Simon Schwartzman, general coordination; Eduardo Krieger, biological sciences; Fernando Galembeck, physical sciences and engineering; Eduardo Augusto Guimarães, technology and industry; Carlos Osmar Bertero, Institutional analysis

6. Policy recommendations

To achieve these goals, the following recommendations are made:

a. To redirect the country's technology policies, in line with the new economic realities.

Technology policies are needed to make it possible for the country to enter a new pattern of industrial growth, centered on increasing levels of competitiveness. On the short run, the policies should be geared to the reorganization and technological modernization of the industrial sector. Then, permanent policies should exist to induce the more dynamic sectors of the productive system to enter a continuous process of innovation and incorporation of new technologies, to follow the rhythm of technical progress in the world economy (box 9). Both approaches require, as the main priority, the incorporation of existing technology to the productive process. Sectorial policies are needed for the reorganization and technological modernization of less efficient parts of the economy, and for the consolidation and expansion of the more dynamic industrial sectors. Support for research and development activities should be selective, and clearly associated with broader processes of innovation based on the transfer, diffusion and absorption of technological competence.
 

Box 9 - Technology Transfer: the new economic policy orientations (Eduardo Augusto Guimarães, 1993) 

The issue of protectionism vs. market competitiveness in scientific and technological development should be dealt in pragmatic, rather than in ideological terms. It is impossible, and it would be tragic, to shield the country from the technological revolution that is taking place in the world. A key element of this revolution is the role of multinational corporations and international trade in the development and diffusion of modern technologies, which makes the issues of technological development and international trade so intertwined. At the same time, a country should not renounce to its instruments of technological and industrial policy, including tax incentives, tariff protection, patent legislation, government procurement and long-term investments in technological projects, in association with the private sector. The aim of these policies should be always to improve the country's scientific and technological competence, and reap the benefits of increased efficiency, productivity and trade. In this context, adequate legislation for patents and intellectual property should be established, with the understanding that they are necessary for the normalization of Brazil's relations with the industrialized countries, although, on itself, such legislation could neither guarantee nor jeopardize the technological advancement of Brazilian industry (Pereira, 1993) (box 10).  

Box 10 - Recommendations about the Patent Law (Lia Vals Perereira, 1993) 

b. To protect the existing pool of scientific competence.

Many of the best R&D institutions and groups are being dismantled by absolute lack of resources, and emergency measures are needed to deter this process. The government should guarantee a stable and predictable flux of resources to its main S&T agencies for their daily routines and "over the counter," peer reviewed research supporting activities. The problem is not just of limited resources, but above all of lack of institutional stability and commitment with the sector, since the resources needed is not very high. Special programs like the PADCT could be used for this purpose (box 11). 

Box 11: Financial Stability and the PADCT (Caspar M. Stemmer, 1993) 

Not only the agencies need to have their resources, but the most qualified research institutions and groups should be preserved in their ability to keep their best researchers and their work. There is a proposal being discussed for several years to establish a network of laboratories and research groups to be supported by government through long-term grants based on past performance ("laboratórios associados"), which requires prompt implementation. Estimates of size and cost of such a network vary, but the scale of the operation is not too difficult to determine. Of the estimated 15 thousand active researchers in the country, about a third, or five thousand, could be included in 200 such "laboratories" of 25 persons each, supported with one million dollars a year in average, or 40 thousand dollars per person, two hundred million dollars in total. This would be the cost of keeping the pool of competence in place, providing a basis from which other policies can be devised. A similar amount will be needed to provide these laboratories with basic equipment and infrastructure. Most of this money is already being spent as salaries by universities and other government agencies, so that the cost of this program would be even less (although the resources for research and infrastructure should be provided in addition to what is needed for regular graduate education). Ideally, the program should compensate for oscillations in salaries, guarantee resources for current expenditures, and provide means for the acquisition and modernization of scientific equipment, regardless of the group's institutional location. Resources should be allocated competitively, under strict peer review evaluation, and for limited periods (typically three to five years). The criteria for allocation should be the laboratories' track records, the quality of their researchers, their ability to get funds from other sources, and their long-term perspectives and projects.

This network of research laboratories should be strengthened by a specific line of support for individual scientists, allowing them to move around to find the best places to use their competence; laboratories could be rewarded for the quality of people they can attract.

