IOZ 03 AGRICULTURAL RESEARCH Sector Policy Paper rf t 'World Bank, June 1981 -~~~~ ~ ~~ -: - - w :14~~~~~~~~~~~~~~ ; :x . :t, . - AGRICULTURAL RESEARCH Sector Policy Paper June 1981 World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. "Few scientists think of agriculture as the chief, or the model, science. Many, indeed, do not consider it a science at all. Yet, it was the first science-the mother of all sciences. It remains the science which makes life possible. And it may well be that before the century is over, the success or failure of science as a whole will be judged by the success or failure of agriculture." -Andre and Jean Mayer AGRICULTURAL RESEARCH Sector Policy Paper June 1981 This paper was initially written for presentation to the Executive Directors of the World Bank. It was prepared by Theodore J. Goering with the editorial support of Emmanuel D'Silva. Agricultural research specialists within the Bank, as well as outside consultants, assisted in the preparation of the paper. AGRICULTURAL RESEARCH Contents Page Summary and Recommendations ............... 5 Chapter 1: Agricultural Research and Economic Development. ........ 12 Agricultural Technology: Indispensable to Growth .12 Chapter 2: Agricultural Research in Developing Countries .16 Size of the Current Research Effort .16 The International Agricultural Research Centers .18 Impact of the Global Research Effort .19 Issues of Organization and Management .24 Chapter 3: Agricultural Research: The World Bank's Experience ......... 34 Types of Bank Support ......... ................... 34 Research Components in Agriculture and Rural Development ........ ......... ... 34 National or State Research and Extension Projects ...... 35 Research Components in Education Projects ..... ...... 36 Support of the CGIAR System ............. ........ 37 Chapter 4: Policy Guidelines ....................... 38 Indicative Investment Program in Research and Extension . 39 Operational Strategies for the Bank ....... ........... 42 Performance of Countries and World Bank Lending .. 46 The Bank and the CGIAR .......... ............... 47 Agricultural Research-In Pictures ....... ............ 49 Text Table Proportions of Planted Area of Wheat and Rice Sown to High-Yielding Varieties, 1976-77 ......... 21 Annexes 1. Expenditure on Agricultural Research by Region and Income Group, 1975 ........................... 58 4 2. Present Structure of International Agricultural Research ............... ..................... 60 3. Summary of Selected Studies of Agricultural Research Productivity ............. ..................... 64 4. Number of Research Scientists per Major Crop by Developing Country .. . ........................ 66 5. National Agricultural Research and Extension Projects Supported by the World Bank, Fiscal 1971-80 ..... 67 6. Estimates of National and International Expenditures on Agricultural Research by Major Commodity ....... 98 7. Summary of Bank Support for Agricultural Research and Extension Activities, Fiscat 1970-78 . . ........ 99 8. Contributions to the CGIAR System, 1972-80 ..... .... 100 9. An Indicative Program of World Bank Lending for Agricultural Research and Extension, Fiscal 1980-84 ................................ 101 10. Toward an Indicative Global Investment Program for Agricultural Research and Extension Systems in Developing Countries .......................... 102 11. Production of Selected Agricultural Commodities in Selected Countries, 1961-78 ................ ... 107 Summary and Recommendations A basic premise of this paper is that carefully implemented agricultural research can be an efficient source of economic growth and is an important contributor to the achievement of key develop- ment objectives. Adequate food for the world's population over the next two decades requires increases in production at a rate of 3 to 4 percent a year in most developing countries and average increases in yield on already-cropped land of no less than about 2 percent yearly. For much of the developing world, it can be concluded, investment must be significantly expanded to improve the capacity to conduct agricultural research if these targets are to be achieved. The objective of the paper is to provide information and policy guidance of relevance to those efforts. Agricultural Research in Developing Countries Worldwide research expenditures on agriculture now total about $5,000 million annually, or roughly three times (in constant dollars) the amount spent in the early 1960s. The portion spent in the developing countries has increased moderately to about one-quarter of the total. Developing countries typically spend only about one-third as much of their agricultural gross domestic product on agricultural research as do developed countries. In terms of research expenditure per person in the agricultural population, disparities between devel- oped and developing countries are even more striking. The establishment of the international agricultural research net- work and the formation, in 1971, of the Consultative Group on International Agricultural Research (CGIAR), gave a major impetus to efforts to enlarge research expenditures in developing countries. This international system now embraces research on almost all major food crops and livestock products that constitute basic food items for some 70 percent of the world's population. Sizable as this effort is, the CGIAR's budget is only about 3 percent of the annual global expenditure on agricultural research. The most significant agricultural research product of recent years has been the high-yielding varieties (HYVs) of rice and wheat developed at international and national research centers. Since the introduction of HYVs in the mid-1960s, more than a third of the area under these two cereals in developing countries has been sown with the HYVs, making HYVs the most widely and rapidly adopted technology in agricultural history. The effects on production have been substantial; rice HYVs, for instance, typically yield up to 40 percent more than the varieties they replace and the incremental yields of HYV wheat may be even greater. Major beneficiaries have been lower-income consumers who typically spend a disproportionate share of their income on food. Effects on income distribution are more difficult to determine and more controversial. Incomes of most producers have risen, although differences in absolute income levels between large-scale and small-scale farmers have widened due to 6 inequality in the ownership of land. The HYVs also have widened income disparities among regions-a reflection of the fact that most of the new varieties are poorly suited to less-favored environments. National research programs are typically the weakest links in the global research effort. Common deficiencies include excessive frag- mentation of research activities among government agencies, low priority accorded to research by governments, and inadequate institu- tional structures for research and extension. The limited research capacity of many national systems also limits the returns to investment in the international research centers. There is need to strengthen the administrative and technical capability of national systems through expanded training programs. Assuming that at least 100,000 hectares of a particular crop are required to justify a significant research effort at the country or regional level, an approximate doubling of the present number of research scientists would be needed in developing countries. Training programs would have to be expanded signifi- cantly if these staffing requirements were to be met and must be accompanied by more generous financial rewards to agricultural research scientists in order to retain them in research activities. External assistance to strengthen national systems must take into account the size of a country's agricultural sector and the current state of development of its research system. Perhaps 10 percent of develop- ing countries already have adequate research skills, good national research programs, and effective linkages with international research institutions. These countries may be able to provide assistance to other countries with weak research programs. Another 10 percent of the developing countries have adequate research expertise, but it fre- quently is poorly organized and managed. For this group of countries, external assistance in research organization and management may be required. Nearly half of all developing countries are large enough to justify and support a balanced national research system but lack essential research infrastructure. The needs of these countries are to develop an effective organization for research, to acquire proper research facilities, and to strengthen the scientific manpower base to conduct research. The remaining countries have very limited research resources and no single crop of sufficient importance to warrant a complete research system. For these countries, the major need is to develop a limited capability for research, largely of an adaptive nature, for a small number of economically important crops. Research is successful only if the improved technologies are adopted by farmers. Where adoption is slow, the explanation may well be that price incentives to producers are inadequate. An attractive economic environment for agriculture, and in particular, remunerative prices for producers, must be a part of the efforts made to strengthen the national agricultural research program. The potential benefits of some of the research have sometimes not been realized because of the failure of governments to provide a suitable economic environment that can encourage the adoption of a technology. 7 Agricultural Research: The World Bank's Experience' The World Bank's support of agricultural research and extension takes several forms: (a) agriculture and rural development projects that contain adaptive research and extension components; (b) national or statewide adaptive research and extension projects; (c) research components in education projects; and (d) financial and administra- tive support of the CGIAR. In fiscal 1977-79, lending for research and extension, as embodied in the first three categories, constituted almost 9 percent of total Bank lending for agriculture and rural development. In recent years, about 30 percent of this proportion has been allocated to research alone. The Bank supports the international research system through its role as co-sponsor and donor to the CGIAR. It supplies the Chairman and a Secretariat that serve both the centers and donors. During 1972- 80, the Bank contributed $54.9 million to the system, equivalent to about 9.5 percent of total contributions. The Bank's contributions, paid from operating profits, have been used to finance the residual amounts needed for the agreed expenditures of the centers after other donors have made their commitments. This contribution to the CGIAR is the only grant support provided by the Bank for agricul- tural research. The Bank's support for agricultural research activities is relatively new, dating only from the late 1960s and early 1970s. Some gains in terms of increases in production benefits are beginning to emerge from the earliest Bank-supported research projects, although the full impact is likely to be felt a few years hence. The World Bank's experience in the field of agricultural research, on the whole, has been positive. More emphasis by the Bank on research has resulted in an increase in the flow of local resources for research purposes and has helped make national policymakers and research administrators more aware of the developmental potential from investing in research. There are, however, some problems-mostly related to the designing of projects-that warrant careful consideration in future support by the Bank for agricultural research. The time frame of a project has emerged as a key consideration in design and implementation. The World Bank's experience suggests that in several projects too much was expected too soon in terms of strengthening research institutions and in achieving practical research results. One element contributing to the lower-than-expected results has been the lack of adequate government commitment to agricultural research in the overall development of the rural sector. The scope of the research components in a number of agriculture and rural development projects appears to have been too small to be effective. In some cases, the components seem to bear no clear 'All references to the World Bank in this paper refer to the International Bank for Reconstruction and Development (IBRD) and the International Development Associ- ation (IDA). The fiscal year of these institutions runs fromJuly i throughJune 30. All money terms are expressed in US dollars. 