy makers consider to stimulate 'normal' behavior, i e systematic product development efforts, with fiscal incentives.

Contrary to D, fiscal incentives for R are justified. Research activities in companies generally generate externalities. One can even easily imagine a situation where the externalities are much larger than the direct benefits to the company. Therefore, incentives for R are not only a compensation for externalities. They may stimulate companies which hesitate to enter into R because of high risk and uncertain benefits. However, it must be pointed out that, contrary to the proposition of neoliberal economists who usually recommend this, the support for basic research definitely discriminates since different industrial branches to a very different extent pursue activities that are formally being counted as research. The formally accounted research and development expenditures are high, for instance, in the pharmaceutical industry (8.7% of turnover) and low, for instance, in the pulp and paper industry (0.3%) (11). However, it is certainly wrong to assume that the pharmaceutical industry is being 29 times more technology intensive than the pulp and paper industry. Rather, the difference is due to the fact that a considerable part of the innovation effort in the pulp and paper industry is part of ongoing activities in process engineering that are not separately counted as R&D.

2) A practical point is linked to the fact that the discipline of Brazilian companies to pay their taxes is fairly limited. There exist various semi-legal means of tax evasion and a general habit to exploit them (World Bank 1994). The saying goes that even leading companies locate some of their activities in the informal part of Brazil's economy. Thus, there is some reason to doubt that Brazilian companies do actually consider fiscal incentives as a stimulus for any kind of activity, let alone innovation. Moreover, it may be assumed that fiscal incentives for R&D only add more complexity to an already overly complicated tax system so that firms perceive them more as part of a more general nuisance rather than as a stimulus.

3) Fiscal incentives that refer to income taxes by definition benefit only those companies which are profitable. However, they are not necessary those which put most effort into R&D. Under the prevailing conditions, those companies which are most profitable are more likely to be most inventive in terms of financial management. Therefore, income tax related incentives will have a very limited effect.

4) Fiscal incentives are not necessarily maintained for a long time period. This can not only be expected due to the erratic macroeconomic management that has prevailed in Brazil in the past. Moreover, the actual costs of R&D incentives are unpredictable (OECD 1994); if lawmakers perceive that these costs become to high, they may be tempted to withdraw the incentives. Economic agents may expect this and therefore will not necessarily change their long-term strategies because of fiscal incentives.

2.2 University-company links

Brazilian policy makers are pinning great hopes on initiatives to stimulate stronger links between universities and companies. Unfortunately, the prospects for this kind of venture are remain limited even under the new economic framework conditions. There are various problems on the supply as well as the demand side.

Supply side. The quality and capability of the universities vary widely. The number of high quality universities is limited. In many federal and most state and private universities the quality of the staff is comparatively poor (Castro 1989). One should not place too many hopes on their ability to give substantial support to private businesses.

Moreover, university researchers apparently show a tendency to direct their research along the lines of research in developed countries. Research activities in universities in developed countries may be influenced by technological problems of companies; but this is not necessarily so, and even if it was it would be the problems of First World companies that compete at the leading edge of technology. However, competing at the leading edge of technology is something that only very few Brazilian companies do.

Then, there remains the question to what extent there are incentives for universities, especially for leading universities or scholars, to enter into cooperative ventures with private companies. Compared to other areas of public services, especially the rest of the education system, the financial situation of the universities has not been too bad, and after the downturn in the early 1990s the situation is improving again, not the least due to foreign funding, especially from the World Bank and the IDB. Paradoxically, foreign aid and funding agencies which are keen to support the Brazilian science system may in the end inhibit stronger university-company links. Foreign donors will predominantly turn towards leading Brazilian scholars with an international reputation and international contacts. These scholars may even find themselves in the comfortable situation of being able to choose between different offers of foreign funds. This certainly will not stimulate their efforts to find funding from domestic companies.

