PHN Technical Note 87-21 1t NUTRITION REVIEW FILE CONY by Alan Berg December 1987 Population and Human Resources Department World Bank The World Bank does not accept responsibility for the views expressed herein which are those of the author(s) and should not be attributed to the World Bank or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank; they do not necessarily represent official policy of the Bank. The designations employed, the presentation of material, and any maps used in this document are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates concerning the legal status of any country, territory, city area, or of its- authorities, or concerning the delimitations of its boundaries, or national affiliation. V11ME COP PHN Technical Note 87-21 NUTRITION REVIEW ABSTRACT The report reviews PHN project and sector work, and Bank-financed research in nutrition from 1976 to 1984, identifies major successes and weaknesses, and recommends future directions for the Bank's work in nutrition. Sixteen sector studies in nutrition are reviewed and. the report recommends that sector work in nutrition be intensified. The report also reviews 55 pieces of nutrition-related research either undertaken or financed by the Bank and recommends that the Bank's future nutrition research program concentrate on operations and project impact, the nutritional effects of agricultural actions and consumer food subsidies. Four free-standing nutrition projects were initiated between 1976 and 1980. The most important contribution of these projects appears to have been their impact on the priority, commitment, policy and program choices, and level of country support for nutrition actions. On the basis of the experience gained in these projects the report recommends three operationally implementable project concepts for future Bank projects, i.e. nutrition to reduce infant and child mortality and morbidity, nutrition to improve human capital formation and labor productivity, and control of micronutrient deficiencies. Nutrition work was also undertaken as components and sub-components within projects in other sectors (i.e. agriculture and rural development, urban development, education, population and health). In general, although there have been successes among the nutrition components in other sectors' programs, by and large the efforts have not been successful. There are several reasons for this including inadequate design, promotion and follow- through, limitations of time and interest of projects staff, increased project complexity and the sectoral organizational structure of the Bank. The Bank's role at this stage should be to help countries develop sound nutrition policies and strategies; make on-going nutrition intervention programs more efficient and mount new affordable programs to address identified nutritional needs; and strengthen the institutional capacity to plan nutrition actions. A good part of this can be handled within health and population projects, either by incorporating nutrition as one of the main themes or by adding nutrition components. In some instances, particularly those involving intersectoral issues, it would be appropriate for nutrition projects to be free-standing. Prepared by: Alan Berg Population and Human Resources Department December 1987 This report benefited from papers wriLten by Joanne Leslie (EDI), T. J. io and Lauren Chester (PHN), and George Beaton, Robert Muscat, James Pines, and Per Pinstrup-Andersen (consultants). It also benefited from the insights and general guidance of an informal working group consisting of Stephen Denning, James Greene, Dean Jamison, Anthony Measham, Emmerich Schebeck, and Mss. Ho and Chester and the support work of Ms. A. Menciano. This paper reflects the consensus of a PHN retreat on the subject and subsequent PHN staff discussions. -- Alan Berg NUTRITION REVIEW Page SUMMARY AND CONCLUSIONS i - vi I. BACKGROUND 1 II. THE BANK-SUPPORTED NUTRITION PROJECTS 4 Brazil Nutrition Research and Development Project 6 Indonesia Nutrition Development Project 13 Colombia Integrated Nutrition Improvement Project 19 Tamil Nadu Nutrition Project 23 III. LESSONS FROM PROJECT OPERATIONS 28 Project Content 28 Project Design 33 Experience in Implementation 34 Costs and Cost-Effectiveness of Nutrition Projects 43 Impact of Nutrition Lending 52 Bank Management of Nutrition Projects 55 IV. NUTRITION IN OTHER BANK-SUPPORTED PROJECTS 56 Projects with Nutrition Components 56 Nutrition Orientation of Projects in Other Sectors 58 V. ANALYTIC ACTIVITIES IN NUTRITION 60 Country Economic and Sector Work 60 Nutrition-Related Research 62 Others' Findings on Nutrition 64 VI. APPLYING THE EXPERIENCE TO FUTURE BANK ACTIONS 65 Project Experience 67 The Issues 70 Future Directions in Nutrition in the Bank 77 APPENDICES A -- Background Papers to This Report B -- Health and Nutrition Indicators in Bank Nutrition Project Countries C -- Bank Nutrition Research D -- Cost-Effectiveness Tables E -- Nutrition Components F ---Economic and Sector Work G -- Relevant Findings from Recent Research and Experience by Others SUMMARY AND CONCLUSIONS Experience has demonstrated that (i) although malnutrition is part of the poverty syndrome, much nutritional improvement is possible without major income increases, and (ii) there are several efficacious, affordable means by which Bank nutrition projects could make an important contribution in this regard. Other findings of a review of four largely experimental nutrition projects include the following: (a) for a variety of reasons the first three nutrition projects were too complex; while malnutrition does indeed have multiple causes, a single project might best be directed at only a few of the causes; (b) despite overcomplexity and slow starts, each of the first three projects (Brazil, Indonesia, Colombia) finished or is. finishing rather well, and Tamil Nadu, with a more focused design that benefited from experience with the three preceding projects, is being smoothly implemented. (The nutrition projects did well in the areas of cost estimates, procurement, loan covenant compliance and project completion; they did less well in disbursements and project management); (c) lack of a single responsible entity within countries for dealing with nutrition increases the difficulty (for Bank and Borrower) of managing nutrition projects; this lack suggests that a premium be put on institution-building as an activity within nutrition projects. Certain project experience in this respect has been encouraging but the diversity of country circumstances precludes general lessons; (d) project experience has shown the practical feasibility of targeting the provision of food and services to maximize nutritional impact and the cost savings that targeting can lead to; and (e) the projects were an important catalyst in attracting both policy attention and resources to the problems of malnutrition; also, they demonstrated the potential of nutrition .projects for contributing substantially to the development of primary health care and family planning programs. Costs and Cost-Effectiveness Compared with earlier style nutrition interventions (which generally had,either indiscriminate coverage or, if targeted, often involved one-on-one treatment and oversight from highly trained professionals), the large-scale concepts tried under the 3ank-assisted project efforts appear to have shown the feasibility of pushing per capita beneficiary costs down to relatively low levels. In some instances these - ii - costs could be met through restructuring of existing programs. However, even if entirely additive, at these levels most of the interventions appear to be affordable if extended to a national scale. The low-cost nutrition education as practiced in Indonesia looks particularly attractive. That it was cheaper than programs requiring food commodities comes as no surprise; the question is whether it is effective. The evidence has shown that nutrition education alone can make a difference in improving nutritional status. Nutritionists have long held out the promise of this possibility; the Indonesian experience is the first time it has been demonstrated in an operational setting. Even at the high end of the cost spectrum, a consumer food subsidy program if targeted properly may be affordable in certain contexts. Vitamin A programs clearly are cost effective. Brazil's preschool feeding and stimulation program is costly relative to other nutrition interventions but may be a less costly way than added construction and other conventional approaches to strengthening an education system. The Tamil Nadu approach is more costly than some and less costly than other common approaches to preventing death. The other benefits of that program, however, go well beyond most of the health approaches. Experience has shown that some governments are ready to commit substantial resources to nutrition and are interested in collaboration with the Bank. Comparative Advantage The Bank's comparative advantage in nutrition derives from the (relatively) large scale of its projects; from its capacity for undertaking sectoral analyses (and for reflecting their findings in Bank dialogue with planning and finance ministries); and from its experience in project planning and management. The Bank lacks comparative advantage in upstream pilot project work (where UNICEF and bilateral lending agencies with field staffs and greater speed in programming have a clear advantage); in financing food aid or redistribution programs (the World Food Program and bilateral agencies are involved here); and in food relief operations (the U.N. High Commission for Refugees and bilateral agencies are involved). Close working ties with other agencies are thus desirable and such collaboration is feasible. Of the agencies, WFP holds special untapped promise. The Bank can benefit also from the technical skills of staff from the specialized agencies of the UN system. Nutrition in Health Projects and Population Projects Future health projects and population projects involving MCH and primary health care should be expected to include nutrition components aimed at preventing and treating severe cases of malnutrition, particularly among mothers and young children. Most commonly this would consist of nutrition education (concerning breast feeding, weaning foods, food hygiene, family food distribution, and related practices), growth monitoring, oral rehydration therapy, and selective supplementary feeding and rehabilitation of seriously malnourished children. The priority assigned to nutrition interventions relative to other interventions needs, of course, to be assessed separately for each project. (The main question is what the absence of nutrition will do to the expected outcomes of the project, and the advantages of including nutrition.) - iii - Nutrition Projects The Bank should be prepared to lend for free-standing nutrition projects. Three operationally implementable project concepts emerged from the review and subsequent discussion. These, of course, are not mutually exclusive. (a) nutrition to reduce infant and child morbidity and mortality and promote child growth; (b) nutrition to improve human capital formation and labor productivity; and (c) control of the major micronutrient deficiency diseases -- iron deficiency anemia, iodine deficiency goiter, and vitamin-A deficiency induced blindness. As malnutrition and undernutrition are major underlying causes of infant and child morbidity and mortality, the provision of food supplements for a limited period to those identified with faltering growth may often be a cost-effe.ctive means of addressing major infant and child health problems. While this type of nutrition intervention -- "food as a medicine" -- may be integrated into a primary health care project, it might also stand on its own as, for example, in the Tamil Nadu project where the health system provides effective referral capability and permits the necessary synergism to be realized. In countries concerned about nutrition of other members of the family as well as the very young, consideration might be given to food. delivery projects aimed at enlarging the family food basket and contributing to the physical and intellectual development of family members, rather than being of a curative nature. The intent would be to increase the availability of food to the household without simply waiting for economic development; indeed, a major rationale for such interventions is that by increasing the production of labor, they will help to increase the rate of economic growth. This type of project might range from institutional feeding to targeted subsidy programs in the commercial market to efforts to improve the efficiency of the food marketing system. Except for the latter, all involve important recurrent costs. Because issues of food subsidy, rationing, and distribution involve important macroeconomic considerations, close coordination with country economists will, be important in formulating possible operations in this area. A third kind of project would deal with the delivery of micronutrients. Major inroads are now possible against nutritional anemia, vitamin A deficiency blindness, and goiter and mental retardation resulting from iodine deficiency. A low-cost technology now exists that makes large-scale fortification of salt with an iron compound feasible and attractive. (In very large countries, a single component project on this may be appropriate.) Sugar can be used as the vehicle for vitamin A, and mass doses are also a well-established approach to meeting vitamin A needs. To avoid duplicating efforts for iodine fortification or mass dose iodine and vitamin A programs the Bank should work with WHO and UNICEF, since they have long been involved in such programs. - iv - Certain components common in the early projects (especially food production, water supply, and food technology) generally will only be included in future nutrition projects when special circumstances dictate a need. Most nutrition projects should provide some form of nutrition education, an area in which the Bank has developed considerable expertise. Training, technical assistance, and development of planning capability (including establishment of nutrition surveillance systems) would commonly be included in projects to help strengthen institutions. Future projects will not have the research orientation that characterized the early projects. Some of these same activities could also be considered as nutrition components in health or population projects or projects in other sectors. The above directions are not new, in the sense that current policies permit them already. They are, however, new in the following ways: (i) the attention given to nutrition in population and health projects would be more systematic and more rigorous than in the past; and (ii) free-standing nutrition projects would be more narrowly focused than the early broad multisectoral projects. The Bank now has concrete models for nutrition projects. In short, as a result of the experience and the review process, concepts have been clarified and interventions have been operationalized. The "food as medicine" and micronutrient project types above clearly fall within the province of PHN. Consumer food marketing, though, is an activity that falls between the cracks in the Bank organizational - structure. PHN is not staffed in this area, but other departments appear at the present time to be no better equipped than PHN to deal with this. The main consumer food marketing activities to date have been undertaken in the experimental nutrition projects and since the impetus for such activity on nutrition and health grounds is most likely to come from PHN, it should be prepared to proceed cautiously in this area. Many countries, particularly those in sub-Saharan Africa, are not yet capable of absorbing nutrition projects of the free-standing type. PIN should be alert, however, to opportunities to include preparatory work (for strengthening institutions and for studies leading to subsequent project development) as components in health and population projects. Also, as many of these same countries are in need and recipients of food aid but often are limited by their logistical capacity to absorb the quantities needed, appropriate Bank projects could include components designed to help expand that capacity. Sector Work Sixteen sector studies of country-specific malnutrition problems and their determinants are reviewed and lessons the Bank has learned in doing this work also are discussed. In nine of the countries the evidence is strong that nutrition problems are serious and in several cases, where time series data are available, the nutrition status of the poorest has been deteriorating. Seasonal malnutrition was found to be of particular importance in nearly half of the countries. - v - One of the surprises emerging from the sector work is the discovery of the size and speed of a shift from traditional locally produced grains (e.g., millets and sorghum) to polished rice and refined wheat, both often imported. Administered prices, favoring refined wheat flour and rice, are a common feature behind these shifts. In some countries micronutrient problems have been detected, possibly stemming from these shifts. This is potentially a major nutrition danger that deserves attention. Sector work in nutrition should be intensified. Nutrition-Related Research While this paper largely is organized to deal with operational experience and its implications, the review of some 55 pieces of nutrition-related research undertaken or financed by the Bank has been an important part of the exercise. Pioneering attempts by Bank staff to disaggregate food intake by income group have led to the conclusion that in the normal course of development, even with a substantial expansion of food production, the nutrition problem is unlikely to be solved; special measures are needed to improve incomes of the poor, including intervention in food markets designed to fill the nutritional needs of the targeted population. Bank research has examined the economic value of improved nutrition. For example, an analysis of the costs and benefits of treating iron-deficiency anemia concludes that with an increase of 10 to 15 percent in hemoglobin, an increase of 25 percent in work output can conservatively be expected from the intervention. Bank studies on the impact of iron-deficiency anemia on labor productivity in Indonesia have made a major contribution in this area. The productivity of anemic workers who were given iron supplements for two months was found to have increased 15-25 percent, depending on the type of work performed. A second set of studies, focused on the productivity effects of food and iron supplementation for workers in Kenya, found that dietary deficiencies are an important cause of anemia, that anemia is an important determinant of worker productivity, and that provision of extra iron and treatment of hookworm are inexpensive and feasible interventions. . Studies of how infant and child malnutrition affects the future earning capacity of children and their contribution to economic growth have focused on the impact of school attendance and performance. Nepalese children whose height was considerably less than the norm were found, after controlling for income differences, to be much less likely to be enrolled in school than normal-sized children. In China low height-for-age was consistently found to be related inversely to performance. Analysis of studies in Brazil showed a useful advantage in subsequent school performance by those who had been part of a preschool nutrition program. Also, a study in Colombia shows significant effects of both nutritional supplementation and home education for mothers on children's diet, growth, and school performance. - vi - Finally, through this accumulated project experience and the concurrent research undertaken by the Bank, a good deal has been learned about nutrition economics, although much remains to be learned. The economic analysis employed by the Bank in conjunction with these projects constitutes a useful advance in the ways of thinking about costs and cost-effectiveness in this relatively new field. PHN operational division chiefs have been asked by the Department Director to discuss the above conclusions with their regional counterparts, to review for their countries the priority and potential for future Bank work in nutrition and to plan their operational work accordingly. I: BACKGROUND Recent completion of the Bank's first two nutrition projects marks an appropriate time to reflect on the Bank's nutrition experience to date and to consider the implications for future work. Such a review takes on added significance.during the present period in which, first,, severe economic pressures facing many countries are likely to have significant nutritional consequences for the poor and, second, there is growing skepticism about the Bank's ability, in general, to design effective poverty-oriented projects. Has the experience provided new 'insights about how to deal with the nutrition problem? Have effective, affordable actions been demonstrated? Does the experience offer ways, during periods of budget tightening, to bring about program efficiencies in nutritionally sensible ways? The first two loan agreements, for the Brazil project signed in October 1976 and Indonesia in March 1977, were preceded by discussions by the Board about the appropriateness of a role for the Bank in nutrition and the nature of that involvement. At the time of these discussions, it was recognized that the problem of malnutrition was substantial. The effectiveness of improved nutrition as a means of reducing the severity of childhood infections and of preventing forms of retardation, blindness, anemia-induced lethargy, and other problems had been established and was, of itself, regarded as sufficient justification for investment in better nutrition. The economic case for investing in nutrition had only recently been advanced and was less clear, but there were beginning indications that malnutrition had negative consequences for economic development. Nutrition actions at that time seldom were directed to the most vulnerable segments of the population. Most programs whose objectives had been deliberately concerned with nutrition concentrated on face-to-face techniques. Mass techniques such as fortification of centrally-processed foods, mass media nutrition education, consumer food subsidies for nutrition purposes, and adjustment of agricultural price policies to get a nutrition effect had seldom been employed. Most countries had little data on either food consumption or spending patterns of households with malnourished members or even on nutritional status. And despite substantial public institutional feeding programs, particularly school feeding programs; there had been little evaluation of their benefits. Rather than limiting itself to research to fill these data gaps or, conversely, moving ahead with full-scale projects in the face of scanty information, the Board decided that the Bank should finance certain high priority experimental actions and their evaluation. At the same time, the Bank would support training and longer-term analyses aimed at more sophisticated operations. In short, a learning-by-doing strategy was adopted. Board members generally favored "beginning cautiously, with a particular emphasis initially on the c6ntinuing dialogue with governments." They expressed concern that work of other development agencies not be duplicated. The management was authorized '"to proceed initially with a few projects from which it would learn" and also to try to introduce nutrition components in projects in other sectors. The Board recognized the amount -2- of technical assistance and staff time per dollar of lending "would be relatively high.. .unlike power financing or transportation where the reverse was true." Recognizing that the Bank "would be unable to achieve significant results, at least in the short run," the Board stressed its contribution "should include a large element of research." (Fifty-five pieces of research have since been completed.) Four nutrition projects were initiated between 1976 and 1980. In a departure from other new sectors, all projects were attempted in major Bank countries -- India and Colombia, in addition to Brazil and Indonesia. (During the years the nutrition projects were initiated, these four countries accounted for about one-third of all Bank lending.) The borrowers all had experienced economic growth but, despite this, continued to face significant malnutrition problems. The reasons varied by country but basically were either that large portions of the populations had been bypassed by the country's economic growth or that the problem was of a scale that decades of substantial equitable growth would be required to meet the needs. Nutrition issues received additional policy-level review in the Bank as one of the Basic Needs studied in 1979-80. A paper summarizing what the Bank had by then learned about nutritionl/ asserted that malnutrition of major proportions existed in most Bank client countries. Nutrition conditions of the poor were no better (and in many countries worse) than when the Bank began discussing nutrition. Increasing incomes and agricultural production was not likely to resolve nutrition problems within a generation in most countries. The basic problem was food energy (or caloric) insufficiency (sometimes complicated by specific nutrient deficiencies), and its adverse effects on development were considerable. Most governments were not reaching the very poor, especially the rural poor, with nutrition benefits, and few central ministries had the resources or the organizational capacity to mount a substantial nutrition effort. More than half of the malnourished in most countries were families of landless agricultural laborers, farmers with land holdings too small to be reached by rural development programs, small-scale fishermen, and the urban unemployed -- in short, those left behind by the development process. While the very young and pregnant and lactating women were the most vulnerable, nutrition problems were found to be important.among people of all ages and both sexes. Not enough food was available to reach them; they did not have economic and sometimes physical access to the food that was available, nor knowledge of how best to use their resources; their health was so poor that they were biologically inefficient in utilizing the food they consumed. The review of the nutrition problem suggested a range of responses: expanding food supplies (increasing production and reducing food losses) in ways that would benefit the poor -- with attention to what is grown, who grows it, what is stored; increasing incomes of the poor; improving marketing systems and adjusting pricing policies to benefit the poor consumer without destroying producers' incentives; introducing special 1/ "Nutrition, Basic Needs and Growth," PHN-AGR, August 1980. -3- feeding programs; educating families to change their food preferences-, preparation practices, distribution of food within the family, and hygienic practices; improving health and environmental conditions (water, sanitation, immunization, diarrhea control); and fine tuning interventions to solve problems caused by specific vitamin and mineral deficiencies. In summarizing the Policy Review Committee's assessment of the nutrition paper, the President of the Bank emphasized "the importance of malnutrition as a development problem and the need for a development institution such as the Bank to deal with it." The review mandated a further analysis of the operational aspects of the Bank's nutrition undertakings. At a meeting in September 1980 the Operational Vice Presidents decided that the scope and complexity of the problem demanded a broadening of the analytic effort. They endorsed a paper2/ that called for "a substantial effort through country economic and sector work to improve our understanding on a country-specific basis of the nature and extent of the problem, where in the chain of nutrition events the weakest links are that need addressi'ng, ways nutrition issues may be better integrated into country operations and, to the extent possible, identification of suitable actions." One to three countries in each region would be treated in some depth in economic and sector work. (By the end of 1983, analyses of varying intensity for 16 countries had been completed.) The 1980 review concluded that nutrition concerns should be explicitly incorporated in agricultural.and rural development project work and that improvement of nutritional status would be an objective and part of the design of appropriate health projects. Free-standing nutrition projects would be undertaken when they were the most appropriate mechanism to achieve a stated nutrition objective. One such new project a year was anticipated. Operational research again was stressed. The present paper discusses what the Bank has attempted thus far in nutrition and what has been achieved. The paper primarily is of an operational bent. Although a sizable body of analytic work has been undertaken during this period (and is briefly discussed in the latter sections, with greater detail in the appendices), the paper largely concentrates on the experience in the four nutrition projects. The main criteria used in assessing the projects are whether what was designed was delivered and what differences the efforts have made in terms of nutrition improvement and in the way that governments address their needs. Other important criteria are whether the actions are replicable and affordable. The subjective judgments concerning the performance of projects (as well as of contributions made by Bank-sponsored research and of the quality of sector work) are based on papers prepared by experts who have had no association with Bank work in the areas they address.3/ 2/ "Toward an Operational Work Program for Nutrition," PHN, September 22, 1980. 3/ Available background papers and other supporting documents are listed in Appendix A. -4- Section II of the paper describes the Bank-assisted nutrition projects, analyzing strengths and weaknesses of each of their components and their impact on target groups and on government food and nutrition policies and programs. Lessons for how the Bank's second generation projects can be made better are discussed in section III, with emphasis on project content and project design. The section discusses managerial requirements for nutrition projects, the pace at which projects can be carried out, organizational arrangements for directing nutrition programs, and what has been learned about the process of evaluation. The costs of interventions, their affordability, and, to the extent possible, their cost-effectiveness also are examined. The section concludes with a discussion of the Bank's management of nutrition operations. Nutrition components and nutrition orientation in Bank projects in other sectors are discussed in section IV and analytic activities (particularly what has been learned from the body of nutrition-related economic and sector work and nutrition-related Bank research) in section V. Section VI exa'mines the implications of the experience for future Bank work. It indicates the kinds of interventions that are technically feasible and those that have actually worked and not worked (of the latter, it suggests which have had some impact, which are promising, which bear watching, and which do not deserve further attention). This concluding section then raises issues that need to be addressed and proposes a course of action. II. The Bank-Supported Nutrition Projects The projects were designed to test different approaches to sometimes different nutrition dificiencies, which makes them difficult to compare. They do, however, share certain common features (see table 1). All included one or more institution-building components and several operational components, among them usually nutrition delivery through primary health care services, some form of supplementary feeding or food subsidy program, and a nutrition education component. The three projects initiated in 1976 and 1977 were heavily multisectoral, with agriculture, water supply and sanitation, and food marketing components sometimes added to more direct nutrition actions. The Tamil Nadu project, which can be viewed as the Bank's first 'second generation' project, was designed to concent-ate on fewer actions. The extent of health and malnutrition problems that the four projects addressed differs widely, but are widespread in all four countries, as reflected in Appendix B. Though the average daily food consumption appears to be quite reasonable in Indonesia, Colombia, and Brazil, a substantial proportion of households in all three countries -5- Table 1: Scope of Bank Nutrition Projects in Brazil, Indonesia, Colombia, and India Description Brazil Indonesia Colombia India Region Pilot areas in Seven neediest of Tamil Nadu Northeast several provinces 22 departments state Dates 1976-83 1977-83 1977-B3 1980-85 Project cost (U.S. dollars) Total 72,000,000 26,000,000 68,900,000 66,400,000 Bank loan 19,000,000 13,000,000 25,000,000 32,000,000a/ Conponents Building nutrition-related institutions (including training) x x x x Supplementary feeding x x ... x Food subsidies x ... x ... Health services with nutrition x x x x Nutrition education ... x x x Anemia control ... x Smal-scale food production (including family gardens) x x x ... Food technology (including reduction of food loss) and quality control x x x Water supply and sanitation ... ... x Food marketing x ... x ... a/ IDA Credit undoubtedly consumes far less than the required level of food.4/ Weaning-age children and pregnant and lactating women may suffer doubly because of the way food is distributed within the family. In Tamil Nadu, where the average per capita food supply is lower, it is likely that an even larger fraction of households suffers substantial consumption shortfalls. 4/ Income distribution is particularly skewed in Brazil, where in 1972 the bottom 20 percent of households had only 2 percent of household income, while the top 20 percent had 67 percent. In both. India and Indonesia in 1976, the bottom 20 percent had 7 percent of household income and the top 20 percent had 49 percent. (World Development Report, 1982; no comparable figures are available for Colombia.) -6- The infant mortality rate, which is strongly influenced by maternal health and nutrition status, is nearly twice as high in Tamil Nadu (and India overall) as in Colombia. The child death rate, which is considered to be a particularly good indicator of the prevalence of child malnutrition, is also high in all four countries, but four times as high in India as in Colombia. Brazil Nutrition Research and Development Project Before discussions on the Bank-assisted project began in Brazil the government had no nutrition policy and little in the way of a national program. A large number of mostly small nutrition activities had been under way for some years and Brazil was establishing a National Food and Nutrition Institute (INAN) as a focal point for nutrition work. The Brazil Nutrition Research and Development Project was aimed principally at developing techniques for countering malnutrition among pre-school-aged children and pregnant and lactating women in low-income families, providing the government with basic information and institutional and manpower capabilities, and developing and producing adequate quantities of certain high-priority inexpensive food products. A major objective of the highly experimental project was to acquire information, through special studies and the evaluation of field components, to enable the Brazilian government to design and conduct an effective national nutrition program. Most of the studies and all of the evaluations were to be undertaken by local research centers. INAN was responsible for coordinating monitoring, research, and evaluation. Throughout most of its first phase (1977-80) the project suffered from management, disbursement, reporting, and evaluation problems, and during 1980 there were doubts that it could reach completion. However, performance subsequently picked up and substitution of a major new component for one that never got under way in January 1981 helped set the project on a steadier course. The project included three field components designed to test new strategies and systems for delivering nutrition, using the existing infrastructure in the agriculture, health, and education sectors; three components aimed at building nutrition-related institutions and.developing an information base; and one component aimed at development and production of low-cost nutritious foods. Field Components The project's field components were all implemented, largely according to plan, and the number of beneficiaries considerably exceeded original targets. The main weaknesses -- in management and evaluation -- arose from INAN's inability to fulfill its coordination role. In spite of the project's difficulties, its operational components were sufficiently well received that all have been integrated (in the same form, or in forms that were modified based on evaluations) into large-scale government programs. -7- The largest field component was undertaken to test the effectiveness of combining the work of agricultural and social (i.e., health and nutrition) extension agents and of using a new special low-interest-rate credit scheme (CAP) to improve food production and consumption as an indirect means of improving nutritional status among 3,000 very low income farm households in the state of Sergipe. Some combination of the services reached about 7,000 households. Impact evaluation of this component, known by the Portuguese acronym PRAMENSE, has been a disappointment. In spite of an elaborate data collection and evaluation plan, no statistically significant evidence has been collected on the direct impact of PRAMENSE on agricultural output, food consumption, or nutritional status for a variety of reasons ranging from an overambitious research design to an inadequate baseline to the migration of control populations because of unprecedented drought. Operational experience with this component has been more encouraging. Agricultural extensionists have succeeded in increasing the use of on-farm storage, improved seeds, animal traction for ploughing, and, to a limited extent, chemical fertilizers. The CAP credit scheme, introduced for the first time through this project, has been the most successful means thus far of reaching the lowest income farmers and sharecroppers who normally have no access to credit and has been incorporated in Brazil's programs throughout the Northeast, including seven Bank projects. Local health associations formed under the project have - constructed thirty health posts and this effort too, along with other activities of the social extensionists, has become a standard part of the Bank's rural development projects in the Northeast. A study of the experience under the PRAMENSE program5/ was among the considerations that contributed to the Bank's understanding of the underlying problems of access to land. (A decision has since been made to include the purchase and redistribution of land in rural development projects in Brazil.) Both the CAP credit and the extension concepts have been institutionalized in the operations of the agricultural extension agency. Another field component, designed to test the effects on nutrition of different levels of and -approaches to consumer food subsidies, involved ten thousand households. Local supervision of both the operation (through large government-owned supermarkets) and the evaluation of this component, known locally as PINS, was excellent. A 1982 evaluation comparing children's growth by subsidy level (subsidies were 60, 45, and 30 percent of costs, plus a variation of the 45 percent subsidy that required a monthly medical examination to qualify) showed the percentage of malnourished children in each of the four subsidy groups was significantly lower (from 11 to 37 percent) than would be expected without the program. The finding of a reduction in malnutrition is consistent across age groups and by different measures of nutritional status. The evaluation did not 5/ Judith Tendler, "Evaluation of Nutrition Research and Development Project - PRAMENSE Component (Sergipe)," September 1979. -8- control for the effects of self-selection or for socioeconomic factors, considerations which will be taken into account in future analysis of the data. Based on what we know thus far, the reduction in malnutrition (which coincided with a reduction in buying power of the general population) "provides fairly strong evidence" that the intervention had a significant, positive effect on the nutritional status of the intended beneficiaries.6/ The PINS dropout rate was found to vary significantly with subsidy level -- 26 percent of those receiving the highest subsidy dropped out after 12 months, 75 percent of those receiving the lowest, 46 percent of those receiving the mid-level, and 63 percent in the mid-level group required to have medical examinations. Evaluations of the component provided valuable practical insight that led to modifications in the program. It was learned, for example, that those needing help most did not benefit as much as expected because they could not put together the cash required twice a month to take advantage of the subsidy. (The extra paperwork required of the supermarkets limited purchase of the subsidized foods to once a fortnight.) Also, for many, distances to the supermarkets were so great that they needed taxis to transport the groceries home. The studies revealed that it would be better to offer subsidies to all families in carefully targeted geographic areas with a very high percentage of low- income households than attempt to distribute coupons to families chosen by arbitrarily set income criteria. They also showed the program would have greater impact if the subsidized food was sold on a regular basis in more smaller neighborhood shops. (These and other lessons, discussed in Lessons from Project Operations, below, were incorporated in the design of the well managed revised program known as PROAB.) The PROAPE component, which provides supplementary feeding and preschool education to children four to six years old through the regular school system, met the project target in 1978 of 4,800 beneficiaries in the Northeast (plus a tightly controlled experiment, CEAPE, in Sao Paulo). By 1983 it had been expanded to 27 states and territories with 876,000 recipients and was being financed by the Ministry of Education (except in Pernambuco and Alagoas, where INAN continues to fund the project to conduct further evaluations). The program has been popular among both participants and implementers and has stimulated a high degree of community participation and contribution, thus simultaneously lowering costs and strengthening links between families and the local schools. The program is said to have opened the door for large-scale public preschool education in Brazil. Programs based on PROAPE have been set up in Colombia and Mexico, and several other Latin American countries are considering them. Evaluation of PROAPE has been fragmented and conclusions about its effect on the nutritional status of participating low-income children and their subsequent school participation rates and scholastic achievement must be drawn from a variety of sources. A careful evaluation of the CEAPE program in Sao Paulo showed that