A competent and prestigious peer review procedure is essential for the project of "laboratrios associados" to work. On the long run, difficult problems of choice between equally competent groups and proposals may appear, requiring decision procedures that go beyond traditional peer review. However, given the current small size of Brazil's scientific community, most competent groups are likely to be supported, without increasing the historical levels of expenditure.

c. To develop a three-pronged policy for S&T development, with clearly distinguished support mechanisms for basic science, applied work and extension and education.

The fact that basic science, applied R&D and high level technical education are very often indistinguishable, and take place simultaneously in the same institutions, does not mean that they should not be treated separately in terms of their supporting mechanisms, working from different perspectives and with different approaches.

Basic science research and education.

Box 12: Basic Resarch: The Laboratory of Synchroton Light (Sérgio M. Resende, 1993)

Basic or academic science, broadly understood as research work that does not respond to short-term practical demands, remains necessary not only for its eventual role as the source of privileged discoveries for applied work, but because of its nature as an indispensable public good. Scientists should be trained on a broad fundamental basis, so that they will not become obsolete in a short time. This purpose is not in contradiction with applied work, but should not be jeopardized by too much emphasis on efforts to help solve short term operational problems of the productive sector. In spite of the growing presence of proprietary knowledge in modern societies, academic science is also expanding,and the resources it can expect from the private sector are not very high(7). The information basic science generates is free for the private sector (although it is paid by society as a whole), and is the main source for the acquisition and spreading of the basis of tacit knowledge that permeates the whole field of science, technology and education. For a leading country, heavy investments in basic science can be thought of as problematical, since their results can be appropriated by other countries and regions for very little cost. For the same reason, investments in basic science in small scientific communities can be extremely productive, since they allow tapping the international pool of knowledge, competence and information. This is the rationale for projects such as the National Laboratory of Synchrotron Light, now under construction by CNPq (Resende, 1993) (box 12).

Besides its eventual impact on the productive sector, basic science can play a fundamental role in enhancing the quality of higher education for engineers and for society as a whole. This role, however, does not take place as a matter of course. Universities have to develop explicit links between their graduate and undergraduate programs; intellectual and financial investments have to be made for the development of materials for science teaching, from handbooks to educational software and experimental kits. When these links and investments exist, basic science becomes more legitimate, and more likely to be supported by society.

Changes are also necessary in scientific and graduate education. Master degree programs should be shortened, and turned either into well-organized courses of professional specialization, or short entrance and leveling programs leading to doctoral degrees. Non-degree, specialization courses should be stimulated, with very little bureaucratic formalism, and as self-supporting as possible.

Globalization requires a profound rethinking of the old dilemma between scientific self-sufficiency and internationalization. The experience of small, high-level scientific communities in countries like Canada, Israel, the Netherlands and Scandinavia, shows that this may be a spurious question. They build their competence through a purposeful effort to be present in the international scientific scene, by the extensive use of the English language, participation in joint research projects, evaluation of their research activities by scientists from other countries, and a constant international flux of students, researchers and information; they are not less developed, and their science less relevant for their societies, for that reason.

The current fellowship programs of CAPES and CNPq for studies abroad need to be revised. Fellowships should be awarded only to first-rate students going to first-rate institutions, and with a clear perspective of returning to productive work in Brazil. Fellowships for doctoral degrees should be combined with "sandwitch" fellowships for doctoral students in Brazilian institutions, and short-term support for training periods in laboratories and companies abroad. Procedures should be devised to constrain the beneficiaries to pay back the money received if they fail to get their degrees or to return to their institutions; fellowships for countries and institutions with poor records of academic achievement for fellows should be avoided. Under these conditions, the current number of fellowships awarded should be maintained and even expanded. The existence of good quality doctoral programs in a given field does not preclude the need to keep a permanent flux of students to the best foreign universities. Provisions should exist for post-doctoral programs both abroad and in Brazil, and to bring top-quality scholars from other countries for extended periods, or even permanent appointments, in Brazilian university and research institutions (De Meis and Longo, 1990).

Applied science

The central feature of applied science is that it has a user, and the knowledge generated in the R&D process tends to be proprietary. The main clients for applied science in Brazil have been the military, the large state-owned corporations and a small section of the private sector, including the export-driven agricultural firms.