Demand side. It is now well documented that the traditional waterfall-model of the diffusion of innovation, from basic research to product development, does not at all match the reality. Rather, there is a interactive process of innovation (Rosenberg 1990, Pavitt 1991). This means that in many cases the waterfall flows bottom-up technical problems in product development stimulate applied research, bottlenecks in applied research stimulate basic research. For instance, research in solid state physics has been stimulated by technical problems of microelectronics companies. More often than not, the physical or chemical fundamentals of a new technology are not well understood when it is being used in radically new products; it is only afterwards that basic research starts to investigate into them.

Moreover, there is evidence that some industries are much more inclined to maintain close contacts with research institutions than others. According to Pavitt's typology (1984), this is most likely to happen in science-based industries like electronics and chemicals. In scale-intensive industries and among specialized suppliers, technological advances are typically generated in-house, or stimulated by users or suppliers. Manufacturers of consumer non-durables rely on suppliers for technological advances.

With a view at Brazil's industrial structure, both findings have important implications for attempts to link university research and companies' product development. There are few high-tech, science-based industries in Brazil the electronics industry has transformed itself into a final assembly industry after the removal of high import barriers (Doria Porto 1993), the pharmazeuticals industry is largely based on reverse-engineering, and biotechnology and genetic engineering are only incipient (Galhardi 1994). Thus, by far most Brazilian companies will not encounter problems that can only be resolved by scientific research. They are facing problems for which solutions are available in other companies and countries. In order to adapt these solutions to their particular circumstances, Brazilian companies will have to put more effort into technological learning and capacity building. There is little reason to believe that universities are of much help in this process (except for the education of engineers). Only in special cases universities are good in scanning the application-oriented technological know- how that is available world-wide, and there is little reason to believe that a university researcher should be better in adapting transferred technology than a company engineer; special cases are higher education institutions which have a very narrow disciplinary focus and exceptionally good contacts with industry anyway. Thus, as a rule the stimulus for firms to enter into joint activities with universities is very low. Looking at things in this way, it is not at all surprising that university-company-links are weak in Brazil.

2.3 Implicit vs. explicit technology Policy

The government's explicit technology policy initiatives are often at odds with the implicit technology policy, i.e. the effects of other policies on technological activities in the companies. Some policies produce factors that affect all firms, others favour medium and large firms and thus discriminate against micro and small firms. (It should be mentioned here that the prevailing notion in Brazil is that a positive correlation exists between firms size and technological effort (12), a notion that is not generally supported by evidence from other newly industrializing and industrialized countries [Acs and Audretsch 1992]). In particular, the following three factors are detrimental to technological development:

Economic instability. An important consequence of the unstable, unpredictable economic environment of the 1980s and early 1990s is a general short-term orientation of companies (World Bank 1994). In Brazilian terms, medium-term means one month, long-term one year. This environment deters technology capability building which is generally a long-term venture.

Taxation. Inter-company transactions are subject to relatively high, cumulative federal and state taxes (World Bank 1991). This means that vertical disintegration and specialization are being punished by taxation policy, and vertical integration is stimulated. This makes it difficult for companies to direct their technological efforts towards certain promising areas. Instead, companies will tend to spread their already insufficient technological effort over a broad spectrum of activities.

Neglect of the education system. Nowadays, it is a generally accepted fact in Brazil that the education system, especially in primary education, is in a poor state (eg Ribeiro 1993, Gomes- Neto and Hanushek 1994). Drop-out- and repetition-rates are high, payment, morale and qualification of teachers is low. Therefore, the average youth will be insufficiently educated. More specifically, there will be a distribution between a small number of fairly good school graduates (who will typically opt for secondary and tertiary education) and a very large number who present an insufficient level of literacy and basic mathematical-technical understanding. It seems plausible to assume that this uneven supply structure is matched by an equally uneven demand structure, where better paying medium and large enterprises are in a privileged position to choose among the better qualified school graduates. Therefore, we may assume that the sad state of the education system introduces another factor that disfavors micro and small enterprises, making it difficult for them to achieve an adequate level of technological sophistication and competitiveness.

