Sri Lanka - Nutrition review

Report No. 7575-CE Sri Lanka Nutrition Review July 13, 1989 Population and Human Resources Division Country Department I, Asia Region FOR OFFICIAL USE ONLY . . a ., .j / 'V o~~~~~~~~~~~I .-,-J . . . ,~~~~~~~~~~~~~~~~~~~~~~~~~~ .).. .k Docoment of the World Bank This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World.Bank authorization.  FOR OFFICIAL USE ONLY ACRONYMS AEU - Adult Equivalent Unit AGA - Assistant Government Agent ARC - Administrative Reforms Committee ARI - Acute Respiratory Infections BCG - Tuberculosis Vaccine CFS - Consumer Finance Survey CPI - Consumer Price Index DD - Diarrheal Disease DHS - Demographic and Health Survey DPT - Diphtheria, Whooping Cough and Tetanus Immunization FNPPD - Food and Nutrition Policy Planning Division FWAE - Farm Women's Agricultural Extension Program GA - Government Agent GDP - Gross Domestic Product IDD - Iodine Defic'.ency Disorders IFPRI - International Food Policy Research Institute JSP - Jana Saviya Progran (People Power Program) MPPI - Ministry of Policy P'anning and Implementation LBW - Low Birth Weight MOPI - Ministry of Plan Implementation MRI - Sri Lanka Medical Research Institute NAFNS - National Agricultural, Food and Nutrition Strategy NFNCC - National Food and Nutrition Coordinating Council NHC - National Health Council OECD - Organization for Economic Cooperation and Development OPV - Polio Vaccine ORT - Oral Rehydration Therapy rMB - Paddy Marketing Board Thriposha - Formulated, pre-cooked supplementary food used in Ministry of Health intervention program UNICEF - United Nations Childrens Fund USAID - U.S. Agency for International Development WFP - World Food Program IThis document has a triesnted distribution and masy be use by recipients only in the perfonnance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.|  SRI LANKA NUTRITION REVIEW Table of Contents Page No. INTRODUCTION AND EXECUTIVE SUMMARY ......................... iii Current Nutrition Problems and Causes . .......... iii Food and Agriculture .. iv Current Interventions .. iv Institutional Arrangements .. v Issues and Main Recommendations . . v I. NUTRITION CONDITIONS AND CONSEQUENCES ...................... 1 Introduction .. 1 Child Nutrition Status .................................. 2 Maternal Nutrition Status .. 3 Hicronutrient Deficiencies .. 4 Recent Trends .. 5 Comparisons ............................................. 5 Functional Implications of Malnutrition . . 6 II. NUTRITION DETERMINANTS ..................................... 7 A. Determinants ............... 7 Household Food Insecurity ............................... 7 Behavioral Factors .. 9 Infectious Diseases .. 10 Most Seriously Affected Groups . . 12 III. NUTRITION IN SRI LANKA'S FOOD AND AGRICULTURE SYSTEM ....... 13 A. Dietary Patterns .. 13 B. The Role of Agriculture .. 14 Rice Development .. 14 Subsidiary Food Crops .. 15 Food Production and Marketing Patterns . . 15 Impact of Pricing .. 16 Procurement and Distribution Policies . . 17 Processing .. 18 Nutrition Effects ..8.................. i IV. NUTRITION PLANNING AND PROGRAMMING ......................... 19 A. Institutional Arrangements .............................. 19 B. Nutrition Objectives, Policies and Strategies . . 20 Objectives ............................................ 20 Nutrition and Related Food and Agriculture Strategy Development .. 21 National Nutrition Action Plan . . 25 NAFNS Implementation Plan .. 27 Implementation Status of Nutrition and NAFNS Plans ...... 28  - it -: V. NUTRITION COSTS AND INTERVENTIONS .......................... 28 A. Nutrition I ntervention Progrsms ......................... 28 Food Stamps ........................................... 29 Thriposha ................. # ........................... 31 On-site Feeding ....................................... 33 Nutrition Education . . . 34 Hea,th-related Interventions ......................... . 35 New initiatives .... .......................... .......... 37 Other Donor Activity. ........................... 38 B. Funding.... .............................................. 28 VI. ISSUES AND OPTIONS: AN ACTION AGENDA ...................... 38 A. Objectives .............................................. 39 B. Policies and Strategies ................................. 39 Food and Agricultare . . . 39 Health .............................................. 41 Nutrition Education ... 41 C. Institutional Arrangements .. ................. ............ 41 D. Programs ............................................. 42 Food Stamps ...... 42 Rice Grading .. 43 Augmented Distribution through Public Assistance Scheme .. 44 Thriposha . ............................................ 44 Kola Kenda ..........* 44 School Feeding ....... 45 Nutrition Villages .. 45 An Optional Approach to Nutrition Villages ............ 45 E. Recommendations ..... 