notwithstanding relatively low energy supplementation (250 to 300 calories compared with 500 calories in PROAPE) 6/ Peter T. Knight, "Brazil Nutrition Project: A Preliminary Evaluation and Lessons for Future Operational Work in Nutrition," May 17, 1982. -9- and far from ideal operating conditions, the level of malnutrition fell from 52 percent to 46 percent among participating children while it increased slightly from 44 percent to 47 percent among nonparticipants; also, school performance scores for the following two years were better and repetition rates were lower among children who had been in the program. The measured impact of PROAPE on physical growth was only marginal, but evaluations showed that 73.5 percent of former PROAPE participants passed in the first and second years of elementary school compared with 59.5 percent in a group that did not participate. And the academic performance of children with two years' exposure to PROAPE was consistently better than the nonparticipating group, ranging from 2 to 21 percent in three variations of the model. Psychometric tests showed statistically significant improvement in PROAPE participants over a six-month period, but it is not possible to know how much of it was due to maturation. Institution-Building The institution-building record of the Brazil project is mixed. Staff training, technical assistance, and support for evaluation of project components and other special studies were assigned to INAN. The main objective -- to enhance the capability of INAN -- was not achieved. Although INAN has survived as a political entity under cautious leadership, it does not enjoy high prestige within either the government or the nutrition community. The lack of government commitment to INAN.is reflected, in part, in the failure to raise it to the status of a' foundation, which in turn makes it more difficult for INAN to offer sufficiently competitive salaries or the institutional leadership and scientific prestige to attract highly qualified nutritionists and economists. Yet the development of a large-scale national nutrition program must be credited to INAN. In June 1983 it was given additional major program responsibilities. A second institution-building effort was the establishment of a nutrition status surveillance unit by IBGE, the Brazilian Institute of Geography and Statistics. The new unit was to undertake a national nutrition survey as a follow-up to a survey of 55,000 families by IBGE in 1974-75 that gathered extensive socioeconomic, demographic, food consumption, and anthropometric data. Political sensitivity to IBGE's earlier findings (for example, approximately 50 percent of urban families and 43 percent of rural families in the Northeast received less than 90 percent of average energy requirements) prevented the following nutrition survey's being started. However, a ten-person nutrition unit was established in (and funded by) IBGE to analyze data from the 1974-75 study.i 7/ An excellent compendium of the unit's work appears in Perfil Estatistico de Criancas e Maes no Brasil, Rio de Janeiro: IBGE, 1982. - 10 - Project funds were used to finance studies of the IBGE data by a special unit established in FINEP (a research management agency that reports to the Secretariat of Planning), to support a study of infant feeding practices in Sao Paulo, and to support 40 studies by university groups. As a result of this experience, a permanent policy-oriented nutrition research program jointly administered by FINEP, INAN, and the National Science Council has recently been established, with ten projects in 1983 programmed at a cost of approximately $370,000. The third institution-building objective was to establish a nutrition unit and a research program on the nutritional implications of agricultural policies in the planning section of the Secretariat of the Ministry of Agriculture (SUPLAN). The nutrition unit was disbanded early on, but many of the proposed studies of nutritional consequences of agricultural policies were subsequently contracted by INAN to the University of Sao Paulo's Institute for Economic Research (FIPE). The project's principal contribution to institution-building has been in focusing interest on nutrition-related activities at both national and local levels and in manpower development. The Bank's May 1982 evaluation concluded that as a result of the project, Brazil today has more trained people and better equipped institutions to deal with nutrition. "Top flight agricultural planners, economists and other social scientists, sparked by project funds, are devoting substantial portions of their careers to nutrition-related topics. And this is reflected in the programs of their institutions."8/ Production of Low-Cost Nutritious Foods and Food Marketing The component that was to have promoted production of low-cost processed and fortified foods of high nutritional value was deleted from the project, largely because of differences between the government and the Bank about subsidized interest rates on loans to the food industry. (This component had been added after appraisal in response to a Bank senior management decision.) The intention had been that loans would be made to small and medium-sized food companies9/ and that to reduce risk for the companies the government would guarantee to take part of the production for use in public institutional feeding programs. One effort, although it involved considerable delay and heavy costs in getting under way, has developed and successfully tested a method for processing cassava to flour that is sufficiently effective to have been replicated by 40 cooperatives or small firms at their own expense. An amended agreement signed on January 14, 1981, and extending the project closing date to June 1983, reallocated the approximately $7 million allocated for development, testing, and production of low-cost 8/ Peter T. Knight, op. cit. 9, Large companies were excluded, in part to assuage the concerns of certain Brazilian officials who saw the component as a ploy to benefit multinational food companies. - 11 - nutritious foods into the expansion of REDE SORAR, a program initiated by Brazil in 1978 to improve the marketing of fooa. Under this program, COBAL (the government's food marketing company, attached to the Ministry of Agriculture) sells certain basic foods to affiliated small and medium-sized private retailers in low-income urban and rural areas who agree to sell the products at specified prices, usually with a 10 percent margin for the retailers. The retailers are provided technical and operational assistance, including help with financing. In general, COBAL claims, the consumer pays 20 to 25 percent less than normal market prices. The competition that REDE SOMAR has created among food retailers in the Northeast, according to one study, has reduced prices of basic food items by an average of 10 percent. Economists have projected cost savings of 53 percent under this project and predict that even under highly adverse conditions its performance would be robust.10! Evaluations are under way of the project's impact on the availability of food to households, its distribution within families, and its effect on nutritional status. Funding under the project has allowed REDE SOMAR to expand to seven states in the Northeast. The program has also since been included in three other Bank projects. The government of Colombia, after studying REDE SOMAR, has asked for Bank assistance in setting up a similar program. Overview The large-scale extension of the operational components of the Brazil project is especially impressive in light of mistakes that can be seen in retrospect. The project design was too complex and ambitious, particularly given the technical and managerial capabilities of INAN. Communication with implementing agencies turned out to be especially ineffective. Weather, ecological, administrative, and socioeconomic factors introduced so much variability into the project that it was impossible to measure the impact of some components. Insufficient attention was paid to the potential problems associated with a weak federal government body executing a project and supervising programs that were essentially being implemented by state or local agencies. Also, there was little attempt to coordinate inputs, management, or evaluation of project components as had been intended. Even though three components were implemented in Recife only the project unit at INAN had knowledge of all three, and it lacked the capacity to compare and integrate the lessons learned from each component to form the basis for a Brazilian nutrition strategy. Eventually, the evaluation specified in the loan agreement had to be contracted out to the University of Sao Paulo's Institute for Economic Research. 10/ Pasquale Scandizzo, "Evaluation of the REDE SOMAR Distribution Program in Recife," November 11, 1980, and Emmerich Schebeck, "Targeted Consumer Food Subsidies Versus Targeted Food Marketing Interventions: Myth and Reality," September 1982. - 12 - In 1982, a report to the ACC Subcommittee on Nutrition of the United Nations on activities to combat malnutrition in Brazil since 1955, concluded that several external agencies were involved in assisting during this period in important ways and that "the World Bank has exerted an outstanding influence by promoting experimentation and innovation on program objectives and strategies through the INAN-World Bank loan agreement. This positive influence was partially handicapped by the resistance of the Brazilian Governmeht to pay adequate salaries for the formation of a qualified 'project unit' expected to develo into INAN's core group on nutrition analysis planning and evaluation." I/ During the project period the Brazilian government significantly increased its own commitment to nutrition. Before the project began, the government spent on nutrition $10 million to $20 million a year, other than for school feeding. From 1977 to 1982, it spent over $400 million on new nutrition programs (and an average of $143 million a year on the school feeding, and $420 million a year -- in revenues forgone -- on a workers feeding program). In 1982, it established FINSOCIAL, a billion dollar a year fund administered by BNDES (the National Economic and Social Development Bank) to provide additional financing for social programs, of which nutrition is one. For 1983, $150 million of the FINSOCIAL funds were budgeted for nutrition programs, about half of this directed to the consumer subsidy activity that had its origins in the project. The resources allocated to nutrition, other than that for the worker and school feeding programs, are still modest relative to the need. Viewed overall, it appears that efficiencies within the currently allocated resources could increase the impact substantially. Measured by the number of people reached (compared to the number projected), the new concepts that have been institutionalized, and the number and size of new programs that are now functioning in Brazil, the project produced noteworthy results. In the effort to better understand the nutritional impact of the interventions, there are encouraging but not yet conclusive results. The project has not succeeded in encouraging government to strengthen INAN as the focus for Brazil's nutrition efforts. On the other hand, the Bank's May 1982 evaluation concludes that "if and when Brazil chooses to launch a major attack on the malnutrition problem, it will be better equipped to move more quickly than it would have been without the project." Although the project did not surmount all of the "considerable risks" foreseen in the appraisal report, on balance it can be viewed as having achieved most of its objectives, albeit belatedly, and having made a useful contribution. After seven years of collaboration on nutrition, the government is interested in maintaining the relationship with the Bank in this field. A second nutrition project has been requested. 11/ Alberto Carvalho da Silva, "Efforts and Resources to Address Malnutrition in Brazil; Government Programs and the UN System," A Consultancy Report prepared for the ACC Sub-Committee on Nutrition of the United Nations, July 1982. - 13 - Indonesia Nutrition Development Project Slow beginnings also marked the Indonesia Nutrition Development Project, which began in March 1977. This in part reflected the Ministry of Health's unfamiliarity with Bank lending and the differences between government and Bank financial procedures. Also, the government's experience in nutrition before the project had largely been limited to an ineffective applied nutrition program. And there was a shortage of trained nutritionists and persons ver8ed in nutrition concepts. The objectives of the five-year project (later extended a year) were to develop measures to improve nutritional status; develop personnel and institutions capable of instituting and managing nutrition programs; and aid the government in the formulation and execution of a national program. The action programs set up under the project were directed at improving energy-protein nutrition of small children, ameliorating iron-deficiency anemia, and changing practices influencing nutritional status. About half of the project funds were dedicated to institution-building activities. Field Components The largest action program was a Nutrition Intervention Pilot Project (NIPP) designed to test new ways of delivering nutrition and health services to children under the age of three and pregnant and lactating women in 180 villages of seven kabupatens (the major subdivision of provinces in Indonesia). The NIPP was a field test of a range of community nutrition interventions (growth monitoring, oral rehydration, nutrition education with emphasis on breastfeeding and weaning foods, home and village gardens designed to increase production of nutritious fruits and vegetables, and small-scale food processing and food storage, along with immunization and, in selected areas, family planning with traditional birth attendants) with the intention that some combination might later be integrated into a national nutrition strategy. It has thus been the object of more evaluation efforts than any other component of the project. Interestingly, it is difficult to draw definitive conclusions about the impact of NIPP on levels of malnutrition in the target communities, in part because of these evaluations. Mid-project evaluations led to significant modifications in the program. Also, some NIPP activities were spontaneously adopted in neighboring villages that were to have served as controls. The initial objective of NIPP was to provide a processed nutritious product made from locally grown foods (BMC) to malnourished children and pregnant and lactating women. Village volunteers (kaders) would monitor the growth of preschool children, distribute food to those who were not growing adequately, and provide nutrition education to their mothers. The. program was to begin in two kabupatens in the first year, expand to two more in the second, and after a mid-project review, add three more by the fourth year. That schedule was followed, but with Bank agreement the program was extended to 43 percent more villages than the 258 planned. - 14 - Poor management at all levels and an unwieldy monitoring system considerably hampered early execution of the program. The mid-project evaluation recommended several substantial changes and by the summer of 1982, significant progress had been made in resolving many of the project's problems. A team of Ministry of Health and UNICEF consultants concluded that NIPP-type short-term rehabilitative feeding, targeted to children who fail to gain weight for three consecutive months and using prepackaged BMC, was preferable to the approaches used in other government nutrition programs. In early 1983, the NIPP approach was adopted for the two major government programs. The NIPP program has demonstrated the operational feasibility of a village-based nutrition rehabilitation effort, managed by the community kaders supervised by health center staff and using the locally produced processed food to rehabilitate seriously malnourished children. It has developed a monitoring and reporting system, portions of which are being introduced into the national program. And it has established the value of a field laboratory to test new operational ideas. - Two studies have been undertaken comparing the experience of children in the NIPP villages to those in the much larger Indonesian National Family Nutrition Improvement Program (UPGK). In one, NIPP children started lower in nutrition status than those in UPGK but showed marginally greater improvement in growth, reaching the same levels by the end of the study. The study shows the NIPP children had participated in the program to a greater extent (for example, attending an average of 23 weighing sessions in 25 months compared to 18 for the UPGK group) and that there was greater change in knowledge and behavior. Interestingly, the average education level of the NIPP mothers was lower than UPGK mothers.12/ The other study showed that the portion of children under three who were judged "well-nourished" (based on weight-for-height standards) rose from 39 to 44 percent in West Lombok (compared to an unchanging 43 percent in UPGK comparison villages) and from 36 to 46 percent in Bojonegoro (compared to an increase from 47 to 48 percent in UPGK villages.13/ As best as can be judged, nutritional impact varied from area to area. In Bojonegoro, the results reflected that local leadership has been more committed and active than in West Lombok. A second field component, a program for prevention and control of anemia, was designed to reach 10,000 workers on three government plantations and 10 small private plantations. The program was based on 12/ Satoto and Mashari Sudjono, "Nutrition Improvement Pilot Project Evaluation: A Cohort Study of Children in Selected Project and Comparison Villages in Indonesia," Tufts University School of Nutrition, SPSS Nutrition Analysis and Evaluation Fellowship Paper, 1983. 13/ Djumadias Abunain and Abas B. Jahari, "Study on Impact of NIPP in Two Kabupatens," Center for Research and Development in Nutrition (CRDN), Bogor, 1984. - 15 - research, undertaken jointly by the.World Bank and the government of Indonesia, that had established that levels of nutritional anemia among adult male workers in Indonesia were extremely high and that had demonstrated that substantial gains in productivity could be realized by treating this anemia.14/ Its objective was to test the logistical feasibility of establishing a large scale-delivery system for iron supplementation and the economic effects of iron supplementation on output and employment. The quick, successful development of the delivery system, plus preliminary evidence of an increase in hemoglobin levels and productivity among workers, led to a fourfold increase in the program's size by the end of 1981. In 1982, 300,000 workers were added at the expense of plantation owners. The 1982 evaluation gave cautious endorsement to the program, noting its extension as evidence of the government's growing interest in nutritional anemia. However, a more careful evaluation of the impact on health and productivity was urged before consideration of further expansion. Such an evaluation is currently under way. The nutrition education component, aimed at molding nutritional behavior, initially was to be implemented in 60 villages in five kabupatens, one of them a NIPP area. The first two years of the program were a preparatory phase, devoted to infrastructure-building, selective training of village volunteers (kaders), administrative preparation, and setting up a growth-monitoring program. The next year was used to test concepts-- households, each including a pregnant or lactating woman or a mother of a malnourished child, met individually with interviewers to agree on a set of dietary modifications that the family then tried out. This trial step helped the nutrition education team dietermine what precisely the program should aim to do -- for example, introduce a homemade weaning food -- and the best media and methods for doing it -- for example, posters for the kaders that mothers were asked to mark every time they performed a prescribed action, or radio spots in dialogue format, using the mothers' own words. One year after the communication strategy went into operation, an evaluation of program and comparison areas showed that the target kaders had learned the program messages and had more specific advice to offer than kader in other nutrition education programs in comparison areas; they were devoting on average nearly 14 hours per month to nutrition work compared to less than 7 hours by workers in in comparison areas. In project villages, 67 percent of households had been visited by a nutrition kader, in 14/ Darwin Karyadi and Samir Basta, "Nutrition and Health of Indonesia Construction Workers," World Bank Staff Working Paper No. 152, 1973, and Samir Basta and Anthony Churchill, "Iron Deficiency and the Productivity of Adult Males in Indonesia," World Bank Staff Working Paper No. 175, 1974. - 16 - comparison villages 44 percent; project village mothers averaged 47 percent correct recall on nutrition messages, comparison mothers 28 percent.15/ After four years of project operation and a year of intense education activity, an evaluation of households confirmed trends of improvement in the nutritional status of mothers and children in the project areas. At all ages mean weights for program children were higher than for comparison children, and at 24 months of age there was a highly significant difference of half a standard deviation between the mean weights. There were half as many moderately and severely malnourished children (those less than 75 percent of median weight-for-age) in the program group. These differences in nutritional status were seen not only across the entire sample but also in each geographic region. Multiple regression analysis indicated that the difference in nutritional status could be attributed to the program rather than to schooling level or other factors. Women applied the knowledge thus gained to feeding their children more of the recommended foods. In program areas 87 percent of the children consumed more than half of the recommended calorie intake and 82 percent of the children consumed more than half of the recommended protein intake; for children in the comparison sample, the figures were 62 and 60 percent, respectively.16! The effectiveness of specific messages was demonstrated, the strongest correlation being between knowing and preparing the weaning food that is important to the nutritional status of five-to-eight-month old infants at a crucial time in their development. The project appears to have overcome one of the largest constraints to improved nutrition, the level of the mothers' education. Children's nutritional status was determined less by the level of maternal education for children in the target group than for those in the comparison sample. This was the most successful of the field programs, the success attributable to thorough research, a carefully conceptualized media strategy, good implementation of the initial phase, and evaluation from the outset. The program had a demonstrably positive effect on the nutritional status of the target population. It provided only nutrition education, utilizing a combination of personal and mass-media contacts. Technical assistance was vital to the development of this component. The government has announced its intention of using the messages and media strategy in its national program. 15/ Marcia Griffiths, Thomas M. Cooke, Richard K. Manoff, and Marian Zeitlin. Nutrition Communication and Behavior Change Component, Indonesia Nutrition Development Program, vol. 5, New York and Washington: Manoff International Inc., 1983. 16/ Ibid., vol. 4. - 17 - The home and village garden component was implemented in 521 villages, including the 258 NIPP villages, and reached at its peak approximately 12,000(of 18,000 expected) households. Extension services, seeds, fertilizer and other production oriented inputs and services were offered for an initial 3 years on a grant basis to establish the necessary demonstration effect. The project also envisaged the setting up of one seed model garden in each village through community efforts on communal land. A 1981 evaluation found that participating families spent more time working in their gardens and produced more and grew a greater variety of vegetables than other garden-owning families in their villages.17/ The majority of participants consumed about 60% of their vegetable production and attempted to market the rest. The main disappointment was that the program could not help the neediest families; the land available to them was just too small to make viable gardens. Also, only a limited number of seed gardens could be established because of lack of suitable community land. The Ministry of Agriculture, which administered the program, is now actively promoting gardens as a regular part of its extension program. No attempt has been made to look at the effect of the garden program on the nutritional status of families participating in NIPP. Institution-Building The major institution-building components of the project were strengthening and expanding staff and facilities of the Center for Research and Development in Nutrition (CRDN) and constructing, equipping, and staffing a new Food Technology Development Center (FTDC). The project also included programs for expanding and improving the training of nutritionists at the Academy of Nutrition and the Nutrition Assistant Training Center and bolstering, through technical assistance, the Food and Nutrition Unit of the Ministry of Agriculture. Strengthening the nutrition research center was a slow process, requiring major technical assistance from foreign consultants; the center is now considered the best of Indonesia's five medical research institutes and the work of some of its staff meets international standards. But CRDN is not yet consistently capable of evaluating nutrition work effectively enough to be useful for the planning agency; that is one of the disappointments of the project. Also disappointing is CRDN's failure to establish an economic and social studies unit on the scale envisioned (due, in part, to its difficulty in attracting economists), thus limiting its role as a bridge to government planning groups, such as BAPPENAS, the central planning agency. In short, the CRDN has made great gains in five years and now has facilities, equipment, and most other prerequisites to meet international standards but has yet to regularly achieve that level. 17/ Pt. Ika Harimurthi, "An Evaluation Study of Home and Village Garden Programme in Nutrition Intervention Pilot Project Areas," Report to Ministry of Health, 3 vols., Jakarta, July 1982, and Emmerich M. Schebeck, "Food Production and Consumption Linkages for the Rural Poor," PHN paper, September 1982. - 18 - A 1982 evaluation team was impressed with the quality of leadership of the Food Technology Development Center, perhaps the first major facility in the world devoted largely to village-level food technology, and by its well-trained and highly motivated staff. But delays in meeting its physical targets had slowed developing technologies, particularly for food storage. Although not without problems, the FTDC has succeeded in establishing machinery in the villages to produce the processed weaning food that is used in the NIPP component. (More recently, FTDC has attracted funding from several other donors; its major project, financed by the Dutch government, deals with food safety.) The number of graduates of the Nutrition Academy more than tripled between 1977 and 1981. The academy is now regarded as one of the outstanding nutrition institutions in Asia and is attracting students from neighboring countries. Only a small fraction of the technical assistance envisioned for the Food and Nutrition Unit in the Ministry of Agriculture was called upon. Plans for development of a strong analytic unit have not been realized. Any serious effort to improve nutritional status must make provision for a continuing research capability and skilled staff. Indonesia now has both nutrition research and training under way, though it may require more time until the research centers can make their proper contribution. Most of the training objectives have been met; the main manpower gap at this stage is at the field management and supervision level. (Plans are under way to expand the Nutrition Academy.to meet this need by building four regional facilities. More Nutrition Assistant Training Centers also are to be established.) Overview . Another major objective of the project was to aid the government in the formulation and execution of a national food and nutrition program, and many background papers have been sponsored by the project toward that end. In February 1983, at a Food and Nutrition Planning Workshop funded by the project, the coordinating minister for economy, finance, and industry gave a strong endorsement for assigning nutrition high priority and expanding nutrition activities. This was reiterated six months later in a meeting that discussed recent work on the nutrition section for the next Repelita, the national five-year plan. The dramatically increased sophistication in dealing with nutrition issues since the last Repelita holds much promise for the current efforts. With the strengthening of the Nutrition Academy and its training activities, the government has a stronger manpower base for expanding operations. Most components of the Indonesia project have been mutually reinforcing. The nutrition research center has taken a role in the evaluation of both NIPP and the anemia prevention and control programs, and it formulated the food used in NIPP's supplementary feeding program. The village-level technology for production of the food was designed and produced by-the food technology center. In addition, the gardening and storage programs were implemented in the same villages as NIPP. Unfortunately, there was essentially no overlapping of the anemia or education components with other field programs. It would have been particularly interesting, as initially planned, to compare the effects of the NIPP and nutrition education programs and of the two combined. - 19 - In almost all of the project components there was a sizable amount of technical assistance that strengthened the programs. A large majority of the assistance came from Third World countries, particularly India and the Philippines, whose experience is relevant to Indonesia's needs. The 1982 evaluation team of international and Indonesian nutrition experts credited the project with "strengthening and expanding the infrastructure for a larger scale nutrition program in Indonesia" and noted the "quite impressive" impact of its action programs.18/ The report linked the greatly increased emphasis on nutrition in the national plan -- which eventually extended nutrition activities to approximately 30,000 villages -- to the project. Other international agencies, particularly UNICEF, have played a very important role in raising the consciousness of the government concerning nutrition as well as in funding nutrition activities. The Ministry of'Health and BAPPENAS, and to a lesser extent the Ministry of Agriculture, have become more conscious of the importance of nutrition and the possibilities for effective interventions. Interest in and sophistication concerning evaluation (clearly one of the project's weak points in the early stages) has not only increased but spread to nutrition activities beyond the project. The evaluation shortcomings, the institutional strengthening still needed, and other (particularly management and coordination) problems should not-detract from the fact that the government has established a strong base in a remarkably short period. Nutrition now plays so large a role in Indonesia (including as an entry point for family planning and women's development activities) that it frequently is referred to as a movement, with few parallels elsewhere. The Bank project is credited in Indonesia with having contributed, directly and indirectly, to this. The government has asked the Bank for a loan for a second nutrition project. Colombia Integrated Nutrition Improvement Project In the mid-1970s the government of Colombia developed a broad, carefully planned strategy to fight malnutrition. It included a nutrition program (PAN) and an integrated rural development program (DRI), both developed by the National Planning Department (DNP) and both assisted by the World Bank. The objectives of the nutrition project, the subject of discussion here, were "to improve nutrition, living standards and the productive capacity of 1.8 million Colombians," to monitor and evaluate the project to ensure its effectiveness and financial feasibility, and to strengthen the program's managerial base. Programs sponsored by several agencies and integrated at both the community and the national level were 18/ Selo Soemardjan, et. al., "Evaluation of the Indonesia Nutrition Development Project - Volume I, Report of the Evaluation Team to the Ministry of Health," Jakarta, December 1982. - 20 - to be targeted geographically to the poorest 30 percent of the populat-ion in 7 of Colombia's 22 departments and in low-income neighborhoods of the federal district of Bogota. Prospects for the project were good. It was being executed through the strong National Planning Department and the nutrition program was a major plank of the presidential platform. Field Components One action program was designed to strengthen the primary health care system. Its community health posts would emphasize preventive rather. than curative care, and locally recruited health workers (promotoras) and auxiliary nurses would carry out nutritional surveys and deliver integrated nutrition and health services. By mid-1983, after a slow start, 544 health posts had been constructed or renovated (the orginal target was 281, which was revised in 1980 to 400 when three additional departments were added to the project area) and 3,480 health workers had been trained (the target was 2,028). The DNP estimated the program reached 3 million people, a Bank supervision mission in November 1982 estimated 1.5 million to be a more realistic number, the latter being slightly short of the target. This component of the project was a catalyst for the primary health care effort in Colombia and is widely reported to have strengthened it markedly. Distribution of food coupons was central to the strategy developed for the nutrition project areas. The coupons, targeted for pre-school-aged children and pregnant and lactating.women in low-income households, were used at food stores in partial payment for a selection of processed foods of high nutritional value. At the peak of the program in 1981, coupons were being used by an estimated 125,000 beneficiaries. While that was only 68 percent of the project target population, Bank consultants regarded it as a respectable achievement for a government embarking on an unusual and complicated program that involved the private food sector working with government agencies. Most beneficiaries participated irregularly, partly because of the uneven issue of coupons and partly because they lacked cash to buy their full share of subsidized foods. Also, price inflation had caused the real value of the subsidy to drop by about one-third. Opposition of a new political administration to the concept of consumer food subsidies ended the program in 1983 before full insights on impact could be obtained. Evaluation results showed that participation in the PAN program -- controlling for family income, the value of food stamps received and other factors -- induced a substantial increase in family food expenditures. Among PAN participants, however, the value of food stamps as a proportion of total income did not affect food consumption, suggesting that the effect of PAN participation on consumption was probably principally through improved nutrition knowledge and awareness. Regression analysis shows that visits to a health center can improve food consumption patterns, raising intakes on the order of 300 calories per person per day. How the increased consumption is distributed through the family and how it affects growth of target individuals are not clear.19/- Another program, designed to increase access to potable water and improve environmental sanitation, was aimed at reducing the prevalence of 19/ Per Pinstrup-Andersen, Evaluation Report for PAN, 1983. - 21 - diarrhea -- a major contributing cause of energy-protein malnutrition, particularly among pre-school-aged children -- and other infectious diseases. By mid-1983,'362 water supply systems had been completed and 119 were under construction (the original target was 360, the 1980 zarget 411); 29,690 latrines had been installed (the original target was 112,000, the revised target 30,000). After controlling for income and other variables, a very clear impact of improved sanitation (brought about by installation of latrines and toilets) has been established both on the occurrence of diarrhea and directly on weights and heights of children. Evaluation data indicate that with every increase of 10 percent in the number of households with latrines, the number of children with nutrition problems would be reduced by 15 percent. (The impact of pit latrines was about half that of septic tanks and flushing toilets.) Households that receive piped drinking water from the project have slightly higher heights and weights than other households, but it is not clear that the differences are statistically significant. (This may be a reflection of the quality of the piped water.)19/ Communities, as envisaged, have borne (in the form of labor) about half the cost of the construction of latrines and about 17 percent of the installation costs of the water supply systems. The National Institute of Health, which is responsible for installing the systems, has been slow in disbursing funds, partly because of a shortage of local counterpart funds. Regional health services and other agencies have had mixed success in operating and maintaining water systems. One part of the nutrition education component -- a mass media program which was the responsibility of the National Planning Department -- got off to a strong start in mid-1978 with a diarrhea information campaign. A second campaign in 1979-80 to promote breast feeding was particularly well planned and implemented. Although it has not: yet been evaluated, the consensus is that it has brought about a considerably more positive attitude toward breast feeding among target families and health center physicians. In general, the project was effective in setting up a permanent working group in each province to develop local nutrition education materials. Production of mass media materials has been impressive and well above targets. However, many of the mass media messages focused on familiarizing audiences with PAN and providing them information about nutrition problems; with the exception of the diarrhea and breast feeding messages, they were not designed to change nutritional behavior. A second part of the education component, administered by the Colombian Family Welfare Institute, provided training for individuals and group discussion leaders for nutrition education programs. The agency reported that 35,400 agricultural extension workers, health workers, nurse auxiliaries, home economists, teachers, shopkeepers, and others had beentrained. Except for the breast-feeding campaign, coordination of the training and the mass media effort and relations between the agencies responsible for them have been poor. Another component, designed to increase vegetable consumption, was to develop a technical package of seed, credit, and fertilizer suitable for home gardening and train extension agents and rural teachers to demonstrate its use. What was planned as a modest program for 20,000 families expanded into a small-scale rural development effort. It proved to be very popular and was gradually broadened to include livestock and fish farming and model sanitary homes. The program, known as PANCOGER, eventually reached over 35,000 families and absorbed more than one-fifth of - 22 - the total PAN budget. A 1982 evaluation concluded that credit was the best-managed activity (89 percent of all recipients kept up on their payments), and that technical assistance and nutrition education activities needed strengthening. The impact on nutritional status has not been evaluated. Because this component had gone so far beyond the original objectives, because its nutritional focus was diluted in the process, and because the subsidized interest rate of 8 percent conflicted with the new government's policy, the program was suspended at the end of 1982. Institution-Building At the national level, the project supported work by the Institute of Technological Research to develop a food quality control system, stimulate investment in commercial processing of nutritious foods, develop small-scale portable food processing equipment, and improve methods of on-farm storage. By mid-1983 the institute had completed 20 of 26 proposed food-technology studies and 3 more were under way. The same institute, collaborating with the National Institute of Health, established 9 regional quality control laboratories and a national quality-control system that is functioning reasonably smoothly. A major institution-building goal was development of a strong nutrition program coordinating group in the National Planning Department, which would have the capacity to organize and evaluate nutrition projects and create a permanent focus for the consideration of nutrition concerns that transcend the responsibilities of individual line ministries. Although the managerial base for planning and executing the nutrition program has been strengthened as a result of the project and the group played a useful role in stimulating progress in regional planning and administration, this goal has not been fully realized. Part of the problem was that the group in DNP lacked the managerial expertise to ensure that the project components would be effectively implemented. The biggest obstacle, however, was the erosion of political support for DNP to play this role. Nor was the group very effective in its role as coordinator. The project was fairly concentrated geographically, so that its components were more likely to reinforce each other than those in, for example, the Indonesian project. Distribution of food coupons by the health centers proved to be an efficient way of reaching target groups and of attracting mothers and children to the centers. But construction of water supply systems and latrines unfortunately was not well coordinated with other components. Overview DNP's. failure to evaluate adequately the impact of the project on nutritional status is especially disappointing given the experimental nature of the project. Nonetheless, many interviews- and field observations led the November 1982 supervision mission to conclude that "there exists a good deal of information indicating that the project has had a salutary effect on health and nutritional status, and that the project represents an imaginative, worthwhile and relatively successful approach to the complex problem of malnutrition." Of the nutrition project, the Bank resident representative in Colombia reports "the performance was actually pretty good and better than large numbers of projects in Colombia." Clearly the - 