8 relationship to the national research systems and objectives. In the case of research and extension projects, difficulties in implementation reflect several factors: inadequate operating budgets for research activities, shortage of qualified administrative and technical person- nel, and project investment periods that are too short to achieve objectives. Research components in education projects, likewise, have occasionally been poorly linked to national research priorities and, in some cases, have tended to duplicate similar research and training activities in agriculture and rural development projects. The Bank's financial support of the CGIAR constitutes a valuable contribution to the work of the international centers; it has helped to ensure that research tasks of high priority are properly funded. The interest of the Bank is to help ensure that the work of the international centers is supportive of the development objectives of member coun- tries and of the efforts of these countries to strengthen their national research systems. To this end, the Bank will continue to play an active role in maintaining the high quality of work at the international centers and in strengthening the linkage between the Bank's project work and research at the centers. Policy Guidelines In a number of countries, national research programs are weak and, therefore, not able to adapt fully and utilize technologies being developed by the international research community. A rational allocation of financial resources for research on a global basis requires substantial increases in funds to strengthen national programs and to complement any further expansion of the international system. A desirable investment target for research for many countries with poorly developed agricultural research systems would be an annual expenditure (recurrent, plus capital) equivalent to about 2 percent of agricultural gross domestic product (GDP). The rate at which annual research expenditures can be increased and effectively utilized will depend on several factors, including the absolute size of the existing system, as well as the relative size (i.e., research expenditure as a percent of agricultural product), availability of qualified technical staff, and the financial capacity of a country to support a larger national research effort. Taking these resource and institutional considerations into account, growth rates for expenditures on agricultural research in developing countries of at least 10 percent a year over the next few years would be appropriate. On this basis, public sector investment in agricultural research and extension among these countries would increase from about $1,600 million in 1979 to $2,730 million in 1984(1979 dollars). This program is ambitious when compared with efforts in the recent past, but the actual achievement will depend on the extent of commitment of the countries concerned and the support rendered by the international community. The World Bank will continue to play a leadership role in helping to mobilize the required resources. The Bank will also lend more money for research projects that are 9 effectively linked to extension and other development projects. If the Bank's total lending for the agricultural sector rises as currently projected, from $2,500 million in fiscal 1979 to around $4,600 million in fiscal 1984, lending for research and extension components could usefully increase (in constant dollar terms) from about $330 million annually to something in excess of $550 million during the same period. By the end of this period, these amounts will be divided about equally between research and extension. A substantial portion of the increase in funds will be used for expanding the production of food crops that are important to small farmers and to lower-income consumers. Agricultural research projects should pay more attention to nutritional aspects. Where appropriate, the research projects can focus on food production systems that are relevant to the diets of the poor; on farming systems in relatively neglected or resource-poor areas; on applied and adaptive research that is directly applicable to farmers' production problems; on the development of effective organi- zational structures for research and extension; and on the examination of the ecological consequences of sustained, high-input production systems. The Bank will also support, on a selective basis, more fundamental or basic research where the potential benefits warrant the long lead time and heavy commitment of resources that frequently characterize these research endeavors. The Bank will increase its support for sectoral lending to national research programs. Where borrowing countries pursue supportive policies, the Bank will be prepared to consider providing continued technical and financial support for a program of 10-15 years in the form of "repeater" projects that build on the initial investment. Among low-income countries, suitable financing arrangements will need to be provided; such arrangements should include funding by the Bank of an appropriate portion of incremental operating expenses of national research projects. An increase in the number of specialized staff is likely to be needed in the Bank in order to support efforts to strengthen national research systems. The Bank and borrowing countries will also need to utilize more fully the services of other national and international organiza- tions, including the international research centers, in the preparation and supervision of projects. The Bank will also provide a more comprehensive treatment of issues concerning agricultural research in its agricultural sector studies. This sector work will address the adequacy of national research systems and examine the appropriate- ness of national economic policies that bear upon the generation and adoption of improved technology. Course material of the Bank's Economic Development Institute (EDI) will be expanded to include a discussion of the organization, management, and the development potential of agricultural research. Institution building will continue to be a basic objective of the Bank in its support of national research systems. There is a need for more careful tailoring of the training component in projects to ensure that: (a) trainees that are selected are properly motivated and return to the sponsoring institution for a set period; (b) reasonable efforts are made 10 to employ qualified national scientists in research projects to provide the continuity of effort essential to good research; and (c) the use of international staff is carefully planned to avoid the problem of having more expatriates than can be effectively utilized. The Bank will urge project authorities to provide terms of service sufficient to attract and retain competent national scientists and to provide special incentives for research and support staff posted in difficult work environments. More efforts will need to be made to establish closer ties between the national research and extension systems within countries and to link national research projects to the work of the international centers and other international programs. Agriculture and rural development projects containing research components must be designed more carefully to ensure that the research results they yield offer effective support to the production components in the project. In national research projects, provision should be made for annual workshops, progress reviews, and periodic external reviews by visiting specialists. The Bank will provide funds, whenever necessary, to support external reviews of large research projects during the implementation phase and at the end of the investment period. Where appropriate, the Bank will also encourage the agency implementing the project to establish effective monitoring and evaluation systems and to arrange to carry out village-level studies to examine the impact of the project on the community. In expanding assistance to national research programs, the Bank will generally require: - The development of an overall strategy for the national research system. The strategy should include consideration of research priorities, organizational structures, physical facilities, staff development, and institutional linkages. Such a strategy could be prepared with the assistance of joint national and interna- tional "task forces" and, if necessary, funds could be provided under a first-phase technical assistance component. - Adoption of a research philosophy that is oriented toward the practical production problems of farmers. Such a philosophy should accept the use of multidisciplinary teams conducting research in different commodity and farming systems, and be backed up by field studies of farmers' circumstances. Research programs must have strong links with extension work to facilitate information flows from research scientists to farmers and from farmers to research staff. - The acceptance by governments of the need to accord appropri- ate priority to the national research effort, to pursue economic policies that are conducive to the generation and adoption of efficient technology, to provide adequate funds for support of the national research system, and to ensure that personnel policies attract and retain qualified staff. The Bank will continue to collaborate with the international centers and encourage complementarity of effort with its lending program for agriculture and rural development. As the donor of last resort for the centers, the Bank will be prepared to supply the residual 11 funds needed to support an expanded CGIAR program up to the present maximum of 10 percent of the total requirements. An active role by the Bank to help ensure that the international centers continue to be successfully managed is important because of the Bank's dominant position as the supplier of external resources for agriculture and rural development, because of its position as one of the largest contributors to the CGIAR budget, and because of its expanding role in support of national agricultural research systems. The Bank will encourage its staff to participate when requested in oversight activities of the international system, including technical reviews of the centers. The recommendations made in this paper, when taken together, establish a solid basis for an expanded research effort that promises attractive economic and social benefits. But the expanded support is likely to be of limited usefulness unless borrowing countries accord high priority to national agricultural research by providing adequate financial support and appropriate economic and