46 ANNEX A - Tables 1-12 ANNEX B - The Ultra Poor ANNEX C - Integrated Community Nutrition Intervention  - iii - SRI LANKA NUTRITION REVIEW Introduction and Executive Summary 1. This report seeks to consolidate the most salient findings of exist- ing analysis to promote and help focus the development of future Sri Lanka nutrition policies, strategies and programs. It results from the work of a Bank nutrition review mission to Sri Lanka in May 1988,1/ and a brief mission by the Task Manager in March, 1989. The findings herein draw on extensive documentation from Sri Lanka, important analysis of 1981-82 and 1986-87 household food consumption and expenditure data and the food stamps program by the International Food Policy Research Institute (IFPRT) and others, agriculture sector work by the World Bank and the Orgatization for Economic Cooperation and Development (OECD) and evaluations of the thriposha program funded by the U.S. Agency for International Development. Current Nutrition Problems and Causes 2. Sri Lanka's record in the field of health is impressive. Health indicators, including life expectancy, infant and child mortality and maternal mortality are significantly better for Sri Lanka than for other countries at similar levels of development. Moreover, an analysis of the most recent household expenditure data (1986-87) indicates that caloric intake for the poorest deciles has increased from very depressed levels in the 1980-82 period (a period characterized by a sharp increase in prices and the aftermath of subsidy elimination) to more acceptable if not yet nutritionally adequate levels not seen among these groups since the 1969-70 period. 3. In light of this good performance in health and the recent improve- ments in food consumption, it is surprising that malnutrition among children and pregnant women should continue to be a major concern. Recent surveys and health service reports indicate that over 25Z of pre-school children are stunted from chronic malnutrition. Over 1OZ are acutely malnourished, which is high, particularly in view of Sri Lanka's low infant and child death rates. Around 5Z are both stunted and wasted. Low birth weight, mainly attributable to maternal malnutrition, occurs in more than an estimated 202 of deliveries. Iron deficiency anemia affects more than an estimated half of low-income women and the birth weights of their children. Iodine deficiency disease remains prevalent in some densely-populated areas. More progress has taken place against immunizable childhood diseases than against morbidity from two leading causes--diarrhea and acute respiratory infections. 1/ The mission consisted of James Greene, principal nutrition specialist, as task manager and three consultants: Dr. John Kevany, public health/nutrition specialist; Mr. James Levinson, economist, and Mr. D. Tharmaratnam, financial analyst.  - iv - 4. Malnutrition in Sri Lanka results from a combination of three factors. The primary one is household food insecurity, which affects predominantly the lowest three income deciles in urban areas and the poorest rural decile through limited purchasing power reflecting low levels of employment and incomes. The others are poor health and family food behavior which compound the corrosive effects of economic deprivation but also directly affect the nutrition status of pre-school children and pregnant and nursing women in less precarious economic circumstances. The continuing levels of stunting in the face of declining proportion of food-insecure households underscores the importance of adequate attention to health and behavioral faccors. Food and Agriculture 5 Food and agriculture have an important nutrition role in Sri Lanka from both consumption and production perspectives. Domestic rice and coconut account roughly for 50Z and 20X, respectively, of average caloric intake. Bread, wheat flour and sugar are the next most important calorie sources. In an attempt to encourage production, current agricultural policy encourages higher, rather than lower, rice prices. The main beneficiaries are relatively small numbers of larger rice producers, since smaller ones for the most part are net purchasers of rice. Sugar prices well above world levels tax the low- income consumer disproportionately. Lack of research and development atten- tion to other field crops such as corn, coarse grains, yams and cassava constrain the development of both incomes from rain-fed land and alternative sources of relatively low-cost calories. Current Interventions 6. Resources for nutrition have tripled since 1987. Sri Lanka now budgets more than an estimated Rs 6.0 billion yearly (around 9X of the current government budget) for nutrition, mainly on three food programs--food stamps, school feeding and thriposha. Food stamps benefitting around half the population cost around RS 3.6 billion yearly. Daily lunches for Sri Lanka's 4,000,000 school children will cost around Rs 2.6 billion yearly. Thriposha, a formulated and