2.4 Some questionable assumptions of the Brazilian technology policy discussion

The transition to more promising technology policy initiatives is made difficult by a number of basic assumptions that are widespread in the Brazilian discussion. Many of these assumptions, however, are highly debatable.

Hierarchic vs. heterarchic governance. By and large, Brazilian policy makers and scholars have perceived technology policy as an activity that is best left to the federal government; some are still dreaming of a competent, strong, insulated state apparatus. For three reasons this notion is misleading. First, currently the capability of the central government to formulate policies is limited (Holanda 1993), and its ability to use any kind of financial instruments is even more limited due to the permanent budgetary crisis. Therefore, proposals for comprehensive policy packages are inappropriate. A report by a special commission of the Brazilian Congress published in 1992 provides an instructive example (13). The problem is not that the policy proposals as such are inappropriate; some are (especially those calling for a continuation of 'market reserves'), but most are perfectly reasonable. The problem is rather that the report contains nine and a half narrowly printed pages with all sorts of policy proposals, but lacks any clear prioritization.

Second, the complexity of economy and society has increased, and so has the level of sophistication of societal actors. The situation today is much different from that in the fifties and sixties when the planning capability of the central bureaucracy may have been higher than the capability of any other group of society. Today, many companies and associations are at least as well informed and technically competent as the government bureaucrats. Furthermore, the government's ability to stimulate any kind of activity by any group of society (and especially by companies) is severely hampered by the mistrust of or even hostility toward the central government that have emerged over the last years among societal actors. Therefore, it is not likely that the traditional model of hierarchic governance will work in the future. Rather, a heterarchic model, i.e. a model where various governmental and non-governmental actors interact, will have to emerge (Meyer-Stamer 1995).

Third, state and local governments naturally have a role to play in technology policy. They are in a much better position to identify the bottlenecks and the specific needs of companies and to implement specific policy measures. However, they have hardly done this until very recently; activities like the state of São Paulo's efforts to stimulate technology transfer between state universities and institutions the Institute for Technological Research (IPT) and industry have been the exception rather than the rule. The Congress report mentioned above does not even mention the possibility of decentralization in technology policy.

Regional development vs. development of regions. Brazilian policy makers are skeptical about the decentralization of tasks and responsibilities because this might widen the gap between more and less advanced federal states. Given the dominant position of the state of São Paulo, this concern should not be neglected. However, so far redistribution schemes on the federal level have been much more successful in creating individual fortunes than societal welfare. In order to reduce the economically dominant position of São Paulo, it makes much more sense to encourage other states, especially in the South, to develop on their respective potential. The recent experience in the northeastern state of Ceará seems to indicate that even under adverse conditions a lot of space for initiatives exists and substantial improvement of economic conditions can be achieved on the state level (Tendler 1994).

Technological autonomy vs. technology capability building. At times, the technology policy discussion still follows the traditional lines of dependency argumentation (14). A number of actors continue to advocate a policy that builds technological autonomy, i.e. a local base that reduces the country's dependency on technology imports (and the problems of access and of predatory pricing that may be linked to them). Although terms like quality, productivity and competitiveness are fashionable today, the notion of a strong interrelation between world market integration, specialization, competition and technological learning is not widely supported.

Technology imports vs. technology capability building. Although technology imports (including capital goods imports) have continued to be low even after the reduction of customs tariffs, Brazilian company executives seem to perceive technology purchasing as the superior alternative to in-house efforts (Fleury et al 1990). They often fail to notice that technology purchases and in-house efforts in technological learning are complementary activities. The full potential of technologies which are new to a company can only be exploited after extensive adapting of the technology as well as the company's organizational structure.

High-tech vs. high-performance industrialization. Many Brazilian policy makers and scholars are fascinated by fancy high technology. They tend to perceive entire industrial sectors as either high or low technology, equating this with high or low growth potential. They have not noticed that most industrial sectors feature low as well as high technology segments (and in fact Brazilian industry tends to operate in most sectors, including computer equipment, in the low technology segment); and that many so-called low technology sectors (eg toys or furniture) have a stronger performance in the world market than some high technology sectors.