Box 13 - Multi-client program in forestry research (João Lúcio Azevedo, 1993) 

Applied R&D should be evaluated in terms of its short-term scientific quality and medium or long-term practical results. When the client is a public institution, such as the military or a state-owned corporation, R&D projects tend to be secretive, large and long-term. Evaluation is very difficult to carry on, since research results are not open to publication, peer review scrutiny, or market competition (Erber, 1993). Lacking appropriate evaluation procedures, applied R&D in the public sector runs the risk of being expensive and of doubtful quality, the same holding for publicly-subsidized R&D in private firms.

There is a clear trend away from these kinds of R&D activities, however. In the current international context, there are limits on what smaller and poorer countries can do in terms of military prowess. In contrast, there is a strong premium on the spreading of education, technical competence and competitiveness through society. Most public corporations are either being privatized or forced to rely on market mechanisms to survive. In both cases, publicly subsidized applied R&D will tend to diminish. The Brazilian experience of subsidized R&D to the private sector is not very good. If loans are granted below the market interest rates, there may be many takers, but the outcomes are often poor. There is a room, however, for special procedures to finance long-term and joint R&D projects that would not otherwise find support through commercial banks. General policies and support mechanisms for applied R&D are difficult to devise, since they refer to an extremely variegated range of activities, and require different combinations of economic, scientific and strategic considerations. A few suggestions, however, can be made:

Research groups in universities and government institutes should be strongly stimulated to link to the productive system and to engage in applied work, while maintaining a high level of academic and basic research activities. It is as unwarranted to expect that all basic science should be linked to production as to assume that they should be kept isolated from each other. There is no reason to believe that applied work would necessarily distract researchers from their basic and academic oriented activities. However, tensions and conflicts of interest may arise, and need to be administered case by case. Links between academic research and the productive system can take place at multiple levels, depending on the capabilities and needs of each side. They can go from help in the solution of short-term problems and difficulties faced by industries, to the transfer and scaling up of innovations produced by research centers for industrial production; and, at the higher end, to the development of large scale, cooperative projects of R&D (Frischtak and Guimarães, 1993). Links can be established either with single institutions or with associations and consortia of users, as in the example of the Institute of Forestry Research of the Universidade de São Paulo (Azevedo, 1993) (box 13). Resources for applied work should not come from the budget for basic activities, but from specific sources in governmental agencies, special programs, private firms, and independent foundations.

Government agencies dealing with matters requiring research work, like in health, education, environment, energy, communications and transportation, should have resources to contract research in universities and research institutions on matters of their interest. This practice should prevail over the tendency of these agencies to create their own research outfits, and their projects should be subject to joint evaluations by peer review and policy oriented authorities. Research institutes and centers in public agencies and state companies should be placed under peer oversight, and required to compete for research support outside their agencies.

The current military projects should come under technical, academic and strategic evaluation with the participation of selected, high quality scientific advisers, and be either streamlined, discontinued, reduced, or converted to civilian projects (Cavagnari, 1993) (box 14).
 

Box 14 - Military research and competitiveness (Geraldo M. Cavagnari, 1993) 

Research programs in applied fields like electronics, new materials, biochemistry and others, should only be established in association with identified partners in industry, which should contribute with their own resources, and be involved from the beginning in the establishment of appropriate objectives; they should be subject to independent evaluations of economic, managerial and scientific feasibility, and monitored on these terms.

New actors should be brought in projects of local and regional development, including local and state governments, business associations, financial institutions, universities and technical schools.

Education

The main challenge for Brazilian science and technology in the coming years is to spread competence for innovation horizontally, to the productive system as a whole, and to increase the educational level of the population. While this is not done, the S&T establishment is bound to relate only to a small part of the country and its economy, with limited resources and relevance.
 