largely take-home supplement distributed mainly through the health system, is targeted to 580,000 malnourished preschool children and pregnant and nursing women. However, all three programs face problems. The food stamps program is poorly targeted; it reaches more people above than below the poverty line. It also appears to have bypassed significant numbers of the very poor. Despi_e a recent doubling of the value of the stamps, the current purchasing power is only around 702 of the original 1979 entitlement. The 3chool lunch program may have social benefits but cannot make up for the effects of malnourishment in early childhood. Thriposha has yet to achieve its full nutritional potential largely because of uneven performance in regard to beneficiary screening, counselling and program management and monitoring. 7. Other supplementary feeding programs operate but with limited nutri- tional effectiveness. Nutrition education activities and health measures against micronutrient deficiencies and non-immunizable childhood diseases which affect nutrition status also appear to have limited impact.  v 8. The Jana Saviva program (JSP) is supposed to eliminate or at least drastically reduce the need for food stamps by providing monthly transfer payments for 24 months to food stamp beneficiary families. The objective is for the recipient families to become financially self-sustaining at the end of that period. However, the RS 28 billion estimated value of the annual JSP consumption transfers are around five times the amount estimated to bring the poorest quintile of the population up to middle-class levels of food consumption. InstituL_onal Arrangements 9. Sri Lanka has had some success in articulating important features of both a national nutrition policy and strategies to implement it. Howe-ter, recent Government reorganization measures represent a prospective step back from Sri Lanka's previously progressive efforts to incorporate nutrition into national development plans and set up an official focal point for that purpose. Until reorganization, Sri Lanka was one of relatively few developing countries to try explicitly to incorporate nutrition into national development plans and set up a government focal point for that purpose. The National Health Council, which consists of Cabinet Ministers and is chaired by the President, and the Food and Nutrition Policy Planning Division (FNPPD) of the Ministry of Plan Implementation (MOPI), were the principal channels for both nutrition advocacy and intersectoral development of national nutrition policies, strategies and programs. FNPPD recently was disbanded, with MGPI incorporation into a new Ministryr of Policy Planning and Implementation. Under earlier arrangements, links between policy and strategy were fragile and efforts to translate strategy into programs were uneven. The recent reorganization weakens the nutrition and food planning framework still further. Issues and Main Recommendations 10. Sri Lanka needs to confront two principal nutrition issues. The first is how to deal with chronic nutrition insecurity at the household level, which particularly affects the lowest income deciles in rural areas and the three poorest urban deciles. The second is how to ensure adequate nutrient availability and utilization by specific target groups, particularly young children and pregnant and nursi.g mothers. Nutrition policies, strategies and programs need to be directed specifically to both problems. 11. Over the longer-term, income and employment gains by the poor will drive durable increases in their food consumption. However, even income and employment generating programs for low-income groups by the poor need to be complemented by other medium-term approaches. 12. A manifest need exists to review, revise, prioritize and quantify existing nutrition objectives and set them in a realistic time fram . Over the next five years, it should be possible through improved program perfor- mance to achieve reductions of 25Z in stunting among younger children and in the incidence of low birth weight and a 30? reduction in the incidence of maternal iron and folate deficiency. The government also may wish to set specific minimum food intake goals as a contribution to both the above object- ives and improved health and nutrition of the most vulnerable segments of the  - vi - population, such as maintaining daily consumption by the ultra-poor 2/ at no less than 902 of estimated calorie needs, which would require more attention to the lowest two urban income deciles. 