23 - project has made a significant contribution in strengthening the primary health care system, which relies principally on locally recruited promotoras and emphasizes better nutrition and other preventive rather than curative care. The government had a well-articulated nutrition policy before the project began. The commitment to nutrition interventions seems to have declined or at least the approach to nutrition problems has now changed, largely because of the different social philosophy of the new political administration that took office in August 1982; it is dropping the food coupon and small-scale food production components. The government, which had agreed to invest $39 million f or nutrition and related work over the four-year project period, had by 1983 spent $88 million. The increase, of itself, is of course not a virtue, but it does reflect the seriousness the government attached to nutrition in those years. In July 1983, the government requested a two-year extension of the project. Although most project targets have been met (and in several cases substantially surpassed) and government expenditures have been well in excess of original projections, an undisbursed balance of about $10 million in Bank funds remains (largely due to the inability in the early years to process withdrawal applications). The new administration hopes that project funds can support its plans to introduce a f.ood distribution and marketing scheme patterned on REDE SOMAR in Brazil. Strengthening, through the project, of the primary health care system led directly to the prospects of another Bank-assisted project (this heavily health-centered but including nutrition), currently under preparation. Tamil Nadu Nutrition Project In 1977 the government of India asked for IDA assistance to improve nutrition conditions in Tamil Nadu and possibly Orissa. Tamil Nadu was far enough ahead of other states in its planning to lead to the decision to concentrate efforts there. A pioneering nutrition systems study, started nearly a decade before with assistance from USAID, had established a strong data base and an understanding about the many local determinants of malnutrition. Also, the state had a long-standing commitment to improving nutrition and to development generally. The original Tamil Nadu request was for a broadly multisectoral project addressing selective aspects of food production, processing, and storage as well as the delivery of nutrition and health services, including supplementary feeding and nutrition education. However, lessons from the three earlier Bank-assisted nutrition projects led to a decision by IDA management to simplify the project and focus on improving the outreach, quality, and efficiency of the state's nutrition and related health delivery systems. - 24 - The project design also reflects lessons from Bank and other donor experience regarding targeting and delivery mechanisms. The implementation period is six years, longer than the earlier projects, with the last year devoted to a major impact evaluation. At full implementation, the project will cover around 9,000 villages in the rural areas of 6 of the state's 14 districts, making it one of the largest nutrition projects in the developing world. In addition to extensive coverage, there are several innovative features: concentration on children aged 6 to 36 months and their mothers as the exclusive target group; development of a sensitive but practical surveillance system to identify at-risk children, with weight-growth velocity as the principal indicator; short-term supplementary feeding to help restore children to acceptable patterns of weight gain; the inclusion of rigorous management information and evaluation systems; and a comprehensive communications program incorporating both mass media and interpersonal techniques. Field Components The project operates through a linked system of village nutrition centers, each staffed by a locally recruited worker, reinforced by health outreach and referral services. The program focuses on surveillance, diarrhea management, deworming, immunization, and vitamin, mineral, and food supplementation. An Empowered Committee-of senior civil servants, chaired by the chief secretary of the state government, is responsible for project oversight. Day-to-day coordination, monitoring, and overall budgeting takes place through an autonomous Project Coordination Office which works closely with implementing agencies, mainly the Directorates of Social Welfare and Health and the state government's Department of Evaluation and Applied Research. The design of the nutrition delivery system draws heavily on experience with the training and visit system for agriculture extension. Workers are thoroughly trained to perform only a few key tasks, supervisory ratios are realistic, and supervision is supportive with an emphasis on in-service training to improve performance. Two operational evaluations are built into the project as formal points for correcting the project's course if necessary. One took place after the first operational year, while project operations were confined to one pilot block of the 170 administrative blocks which are to be covered by the project. The second review is currently under way, now that the project is at midpoint, and covers 78 blocks or 46 percent of the project total. Initial results from both the pilot block and broader implementation have been extremely encouraging. The program is well-organized and well-administered. Workers know what their jobs are and understand their priorities. They receive frequent supervision, as had been planned, and they have adequate supplies for their work. The major questions were whether the high quality of training, supervision, and administration could be sustained as the program expanded, and whether the program would work in a state with less administrative capacity than Tamil Nadu. - 25 - As of April 1983, trained community nutrition workers were installed in more than 4,000 community nutrition centers. The former secretary of the Indian Planning Commission reported that the recruitment and training of the workers and supervisors for the program had been "quite thorough and will serve as models" to other government programs. "Each of the cadres," he wrote, "knew and performed their tasks well."20/ For every ten workers there is one supervisor, which ensures visits twice a month to each center, allows careful monitoring of record keeping, and helps to maintain a high level of motivation. A recent Bank study analyzing the potential use of the training and visit system in primary health care ranks the Tamil Nadu effort considerably above the other programs studied, concluding "in terms of having the prerequisites for success, the Tamil Nadu Nutrition Project is extraordinarily impressive."21/ Participation rates at the monthly weighing sessions used to monitor growth and identify children who need supplements have been impressive, averaging around 85 percent in all project areas and 90 percent where the program has been under way for over a year. Community enthusiasm is reflected also .in the formation of Women's Working Groups in most project villages and their interest in obtaining the supplement as a regular food for their weaning-age children. The use of a wheat- rather than rice-based supplement, prepared as a traditional snack food (laddu) and provided for 90 days with the possibility of extension for another 90 days, reduced the tendency of families to.view the supplement as a substitute for food given at-home.22/ The effects appear to have been dramatic. The portion of children requiring feeding is 43 percent of that when the program began. In the face of a year of bad drought, severe and moderate cases of malnutrition have dropped from a baseline of 19 percent to 12 percent; in control areas during the same period, cases have gone up from 16 percent to 30 percent. Relapse rates have been under 25 percent. Estimates based on evaluations for the early years of experience indicate that as a result of the completed project (or over a 21-month period if based on a statewide level with three times the project population) there will have been 107,000 20/ Asok Mitra, "Tamil Nadu as a Nutrition Delivery Laboratory," November 12, 1982. 21/ Richard Heaver, "Adapting T&V to Primary Health Care," PAS, May 1983. 22/ In Tamil culture, a food that does not contain rice is nol: considered to be a meal. - 26 - fewer severely malnourished children and 76,000 fewer moderately malnourished. Through improved nutrition status, over 12,000 deaths will have been averted.* The project also included activities to reduce the prevalence of vitamin A deficiencies; at first a shortage of supplies delayed this program. Early supply shortages also interfered with the provision of iron and folate to pregnant and lactating women, an activity that has been turned over to multipurpose health workers. However, the government has established a new drug procurement system (based on recommendations arising from a drug supply study under the project), which should speed up supplies of mineral and vitamin supplements as well as drugs and vaccines. The rural health services component, calling for construction, furnishing, and equipping of 1,600 health subcenters, training of 2,500 female multipurpose health workers and 800 health visitors, and construction of training facilities and training of instructors has been reasonably successful. Virtually all of the scheduled 361 subcenters in Madurai district have been completed and handed over to the Health Department. A full complement of multipurpose health workers and supervisors was in place by October 1982; statistics on ante-natal registrations, maternal tetanus inoculations, and assisted deliveries indicate increased health outreach. .During the first year of implementation in the pilot block, coordination was -weak between nutrition workers and supervisors, who come under the Social Welfare Department, and health workers and health visitors, who come under the Health Department. The government took prompt action which resulted in steady improvement in relations and coordination now appears to be smooth and reasonably effective. The major continuing problem is the health referral system for children who do not adequately --gain weight while receiving the food supplement.23/ Though the Tamil Nadu health system is good compared to others in India, the poor performance of the health system has adversely affected the project. The slowest of the components to get under way was the nutrition and health communications program. Delays in finding staff and in ironing out differences of approach toward nutrition education between the government of Tamil Nadu and IDA staff initially held up the mass media campaigns; only one, using radio broadcasts and printed material to teach * Between the time this Nutrition Review was written and the OVP meeting that requested its circulation to Bank staff, a mid-project evaluation was completed that suggests the project impact will be considerably greater than is reflected above and in the later discussion on costs and cost-effectiveness (pp. 43-52). It now appears that there is at least 40% less severe malnutrition among 1-to-3 year olds in project villages than in control villages. By the age of 5, children who have been through the program are a significant 1.75 kilograms heavier than children from control villages. 23/ Community nutrition workers usually refer these children to multipurpose health workers who routinely suggest a physician's examination at the primary health center. Parents appear reluctant to take their children to the primary center, principally because of its distance from the village. The government has agreed to a sample survey of such children to determine whether there is a consistent pattern of health or socioeconomic problems. - 27 - better management of diarrhea, was undertaken during the first two years of the project. By the third year, the program was working well. The project support communications (development of teaching materials for training workers) undertaken as part of this component has functioned well throughout. Pre-project workshops for nutrition delivery staff were particularly useful. Institution-Building A great deal of attention was paid to the design of monitoring and evaluation systems for the project and to their refinement. These have done well in comparison with earlier Bank nutrition projects and most service delivery projects in general. Community nutrition workers have kept regular and accurate records (with regular and frequent supervisory visits). Concern that the amount of data might prove unmanageable has led to surveys in 45 different nutrition centers each quarter to determine what information is necessary to ensure reliable monitoring. Though there were delays and inadequacies in analysis of data comparing the pilot and experimental blocks, the right questions were addressed and data were reworked in accord with detailed suggestions by supervision missions. The high level of commitment on the part of nutrition workers and the high visibility of the project has contributed to the high participation rates among the target children. The apparent success of the project in reducing the prevalence of severe malnutrition is-fundamentally due to the targeting of services to weaning-age children; it could not have been accomplished without regular weighing to identify malnourished children, regular supplies of a well-accepted supplementary food, and effective education of families to increase the home feeding of the malnourished child. The project has demonstrated that, even when they come under different government agencies, health and nutrition workers can cooperate and their work can be mutually reinforcing; this may be a more successful formula than overburdening one community level health worker with many different responsibilities. The project reflects the high level of government commitment to alleviating malnutrition in Tamil Nadu. The secretary of social welfare credits the project with having led to a recognition of the value of nutrition surveillance, acceptance of limiting the number of tasks assigned to community workers, and an understanding of the importance of good supervision. The national child development program has adopted the project's approach for training.of Tamil Nadu workers along functional lines. Although the project is well regarded by state officials, there is a danger that the state government's attention, and possibly resources, may be diverted by the chief minister's highly publicized and politically controversial $300 million a year mid-day meal program that began in July 1982 for children aged two to fifteen years. Since children in the nutrition project participate in the mid-day meal program, the evaluation results of the nutrition project may suffer. And if the two programs should be merged and there is a shift toward dependence on regular feeding by the state, family responsibility for child feeding would be diminished. Meanwhile, at the national level, the Tamil Nadu project is attracting favorable attention, and there have been discussions with the central government about the possibility of IDA funding for similar projects in other states. Also Tamil Nadu, itself, has requested a second project. - 28 - III. LESSONS FROM PROJECT OPERATIONS The four largely experimental projects were designed, in part, to draw lessons for future nutrition-related activities. Experience to date is sufficient to provide some general directions, at least within the settings in which the experiences took place. Project Content Food Distribution, Supplementation, and Consumer Food Subsidies The food subsidy programs in Brazil and Colombia and the feeding programs in Brazil, Indonesia, and India used a variety of approaches to identify the at-risk populations and to target nutrition services. The emphasis on targeting is an important break from the past, where mass coverage was the norm. The main lesson is that some forms of targeting are feasible and in some cases do indeed lower costs. The food subsidy programs in Brazil and Colombia promoted the idea of selective participation. The pilot project in Recife for distributing food through several government-run supermarkets based participation in a coupon program on income criteria; it demonstrated the difficulty of targeting by income in a setting where income reporting can be very arbitrary. The program showed that food coupon programs are more effective at reducing levels of child malnutrition if the subsidies are high enough to sustain participation. It also showed that heavy bookkeeping and related administrative cost is required to conduct an effective food coupon program, that downpayments for coupons pose a barrier to the lowest income group, and that the system must adapt to the frequent small purchases that low-income families are forced to make. Building on lessons from the evaluations, the Brazilian program was modified, with apparent success, to reach very low income neighborhoods without requiring coupons or downpayments. Commonly consumed basic foods now are subsidized for all customers of many registered small neighborhood stores in selected poverty areas. Any leakage of benefits to those not in need (judged by Bank staff to be very small in Recife) is considerably less expensive than administering the cumbersome coupon program. The revised system has also made possible frequent purchases of small quantities. (The system presumably would not be as effective in areas where the poor are less concentrated geographically than in Recife.) In Colombia, poverty areas of the country were identified as part of the national development plan, and targeting of its food subsidies was narrowed to households in those areas where there was a child under five or a pregnant or nursing woman. Costs were lowered by reducing the number of beneficiaries, and nutrition education furthered by focusing attention on those with special nutrition needs and by encouraging consumption of certain nutritious weaning foods. Very little leakage or fraudulent use was apparent. The coupons seem to have been better accepted in urban than in rural areas (75 percent of eligible urban households collected their coupons and 90 percent of the coupons were exchanged compared with 30 percent and less than 50 percent in rural areas). Targeting in direct feeding programs was best achieved through the weight monitoring program set up in Tamil Nadu to screen children for - 29 - admission to a feeding program when their growth faltered and release them once weight increased satisfactorily. Village workers motivated by good training, good supervision, and good project support materials kept participation rates at 85 percent in the monitoring program, even when the majority of the children received no food supplement. The food cost was significantly below that of most feeding programs that aim at reaching pre-schoal-aged children. Mothers apparently do not use the supplements as an excuse to reduce their children's intake at home, apparently because of a sense of shame if the child does not show a weight gain (the same sort of discomfort that, along with inconvenience, keeps affluent nonresidents from buying at subsidized shops in poverty neighborhoods in other projects). On a related controversial issue in the nutrition field -- processed vs. nonprocessed foods -- experience from the projects confirms that processed weaning food mixes can be an important element in the success of supplementary feeding programs and can promote nutritionally beneficial changes of behavior. Their development, however, must be based, on careful study of food practices, local availability of ingredients, and purchasing patterns of the target groups. Beyond weaning foods, special processed products (such as the enriched pastas provided in the Colombia project) are only likely to have a nutritional impact where there is clear evidence that malnutrition is mainly the result of imbalanced nutrient intakes rather than simply a low intake of everything, which is more common. The attraction of food as an incentive to encourage participation in other development activities was seen both in Brazil and Colombia. When food supplements were temporarily unavailable in some PROAPE schools in Pernambuco, participation in the preschool stimulation program declined 30 percent. In Colombia the availability of coupons at health centers led to a marked increase in the utilization and effectiveness of the primary health care system, discussed below. In one of the most important findings from the Bank's experience, the REDE SOMAR program in Brazil showed that it was possible to reduce food prices for low-income families by increasing efficiencies in the food marketing system. Moreover, lower REDE SOMAR prices stimulate through price competition reductions in food prices at nonaffiliated stores. Bulk price purchasing by the government food marketing agency to meet the needs of this financially self-sustaining program also lowers costs of food obtained at the same time for other social programs that are not self-sustaining. Linking Nutrition Delivery with Family Planning and Primary Health Care Nutrition's contribution to reduced fertility has generally been seen via the route of better-nutrition lowering infant and child deaths which, in turn, lessens the demand for more children. In the Bank's experience, strong complementarities have also been seen between the delivery of nutrition services and the delivery of family planning services. In Indonesia-the family planning agency credits monitoring and related nutrition actions with having provided an important inducement for villagers to get together to discuss and become involved in family planning activities. In the design of a follow-on project now under prep'aration for the Bank, nutrition at the village level is closely linked to family 30 - planning services. Because of their village base and close contact with mothers, community nutrition workers from the Bank-assisted project in - Tamil Nadu have been cited by family planning staff there as the best sources for successfully identifying family planning acceptors. In Thailand nutrition as designed in a Bank population project has been used as the opening wedge to promote (as part of the national village rural poverty effort) an expanded primary health care system, which is itself an important element in the family planning program. Perhaps the greatest potential for linking health and nutrition delivery has been demonstrated in India and Indonesia. The weight-monitoring and selective feeding approach for screening and servicing nutritional needs used in these two projects requires the establishment of a network of village-based workers. In Tamil Nadu the full-time community nutrition worker provides simple health services (e.g., deworming, diarrhea control, provision of vitamin A and iron and folic acid) and at the weighing sessions and through regular home visits promotes use of health centers. Children identified as malnourished are at risk also from infection or other health problems. In turn, the health system provides supporting services to reinforce nutrition interventions. Simply monitoring weight gain turns out to be more cost-effective than the commonly used weight-for-age as a measure of nutritional status.24/ This "growth velocity" approach also has high educational value, increasing parents' awareness of the nutritional needs of very young children and providing an easily understood signal that a faltering child needs special attention. The Tamil Nadu experience has demonstrated the feasibility of instituting a village-level nutrition surveillance and rehabilitation program that has a positive effect on nutrition status at an affordable cost. In Indonesia nutrition interventions at the village level are regarded by government as the entry point for delivery of a broader range of health services. A study of the Indonesian nutrition project demonstrated that such village-level services were likely to offer a more efficient and more equitable use of resources than comparable services offered higher up in the health system, such as a sub-district health center.25/ A comparison was made of services (growth monitoring, nutrition education and oral rehydration) provided through the village-level nutrition education component of the project with services provided by well baby clinics at sub-district health centers. The cost per child served in the village-level program was 60 percent less and a substantially higher proportion of low-income people used the village services.26/ 24/ The probability analysis employed by the Bank in these determinations -- from knowing the best time of day to weigh a child for getting the most accurate measurement to the probability of bypassing a malnourished child -- constitutes an important advance in the sophistication of planning direct feeding programs. 25/ Peter A. Berman, Equity and Cost in the Organization of Primary Health Care in Java, Indonesia, Unpublished PhD dissertation, Cornell University, 1984. 26/ 73 percent of all low-income children participated in the village program compared with under 9 percent participation in the comparable health center program. - 31 - In Colombia, food coupons substantially increased attendance at health centers, bringing needy families into the orbit of the public health system, and the centers proved they could manage a food coupon system with few difficulties. (Clearly it was more manageable than direct food distribution, common in health centers elsewhere.) Integrating a food subsidy into the primary health care service helped people recognize that good nutrition, particularly of the target groups of young children and pregnant and nursing women, was essential to good health. Food coupons, in a sense, served as health center "prescriptions" to prevent and cure a "disease" known as malnutrition. In Brazil, requiring a health check-up for eligibility in some locations in the subsidy program caused participation to drop off so steeply that the requirement was dropped. Health and nutrition activities seem to be compatible services whether delivered by a community worker or separate health and nutrition workers, so long as tasks are well defined, cooperation encouraged, and the supply of vaccines, drugs, nutrients, and food supplements adequate and regular. Nutrition interventions intended to benefit the poor will be more effective to the extent that they recognize the opportunity costs of participation and limit *the number of changes they require in the normal habits of the benefiting population. Linking Nutrition and Education Nutrition education appears to be most effective when designed to modify specific behavior. Working with intended audiences, allowing them to try different alternatives and to formulate new ones, gave both relevance and specificity to the Indonesian project strategy. (One of the keys to success was the work of a nutrition anthropologist who lived in Javanese villages during most of the 14-month period of program formulation. She was able to determine, for example, that the reason most children were underweight even though breastfeeding was nearly universal was that most village women nursed with only one breast.) The program's objectives were based on what people could and would do, they addressed a few priorities, and they were transmitted easily and effectively by village workers in home visits and growth monitoring sessions and these efforts were reinforced by radio. The project was highly successful because it was built on the use of resources that already existed in the community. That principle would seem to apply regardless of a program's context. ' The Tamil Nadu field-level program emphasizes village-level monitoring of growth as a form of nutrition education, with short-term supplementary feeding as a corollary rather than dominant measure. It demonstrates that such programs are feasible, acceptable to the community, and apparently effective in terms of impact. The Brazilian experience indicates that nutrition education for mothers, along with providing nutrition and intellectual stimulation in a school setting for preschool children, contributes to the improvement of nutrition and subsequent school performance. Small-Scale Food Production The garden programs in Colombia and Indonesia were popular and successful in reaching the rural poor, and in the credit component of the - 32 - Colombia project the default rate has been extremely low. The Indonesian experience demonstrated, however, that since such a program only benefits those who have land to devote to gardens, there is no benefit for those most in need of improved nutrition. That program also showed the importance of community seed gardens. The number of seed gardens in the project was well below that envisioned but from even the limited experience it was seen that the concept is worth replicating. Although the Indonesia evaluation showed that families participating in the home garden programs grew more vegetables than comparison families, the nutrition impact of those gardens and other small scale food production programs is, unfortunately, not known. (The closest information comes from a nutrition component in a Mauritius Rural Development Project where 97 percent of the 2,300 families involved in a garden program claimed it had increased their food consumption and that their incomes, from sale of garden surplus, had gone up 5 to 10 percent.) The evaluation of the Brazil project highlighted the importance of animal production in the survival strategies of small farmers and even sharecroppers, though it was intended only to promote crop production. It also showed that in areas subject to frequent drought, crop insurance was needed as a stimulus for modern inputs to be purchased on credit. The project's experience also highlighted the limitations of a rural development strategy for Brazil's Northeast that does not provide access to land for the rural landless. Water and Sanitation The Bank's experience with water and sanitation in nutrition projects is largely limited to the Colombia project but there is evidence from that project that improved sanitation can lead to improved nutrition, as measured, first, in reduction of diarrhea and then in improved heights and weights. The experience showed that septic tanks and flushing toilets were approximately twice as effective as pit latrines. The project has done little to clarify uncertainties concerning the relationship of piped water and health and nutrition status. Participation of the Community All four projects mobilized, with some degree of success, both resources and volunteers in the community -- leading to construction of health posts in Brazil; construction of water and sanitation facilities in Colombia; mothers' working groups in Tamil Nadu; and nutrition education through village nutrition workers in Indonesia -- which went considerably beyond the standard concept of community participation through meetings to discuss nutrition education. Local participation helped defray costs and improve the projects' quality. In the PROAPE preschool feeding component in Brazil, mothers regularly assisted teachers, thus contributing to the system's development while learning about nutrition themselves. Fathers and other community members donated time and materials for construction of feeding and play centers, reflecting their enthusiasm for the program. In Indonesia, communities that were not part of the nutrition program were willing to organize themselves and raise the money necessary to establish similar programs in their own villages. - 33 - Generally, participation by members of the community generated enthusiasm in the project, raised nutritional awareness, and improved the chances that project activities would continue after the project ended. On the other hand, part-time volunteers with limited skills take more time than regular workers to get a job done, there are more of them to be trained, tasks must be fewer and simpler, and quality control and supervision are more difficult. And in areas where not enough volunteers are available, services will suffer. There are cost savings in a volunteer program but there also are costs to both the project and the participants (see Costs and Cost-Effectiveness of Nutrition Projects, below) that should be weighed when a volunteer system is being considered. The same applies to the desirable goal of community involvement in determining the content of a project. Delays this caused in preparation of a nutrition project requested by the Senegalese government contributed to the project's eventually being dropped.27/ (This raises the broader issue of how to deal with the tension between the Bank's role as a mover of money and its role as an alleviator of poverty.) Opportunities for what commonly is called community participation are culture-specific. The concept assumes that autonomous, rural communities exist in some democratic form and that they can be mobilized to manage their own collective welfare in an egalitarian manner. These assumptions are not valid for many traditional societies, and striving to incorporate community participation where it is alien to the culture may be an elusive goal. Where extensive community participation is deemed feasible, experience from the Bank's nutrition projects suggests that the main ingredients for success are appropriate training and adequate supervision. Project Design Project Complexity The initial three nutrition projects, as originally planned, were too complex, trying to test numerous approaches and optimistically assuming a high degree of management and organizational skills. They were administratively cumbersome, cutting across organizational lines in governments and in the Bank. Communications and coordination among agencies were poor, particularly at the headquarters level; much less so at the field level. Part of the complexity was created by "productivity" components, added in response to discussions in the Loan Committee and at the Board. In stressing the need for directly productive components (such as food gardens in Colombia and Indonesia and food industry activities in Brazil), a key point about the value of nutrition projects was missed, i.e., better nutrition makes all other sectors more productive. 27/ Sufficient time was built into the Indonesian nutrition education program to involve the community effectively in determining content, but this followed project approval and initiation. - 34- A less ambitious, more narrowly focused project consisting of no more than three or four well-integrated nutrition interventions and carefully designed to limit needs for managerial skills is more likely to be successfully implemented. Although nutrition projects should be considerably more limited in scope than the first three projects were, experience in Tamil Nadu demonstrates that they need not be monosectoral. The Colombia project also demonstrated the benefit of an intersectoral effort -- in this case the substantially increased use of primary health care facilities. Project Length The early projects, planned to last four or five years, were unrealistically short to demonstrate the expected changes in nutritional status. Although there are exceptions (Tamil Nadu, with its advanced planning and management skills, being the best example), many first projects that depend on newly established institutions should have lesser aims or may require six to eight years to show a nutritional impact for large population groups. Also shorter "preparation" projects may sometimes be appropriate to help those countries lacking a sufficient base in nutrition (of which there are many) to mount larger, operationally oriented projects. The Project Package A blueprint can not be written for a standard nutrition project. In a socially coherent country like Indonesia, different mechanisms can be used than in a pluralistic society like Tamil Nadu, whefe it is difficult to get villagers together. Differences in nutritional needs, causes of malnutrition, sophistication levels, trained manpower, and political system dictate differences in approach, type of delivery system, and ratio of servicing staff to beneficiaries. It is possible, however, to suggest the kinds of things that analysis should search for and the method of the search, and the kinds of packages for addressing different types and causes of malnutrition under different circumstances (see Applying the Experience to Future Bank Actions, below). Project Refinement Because political interest in subjects such as nutrition can fluctuate, projects may not emerge if their planning period is as drawn out as a new field of activity would seem to require. In Indonesia, the government planning agency, BAPPENAS, decided to push ahead without all the answers because it believed its program was moving in the right direction and that (in the words of the assistant director) "seizing the opportunity while there is interest and resources" would pay the greatest dividends ten years later. Such circumstances call for starting out on a pilot scale and expanding later, as best exemplified by the Tamil Nadu project. Also, projects should be flexible enough to take advantage of unanticipated opportunities that emerge during implementation. Experience in Implementation Since nutrition does not fit easily into any existing ministry, should a technical ministry like health or agriculture, a supraministerial - 35 - body like a planning ministry, or a specially assigned interministerial commission be given responsibility to integrate nutrition into government machinery? And should nutrition projects be managed by existing agencies or by new project units? The Bank has participated in operations with a variety of approaches that offer some general guidance. Institutional Framework In Colombia the National Planning Department, which had control of investment budgets, made some progress in getting old-line agencies to work together (and prompted the health ministry to move toward better primary care, even though at times there was resistance from parts of the bureaucracy). Focusing diverse ministries' efforts requires strong direction and coordination by a bureaucratically separate entity not committed to parochial interests. It is essential that ministries participate in some kind of governing body for the entity, but the entity must be more than a sum of the participants. Experience in Brazil showed the difficulties of asking a health ministry agency (INAN) to plan and coordinate actions of other line agencies; as in most countries, health staff in Brazil often are not chosen for their administrative effectiveness. Despite the wide representation on the governing board of the Brazilian nutrition coordinating agency, it sometimes had difficulty getting from the member ministries the responses contemplated by the project. Some of the same problems were seen in the Indonesia project. A special unit established solely to administer a project, as in Tamil Nadu, met with more success. Management problems were simpler than those of coordinating agencies. Yet the absence of a permanent coordinating agency, which might cause problems when the project units no longer exist, made the government take care to integrate project activities into existing institutions. During the period these projects were being implemented, several other countries established multiministerial coordinating mechanisms; generally they were not successful. The need for an administrative focal point to run projects is clear. A lead ministry (or under some circumstances a supraministry) appears to be the best approach,, its exact placement depending on the nature of the project and governmental practices and preferences. A project that is innovative and involves several executing agencies requires a coordinating body, at least initially, located in an agency that is respected and has power.28!' More important for project effectiveness than organizational form is the support and involvement of politically powerful decisionmakers. The importance of governmental commitment, the fragility of it, the difficulties of assessing it, were special problems in this new area of Bank lending. The Tamil Nadu project unit brought ministries of social welfare and health together effectively through an empowered committee, chaired by the state's chief secretary. This high-level support increased the ministries' willingness to cooperate. 