The role of multinational companies. The multinationals who have affiliates in the country are often still perceived as predators who just benefit from the local economy without giving technology in return. It has not come to the mind of many Brazilian actors that this behavior is not due to the very nature of multinationals; rather, it was exactly the import substitution policy that allowed the multinationals to sustain a low level of technological sophistication (Gonçalves 1993). Thus, they behaved in no way different from national firms; and just like them, they are targets of the whip of foreign competition that has been introduced by the opening policy.

3 New approaches to technology policy

In the process of transition from import substitutions, many countries have chosen to refrain from technology policy deliberately, as they believed in the neo-liberal laissez-faire gospel (15), or by default, since they lacked the capacity to formulate and implement it (16). Brazilian policy-makers seem determined not to follow these examples. Yet they face the problem that the fundamental change in the industrial development model also requires a fundamental change in technology policy. A new approach to technology policy requires a new underlying concept; I will outline some possible elements in the first part of this final section. Furthermore, there are a few clear priorities when it comes to defining concrete measures; I will discuss them in the second part.

3.1 Framework conditions for technology policy

With the end of import substitution one of the most importants obstacles to the emergence of a national system of innovation has been removed. Another important obstacle has been addressed with the macro- economic stabilization policy that was launched at the end of 1993. The first measure increased the pressure on industrial firms to become more competitive and thus extinguished the possibility of idiosyncratic technological behavior, the second set out to create the conditions for firms to adopt a longer-term view in terms of corporate strategy, including the strengthening of technological capability on the firm level. However, other obstacles remain which will have to be addressed in the near future.

Redefine the division of tasks between the Union, the federal states and the municipalities. Given the weakness of the central state, the increasing competence of other societal actors, and the advantages of decentralized policies, a redefinition of the division of tasks between the Union, the federal states, and the municipalities is an important and urgent issue in Brazil. This does not necessarily require an unilateral act of the central government or a reform of the constitution. Rather, the fact that the central government has a limited capability to deal with the problems of the country, be they current or structural, opens a window of opportunity for state and local governments.

In fact, in most industries the regional or local level is the most adequate level for supportive state action; in most industries competitive advantage is highly localized (Porter 1990). Thus, there exist many possibilities for state and local government to formulate technology policy measures jointly with societal actors, in particular private business. There is also space for private business associations to set up technology policy initiatives. In some cases, this requires the change of laws on the federal level. Many, however, can be adopted by state and local governments even under current circumstances.

Yet it should be pointed out that the uncertainty of local and regional actors about possible interventions and policy initiatives by the central government may actually undermine the propensity to formulate local or regional technology policies. The central government should therefore clearly restrict its range of activities to those areas where local initiatives are complicated (eg the support of newly emerging industries, or the formulation of programmes for branches that are scattered all over the country), and it should seek to encourage and stimulate local and regional initiatives.

Stimulate stakeholder's dialogues. Another great but nevertheless unavoidable challenge is a new approach to policy formulation (Esser et al 1993). Until today, hierarchic governance is the prevailing governance modus in Brazil. The attempt of the Collor government to introduce fundamental changes without any significant effort in consensus building was the rule rather than the exception. Any effort to build something like a national consensus on the federal level, involving all important actors, appears like a Herculinean task; probably, a much deeper crises is required to facilitate anything of this kind. At the moment, is does already appear as a major challenge to mount an effort in consensus building at the municipal or state level.

Nevertheless, it should be worth the attempt. State governments might consider to introduce incentives that stimulate consensus building attempts at the municipal level, e.g. by financing the hire of non-local persons as moderators. External moderators who do not have any stakes in local politics may play an important if not crucial role in stimulating and managing such dialogues.

Build advanced factors. Technological advance is a means to enhance competitiveness; and competitiveness rests in the terminology of Michael Porter (1990) on the capacity of business associations and the state to provide advanced factors, i.e. an infrastructure that is particularly aimed at necessities of local industry (especially in terms of communication, transport, and energy supply); technology consultancy and extension services; vocational and technical schools, etc.