13. Consideration should be given to focussing food and agriculture strategy on increasing both food consumption or incomes of the poorest deciles and the availability of lower-cost nutrients generally in short supply. In that regard, three approaches might be considered which would dispropor- tionately benefit the most vulnerable. First, efforts to lower rice prices to consumers through increased efficiency of production and post-harvest techno- logy might be explored. The latter would include improvements in the rela- tively inefficient milling industry where, for reasons not entirely clear, market factors nave not led to mode-nization. Second, reducing sugar prices approximately to world levels woul _mprove access to an important calorie source for the poor, but the implications of such a move on the balance of payments and on government revenues should first be studied. Third, increas- ing the output and marketinig of subsidiary field crops such as corn and manioc could generate additional employment and incomes as well as provide an optional source of relatively low-cost calories. Successful diversification, however, would require improved research and extension services and would not necessarily succeed unless demand and relative agricultural output prices provided adequate incentives. 14. Nutrition also needs to be made a more central focus of the primary health care agenda, mainly by more vigorous execution of current programs. Nutrition education strategies need to be developed to improve feeding prac- tices for weaning children and pregnant and nursing women; to counter iron and folate, iodine and subclinical Vitamin A deficiencies; to institutionalize home management of diarrhea, and to combat infectious disease. 15. Consideration also should be given to four program measures for those at high nutritional risk. First, the Government needs to reconsider the scope and costs of its apparent plans to replace food stamps with JSP transfer payments to food-insecure households. The cost of the proposed JSP consumption component i;- estimated at around Rs 28,000 million per year. JSP eligibility extends well beyond the relatively limited number of around 800,000 families at nutritional risk, according to the 1986-7 Consumer Finance Survey (CPS). A well-targetted food stamp program for them alone would cost less than RS 1,000 million per year. These and other families comprising the poorest quintile of Sri Lanka's population could be brought to middle-class food consumption levels for an estimated Rs 5,600 million yearly, less than one-fifth the annual cost of the proposed JSP consumption corponent. Even with JSP investments, it is likely that some poor families would continue to need food support after two years. A targetted food stamp program based on JSP survey procedures might be a relatively efficient way of meeting those residual needs. 16. Second, introduction of cosmetically inferior grades of rice into the market, if acceptable, would provide cheaper but equally nutritious options to the present range of rice choices for low-income consumers and could prove 2/ Households which spend more than 80X of their income on food but consume less than 80t of daily calorie requirements.  - vii - cost-effective in raising their consumption levels. The cost of grading at the mill could be passed along to other consumers through higher prices for the better-quality varieties. 17 . Third, the thriDosha program should be consolidated and strengthened. Improvements are needed in growth monitoring, beneficiary selection and family use of thriposha against growth failure, which in turn requires &.ore effective nutrition education. 18. Fourth, the Government should consider the development of a community-based program to counter faltering growth among children under three years of age, low birth weight and iron and folate deficiency in pregnant women. Substantial improvements in child nutritional status are likely to result from concentration on a combination of growth monitoring, health interventions, behavioral change and short-term food supplementation. The incremental cost of such a national program would be around RS 140 million. A pilot effort is being developed under the IDA-aided Health and Family Planning Project (Cr. No. 1903-CE). 19. Finally, the Government needs to re-establish a national focal point for food and nutrition planning and advocacy. The Administrative Reforms Committee recommendation to establish a Nutrition Policy Council under the Primie Minister, serviced by a Nutrition Policy Bureau, has considerable merit. If nutrition policy and strategy continues as a National Health Council (NHC) responsibility, then a separate division in MPPI would be the most appropriate locale for either a re-established FNPPD or a successor division to service NHC's nutrition aBende aud carry out technical analysis. Moreover, there are two important tasks which the Government should carry out. First, it should promptly complete analysis of the 1988 national nutrition survey, which (except in the uncovered North and East) is expected to provide the first reliable data set for trend analysis of nutrition status since the last national survey ten years ago. Second, it should undertake detailed analysis of food expenditure and income data in the 1986-7 CFS as a basis for defining optional strategies to meet the nutrition needs of food-insecure households. 