other policies that encourage the adoption of improved technologies and that, above all, establish hospitable work environments for national research scien- tists and administrators. Chapter 1: Agricultural Research and Economic Development In most countries, technological change in agriculture has been the mainspring of economic transformation and growth. From the intro- duction of irrigated agriculture some 8,000 years ago, through the first use of the wooden plough perhaps 6,500 years ago, to the changed farming practices in 18th century Europe and the mechanical and biochemical innovations in this century, new technology in agricul- ture has been the mechanism that permitted production of food to be increased and, simultaneously, capital and labor for nonagricultural activity to be released. Agricultural Technology: Indispensable to Growth It is now widely accepted that a steady increase in agricultural productivity through technological change is indispensable to sus- tained economic growth. It is notable that increases in productivity in the past 100 years have come largely from the application of science- based farm technology, and from changes in management and inputs developed through organized research. Because most countries are now running out of good agricultural land, it is essential that agri- cultural research generate new technologies that will permit higher- yielding crop and livestock production if malnutrition is to be re- duced, increased food costs are to be avoided, and economic growth not threatened. It is also clear that effective agricultural research must be accompanied by strong extension systems and other agricul- tural support services if these production gains are to be realized. In the mid-1930s, average grain yields in the industrialized countries were approximately equal to those in developing countries (1.1 tons per hectare). By the mid-1970s, yields in the developing countries had increased to 1.5 tons, while production per hectare in industrialized countries had more than doubled-to 3 tons. While growth in yields reflects many social, economic, and ecological factors, the differences between the productivity of developed and developing countries can be attributed, at least in part, to the size and effectiveness of national agricultural research efforts. In this context, agricultural research is defined simply as a systematic effort to develop new methods (technology) to increase agricultural productivity or technical efficiency. These methods can include socioeconomic research, as well as conventional field and laboratory work by agricultural scientists. It is useful to divide research activity into two functional compo- nents-science and technology. Science is taken to be that research activity which results in the generation of knowledge with varying degrees of applicability to immediate problems. In contrast, the development of technology in agriculture is the research activity, based on scientific knowledge, that results in a mechanical, biological, or institutional innovation. Competent research scientists move freely 13 between science and technology. Basic research is essentially the generation of knowledge, while applied research is more akin to the generation of technology. Both basic and applied research represent extremes on a spectrum of research activities, and, in practice, the resolution of a particular research problem may utilize both types of research. The agricultural research that is needed in any particular country is determined, in part, by the unique soils, climate, and social and other conditions that can be found there. The cost of research can sometimes be reduced by the transfer of technology from one country to another or from an international center to a national research system. For the developing countries, however, the vast majority of which have tropical or subtropical climates, effective transfer of technology from the developed countries, which, by contrast, largely have temperate climates, is frequently difficult. Sociocultural differ- ences are also sometimes important. Accordingly, the developing countries have particular research needs that must be met mainly through research carried out in the agroclimatic and socioeconomic circumstances in which the resulting technology is to be employed. Other characteristics of the agricultural sector in developing countries also determine the spatial and institutional framework within which agricultural research must be organized. Agriculture comprises, in most countries, large numbers of small producers, many of whom have relatively little economic or political power. Individual producers do not undertake research because it is too expensive in relation to their operation and because any resulting benefits are likely to be widely shared and not captured only by those who finance the research. Since the increased production from a new technology generally causes costs to decline, a substantial share of the long-term benefits of technological change goes eventually to consumers. Be- cause social benefits are potentially large, and because producers are frequently unwilling or unable to finance it, agricultural research in most developing countries is mainly a public sector activity. How benefits are distributed among farmers and the extent to which they are shared by others depend largely on the characteristics of a country's research effort. These characteristics-type, emphasis, and geographical location-in turn, are decided mainly by those who design and implement international and national research programs. If the main goal of agricultural research is greater output through increased productivity, the determination of research priorities and strategies is a relatively straightforward matter. But when research is also viewed as an instrument for achieving broader socioeconomic objectives, the problem of selecting appropriate strategies becomes more difficult. Successful agricultural research produces knowledge or improved material (e.g., seed). This knowledge and material may be fed back into the research process for further development, or it may be released to farmers as new technology. Several direct contributions flow from such technology, including: (a) increased technical efficiency (e.g., increased crop production per hectare); (b) a change in the characteristics or composition of 14 agricultural output (e.g., types of plants more suitable to mechaniza- tion or improved amino acid composition in the protein content of a crop); or (c) reduced production risk (e.g., lessened variability in yields). Any of these results, affecting production, input use, and consumption, may contribute to the achievement of national develop- ment objectives through changes in, for example, farm income and its distribution, consumer incomes, nutrition, and foreign exchange earnings. Institutional factors, in addition to environmental and economic forces, are important determinants of the process by which technology is produced by the research establishment and diffused among users. Central elements in this process are the economic gains or losses to particular interest groups in society that result (or are expected to result) from a new technology. Interest groups may include govern- ments, commercial farmers, agrobusinesses, subsistence farmers, landowners, landless agricultural workers, urban consumers, and exporters. Technological choices relate to research priorities among commodities (e.g., food vs. cash crops; cereals vs. livestock); among regions (irrigated land vs. dry land; fertile lands vs. marginal lands); and among technological biases (land-saving biochemicals vs. labor- saving mechanical devices; capital-intensive technologies vs. those that use intermediate mechanical technologies). Both the supply of, and demand for, technology are influenced by the socioeconomic and political structures of a country. The demand for a particular technology results from the actual or perceived economic gains it provides to specific groups of users. These gains are determined largely by the physical impact of the innovation in terms of its effects on yields and the extent to which the innovation is diffused among producers. The extent of diffusion is determined largely by the nature of the technology and the position of the interest group in the country's socioeconomic structure (the land tenure system, degree of access to agricultural support services, and so forth). The supply of technology, on the other hand, may be influenced by the relative political power of particular interest groups, as pressures on the political structure induce the allocation of funds and human capital to follow particular lines of research. The resulting technology operates through the socioeconomic structure to produce gains (or losses) to particular groups who may thereby experience an increase (or decrease) in economic and political power. The design of national research systems should, therefore, ensure that political and economic forces are appropriately reinforced by government in order that the research needs of all farmers are adequately reflected in the choice of research priorities. This view of the process by which agricultural technology is developed and disseminated facilitates the necessary distinction be- tween means and ends (or objectives). The first level of research results pertains to the generation of knowledge and materials and is clearly a set of means to other objectives (except in the rare instances when research is carried out for its own sake). A second level, pertaining to technical efficiency, product characteristics, and produc- 15 tion risk, comprises the working objectives of the research scientist. For the manager of the research effort and for society as a whole, however, this second level indicates various means to more basic social goals. Thus, while increased production may be an immediate research objective, it is not necessarily among the final goals-as seen by government-of agricultural research. In similar fashion, im- proved income distribution may be an important goal for a govern- ment, but it cannot serve as the wrorking objective of the research scientist. Goals set at higher levels by government properly relate to economic growth, equity, nutrition and, finally, to the welfare of society. The ultimate objective of a country's agricultural research program must be to enhance the welfare of its human population and, in the final analysis, the value of the research effort must be measured by the extent to which it contributes to that objective. Chapter 2: Agricultural Research in Developing Countries The perception that technological change can be an efficient source of growth in traditional agriculture, while of relatively recent origin, has been a major factor in the notable increase of the agricultural research effort in developing countries in the past two decades. In consequence, there is, for the first time in history, an emerging global agricultural research "system" that has, as its primary focus, in- creased food production for the developing countries. This system, while deficient in some respects and, in general, not well coordinated, is composed of several major parts, including national research organizations in the developing countries; regional or intercountry research programs in developing countries; international agricultur- al research centers; and research organizations-including private firms-in the high-income countries. Size of the Current Research Effort Global expenditure on agricultural research has grown sharply in recent years, as has the portion spent in the developing countries. Worldwide research expenditures