It is important to point out that these are only partially public tasks. Business associations can play an important role in this field (as they have done in Brazil in vocational training), eg by fostering company modernization programmes, by diffusing information about new technologies and new organizational techniques, by stimulating interactive learning between companies, and by stimulating information exchange about export markets and export techniques. Private consultancies are another important actor, especially in the diffusion of technological and organizational know-how.

Since the technological gap between average local firms and local leading firms, not to mention global leaders, is big, the build-up of technology extension services is a primary task. It may be worthwhile to re-evaluate the experiences with agricultural extension services which seem to have worked reasonably well in the past (it may be mentioned here that the recent U.S. discussion on industrial technology extension draws heavily on the experiences with agricultural extension). In industry, extension services should only maintain a small administrative and technical apparatus. Their main function is to work as a mediator that organizes the contact between companies and consultants, university experts, other companies, retired executives and other knowledgeable persons.

The build-up of advanced factors is a particularly important task in those areas where static comparative advantages abound. Brazilian companies show a certain tendency to rest on static advantages; e.g., they rely on cheap inputs and put little effort into the improvement of production processes to make the best of these inputs. Some so-called traditional industries could perform much better if they changed this pattern.

Building advanced factors may turn out to be a self-financing effort. Over the last years, federal states have increasingly introduced tax exemption schemes in order to attract new investors; and municipalities have introduced levy exemptions in order to achieve the same goal (Lasmar 1994). The obvious result has been a race to lower taxes and levies, thus reducing the available resources for states and municipalities. This tends to weaken locational factors, as for instance resources for the maintenance of the physical infrastructure are lacking. If, on the other side, states or municipalities succeeded in the build-up of advanced factors (financing this e.g. by restructuring the budget), these factors will appear much more attractive for firms than tax exemptions. Thus, states and municipalities may substitute tax exemptions by advanced factors.

Stimulate the participation of trade unions in dialogue and modernize capital-labour relations. To a large extent, innovation means incremental innovation. Incremental innovation is a result of continuous efforts to improve production processes and products. This is not exclusively nor even predominantly a task of engineers and technicians. Rather, it requires conscientious, responsible behavior of qualified workers. The autocratic, Taylorist organization patterns that prevail in Brazilian industry do not facilitate this kind of behavior (Carvalho 1993). Thus, the modernization of capital-labour relations is an important prerequisite for the modernization of industry. The participation of trade unions in consensus building can help in this process.

3.2 Specific measures

Stimulate the emergence of networks of enterprises. Specialization is a key to competitiveness. Specialization is not compatible with a high level of vertical integration. Reducing vertical integration requires the establishment of close relations between companies. Close relations rely on trust. This can be strengthened or weakened by the local culture and political structure (Storper 1993). Efforts to build a political consensus that include private business may thus have the important side effect of encouraging the build-up of a denser network of inter-enterprise relations.

Furthermore, the tax system ought to be changed so that it no longer deters transactions between firms.

Stimulate R, not D. The adequate way of stimulating innovative product development in companies is to put them under competitive pressure while at the same time supplying them with the advanced factors they need. Fiscal incentives cannot substitute for this. Fiscal incentives may be considered to support genuine research activities in companies; however, due to the problems mentioned alternative forms of support, for instance project-based financing, may be preferable.

Introduce incentives for universities to strengthen links with companies. Even if the universities can only play a limited role in the technological modernization of Brazilian industry, any existing potential should be used. This potential may exist in fields like ad-hoc-consultancy, that is in helping firms solve one limited problem at a time, or in areas like metrology and quality assurance where the existing institutional structure apparently leaves much to be desired (Ramos 1990). Increasing the involvement of the universities may require a change in their incentive structure, especially in terms of financing.

Strengthen the links between universities and vocational schools. Universities may play an important role in another field of technological upgrading, i.e. in the upgrading of the vocational training system. The level of education of teachers in vocational schools is often limited. Universities should offer special courses that provide special training and upgrading of the qualifications of vocational school teachers.