20. The above recommendations are mutally reinforcing rather than mutally exclusive. The proposed package of nutrition interventions would yield substantial savings over the cost of existing programs through better targeting of the food stamps and school lunch programs. The relatively modest incremental expenditures on agricultural research, nutrition education and micronutrient interventions would have large pay-offs in terms of nutritional status. In sum, the proposed package would be cost-effective as well as technically and managerially feasible. Its total cost of under than Rs 3,000 million would be less than 402 of current nutrition outlays. Effective implementation would permit Sri Lanka substantially to improve the nutrition standards of its most vulnerable groups at a sustainable cost of less than two percent of the recurrent government budget.  I. NUTRITION CONDITIONS AND CONSEOVENC S Introduction 1.1 A paradoxical situation exists in respect of maternal and child nutrition in Sri Lanka. Overall health conditions have improved substantially over the past two decades. Steady development of social sector services has been accompanied by a progressive increase in the coverage and, in respect of at least some services, the quality of primary health care. High levels of formal education and an effective population policy, leading to a consistent reduction in fertility, have underpinned these advances. As a result, national health indicators are remarkably good in comparison to other countries of similar income levels (see Table 1, below). However, significant variations exist within Sri Lanka; estimated IMR for the escate sector in 1986 was 49.6, almost twice the rate of the rest of the country, while neonatal mortality comprises 60-702 of total infant deaths, suggesting suboptimal maternal health conditions. Table 1.1: COMPARATIVE HEALTH INDICATORS BY GNP LEVEL Country GNP Per capita Life Expect. IMR la CDR /b MHR /c (in US$, 1986) at birth (1986) (1985) per 100,000 (1986) live births (1980-84) F H Ghana 390 56 52 89 11 1,100 Sri Lanka 400 72 68 29 2 90 Mauritania 420 49 45 130 25 119 Senegal 420 49 46 137 27 530 Liberia 460 56 52 87 23 173 Yemen PDR 490 51 49 142 50 100 Indonesia 490 58 55 87 12 800 /a Infant Mortality Rate. lb Child Death Rate. /c Maternal Mortality Rate. Sourcest World Development Report, 1988, 1987. IBRD. The State of the World's Children 1988. UNICEF 1.2 Despite impressive gains in conventional health indicators based on mortality and service coverage, substantial malnutrition persists throughout Sri Lanka. More sensitila than mortality as an indicator of health and social conditions, malnutrition particularly among pre-school children and pregnant women is a serious and unresolved problem.  Child Nutrition Status 1.3 Several national nutrition surveys since 1975 provide a useful basis for assessing conditions and examining trends. Special studies in population subgroups have examined specific problems including anemia, goiter and Vitamin A deficiency. Two recent studies on maternal and child health in the dry zone and on perinatal and neonatal mortality provided further Information confirm- ing existing service statistics 11 1.4 The dry zone study indicated that parity, maternal stature and educa- tion, ante-natal care, place of delivery and maternal tetanus toxoid were strongly associated with variations in perinatal and infant mortality. The mortality study indicated major gaps in home visiting patterns of public health midwives. However, it also fouad that birth attendance by untrained personnel did not have a serious adverse effect on perinatal and neonatal mortality, suggesting that untrained workers use safe delivery practices. 1.5 The most recent national effort was a 1987 Demographic and Health Survey (DBS),21 which included anthropometric assessment of nearly 4,000 children 3-36 months old in all parts of Sri Lanka except the troubled North and East. The results were expressed in terms of stunting measured by height for age and representing chronic and longer-term malnutrition, wastlng v.sasured by weight for height and representing acute, short-term malnutrition and combined stunting and wasting representing a severe degree of past and current nutritien deprivation. 1.6 The national prevalence of stunting was 27.51 with variations among regions from 18.9 to 42.12. It was highest in the estate sector at 602 and lowest in urban populations. Stunting increased cumulatively with age and was highest with birth intervals less than 2 years. It was substantially higher for children of mothers with no education (50.52) tian for those with even primary education (34.32) and lowest for those with more than secondary educa- tion (15.4Z). Many of these variables are highly intercorrelated and indicate that stunting i. most severe in the work force of the estates in the south- central region. 