may now approach $5,000 million a year, or roughly three times (in constant dollar terms) the amount spent in the early 1 960s. The portion spent in Latin America, Asia, and Africa has risen steadily, but still accounts for no more than one- quarter of the total. More than one-third of the total continues to be invested in North America and Oceania. Research expenditures by the developing countries are supple- mented to an important degree by external assistance from mul- tilateral and bilateral donors and by private foundations. The private sector, including multinational concerns, has been an important source of research investment in developed and developing countries, particularly research devoted to export crops and research related to modern production inputs such as hybrid seeds, fertilizers, pesticides, and veterinary supplies. Much of the assistance provided by bilateral donors supports research programs in certain commodities or in a particular discipline. About one-third of the support provided by major bilateral donors goes to strengthen national or international research institutes, while two-thirds is dedicated to research in selected disciplines. France is an exception; much of its external support for research goes to basic research and commodity-oriented research, particularly to franco- phone Africa. An important part of this support takes the form of paying the costs of French scientists who work in locally managed research stations. The Food and Agriculture Organization of the United Nations (FAO) has a large research effort at both the program and discipline levels. In 1977, the FAO provided support to 58 projects designed to establish or strengthen national research institutions. Many of these 17 projects were funded by the United Nations Development Pro- gramme (UNDP), with the FAO providing the technical personnel. About two-thirds of the projects in the 1974 listing of FAO-supported research activities in developing countries involved expenditures of less than $100,000 per project. A recent review by the International Food Policy Research Insti- tute (IFPRI), covering some 65 developing countries, suggests that serious problems exist of understaffing and underfinancing in agricul- tural research, particularly in low-income countries that have impor- tant agricultural sectors. Among the low-income countries, the equivalent of only 0.26 percent of the agricultural gross domestic product (GDP) was spent on agricultural research in 1975; in middle- income and high-income developing countries, which generally have relatively smaller agricultural sectors, the equivalent of 0.42 percent and 0.33 percent, respectively, was spent. The comparable figures in developed countries are in the I percent to 2 percent range. In terms of money spent on research for each person whose livelihood depends on agriculture, disparities are even more striking: the estimate of $0.26 annually in the low-income countries is only about one-fifth the level in the middle-income and high-income developing countries. Annex I provides estimates of research expenditures by region and income group in 1975, and indicates the number of agricultural research scientists in each of those regions. Among the 65 countries surveyed, the number of professional staff engaged in agricultural research (about 23,000) is hardly more than the number that exists in Japan alone. Again, there are sharp differences between low-income and higher-income developing countries. Among the former, there is an average of 16 research workers per million people in the agricul- tural population; among middle-income and high-income developing countries, the numbers are 50 and 62 scientists per million, respec- tively. Nearly two-thirds of all agricultural research workers in these 5 countries are located in only seven countries.' Annex 4 indicates the number of research scientists per major crop by developing country. This general underfunding of agricultural research in developing countries is also suggested by the extent of support services available to research scientists-as approximated by the ratio of technicians to research scientists. This ratio is in the range of 2:1 to 4:1 at the international research centers and in many developed countries. In contrast, among 30 developing countries for which data (1974) are available, only four had more than two technicians for each agricul- tural scientist, and over half had fewer than one. This somber assessment of the state of agricultural research in developing countries notwithstanding, evidence suggests that most developing countries have continued to increase the absolute levels of financial resources committed to such resea-ch. Particularly large increases in expenditures have been registered by Argentina, Bangla- desh, Brazil, Colombia, India, Iran, Mexico, Morocco, Nigeria, Peru, and the Philippines. Research efforts in a number of these countries 'Bangladesh, Brazil, India, Mexico, Nigeria, the Philippines, Thailand. 18 have been supported by the World Bank. Despite these gains in absolute terms, however, the rate of growth of public expenditure on agricultural research declined between 1958-64 and 1968-74. Although the necessary complementary expenditure for agricul- tural extension has also increased in both richer and poorer countries, expenditure patterns for research and extension differ widely between these two groups. Developed countries spend about four times as much for research as for extension; in the developing countries, money spent on extension is much greater than expenditure on research. The reasons for these differences are not always clear; they may relate, however, to differing perceptions of profitability from these two types of expenditures. The greater political "visibility" of money spent on extension may also account for the larger investment in this activity in some developing countries. The International Agricultural Research Centers International investment in agricultural research dates from the colonial era in many tropical countries. There, the focus of such efforts was directed principally at improving the productivity of crops-rub- ber, tea, sugarcane, and bananas-that the temperate-zone colonial powers imported. In the immediate postwar period, some investment in food crop research was made by governments and private groups. Much of the interest in international agricultural research in the past decade stems from the successes of IRRI, the International Rice Research Institute, established in the Philippines in 1960, and CIMMYT, the International Maize and Wheat Improvement Cen- ter, which began operations in Mexico in 1963. Both received major financial support during their early years from private foundations. A major impetus for the establishment of the international research system came in 1971 with the formation of CGIAR, the Consultative Group on International Agricultural Research. Establishment of the CGIAR represented a marked change, in several important respects, in the philosophy toward investment in agricultural research. First, it provided an international stamp of approval of the need for new agricultural technology. Second, it emphasized research on food crops for domestic consumption, thereby correcting the earlier research imbalance that emphasized export crops. And finally, it provided an effective means to increase bilateral and multilateral funding of international research. From 1960 through 1979, about $700 million (constant 1977 dollars) have been spent by the international centers. The CGIAR budget for 1979 exceeded $100 million and was provided by 18 donor countries, four foundations, and nine other national and international donor institutions. As a result of rapid growth in its size and coverage, the international system of the CGIAR now embraces research on almost all major food crops and livestock products that constitute basic food items for some 2,900 million people. Sizable and important as this effort is, it should be emphasized that the CGIAR's budget accounts for only about 3 percent of the annual global expenditure on agricultural research, and is less than what is spent annually on 19 agricultural research in Japan, where only 6 million hectares (less than 1 percent of the cultivated area in developing countries) are under cultivation. Annex 2 describes the present structure of the international agricultural research network. Impact of the Global Research Effort Economic returns to research The larger amounts of money and trained manpower committed to agricultural research in recent years reflect, in part, the high actual or expected rates of return on such investment. Many studies confirm these returns, although data are limited in most of these analyses, and methodological errors may, to some extent, distort results.2 Evidence also suggests that returns on research are sensitive to the time lag between the development of technology and its adoption-a factor that highlights the need to strengthen a country's extension service as its research capabilities expand. It also implies that, in some countries, a large investment in research may be premature if no adequate extension service exists to transplant laboratory and field test suc- cesses to the farm. Nevertheless, the studies do tell a consistent story: realized rates of return are much higher, generally two to three times greater, than likely returns from most alternative investment oppor- tunities in the countries concerned. The results are so robust that they must be taken seriously by policymakers and research managers alike. High rates of return have been realized by research conducted by both national and international research organizations (see Annex 3). Economic returns to particular research activities typically exceed 20 percent a year and frequently are greater than 40 percent. The internal rate of return on investment by IRRI in rice research through 1975 is estimated to be about 80 percent. Roughly similar returns have been estimated on investment in rice research by CIAT, the International Centre for Tropical Agriculture-located in Cali, Colombia-and on national programs in Colombia. With the plausible assumption that opportunity costs of capital in most developing countries are in the 10 percent to 15 percent range, underinvestment in agricultural research appears to be substantial, and considerably more money could be invested in this activity with the expectation that returns would exceed both the opportunity costs of capital and the returns from most feasible alternatives in rural 'For several reasons, rate of return estimates may be biased upward: (a) full costs of developing and utilizing an improved technology may not be included in the analysis; (b) costs of unemployment of resources induced by research may be omitted; (c) inappropriate welfare weights may have been assigned to gains and losses in calculat- ing research returns; (d) cost of research failures-"dry holes"-may not have been properly included; and (e) costs of supporting programs for the development and adaptation of technology may have been understated. For a fuller discussion, see G. Edward Schuh and Helio Tollini, Costs and Benefits of Agricultural Research: The Slate of the Arts. Staff Working Paper No. 360 (Washington: World Bank, 1979). 