Do not pin too much hope on multinationals. Brazilian actors do not only have a certain tendency to fear multinational companies. They also show a tendency to pin a lot of hopes in terms of industrial modernization on them. Unfortunately, there is little evidence to support this hope. Multinationals will not modernize Brazil as they modernized Mexico; the latter case was special and can only be explained by the geographic location of the country. As a base for exports to the world market, Brazil's attractiveness is limited. It is somewhat better positioned regarding the regional market; but the regional market is largely identical with Brazil's domestic market. As long as this does not change, those multinationals already present will hardly feel inclined to mount heavy modernization efforts; and those which are not present will prefer to wait and see.

In fact, it will probably work the other way around: Once the Brazilian actors will have succeeded in getting their economy on track again, multinationals will start to invest in Brazil in order to benefit from a striving local market, and in order to participate in the exploitation of advanced factors.

Footnotes

(1) The article is based on interviews that I conducted with Brazilian policy-makers and researchers in May 1993 and June 1994. For critical comments on an earlier draft I am indebted to Maria Ines Bastos, Ruy de Quadros Carvalho, Joao Carlos Ferraz, Afonso Fleury, Hans Mathieu, Hubert Schmitz, and an anonymous referee.

(2) The problems have already been observed, for instance, by Sagasti (1978). However, he did not trace this back to the prevailing incentives but rather linked it to some policy failure.

(3) This has been claimed, for instance, by Frischtak and Guimarães (1993), Castilhos (1992) and Villaschi (1992).

(4) Marcovitch (1990), p. 106, and Dahlman and Frischtak (1993), p. 426, respectiveley.

(5) Calculated from data in Villaschi (1992, p 53) and UNIDO (1993, p A-18).

(6) See Castro (1989) for a detailed discussion.

(7) For a detailed discussion of the history and profile of these agencies see Dahlman and Frischtak (1993).

(8) This has been tentatively suggested by Nelson and Rosenberg (1993) and Nelson (1992) and more explicitely been formulated by Soskice (1994).

(9) For an overview see UNIDO (1992a, 1992b).

(10) 'Incentivos fiscais: Investimentos em P&D devem crescer muito em cinco anos, diz ministro', Gazeta Mercantil, 18.07.1994.

(11) data for OECD countries; see OECD/TEP (1992), p. 32.

(12) A special commission of the Congress even proposed a pro-trust- rather than an anti-trust policy in order to create large firms capable of financing R&D (CPMI undated, p 204. The commission conducted its hearings in 1991 and presented the report in 1992).

(13) CPMI (undated).

(14) See, for instance, CPMI (undated), in particular p 203 ff, paragraphs 13, 24, 28, and 41.

(15) For instance, this apparently was the case in Mexico and Argentina (Perez-Nunez 1994, Nun 1993).

(16) This applies to countries with severe structural deficiencies like Bolivia (Messner 1993) or Tanzania (Gocht and Meyer-Stamer 1993).

References

ACS, Z.J., und AUDRETSCH, D.B. (1992). Innovation and Technological Change: The New Learning. Berlin: Wissenschaftszentrum (FS IV 92-5).

BASTOS, M.I. (1993). State Autonomy and Capacity for S&T Policy Design and Implementation in Brazil. Maastricht (mimeo).

CARVALHO, R.Q. (1993). Projeto de Primeiro Mundo com conhecimento e trabalho do Terceiro? Estudos Avançados, No. 17, pp.35-80.

CASTILHOS, C.C. (1992). O sistema brasileiro de inovação: uma proposta de configuaração. Ensaios FEE, Vol. 13, No. 1, pp.88-114.

CASTRO, C.M. (1989). What Is Happening In Brazilian Education. In E.L. Bacha, H.S. Klein (Eds.), Social Change in Brazil, 1945-1985. The Incomplete Transition (pp. 263-309). Albuquerque: University of New Mexico Press.