1.7 The prevalence of wasting was 12.92 nationally with regional varia- tions ranging from 9.9 to 16.82. Frequency rose sharply at 12-23 months to 19.32 and was relatively independent of birth intervals. In relation to maternal education, prevalence declined consistently from 15.2Z for no educa- tion to 11.8S for more than secondary education. However, the wasting and stunting figures among the well-educated also are high, suggesting that nutri- tion education is quite weak. 1.8 The prevalence of wasting in children on estates war remarkably low. With a frequency of 7.12 versus 13.62 and 13.42 in the rural population and 1/ Profiles of Women and Infants in the Drv Zone of Sri Lanka (1988). Perinatal and Neonatal Mortality - Some Aspects of Maternal and Child Health in Sri Lanka (1986). Department of Census and StatisticsJUNICEF. 2/ Sri Lanka Demographic and Health Survey 1987. Institute for Resource Development/Westinghouse and Department of Census and Statistics, Ministry of Plan Implementaticn, Government of Sri Lanka, May, 1988.  Col.mbo, respectively, children In the estate sector had about half the rate of wasting but double the rate of stunting of the general population. These figures suggest that health and nutrition conditions In the estate sector were very poor in 1984185 and improved dramatically from the second half of 1986. Increased employment levels and wages in the estate sector occurring shortly before the survey, combinea with progressive improvements In health and sani- tation conditions, might account for such anomalous findings. The alternative explanation is that the severity of stunting was such that the nutritional energy and nutrient neede of these children were reduced to a level consistent with the food bupply available to them within the family. This degree of adaptation to dietary limitations would seem unlikely to occur without other health consequences such as rapid increases in general morbidity and morta- lity. The overall prevalence of combined stunting and wasting, representing acute malnutrition superimposed on chronic effects, was 4.72; no geographic breakdown for this category is ovallable. 1.9 The 1987 survey is not strictly comparable to two earlier island-wide nutrition surveys, in 1975-0 and 1980-2, which sampled children aged 6-59 months. Additionally, the 1980-2 national survey used a different sampling frame and analytic methodology from the one in 1975-6. Therefore, it is difficult to make definitive statements about nutritional trends in Sri Lanka until the results of the 1989 national survey, which will be comparable with the 1980-82 effort, become available. Wasting and stunting among rural and estate children combined in 1987 was 11.31 and 24.92, respectively, against 1980-81 figures of 10.6Z and 24.4S, respectively. Despite age, geographic and other sampling differentials, tbs 1987 DHS findings imply that nutrition conditions are not improving for Sri Lanka's children. 1.10 A receht survey of school children in Colombo Municipality showed the rate of stunting for a random sample of children 7-10 years at around 20.51 and that of wasting 9.31, indicating that the effects of early stunting persist at later ages. The latter rate, indicatiag acute malnutrition, is high for an age group generally considered to be at low risk. Maternal Nutrition Status 1.11 Maternal nutrition conditions in Sri Lanka can only be assessed from birth weight data and from a series of imall studies carried out at different times in various low-income subgronpe. One study reported 23X of women in early pregnancy with height below 150 cm and 241 below 40 kg weight, indicat- ing past and current malnutrition associated with obstetrical risk.3/ Pregnancy weight gain in deprived urban women has been reported ranging from 0-3 kg, against a minimally desirable level of 7 kg. Another study reported a mean weight gain of 4.5 kg associated with mean birth weights of 2.6 kg and low birth weight (LBO) rate of around 301. l In 1986, the national level of L9W was reported to be around 20-251, rising to 332 in the estate sector. For the 448 reported infant deaths in the estate sector in 1986, the average birth weight was 1.9 kg. The average birth weight of infants delivered in major maternity units has declined from 2.9 kg to 2.7 kg in recent years. suggesting that this problem is increasing. 3/ Profiles of Women and Infants, etc., oD. cit. 41 Lassalean Communitv Education Services Arnual ReDort 1985. Colombo.  -4- 1.12 It is estimated that about 802 of reported LBW is attributable to fetal growth retardation, caused principally by maternal malnutrition and malaria; in 1986, this amounted to about 57,500 low-weight births, or around 17? of all live births. Clinical observations indicate that maternal malnu- trition is widespread, at least in the lower socioeconomic strata, and contri- buted to infant morbidity and mortality. The high (60-70Z) proportion of neonatal to all infant deaths supports this conclusion, particularly given the high level of institutional deliveries and low level of neonatal tetanus. Micronutrient Deficiencies 1.13 Anemia is highly prevalent in women of reproductive age although the wide variation of measurements and criteria applied make it difficult to quan- tify the problem precisely. Special studies and service reports indicate prevalences of 50-702 in pregnant women attending antenatal clinics. The estate sector reported 79? low hemoglobin levels in pregnant women for the last reading before delivery; however this may not have been adjusted for normal hemodilution. Anemia, expressed as low hemoglobin levels, is also reported to be prevalent in preschool children, with rates of 60-702 reported, and may relate particularly to intestinal parasitism. Most reports do not discriminate between types of anemia, but macrocytic forms are reported. Low dietary iron, Vitamin C and folic acid intakes, combined with high prevalence of malaria and intestinal parasitism, are the principal causes. 1.14 Vitamin A deficiency existed in the past as a public health problem but now appears to be confined, at least in its clinical forms, to discrete areas and subgroups. Diagnostic criteria for clinical Vitamin A deficiency are frequently imprecise and special training of observers is not often under- taken, making it difficult to quantify the severity of this problem. Of greater importance may be the contribution of subclinical Vitamin A deficiency to morbidity and rortality from infectious disease. Experience in Indonesia indicates that morbidity and mortality from acute respiratory infections and from diarrheal disease are substantially reduced by Vitamin A supplements in populations where deficiency is prevalent. (The Nutrition Department of the Medical Research Institute is planning a study of Vitamin A deficiency in Galle which would include biochemical assessment. This would provide impor- tant information on the prevalence of subclinical states and indicate whether routine megadose supplementation might be effective in reducing infectious disease levels in young children.) 1.15 Iodine deficiency disease (IDD) is a public health problem, particu- larly in Kalutara, Kegalle and Matale districts where a recent regional survey of school children showed 10.52 with prominent goiter (Grade 3) and 21.92 with substantially enlarged thyroid glands (Grade 2); only 352 had normal glands. However, IDD also occurs in a number of other districts including, Galla, Hambantota, Kandy, Moneragala, and Ratnapura. An unusual feature of the problem in Sri Lanka is the prevalence of goiter in coastal communities with access to marine foods which should provide an adequate iodine intake. However, many of the affected populations are reportedly very poor fishing communities who sell all or most of their catch to purchase rice and other staples.  -5- Recent Trends 1.16 Between 1981-2 and 1986-7, food consumption appears to have increased significantly for the very poor, and unvevenly for the next lowest deciles. However, in recent years, underemployment has risen substantially, unemployment In now at an estimated 18-201 o: the work force and around 25Z of the population, concentrated In the rural sector, is below the poverty line. The poorest and most vulnerable groups remain as identified above, with special reference to dry-sone farmers, new settlers and small food-crop farmers, augmented by small fishermen. Significantly, poverty also is reported among segments of plantation workers and in the informal urban sector. Geographically, lt is questionable whether the northern areas, disrupted by continued conflict, have managed to maintain their previously reported comparatively high calorie consumption levels. Additionally, over the past three years food prices have continued to rise faster than wages. The health system and nutrition education have made steady, but not remarkable, progress and no extensive nutrition initiatives have become operational. Coamarisons 1.17 Malnutrition in Sri Lanka, as measured by the 222 of children with wasting between 12-23 months, is relatively low compared to other low-income countries with high Infant and child mortality rates such as Nepal (272), Burma (482) and India (372). lowever, it is substantially higher than In lower mlddle-income countries (98$400-1,600 per capita) with comparably low infant and child death rates, such as Malaysia (61) Indonesia (172), Philippines (141) or Thailand (181). Table 1.2 below, compares Sri Lanka's nutrition Indicators with those of other countries at similar GNP levels and indicates that Sri Lanka's nutrition progress ls much less noteworthy than its health accomplishments. These comparisons underline the conclusion that Sri Lanka has made exceptLonal