20 areas. Though research failures are not uncommon, and although money is frequently wasted through poor project design and imple- mentation, there are probably few alternative investment opportuni- ties to which national and international funds could be dedicated that could so consistently yield returns as high as those from investment in carefully designed and managed agricultural research programs. It is ironic that, despite this promise of high returns, investment in agricultural research continues to be accorded low priority in many developing countries. Less is known of the returns to investment in agricultural exten- sion. There is some evidence, however, that they have been less than those from good research, although returns to newer extension approaches, such as the Training and Visit (T and V) system,3 appear to be attractive. The growth of research expenditure has been paralleled by an interest in the economic and social impact of that investment. The effects on agricultural production and productivity are perhaps the most dramatic of possible impacts. They are also the ones most easily measured. Other socioeconomic consequences, probably less well understood and more controversial, grow out of: (a) differential rates of adoption of new technology between large and small farmers, landowners and tenants, and between agriculturally richer and poorer regions; (b) effects on income distribution between producers and consumers, among producers in the same region, and between employer and laborer; and (c) effects on employment and nutrition. Production High-yielding varieties (HYVs) of rice and wheat, the most sig- nificant agricultural research product of recent years in global terms, were generally introduced in 1965-66.4 Twelve years later, 55 mil- lion hectares, or more than one-third of the area under cultivation in these two cereals in developing countries, were being sown to high- yielding varieties, thus making it the most widely and rapidly adopted technology in agricultural history. The major impact of HYV technology on the production of cereals has been in the form of increased yields. Many studies suggest that yields of HYVs of rice typically yield 40 percent more than the traditional varieties they replace. In the case of wheat, the average 'The T and V system aims at increasing incomes mainly of smallholder farmers through the use of simple techniques of better crop husbandry. These techniques often call for little or no increase in cash inputs. See Daniel Benor and James Q. Harrison, Agricultural Extension: The Training and Visit System (Washington: World Bank, 1977). 'High-yielding, semi-dwarf varieties of rice were inroduced in China some five or six years earlier. 21 yield increase is considerably greater. Production growth rates of rice and wheat have accelerated, in part, as a result of the HYV technology. Production of rice in developing countries has grown by about 2.4 percent a year since the advent of the HYVs, as compared with 0.9 percent in the early 1 960s. Wheat output has grown by about 4 percent annually in recent years, as compared with 2.4 percent in the earlier years. During the same period, production in developing countries of maize, sorghum, millet, and other coarse grains fell. Table 1 Proportions of Planted Area of Wheat and Rice Sown to ligh-Yielding Varieties, 1976-77 Wheat Rice Total Region Percent Asia 73 30 41 Near East 17 4 17 Africa 23 3 7 Latin America 41 13 31 Total 44 27 35 Source: Dana G. Dalrymple, as reported in "The Consultative Group and the International Agricultural Research System: An Integrative Report,' CGIAR Secretariat, September 1978. Income distribution and employment The impact on income distribution of changes in yields and production resulting from public investment in research has become a major concern in both the planning of research investment and the assessment of its impact. The mechanism through which technical change affects income distribution is complex and the impact of the change has many facets. Any meaningful discussion of this subject requires consideration of several related isues: (a) the impact on distribution among producers in a particular region; (b) the distribution of gains between landholders and laborers; (c) distribu- tion between regions; and (d) the distribution between producers and consumers. Because the new technology lowers the unit costs of production, net incomes of all producers who adopt it can be expected to rise. But because large-scale farmers generally acquire knowledge of such technologies more quickly, and because they have better access to the working capital needed to utilize these technologies more fully, they capture the early and largest gains from innovation. While absolute income levels of smaller farmers would also increase with the adoption of technology, the lag in their income gains means that, at least in the short run, relative distribution of income worsens as between large- 22 scale and smallholder farmers. Available evidence suggests that smallholder farmers ultimately adopt technologies nearly as com- pletely as do large-scale farmers, thereby raising the incomes of all those who embrace the technologies. While relative income distribution is unchanged, absolute differences in income widen largely as a result of existing differences in the distribution of productive assets. Thus, if the technology made available is neutral to the scale of farming operations and is adopted by all classes of landholders, there is no reason to expect that the relative distribution of incomes should be significantly altered. But, in the absence of effective redistribution of productive assets, of which land is typically the most important, absolute income differences will widen. This may have occurred in many developing countries in the past decade. The preceding analysis assumes that all farmers, large and small, are landowners. In some regions, however, tenant farmers represent an important group of producers. Relatively little is known of the impact of new technology (such as the HYVs) on landlord-tenant contractual relations and on the distribution of resulting gains (or losses). Some evidence suggests that as new varieties are adopted, the area under tenancy has declined, and that there has either been some tenant displacement or that the gross product that must be paid as rent has been increased. Where land is scarce-an increasingly common situation in developing countries-tenants are likely to be adversely affected by the introduction of HYVs unless special provisions are taken to protect them. The HYVs have tended to increase labor requirements per hec- tare (by perhaps as much as 15-30 mandays/ha/crop) in areas where techniques of land preparation and harvesting have remained essentially unchanged. As a result, some gains in income have ac- crued to labor. Where the introduction of HYVs was accompanied by greater mechanization, as has sometimes been the case, labor has been displaced. The increase in the demand for labor appears to have had relatively little impact on agricultural wages, a reflection of several factors-existing underemployment, labor mobility in many countries, and the continued and rapid expansion of the rural labor force. The principal impact on employment of the HYVs may lie in employment off the farm that has been created by the increased demand for nonfarm products resulting from higher rural incomes. Typically, a substantial portion of the increased income resulting from the HYVs is spent by farm families on nonagricultural products such as textiles, footwear, and furniture, all of which have relatively high labor inputs in their manufacture. The impact of technological change on the distribution of income between producers and consumers is probably the most powerful and pervasive of the distributional consequences of the HYVs and of other cost-reducing technologies in agriculture. Gains from technology accrue both to those who produce a commodity and those who consume it. Distributional impacts hinge on the income levels of major consumers and the response of consumption of the product to changes 23 in price. Because the demand for basic foods, such as HYV cereals, tends to be inelastic with respect to price, any significant increase in output will lower the real price of food. Since low-income consumers usually spend a higher proportion of incremental income on food than do high-income groups, lower food prices confer disproportionately high benefits on the poor, among whom are generally included agricultural laborers. Recent studies illustrate these points. With the introduction and rapid diffusion of dwarf rice varieties in Colombia, the real price of rice to consumers fell by 15 percent to 20 percent. Since rice is the basic source of calories in that country, low-income consumers captured a larger relative, as well as absolute, share of the gains to investment in rice research. It is estimated that, although the lowest income group in Colombia receives only about 4 percent of national household income, it captured 28 percent of the economic benefits of the research program. The Colombian experience is a good example of the manner in which agricultural research may effect a favorable change in the distribution of income among consumers. Results such as these have obvious implications for the choice of research priorities among commodities. In general, if the main goal of technology is to benefit consumers in the form of lower prices, research emphasis should be on agricultural commodities with relatively low elasticities of demand, such as foodgrains, beans, and root crops. On the other hand, if technology is to benefit primarily a country's producers, research should then be directed toward products with higher elasticities of demand, such as cash crops, export crops, and crops with attractive fixed prices. The socioeconomic impacts of agricultural research, as embodied in the HYVs, therefore, are complex and can only be determined through careful, case-by-case analysis. The introduction of HYVs of wheat and rice has had a positive effect on food production in developing countries. Major beneficiaries have been lower-income consumers. Incomes of those who have adopted HYVs have generally risen, although differences in absolute income levels between large- scale and- smallholder farmers have widened due to the unequal ownership of productive assets. The direct employment effect, in most circumstances, appears to have been marginally positive, and may be more significant in the rural nonfarm sector than in agriculture. Wage levels have increased in a few areas, but less rapidly than land values, thereby worsening the relative income distribution between landowners and laborers. The total availability of protein and calories appears to have exceeded that which would have prevailed in the absence of the HYVs, although full knowledge of the nutritional impact of HYVs requires more information on patterns of consumption. On the negative side, there is little doubt that the technology embodied in the HYVs now available has widened regional income disparities-a reflection of the fact that many of these varieties are not suited to less-favored environments. Disparities in rural incomes have also increased for the resource-poor farmers who inhabit those ill- 24 suited areas. The number of these farmers is substantial. In Asia, for example, only 20 percent of the area under rice is irrigated. Throughout the world, the small-scale producers of rainfed and deepwater rice, of millet and sorghum, and of many of the pulses, tubers, and root crops have yet to benefit substantially from the new technologies. Agricultural research, as an adapter or generator of new technology, has been a powerful means to realize objectives of economic growth, particularly in better-endowed agricultural areas; but even there, it is a relatively blunt instrument when used in pursuit of distributional goals among agricultural producers. Issues of Organization and Management In many cases, national research programs in developing