CPMI (undated). Causas e dimensões to atraso tecnológico. Relatorio final. Brasília: Comissão parlamentar mista de inquérito.

DAHLMAN, C.J., und FONSECA, F.V. (1987). From Technological Dependence to Technological Development: the Case of the Usiminas Steelplant in Brazil. In J.M.Katz (Ed.), Technology Generation in Latin American Manufacturing Industries (pp. 154-182). Basingstoke: Macmillan.

DAHLMAN, C.J., und FRISCHTAK, C.R. (1993). National Systems Supporting Technical Advance in Industry: The Brazilian Experience. In R.R.Nelson (Ed.), National Innovation Systems. A Comparative Analysis (pp. 414-450). New York, Oxford: Oxford University Press.

DORIA PORTO, J.R. (1993). Competitividade do Complexo Eletrônico. Estudo da Competitividade da Indústria Brasileira. Campinas (mimeo).

ESSER, K., HILLEBRAND, W., MESSNER, D., und MEYER-STAMER, J. (1993). Systemic Competitiveness - Key Policy Issues. Berlin: German Development Institute (mimeo).

FLEURY, A. et al. (1990). Panorama Atual e Perspectivas da Indústria de Equipamentos para Automação no Brasil. São Paulo: Fundação Carlos Alberto Vanzolini.

FRANCO, G.H.B. (1993). Brazilian Hyperinflation: The Political Economy of the Fiscal Crises. In M. D'Alva G. Kinzo (Ed.), Brazil - The Challenges of the 1990s (pp. 24-35). London: British Academic Press.

FRISCHTAK, C., und GUIMARÃES, E.A. (1993). O Sistema Nacional de Inovação. V Forum Nacional, São Paulo, 3.-6.5..

FRISCHTAK, C.R. (1992). Learning, Technical Progress and Competitiveness in the Commuter Aircraft Industry: An Analysis of Embraer. Washington: World Bank (Industry Series Paper No. 58).

GALHARDI, R.M.A.A. (1994). Small High Technology Firms in Developing Countries. The case of biotechnology. Aldershot etc.: Avebury.

GOLDIN, I. (1990). Agriculture and Economic Crisis. Lessons from Brazil. OECD-Observer, No.161.

GOMES-NETO, J.B., und HANUSHEK, E.A. (1994). Causes and Consequences of Grade Repetition: Evidence from Brazil. Economic Development and Cultural Change, Vol. 43, No. 1, pp.117-48.

GONÇALVES, R. (1993). Estratégias dos oligopólios mundiais nos anos 90 e oportunidades do Brasil. Estudo da Competitividade da Indústria Brasileira. Campinas (mimeo).

HAWRYLYSHYN, G. (1991). Brazil - leader in deepwater technology. Petroleum Review, January, pp.32- 35.

HOLANDA, N. (1993). A crise gerencial do Estado brasileiro. In J.P. dos Reis Velloso (Ed.), Brasil: a superação da crise (pp. 163-176). São Paulo: Livraria Nobel.

LASMAR, J.O. (1994). Descentralização da política industrial: o novo papel dos estados e municípios. Simpósio internacional sobre política industrial, São Paulo e Rio de Janeiro de 07 a 10 de junho.

MARCOVITCH, J. (1990). Política Industrial e Tecnológica no Brasil: Uma Avaliação Preliminar. Pensamiento Iberoamericano, No. 17.

MESSNER, D. (1993). Stärkung technologischer Kompetenz in Bolivien. Berlin: Deutsches Institut für Entwicklungspolitik.

MEYER-STAMER, J. (1995). Governance in the Post-Import Substitution Era: Perspectives for New Approaches to Create Systemic Competitiveness in Brazil. Brighton (IDS Discussion Paper, No. 349).

NELSON, R.R. (1992). National Innovation Systems: A Retrospective on a Study. Industrial and Corporate Change, Vol. 1, No. 2, pp.347-74.

NELSON, R.R., und ROSENBERG, N. (1993). Technical Innovation and National Systems. In R.R.Nelson (Ed.), National Innovation Systems. A Comparative Analysis. New York, Oxford: Oxford University Press.