28/ The organizational problem is less difficult, of course, if the project's components fit neatly in a single ministry. But multisectoral projects may, in some situations, be the most helpful means of addressing a nutrition problem. - 36 - Another vital ingredient is the interest and support of ministries expected to participate in a project. Without building good relations, project management staff will remain outsiders, confronted with the defenses and inertia of bureaucracies not committed to cooperation. Capacity to build commitment must become as important as technical qualification in hiring project staff. How integrated should services be? The Tamil Nadu project risked the conflicts and complications of setting up a separate system of community nutrition workers in order to avoid overburdening health care workers. The project demonstrated that where cross-departmental cooperation was carefully nurtured, parallel and complementary systems can work. If tasks are well defined and cooperation is stressed during training and by supervisory staff, health and nutrition workers can work together effectively even when they come under the authority of different line agencies. Management Requirements Good management can make the most of an unfortunate organizational choice, and the best organizational arrangements make little difference if management is not competent. Delays in implementation, the need for reprogramming, and the heavy supervision required in certain of the projects reflect a widespread lack of management capability in agencies likely to implement nutrition projects, and indeed in many development projects (51 percent of all Bank projects have moderate or major managerial problems). Development of managerial capability required time that project circumstance generally did not permit. Moreover, the need for developing management skills was not given the attention it deserved in project training plans. The complexity of the nutrition projects, except in Tamil Nadu, generally strained management capabilities of implementing agencies. (Bank staff were overly optimistic in their expectations.) Had the Brazilian and Colombian groups been free to concentrate on fewer tasks, they might have performed better, and more careful attention to regular communication with participating ministries might have opened opportunities for early resolution of differences that interfered with implementation and coordination of the projects. Brazil's INAN, for example, did not circulate the Bank's appraisal report among participating agencies and neglected to communicate at all with some of them during the project's first twelve months. The Colombian staff was overextended, with the project going into too many areas too quickly. Although far more investment was carried out as a result than would have been if the project had expanded more slowly, more systematic expansion could have led to better services and perhaps more demonstrable effect. In contrast, Tamil Nadu managers were able to give more effective direction; the gradual evolution of routine systems helped with coordination among agencies and the systematic acquisition of necessary skills. In a number of situations, failure to pay attention initially to essential administrative underpinnings at all operating levels precluded - 37 - effective management. Preparing budget proposals, writing contract specifications, and monitoring performance of nutrition projects through a management information system, for example, require special knowledge and skills. And Bank requirements compound management difficulties. The Project Launch workshops introduced by the Bank's Urban Projects Department is a useful orientation concept for first nutrition projects. The value of pilot activities was seen in Tamil Nadu and Indonesia where they helped to expose implementation problems, so that the subsequent larger projects were more sensitive to bureaucratic realities, political constraints, and local needs. An underlying strength of pilot projects was that they were undertaken by agencies responsible for the larger operations and their procedures were replicable on a large scale. In the places where nutrition programs have been successfully launched, strong entrepreneurial skills have invariably been evident. The ability to serve as catalyst and promote the involvement of several agencies and disciplines in a program is more important than sophisticated scientific or technical capacity. People who have demonstrated abilities to initiate and effectively manage large programs (and, preferably, who have a grounding in the general nutrition field) are better candidates for leadership than those of scientific prominence whose work has been largely restricted to research. The central group must of course establish its professional credibility. But there are different ways of doing this. .Colombia and Brazil relied on sophisticated technical proficiency, failing to recognize that field staffs needed simpler, more practical assistance, of the kind provided in Indonesia and Tamil Nadu. Quantitative Indicators of Implementation As the Bank now uses a variety of quantitative indicators to rank the progress of projects, experience in nutrition can be crudely compared with that of other Bank sectors (see figure 1).29/ The trend in all but one (Colombia) has been regular improvement. The average performance rating of the four nutrition projects, on a scale of I (for problem-free or minor problems) to 3 (for major problems), was 1.3. Bank average is 1.8. The 37 field reports for the nutrition projects reflect early problems in overall performance in three of the projects -- one, Brazil, was regarded as a problem project -- but show improvement over time. These findings confirm the impression that although start-up problems may be considerable, once a team becomes familiar with the tasks of implementation, projects progress without exceptional problems. Most of the quantitative targets, from construction of facilities to staffing and training to delivery of services, have eventually been met. Project management also has been improving over time. "Compliance with loan conditions" was generally good after the first year (rated 1.4, compared to a Bankwide average of 1.7). Delays of audit 29/ Performance ratings represent the subjective assessments in. reports of project officers on return from supervision missions. They do not take into account how long the project has been in operation. The data used here, for fiscal years 1981-82, generally reflect earlier years of operation for the nutrition projects than for other projects from longer established sectors. - 38 - Figure 1: Performance Ratings of Nutrition Projects Compared with Ratings of First Three Projects in Other Sectors and Bankwide Ratings 2 7575- 2 so- 2 2- 2 2- 2 0062 t 75- 257- Completion Management 3 M 2 7- 2 75 2920 22 - 25 22 2 2 00- 2MD 1 75- 1 25 costs Procurement 2 75 2 so- 2 2S- B__ ualkwdeveon projects Rankings: 1.00 - Problem free or minor problems Disbursements 2.00 - Moderate problems 3.& U - Major problems - 39 - reports in two projects were the most frequently mentioned difficulty. Procurement performance (rated 1.3) showed problems in the early years but became problem-free in the third year once procedures had been sorted out. Project reporting was uneven (averaging 1.7) but also became generally stronger as the projects progressed.. Performance of consultants was rated uniformly high. Consultants from Third World countries played an important role in providing technical assistance for the projects. Evaluation of their performance confirmed that it is productive to apply the experience of consultants fron countries whose conditions and work problems are close to those of project countries. In Indonesia, four of every five of the consultants to the food technology institute, nutrition academy, and project secretariat were from developing countries, as were a third of those who aided the nutrition research institute. The disbursement performance of nutrition projects is short of projections but nonetheless is now about twice as good as other projects in the PHN portfolio (an average 30 percent disbursement shortfall from the appraisal target, compared with 58 percent) and ranks well with first projects in other sectors. An examination of the first five years' disbursement experience of the first three nutrition, population, urban and rural development projects finds that throughout the period nutrition ranks below urban projects and above rural development and population projects. By year five, 75.4 percent of nutrition projects had been disbursed compared with 88.6 percent for urban, 51.3 percent .for rural development and 33.4 percent for population. As in first projects in most other sectors, disbursement was poor in the beginning but ended strong. For example, Colombia's early difficulties in processing paperwork limited its reimbursements from the Bank to $15.2 million of the $28.1 million for which it would have been eligible. Procedural problems also severely hampered the beginnings of the Indonesia project. On average, it took 15 months for the projects to work out of the major-problem category and an additional 16 months to reach problem-free status. The actual costs of the completed projects compared to estimates before loans were agreed, including two still under implementation, were rated 1.1, considerably better than the Bank norm. There have been no reported cost overruns among the nutrition projects, whereas over 40 percent of projects Bankwide had cost increases of 25 percent or more, and 26 percent exceeded estimates by 50 percent or more. The time required for completion of the nutrition projects shows an average slippage of 19.5 months from original estimates (table 2). Bankwide, more than two-thirds of the projects require more than this amount of additional time. In 1982, 54 percent of projects required over 50 percent more time than was estimated, and 26 percent took at least twice as long as originally estimated. - 40 - Table 2 Time Required to Complete Bank Nutrition Projects Project Completion date Project effectiveness Estimated 1984 date at appraisal estimate Slippage Brazil 12-31-76 12-31-80 6-30-83 30 months Indonesia 4-01-77 3-31-81 3-31-83 24 months Colombia 3-09-78 6-30-82 6-30-84 24 months Tamil Nadu 8-05-80 6-30-85 6-30-85 none Of the four nutrition projects, the Colombian project encountered the greatest difficulties, particularly in disbursements. Management performance and reporting were problem areas for all; the best areas were cost estimates, loan compliance, and procurement. Compared to projects in other sectors and to the overall Bank average, the nutrition projects held their own or better in the areas of cost estimates, loan compliance, and vroiect comoletion and were wpak.r in 0ha :ara , , management, and reporting. Overall, the implementation of nutrition projects has not-been substantially different from that in other sectors. The reports of project officers acknowledge problems, particularly start-up delays, but the trends are encouraging, showing marked improvement well before project completion. The areas in need of improvement fit within overall Bank patterns, the most important being the need for greater realism in scheduling project activities and in establishing project goals. Institutional and management issues will require continued major attention. Monitoring and Evaluation As in other sectors, monitoring and evaluation components of the nutrition projects proved difficult. Generally, the value of assessing performance as a management tool was not appreciated. Neither the recording and analysis of routine data for monitoring operations nor special studies to evaluate impact were viewed as integral parts of the project's implementation. Some managers saw evaluation as politically threatening (leading in one instance to a "dress rehearsal" for beneficiaries of questions that might be expected to be asked by evaluators), others as esoteric research, and others as a minor aspect of project funding. It was often seen as something to be done for the Bank rather than as a useful tool of management. -41 - Even under ideal conditions, project evaluation is a technically difficult activity, and particularly so in social sectors where behavioral factors are critical measures.30/ Early experience in the nutrition projects (and during the same period in rural development projects) showed that only the simplest evaluation designs should be used for large-scale projects. The evaluation study for the rural component (PRAMENSE) of the Brazil nutrition project, for example, would have been formidable to experts in a well-controlled situation; complicated by an unanticipated drought that, in turn, led to migration of a large part of the project sample, the evaluation exercise produced few useful results. Both Bank and Brazilian staff, meantime, had neglected opportunities to collect routine information that would have been useful in the operations of the extension agency.31 The Bank's supervision reports, especially for Brazil. and Colombia, frequently included exhortations for more attention to evaluation, suggestions for additional data and analysis, and technical guidance for responding. Less ambitious aspirations might have produced better outcomes. In their zeal to satisfy the more, sophisticated data needs of the impact studies, central monitoring units failed to build the capacity in field units to collect simple, operating data routinely and accurately or the capacity in managerial offices to interpret and respond to the data. Once their biases were perceived by operational staff, the evaluatots' effectiveness was .diminished. Much of the monitoring and evaluation work in the nutrition projects is either behind schedule or has fallen.below acceptable standards. There are examples, however, of work that is of good quality (Indonesia's evaluation of the nutrition education component and its revised system for monitoring NIPP, monitoring and evaluation in Tamil Nadu, Colombia's evaluation of the primary health care program, and Brazil's evaluation of PINS that resulted in the program's modification). Efforts to institutionalize the art and practice of evaluation generally have been much slower than expected and, in some cases, have fared poorly. Helping countries expand their evaluation horizons and skills required more time than the nutrition projects permitted. Sensitivity to the need for monitoring and evaluation did increase and substantial 30/ A recent review of the health literature concluded that "for all practical purposes it has been virtually impossible to measure any impact (on mortality) of the existing formal medical system as an independent variable." W. Henry Mosley, "Will Primary Health Care Reduce Infant and Child Mortality? A Critique of Some Current Strategies, with Special Reference to Africa and Asia," paper for IUSSP Seminar on Social Policy, Health Policy and Mortality Prospects, Paris, February 28, 1983. 31/ Also, proper evaluation of projects depends on using some segment of the population for control purposes; but their exclusion fErom program benefits raises political and moral problems that governments (e.g., Brazil in the PROAPE component and Indonesia in NIPP) are not always willing to accept. - 42 - progress was reported. And the mistakes and problems of each earlier project served as guides in planning for monitoring and evaluation of subsequent projects. Hence the contrast between the complex, research-oriented design of the Brazil evaluation component and the simpler, more effective design of the Tamil Nadu component. Lack of commitment to and understanding of the evaluation process among local managers, absence of analytical skills, technical difficulties, neglect of monitoring, and lack of understanding of the related roles of monitoring and evaluation are common problems -- for instance, in rural development projects in one region "information has seldom been used effectively".32/ In other sectors the impact on well being of project beneficiaries is seldom evaluated;33/ process indicators commonly are used as proxies to measure impact. The hope of assessing the changes in physical growth resulting from nutrition projects exceeded the evaluation goals of other sectors by a considerable margin. In future nutrition projects it is important to keep expectations at a reasonable level, as in Tamil Nadu. The Bank should insist on a sound and workable monitoring and evaluation structure and approach at the outset, but also be sensitive to the limitations and the need for flexibility to make adjustments. Data, collected routinely from the outset of a project, must focus on what is needed for analyses. Collecting data on everything that appears relevant is costly and may neglect what is most needed. Intermediate outcomes should not .be neglected since they are of importance to operating staff and thus to their appreciation of the value of sound monitoring and evaluation. Future Bank-assisted nutrition projects should concentrate initially on process evaluation for determining whether systems can deliver on an operational scale the services that earlier demonstrated an impact in well-controlled, small-scale scientific studies. In the longer term, efforts should aim at evaluating the impact of projects on nutritional status of the targeted population. This is the same pattern now espoused for health projects. Bank staff, consultants, or outside research institutions may be needed to show how the work should be done and to train local staff and strengthen local institutions. The Bank may have to fund this work separately. 32/ Guido Deboeck and Ronald Ng, Monitoring Rural Development in East Asia, World Bank Staff Working Paper No. 439, October 1980. 33/ Robert Ayers, Banking on the Poor: The World Bank and World Poverty, MIT Press, 1983. -43 - Costs and Cost-Effectiveness of Nutrition Projects Costs The annual financial cost of Bank project components on which aggregated data are available ranged from $49,000 for the Indian nutrition delivery component at the block level to over $4 million for the Colombian national food subsidy program. Many obvious factors contribute to these large differences -- the size of the project, the extent of targeting, the size of the food transfer, and so on. Costs also vary by country, because of differences in wages, food prices, and food consumption patterns. To get insights about the importance of these factors it is useful to look at cost from a number of perspectives, as reflected in table 3. In addition to comparing total costs, efforts have been made to compare the cost per beneficiary, the cost for servicing a population group of 100,000 people, and the cost to deliver 1,000 calories. Attempts also have been made to compare the costs of project interventions with similar interventions outside the projects and to assess their affordability if the components were to be expanded nationwide. Interpretation is complicated because of different types of projects with different aims and different outcomes. Even then, because of country differences, costs would not necessarily be replicable from one project to another. Another caveat: although much of the data is quite good for projects of this type, its quality is uneven. If projects are compared in terms of annual cost per beneficiary, the least expensive by far is the Indonesian nutrition education program at $2, the pain reasons being that no food transfers are involved and much of the field work is done by volunteers. At the other end of the scale are Brazil's preschool feeding and stimulation program (PROAPE) and the Colombian and Brazilian consumer food subsidy programs. The differences between the two subsidy programs, at $35 and $21 respectively, reflect in part the larger number of health and nutrition services built into the Colombian costs. A major drawback of the cost per beneficiary measure is that it does not reflect the volume and quality of services provided and improvements brought about by the program. In India, as the yearly overhead cost per child receiving project services declined over time from $12 to $8, the cost per child fed increased from $21 to $38 as fewer children required the rehabilitative feeding. Also, the per beneficiary costs here, as with the other measures used, include only the direct public sector costs of the program. They do not capture hidden costs incurred for transportation and time spent by the beneficiary in using the service, the purchase price of subsidized foods, and additional food that an education program convinces people they should buy. Efforts also were made to compare the costs of servicing a standard-sized population, using 100,000 people as a common denominator and assuming constant returns to scale. (These costs pertain not to 100,000 beneficiaries but to the total population from which beneficiaries are selected.) Thus food subsidies initiated in Brazil and Colombia cost $399,000 and $438,000, respectively, per 100,000 persons, while the weighing and feeding programs in Indonesia and India cost $98,000 and $75,000. The differences primarily reflect the size of the target group as a percentage of the total population, the effectiveness of the coverage of the target group defined, the length of time people were fed, the amount of food and other services per beneficiary, and country cost factors. Table 3. Financial Costs of Sate Qqxmients in tink-Supported Nutrition Projectsa/ Anual canponent cost Total project Anasal asqxnnt cost per 100,0) (thousands of dollars) Annual cost per cost to deliver population (thisakis of dollars) Non- Recurrent Population in Number of beneficiaries 1,0(0 calories Non- Recurrent Project. component, ad date recurrentb/ Food Other Total area served beneficiaries (dollars) (dollars) recurrent Food Other Total Brazil Food Subsidy (PINS), 1980 0.04 722.93 151.59 874.56 219,(XX)c/ 41,026 21.32 0.30d/ 0.02 330.11 69.22 399.34 Preschool feeding and education (PiGAPE), 1980 6.86 435.45 594.06 1,036.37 Not available 22,298 46.48 0.53e/ . . ... ... Colanbla fy t/ Food subsidy, 1981 ...f/ 2,652.84 1,551.38 4,204.22 960,0(X) 120,(X) 35.04 0.799/ ...f/ 276.34 161.60/ 437.93f/ Indonesia Weiglng and feading program (NIPP, Bojonegoro area), 1982 387.73 76.71 74.22 189.66 194,11) ... ... ... 19.96 39.54 38.26 97.76 - weighing, ad screening ... ... ... ... ... 14,886 12.74 ... ... - feeding ... ... ... ... ... 2,307 56.01 ... ... Nutrition education, 1977-81 Initiation phaset/ 83.38 0.00 63.26 146.64 225,(O) 37,272 3.94 Not applicable 37.06 0.00 28.12 65.18 Expansion Phase47 36.18 0.00 40.30 76.48 225,(X) 37,272 2.05 Not applicable 16.06 0.00 17.91 33.99 Tamil Nadu Weighing and feeding program (nutrition delivery services, block level, Madural district), 1982 3.68 43.09 28.29 75.06 X100,(XX)k/ ... ... ... 3.68 43.09 28.29 75.06 - weigdng and screening ... ... ... ... ... 6,084 12.34 ... ... ... ... ... - feeding ... ... ... ... ... 3,615 20.76 ... ... ... ... ... 1985 projected, at 1982 prices 3.68 17.24 28.29 49.21 10000k/ .... 0.721/ 3.68 17.24 28.29 49.21 - weigidng ... .. ... ... ... 6,084 8.09 ... ... ... ... ... - feeding ... ... ... ... .1,01 37.82 ... ... ... ... a/ For detailed component costs, see T,J. 110, "EconivC Issues: Costs Affordability and Cost Effectiveness,- PlIN, 1983. b/ Anuailized over the respective project component periods which vary from 3 to 5 years. See inlividual cwst tables for details. c/ Estimate derived by tripling the 10,000 families of 2 ninnig salaries or less targeted for the subsidy program since around 1/3 of families in the project area fall in that category. then iltiplying by 7.3, tl average family size in the lowest income group in urban Northeast Brazil. d/ Based on 600 calories per person per day at 50 percent subsidy for 27,000 recipients for the full year (i.e., sane of the 41,026 beneficiaries participated for less than full year). e/ Basai on 500 calories per person per day at 100 percent subsidy for 14,055 children for 217 days, 4,270 for 131 days, aid 3,973 for 60 days. f/ No information is available on rtnAmecrrent costs for 1981 or previous year. Other recu'rent costs lly iloKde SIW training costs, ilicli should fall under inorecurrent costs. y/ Based on 5,227 tons of food, the value of oaupons redeemd. 1/ Includes fixed costs. j/ Excludes fixed costs. k/ Average block size is 100,(MX). - 45- In Brazil, entire families, selected by income level, were fed; in Colombia, children under five and pregnant and lactating women; in Indonesia and India, children under five and under three respectively, who were screened before being fed and released when their nutritional status improved. Since the latter two programs were much more selective, food costs were kept to a minimum (annual food costs in India began at $43,000 and are estimated to go down to $17,000 as there become fewer malnourished children who require feeding; in Brazil, they were $330,000). And because fewer people were reached and less food was involved in the narrowly targeted programs, the administrative, delivery, and other recurrent costs were kept lower, even though screening imposes additional costs. However, recurrent costs are a higher proportion of costs in highly targeted (e.g., Tamil Nadu) direct feeding than in more generalized (e.g., Brazil) food subsidy programs. The need to weigh children monthly requires more contact with beneficiaries, but it also provides the opportunity to include in the program nutrition education and primary health care for the children and to spread costs commensurately. Recurrent costs also are a higher proportion of the Colombian subsidy program because of the requirement (and related benefits) of collecting food coupons at the health facilities. Nutrition education as implemented in Indonesia is the least expensive of the components to service a population of 100,000, listed at $65,000 annually during initiation and $34,000 during expansion. Again, this low level of costs is expected since no food transfers are involved. The difference between initiation and expansion phase-expenditures is large, mainly because development of the message is an iexpensive initial cost that need not be repeated during expansion. The cost of transferring 1,000 calories also is estimated in table 3.36/ As would be expected, costs rise by the degree of targeting -- from $0.30 to transfer 1,000 calories for the Brazil subsidy program to $0.53 for Brazil's more targeted preschool program to $0.72 for Tamil Nadu's feeding program, the most tightly targeted of the efforts. The Colombian subsidy program at $0.79 is the highest. Part of this, as noted, reflects nonfood services. But costs also are higher because of the nature of the food. One reason the cost of transferring 1,000 calories in the Colombian program is greater than the Brazilian subsidy program is the much heavier use of more expensive processed foods in the Colombian program. 34/ The cost of transferring calories free of charge can be estimated by simply dividing project costs by the number of calories transferred. Where food prices are subsidized, the number of calories transferred free of charge can be estimated as the total transfered minus a portion equal to the proportion of the market price paid by the target group. For a program providing a-60 percent price subsidy, 600 of each 1,000 calories transferred would be free of charge. The estimates can only be considered as orders of magnitude, given shortcomings in the data. Since all the programs included components other than food, it is incorrect to charge total program cost to food transfers. But, it is also incorrect to assume that cost of the food represents the total cost of the program since it excludes a number of other costs necessary for food transfers. - 46- In addition to comparing different components in the project, it is possible to compare costs of nutrition components with similar programs outside the project. Examination of Brazil's PROAPE and three other preschool feeding programs in the state of Alagoas in Brazil, for example, shows that PROAPE at $46 costs significantly less than the others -- 55 percent less in one case and less than a third the cost in the two others -- largely because it relies on support from mothers and others in the community, does not need elaborate kitchens and running water, and uses lower-cost foods. Generally, the cost per beneficiary in all of the Bank-assisted programs ranks favorably with feeding programs in other countries, which range from $11 to over $200.35/ To what extent would interventions be affordable for governments if extended nationwide? This would depend on a number of factors, not the least being the importance a government assigns to the malnutrition problem and its willingness to commit resources in this area. Clearly, an intervention such as Brazil's REDE SOMAR food marketing program is affordable, as it is self-supporting. Similarly, a program such as Indonesia's anemia effort, in which costs could largely be borne by employers who receive substantial production benefits, does not present major financial problems. Table 4 shows the costs of expanding other programs nationwide, with those costs expressed as percentages of the national budgets. These, o,f course, are only very rough estimates becaise they assume that the costs per beneficiary observed in smaller programs would continue in larger ones. This could well result in either an underestimate of costs (because additional groups may be harder to reach and have less eager demand for the services) or, perhaps, an overestimate (because of economies of scale). These costs assume the programs all are additive to a budget. In fact, for many countries existing nutrition budgets might usefully be restructured toward more efficient use of resources.36/ 35/ See G. H. Beaton and H. Ghassemi, "Supplementary Feeding Programs for Young Children in Developing Countries," report for UNICEF and the ACC Sub-Committee on Nutrition of the United Nations, October 1979, table 12, 45. Analysis during a recent nutrition sector mission of Bangladesh's institutional feeding program found per child costs there to be $248 a year, most of this however in the form of foreign food donations. 36/ Also, portions of some programs can be borne by the consumer or covered through cross subsidies. A means of limiting the costs the poor incur has been tested in India, where the government for a period required fortification of atta, the mesh of wheat used to make the chappati, the staple of the north Indian diet. Through its price controls, the government encouraged the miller to spread the fortifying costs to consumers of all processed wheat products; upper income groups, who consumed mostly refined wheat flours, carried much of the expense while the fortified atta cost the beneficiaries less than they would otherwise pay. - 47 - Many low income countries are now spending 6 percent of their budgets or less on health and nutrition together. On the other hand, some countries have devoted substantial portions of their budgets to food subsidy programs, generally not perceiving or budgeting them as nutrition programs. In 1975, such programs accounted for 21 percent of Egypt's total government expenditures, 19 percent of Korea's, 16 percent of Sri Lanka's, and 12 percent of Morocco's. In 1981, the urban food subsidy in China accounted for 13 percent of government expenditures. Although for many countries these levels have proven too high to sustain in the long run, the examples suggest substantial commitments are sometimes possible. If reallocation of existing resources or some form of cost recovery were not possible, costs for the actions taken under the projects would range from 0.1 to 2.6 percent of national budgets. India's nutrition delivery component, for example, would cost 1.0 percent of its budget,37/Indonesia's NIPP would cost 0.38 percent and its nutrition education component 0.1 percent (see table 4). Colombia's food subsidy program would be considerably higher at 1.92 percent and Brazil's, at 2.6 percent, would be higher still. If, however, the Brazil program were directed, as would probably be the case, to the urban poor rather than all poor on whom the calculations are based, the costs involved would fall to about 1.5 percent. A national program would actually cost less than projected, since the calculations are based on the distribution of the poor in the Northeast and there are relatively fewer poor in most other parts of the country. The cost would be still less if the project were restricted to poverty areas, as is the case in Colombia where the design was to reach 30 percent of the population through geographic targeting. Cost-Effectiveness To more fully assess the appropriateness of nutrition interventions, it is of course necessary to attempt to quantify their effects or benefits and then relate such measures to costs. A considerable amount of attention has been given to attempting to measure effects in these projects. In principle it would have been desirable to have obtained complete information on all the effects or benefits indicated in table 5. In practice, information of this sort is never obtained comprehensively and in the case of these participating projects there was greater success in some areas than others. Fortunately, findings of Bank research conducted during the same period on impact of nutrition have enriched our understandings of the benefits (especially those examining the relationships of nutrition to productivity and education) that might be expected from nutrition interventions. (See Appendix C for a discussion of research findings.) 37/ Based on actual experience through 1983, the cost of the Tamil Nadu project would be lower, 3/4 of 1.0 percent of the revenue budget. -48- Table 4. Annual Recurrent Costs for Expansion of Nutrition Programs to Entire Country, as Percent of Total National Budget Annual recur- rent cost for Project cost Total Expansion Popdlation to population of for expanded national cost as be served by 100,000 coverage budget percent expansion (thousands of (millions of (millions of of national Project, component, and date (thousands)' dollars)a/ dollars) dollars) budget Brazil Food subsidy (PINS), 1980 118,332 399.34 472.55 18,183 2.60 For urban areas only 66,685 399.04 266.30 1.46 Colombia b/ Food subsidy, 1981 7,928c/ 437.94 34.72 1,807 1.92 Indonesia NIPP, 1982 122,722 77.80 95.48 25,052d/ 0.38 Nutrition education, expansion phase, 1981 120,507 17.91 21.58 22,002 0.10 Tamil Nadu Weighing and feeding program, 1982 705,725 71.38 503.75 31,290 1.61 1985 projected, at 1982 prices 45.53 321.32 1.03 a/ From table 1. b/ Expansion to all designated "poverty" areas in country (30% of total population). c/ Total population of Colombia was 26,425,000. d/ Projected. Quite good information has been collected for most of the components on process. For impact measurements, several of the Bank project components provide data for analyzing cost-effectiveness. In Indonesia, the nutritional status of children up to 24 months old in five areas where nutrition education under the project was offered can be compared with that in five areas that receivNed different programs of nutrition education and other nutrition inputs from three major government programs, including NIPP. One year after the full implementation of the communications strategy, there were significant differences in percentages of malnourished children, as measured by weight for age: -49- Percent of children in Nutrition Percent of standard education Comparison weight-for-age villages villages Less than 60 0.0 1.2 60-74 10.0 18.2 75-89 42.4 40.0 90+ 47.6 40.6 There were no significant differences between the villages before the program and no other factors were found to explain the improvements. The NIPP component cost $97,760 annually to cover a population of 100,000, the nutrition education $65,180, or $33,990 excluding fixed costs. If the control group outcomes are assumed to be representative of NIPP outcomes, the nutrition education approach appears to be the more cost-effective approach, producing a greater impact at a lower cost. Based on cost estimates and the finding that 40 percent of the children in the nutrition education program were growing more rapidly than those in the comparison group, the cost per child with improved nutrition status was $9.85 per year during the pilot phase and has been projected at $5.15 a year for an expanded program. Data from Tamil Nadu on numbers of malnourished children can be used to compare the project with other nutrition projects, while data on the change in number of deaths can be used for comparison with other health projects.38/ If the current program quality is maintained throughout the project,* there will be about 107,000 fewer cases of severe and serious malnutrition, and an estimated 76,000 fewer cases of moderate malnutrition. Calculations based on the probability index for children 1 to 36 months old in Punjab and a reasonable set of assumptions concerning external influence show the number of deaths averted by the project is 12,385 (see tables 1 through 3 in Appendix D). But that figure excludes children below 6 months old (because they were not counted in the baseline survey) and those 37 to 60 months old (because their death rates are not available); thus the estimate of total deaths averted is low and hence the cost estimate per death averted is high. The cost of removing a child from malnourishment (depending on the severity) ranges from $33 to $126 per year ( during the initial project period it was $51 to $193). These Tamil Nadu costs are strikingly lower than those of well-documented, more traditional nutrition rehabilitation programs in Haiti, where improvement in weight in comparison to control 38/ Project benefits, measured in the changes in numbers of children malnourished, are calculated by comparing changes since a baseline survey in the control and pilot blocks. The number of deaths averted is estimated by applying rates from a probability,index of death by nutritional status for Punjab children to estimates of changes in the number of children in each nutritional status group. See A. A. Kielmann and C. McCord, "Weight for Age as an Index of Risk of Death in Children," Lancet, June 10, 1978, pp. 1247-50. * See footnote (*), page 26. - 50 - children was estimated to cost $600 per child and the cost to eliminate a case of third-degree malnutrition compared with control groups was $3,600, and in Guatemala where the cost for the latter was $5,300.39/ Although mortality prevention is not necessarily the principal objective of nutrition programs, the costs of the Tamil Nadu project are within the range, albeit the higher end of the range, of more conventional health approaches to reducing mortality. The cost per death averted in the Tamil Nadu program is estimated at $1,482 per year ($2,260 in the earlier phase). These figures, particularly the longer-run cost figures, are generally lower than cost per death prevented through malaria eradication, polio immunization, community water supply and sanitation, and hospital care. Additional and probably more important, if as yet unquantified, benefits of nutrition programs include reduction in morbidity and improvements in the quality of life generally as well, for malnutrition is a state of ill health that detracts from the quality of life (see table 5). Also, the demonstrated relationships between malnutrition and productivity and between malnutrition and education portray obvious added benefits from better nutrition. Even within the narrow context of mortality control, a nutrition intervention aimed at children below three years of age has added benefits. A reduction in mortality of that age group, for example, is likely to result in a reduction in subsequent births, making nutrition a potential family planning intervention. Table 5. Expected Benefits of Interventions that Increase Food and Nutrient Intake of Beneficiary Reduced Reduced Improved Improved severity and infant Beneficiary physical mental dev- Increased duration of Reduced and child development elopmnt a/ productivity disease undernutrition mortality Child Preschool, 0-3 years X X ... X X X Preschool, 4-6 years X X ... X X ... School-age, in school X X. ... X X ... School-age, out of school X ... X X X Adult Male ... X X X.. Female, pregnant or lactating ... ... X X X X Female, nonpregnant, nolactating ... ... X X ... a/ Resulting fran increased activity levels and mental alertness. 