Box 15: Human resources development in South Korea (C. M. Castro and J. B. Oliveira, 1992) 

Policies for science and technology should not wait for educational reform, but they cannot be expected to succeed without profound transformations in the educational system as a whole, through increased access to educational opportunities, quality improvement of basic and secondary education, strengthening of technical education and diversification and better use of public resources in higher education. Questions of educational policy lie beyond the scope of this document. However, a few items should be stressed regarding the interfaces between the educational and the S&T sectors:

Technical Education. Brazil has maintained a wide gulf between education for the academic professions, including engineering, and middle-level professional training, the first provided by universities, the second by federal and state technical schools (and also by industry and commerce through their own institutions, SENAI and SENAC). The knowledge-intensive basis of modern industry and services requires the development of general skills for the technician, and proximity with industry for the institutions trying to provide technical education through formal course programs (Castro and Oliveira, 1993) (box 15). Brazil has lagged behind the worldwide trend of developing post-secondary, short-term course programs as alternatives to conventional university education. The expansion of post secondary, technical education, developed with close links to industry, should become a central task for public universities and state governments. Although more difficult at the beginning, this new emphasis could prove more useful than the sheer expansion of evening course programs that became mandatory for public universities in the recent past; and more realistic, in budgetary terms, than the proposed multiplication of federal technical schools operated by the Ministry of Education.

Extension work and continuous education. University research deparments and institutes should be stimulated to become more intensely involved with extension work and continuous education. An important activity in this regard is the systematic translation into Portuguese or rewriting of science education textbooks and engineering documents (handbooks, standards, technical manuals for craft and skilled workers) needed for general use in industry and education. These activities already take place in some universities, but are usually considered of low prestige, and inimical to academic excellence. It does not have to be so. High quality research centers can attract more resources, increase their relevance and involve more people through extension activities. Institutions with little to offer in terms of research can gather strength and recognition, and provide their students with significant opportunities for practical training. Since most of these activities can be paid for by users, they do not require much in additional resources, but there should be means to provide rewards, incentives and recognition for this type of work.

Teaching of science and technology. Academic departmemts in universities should take more responsibility for undergraduate education. The current departmental structure tends to leave undergraduate career programs without intellectual leadership, and undergraduate teaching is often seen as a burden by professors engaged in graduate education and research. Incentives should be created to stimulate researchers to get involved with undergraduate education, by writing textbooks, bringing undergraduates to their research projects, and participating in the upgrading of their course programs. Fellowships for undergraduates ("bolsas de iniciação científica") should be expanded, and successful involvement in undergraduate education should be added as a criterion for CAPES in its evaluation of graduate programs.

General education. Most undergraduate education in Brazil, as elsewhere, is in fields like administration, languages, social sciences and the humanities. They can be considered "general education" course programs, since they have little knowledge specificity, and are supposed to provide the students a broad spectrum of cultural, social and historical disciplines. There is a tendency to see these "soft" fields as a waste of time and resources, under the assumption that they are not directly relevant to the production of goods. However, general skills, social and cultural abilities are central components of modern economies and societies, characterized by intense flows of information and communication, the continuous expansion of services and a shifting social and economic environment. The situation of neglect should be reversed, with the graduate and research sector taking responsibility for improving the quality of secondary and undergraduate general education, through direct involvement in the production of good quality textbooks, the development of curricula and new teaching methods. Here again, adequate procedures should be devised to make these activities more rewarding and prestigious than they have been so far.

Distance learning. Modern technologies for distance learning have not been adopted in Brazil except in some isolated experiences in basic education. A systematic effort should be made to incorporate the international experience, and a few universities should be stimulated to begin pilot projects using the newly available instruments, from computers to electronic mail.

d. Infrastructure for information and knowledge dissemination.

New and systematic procedures to incorporate technology into the industrial process should be developed, with strong emphasis on the development and dissemination of norms and standards, information, and procedures for technological transfer and quality improvement. Brazil has several institutions created for these tasks, such as the National Institute of Metrology, the National Institute of Intellectual Property and the Brazilian Institute for Scientific and Technological Information (IBICT). These institutions, however, have lived in a no-man's land between the researcher in academic institutions and the productive sector, both of which were either protected from external competition, or linked directly to their own sources of information and technology outside the country. Without closer interaction with the users of their services, these institutions tended to become rigid and bureaucratic, weakening still further their links with the scientific and productive sectors.