NUN, J. (1993). Argentina: The State and Scientific and Technological Activities. Paper presented at INTECH Seminar, Maastricht, March.

NUSSENZVEIG, H.M. (1992). Research Funding in Brazil: A Case History. Technology in Society, Vol. 14, pp. 137-149.

OECD (1994). Science and Technology Policy Outlook: Part III -Selected Issues. Paris: OECD (DSTI/STP(93)19/REVol. 1).

OECD/TEP (1992). Technology and the Economy. The Key Relationships. Paris: OECD, The Technology/Economy Programme.

PAVITT, K. (1991). What makes basic research economically useful? Research Policy, Vol. 20, pp.109- 119.

PBQP (1992). Pesquisa qualidade e produtividade no meio empresarial. Brasília.

PEREZ-NUNEZ, W. (1994). Latin America's experience with technology policies: Current situation and prospects. International Journal of Technology Management, Vol. 9, No. 3/4, pp. 332-50.

PORTER, M.E. (1990). The Competitive Advantage of Nations. New York: The Free Press.

RAMOS, A.W. (1990). Reavaliação do sistema nacional e da infra-estrutura pública de normalização técnica, metrologia e qualidade. Campinas: Unicamp, Instituto de Economia (mimeo).

RIBEIRO, S.C. (1993). Educação e cidadania. In J.P. dos Reis Velloso (Ed.), Educação e modernidade (pp. 137-164). São Paulo: Livraria Nobel.

ROSENBERG, N. (1990). Why do firms do basic research (with their own money)? Research Policy, Vol. 19, pp.165-174.

SAGASTI, F. (1978). Science and Technology for Development: Main Comparative Report of the Science and Technology Policy Instruments Project. Ottawa: IDRC.

SCHWARTZMAN, S. (coord.) (1993). Ciência und Tecnologia no Brasil: um nova política para um mundo global. São Paulo: Fundação Getúlio Vargas.

SERCOVITCH, F.C. (1984). Brazil. World Development (Special Issue on Technology Transfer from Developing Countries), Vol. 12, No. 5/6, pp. 575-599.

SOSKICE, D. (1994). Innovation Strategies of Companies: A Comparative Institutional Approach of Some Cross- Country Differences. In W.Zapf, M.Dierkes (Ed.), Institutionenvergleich und Institutionendynamik. WZB-Jahrbuch 1994 (pp. 271-289). Berlin: edition sigma.

STORPER, M. (1993). Regional 'Worlds' of Production: Learning and Innovation in the Technology Districts of France, Italy and the USA. Regional Studies, Vol. 27, No. 5, pp.433-55.

TEITEL, S. (1987). Towards An Understanding of Technical Change in Semi-Industrialized Countries. In J.M.Katz (Ed.), Technology Generation in Latin American Manufacturing Industries (pp. 94-115). Basingstoke, London: Macmillan.

TENDLER, J. (1994). Good Government in the Tropics. Cambridge, Mass.: MIT (mimeo).

TEUBAL, M. (1984). The Role of Technological Learning in the Exports of Manufactured Goods: the Case of Selected Capital Goods in Brazil. World Development, Vol. 12, No. 8, pp.849-865.

UNIDO (1992a). Brazil's Industrial Policy: An Assessment in the Light of the International Experience. Vienna (PPD.226).

UNIDO (1992b). Industrial Competitiveness in Brazil: Trends and Prospects. Vienna (PPD.225).

UNIDO (1993). Industry and Development. Global Report 1993/94. Vienna.

VILLASCHI, A. (1992). O Brasil e o novo paradigma tecnológico de desenvolvimento econômico mundial. Ensaios FEE, Vol. 13, No. 1, pp.43-87.

WORLD BANK (1991). Brazil. Economic Stabilization with Structural Reforms. Washington (Report No. 8371-BR).

WORLD BANK (1994). Brazil. An Assessment of the Private Sector. Washington (Report No. 11775- BR).