39/ Beaton and Ghassemi, op. cit. - 51 - The cost of preventing a case of blindness through vitamin A - distribution has not been computed for the Bank-assisted projects (three include vitamin A) but similar delivery in a successful vitamin A program in Bangladesh is estimated to cost about $350 for each case of blindness prevented (in addition to greater resistance to infection and other benefits from vitamin A). This compares to an estimated $6,700 to prevent a case of blindness in the Onchocerciasis Control Program (which also provides other benefits).40! In Brazil the principal benefit sought in the PROAPE program combining nutrition and educational inputs for preschoolers was an improvement in cognitive development and subsequent school performance. For PROAPE 93 percent of the 22,298 children in the sample completed the first year of school, 68 percent of whom passed; in the second grade, 94 percent of 14,408 remaining children completed the year and 79 percent of these passed. Comparable figures among non-PROAPE children are 91, 57, 94, and 62 percent. The total cost (including preschool services) per second grade graduate who had been in the PROAPE program was $442. This compares with a cost per second grade graduate who had not been in PROAPE of $494. (See Appendix D, table 4.) A strategy that combines investment in PROAPE with schooling, therefore, appears on the basis of these results to be more cost-effective than investment in schooling alone. In short, compared with earlier interventions (which generally had either indiscriminate coverage or, if targeted, often involved one-on-one treatment and oversight from highly trained professionals), the large-scale concepts tried under the Bank-assisted project efforts appear to have shown the feasibility of pushing per capita beneficiary costs down to relatively low levels. In some instances these costs could be met through restructuring of existing programs. However, even if entirely additive, at these levels most of the interventions appear to be affordable if extended to a national scale. The low-cost nutrition education as practiced in Indonesia looks particularly attractive. That it was cheaper than programs requiring food commodities comes as no surprise; the question is whether it is effective. The evidence has shown that nutrition education alone can make a difference in improving nutritional status. Nutritionists have long held out the promise of this possibility; the Indonesian experience is the first time it has been demonstrated in an operational setting. Even at the high end of the cost spectrum, a consumer food subsidy program if targeted properly is affordable in certain contexts. Also, in those countries adopting policies aimed at achieving more equitable distribution of income and for whom cash transfers are politically unacceptable, the targeted consumer food subsidy looks to be a 40/ Michael Mills, "Cost-Effectiveness of Food and Nutrition Intervention Programs," background paper for Bangladesh Food and Nutrition Sector Review, PHN, 1984. - 52 - financially feasible and politically more viable means of working toward that end. (Apart from whatever income is transferred in such a program is the question of whether it increased food consumption. In Colombia, the only Bank project in which this was examined, it appears to have done so.) Vitamin A programs clearly are cost effective. Brazil's preschool feeding and stimulation program is costly relative to other nutrition interventions but may be a less costly way than added construction and other conventional approaches to strengthening an education system. The Tamil Nadu approach is more costly than some and less costly than other common approaches to preventing death. The other benefits of that program, however, go well beyond most of the health approaches. Finally, through this accumulated project experience and the concurrent research undertaken by the Bank, a good deal has been learned about nutrition economics, although much remains to be learned. The economic analysis employed by the Bank in conjunction with these projects constitutes a useful advance in the ways of thinking about costs and cost-effectiveness in this relatively new field.41/ Impact of Nutrition Lending An estimated 2.7 million people were reached directly by the projects (11 percent more than planned) as the following figures show: Beneficiaries* Project Planned Actual Brazil 97,000 108,000 Indonesia 150,000 700,000 Colombia 1,800,000 1,500,000 India 420,000 426,000 (*These country figures constitute the total of the number of beneficiaries referred to for each component discussed in earlier pages.) Since other programs in the four countries (including those sponsored by other external agencies, particularly UNICEF and USAID) were under way simultaneously, it is often difficult to link outcomes with individual actions or institutions. However, each of the Bank-assisted projects appears to have had some effect on policies, programs, and institutions -- in addition to the impact on target groups, covered in the descriptions of country projects. Impact on Policies and Programs Some of the Bank-assisted activities have been expanded, even to a national scale, and certain of them have been taken up, or are being planned for, by other countries. In Brazil, PROAPE, which began in a 41/ See background paper to this report by T. J. Ho, "Economic Issues: Costs, Affordability and Cost Effectiveness." - 53 - single state, has been extended at government expense to every state and territory; Colombia and Mexico are planning similar efforts, and Bolivia has shown interest. The consumer food subsidy has broadened (with modifications made as a result of the initial evaluation) from an experiment in Recife to five cities and 1.2 million beneficiaries and the budget has increased more than tenfold. REDE SOMAR has been included in three other Bank projects, and the government of Colombia is introducing a large program modeled on it. The short-term, seasonal credit (CAP) to very poor farmers in Sergipe has been picked up as part of seven of the ten Bank-assisted rural development projects in Brazil's Northeast, and social extension agents in 15 Bank-assisted rural extension services now perform the kinds of nutrition services introduced in the Sergipe component. The Sergipe model has become an integral part of the thinking of the rural extension agency. And the finding from the Sergipe evaluation that those the project was trying to help seldom had access to land contributed to the Bank's decision to include land purchase in rural development projects. In Indonesia, most of the elements tested in combination for the first time in NIPP (e.g., growth monitoring, diarrhea control, nutrition education, immunization, supplementary feeding, family planning) have been adopted as the standard package for the national program. The principles and approaches of the nutrition education component are being incorporated into a program now covering 30,000 villages, and expected to double in the next five years. Gardens have become a regular part of the agricultural extension service program as a result of the home and village gardens component. The government has adopted the NIPP concept of retaining "village laboratories" to test new program ideas. And the project's supplementary feeding approach will become the standard for all other government village feeding efforts. One of the more impressive outcomes of the project in Indonesia has been the spontaneous response of villages. Communities near NIPP villages have organized themselves and established similar programs with their own resources. (One regency, Bojonegoro, has arranged funding from local sources to permit NIPP to continue there.) The four governments are making unprecedentedly large budgetary commitments for nutrition actions. They put up $148 million of the $233 million for the four projects and were willing to borrow for the rest. They also spent considerably more than that for nutrition, albeit not always effectively. In several instances, dialogues in conjunction with projects helped to discourage initiation of questionable actions that would have involved still larger expenditures. Impact on Institutions The record of the nutrition projects in strengthenin institutions is mixed, as is the case for most Bank projects ./41 In Indonesia, for instance, the contribution of the.project appears to have been considerable for several key institutions; the Brazilian nutrition 42/ Sixty-one percent of Bankwide projects have significant difficulties in institution-building. -54- project, by contrast, seems to have had little effect on the workings of the national nutrition agency, INAN. The institution-building strategy included sizable training and technical assistance elements to broaden capacity to plan and operate programs. Training has covered a broad range of skills and sophistication. In Indonesia, for example, the head of nutrition in the national planning agency was completing a Ph.D. in nutrition planning at the same time village kaders were learning about nutrition education. The project also funded 24 long-term fellowships and 180 short-course grants for technical specialists, training of kader supervisors, and training of nutritionists. Indonesia's Nutrition Academy, whose facilities and staff were improved, has increased its annual number of nutritionist graduates from 60 to 200 and outside evaluators have praised the quality of their preparation. A national staffing plan.for nutrition, developed in the project to accompany the Fourth National Development Plan, is being implemented on schedule. (The drafting of the food and nutrition chapters of the Plan was the responsibility of the officer mentioned above who received his doctorate under the project. Upon receiving his degree, he was given responsibility for both health and nutrition in the national planning agency.) The execution of such training activities and the subsequent performance of those trained offer some of the more encouraging accomplishments of most of the projects (Brazil being the exception) and confirm the importance of developing the skills needed in a nutrition program. Thus Indonesia's intention of dramatically expanding nutrition activities in its next five year plan appears realizable. Nutrition services not previously available are now being provided by 3,500 Colombian primary health workers, 250,000 Brazilian preschool teachers, 1,800 Tamil Nadu community health workers, and more than 3,000 Indonesian village workers. The institution-building strategy also included development of an evaluation capability. The evaluation studies themselves were uneven, and in some cases the performance was disappointing, but much useful experience has been gained. In all but the first project, but most effectively in Tamil Nadu and Indonesia, management information systems have been deve- loped to monitor nutrition status and program progress. Also in all of the projects, sizable numbers of social scientists have been engaged for the first time in nutrition issues and progress has been made toward a general appreciation of evaluation. In Brazil, for example, there is a growing interest in setting aside for evaluation a fixed percentage of budgets of future nutrition projects funded by the National Economic and Social Development Bank (BNDES). In Indonesia, long-standing activities outside the project are for the first time being evaluated. The least successful element of the institution-building strategy has been the strengthening of entities responsible for directing or coordinating nutrition activities. The endemic problems of weak organizations in social sectors and the inherent difficulties of improving internal organization, procedures, and staff, along,with the demands on institutions to take on new roles and responsibilities, clearly are more difficult and time-consuming than were anticipated. - 55 - Personnel shifts are an important, seldom reported part of institution building. In two countries top-level changes were made because of the projects, leading to marked improvements that will be felt well beyond the projects. The Indonesian who came to prominence with the opportunity to build and direct the food technology institute recently, in addition to his FTDC role, was named Advisor on Food Technology to the Minister of Agriculture. The nutrition projects also helped strengthen disparate nutrition communities. The projects played an important catalytic role in providing opportunities and a rallying point for existing nutrition activities. New groups, including disciplines not before involved in nutrition, have been given reason to address nutrition issues and now form part of the nutrition constituencies of their countries. The Bank's strategy has given heavy emphasis to sensitizi-g senior policy officials to the need for explicit attention to nutrition and institutions to deal with it. In all four of the countries, finance or planning ministers (and in three, heads of state) have made important pronouncements that helped make nutrition improvement a legitimate goal of social policy. (In three countries, this was new; again, India had taken a strong stand earlier.) Bank Management of Nutrition Projects Nutrition projects have been supervised more efficiently with each successive project, as experience has grown within the Bank. Supervision costs for the Tamil Nadu project, for example, are 36 percent less than they were for the Brazil and Indonesia projects at a comparable stage of execution. The heavy supervision reflects, in part, the considerable technical assistance provided to borrowers who lacked experience with approaches being implemented.43/ The impact of this extra supervision was most notable in the Indonesia and Tamil Nadu projects where early delays gave way to timely and adequate execution of activities. In Brazil, the best counter example, the Bank offered sporadic and variable supervision and simple problems were too often allowed to produce inordinate consequences. The project did not receive the sustained attention needed, being at one time or another the responsibility of six different project officers. At one period, the project went for nearly a year without a supervision mission.- The Colombia project for periods also was undersupervised as a result of preoccupation with supervision costs. The annual supervision time for nutrition projects was 9.8 percent higher than for urban projects and 19 percent higher than for rural development projects. More staff weeks were required for field supervision on each of the nutrition projects than-on the first urban projects in 43/ The heavy technical assistance requirements were anticipated during project preparation and funds were allocated in the projects for this purpose. With the major exception of Indonesia, however, the countries often were unable or unwilling to tap outside expertise. - 56 - Brazil and India (and slightly more in Indonesia), while the urban project required more time in Colombia. Just as supervision became better with successive projects, projects were prepared, appraised, and presented to the Bank's Board of Executive Directors for approval more rapidly with each project. The staff time involved in processing a project from identification to Board presentation dropped from 165 staffweeks for the Indonesia project to 99 staffweeks for Tamil Nadu. This compares to an average 177 staffweeks for all urban projects and 220 staffweeks for integrated urban projects. Also, the elapsed time from loan signing to loan effectiveness has declined, averaging 2.8 months for the four nutrition projects. Bankwide average is 5.2 months. In short, with experience, substantial efficiencies have been achieved. Future nutrition projects could be appraised and implemented more quickly and economically, both to the borrower and the Bank, without undue risk to project and sector objectives. One lesson that emerges from the management experience is that an on-the-scene agent of the Bank could help greatly to deal with small day-to-day issues that, when not attended to, can lead to major difficulties. This is particularly important during the start-up period of a project. Such an agent might be a local recruit (perhaps part-time) of the resident mission or a staff member of a sister agency, for example, UNICEF.44/ IV. NUTRITION IN OTHER BANK-SUPPORTED PROJECTS Projects with Nutrition Components For a two-year period in the late 1970s, the Bank concentrated operational energies on incorporating nutrition components and subcomponents into projects in other sectors. These required considerably less time to prepare and supervise than free-standing nutrition projects; yet they appeared to offer a means of introducing nutrition ideas to national leaders and to achieve the broader nutrition objectives of nutrition projects. Nutrition work undertaken in projects in other sectors is reflected in table 6 and the actions are summarized in Appendix E. The nutrition activities in 23 urban projects have generally been health-oriented -- training paramedical workers, surveys or surveillance of nutrition, and addressing specific nutrient deficiencies. Fifteen agriculture and rural development projects have included either home gardens, small fish ponds, food fortification, a food technology center, growth monitoring and treatment of nutritionally vulnerable groups, nutrition education, or surveys and training for future programs. A few 44/ The need in some cases to augment resident field staff-for such projects is discussed in "Focus on Poverty," a Report by a Task Force of the World Bank, February 1983. - 57 - projects contain food for work and other food aid components undertaken in conjunction with the World Food Program, although the extent to which these are nutritionally oriented is not clear. The main orientation of nutrition work in 12 population and health projects has been on growth monitoring and nutrition education. Nutrition education also is sometimes included in education projects. Also school lunch programs have been included in two education projects and another is being prepared. Many of the nutrition actions have been designed to complement primary health programs that are, in turn, components of larger projects. Table 6: Nutrition Components and Subcomponents in Other Projects (By Region and Type of Project) Agriculture and Urban Region Rural Development Development Education PHN Total East Asia 5 5 ... 5 15 South Asia ... 6 ... 1 7 East Africa 2 5 ... ... 7 West Africa 1 1 1 1 4 Middle East and North Africa 2 ... 1 3 6 Latin America 5 6' 1 2 14 TOTAL 15 23 3 12 53 The Lessons Although there have been successes among nutrition components in other sectors' programs, implementation generally has been difficult. Nutrition components rarely exceed 5 percent of total project costs and receive little country attention, particularly if administrative or managerial problems of infrastructure or of larger components impede implementation of the overall project. For the same reason, nutrition components often are poorly supervised by Bank staff. Monitoring and evaluation may also be insufficient because the size of a nutrition component does not appear to merit it and because mission members often lack the technical familiarity, the necessary time, and sometimes the interest to adequately supervise the component. This lack of technical familiarity also means that mission members can provide only limited assistance and guidance to those trying to implement the component. Urban projects .have tried to encourage community participation in neighborhood improvements, including nutritional efforts, hoping to - 58 - strengthen community cohesion and contribute to cost recovery, maintenance, and organization of social services. Where community support has been effective it has been time-consuming to build, but neighborhood politics have often precluded such efforts.' / Experience in urban projects has led to a gradual simplification of projects, with fewer special health and nutrition components. Though to date the preparation and supervision of nutrition components has been relatively inexpensive, they have not necessarily been addressed to the most important problems or the neediest geographic areas, but those dictated by the nature and location of the parent project. In most instances they have done nothing to promote policy discussion on nutrition or institution-building. Nutrition Orientation of Projects in Other Sectors Agriculture and Rural Development Building nutrition considerations into the design,of appropriate agricultural and rural development projects is potentially more important than adding nutrition components. The 1980 review of nutrition by the Policy Review Committee mandated that (a) when economic and sector analyses conclude that agriculture-related interventions can make a central contribution to meeting chronic nutrition needs, projects should be considered toward that end, and (b) in projects aimed at improving the well-being of low-income groups, nutrition should be added as an explicit objective, where feasible.406/ The principal aim was to encourage analysis and broaden the scope of project planning to include nutrition considerations such as the choice of crops to emphasize in research, storage, and extension projects.47/ AGR was to develop a capability to guide the work of regional projects divisions responsible for agriculture and rural development to insure that projects were designed and implemented in such a way as to have the maximum feasible beneficial impact on nutrition. At a minimum, projects were to be examined to insure that their effect on food supply, food prices, or incomes did not contribute to a worsening of nutritional status (including the status of groups who were 45/ Learning By Doing, World Bank Lending for Urban Development, 1972-1982, World Bank, 1983. 46/ That depends on whether maximization of nutrition benefits has negative effects on other project goals that more than offset gains. Modest reorientation of project design may have significant nutrition effects without causing unacceptable changes in other areas. Determining acceptable trade-offs among conflicting goals is, of course, a policy question, but the trade-offs must be explicitly considered. 47/ E. g., as called for in the Agriculture Research Policy Paper (1980): "Bank-supported agricultural research projects will give increased attention to nutritional implications of food production systems and, where appropriate, include a focus on foods of major importance to the diets of the poor." - 59 - not direct project beneficiaries). Projects judged to have potentially- deleterious effects were either to be reoriented or to have nutrition components added to offset the negative effects. Guidelines were developed48/ and several projects were designed with food consumption concerns given explicit attention. The Second Nepal Rural Development Project was based on increasing food consumption levels to meet-requirements and the project components flowed from this. The Southern Highlands Project, in Papua New Guinea, which caused a shift from subsistence to cash crops, was modified to provide extension services to increase production in family food gardens and to include other assurances that the modernization effort would not be nutritionally negative. The South Kelantan Resettlement Project in Malaysia was designed to help settlers meet their nutritional needs during the years before rubber trees could be tapped -- a portion of land was withheld for food crops, the government helped to build and stock community fish ponds, and nutrition education was offered. However, agricultural projects generally have not had nutrition integrated into their framework. Bank agricultural staff, as a rule, either believe that malnutrition is for the most part being addressed through increasing food supply, and nutritional issues that go beyond are social welfare problems outside their province, or do not have time (and, in some cases, the training to feel comfortable) to deal with it. Malnutrition is sometimes used as justification for agricultural projects in appraisal reports, with the projects' effects on nutrition assumed to be beneficial. But nutrition goals rarely are more explicitly considered. Health Projects Since inadequate nutrition is so much part of the health problem and because the interaction of malnutrition and infection has more serious results than the combined effects of the two working independently, it seems logical to include nutrition in the planning and implementation of health programs. Malnutrition commonly has been identified in the Bank's country health sector analyses as a major factor in poor health status. Yet, actions of a sizable scale to counter the nutrition problems have not regularly been incorporated in health projects. (Small nutrition elements have been; see Appendix E.) This partly reflects the nature of the projects (e.g., largely infrastructure projects would not appear to be 48/ See Nathan Koffsky, "Nutrition in Agriculture Sector Work Guidelines," PHN and AGR, 1982, and Per Pinstrup-Andersen, "Nutritional Consequences of Agricultural Projects: Conceptual Relationships and Assessment Approaches," World Bank Staff Working Paper No. 456, May 1981. Also see Shlomo Reutlinger, "Nutritional Impact of Agricultural Projects," AGR/ARU 14, August 1983. - 60 - compatible with nutrition components49/), the sense that solving malnutrition problems falls largely outside health jurisdiction, concern that adding a nutrition dimension would add unmanageable complexity to the project, or a lack of familiarity with the techniques and tools to incorporate nutrition.. Although there are exceptions, by and large the efforts to add a nutrition dimension to other activities in the Bank have not been successful. There are probably several reasons for this including inadequate design, promotion and follow-through, limitations of time and interest of projects staff, and concerns by-staff about decreasing rather than increasing project complexity. It also, however, reflects the way the Bank is structured and staffed along sectoral lines (nutrition being a multisectoral subject) and the Bank's reward system. Without addressing these latter issues, it is not likely that more major inroads on nutrition through other sectors can be expected. V. ANALYTIC ACTIVITIES IN NUTRITION Country Economic and Sector Work The decision was made in December 1980 for nutrition to be given more emphasis in the Bank's county economic and sector work. This was to provide a context for operational work, to better inform ourselves for incorporating nutrition in the policy dialogue with member countries, and to provide a form of technical assistance. Nutrition sector guidelines were drawn up, and 21 reviews dealing with the nutrition problems of 16 countries (Brazil, China, India, Indonesia, Lesotho, Mauritania, Nigeria, Pakistan, Peru, Philippines, Rwanda, Thailand, Upper Volta, Yemen Arab Republic, Zambia, and Zimbabwe) have been completed.50/ The reviews are of varying depth, a few scan the subject only lightly. An assessment of the sector work suggests that twelve of the papers provide useful information for formulating Bank projects, five are 9/ In fact, early consideration of nutrition services might influence the nature and design of facilities to be constructed and the type of equipment to be purchased -- just as it should influence training programs, logistic systems, and health information systems. (See PHN's "Incorporating Nutrition into Health Projects" guideline, June 1982.) 50/ Summaries of country reports are available as part of the background panpr (see Appendix A). Sector reports not completed as of September 1983, e.g., that of the nutrition sector mission to Bangladesh and the nutrition portion of the PHN sector mission to Uganda, were not included in this review. Nor is sector-type work included that was part of project preparation, e.g., Senegal, or was financed under a project loan, e.g., Brazil. - 61 - fair guides, and three are poor. Some of the studies provide new insights for governments or into particular situations. In eight countries the work has contributed to at least some modest extent to government policy or to the Bank's or other donor's project formulation. (In six cases it is too early to tell whether studies were useful, and in one, the brief sector scan enabled PHN to decide not to explore further at this time.) The sector work on Brazil appears to have contributed to national planning and was explicitly acknowledged by Brazilian officials and used in policy documents. The Brazilians recently requested an update, in light of current austerity measures. The Bank's sector work is being used by Indonesia for planning the food and nutrition chapter in its five-year plan. Sector work led the Yemen Arab Republic to establish a nutrition planning unit, was partly responsible for more emphasis on relatively neglected crops in a Bank-aided agriculture project in Zimbabwe, helped encourage Nigeria to carry out a national nutrition survey that: will be followed by discussions with the Bank on a possible nutrition project, led to plans for a similar survey in Upper Volta, and has apparently been taken into account in the recent change of commodity mix of Morocco's consumer food subsidies. The Bank's sector work contributed to the design of the Philippine Food and Nutrition Plan (which was the main agenda item at a meeting of the donor consortium) and led to establishment of a nutrition planning unit in the Ministry of Agriculture and a follow-on national food consumption-nutrition survey (aided by Bank agriculture project lending). The Minister of Agriculture credits the Plan for leading to current pilot work in consumer food subsidies. In China, the sector work has both broadened the look at the nutritional value of different foods in agricultural work and led to inclusion of malnutrition as a health problem in discussions with the Chinese. The diversity of the countries and their nutritional problems, and of the scope, intensity, and quality of the papers, makes generalizations difficult. However, in nine of the countries the evidence is strong that nutrition problems are serious and in several cases, where time series data are available, the nutrition status of the poorest has been deteriorating. Seasonal malnutrition was found to be of 'particular importance in nearly half of the countries. One of the surprises emerging from the sector work is the discovery of the size and speed of a shift from traditional locally produced grains (e.g., millets and sorghum) to polished rice and refined wheat, both often imported. Administered prices, favoring refined wheat flour and rice, are a common feature behind these shifts. In some countries, e.g., Zambia and Zimbabwe, micronutrient problems have been detected, possibly stemming from these shifts. This is potentially a major nutrition danger that deserves attention. Other general findings from economic and sector work about the malnutrition problem and its determinants are covered in Appendix F, as are the experiences of countries that have tried to deal with the problem. Lessons the Bank has learned in doing this work also are discussed. - 62 - In brief, examination of nutrition in sector work and in country economic work has been somewhat more limited than had been hoped and mixed in performance, but clearly improving with experience. The quality of recent work is such that it can form one of the blocks for building country knowledge on which economic and social policy as well as nutrition lending are based. Nutrition-Related Research While this paper is organized to deal with operational experience and its implications, the review of Bank research in this field and findings by others during this same period has been an important part of the exercise. Some 55 pieces of nutrition-related research undertaken or financed by the Bank in recent years has provided much that is useful for operational activities: information on the 'nature and causes of malnutrition, innovation in estimating social benefits and costs, insight into the characteristics of food supplementation projects, understanding of the design and targeting of consumer-oriented food-price subsidies, guidelines for integrating health and nutrition objectives in project design, and criteria for determining nutrition's influence in the choice and design of agricultural projects. These studies are discussed in Appendix C. Also an annotated bibliography is available. Among the highlights are the following. Pioneering attempts by Bank staff to disaggregate food intake by income group have led to the conclusion that in the normal course of development, even with a substantial expansion of food production, the nutrition problem is unlikely to be solved; special measures are needed to improve incomes of the poor, including intervention in food markets designed to fill the nutritional needs of the targeted population. Bank research has examined the economic value of improved nutrition. For example, an analysis nearing completion of the costs and benefits of treating iron-deficiency anemia concludes that with an increase of 10 to 15 percent in hemoglobin, an increase of 25 percent in work output can conservatively be expected from the intervention. Bank studies on the impact of iron-deficiency anemia on labor productivity in Indonesia have made a major contribution in this area. The productivity of anemic workers who were given iron supplements for two months was found to have increased 15-25 percent, depending on the type of work performed. The increase in productivity was estimated to be worth $260 for each dollar invested in iron supplementation. These economic findings led to the inclusion of efforts to reduce iron deficiencies in the Bank's nutrition loan to Indonesia. - A second set of studies, focused on the productivity effects of food and iron supplementation for workers in Kenya, found that dietary deficiencies are an important cause of anemia, that anemia is an important determinant of worker productivity, and that provision of extra iron and treatment of hookworm are ,inexpensive and feasible interventions. Food supplementation increased road workers' productivity by 13 percent. Studies of how infant and child malnutrition affects the future earning capacity of children and their contribution to economic growth have - 63 - focused on the impact of school attendance and performance. Nepalese - children whose .height was less than 70 percent of the norm were found, after controlling for income differences, to be 14 times less likely to be enrolled in school than children whose height was 85 percent of the norm and 228 times less so than children whose height was equal to the norm. These findings are consistent witl those on data collected by Bank staff from three provinces in China where low height-for-age was consistently found to be related inversely to performance. Analysis of studies in Brazil showed a 15 percent advantage in school performance by those who had been part of a preschool nutrition program. Analysis of data from a study in Colombia shows significant effects of both nutritional supplementation and home education for mothers on children's diet, growth, and school performance. Subsidized food prices and distribution policies in Sri Lanka, India, Pakistan, Bangladesh, and Indonesia were found to have eliminated between 16 and 59 percent of the calorie gaps -- the quantities required to permit an adequate intake for everyone -- in those countries. The Sri Lanka rice ration program, according to one study, was directly and dramatically linked to food and nutritional adequacy and low death rate. Much of the Bank research emphasizes the interaction of health and nutrition. Research on parasitic infections accompanied studies of iron-deficiency anemia in Indonesia and Kenya. One study found that heavy ascaris infections caused as much as 20 percent of ingested calories to be lost, greatly contributing to malnutrition and reducing the effectiveness of food supplementation programs. The benefits of deworming undertaken twice a year, the study concluded, could be as high as ten times the cost. An examination of ten field projects that combined nutrition and health services concluded that such programs in the hands of able administrators and in populations as large as 60,000-70,000 can reduce infant and child mortality by a third to a half within one to five years, at a cost of less than the equivalent of 2 percent of per capita incomes. While some of the Bank research is useful primarily for Bank decisionmaking, the bulk of the research findings adds to the general stock of knowledge. Some of the work has led to sustained and important international discussions that have helped to clarify the issues. Bank researchers are said to have made particularly important contributions in developing ways to gauge the magnitude of energy deficiencies on the basis of average food availability data; the nutrition effects of price policies and agricultural projects; the impact of infant and child nutrition interventions on school performance and future earning capacity; the impact of adult nutrition interventions on labor productivity and the impact of labor productivity on food need; and the social benefits of nutrition interventions. Overall, the volume of nutrition research has been substantial and the work productive. It has not been possible to quantify the costs involved or to compare costs and output to other areas. - 64 - Others' Findings on Nutrition While the Bank has been undertaking its initial work in nutrition, new knowledge, relevant to Bank operations, has emerged from the broader research community and pertinent operational experience has been gained by governments and other aid agencies in this relatively new field. This is summarized in Appendix G but several items deserve special mention. At the time that Bank nutrition projects were designed, the prevailing assumption of the scientific community was that increases in a malnourished child's nutrient intake (or decreases in his nutrient losses) would lead to growth. Accordingly, changes in heights and weights were used as key indicators of project benefits. Recent evidence that energy consumption that does not show up in growth is used in increased physical activity suggests that the measures used in evaluating nutrition projects probably did not capture a good portion of the benefits derived. Lower levels of physical activity are now seen as an accommodation to lower energy intake. In children this may be reflected in cognitive and affective development, and in societies the accommodations may set the norms for activity. - Intergenerational effects of malnutrition now are being documented. Nutritional status is a cumulative manifestation of deprivation, including parental malnutrition, that must be dealt with in terms of decades rather than single years. Mothers of short stature due to chronic childhood malnutrition are more likely to have small babies who, in turn, are more likely to be malnourished and likely candidates for early death. Nutritional anemia, probably the most widespread of the micronutrient deficiencies, affects about half of the women in developing countries as well as sizable portions of the male and child populations. It has recently been demonstrated that it can deleteriously alter brain chemistry and function and thereby influence human behavior itself. Fortunately, a breakthrough in food technology by Indian scientists now makes it possible to fortify a nation's salt supply with iron, at a cost of $0.05 to $0.09 a year per person. Research in Central America has demonstrated that fortifying sugar with vitamin A can reduce nutritional blindness and the severity of measles. Also, in recent years, vitamin A has been given in high-potency periodic doses at reasonable costs. In Bangladesh, which has a highly focused village-level administrative set up, distribution of capsules substantially reduced vitamin A deficiency. About 2,500 cases of blindness there are estimated to be prevented each year. Research in Zimbabwe and India has shown that children with the greatest weight deficit show the greatest response to feeding programs. Thus, directing feeding programs to children with third-degree malnutrition or those "at risk" should be highly cost-effective. However, maternal supplementation looks to be the most cost-effective means of reducing child mortality and improving the nutritional status of those who survive, as well as protecting the health of the mother, which is indirectly essential for child health and survival. -- 65 - Programs encouraging farm families to switch from subsistence to cash crops have on occasion been shown to have negative nutritional consequences. They are said not only to change incomes and relative prices but the distribution of income and budget control within households, exposure to expensive foods and nonfoods, the position of women and demands on their time, and the pattern of ownership of resources. Generally, however, evidence on cash cropping is too scarce and too fragmented to be of much use at this stage for the design of policies and projects. In many areas, malnutrition and nutritional damage have been discovered to vary significantly with the seasons, a finding that has important implications for the design and timing of feeding programs. Much has emerged in recent years from the analytic activities of Bank staff that should enhance future nutrition operations. (Some of these understandings already are reflected in the Tamil Nadu projeci) The combination of what has been learned from the Bank's project work, along with sector work and research -- and findings from the outside nutrition community during this period -- places the Bank's staff in a considerably stronger position than it was at the outset, when there were meager information and experience on which to mount nutrition operations. VI. APPLYING THE EXPERIENCE TO FUTURE BANK ACTIONS Is malnutrition a development problem? If so, are there things that can be done about it? And things that the World Bank can do about it? These were the questions that concerned the Executive Directors during the initial discussions on nutrition in the mid-1970s. Now, nearly seven years after the first project was approved and some 75 pieces of research and other analytic work later, it is clear that the problem merits attention as a development issue. Bank research has shown that malnutrition affects not: only mortality but learning, school performance, and productivity. In several forms it affects substantially more people than do diseases that international programs have been mounted to fight -- there are four times as many cases of serious vitamin A deficiency as onchocerciasis, nine times as many cases of iron deficiency anemia as of schistosomiasis, and the number of cases of energy-protein malnutrition probably exceeds many other diseases combined. Moreover, the nutrition problem for many countries worsened between 1970 and 1980. Twenty-six of ninety countries examined had a decline in per capita calorie consumption. An additional twenty had an annual compound growth over the decade of less than 0.5 percent, some being at a virtual standstill (e.g., Brazil at 0.2 percent). (Because of data limitations, the numbers presented in figures 2 and 3 are national averages. To the extent that income distribution is deteriorating, as it probably is, particularly in the poorer countries, the figures understate the problem.) The near-term outlook is grim. The malnourished generally are landless or have too little land to be assisted,by the Bank's -66 - Figure 2. Number of Countries at Different Levels of Per Capita Daily Calorie Consumption in 1970 and Rates of Change During 1970-1980 ;IM I Iss) I t*I Z..i 26 to 20 L--*1 14 30 . ~ ~ ~~ Is pal t1ai_----le ll... ia I Base P.rtod Daily Calorie Supply *.,g I / jnlj / "u I / tA / Il j 198t o / ast I / 41- 1ls 2400 calories and over l as / IsII / *I I /./I I last I l St/ / J: / I~ / ISI I IM1/ 1301 1 z// -/la dl I l i I L 1 l 26 2200 to 2400 11l I l101 1 . 1 IA / last I as81 1 asi / Iasi 19A!i .. A/Iasi~/ lost I IM~/ last I Ilst/ ~ I I Ia 1 1 *l / 3 12I I 24 1141 I - 1aI - Afz * - I- - I las A / lag A / Z.*/ / Ifll / / Ia / RAI / ea: H / Ical / 2000 to 2200 / Ias i / Ial I / Iai I **.. ]-I / / A981/ / 1991/ / I!SA/ 4*/I 1991/ / 7// sIa / d/7 L-/l / Iasi I Y less than 2000 / as I /lot I Ia I P / IPSI / 7 12 < 0 0-o.5 0.5-1.0 1.0 * Annual Rates of Change in Per Capita Consumption (%) Figure 3. Population (in Millions) Living in Countries at Different Levels of Per Capita Daily Calorie Consumption in 1970 and Rates of Change During 1970-1980 994 349 316 476 Base Period Daily Calorie Supply / lt A / / / / IsI / ._. . 2400 calorie and over/ .. / tast a ..;/l / Z/I/ / to / 2S0 / 85 / 93 / 438 / *-.* / / / / / - - -- -- - / --- - -- - - -* / . - 2200 to 2400 / lIasi / ../l / L./ / ./ lis/ /1 1M I/ I 117 / 54 / 04 / 42 /297 .*1 *. / / ..-. / .