To reduce this problem, users should play a much stronger role in the definition of the goals and practices of these institutions. A well organized and properly funded knowledge infrastructure is necessary to assure the easy access of scientists to libraries and data collections in the country and abroad. Significant advances have ocurred in the last few years, through the gradual generalization of access to Internet and similar networks for Brazilian universities and research groups, and the development of computerized library catalogs in some of the main universities. Now it is necessary to make these links more widely used, more effective and more transparent to the individual researcher, and to establish means to bring the documents and data to the scientist's working place. A coherent policy for creating, maintaining and expanding these information resources is needed, and should rely on the competence developed by FAPESP, the National Laboratory of Computer Science Research (LNCC), the Institute of Applied and Pure Mathematics (IMPA), and other groups that have built and developed the current facilities.

e. Institutional reform.

For these policies to be carried on, governmental agencies for S&T policy should become smaller, more flexible and more efficient. The Brazilian agencies for science and technology are considered more efficient, and less plagued with problems of political patronage and bureaucratic formalism, than most of Brazil's civil service. However, with a few exceptions, the general evaluation of the main government agencies is not very positive. CNPq has grown to be a large bureaucracy, going from 1,502 employees in 1988 to 2.527 in 1992, about half of them without a university degree (Barbieri, 1993, table 2). Its administrative expenditures have varied enormously throughout the years, and most of its resources are now given to fellowships. Researchers and fellows applying for its resources complain constantly about the difficulties in getting information and receiving their fellowships and grant money on time. The agency could never establish a competent information system about its own activities, and there is little or no follow-up of the results of its investments in research and graduate education. FINEP also increased its bureaucracy to about 700 employees, while its resources shrank (Frischtak, coord, 1993). There are no established dates for submission of projects and proposals nor public announcement of awards. Without systematic peer review, there is no information on how decisions are made. In both cases, the problems were compounded by budgetary uncertainties. The agencies do not know how much money they will have at any given time, and their decisions are often based on expectations that are not fulfilled. Finally, these agencies have not established adequate procedures to receive proposals in constant values and protect their grants from inflation. The consequence is that, when a project is finally approved, its value is significantly lower than when it was presented, and still lower when the money is received and spent.

Table 3: CNPq, Budget according to main lines of activity, 1980-1992. Millions of US$ of 1992

In contrast, FAPESP, in the state of São Paulo, and CAPES, at the Ministry of Education, are perceived as success cases. FAPESP works almost exclusively through peer review, its administration is very small, communication with applicants is very efficient, its grants are fully corrected for inflation, and has well-designed follow-up procedures for its grants and fellowships (box 16). CAPES suffers some of the difficulties of being within the Ministry of Education, but has an established tradition of peer review assessments. Its leadership has been always recruited among persons of good academic standing, and its bureaucracy remains small.  

Box 16 - The FAPESP model (Milton Campanário and Neusa Serra, 1993) 

These experiences provide a basis for the following suggestions for institutional reform:

• The Ministry of Science and Technology should restrict its role to matters of policy, financing, assessment and evaluation, without carrying on R&D activities under its direct administration. Although a science adviser or an equivalent cabinet-level position for science and technology is clearly necessary, the very existence of a formal ministry of science and technology, with all its overhead costs and exposure to political patronage, should be reexamined.
• The existing system of federal institutions for scientific and technological support should be evaluated in terms of its ability to perform the functions needed by the sector: support for basic research, support for applied projects, large and small research grants, fellowship and training programs, scientific information, norms and standards, and others. Brazil needs a federal agency to provide long term, sizeable grants for institutions and cooperative projects. This was the role played in the past by the National Fund for Scientific and Technological Development (FNDCT), administered by FINEP. Whether these resources should be managed by FINEP, CNPq or by a new institution should be examined as part of a broad review of the roles, jurisdiction and competencies of the existing agencies.
• Financing agencies should be organized as independent, state owned corporations, and free of formalistic and bureaucratic constraints. They should be placed under strict limitations regarding the percentage of their resources they can spend on administration, and should be supervised by high-level councils with the participation of scientists, educators, entrepreneurs and government officers. They should rely on external advise for their decisions, and their bureaucracy should be limited to the minimum.

Research institutions and public universities should not be run as sections of the civil service. They need to have the flexibility to set priorities, seek resources from different public and private agencies, and establish their own personnel policies. While this is not changed, there is always the alternative of developing hybrid institutions with flexible mechanisms coexisting with more rigid procedures (the Brazilian academic community has some experience in this). Universities should develop appropriate settings for interdisciplinary work in new fields such as biotechnology and artificial intelligence (Carvalho, 1993; Silva, 1993).