countries are among the weakest links in the global agricultural research effort. This is particularly unfortunate, because national research may also be one of the most important components. Despite wide differences in research capabilities among countries, a number of common weaknesses can be identified. A pervasive problem is the inadequate appreciation by govern- ments of the role that effective research plays in agricultural growth and development. This attitude has several consequences. Low priority, in terms of budgetary and manpower support, is given agricultural research by governments. Deficient organizational structures for research and extension result in fragmentation of effort, weak linkages between research and extension, and the generation of research results that have little practical value to farmers. Training programs for research scientists and technicians are generally deficient. Country economic policies, which frequently emphasize low food costs, may discourage the adoption by farmers of improved technologies. Organizational structure No single model can be recommended as the most effective approach to the organization of a national research effort, as situations in individual countries are unique. Nevertheless, a review of experiences in several countries suggests that several basic principles are widely applicable in the design of effective research programs: (a) research goals must be clearly stated; (b) there must be continuity in goals, management policies, and supervision of the research program; (c) the research agency must have an acceptable degree of autonomy; (d) the national research effort must be provided with technical staff and financial support consistent with the severity of the problems that limit agricultural production; (e) there must be a continuous flow of information to the research staff about production problems confronted by farmers; (f) simultaneously, mechanisms must be provided to permit research results to flow to farmers, with emphasis on onfarm testing; and (g) the government should encourage arrangements that facilitate the coordination and cooperation of a 25 country's total research establishment (public or private) in pursuit of common objectives. Research organizations in a growing number of countries embody these principles. In these systems, national research organizations are frequently structured on the basis of major commodities (crop and animal) and production Jactors or "special problems" research (soil, water, engineering, national resources, processing). A socioeconomic research division is an integral part of such a national research system. Policy guidance to the organization is provided by a high-level board that represents the various interests in agricultural development. The research organization, as a semi-autonomous public agency, is provided with considerable administrative flexibility. Subject matter specialists, working at research stations, provide linkages with regional extension services, area development projects, and commod- ity production programs. A network of research stations is designed to meet the needs of the major commodities and the principal agro- ecological areas. The generation (research) and diffusion (extension) of agricultural technology should be closely linked. Separation of these activities-in time or organizationally-usually disrupts the process of technologi- cal transfer and reduces the effectiveness of the total research effort by limiting its relevance to farmers. Fortunately, the importance of this linkage is increasingly being recognized and, as a result, several national research and extension systems are being reorganized. Expanding the use of the Training and Visit system of extension, where it is applicable, is an additional and valuable means of providing the essential linkage between those two activities. The system recognizes that the views of farmers are as important to effective research as is the flow of technical information to farmers. It also encourages essential on-farm testing of new technology. In many cases, up to 80 percent of the total research effort can be carried out in fields away from the central or regional experimental station. Other approaches can be taken to foster the essential link between research and extension. These activities can be combined organiza- tionally under one director. Or, based on experience in several coun- tries, an effective organizational arrangement may be to establish research and extension activities in an autonomous, or semi-autono- mous agency-in effect, creating an agricultural institute of technol- ogy with responsibilities for both the generation and diffusion of technology. Career development and training of researchers and technicians Experience suggests that the extent and effectiveness of the diffu- sion and adoption of technology across international boundaries de- pends heavily on the abilities of national research systems. The country without the capacity to carry out research on its own benefits very little from research done elsewhere. A developing country's ability to screen, borrow, and adapt scientific knowledge and tech- nology requires essentially the same research capabilities as those 26 needed to generate new technology. Yet, few national systems have so far developed the administrative and technical capabilities to ab- sorb and adapt, in an effective way, knowledge and technology that is becoming available to them from work at the international centers and research institutions in developed countries. This has implications for both the numbers of agricultural scientists needed and the type of training most appropriate for the work they will be expected to do. A high degree of conjecture is involved in estimating the number of research workers needed in developing countries. Weaknesses in base data on current numbers of research scientists add to the difficulty. One approach is based on the area under a particular crop. Assum- ing that at least 100,000 hectares of a crop are needed to justify a significant research effort at either the country or regional level, it is possible to develop crude estimates of the number of research scientists required globally. In 1975, there were approximately 400 areas of this size in developing countries planted to major food, fiber, or oilseed crops. Full research coverage at present in the developing countries (excluding China, for which relevant data were not available) on the basis of a headquarters station team of 20 scientists and 4 regional stations, each with 6 scientists, would today require at least 24,000 research scientists for crop research alone, or some 4,000 more scien- tists than were in place in the mid- 1 970s. If additional research needs in livestock, fisheries, and agricultural economics are added, at least a doubling of the number of scientists for national research programs would be necessary. If this doubling of research staff were to be achieved over a 15-year period, the annual increment needed in the next few years would approach 4,000 trained staff.5 Additional num- bers would be needed for expanded agricultural extension programs as well, although some of these personnel would probably require training of a shorter duration. These figures can be compared with the estimated 7,000-8,000 students from developing countries now in training at the graduate or postgraduate level in technical agriculture or related subjects.6 Probably less than one-half of this total number complete their train- ing each year, and many are likely to find nonresearch-related em- ployment. In addition, not all trained agriculturalists can be ex- pected to perform well as research scientists. Significantly expanded training programs would be needed to meet the research require- ments. There also is a need to expand the type of training that develops a capacity to generate, adapt, and apply technology. Many universities 'This assumes a 10 percent annual rate of staff replacement and a 30 percent "dropout" rate for new graduates. These are approximate figures provided by the International Food Policy Research Institute (Washington, D.C., U.S.A.). Some estimates from other sources are higher. Of this total, the training of 1,500 is funded by bilateral agencies or private founda- tions. Some 800 are trained at the international or regional research centers, 400 in FAO/UNDP and multilateral/bilateral projects, and 300 in World Bank-assisted and regional development bank-assisted projects. 27 concentrate on theoretical principles in particular disciplines; a sound training program requires that scientists learn, in addition, those research skills that will permit them to apply their specialized knowl- edge as members of multidisciplinary teams whose primary objective is to generate efficient agricultural technologies. One example is the training program of the International Maize and Wheat Improve- ment Center that emphasizes training for the development of technol- ogy-to produce a "critical mass" of trained scientists capable of conducting effective crop improvement programs in their respective countries. A related issue is one that involves finding the appropriate balance among training activities. Several levels of skills and disciplines can be identified as needed in national research systems. These include senior managerial and planning staff, scientific workers of various disci- plines, intermediate-level technicians, managers of experimental sta- tions, and production agronomists to link research and extension programs. Among the scientific disciplines, severe shortages of trained personnel exist in soil microbiology, tropical pastures and forestry, agricultural engineering, and socioeconomic research skills, to name just a few. Additional training opportunities for national research planners, administrators, and experimental station managers are also badly needed. It is important that developing countries analyze their manpower needs in agricultural research, both with regard to num- bers and the type of training required, and to devise manpower development plans based on national research priorities. An additional factor contributing to the limited effectiveness of research systems in some developing countries is the professional orientation of research workers. Not infrequently, higher-level train- ing concentrates on academic research of purely scientific interest (the science component of research activity defined in Chapter 1), rather than on the production problems of farmers (the technology develop- ment function). Research scientists in developing countries frequently have adopted the approach of the public research establishment (uni- versities and their corresponding academic and scientific professions) of developed countries that tends to focus on the generation of knowl- edge rather than the generation of technology. In developing coun- tries, opportunities for research scientists to choose between financial and professional rewards are limited. In such an environment, it is almost inevitable that professional rewards, gained through academic research and publication in journals, will exert an important influ- ence on the type of research preferred by overseas-trained scientists when they return home. Whenever possible, research in support of dissertation subjects should be undertaken in a scientist's home coun- try or on a problem of direct relevance to his country. While the generation of both knowledge and technology is needed in agricultural research systems, it is important, in the organization of many national research systems, to give relatively greater emphasis, than has been done so far, to the generation of technologies