-- / Iasi A /--/I // L/ ./ 2000 to 2200 1as1I loel I / I./1 / Ias i / la Ias / / / / isl A 147 45 / 127 / 112 / 451 /Jae) I /Z// ./I / tool I /// Iasi I lcs than 2000 / ee / / *.s / I I / l/ d/ / lf/ / 700 / / 20 / 229 /99 .n< 0 0-0.5 0.5-1.0 .0 Annual Rates of Change in Per Capita Consumption (Z) - 67 - projects. The Bank's food security policy study now under way indicates that special efforts will be required to reach this group. From the Bank's analytic work, something has been learned about malnutrition and its determinants, and experience has been gained in dealing with the problem. Also, the Bank's work and that during the same period by others has shown there are several efficacious, affordable means of addressing;malnutrition. Experience has demonstrated that Bank nutrition projects can make a contribution. The first two nutrition projects, Brazil and Indonesia., began slowly and had serious difficulties but finished strong. Most of their targets were met, as were those of the third project, Colombia, which is still under way. The fourth project, begun in Tamil Nadu in 1980, appears at mid-point to be meeting all of its performance objectives. In numbers of people who benefited, physical structures completed, and other commonly used process indicators, the four projects were very successful. In delivery of services, most of the components met their targets, and in all an estimated 2.7 million people benefited from the projects -- 11 percent over the estimates in Bank appraisal reports. The efforts appear to have penetrated more deeply in the poorest segments of the population than many Bank projects. Policy impact, as measured by government resources committed and policies and programs changed, was substantial in several cases. Less than had been hoped was learned about the nutritional impact of the projects, although some highly useful data emerged. In short, following rocky beginnings,. the nutrition projects made a creditable showing. An understanding of what went wrong (and right) -- and why -- is essential to building a new strategy. Project Experience Implementation. The interventions that worked most smoothly were the nutrition services delivery in Tamil Nadu, nutrition education in Indonesia, preschool feeding and stimulation (PROAPE) in Brazil, anemia control in Indonesia, and the food subsidy and marketing efficiency programs PINS (and its modified successor, PROAB) and REDE SOMAR in Brazil. The efforts that were most difficult to implement and for which there is least to show were those trying to build nutrition considerations into agricultural planning in Brazil and Indonesia. In general, institution-building efforts fell short of expectations, particularly those relating to building capacity for monitoring and evaluation. Nutrition impact. The most striking impact is seen in Tamil Nadu, where severe malnutrition is 40% less common among 13-to-36 months old children in project villages than in control villages. (By the age of five, those children in participating villages are a nutritionally significant 1.75 kilograms heavier than those who are not.) Successfully conducted evaluations also have demonstrated improvement in nutritional status from the nutrition education campaign in Indonesia and the sanitation program in Colombia. Results look promising-but more needs to be known about the anemia control program in Indonesia and the marketing interventions (PINS/PROAB and REDE SOMAR) in Brazil. The gardens components in Colombia and Indonesia have led to increased production but - 68 - they and the nutrition efforts built on agricultural extension in Brazil have not lived up to expectations in terms of nutrition impact. Such efforts by others bears watching. So do sanitation components. Food processing components in Brazil and Colombia, and perhaps in Indonesia, do not require further attention. (Fortification of processed staple foods, however, would deserve high priority.) Water supply components, as practised in the Colombia project, could be deleted from future nutrition projects, except under circumstances where there is assurance that the pre-conditions exist for the water to have a health and nutrition benefit. Impact on policies and programs. The activities with the greatest influence on government policy and programs were the nutrition education, anemia control, NIPP, and training components in Indonesia, nutrition surveillance and training in Tamil Nadu, and the preschool (PROAPE), extension (PRAMENSE), and marketing programs (PINS and PROAB) in Brazil. Those with the least influence were the efforts trying to build nutrition considerations into agriculture policymaking, as in Brazil and Indonesia. The greatest impact on institutions (aside from the construction of facilities) was on the health ministry in Colombia (through the primary health component) and the preschool education activity, the food marketing system, and agricultural extension program in Brazil; all of these have been substantially expanded or reoriented or both as a result of experience under the projects. Program costs. Costs for programs vary dramatically, from Brazil's self-supporting food marketing program and Indonesia's very low-cost nutrition education effort at one end of the spectrum to the consumer food subsidy programs at the other. Those components that would, if expanded on a national scale, consume under 2 percent of the national budget are the Indonesian nutrition education and NIPP programs, the Tamil Nadu nutrition services, the Colombian consumer food subsidy program, and the Brazilian food marketing program; the Brazilian consumer food subsidy program is on the margin. At 1 percent, the two Indonesian programs and Brazil's food marketing program are viable; Tamil Nadu is close to the margin; the Brazilian and Colombian subsidy programs are not. Comparative costs. Although there are exceptions (such as the Indonesian nutrition education component), in most cases not enough information is available to say that one Bank-assisted program is less expensive than another for the benefit received. In large part this is because such a comparison presumes a commonality of objectives. It does appear from information available that some of the programs can be cost-effective relative to other interventions on which governments already are spending large amounts. Thus major opportunities exist for efforts aimed at efficiencies. Process. A number of lessons for future nutrition work have been learned from the Bank's experience with projects, research, and sector work. (Some of these lessons can be applicable to projects in other - 69 - sectors.51/) Enough is now known about process to devise systems that can deliver goods and services satisfactorily. The projects were able to reach people with better nutrition more efficiently (at a more acceptable cost) than earlier programs. Much has been learned about targeting (by area, by age, by income, by commodity, by season). One of the main findings is that appropriate targeting techniques can substantially lower financial and administrative costs of consumer programs and increase their nutritional effectiveness. Project evaluation. A great deal has been learned about how to monitor and evaluate the impact of projects on nutrition status, and less about the impact itself. The most knowledge was gained from projects that included funds for outside evaluation experts to help on impact studies, as in the Indonesian nutrition education component, or whose evaluations were directly aided by Bank staff, as in Tamil Nadu. Considerably more has been learned about the making of nutrition policy; projects have affected development plans and programs of the borrowing countries and have demonstrated what can be achieved in countries that have at least a modest commitment in this area. Something also has been learned about how the preparation of a loan can influence government policies. The resource leverage of nutrition projects has probably exceeded that of a number of other sectors in which the Bank operates. Project design. The main design lesson (just as other sectors learned during the same period in their early operations) is to reduce the number of project components and the complexity of administration. The multiplicity of potential nutrition actions should not be allowed to dictate complex projects. Analysis to define project -content should be fairly broad but the resulting projects should not be expected to address all the factors that affect nutritional status. Everything relates to everything else, but not equally. Actions should be sharply focused on a few critical needs, but recognizing that these sometimes cut across sectors. Nutrition aims. It is unlikely that a precise prescription for nutrition project objectives can be established (given vast differences in cultures, ecological zones, experience, sophistication levels, and so on), but common guidance can be provided on what to search for in analysis and on the types of measures for addressing different types and causes of malnutrition under various circumstances. Given the diversity of conditions and fast pace of change in the nutrition field, projects must be flexible enough to take advantage of unexpected opportunities. Project length. Projects should probably run for a minimum of six years, if demonstrated improvement in nutrition status of large 51/ For instance, the approaches for nutrition education in Indonesia would be useful for health and family planning education. The training concept employed in Tamil Nadu is one of several features of that project that are applicable to health work. Brazil's PROAPE experience has implications for education projects. The targeting lessons learned in the nutrition projects may be useful to several sectors. - 70 - population groups is an expected outcome. Although something was learned about impact from the projects, some components in the early projects lasted such a short time that it was impossible to measure their impact on nutritional status or learn a great deal from the experience. It should not have been a surprise that governments take time to get services running smoothly when projects require a new philosophy, a new design concept, a new delivery system, a new organization, and new staff. More also is now known than at the outset about the time and difficulties involved in generating reliable data. These findings are largely in accord with conclusions by OED from studies in other sectors. Project performance. The nutrition projects when judged by disbursements and other quantitative measures have improved over time. They generally stack up well against Bank projects in other sectors, particularly first projects. They have outperformed or equaled Bankwide projects in estimating costs and project completion, compliance with loan conditions, and procurement and have done less well in disbursements, management performance, and reporting. In comparison to the first four projects in other sectors or first projects in countries with nutrition projects, the picture is better. Enough has been learned about Bank management of nutrition projects that preparation and supervision of the fourth project was considerably more efficient than the first. Country interest. Experience has shown that some governments are ready to commit substantial resources to nutrition and are interested in collaboration with the Bank. Ten countries have expressed interest in nutrition projects. A request for a follow-on loan, of course, does not necessarily reflect satisfaction with a first project performance. However, it is worth noting that in Brazil; Indonesia and Tamil Nadu second nutrition projects have been requested, and that there is under preparation in Colombia a health project that includes nutrition and was made possible by groundwork laid by the first nutrition project. The Issues Several main issues need to be addressed before considering future Bank directions in nutrition. Does the Bank Have a Comparative Advantage in Nutrition? Questions were raised in early Board discussions on nutrition whether the Bank had a comparative advantage for working in this field. Also, concerns were expressed that the Bank's involvement not duplicate work supported by other agencies. What has experience shown? First, the nutrition roles of the Bank and of other agencies generally differ over time. Before the Bank comes in to help finance nutrition efforts, the country in many instances will have received help for several years from other agencies, primarily in technical assistance and training. The recipient government will have developed a group of professional nutritionists and have accumulated some operational experience. - 71 - Second, other agencies primarily concentrate their aid on technical aspects of the nutrition problem; the Bank's emphasis has been on planning and programming, in keeping with Bank project practice. This discipline has been a forceful tool in helping to bring order to ad hoc nutrition activities. Closely related and also important is the Bank's heavier emphasis on policies relating to financing, pricing, and management and related institutional issues. Third, the Bank's analyses before a project is undertaken are much broader than those of other agencies, which generally have a functional orientation. In such analysis, as in other sectors, the Bank's comparative advantage is in the relatively untapped area of economic and systems aspects of nutrition work. The Bank also has an advantage in access to economic data, including that generated by the projects. Fourth, Bank funds are much larger than those of other agencies. Combined with the Bank's analysis and policy dialogue, they potentially have a substantially greater impact on the effectiveness of nutrition programs and policies. Also, the Bank's nutrition lending increases the importance and policy profile of nutrition. The Bank's access to the ministries of finance and planning has brought the problem of malnutrition to their attention and encouraged them to focus more on issues of identifying problems, priorities, financing, policies, and institutional arrangements. The Bank lacks comparative advantage in upstream pilot project work (where UNICEF and bilateral aid agencies with field staffs and greater speed in programming have a clear advantage); in financing of food aid or redistribution programs (the World Food Program and bilateral agencies.are involved); and in food relief operations (the UN High Commission for Refugees and bilateral agencies are involved). Close working ties with other agencies are thus desirable and such collaboration is feasible.52/ Of the agencies, WFP holds special untapped promise. The Bank can benefit also from the technical skills of staff from the specialized agencies of the UN system. 52/ Through interagency committees, there is a fair understanding of what others are doing. Also, it is not uncommon to include UNICEF and WHO staff on Bank nutrition missions. Cooperation among UN agencies has been unusually successful in the case of nutrition, having been effectively fostered. by an active ACC Sub-Committee on Nutrition (SCN). In recent years a fair degree of consensus has developed among multilateral and bilateral donor agencies on the objectives to be achieved in nutrition and confidence in the array of interventions. See Robert J. Muscat, in consultation with the Advisory Group on Nutrition, "Malnutrition in the 1980s: Roles and Operations of the International Agencies," a report to the Sub-Committee on Nutrition of the UN Administrative Committee on Coordination, ACC-SCN, December 1982. -72- Does "what works" in nutrition mesh with the Bank's comparative advantage? The evidence presented suggests that it does. Moreover, the Bank's contribution to the alleviation of malnutrition is in drawing attention to the problem, assisting in planning, furthering the development of programming discipline, and providing resources to make significant nutrition interventions (often innovative ones) possible.53/ In all these areas, there exists a wide gap between what needs to be done and is being done by others. Should the Bank continue lending for nutrition projects? Because of the growing size of the malnutrition problem and the consequences of better nutrition for human capital development and fertility reduction, the Bank's management has recognized malnutrition as an important development problem, and one the institution should address. If the Bank is to make more than a peripheral contribution to nutrition, it sometimes must have projects directed primarily to nutrition goals. Experience has shown that nutrition projects focus attention on malnutrition problems and attack them in ways that, at least for now, other Bank vehicles have not been able to. Nutrition components of projects in other sectors generally are not large enough to command sufficient attention of either governments or Bank staff to make a consequential difference. Also, other projects (health being the main exception) are often not directed either at the problem groups or to the areas where nutrition action is most beneficial. No question that the primary health care system is an important vehicle for introducing nutrition services and, when possible, should be used. But the health system has significant limitations in scope and coverage. The deteriorating financial situation in many countries in recent years, with obvious implications for the food consumption and nutrition levels of the poor, has generally swamped the benefits of explicit nutrition efforts through health facilities. Depending on the health system alone to respond sufficiently to accommodate such levels of need will only create disillusionment and risk neglect of needed actions in other areas. Nutrition projects have proved viable and they have demonstrated the power to sensitize policymakers, help mobilize a constituency, help catalyze other nutrition actions and institutionalize project objectives in ways that lead to much larger programs. Nutrition projects have been small in relation to the problems they have attacked, but they have been catalytic. Though the operational lessons, good and bad, from the Bank's experience are stressed in this report, the most important contribution of the free-standing nutrition projects appears to have been their impact on the priority, commitment, policy and program choices, and level of country support for nutrition actions. In a number of countries, substantial expenditures already were and are being made for nutrition purposes, and generally there is room for major increases in financial efficiency. Brazil now spends over $900 53/ The problem is compounded by nutrition scientists who strive for a much higher degree of exactitude (and publicly flagellate themselves for failure to find it) than is seen in evaluation efforts in other sectors. - 73 - million a year for nutrition programs, more than half of it on worker and school feeding programs, both questionable nutrition priorities in the Brazilian context. Similarly, Tamil Nadu now spends over $300 million a year on an untargeted midday meals scheme. A main purpose of Bank nutrition projects should be to increase the efficiency of existing programs. Also, as many countries face IMF and other pressures to cut out or s.harply cut back consumer food subsidies and other social programs (and have their own worries about where to cut the budget), the Bank, through a nutrition project, may be helpful in -encouraging nutritionally sensible ways -- at least in avoiding nutritionally damaging ways -- to make those cuts and in demonstrating alternative routes to meet the needs. Nutrition projects are likely to be the best means of dealing with intersectoral nutrition issues that do not fall in the exclusive domain of either agriculture or health ministries or of seizing opportunities to work with, say, social welfare or marketing agencies. In some circumstances, a nutrition project may be appropriate because the nutrition system is more fully developed than the health system, or because adding nutrition would overly complicate management of a health project, or because enlarging government's interest in and capacity to deal with nutrition is a main purpose of the project. Goals for the Bank Nutrition Work Should nutrition efforts be oriented principally to reducing infant and child mortality and its concomitant effect on fertility? Or should there be broader concerns that extend to educability and other developmental aspects of the child population and to the productivity and enhancement of adult lives. The decision is the country's, but the Bank's policy dialogue should stress the potential return from investments in all these areas, emphasizing the more promising directions that may emerge from the epidemiological-type sector analysis. For operational purposes, this could translate into efforts aimed at reducing the number of cases of severe and moderate malnutrition in children under three years of age (and thereby the risk of death and the limitations on human development and productivity); reducing micronutrient deficiencies that are debilitating and costly for the whole population; and increasing the amount of food available to low-income households so that their consumption improves enough to increase parents' incomes, improve children's school attendance, and help the family generally to live healthy and productive lives. Evaluation of Impact What has been learned from evaluations of field operations is encouraging but not conclusive. In many components of the first three projects, impact evaluations were not adequate; lessons from this experience incorporated in the fourth project are producing much better evaluations. Measuring the impact and economic benefits of human resource projects is generally difficult, and measuring the effects of nutrition projects is more difficult than most, particularly in operational settings. The difficulty is especially great for development agencies and governments accustomed to basing their decisions on engineering criteria and financial analyses. - 74 - Since nutrition is a relatively new area still full of unknowns, currently available .measurements may not always be focused on the right thing. For instance it is now accepted that anthropometrics, although still very useful, do not capture all the benefits they were thought to reflect only a few years ago. (See Appendix G.) Evaluation may thus only partly reflect project impact. (Results in terms of growth can be expected to show up in programs addressed to very severe malnutrition, but not necessarily in those aimed at moderate cases, which affect a much larger part of the population and are therefore likely to have greater effect on economic activity.) While all short-run effects may not be known within the lifetime of certain components of the nutrition projects, the long-run, intergenerational effects of malnutrition are unlikely to be known until far beyond the conclusion of the projects. Either one must assume there is little value in these programs, withhold judgment until perhaps the mid-1990s when better measurement tools should be available, or accept the experience and analysis that suggests that the malnourished will benefit from the food and nutrients provided, until more refined judgments can be made. The choice obviously should be the latter. That economic effects of nutrition interventions are not measurable with the precision claimed in other sectors does not deny the significance of the malnutrition problem or diminish the reality that malnutrition is a major contributor to susceptibility to many of the diseases that vast resources are devoted to -eradicating. In spite of the uncertainty in measuring results in an operational setting, the evidence to date is highly suggestive and some governments appear ready to make substantial program and policy commitments to nutrition. Behavioral change and differences in amounts of food consumed must be carefully monitored so that programs can be designed around the most cost-effective, affordable means to bring about these changes. Should the Bank Support Consumer Food Subsidies? One potential nutrition intervention, the consumer food subsidy, is so contentious that it merits discussion as an issue. The high fiscal costs of explicit subsidy programs tempt governments either to set low prices for domestically produced food or to rely on excessive foreign food aid, both of which discourage agricultural production. Certain forms of subsidies are difficult to administer, and those in place are difficult to cut back because they build up political constituencies. But research arid sector work show that few other nutrition interventions offer the possibility of substantially and reasonably quickly alleviating the widespread malnutrition and nutrition-related deaths among poor people in developing countries. In several countries, subsidy programs have helped substantially to reduce malnutrition and its consequences among the urban poor. Moreover, Bank staff analysis shows that the financial and administrative problems associated with such programs can be reduced if the programs are carefully targeted to help the needy. Reaching the rural poor with such programs unfortunately remains a difficult problem. - 75 - Consumer food subsidy programs also often provide a politically palatable instrument of policy to governments whose political objectives include reduction of levels of absolute poverty. While redistribution of financial assets or of income per se might in principle be expected to achieve the same distributional objective, and perhaps in some case more efficiently, political and fiscal realities in most countries sharply curtail the practical potential for financial redistribution. In many countries, on the other hand, the appeal of hunger alleviation makes consumer food subsidies more politically acceptable.54! Whether consumer subsidies are appropriate is a matter that should be considered separately from whether the Bank should support a particular subsidy program. Country-specific analysis by Bank staff, as a regular part of economic work, will indicate whether, in the context of broader pricing and equity issues and development needs, the subsidy approach is deemed sensible (i.e., could be set in place, without adverse effects on agricultural production and with feasible financial and admfnistrative costs, as part of a strategy to deal with malnutrition). The Bank should discuss with the government how such programs can be made cost-effective and their potentially adverse effects minimized. Most developing countries already spend a great deal on subsidies, some of which have no nutritional effects, or even perverse effects. Those programs might be altered to provide greater nutritional impact by shifting the commodity mix, improving the targeting, or extending the outreach. Also when countries for austerity purposes cut back such programs, Bank staff should help search for ways, within the confines of adjustment policies, to prevent the food consumption levels of the poor from getting worse. The Bank should help to make food subsidy programs as nutritionally effective, equitable, and cost-effective as possible, but without now financing the subsidy itself. Instead the Bank could finance the cost of services, facilities, equipment, and evaluation for the program. The Bank should consider various possible means of coordinating with the World Food Program and other organizations involved in, the supply or financing of food. Bank assistance in this area should be largely restricted at this stage to countries that already are committed to such programs on a large scale and where improvements are possible, e.g., concentration on very low-income groups instead of all income groups, on regional rather than countrywide programs, and on seasonal rather than year-round aid. However, the Bank should consider requests for financing experimental projects elsewhere to test the approach -- always remembering that such programs, once started, are politically difficult to stop. 54/ Also, nutrition-type social sector projects have been judged to be "less likely to arouse elite opposition than projects meant: to assist the agricultural production of the poor" and "their benefits, moreover, may be less readily appropriable by elites." (Judith Tendler, "Rural Projects through Urban Eyes: An Interpretation 6f the World Bank's New-Style Rural Development Projects," World Bank Staff Working Paper No. 532, 1982). - 76 - Are Nutrition Projects Replicable? We have seen that there are several nutrition actions with demonstrated benefits that, within the settings tried, are technically feasible, cost-effective, and affordable. Whether they can be mounted and effectively administered on a large scale depends, most importantly, on the existing infrastructure. In the food marketing interventions and the food delivery part of the preschool program in Brazil, for example, which are looked after largely by an experienced and effective food marketing agency (COBAL), prospects are good. Where such an infrastructure does not exist, only time will tell whether nationwide expansion can be achieved. In the Tamil Nadu program and the nutrition education program in Indonesia, prospects are promising. Can lessons learned from the four projects be applied in other, less experienced places? Clearly, certain of the more successful efforts (for instance, NIPP in Bojonegoro and nutrition education, also in Indonesia) depended on unusual local activism; tighter and more expensive Tamil Nadu-type management systems may be needed to replicate and sustain such programs in areas of less commitment. Also the limitations in absorbing external assistance in some countries, particularly in Africa, pose special problems. However, while specifics of the intervention programs will depend on particular characteristics of a country (such as sophistication of staff, monetization of the society, extent of urbanization and infrastructure), the type of analysis leading to development of new projects and the underlying principles of the interventions (for instance, those learned from the nutrition education component in Indonesia) are transferable. Interventions based largely on introducing simple technologies are more easily transferable, provided the basic conditions (centralized processing for salt fortification, for example) are met. Institutional Link for Nutrition Work There is no obvious governmental institutional framework for nutrition. It may fit into a country's health, agriculture, social welfare, or, under special circumstances, planning ministry. Working through the planning department in the early years of the Colombia project was probably the most promising Bank experience, but the appropriate institution will be different in each country. Whether nutrition services should be integrated with others at the field level or operated separately depends on preference and the situation in the individual country. Common thinking in the health field calls for integrated services; the Tamil Nadu experience indicates that that approach also can be workable. Although integration of service will be normally expected, there is no prima facie reason to conclude that it is unreasonable to have a separate nutrition cadre. - 77 - Future Directions in Nutrition for the Bank Results of the Bank's research and sector work suggest that certain strategies for improving nutrition merit high priority in most countries, particularly accelerated growth in the income of the poor and, with few exceptions, in food production. These, already being major aims of the Bank, do not require special attention in the context of nutrition work. However, direct attention must be given to bridging the gap for low-income groups until their incomes and education bring their food consumption levels and nutrition practices up to their nutritional needs and to helping those outside the mainstream of growth. A great deal can be achieved through the Bank's policy work and orientation of projects in other sectors, particularly agriculture, population and health. Efforts had been made to increase the importance of food consumption issues in economic and sector work, to encourage consideration of agricultural projects with a food consumption objective, and to introduce nutritional components to compensate for potentially negative effects of agriculture projects (particularly those in which the Bank supports expansion of cash crops among semisubsistence farmers). Although modest gains have been made, these efforts, for the most part, have not been emphasized or institutionalized. This reflects the - perception among many that nutrition problems are largely macroeconomic concerns calling for fundamental changes in income and income distribution. A clearer definition of agriculture's role is likely to emerge from work now under way on the food security paper. PHN's role at this stage should be to help countries develop sound nutrition policies and strategies; make current nutrition intervention programs more efficient and mount new affordable programs to meet identified needs; and strengthen the capacity to plan nutrition actions and to evaluate their effects. Institutional issues should be given particular attention. A good part of this can be handled in the context of health projects and population projects, either by incorporating nutrition as one of the main themes of such projects or by adding nutrition components. On the operational side, a number of what are thought of as nutrition interventions are integral parts of primary health care -- for instance, growth monitoring, nutrition education, oral rehydration therapy, and food supplementation. In some instances, nutrition actions can easily be labeled as primary health care projects, but in others labeling them as nutrition projects may be an important means of raising sectoral issues -- a health project is a poor vehicle for raising issues of food distribution policy, and certainly for raising nutrition-related agricultural issues. In some instances, particularly those involving intersectoral issues, it may be most appropriate for nutrition projects to be free-standing. - 78 - Nutrition Project Content Insights gained from project experience, as well as from sector work and research, by the Bank and others, can enrich the next round of projects. Among the many forms of malnutrition are four deficiency diseases that are widespread, have severe consequences, and can be treated if trained people, funds, and political commitment are available: energy-protein malnutrition and the diseases due to deficiencies in iron, iodine, and vitamin A. From the myriad of possible actions to address these problems, the Bank should carve out areas for project concentration where links to economic development needs and the previously cited goals have been established and where the problems are not likely otherwise to be addressed on the scale required. In the next five years, PHN should concentrate part of its efforts on projects designed around the Tamil Nadu model. )5/ Where serious malnutrition is prevalent among the very young (and, relatedly, pregnant and nursing women), this model offers a cost-effective means of identifying the nutritionally needy and helping them by providing food supplements for a limited period. The concept -- almost 'food used as medicine' -- would be generally most feasible for middle-income countries. When country concern extends to other members of-the family (and to broader distributional issues), projects should be designed to enlarge the family food basket, particularly during periods of seasonal hunger. (The intent would be to increase the availability of food to the household without waiting for economic development.) Food delivery projects can range from institutional feeding to targeted subsidy programs in the commercial market to efforts to improve the efficiency of the food marketing system.56/ Involvement in food delivery projects does not necessarily suggest that the Bank finance the purchase of food. In institutional feeding programs, untapped collaborative opportunities exist with WFP in which Bank resources could be devoted to strengthening institutional capacity. In countries, particularly in sub-Saharan Africa, that lack adequate transport, storage, and other infrastructural prerequisites to take advantage of donated food in the amounts they require, Bank nutrition projects could be helpful in meeting these needs. For the urban poor, who usually buy food in small quantities and in its highest priced form at the end of the food system, the Brazilian REDE SOMAR program is an especially promising approach that generates income benefits for target recipients by passing along the gains 55/ Following a 'Tamil Nadu model' should not, of course, involve exact replication. Not all the factors found in Tamil Nadu (e.g., a health system in which referrals can be handled) will be found elsewhere; concepts and project principles, however, may commonly apply. 56/ Except for the latter, important recurrent cost implications are involved. Issues of food subsidy, rationing, and distribution also involve important macroeconomic considerations, so close coordination with country economists will be important in formulating possible operations in this area. - 79 - of increased market efficiencies. More needs to be learned on this. Consumer food marketing, a potentially important but largely neglected part of development programs, is an example of an activity that falls between the cracks in the Bank organizational structure; the main work on consumer food marketing has been undertaken in the experimental nutrition projects. A third kind of project should deal with the delivery of micronutrients. Major inroads are now possible against nutritional anemia, vitamin A deficiency, and iodine deficiency. A low-cost technology now exists that makes large-scale fortification of salt with an iron compound feasible and attractive. (In very large countries, a single component project on this may be appropriate.) Sugar can be used as a vehicle for vitamin A, and mass doses are also a well-established approach to meeting vitamin A needs. To avoid duplicating efforts in mass dose vitamin A and in iodine fortification programs, the Bank should work with WHO and UNICEF, since they have long been involved in such programs. The relative importance of the three types of projects should depend on country circumstances. Certain components common in earlier projects (especially food production, water supply, and food technology) generally will not be needed in future nutrition projects, although special circumstances may at times dictate their inclusion. Most projects should include some form of nutrition communication and education, an area in which the Bank has developed considerable expertise. This can be used to encourage breast feeding and preparation of appropriate weaning foods and to promote food marketing programs. Nutrition anthropologists can work effectively in helping communities to design nutrition messages and set up education programs. Mass media can be used to complement these efforts and to reach less socially cohesive groups. Training, technical assistance, and other actions to help strengthen institutional capability to analyze, plan, and prepare, manage, and evaluate operations should commonly be included in projects. Evaluation is so important that no Bank-financed project should be without an evaluation component, but future projects should be operationally rather than research oriented as the first three projects were. The major criterion in PHN's consideration of nutrition project proposals should be the project's contribution to a soundly conceived nutrition plan. The country applying for assistance must have an expressed interest in and willingness to commit substantial resources to nutrition and an ability to reach or subsequently extend the project to large numbers of the nutritionally needy. In countries with little nutrition experience, emphasis should be on projects that lay the groundwork for more extensive interventions. Other departments of the Bank may be better equipped to take the lead in some nutrition projects, provided they are willing. Preschool programs like Brazil's PROAPE program aimed at broader developmental needs of young children at times may more appropriately be undertaken by the education divisions since the ministry of education is likely to be the - 80 - implementing agency and school facilities and staff will provide the backbone for the program. (The PROAPE concept is based in part on the notion that there has been, in the past 20 or 25 years, a dramatic growth in investment in primary school education in developing countries without a concomitant improvement in the academic performance of children and that the problem may not be with the quality of education but with the quality of students.) Agriculturally oriented nutrition interventions like those in Sergipe, Indonesia and Colombia would best be handled by agricultural divisions, unless there is some -demonstrable comparative advantage in having them carried out within the framework of a nutrition project. Nutrition Components in Other Projects Nutrition needs to be more systematically included and given greater emphasis in PHN's health projects and population projects. Most commonly this should consist of growth monitoring, nutrition education (concerning breast feeding, weaning foods, food hygiene), oral rehydration, and selective supplementary feeding and rehabilitation of seriously malnourished children. Micronutrient components also may be considered. The priority assigned to nutrition relative to other interventions must, of course, be assessed for each project. (The main question is what the absence of nutrition will do to the expected outcomes of the project, and the advantages of including nutrition.) The apparent cost-effectiveness of the nutrition package under most circumstances is such that it should be mandatory in all primary health care and MCH work. Many countries, particularly those in sub-Saharan Africa, are not yet capable of absorbing free-standing nutrition projects other than perhaps infrastructural projects aimed at increasing their capacity to use food aid. Opportunities should be sought to include preparatory work (for strengthening institutions and for studies leading to development of nutrition projects) as components in health and population projects and in technical assistance projects. Also, just as in free-standing projects discussed above, since many of these same countries are in need and recipients of food aid but often are limited by their logistical capacity to absorb the quantities needed, appropriate Bank projects could include components designed to help expand that capacity. Certain of the nutrition activities may be considered for inclusion as components of rural development, urban development, and education projects. They are most likely to be successful where there is a strong political interest in improving nutrition, where the components are sufficiently large to attract the attention of Bank staff, and where local managers are capable of implementing multisectoral projects. Experience suggests that nutrition components should only be considered for non-PHN projects if they are directly related to basic project objectives, can compensate for changes caused by the project that may be nutritionally negative, can be used to prepare for a future free-standing nutrition project, or can contribute to expanding programs already developed in the country. - 81 - Future Analytic Work Country economic and agricultural sector work should regularly include collection and analysis of data on food consumption. And continued emphasis should be given to nutrition sector work. Considerably more is now known about how to do nutrition sector work than three years ago when the Bank's nutrition efforts began to focus on it more heavily. (See Appendix F.) Results have been limited and performance mixed. But recent work in, for example, China, the Philippines, Zimbabwe, Bangladesh, Rwanda, and Zambia is of a quality that adds to country knowledge and contributes to economic and social policy. To inject this knowledge into policy dialogues and operational follow-ons will require stronger organizational links within the Bank and additional staff to insure that sector work findings are known and exploited. Nutrition sector work has too often been undertaken by staff outside the responsible operational division. As a result, there has not been the continuity of interest to insure that sector work findings lead-to appropriate next steps. Nutrition-related research conducted or sponsored by the Bank has reached the stage that it can in future be more sharply focused on specific problems and be more closely integrated with operational activities. Priority should be given to impact research which ideally should be linked to the monitoring and evaluation components of nutrition projects. (Experience shows that a great deal more is likely to be learned about project impact if qualified Bank staff or.consultants take part in evaluations.) In both performance of the project at hand and design of future projects, training and experience in project monitoring and evaluation should be provided. Second priority is for research directed at the nutritional effects of agricultural actions that are assumed to be nutritionally beneficial but may well not be.57/ (At a minimum, agriculture projects that contain evaluation components should, when feasible, include changes in food consumption as a measure of project performance.) Third is for research directed at determining the nutritional effects of current consumer food subsidy programs and the design of such programs to increase nutritional effectiveness. Most of the remaining research should relate to the consequences of nutritional deprivation (e.g., the extent to which malnutrition contributes to poor health), longer range effects of remedial actions (e.g., the extent to which an early-age supplementary feeding program, as in Tamil Nadu, affects subsequent educability and productivity), and the priority to assign nutrition interventions vis-a-vis other types of interventions that have the objective of improving health status. Organizing for Action Any implementation of action-oriented conclusions of this paper are highly contingent on how the Bank is organized for the effort and the 57/ Terence H. Martin, "Nutritional Consequences of Agricultural Development Projects, A Survey of Experience," report prepared for the World Bank (AGR and PHN), July 1983. - 82 - resources made available.58/ Nutrition does not fit neatly into the Bank's way of doing business. In principle, it should be possible to parcel out the elements of a nutrition program to different Bank departments to handle. But just as governments are structured along sectoral lines, the Bank is structured by sectoral departments whose priorities and rewards make it difficult to deal with issues outside their specialization. It is not by chance that agriculture and health staff have shied away from the nutrition problem, given their more immediate and more direct responsibilities. For PHN staff, there is limited attraction in involvement in sector work that may end up suggesting that the best investment opportunities lie in agriculture, and for agriculture staff the reverse is true. The perceived difficulties involved in solving nutrition problems have made them low priorities in both departments (and, in turn, in recruitment). Agriculture and health staff seem unlikely to address the problems independently. In PHN there is added pressure in the use of staff time, given senior management priority for population projects and country demand for health projects. Some way must be devised for nutrition to reach into the agriculture and health departments.without being swallowed by them. Actions to Be Taken 1. The evidence presented here should lead to more confidence that the malnutrition commonly identified in PHN's epidemiological analyses as a major contributing cause of.mortality and poor health status can be usefully addressed in the context of those health projects and population 58/ Nutrition was first established in the Bank's Population Projects Department. With increased understanding of the broader types of interventions required, and increased emphasis on productive elements within nutrition projects, nutrition was shifted to Agriculture and Rural Development (Central Projects) under the Assistant Director responsible for rural development projects. Late in 1980, with the establishment of a health lending program, most nutrition activities and six of nine positions were transferred to help form a critical mass of staff in the new Population, Health and Nutrition Department. The nutrition staff, which had formed a separate division in Agriculture, was integrated with health and population staff into geographic divisions in PHN. In the new department, which was preoccupied with developing a health program, the nutrition program lost the modest momentum that had been developed. The paper, "Toward an Operational Work Program," in December 1980 called for the establishment of a focal point for nutrition within each region and the development of a regional nutrition program. However, the staff designated for nutrition activities were made coordinators of overall population, health, and nutrition actions and nutrition programs as such were never developed. - 83 - projects that involve MCH and primary health care. Such projects should prima facie include nutrition components aimed at preventing and treating severe cases of malnutrition, particularly among mothers and children (e.g., through growth monitoring; nutrition education on breast-feeding, weaning foods, and food hygiene; diarrhea control; and selective supplementary feeding for rehabilitating seriously malnourished children). The normal expectation should be that these actions would be included and, when appropriate, improved nutrition should be one of the main thrusts of such projects. (The priority assigned to nutrition interventions relative to other intervention needs, of course, to be assessed separately for each project. The main questibn is what the absence of nutrition will do to the expected outcomes of the project, and the advantages of including nutrition.) 