No research institution receiving public support, and no government program providing grants, fellowships, institutional support or other resources to the S&T sector, should be exempt from clear and well-defined procedures of peer evaluation, combined, when necessary, with other types of economic and strategic assessments. Peer review procedures should be strengthened by the federal government, made free from pressures of regional, professional and institutional interest groups, and acquire a strong international dimension. For instance, research proposals could be easily distributed to international referees through electronic mail.

It should be the task of the Ministry of Science and Technology to stimulate such reforms in other branches of the federal government. The Ministry of Education should have a particularly important role in maintaining the quality and the autonomy of the research groups in the federal universities.

It should be also the task of the Ministry of Science and Technology, in cooperation with the Ministries of Finance and Foreign Affairs, to keep the channels open for international cooperation between Brazil, international agencies and institutions, and the international scientific community. The World Bank, the International Development Bank and the United Nations Development Program have played important roles in providing resources for capital investment, research support and institutional development of Brazilian institutions. This presence should be maintained not only because of the resources involved, but because of what they bring in terms of international perspectives and competence. In the future, such agencies could be very helpful in a process of institutional reform. As a rule, cooperation among scientists, research institutions and private foundations in different countries is established directly, and need the support, but not the interference, of governmental agencies.

f. Goal-oriented projects

The broad changes suggested in this document do not preclude the adoption of well-identified projects linking science, technology and the productive sector, aimed at the strengthening of specific fields and orientations in the natural and social sciences, the establishment of instruments for S&T diffusion and education, and others. Proposals of this kind are present in the background studies commissioned for this study, and will emerge as a matter of course whenever decisions have to be made about the allocation of scarce resources, the beginning of new lines of work, and the phasing out of others. It is necessary to develop a list of main areas of established competence and social relevance to be the focus of future investments; to identify areas that should be phased out, or reduced; and special weaknesses and competencies in need of strengthening and support. Two very broad principles should preside this policy-making process:

It would be inappropriate to assume that science and technology would develop and increase its usefulness in a peripheral region if left alone to respond to market mechanisms of economic and scientific competition. The distribution of science and technology, as it is well known, is very lopsided in any given field and region, and the current expansion of communications and trade is leading to further concentration of knowledge, competence and technical resources. Policies are needed for general and technical education, for bringing flexibility and introducing accountability in public-supported institutions of higher education, research and development, and for introducing standards of quality. The trend toward concentration cannot be stopped by policies of isolation and self-sufficiency, or subsidies to second-class institutions and research groups. But it is not a zero-sum situation. As the world gets more integrated, information circulates, and the knowledge basis increases, there are new opportunities to be grasped. For this, they have to be properly perceived and understood, and adequate investments in education should exist.

Brazil has had some experience of integrated programs covering specific areas of interest, such as tropical diseases, natural resources, energy and computer sciences. PADCT has followed a similar pattern, by choosing a few, selected areas for support, and assigning a given fraction of its resources to them. An integrated program would have, ideally, resources for combined activities of basic and applied research, graduate education and training. For the fields chosen for such programs, the benefits seem obvious, since their resources are guaranteed, and the links between basic and applied research and education can be made more coherent (box 17). There are, however, three pitfalls for such programs, which should be avoided. First, they run the risk of isolation. As with any applied project, integrated programs need to have clearly identified and active partners outside the research and education sector - be they the health ministry, the electronics industry or public utility companies. In fields that are economically relevant, they should be linked to specific industrial policies, and the participation of relevant business leaders. When this pattern is not present, the results of the integrated project are not used, and the effort can be wasted. Second, there is always the temptation to distribute the R&D resources arbitrarily among programs, creating unwarranted imbalances. Third, self-contained programs are prone to shun peer review evaluations and give excessive protection to a few institutions and research groups, disregarding excelence in favor of selected subjects or problems. If these difficulties are considered if there are clearly identifiable partners in government and industry, if there are no arbitrary block assignations of resources, and if peer review procedures are kept in place integrated programs can be important and forceful instruments for improving the country's S&T capabilities.
 

Box 17 - A goal-oriented project: strategic development of computer science (Carlos Lucena, 1993) 

Note

7. Only about 15% of the resources for "public good" research in the US comes from the productive sector (Aron Kupperman, private communication).

---