that can be of direct benefit to farmers. For this to happen, a reorientation of the national research effort is frequently needed. But, more importantly, 28 it requires a system of monetary incentives to competent research scientists so that their efforts toward the generation of technology-a research activity where professional rewards in many developing countries are limited-might be encouraged. The problem of smallness There is no clear agreement on the volume of agricultural produc- tion necessary to justify a minimally adequate research program in a particular country. Many small countries do not need, nor can they support, the comprehensive research systems appropriate for larger countries. But they frequently do require smaller research programs that can focus on farming systems and commodities of major impor- tance. Moreover, even small countries may have several agro-ecologi- cal zones with different farming conditions. Flexibility, therefore, is needed in considering research needs to ensure that the national program is tailored to a country's resources in trained manpower, research facilities, and equipment. A first approach might be one that considers the development of one or two research stations in the most important agricultural regions of a country. A small, multidisciplinary team of scientists, together with adequately equipped support staff, could work on the major problems facing farmers of the region. If experimental stations already exist, external support may be required to upgrade their programs and facilities, or to provide additional research stations in other important regions. It is particularly important that funding for professional- level and inservice training for research scientists and administrators is provided in these circumstances. In the smaller or poorer countries, external support for much of the incremental operating expenditure may be needed for the duration of the project. A clear understanding of how such costs are to be met in subsequent years would be required. Small countries generally cannot afford to conduct basic research, except possibly at a university; instead, they should devote most of their efforts to applied research at an experimental station and on farms. Much of the information that is needed may have been devel- oped elsewhere with practices and materials that can be adapted to local conditions. The resources of the international research centers are particularly important to these countries, and they should be fully utilized as a source of improved practices, materials, and training. Greater regional collaboration on research matters may also be help- ful, particularly in cases where similar agro-ecological conditions exist within regions, and in regions where common development problems prevail. In Central America and Panama, for example, there are good opportunities to strengthen the generation and transfer of technology through improved intraregional and extraregional lin- kages among institutions and programs having common goals. With greater coordination, many of the specific technological needs of the region, such as those pertaining to major cereals, food legumes, root crops and fibers, livestock, farming systems, and seed production, could be met through linkages of this type. These include linkages to 29 research systems outside the region (such as to the international centers), to regional research organizations such as CATIE (the Center for Research and Education in Tropical Agronomy, Turri- alba, Costa Rica) and IICA (the Inter-American Institute of Agri- cultural Sciences, San Jose, Costa Rica), and to national research systems. These efforts must be complemented by local testing and adaptive research. Regional difJerences External assistance to strengthen national systems can be effective only if differences in size, present state of development, and existing research capacity are recognized and taken into account. For many countries, intracountry differences must also be considered in the design of national programs. Some important regional characteristics can be identified: - Countries in the Eastern Africa region tend to have weak na- tional research organizations, ineffective links with extension services, and a scarcity of financial resources to support a na- tional research program. The serious shortages of qualified re- searchers are exacerbated by attractive employment opportuni- ties in the neighboring Arabian Gulf states. In some countries, there is a reluctance to borrow funds for research. Instead, countries expect that funds will be provided as technical assis- tance on a grant basis. - In Western Africa, many of the countries suffer from a chronic shortage of funds to establish and operate effective national research programs. The number of qualified national research personnel is very limited, and the national research efforts are fragmented and uncoordinated. Much of the research is along traditional, discipline-oriented lines, and is not directly related to solving problems in food production. Though extension is frequently a part of area development programs, it is rarely linked to a national research program. The difficulty and high cost of hiring expatriate scientists add to the problems of mount- ing an effective national research program. These difficulties have been partly overcome by the provision, on a grant basis, of expatriate research personnel through bilateral technical assis- tance programs. - Few generalizations can be made for the diverse Latin Ameri- can and Caribbean region. Several Central American countries are characterized by shortages of trained personnel and poorly articulated links between research and extension. In the Caribbean, these problems are further accentuated by the problems of smallness (noted previously). In several of the larger countries of Latin America, the number of qualified personnel is generally adequate relative to needs, and key research infrastructure is in place. However, in some of these countries, the low priority given by governments to agri- cultural research has resulted in poor morale in national 30 research organizations and the loss of some available qualified personnel. - In the Middle East and North Africa region, the state of development of national research systems varies widely among countries. In the more developed countr-ies of the re- gion, sound national systems are in place and are functioning reasonably well. In others, very little systematic research has been carried out, and governments attach little importance to increased efforts in this field. In the less developed countries in the region, weak research organizations and shortages of qualified personnel make expansion of the national research effort particularly difficult. - Among the larger countries of South Asia, the number of qualified personnel may be adequate to carry out an effective national research program. However, inadequate physical facilities tend to limit a program's effectiveness, as does a shortage of operating funds. Linkages between research and extension are improving. In the smaller countries of the region, the research infrastructure is less well developed, and there are relatively fewer trained personnel. - Several countries of the East Asia and Pacific region have a long history of research that is oriented toward commodities and export crops; the research has been both well funded and effective. Foodcrop research is less well developed and has tended to receive less attention from governments. Most of the countries of the region suffer from a fragmentation of research effort, weak links with extension, and inadequate numbers of staff. In a number of the larger countries, national research programs are being reorganized and strengthened. Most of the developing countries can usefully be divided into four categories on the basis of the characteristics mentioned earlier. I . A small number of these countries, perhaps fewer than 10 percent of the total, have generally adequate levels of man- power available for both the management and conduct of research and have developed, or have taken substantial steps toward developing, a well-organized research infrastructure. The national research establishments of these countries, which generally have effective linkages with research in- stitutions in developed countries and with the international research centers, are largely self-supporting and have little need for external technical assistance-either at the planning or operational levels. These countries may well be a source of expertise for others. 2. Another 10 percent of developing countries typically have generally adequately trained staff, but their research activities are frequently fragmented, uncoordinated, and isolated from the development process. These countries frequently require assistance in organizing an appropriate administrative struc- ture for research, in managing the national research efforts, 31 and in supplying specialized skills in key areas. 3. Though nearly half of all developing countries are large enough to require a balanced national agricultural research system, they lack, nevertheless, essential infrastructure and usually have only a limited number of crop research pro- grams. The needs of these countries generally are twofold: to develop an effective organization and management system for research and to expand and strengthen the scientific personnel base to conduct research. 4. The remaining countries-about 30 percent in all-are those with very limited research resources and no single crop of sufficient importance to warrant a comprehensive research system. The major need of these countries is to develop a limited research capability, in terms of trained manpower and physical infrastructure, to carry out research, largely of an adaptive nature, for a small number of commodities that are of greatest economic importance to them. Agricultural research and national economic policy Farmers will adopt new technologies only if they perceive that it is in their economic interest to do so and if necessary support services are adequate. Adoption of technology by farmers, rather than its mere development, must be a basic measure of research effectiveness. Demand for the HYV technology made available in the late 1960s and early 1970s was high, and adoption was correspondingly rapid, because this technology was unusually efficient in physical terms and because attractive price relationships existed between inputs (largely fertilizer) and output. When new technologies are disappointingly slow to be adopted, as is often the case on small farms, the explanation may well lie in the perceptions by smallholder farmers that few compelling economic reasons exist for adoption. Promised increases in production may not be great enough to make it worth their while, or price relationships between inputs and outputs may be unattractive. The basic problem is often the government's price policy, which is to keep food (and farm) prices low for the benefit of politically stronger and more vocal urban groups. Persistent underpricing of foodgrains in many coun- tries is an important force that limits both the demand for research and the diffusion of available technology. Distortion of input and output prices affects farmers' short-term decisions regarding the use of inputs and the commodities they will produce; over a longer period of time, it also determines research priorities and, thus, influences the future supf ly of technology. Empirical work supports these conclusions. Analysis suggests that India, for example, may have substantially overvalued wheat in the domestic economy and undervalued rice, thereby distorting