2. Because of the largely experimental nature and shaky start of self-standing nutrition projects, lending for nutrition has, in effect, been held in abeyance until results were in on the early projects. That these projects have made a contribution is now clear and new projects are likely to accomplish more as a result of the Bank's earlier investment in this learning-by-doing exercise. This already is markedly evident in Tamil Nadu, the first of the Bank's second-generation projects. There should now in appropriate countries be renewal and expansion of -lending for three types of free-standing nutrition projects -- food as medicine (Tamil Nadu-type projects); projects relating to food distribution, including- projects to improve efficiencies in the food marketing system; and micronutrient programs. In the context of developing projects, special attention 'should be given to encouraging borrowers to consider restructuring current expenditures for nutrition. 3. Many of the countries in greatest need, particularly most countries in sub-Saharan Africa, are ill equipped to undertake nutrition projects or major nutrition components in other projects. Special efforts should be made to develop their capacity to undertake nutrition work, through small components in health and population projects, and technical assistance projects. 4. Work on nutrition issues in economic and agriculture and health sector work, as in nutrition sector work per se, should be intensified. 5. The nutrition research program in the Bank should concentrate on operations and particularly on project impact. Because governments are less interested than the Bank in such work, direct Bank involvement in evaluation, including some funding, is necessary. Other priority topics are the nutritional effects of agricultural actions and of consumer food subsidies. A five-year nutrition research program should be drawn up. 6. Training sessions on nutrition should be available for Bank staff, especially in PHN and AGR. 7. Additional attention to nutrition should be considered for the programs of the Economic Development Institute, the first step being the October 1984 food policy seminar. Consideration should be given to a separate course on nutrition issues for public health planners. - 84 - 8. Finally, consideration should be given to having nutrition and food .security as the theme of a World Development Report in the late 1980s. (The 1982 report concentrated heavily on the production side of the food problem.) Through a WDR, the Bank's analytic skills can make a contribution to the understanding of the problem at a policy level. An atmosphere of uncertainty has clouded operational work in nutrition in recent years, in part because efforts to improve nutrition have not lived up to expectations. But this attitude appears to be changing, sparked by UNICEF's recent major thrust in this field, by PAHO's assignment of nutrition as a top priority, by the Bank's ongoing work on food security that underscores the importance of work in this area, and by the very encouraging evidence from recent evaluations of Bank nutrition projects, particularly Tamil Nadu. The nutrition projects did not do everything they were supposed to do -- in the first three projects the ambitiousness of the objectives obscured the achievements -- but on balance, they clearly have been worth doing. Substantial portions of nutrition projects have been successes -- countries have taken them up and enlarged them on their own, and so far three of the four have requested second projects. More important, these countries are now looking at nutrition in a different way. - As a development institution committed to the alleviation of poverty, the Bank has acknowledged a responsibility for dealing with the problems of malnutrition. Toward this end, much can be achieved through Bank projects aimed at increasing incomes. For many population groups, however, this route is too slow. The experimental nutrition project period, the first full operational project in Tamil Nadu, and the subsequent phase of concentrating on sector and other analytic work have all led to recognition that malnutrition is an effect of poverty that need not wait until incomes increase three or four times to overcome the problem. The task is difficult, but experience shows that the Bank's preliminary work in this area and evolutionary variants of this work can make a contribution. Thus far, very modest resources have been devoted to the attempt to solve the problems. If the Bank is to make a contribution of consequence, a moderate increase in resources will be required. In the early-Board discussions on nutrition, subsequently verified by research, it was recognized that "attention to food production of itself is insufficient to satisfy food needs." Approximately 25 percent of the Bank's lending and staff is devoted to food production; attention should now be given to assuring that food is used efficiently to meet hunger needs and to take related actions aimed at alleviating malnutrition. APPENDIX A Background Papers to this Report 1 Ahmad, Nazir "Annotated Bibliography of Nutrition-Related Research by the World Bank" Beaton, George "Food and Nutrient Distribution Programs: Where Are We Now?" Chester, Lauren "Nutrition Components" Ho, T. J. "Economic Issues: Cost, Affordability and Cost Effectiveness" Knight, Peter "Brazil Nutrition Project: A Preliminary Evaluation and Lessons for Future Operational Work in Nutrition in Brazil", May 17, 1982. Leslie, Joanne "A Review of the World Bank Nutrition Projects in Brazil, Indonesia, Colombia and India." Measham, Anthony "Colombia Nutrition Project", December 28, 1982. Muscat, Robert "Review of Nutrition Economic and Sector Work" Pines, James "Monitoring and Evaluation" and "Organization and Management Requirements" Pinstrup-Andersen, Per "A Review of Nutrition-Related Research Undertaken or Supported by the World Bank" Soemardjan, Selo, "Evaluation of the Indonesian Nutrition et. al. Development Project", Volumes 1 and 2, Report of the Evaluation Team to the Ministry of Health, Republic of Indonesia, December 1982. I/ All are PHN papers, unless otherwise noted. APPENDIX B Health and Nutrition Indicators in Bank Nutrition Project Countries India Indicators (Tamil Nadu) Indonesia Colombia Brazil Per capita GNP, 1980 (dollars) 240a/ 430 1,180 2,050 Government expenditures on health 1979 (dollars) per capita 2a/ 1 28 13 Percent of all expendi- tures, 1979 1.7a/ NA NA 8.5 Daily energy supply Calories per capita, 1977 2,036b/ 2,272 2,364 2,562 Percent of full requirement 85b/ 102 98 111 Infant mortality rate, 1980c/ 120d/ 93 64 77 Child. death rate, 1980e/ 17f/ 11 4 7 Sources: World Development Report, 1982; World Bank Health Policy Sector Paper, 1980; World Bank Government Finance Statistics Yearbook 1982; and Tamil Nadu Integrated Nutrition Project Proposal, prepared by the Government of Tamil Nadu, 1978. N.A. Not available. a/ India-wide. !;/ 1974 data for Tamil Nadu. c/ Number of deaths per thousand infants up to one year old. d/ 1978 data for Tamil Nadu. e/ Number of deaths per thousand children one to four years old. Y/ India-wide figure; Tamil Nadu figure for one to five year olds in 1978 was 9.4. APPENDIX C Page 1 of 12 BANK NUTRITION RESEARCH 1. In recent years, the World Bank has undertaken or supported 55 pieces of nutrition-related research.1/ Its goal has been to improve understanding of the nature of the nutrition problem and how policies and projects can deal with it, and to add to the general stock of knowledge regarding the impact of policies, programs, and projects on nutrition and the effect of nutritional improvements on human capital and economic growth. Most of the research has been socioeconomic, covering many health, food, agricultural, and broader economic policies and projects in which malnutrition (as outcome or cause) has had an explicit and major role.2/ Nature and Magnitude of Malnutrition 2. A pioneering attempt to disaggregate food intake by income group on a global scale estimated that 56 percent of the population of developing countries had inadequate calorie consumption.3/ In an expansion of the analysis which examined the incidence of malnutrition among children it was, as expected, found to be higher among young children than for the rest of the population -- in Latin America, 27.4 percent among individuals above the age of ten and 55.4 percent among children below that age,.4/ An in-depth analysis for urban India showed similar results. 3. Another study in this series, using an improved estimation method, concluded that in 40 developing countries, 65 percent of the 1 An annotated bibliography is available of all research cited in this section. 2/ Areas excluded from this review are food production, buffer stocks, and other food security factors (see Shlomo Reutlinger, "World Bank Research on the Hunger Dimension of the Food Problem," Research News, World Bank. Winter 1981/1982, pp. 3-9, for a review of Bank research in this area). Poverty and basic needs, general food price policies, population, and food demand and supply are also excluded, as are studies undertaken as part of the four nutrition projects described in this paper. Bank sponsored books on nutrition that go beyond research also are omitted (e.g., James E. Austin, Confronting Urban Malnutrition: The Design of Nutrition Programs, World Bank Staff Occasional Paper No. 28, Baltimore and London: Johns Hopkins University Press, 1980 and C. Peter Timmer, Walter P. Falcon and Scott R. Pearson, Food Policy Analysis, Published for the World. Bank, Baltimore and London: The Johns Hopkins University Press, 1983. 3/ Shlomo Reutlinger and Marcelo Selowsky, "Malnutrition and Poverty: Magnitude and Policy Options, World Bank Staff Occasional Paper No. 23, Baltimore and London: Johns Hopkins University Press, 1976. 4/ Marcelo Selowsky, "Balancing Trickle Down and Basic Needs Strategies: Income Distribution Issues on Large Middle-Income Countries, with Special Reference to Latin America." World Bank Staff Working Paper No. 335, June 1979. APPENDIX C Page 2 of 12 populations consumed less than recommended requirements.5/ These studies suggest that in the normal course of development, even with a substantial expansion of food production, the nutrition problem is unlikely to be solved; special measures are needed to improve incomes of the poor, including intervention in food markets designed to fill the nutritional needs of the targeted population. One study estimated that with no change in income distribution or in real food prices, and assuming a continuation of past rates of economic growth, it would take about ten years to raise 5 percent of the population above malnutrition levels in Pakistan, 45 years in Morocco, and 25 years in Sri Lanka and would be nearly impossible in Bangladesh.6/ 4. These and other Bank studies point out that while expanded food production is essential to eliminate malnutrition, it is not sufficient.7/ Policy should focus on the ability of the poor to meet their food and other basic needs. Expanded food production may help to lower food prices for the urban poor, and generate income among the rural poor, but other means must also be found to boost real income. 5. A number of efforts to collect information in other areas are of direct utility in the design of nutrition projects and policies. They include household consumption and expenditute surveys in Brazil and Indonesia; similar surveys for the cities of Bogota and Cali, Colombia, for 5/ Shlomo Reutlinger and Harold Alderman, "The Prevalence of Calorie-Deficient Diets in Developing Countries," World Bank Staff Working Paper No. 374, March 1980. 6/ Odin Knudsen and Pasquale Scandizzo, "Nutrition and Food Needs in Developing Countries." World Bank Staff Working Paper No. 328, May 1979. 7/ See Alan Berg, Malnourished People: A Policy View, Poverty and Basic Needs Series, World Bank, June 1981 and Shlomo Reutlinger, "Malnutrition: A Poverty or a Food Problem?" World Development Vol. 5, August 1977 (World Bank Reprint Series No. 47). APPENDIX C Page 3 of 12 the Terai region of Nepal, and for East Java; and food balance sheets for China.8/ 6. The Bank has also supported research to consider other methods for estimating the magnitude of the nutrition problem and for analyses based on those approaches.9/ This work, following non-Bank work by Sukhatme and FAO10/, has led to more conservative estimates among some observers of the magnitude of existing energy deficits among developing-country populations. It highlights the need to consider energy deficits and efforts aimed at their reduction within the broader context of health, sanitation, and poverty. The Economic Value of Improved Nutrition 7. Malnutrition affects labor productivity, motivation, the activity levels and learning capacity of children, and thus both the well-being and earning capacity of the poor. It is important to show what the effects of nutritional improvements are if economic growth is to be used to justify projects and policies aimed at nutritional improvements. 8. World Bank studies on the impact of iron-deficiency anemia on labor productivity in Indonesia have made a major contribution in this area. The productivity of anemic workers who were given iron supplements for two months was found to have increased 15-25 percent, depending on the 8/ Peter T. Knight, Dennis Mahar and Ricardo Moran, "Health, Nutrition and Education," In Brazil: Human Resources Special Report, annex 3, Report No. 2604-BR, World Bank, July 1979; Dov Chernichovsky and Oey Astra Meesook, "Patterns of Food Consumption and Nutrition in Indonesia," World Bank Staff Working Paper No. 670, September 1984; Rakesh Mohan, Jorge Garcia, and M. Wilhelm Wagner, "Measuring Urban Malnutrition and Poverty: A Case Study of Bogota and Cali, Colombia," World Bank Staff Working Paper No. 447, April 1981; and Joanne Leslie, Reynaldo Martorell, and Peter R. Moock, "Characteristics and Determinants of Child Nutritional Status in Nepal," Discussion Paper No. 82-15, Population and Human Resources Division, December 1982; T. J. Ho, report on nutrition in East Java, PHN, forthcoming; Dean T. Jamison, Teresa J. Ho, and F. L. Trowbridge, "Food Availability and the Nutritional Status of Children in China," Discussion paper No. 81-26, Population and Human Resources Division,October 1981; and Alan L. Piazza, "Trends in Food and Nutrient Availability in China, 1950-81," World Bank Staff Working Paper No. 607, September 1983. 9/ T. N. Srinivasan, "Malnutrition: Some Measurement and Policy Issues," World Bank Staff Working Paper No. 373, February 1980; Michael Lipton, "Poverty, Undernutrition, and Hunger," World Bank Staff Working Paper No. 597, April 1983). 10/ P. V. Sukhatme, "Malnutrition and Poverty," Ninth Lal Bahadur Shastri Memorial Lecture, Indian Agricultural Research Institute, New Delhi, January-27, 1977; and FAO, The Fourth World Food Survey, FAO Nutrition Report Series No. 10, 1977, and Agriculture: Toward 2000, FAO Conference Report e79/24, 1979. APPENDIX C Page 4 of 12 type of work performed.11/ The increase in productivity was estimated- to be worth $260 for each dollar invested in iron supplementation. These economic findings led to the inclusion of efforts to reduce iron deficiencies in the Bank's nutrition loan to Indonesia. 9. A second set of studies focused on the productivity effects of food and iron supplementation for workers in Kenya found that dietary deficiencies are an important cause of anemia; that anemia is an important determinant of worker productivity; and that provision of extra iron and treatment of hookworm are inexpensive and feasible interventions. Food supplementation increased road workers' productivity by 13 percent. Similar studies in India found no relationship between road workers' current food intakes and output but found a positive relationship between output and good nutrition status, as reflected by weight for height and arm circumference.12/ There are indications in studies now under way that benefits from reduction of iron deficiency anemia exceed costs in some cases by wide margins.13! Improved Nutrition and Human Capital Formation 10. Studies of how infant and child malnutrition affects the future earning capacity of children and their contribution to economic growth have focused on the impact of school attendance and performance. Children in the Terai region of Nepal whose height was less than 70 percent of the norm were found, after controlling for income differences, to be 14 times less likely to be enrolled in school than children whose*height was 85 percent of the norm and 228 times less so than children whose height was equal to the norm.14/ A similar relationship was found for weight, where the probability increased from 4 to 13 times less. The relationship between 11/ Samir S. Basta and A. Churchill, "Iron Deficiency Anemia and the Productivity of Adult Males in Indonesia," World Bank Staff Working Paper No. 175, April 1974). Also see Samir S. Basta, Darwin Karyadi Soekirman, and Nevin S. Scrimshaw, "Iron Deficiency Anemia and the Productivity of Adult Males in Indonesia," American Journal of Clinical Nutrition, April 1979, pp. 916-25. 12/ Michael C. Latham and Lani S. Stephenson, "Kenya: Health, Nutrition and Worker Productivity Studies," Reports to the World Bank, January 1981 and Michael C. Latham, and Lani S. Stephenson, "Costs, Prevalence and Approaches for Control of Ascaris Infection in Kenya," Journal of Tropical Pediatrics, No. 26, December 1980), pp. 246-264. Also see "Effect of Health and Nutrition Status of Road Construction Workers in Northern India on Productivity," World Bank Technical Memorandum No. 4, 1975. 13/ Henry M. Levin, "A Benefit-Cost Analaysis of Anemia Reducation," PHN, forthcoming. Also see-Susan Horton and Timothy King," Labor Productivity: Un Tour d'Horizon," World Bank Staff Working Paper No. 497, October 1981, which concludes that "the evidence that iron status affects productivity in heavy manual labor...seems convincing." 14/ Joanne Leslie and Peter R. Moock, "Childhood Malnutrition and Schooling in the Terai Region of Nepal," Discussion Paper No. 82-14, Population and Human Resources Division, December 1982. APPENDIX C Page 5 of 12 height and weight and school enrollment was much stronger for boys than girls. These findings are consistent with those on data collected by Bank staff from three provinces in China where low height-for-age was consistently found to be related inversely with performance, as measured by the number of grades a child was behind his age group.15! 11. Analysis of studies in Brazil showed a 15 percent advantage in school performance by those who had been part of a preschool nutrition program.16/ Analysis of data from a study in Colombia shows significant effects of both nutritional supplementation and home education for mothers on children's growth, diet, and school performance.17! 12. A study of the impact of infant and childhood feeding programs on earning capacity after children have grown up, comparing a child's relative weight and IQ to the IQ of adults and their incomes, concluded that the income gains generated by improvements in infant and child nutrition are likely to be large relative to the cost of the improvements.18/ Further work suggests how countries might arrive at estimates of economically sound levels of investment in improving the "quality" of children through nutrition, schooling, and related projects.19/ Food Policies and Malnutrition Problems 13. Bank research adds significantly to the growing literature on how subsidized food prices and food distribution may reduce malnutrition without negative effects on food production. Subsidized food prices and distribution policies in Sri Lanka, India, Pakistan, Bangladesh, and Indonesia were found to have eliminated between 16 and 59 percent of the calorie gaps -- the quantities required to permit an adequate intake for 15/ D. T. Jamison, "Child Malnutrition and School Retardation in China," Discussion Paper No. 81-27, Population and Human Resources Division, September 1981. 16/ Benicio D'Aquino and others, "Avaliacao Antropometrica de Eficacia da Suplementacao do Pre-Escolar," Revista De Saude Publica, Vol. 15, December 1981. 17/ M. G. Herrera and C. M. Super, "School Performance and Physical Growth of Underprivileged Children: Results of the Bogota Project at Seven Years," Report to the World Bank, June 1983. 18/ Marcelo Selowsky and Lance Taylor, "The Economics of Malnourished Children: An Example of Disinvestment in Human Capital," Economic Development and Cultural Change, Vol. 22, No. 1, October 1973, pp. 17-30. 19/ Marcelo Selowsky, "Nutrition, Health and Education: The Economic Significance of Complementarities at Early Age," Journal of Development Economics, Vol. 9, 1981, pp. 331-46. See also Marcelo Selowsky, "Target Group Oriented Food Programs: Cost Effectiveness Comparisons," American Journal of Agricultural Economics, Vol. 61, December 1979, pp. 988-94 (World Bank Reprint Series No. 127). APPENDIX C Page 6 of 12 everyone -- in those countries.20/ The economic cost of programs was divided between the agricultural sector and government, with concessional food aid greatly reducing actual fiscal costs in some countries. The share borne by the agricultural sector varied, with Indian farmers carrying a large share of costs and farmers in Sri Lanka a relatively small share. 14. A study of a ration-shop scheme in India concluded that "in spite of the fact that it tends to distort prices from the competitive market values, the complex tradeoffs in the present system ultimately result in- positive net social benefits for the country, if it is assumed that there is reasonable social concern for minimum food consumption of the poor."21/ The authors advocated expansion of the system if more stringent and effective eligibility requirements could be imposed and that ration shops be set up in rural as well as urban areas. 15. The Sri Lanka rice ration program, according to one study, was directly and dramatically linked to food and nutritional adequacy and low death rate.22/ A subsequent examination of Sri Lanka's food stamp program concluded that maintaining a constant stamp value in the face of food price increases would reduce the real value of the stamp and result in reduced food consumption and deterioration of the nutritional situation; stamp values, the author proposed, should be fully indexed to price increases and the first group to benefit should be households in the bottom 20 percent of the income distribution.23/ An analysis of food policy in Zambia, Bangladesh, and India concluded that public interventions "are frequently of benefit only to urban groups (but often including the poorest among them)." The.study also concluded that "there is*scope for manipulating the supply and prices of basic food staples to produce cheap calories for the poor."24/ 16. A review of the effects of agricultural policies and projects on nutrition concluded that "the evidence makes a convincing case that some, perhaps many, agricultural projects have had adverse nutritional outcomes." Without quantification of the results, however, it is difficult to show how important those effects are or how significant one cause may be to 20/ Pasquale Scandizzo and Judith Graves, "The Alleviation of Malnutrition: Impact and Cost-Effectiveness of Official Programs," World Bank AGREP Division Working'Paper No. 19, January 1981. 21/ Pasquale Scandizzo and Gurushri Swamy, "Benefits and Costs of Food Distribution Policies: The Indian Case," World Bank Staff Working Paper No. 509, 1982. 22/ Paul Isenman, "Basic Needs: The Case of Sri Lanka," World Development, Vol. 8, March 1980, pp. 237-58. 23/ Marcelo Selowsky, "Food Prices and the Indexing of the Food Stamp Program in Sri Lanka: An Evaluation of the Trade-offs," mimeo, April 1980. 24/ Edward Clay and others, "Food Policy Issues in Low-Income Countries," World Bank Staff Working Paper No. 473, August 1980. APPENDIX C Page 7 of 12 another.25/ Clearly, additional research is needed in this area. The effects of the dairy development scheme in two Indian states on food - consumption and nutrition are now being studied.26! 17. A survey of efforts to incorporate nutrition concerns in the preparation of agricultural projects concluded that although a number of agencies show an interest, "very little has yet been accomplished." The study recommends a strategy for such incorporation.27! Nutritional Impact and Cost-Effectiveness of Nutrition Programs 18. Much of the Bank research emphasizes the interaction of health and nutrition. Research on parasitic infections accompanied studies of iron-deficiency anemia in Indonesia and Kenya. One study found that heavy ascaris infections caused as much as 25 percent of ingested calories to be lost, greatly contributing to malnutrition and reducing the effectiveness of food supplementation programs. The benefits of deworming undertaken twice a year, the study concluded, could be as high as ten times the cost.28/ 19. An analysis of data from the Indian Narangwal project found that there was a strong relation between food intake and anthropometric status and that nutritional care reduced perinatal mortality and mortality among children below two years of age. This study measures the costs of averting death at different early ages with nutrition services, of averting a day of illness, of gaining an extra centimeter in growth, and of increasing a percentage point in psychomotor scores.29! 20. An examination of ten field projects that combined nutrition and health services concluded that such programs in the hands of able administrators and in populations as large as 60,000-70,000 can reduce infant and child mortality by a third to a half within one to five years, at a cost of less than the equivalent of 2 percent of per capita incomes, an amount no greater than the current cost of health care. The authors recommended expanding "beyond the initial carefully controlled field 25/ Terence H. Martin, "Nutritional Consequences of Agricultural Development Projects, A Survey of the Experience," report prepared for the World Bank, (AGR and-PHN) July 1983. 26/ Odin Knudsen, Roger Slade and Per Pinstrup-Andersen, forthcoming. 27/ Per Pinstrup-Andersen, "Nutritional Consequences of Agricultural Projects: Conceptual Relationships and Assessment Approaches," World Bank Staff Working Paper No. 456, May 1981. Also see Pasquale Scandizzo, "Analyzing Nutrition Effects of Agricultural Projects," AGR Interim Guidance Note No. 7, July 1981. 28/ L. Latham, M. Latham and S. S. Basta, "The Nutritional and Economic Implications of Ascaris Infection in Kenya," World Bank Staff Working Paper No. 271, September 1977. 29/ Arnfried Keilman, Carl E. Taylor, Cecile De Sweemer, and others, Child and Maternal Health Services in Rural India: The Narangwal Experiment, Volume 1, Integrated Nutrition and Health Care, A World Bank Research Publication, Baltimore and London: The Johns Hopkins University Press, 1983. APPENDIX C Page 7 of 12 another.25/ Clearly, additional research is needed in this area. The effects of the dairy development scheme in two Indian states on food - consumption and nutrition are now being studied.26! 17. A survey of efforts to incorporate nutrition concerns in the preparation of agricultural projects concluded that although a number of agencies show an interest, "very little has yet been accomplished." The study recommends a strategy for such incorporation.27! Nutritional Impact and Cost-Effectiveness of Nutrition Programs 18. Much of the Bank research emphasizes the interaction of health and nutrition. Research on parasitic infections accompanied studies of iron-deficiency anemia in Indonesia and Kenya. One study found that heavy ascaris infections caused as much as 25 percent of ingested calories to be lost, greatly contributing to malnutrition and reducing the effectiveness of food supplementation programs. The benefits of deworming undertaken twice a year, the study concluded, could be as high as ten times the cost.28/ 19. An analysis of data from the Indian Narangwal project found that there was a strong relation between food intake and anthropometric status and that nutritional care reduced perinatal mortality and mortality among children below two years of age. This study measures the costs of averting death at different early ages with nutrition services, of averting a day of illness, of gaining an extra centimeter in growth, and of increasing a percentage point in psychomotor scores.29! 20. An examination of ten field projects that combined nutrition and health services concluded that such programs in the hands of able administrators and in populations as large as 60,000-70,000 can reduce infant and child mortality by a third to a half within one to five years, at a cost of less than the equivalent of 2 percent of per capita incomes, an amount no greater than the current cost of health care. The authors recommended expanding "beyond the initial carefully controlled field 25/ Terence H. Martin, "Nutritional Consequences of Agricultural Development Projects, A Survey of the Experience," report prepared for the World Bank, (AGR and PHN) July 1983. 26/ Odin Knudsen, Roger Slade and Per Pinstrup-Andersen, forthcoming. 27/ Per Pinstrup-Andersen, "Nutritional Consequences of Agricultural Projects: Conceptual Relationships and Assessment Approaches," World Bank Staff Working Paper No. 456, May 1981. Also see Pasquale Scandizzo, "Analyzing Nutrition Effects of Agricultural Projects," AGR Interim Guidance Note No. 7, July 1981. 28/ L. Latham, M. Latham and S. S. Basta, "The Nutritional and Economic Implications of Ascaris Infection in Kenya," World Bank Staff Working Paper No.-271, September 1977. 29/ Carl E. Taylor, Arnfried Keilman, Cecile de Sweemer, and others, Malnutrition, Infection, Growth and Development: The Narangwal Experience, in press. APPENDIX C Page 8 of 12 experiments ... to serve significantly larger numbers of people."30/ 21. A study of the economics of supplementary feeding concluded that "supplementary feeding, if it results in at least a 10 percent improvement in mortality rates and a relatively modest 10 percent increase in productivity, is economically justifiable."31/ The study includes recommendations to keep leakages low -- the ration should be near the complete nutrition requirements of the target children; they should be fed on site, and for only two to three months. Those were features of the Tamil Nadu nutrition project whose economic and social benefits were estimated to range between 10 and 22 percent. 22. A study of the Chilean supplementary food program found that calorie and protein intakes by young children (5-23 months of age) were greater than what would be expected in an income-transfer program of the same size.32/ This study provides valuable empirical evidence on the unresolved nutrition question of whether the.income should be transferred as food or cash. Choice and Design of Policies and Projects 23. Work also has gone on with a prospective rather than retrospective orientation. Reutlinger and Selowsky in estimating the fiscal-cost of achieving a given consumption increase in various food subsidy programs concluded that certain "target group oriented food programs in urban areas and programs to assist low-income farm families to increase and stabilize production of food for their own consumption can be more cost-effective than outright income distribution."33/ 24. A follow-up study examined a number of programs aimed at increasing the number of calories consumed by children in malnourished households. The author concluded that in most of the programs the effects were "no larger than an equivalent income transfer."34/ Cash transfer programs, however, are difficult politically to introduce and maintain. They are also likely to be difficult to control, and a higher proprotion of the transfer is likely to miss the target group. 30/ Davidson R. Gwatkin, Janet R. Wilcox and Joe D. Wray, "Can Nutrition Efforts Make a Difference?" in Malnourished People: A Policy View. 3i/ Odin K. Knudsen, "Economics of Supplemental Feeding of Malnourished Children: A Case of Leakages, Benefits and Costs," World Bank Staff Working Paper No. 451, April 1981. 32/ Lloyd Harbert and Pasquale L. Scandizzo, "Food Distribution and Nutrition Intervention: The Case of Chile," World Bank Staff Working Paper No. 512, May 1982. 33/ Reutlinger and Selowsky,"Malnutrition and Poverty." 34/ Selowsky, "The Economic Dimensions of Malnutrition in Young Children." The findings of Harbert and Scandizzo, "The Case of Chile," challenge this idea. Further research in at least six countries on this matter is currently under way outside the Bank. APPENDIX C Page 9 of 12 25. A study of the cost of eliminating energy gaps in India, Senegal, and Colombia over a period ^f 20 years showed that specific subsidies were less costly than general subsidies. Use of locally produced staples traditionally consumed by the poor helped to reduce subsidy costs and to concentrate subsidies on the poor.35! 26. A 1979 study introduced the idea of a social demand function to take account of any difference between the market and the social value of a particular good in benefit-cost estimation.36/ (This carries forward efforts in the Bank to develop methodologies for social project analysis.37/) The social demand function was used in an empirical analysis in 1980,38/ and in quantifyini the benefits of a food supplementation scheme for Tamil Nadu. 9/ A 1981 study developed a conceptual framework for estimating nutritional effects of agricultural and 35/ Dennis H. Wood and Christopher J. Roesel, "Analysis of Government Costs of Selected Strategies to Achieve Minimum Calorie Standards for a Defined Target Group in Three Developing Countries," Background paper prepared by Devres, Inc. for The World Development Report, May 10, 1982. 36/ Marcelo Selowsky, "Balancing Trickle Down and Basic Needs Strategies: Income Distribution Issues on Large Middle-Income Countries, with Special Reference to Latin America," World Bank Staff Working Paper No. 335, June 1979. 37/ Lyn Squire and Herman G. van der Tak, Economic Analysis of Projects, A World Bank Research Publication, Baltimore and London: Johns Hopkins University Press, 1975. 38/ Pasquale L. Scandizzo and Odin K. Knudsen, "The Evaluation of the Benefits of Basic Need Policies," American Journal of Agricultural Economics, February 1980, pp. 46-57. 39/ Knudsen, "Economics of Supplemental Feeding." APPENDIX C Page 10 of 12 rural development projects and policies.40/ It suggested that: where a trade-off exists between nutritional and other economic benefits of the projects, modifications be made to meet nutritional goals only if the cost is less than the cost of achieving a given nutritional improvement by means of the most cost-effective alternative. This approach to analyzing marginal cost-effectiveness to determine the extent to which nutrition concerns should be incorporated in the design of agricultural projects is elaborated in a 1983 study.41 27. A study of food consumption in Indonesia used data from a national household expenditure survey to estimate the elasticity of demand for food, with respect to income and price.42/ Another study went beyond conventional approaches to estimate demand functions for calories for a set of developing countries on the basis of the theory of characteristic demand.43/ 28. To help.resolve the issue of how to measure nutrition effects, a review of various types of indicators that might be appropriate for household surveys in developing countries discussed advantages and limitations of each.44/ One Bank staff member has developed a tape to measure arm circumference that is inexpensive and increases the reliability of this kind of measure of nutritional status from 35 percent to 85-96 40/ Pinstrup-Andersen, "Nutritional Consequences of Agricultural Projects." 41/ Shlomo Reutlinger, "Nutritional Impact of Agricultural Projects," Prepared for ACC/SCN Workshop on Nutrition in Agriculture and Rural Development Projects, Castelgandolfo, Italy, February 1983. 42/ Chernichovsky and Meesook, "Patterns of Food Consumption." 43/ Odin K. Knudsen and Pasquale L. Scandizzo, "The Demand for Calories in Developing Countries," American Journal of Agricultural Economics, Vol. 64, February 1982, pp. 80-86. 44/ Reynaldo Martorell, "Nutrition and Health Status Indicators: Suggestions for Surveys of the Standard of Living in Developing Countries," Living Standards Measurement Study, Working Paper No. 13, World Bank, February 1982. APPENDIX C Page 11 of 12 percent (depending on the age group).45/ The tape has been used in Bank-assisted development projects in Paraguay, Ecuador, and Bolivia and in the health project in Peru. Several studies of household behavior provide useful guidelines for policy and project design because they assist in anticipating how households will react to changes in food prices, incomes, demand for women's time, and the like.46/ An Assessment of the Research Contribution 29. In addition to its own research, the Bank has helped in funding nutrition research at institutions such as the International Food Policy Research Institute. Bank staff have participated in such projects as the National Academy of Science's world food and nutrition study4 / and food aid study.48/ The Bank has played an active role in organizing an interagency collaborative research program for the Sub-Committee on Nutrition of the U.N. system and for this program is sponsoring papers on nutrition education, nutrition and child survival, and nutrition effects of policies leading to increasing marketable surpluses. 30. Bank research has provided much that is useful for operational activities: information on the nature and causes of malnutrition, innovation in estimating social benefits and costs, insight into the characteristics of food supplementation projects, understanding of the design and targeting of consumer-oriented food-price subsidies, guidelines for integrating health and nutrition objectives in project design, and criteria for determining nutrition's influence in the choice and design of agricultural projects. 4/ Oscar Echeverri, forthcoming. 46, Howard N. Barnum and Lyn Squire, A Model of an Agricultural Household, World Bank Staff Occasional Paper No. 27, Baltimore and London: Johns Hopkins University Press, 1979; Dov Chernichovsky, "An Economic Theory of the Household and Impact Measurement of Nutrition and Related Health Programs," in Robert E. Klein, ed. Evaluating the Impact of Nutrition and Health Programs, New York: Plenum Publishing Corp., 1979; and Inderjit Singh and Lyn Squire, "A Model of the Agricultural Household: Some Implications for Nutrition Policies in Rural Areas," paper presented at conference on the Economics of Nutrition-Oriented Food Policies and Program, Bellagio, Italy, September 1978. 47/ Alan Berg, C. Peter Timmer, et.al., Nutrition, in National Research Council, Commission on International Relations, World Food and Nutrition Study: Supporting Papers, Vol. 4, Washington, D.C.: National Academy of Sciences, 1977); also published by the U.S. Senate Committee on Nutrition, 95th Congress, First Session, U.S. Government Printing Office, 1977. 48/ Shlomo Reutlinger, "Nutritional Cost-Effectiveness Considerations," in National Research Council, Committee on International Programs, Nutritional Analysis of Public Law 480 Title II Commodities, Washington, D.C.: National Academy Press, 1982, pp. 54-107. Also see Reutlinger's "Project Food Aid and Equitable Growth: Income Transfer Efficiency First!" Discussion Paper No. ARU-13, Agriculture and Rural Development Department, August 1983. APPENDIX C Page 12 of 12 31. While some of the Bank research is useful primarily for Bank decision making, the bulk -of the research findings adds to the general stock of knowledge. Bank researchers have made particularly important contributions in developing ways to gauge the magnitude of energy deficiencies on the basis of average food availability data; the nutrition effects of price policies and agricultural projects; the impact of infant and child nutrition interventions on school performance and future earning capacity; the impact of adult nutrition interventions on labor productivity and the impact of labor productivity on food need; and the social benefits of nutrition interventions. - 32. The Bank has been catalytic in promoting outside research that is relevant to Bank needs and has in turn been influenced by the scientific community in setting the priorities for Bank research. The benefits of close interaction with the outside research community are the result primarily of the Bank's maintenance of an active research program and highly competent research staff. Although it has not been possible to quantify the costs involved, the impression of those responsible for managing the Bank's research is that the volume of nutrition research has been high for the money spent and the work highly productive for the cost, as compared to other areas. APPENDIX D Page I of 4 COST-EFFECrIVENESS TABLES Table D-1A. Changes in Distribution by Nutritional Status (Weight-for-Age) of Children 7-60 Montbs old in Tamil Nadu, Pilot. and Control Blocks Weight-for- Pilot (with project) Control (wLth project) age statusal Baseline Outcome Change Baseline Outcone Change Normal 16.0 13.4 - 2.6 26.2 20.6 - 5.6 I grade 30.6 39.0 + 8.4 33.0 33.0 0 II grade 36.1 34.2 - 1.9 25.9 30.7 + 4.8 III & IV grade 17.3 13.4 - 3.9 14.8 15.6 + 0.8 Number in sample 526 544 587 436 a/ Based on standards set by the Indian Council of Medical Research. "Normal is above 90 percent of standard; Grade I, 80 to 90 percent; Grade II, 70 to 80 percent; Grade III, 60 to 70 percent; and Grade IV, below 60 percent Source: Tamil Nadu pilot block evaluation report. Table D-1B. Estimated Reduction in Numbers of Children Malnourished in Tamil Nadua/ Initial distribution Final distributioti Difference With project Without project due to Weight-for- Percentb/ Number Percentc/ Number Percentd/ Number project age status (thousands) (thousands) (thousands) (thousands) Normal 21.1 1,025 18.5 898 15.5 753 + 145 I grade 31.8 1,544 40.2 1,952 31.8 1,544 + 408 II grade 31.0 1,505 29.1 1,413 35.8 1,738 - 325 III & IV grade 16.1 782 12.2 592. 