perceived rates of return to those respective research programs and affecting the relative rates at which new wheat and rice technologies were 32 adopted.7 Other work indicates that Brazilian trade and economic policies lowered the domestic price of agricultural products relative to fertilizer, resulting in a less than optimum use of fertilizer and an underutilization of a significant amount of research on soils and fertilizers. Further, because it was not profitable under prevailing price relationships to fertilize corn, the adoption of hybrid corn was limited. As a result, the economic return to research on hybrid corn was relatively low. In short, because economic policy caused Brazil to undervalue its research activities, less than optimal levels of re- sources were invested in research.8 A country's agricultural research program must be developed with appropriate regard for more general economic policies that determine the extent to which incentives exist to adopt new technol- ogy. Decisionmakers responsible for the development of a country's research system must also be concerned with economic policies. This point has not always been sufficiently appreciated. Too often the tendency has been for national economic planners to ignore the impact of their policies on the research establishment and for re- search managers to retreat into the more comfortable confines of the research institute. One consequence is that an efficient source of economic growth (agricultural research) is sacrificed and food defi- cits are larger than they need be. National systems and the international centers The serious deficiencies that exist in the research systems of many developing countries have been major reasons for the steady expan- sion of "off campus" activities by the international agricultural research centers. These outreach activities now involve some 200 staff of international centers who are stationed in more than 40 developing countries. The need to strengthen national research sys- tems has also generated some new institutional initiatives, including the privately funded International Agricultural Development Ser- vice (IADS) and, more recently, the International Service for Na- tional Agricultural Research (ISNAR), which is being funded through the CGIAR. The international centers are not intended to substitute for na- tional research systems, although, in some cases, they have assumed institution-building roles because the national systems were so weak as to impede effective collaboration. Nevertheless, institution build- ing, except through training of national staff at the centers, is not a function of the centers, nor is it a field in which the centers have a significant comparative advantage. Despite rapid growth in recent years, the outreach activities of the international centers can be 'V. A. Sukhatme, "The Utilization of High-Yielding Rice and Wheat Varieties in India: An Economic Assessment," unpublished Ph.D. dissertation (Chicago: Univer- sity of Chicago, 1976). 'G. Schuh and H. Tollini, CosisandBenefits of Agricultural Research: The Stateof the Arts. Staff Working Paper No. 360 (Washington: World Bank, 1979). 33 expected to grow much less rapidly in the future. There are two reasons for this: research capability has expanded in some national systems, and concern has increased among CGIAR donors over the possible adverse effects that outreach activities might have on the research efforts of the centers. An effective ISNAR will also reduce the need for the outreach efforts of the international agricultural research centers. As national research systems become stronger, the role of the international agricultural research centers may change in other ways, as well. Since national systems can be expected to take over more of the task of produc'ing new technology (as contrasted with the previous emphasis on adapting and verifying technology produced at the international centers), the international centers could function more as clearinghouses for highly specialized knowledge and genetic material. The training of national research workers will continue, however, to be an essential function of the international centers. By attaching importance to the task of "training trainers," the centers will be able to maximize their impact on the national system and the basis for larger, more effective training activities in those systems will be better established. The international centers might also provide highly specialized training in particular disciplines or prob- lems. Evaluation of research activities As the financial resources committed to agricultural research in the developing countries become larger, basic questions will arise concerning the effective use of these funds. Simply put, agricultural research, like all economic activities, competes for the allocation of scarce resources. In determining the returns to extension projects, the World Bank first estimates the increase in agricultural produc- tion that is necessary to generate an internal rate of return that is at least equal to the opportunity cost of capital. A judgment is then made as to the feasibility of achieving such an increase. This methodology, while very rudimentary, provides some basis for deter- mining a project's worth. The weaknesses of this approach relate to the uncertainties of estimating quantities and values of future output and to the narrow analytical focus that considers increased output as the sole objective of an investment. Continued efforts are needed to refine this procedure and, where possible, to extend this approach to research activities and to develop approaches that provide more information on the impact of research investment on other socioeco- nomic variables, such as income distribution and nutrition. Chapter 3: Agricultural Research: The World Bank's Experience Support of agricultural research activities received low priority in World Bank lending until the mid-1960s. This was partly a reflec- tion of: (a) the view that technology already available, if properly applied, was adequate to achieve production objectives; (b) the commonly held opinion that the major factor inhibiting rural growth was inadequate capital investment, not a shortage of good technol- ogy; and (c) the fact that Bank lending was only gradually becoming concerned with integrated development and farm-level inputs. In more recent years, increased funding for research reflects both the Bank's greater emphasis on lending for agriculture and rural devel- opment and the recognition that improved technology suitable for low-income producers, especially for use in projects located in areas with difficult agro-ecologic conditions, is relatively scarce. In the period fiscal 1977-79, lending for agricultural research and exten- sion constituted almost 9 percent of total Bank lending for agricul- ture and rural development. About 30 percent of this lending has been for research alone. Types of Bank Support For purposes of discussion, four general types of Bank support for agricultural research may be identified: (a) lending for agriculture and rural development projects that have adaptive research and extension components; (b) lending for national or statewide research and extension projects; (c) lending for education projects that include agricultural research components; and (d) financial support to the CGIAR. Lending for national or statewide research and extension projects has been both the largest and fastest growing category, with 29 projects, totaling $721 million in Bank commitments, approved during fiscal 1970-80. During the period fiscal 1970-80, $171 mil- lion was lent to support research activities in 312 agricultural and rural development projects, while $51 million was lent to support agricultural research components in 18 education projects (see An- nex 7). Research Components in Agriculture and Rural Development Research components in Bank-assisted agriculture and rural de- velopment projects are designed to adapt or develop knowledge or techniques for specific agro-ecological areas and improve national research capabilities through training and technical assistance. Many of these components originate in the need to develop improved technological packages for use in rural development projects. Their inclusion helps ensure that appropriate research will be carried out at the farm level. A particular focus has been to develop relatively 35 simple technologies that require a minimum of purchased inputs. Unless a country has a well-established national research strategy, the incorporation of such components can: (a) result in incomplete research coverage of subsectors and regions within a country; (b) lead to unnecessary duplication of research efforts; (c) generate research that has no enduring impact in terms of significant technological or institutional development; and (d) cause the work at research facilities to be diverted to those activities of importance only to particular projects and away from problems and opportunities that may affect large numbers of rural people who do not live in the project area. Since the need for supplementary research to refine the techaiical packages that are incorporated in projects is indisputable, it is inmportant that such research be done in the context of a national research strategy. The creation of such a strategy can be best achie!ved through projects that build or strengthen a national or state agricultural research system. Just as there are sectoral standards that are applied to the design and levels of various infrastructure and service projects, so, too, it is imperative that sectorwide standards be established and adhered to in the design of strategies and project lending for agricultural research. An example of such standards is the number of support staff to be provided to each research scientist. Some research components in agriculture and rural development proj,!cts appear to have been poorly designed or are too small, either in scope or in financial terms, to be effective. Some have neither specified clearly what the research problem is nor how the problem is to be tackled. Because of the time lag between research work and practical results, research financed through some project compo- nents has not yet been able to provide the expected technical under- pinr ing for the main project. These components are frequently less careFully supervised and the need for their timely implementation rece .ves less emphasis than do the main project components. Because the 3eed for adaptive research in these projects is great, research com?onents will continue to be required. But their design, time phasing, and linkages to the national research system warrant greater attention. Nati.onal or State Research and Extension Projects Projects that seek to upgrade national (or statewide) agricultural research and extension programs have varied widely in emphasis- from pure research at one extreme to mainly extension support at the othe:-. These projects usually involve substantial restructuring of existing government research and extension services, including a redirection of their goals and an increase in their staffing and fund- ing. The projects have stressed the establishment of research work on speci alized commodities and farming systems that is tailored to spe- cific agro-ecological areas and integrated into the national research systems. In the preparation of these projects, the Bank has in- creaRLD BANKC . - - a -- .- He4quarters -< r_ 1 -18 i8-HS t, N.W. t~ ~~ahntz a - 2O4), 1JSA. ^ n -~~t1ehR (202) 477-1234 Si att>'˘_ WASjJfNGTONDC t - - ~~~Eurpe OflDc-e-- ., , [oks~~?llsPis, Eilda -c& -1v , Ma- noUCb 3-.home Chiyoda-ku,-TokYo0,Jaa ::f f : f t~~~