16.9 821 - 229 a/ Number of children 7-60 months in Tamil Nadu = 4,855,500, assuming they constitute 15 percent of the total Tamil Nadu population in rural areas of 32.37 million. b/ Assuming proportion of children in each nutritional status grade is the average of observed populations in pilot and control villages in evaluation study. c/. Assuming changes observed in pilot block. d/ Assuming changes observed in control block. APPENDIX D Page 2 of 4 Table D-2A. Changes in Distribution by Nutritional Status (Weight-for-Age) of Children 7-36 Months old in Tamil Nadu, Pilot and Control Blocks Weight-for- Pilot (with project) Control (with project) age status Baseline Outcome Change Baseline Outcone Change Normal 18.5 13.3 - 5.2 23.6 23.1 - 0.5 I grade 27.9 32.1 + 4.2 33.7 29.9 - 3.8 II grade 34.8 35.5 + 0.7 26.2 27.7 + 1.5 III & IV grade 18.8 19.1 + 0.3 16.4 19.3 + 2.9 Number in sample 276 293 347 264 Source: Tamll Nadu pilot block evaluation report. Table D-2B. Estimated Reduction in Nunber of Deaths Due to Project in Tanil Nadua/ Change in Initial number of distribution Final distribution Difference due to project deaths per With project Without -roject Unadjusted for Adjusted for Risk of year due to Weight-for- Percentb/ Itber Percentc Number Percentd/ Nunber nonproject factors nonproject factors deathe/ project age status (thousands) (thousands) (thousands) (thousands) (thousands) (thousands) Normal 21.0 612 15.8 460 20.5 597 - 137 0 0.006 0.01 I grade 30.8 897 35.0 1,020 27.0 787 + 233 + 180 0.010 + 1.800 l grade 30.5 889 31.2 909 32.0 932 - 23 -76 0.036 - 2.736 III & IV grade 17.6 513 17.9 521 20.5 597 - 76 - 107 0.107 -11.449 -12.385 a/ Nuber of children 7-36 nonths in Tanil Nadu = 2,913,300, assuming they constitute 9 percent of total Tanil Nadu population in rural areas of 32.37 million. b/ Assuming proportion of children in each nutritional status grade is the average of observed populations in pilot and control villages in evaluation study. c/ Assuming changes observed in pilot block. d/ Assuming changes observed in control block. e/ Risk of dying within a year for children aged 6-36 nonths, by nutritional status at beginning of year. Estimated from A.A. Kielmann and - - .- - - --- " I - n.a IA I Q7k rn. I747-50. APPENDIX D Page 3 of 4 Table D-3. Measures of Cost Effectiveness of Tamil Nadu Nutrition Project, Nutrition Delivery Component Cost Measures 1982 1985 (projected, at 1982 prices) Total annual cost of component (statewide coverage)a/ Nonrecurrent ,costs annualized over four-year projet period $ 1.37 million $ 1.37 million Annual recurrent costs 26.62 million 16.98 million Total annual cost $ 27.99 million $18.35 million Nutrition status Cost per child removed from malnourishmentb/ $ 193 $ 126 Cost per child removed from moderate and severe malnourishmentc/ $ 51 $ 33 Cost per child removed from severe malnourishmentd/ $ 122 $ 80 Child survival Cost per death avertede/ $ 2,260 $ 1,482 a/ From main report, table 3. Statewide costs are 373 times blockwide costs (373 blocks in Tamil Nadu). b/ Reduction in number of children malnourished (Ist-IVth degree) due to project = 146,000; from table 1. c/ Reduction in number of children moderately and severely malnourished (IInd-lVth degree) due to project 554,000; from table 1. d/ Reduction in number of children severely malnourished (Ilird - IVth degree) due to project = 229,000; from table 1. e/ Number of deaths averted through improved nutritional status = 12,385; from table 2. APPENDIX D Page 4 of 4 Table D-4. Cost Effectiveness of PROAPE Preschool Feeding and Stimulation Program I. First and Second Grade Drop-out and Passing Rates for PROAPE and non-PROAPE Students (percent) Dropped out Passed PROAPE Non-PROAPE PROAPE Non-PROAPE 1st Grade 7 9 68 57 2nd Grade 6 6 79 62 II. Estimated Additional First and Second Grade Graduates due to Project No. of Children = 22,298 Difference due PROAPE Non-PROAPE to program A. First Grade Graduation Rate (0.93)(0.68) 0.63 (0.91)(0.57) = 0.52 0.11 No. of Graduates 14,048 11,595 2,450 B. Second Grade Graduation Rate (0.94)(0.79) 0.74 (0.94)(0.62) = 0.58 0.16 No. of Graduates 10,396 6,725 3,671 III. Cost Effectiveness of PROAPE: Cost per Second-Grade Graduate PROAPE Non-PROAPE Project Cost 1,036,370 - Cost for First Year Graduate ($98x22,298)a/ 2,185,200 2,185,200 Cost for Second Year Graduate ($98 x First Grade) 1,376,700 1,136,310 Total Cost for First Two Years 4,596,270 3,321,510 Cost per Second Grade Graduate $442 $494 a/ The estimate is based on cost per primary student per year in Nicaragua, assuming costs are about the same in Northeast Brazil and that the efficiency of the two systems is similar; first-grade repeaters are excluded. (Nicaragua Academy for Educational Development data, 1975.) APPENDIX E Page 1 of 3 NUTRITION COMPONENTS Urban Development Projects 1. Most urban slum upgrading projects coatain provision for building, equipping, and expanding health posts, clinics, and centers that are to be used for diagnosis and treatment of malnutrition and for nutrition education programs in conjunction with standard health services. Community centers built under the projects are also sometimes used for nutrition education. Some projects provide for the construction of markets to improve access to food or to help lower food prices. 2. A $1.1 million nutrition pilot project in the Tanzania Naitional Sites and Services Project included two nutrition and garden centers and ten demonstration gardens. To encourage fruit tree and vegetable cultivation and poultry keeping, extension services, nutrition education, and cooking classes were established. Seedlings, insecticides, and tools were sold in the centers. The program was popular and the centers have become self-supporting. Success has been attributed to the dedication and inventiveness of the horticulturalist who ran the project. 3. A health and nutrition program in the Second Calcutta Urban Development Project involved identification of malnutrition, followed by supplementary feeding and nutrition education in nutrition clinics. The government has extended it to a population of 2 million bustee (shanty) and slum dwellers, a tenfold increase, in the Third Calcutta Project. 4. In a component of the Bolivia Urban Development Project, high school students followed up patients to see that those who needed further nutrition or health care returned to clinics. The students' commitment to the task and the low administrative costs apparently contributed to success, leading to a request to expand the component. Rural Development Projects 5. Nutrition status surveys in the Second Fisheries Project in the People's Democratic Republic of Yemen and the Grain Storage and Processing Project in the Yemen Arab Republic have led to follow-on activities. Nutrition extension in the Puno Rural Development Project in Peru is well executed and the number of villages serviced has increased 50 percent from original estimates. In the nutrition component of the Philippines Rainfed Agricultural Development Project (Iloilo), anemia control and diarrhea management have exceeded their targets and monitoring and evaluation activities are well implemented. The Volta Regional Agricultural Development Project in Ghana has expanded home extension work (including promotion of home gardens and reduction of food waste) to five districts. All Northeast Rural Development projects in Brazil include social extension, part of which is nutrition education. The Philippine Agricultural Support Services Project has financed the establishment of a food and nutrition unit in the Ministry of Agriculture to assess nutritional needs and to design policies and programs based on the Philippine Food and Nutrition Plan. 6. In Mauritius, there are 2,300 kitchen gardens in a component of the Rural Development Project (many more were lost in 1982 cyclones) and a considerable demand for additional government assistance. All but 3 APPENDIX E Page 2 of 3 percent of the families with the gardens claimed their food intake and their incomes (from sale of surplus foods) had increased. A companion rabbit-rearing project was somewhat less successful (the rabbits' mortality rate was 47 percent). Nutrition improved directly in 39 percent of the cases, and indirectly from an average income increase of 10 percent from rabbit sales. 7. The nutrition component in Rwanda's Second Rural Services Project (BGM) includes explicit support for monitoring malnutrition, its causes and the impact of remedial measures in the project area. This could serve the country and the Bank as a field laboratory for exploring alternative approaches to nutritional improvement. Population and Health Projects 8. Because of the way health projects are constituted, and in order to minimize the number of their components, few include full nutrition components. Many health projects are concerned primarily with developing infrastructure or improving management. However, projects do commonly include nutrition work (as reflected in table E-1) as part of other programs. Assessment of nutrition status is part of community health care under the Peru Health Project. Nutrition education is part of the cotaponent to strengthen basic health services in the Senegal Health Project, and part of broader communications components in second projects in Tunisia, Philippines, and Indonesia. (The distinction between nutrition and health components is often blurred. For example, oral rehydration was devised as a nutrition intervention and is now a standard part of many primary health care programs, as is growth monitoring, which was devised by the nutrition community.) 9. Certain projects have discrete nutrition components. A major and apparently effective mass media breast feeding promotion was conducted under the Jamaica Population and Nutrition Project. In India Population I, nutrition education was shown to have positive effects on nutrition Table E-1: Bank Population and Health Projects that Include Nutrition Actions Breast Growth Oral rehy- feeding Nutrition Supplemen- Project monitoring dration promotion education tary feeding India I ... ... ... x x Jamaica II ... ... x x Malaysia I ... ... ... x... Malaysia II ... ... x x ... Peru I x x x x Philippines II x x x x ... Senegal I ... x ... x ... Thailand I x ... ... x x Tunisia II ... x x x ... Yemen Arab Rep. I ... ... ... x ... Yemen PDR I ... x x x ... APPENDIX E Page 3 of 3 awareness and dietary practices but distribution of food supplements to have minimal nutrition impact. A nutrition component in the Health Project in the Yemen Arab Republic has established a nutrition unit: to plan and coordinate food and nutrition activities. 10. One of the more successful nutrition components of a PHN project is in Northeast Thailand, where villagers have formed nutrition cooperatives (90 at the outset of the project) to grow or purchase legumes and other ingredients, then roast, grind, and package them as weaning foods. The packets are sold in the village as part of a growth monitoring program and given free to children with severe malnutrition. When all severe malnutrition in a village has disappeared, the cooperatives can market the food in nearby villages. Some cooperatives are becoming self-sustaining. Although firm impact data are not yet available, there are clear signs of decreasing malnutrition resulting from the programs. The government's Fifth Plan includes expansion of the program to 14,000 poverty hamlets, representing about one-third of rural Thailand. Education 11. The Northeast Brazil Basic Education Project finances the construction and equipping of warehouses and purchase of vehicles to improve the efficiency of the school feeding program. The quality of school feeding also receives attention in the Second Syrian Arab Republic Education Project, as does nutrition education. Nutrition education also receives attention in the Togo Education Project. APPENDIX F Page 1 of 4 FINDINGS FROM ECONOMIC AND SECTOR WORK Nature, magnitude and causes 1. In nine of 16 countries studied, the evidence is strong that nutrition problems are serious -- that is, moderate to severe among a significant fraction of the child population, with adult undernourishment among some nontrivial fraction of the population compounded by micronutrient deficiencies. Lack of sufficient energy in the diet is most commonly identified as the major prob.lem.1/ For some countries (e.g., the Philippines, Zimbabwe) the problem is not only the number of calories but also their density. Information on the numbers suffering from micronutrient deficiencies generally is too thin to draw conclusions about their importance. 2. In a few countries where consumption has been rising, a portion of the population (estimated at 15 percent in Indonesia, certainly higher in others) is lagging badly and may require special help, given the lag time before general economic growth could be expected to raise incomes sufficiently to erase energy deficits. 3. In many cases it is clear that increases in income are a necessary condition for any improvement in the nutritional status of adults and important to children's status. Since for most countries the largest nutritionally distressed groups appear to be rural, the greatest: potential impact of public policy appears to be in the area of food production (especially where a big fraction of the income of malnourished families consists of food produced for their own consumption -- e.g., Rwanda, China), rural employment and farmgate food prices, supplemented by targeted attention to malnourished children. 4. In several other cases, malnutrition does not appear to be income bound. In Brazil, for instance, food acquisition levels have not responded closely to increases in income (malnutrition is observed among low-income families where discretionary income goes into TVs and other durable goods), and in the Yemen Arab Republic deleterious child feeding practices and poor dietary selection and preparation habits persist. 5. A common problem of considerable importance is the tension between a pricing policy that encourages domestic food production and a nutrition policy that aims at keeping retail prices of food grains at a level the poor can afford. Pakistan, with a two-track subsidy system benefiting both producer and consumer, seems to have come closest to dealing effectively with this dilemma. 6. Current Trends. In some cases (e.g. Peru, Yemen Arab Republic), time-series data have made it possible to conclude that the nutritional status of the poorest has been deteriorating even though this does not show up in health data. In two important cases (Indonesia, Nigeria), food price changes are identified as causing nutritional deterioration. (Income movements are also suggested.) High priority should be given to substantiating these trends. Especially where there is reason to think that fundamental economic factors may be pressing downward 1/ Malnutrition is sometimes the result of excess nutrients; this paper does not deal with those problems. APPENDIX F Page 2 of 4 on food consumption in the next few years, and where negative indications are strong (in Brazil,.Zambia, Yemen Arab Republic, Lesotho, and possibly Rwanda and Upper Volta), PHN should carefully follow the formulation and implementation of food policy and supplementary measures. 7. Seasonal Malnutrition. Seasonal food stress was found to be important in Zambia, Indonesia, Lesotho, Zimbabwe, Rwanda and Upper Volta. Whether the problem stems from changes in income (seasonal declines in effective demand) or swings in food supply is not clear. Solutions cannot be proposed until the relative contributions of the supply and demand sides to the problem are established. 8. Shifts to Imported Foods. In a few countries a rapid shift in consumers' preferences to rice (and sometimes wheat) or other imported foods is sweeping away consumption of traditional, locally produced grains (such as millet and sorghum in Nigeria and Yemen). These countries face reduced food security, a decline in expenditure on domestically produced grains (and in rural income), and possible micronutrient problems stemming from the shift to highly processed grain (e.g., in Zambia-and Zimbabwe). Administered prices favoring rice and refined wheat flour are a common feature behind these shifts. This is potentially a major nutrition danger that deserves attention as a general problem. 9. Breast Feeding. Only in the sector studies for Brazil, Zimbabwe, and Zambia is there reference to changes in breast feeding. Significant declines in breast feeding would not be expected in much of Africa and in rural Asia, but in countries with large urban populations, they would. Statistical information is seldom available on the problem. But informed judgments can be made from import data-on infant formula and from the observation of professionals, for instance, in urban clinics or women's organizations working with the poor. It is not clear from the sector work whether the problem has been overlooked or its extent has been exaggerated. 10. How Interventions Have Worked. In several extensive treatments of the effects of agricultural pricing policies on food consumption levels, or of subsidy policies, there are no clear conclusions about efforts to manipulate prices for nutrition objectives. Household studies are used to show how much food purchases vary in response to changes in prices. Where possibly negative effects of agricultural production strategies on malnourished groups are considered, the sectoral reports do not clearly demonstrate the relation between specific policies and consumption outcomes. 11. Despite the recognized power of the mass media to affect tastes and behavior of many types, it is striking that with few exceptions (e.g., Zambia) no mention is made of efforts to use mass media to change nutritionally deleterious .behavior (diet and child feeding practices) that appears to be independent of household budget constraints. The study of the nutritional adequacy of food consumption and its socioeconomic correlates in Indonesia found that relatively affluent households with relatively educated homemakers opt for expensive diets even at some sacrifice in terms of needed nutrients. 12. In countries where government commitment has enabled development of reasonably effective delivery systems, direct nutrition interventions appear to have good possibilities. In Zimbabwe, for example, the sector APPENDIX F Page 3 of 4 work pointed up the potential importance of directing nutrition education efforts to men, given their control over family resources, as well as to women as currently is the case. Sector reviews identify successful interventions in Brazil where companies are required to make provision for breast feeding by women employees with infants, in Zimbabwe's supplementary feeding programs for children, in Rwanda's network of nutrition centers (for surveillance, rehabilitation, supplementation and applied nutrition activities), in Pakistan's consumer food subsidies, in China's support of a general minimum consumption level, and in Zambia's mass media effort on child feeding practices. They also identify unsuccessful interventions, such as protein enrichment in India and consumer subsidies on the wrong foods in Zimbabwe. 13. Where governments are contemplating or are already conducting programs that offer food supplements, subsidies, food for work, or other potentially large scale interventions, the paper comparing the Bangladesh interventions and the approaches discussed in the background paper on costs2/ offer useful methodology for reaching judgments on relative co,st effectiveness of alternative programs. 14. The sector reviews give a strong impression of nutritionists' heavy reliance on other disciplines and other areas of organizational responsibility. Nutrition authorities have a relatively narrow track for delivering services independently. The sector studies emphasize the importance of infusing nutritional considerations into areas that have an impact on nutrition and of developing food and nutrition focal points and- coordinating processes in government. There are some exceptions, however, where nutrition programs have proceeded independently, even ahead of the implementation of primary health delivery systems (e.g., Zimbabwe, Rwanda and Indonesia). Lessons for Doing Sector Work 15. A format should be developed that properly prepared analysts can use to scan the spectrum of nutrition issues to identify aspects that would be worth a more probing and time-consuming effort. The Bangladesh report, currently under preparation, is an example of such a focused effort. Since striking similarities in nutrition problems and causes exist among some groups of countries, a short list of specific things to look for should be added to the guidelines. Sector analysis teams should have leeway in their methods, however, since there are, not surprisingly, great differences among countries. 16. In large countries where substantial regional diversity is found in health systems, food productivity and distribution, income level, seasonal malnutrition, etc., follow-up studies should focus on areas where malnutrition is most severe. Countrywide analysis should then be limited to broad policy questions. 17. Nutritional analysis should capture key determinants and identify those subject to some change at acceptable cost levels. Income and food price are powerful factors that must normally be approached a: a national level, but community and household variables are even more important in 2/ T. J. Ho, "Cost, Affordability and Cost Effectiveness." APPENDIX F Page 4 of 4 some cases. In countries such as Thailand where the major explanations for malnourishment, and major leads for intervention and policy, are found at the lower structural levels, the national bird's eye.view that works downwards would be less important. To the extent that significant determinants were identified at provincial or national levels, the analysis would broaden to cover aggregate questions. The bottom-up approach would identify operational conclusions more quickly and surely and would be more cost-effective in use of staff time. 18. Because of the often limited amount of data available -- and the need to generate new data -- nutrition sector work generally should begin with a brief mission that poses the issues and makes arrangements for collection of daca which then would be available for a subsequent mission to evaluate. 19. Current data can be obtained rapidly, as demonstrated in Zimbabwe, by sending brief questionnaires to health field staff (government, voluntary agency, and religious mission doctors) and agricultural extension staff asking for their judgments about nutrition conditions and causes; mobilizing medical school students to undertake spot nutrition surveys in different parts of the country; and adding the collection of data on food consumption and nutrition status to studies already planned on agricultural production and income. 20. Where the feasibility of sector work is not clear, a few days spent in the field by a staff member or consultant visiting a country for other reasons can be helpful in addressing the questions necessary in making that determination. The very brief effort that went into the Mauritania analysis was sufficient to conclude that further Bank attention to nutrition was unlikely to be productive. 21. Nutrition sector work needs to take place not just in preparation for projects but in the process of projects. Since a main point of nutrition projects has been to help governments identify and carry out nutrition policies, sector work during a project helps the Bank shore up its side of the dialogue. The Indonesia project provides the best example of how nutrition sector work by Bank staff during a nutrition project helped make possible a continuing policy dialogue throughout much of the life of the project. 22. Finally, two important pre-conditions appear from the experience to be necessary for successful sector work. First is the prospect that funding will be available for a project, which generally has not been the case. Second is the need for the sector work to be controlled by a staff member who is engrossed in it and can provide the required continuity. Several otherwise promising opportunities were lost when the work was done solely by consultants or Bank staff outside the responsible operating division. APPENDIX G Page 1 of 7 RELEVANT FINDINGS FROM RECENT RESEARCH AND EXPERIENCE BY OTHERS 1. While the Bank has been undertaking its initial work in nutrition, new knowledge has emerged from the research community and operational experience has been gained by governments and other aid agencies in this relatively new field. Work relevant to Bank interests should, of course, be reflected in Bank plans for future nutrition actions. Nutrient Requirements 2. Scientists have directed attention increasingly in recent years to the effect of malnutrition on function. This research has increased understanding of energy (calorie) need and expenditure. At the time that Bank nutrition projects were designed, the prevailing assumption. of the scientific community was that increases in a malnourished child's nutrient intake (or decreases in his nutrient losses) would lead to growth and in pregnant women's to increased birth weight. Accordingly, changes in heights and weights were used as key indicators of project benefits. Recent evidence that energy consumption that does not show up in growth is used in increased physical activity suggests that the measures used in evaluating nutrition projects probably did not capture a good portion of the benefits derived.l/ 3. Lower levels of physical activity are now seen as an accommodation to lower energy intake. In children this may be reflected in cognitive and affective development, and in societies the accommodations may set the norms for activity.2/ 4. New knowledge of nutrient requirements, with implications for the kinds of foods to be included in intervention programs, suggests that most traditional mixed diets, if taken in adequate quantities, would meet protein needs. In areas where tubers such as cassava are a staple, and possibly in some rice-eating areas, protein intakes may be marginal. , George Beaton, "Energy in Human Nutrition," W. 0. Atwater Memorial Lecture, Western Hemisphere Nutrition Congress, Miami Beach, Florida, August 8, 1983. 2/ Thus energy requirements would include consideration of human functions described in terms of desirable activity levels and other behavioral patterns that may differ from societal norms. Sizable work on 'functional consequences of malnutrition' (on duration and severity of disease, cognitive development, productive capacity, reproduction and social performance) is now under way (much of it financed by the U.S. Agency for International Development), but is unlikely to result in measurement devices suitable for operational use within this decade. APPENDIX G Page 2 of 7 Otherwise, the concentration on processed protein foods and seeds bred for higher protein levels appears less important than it did a decade ago.3/ 5. Caloric density has also emerged as an issue of particular importance to the very young. Sometimes where sufficient quantities of calories are available, they may not be ingested because of the bulk of staple roots or grain that must be consumed. Caloric density was identified in the Bank's sector work as the major cause of malnourishment among Filipino children and a program was designed to increase their intake of calorie-dense vegetable oils. Nutrition-Health Links 6. Recent research confirms and gives greater precision to earlier understandings that malnutrition contributes to infectious diseases and that various parasitic, bacterial and viral diseases, in turn, increase energy demands and decrease food absorption. The problems are compounded by anorexia and cultural practices that limit the amount of food during illness. Even mild infections lead to anorexia and are cause for providing extra food for growth following a disease. This has implications for design of nutrition education, food supplementation, and integrated nuitrition and primary health care programs. 7. In Bangladesh a diarrhea-nutrition research program found that adequate calorie intake during acute illness was critical in preventing malnutrition. Researchers developed a cheap oral rehydration solution using local ingredients. Their concept is now being promoted worldwide by UNICEF and WHO. 8. Intergenerational effects of malnutrition now are being documented. Nutritional status is a cumulative manifestation of deprivation, including parental malnutrition, that must be dealt with in - terms of decades rather than single years.4/ Mothers of short stature due to chronic childhood malnutrition are more likely to have small babies who, in turn, are more likely to be malnourished and likely candidates for early deaths. 9. In rural Guatemala, where good medical care was available, a research project found that the risk of death for babies weighing less than 2,500 grams at birth was four times higher than for heavier babies. In Bangladesh and India, studies show the mortality rate among children below 60 percent of standard weight-for-age (third degree malnutrition) is three to four times higher than among better-nourished children. 3/ Nutrient requirements and indeed whether moderate growth failure itself is in any way disadvantageous has been the source of considerable debate in the nutrition community. A soon-to-be-published report by a WHO/FAO/UNU Expert Committee established to study this will show that nutrient requirements have not changed substantially. / Lincoln Chen, "Introducing Nutrition into the Rockefeller Foundation's Agricultural and Health Programs," paper for conference on food policy analysis, Bellagio, Italy, November 1-5, 1982. APPENDIX G Page 3 of 7 10. A study in Malaysia has demonstrated that breast feeding is particularly critical in areas where impure water is consumed. And - research is beginning to provide insight on nutritional thresholds. An Indian study, for instance, indicates that for a child at 70 percent of the median for development, immune capacity is markedly decreased. 11. Government nutritionists in Bangladesh and China have successfully addressed two debilitating and sometimess deadly diseases. In the future, the million-or-so South Asian cases of lathyrism (a total paralysis of arms and legs caused by eating an inexpensive, drought resistant pulse -- kessari).can be prevented by adding vitamin C to the diet. Tens of thousands of Chinese cases of keshan disease, commonly leading to cardiac arrest among young children, have been reversed by providing the trace mineral selenium, and the mortality rate has already been reduced to less than 10 percent. New Technologies 12. Nutritional anemia, probably the most widespread of the micronutrient deficiencies, affects about half of the women in developing countries as well as sizable portions of the male and child populations. It has been demonstrated that it can deleteriously alter brain chemistry and function and thereby influence human behavior itself. Fortunately, a breakthrough in food technology by Indian scientists has made it possible to fortify a nation's salt supply with iron, at a cost $0.05 to $0.09 a year per person.5/ 13. - Research in Central America has demonstrated that fortifying sugar with vitamin A can reduce nutritional blindness and the severity of measles. In recent years, vitamin A has been given in high-potency periodic doses at reasonable costs (about $0.10 for the capsules). In Bangladesh, which has a highly focused village-level administrative set up, wide-scale distribution of capsules substantially reduced vitamin A deficiency. About 2,500 cases of blindness there are estimated to be prevented each year. 14. For several decades, salt fortified with iodine has been used effectively to control endemic goiter and reduce endemic cretinism and deaf-mutism (and, more recently, understandings have advanced about subtle effects on mental and behavioral development of even relatively mid iodine deficiencies). These programs are highly cost effective ($0.08 per target person per year or $0.20 per case of goiter prevented), although many countries that have legislation calling for iodization do not have effective programs. Where salt is not centrally processed, a mass dose attack using an intramuscular injection of iodized oil (at $0.24 per person) is now available. 15. Some modest nutritional advances have been made in agronomic plant breeding, in addition to the well-publicized increases in yields. The Asian Vegetable Research and Development Center in Taiwan has developed a high carotene sweet potato and work in Latin America has progressed on increasing the protein content of cassava. 5/ Opportunities with older technologies also exist to fortify centrally processed grain products with vitamins and minerals at a cost of about $0.08 per person a year. APPENDIX G Page 4 of 7 Food Supplementation 16. Researchers in Guatemala and India (but not in the Gambia) have found that supplementing the diets of undernourished pregnant women significantly reduces stillbirths and deaths in the first month after birth and improves birth weight. Maternal supplementation looks to be the most cost-effective means of reducing child mortality and improving the nutritional status of those who survive, as well as protecting the health of the mother, which is indirectly essential for child health and survival. 17. Research in Zimbabwe and India has shown that children with the greatest weight deficit show the greatest response to feeding programs. Thus, directing feeding programs to children with third degree malnutrition or those "at risk" should be highly cost-effective.6! 18. In many areas, malnutrition and nutritional damage have been discovered to vary with the seasons, a finding that has important implications for the design and timing of feeding programs. 19. Many feeding programs have been evaluated but their methodological shortcomings generally limit the usefulness of their conclusions. Among the better but still limited studies are of large preschool food distribution programs in Sri Lanka and the Philippines which reportedly brought physical improvements to 30 percent and 38 percent of the children respectively. School programs in India, the Philippines, and Upper Volta were judged to reduce malnutrition by 24 percent, 22 percent, and 20 percent. Impressive nutrition results of mass child feeding during famine and civil strife have been well-documented in India, Bangladesh, and Zimbabwe, but such programs at other times generally have been less impressive in terms of weight gain. 20. In Chile, a combined feeding and psychomotor stimulation program is credited as the major influence in a dramatic decline in second- and third-degree malnutrition, which now affects less than one percent of the child population. Previously, among the approximately 5,000 marasmic children under six months of age each year, 96 percent died. Chile's treatment includes an average four-month stay in nutrition rehabilition centers for those who need it. A computerized growth monitoring program covers 88 percent of the children in the country (the remainder are mostly children of high income families who do not use government health services). Such a program depends on an effective health service and at this stage is limited to relatively affluent countries. Costa Rica also has a successful nutrition program based on a national nutrition surveillance system. 21. Brazil's Federal District has a combined feeding and day-care program that claims to have reduced infant and child mortality to 15 percent of what normally would be expected. That achievement is attributed both to the food and extra care given at the center and to the nutritional effects of the extra income that mothers earn in their time now freed for employment. 6/ Characteristics that identify at-risk children are: sibling of a malnourished child, breast and milk failure of mother, repeated diarrhea, sixth or later child in birth order, birth weight less than 2,500 grams, single-parent family, death of brother or sister under twelve months, a twin, and very low economic status. APPENDIX G Page 5 of 7 22. In the United States, a feeding program directed at women, infants, and children reduced iron deficiency anemia markedly in low-income families. And in Korea an applied-nutrition program has significantly increased consumption of major nutrients over an eight year period; the heights and weights of those in program villages have also increased significantly, but that cannot be attributed to the program alone. 23. In Brazil, sugar cane workers who received lunch were found to be 26 percent more productive than those who did not; their employers are now regularly providing lunches to all workers. (The Brazilian government offers tax incentives valued at $534 million a year to firms that provide meals to workers, although there is little understanding of the effects.) In Guatemala a study of sugar cane workers showed no effect of improved nutrition on labor productivity, but the levels of the workers' spare-time activity increased. The record seems pretty clear about the general importance of energy intake to production, but also about the inadequacy of food supplements for workers as a means of producing immediate increases in output. The importance is suggested largely by the relationship between body size and output. Food supplements, however, rarely appear capable of making up'in a short period for damage suffered by workers' bodies during a long period of previous nutritional abuse, or of disturbing whatever adaptation to nutritional deprivation may have taken place.7/ Nutrition Education 24. Among the few nutrition education programs that have been evaluated, a small number show substantial effects. Nutrition education in company with food supplements in Morocco produced greater changes in nutritional status than did supplementation alone. Media-based projects in the Philippines, Tanzania, Honduras, Micronesia, and the Gambia have reached large audiences who display new knowledge and changed attitudes; the projects report some changes in practice. None of these should suggest that nutrition education can routinely be done effectively on a large scale. However, a review of evaluations of 15 projects using mass media to teach nutrition in low-income countries concluded that such projects can reach large numbers of people (up to several million) in a relatively short period of time and that 10 to 50 percent of those reached will remember the main message of the program, at least during or immediately after the project.8/ Nutritional Effects of Agricultural and Price Policies 25. Programs encouraging farm families to switch from subsistence to cash crops have on occasion been shown to have negative nutritional 7/ Davidson Gwatkin, "Does Better Health Produce Greater Wealth? A Review of the Evidence Concerning Health, Nutrition and Output," A.I.D., August 1983. 8/ Joanne Leslie, "Evaluation of Mass Media for Health and Nutrition Education" in Health Education by Radio and Television, Ed. by Manfred Meyer, Munich, K.G. Saur, 1981, pp. 198-223. APPENDIX G Page 6 of 7 consequences.9/ They are said not only to change incomes and relative prices but the distribution of income and budget control within households, exposure to expensive foods and nonfoods, the position of women and demands on their time and the pattern of ownership of resources. In Kenya production of sugar and tea in some instances was found to have a negative effect on nutrition status; production of foods that could be eaten as well as sold and that boosted income had a positive effect.10/ In Papua New Guinea several studies have pointed to possibly deleterious effects on nutrition from introduction of cash crops. Generally, evidence on cash cropping is too scarce and too fragmented to be of much use at this stage for the design of policies and projects.11/ 26. Consumer-oriented food-price subsidies are widespread and, although not focused specifically on improving nutrition, may contribute to such improvement where lack of purchasing power is a .major reason for malnutrition. Studies in Kerala, Sri Lanka, Bangladesh, Pakistan, and Mexico indicate that consumer subsidies have had a significant effect on food consumption. In Pakistan, in the second lowest income group, increments of 8.6 percent of the calories required and 24 percent of the protein required were attributed to participation in the subsidized ration system. Similarly, in Sri Lanka, the ration program provided approximately 20 percent of the total calorie intake of the lowest income group. 27. An evaluation of the Pakistan system did not find a significant decrease in the prevalence of malnutrition in pre-school-aged children within participating households, nor did an evaluation of the Conosupo program in Mexico. In Sri Lanka, however, the increase in child mortality by 70 to 80 percent during the 1974 food shortages has been attributed to the decline in availability of subsidized foods to the low-income families on tea estates. The subsidized food ration program in China also appears to be part of the reason for that country's relatively good nutrition condition. 9/ See Patrick Fleuret and Anne Fleuret, "Nutrition, Consumption, and Agricultural Change" in Human Organization, vol. 39, no. 3, 1980, pp. 250-60. 10/ Jon Hitchings, Agricultural Determinants of Nutritional Status Among Kenyan Children with Models of Anthropometric Growth Indicators, doctoral dissertation, Food Research Institute, Stanford University, June 1982. 11/ Research is increasing. AID is sponsoring a series of studies of the nutritional effects of selected agricultural policies, and IFAD is supporting analyses by IFPRI of the effects of technological change in agriculture on food consumption and nutrition. APPENDIX G Page 7 of 7 28. Several researchers have concluded that subsidizing "inferior" or low-status foods is one way to reach the poorest households.12/ The - potential effect of any consumer price policy will be enhanced if the subsidy can be applied to foods normally consumed in large amounts by the malnourished population but not eaten by other income groups. In Bangladesh, it was found that when low-priced sorghum was available through the ration shops, the poorest households chose sorghum over the higher priced, generally preferred rice or wheat. For each dollar spent, a larger quantity of sorghum than of rice or wheat could be purchased, and the effect on household food intake was thereby greater. 29. Trying to reach nutritionally vulnerable individuals within households remains a problem. Research now is under way in several countries that may provide better understanding of how food is distributed within families. 12/ Rezaul Karim, Manjur Majid and F. James Levinson, "The Bangladesh Sorghum Experiment," Food Policy, Vol. 5, no. 1, February 1980, pp. 61-3, and C. Peter Timmer, "Commodity-specific Orientation to a Nutrition-related Food Policy," in Malnourished People: A Policy View.