progress with relatively limited economic resources in reducing mortality and severe disease in infancy and childhood but has not managed to bring about corresponding improvements in nutrition conditions or in controlling morbidity from nonimiunizable infections of childhood. Table 1 2: SRI LANKA, COMPARATIVE CHILD NUTRITION INDICATORS BY GNP LEVEL S children under 5 GNP with mild-moderate/ per capita I low severe malnutrition 2 of wasting (in US$, birth (1980-86 or latest (1980-86 or latest Country 1986) weight available year) available year) Ghana 390 17 2317 28 Sri Lanka 400 20 25/10 22 Mauritania 420 10 30/10 NA Senegal 420 10 20/NA 20 Liberia 460 NA 31/4 NA Yemen PDRY 490 12 32/8 36 Indonesia 490 14 27/3 17 Sources The State of the World's Children, 1988. UNICEF.  - 6- Functional Implications of Malnutrition 1.18 In deciding the appropriate national priority to assign to the control and prevention of Sri Lanka's malnutrition, the functional signifi- cance of its different forms needs to be assessed. The functional effects of maternal malnutrition are substantial and relatively easy to assess. Limited weight gain in pregnancy is closely correlated with fetal growth retardation and low birth weight, which in turn carries a greatly increased risk of serious morbidity and early infant mortality. Such a process is wasteful in biological, psychological, social and economic terms and represents a clear priority for i.tervention. 1.19 Wasting, particularly in the post-weaning period, is associated with decreased immunocompetence resulting in increased frequency and severity of common infections such as diarrheal disease and acute respiratory infections. Stunting represents an ultimately undesirable alaptation to limited dietary intake. The degree to which this affects physical, mental and social perfor- mance cannot easily be defined and will vary with community and family condi- tions. It is certain, however, that the physical and psychological environ- ment in which protein-energy malnutrition develops results in social disadvan- tage, often of a permanent nature. 1.20 Severe anemia affects both physical and mental performance and may reduce the effectiveness of educational inputs in children. It also represents an obstetrical risk in pregnancy and adversely affects fetal growth. At the same time, there is considerable capacity for adaptation to minor decreases in hemoglobin levels, particularly if tlese occur over an extended time period. Thyroid enlargement in endemic goiter is initially a successful adaptation to low levels of iodine intake. As the condition becomes more severe, thyroid function is compromised and mental and physical performance are adversely affected. Maternal iodine deficiency is associated with an increased risk of cretinism in the offspring, which constitutes severe loss of physical, mental and social function. Recent evidence also suggests that subclinical nongoiterous iodine deficiency impairs mental function. Vitamin A deficiency in severe forms causes eye conditions that result in loss of acuity and eventually in blindness. Lesser degrees of deficiency are associated with increased morbidity from infectious diseases. II. NUTRITION DETERMINANTS A. Determinants 2.1 Malnutrition in Sri Lanka results primarily from chronic household food insecurity exacerbated for some population groups by two sets of contri- butory factors: food-related behavioral patterns and morbidity. Household food insecurity arises from low incomes and limited purchasing power, the product of high underemployment and rapidly rising food prices. In urban areas, it reflects job uncertainty and high prices of essential commodities. In rural areas, it reflects landlessness or subsistence agriculture in which families continue to be net purchasers of food.  -7- Household Food Insecurity 2.2 Periodic national surveys of consumer finance clearly demonstrate the relationship between ecenomic status and food intake. Both the Labor Force and Socioeconomic Surveys carried out by the Department of Census and Statistics in the Ministry of Plan Implementation and the Consumer Finance Surveys cerried out by the Central Bank show that purchase of food commodities (or their imputed value if they are home-grown) represents, by a wide margin, the major budgetary expenditure for Sri Lankan families. Food expenditures averaged roughly 652 of total expenditures for the population as a whole in 1986/87, and 70X for the poorest two deciles, figures generally associated with poverty threshholds in low-income countries. There is at the same time substantial variation in the estimated cost of calories consumed, ranging from just over SL Rs 5.4 per 1,000 calories in the poorest decile to just over SL Rs 10 in the highest.5/ This reflects, as would be expected, a substantial increase in expensive protein-rich foods as income increases. Meat consumption, for example, increases thirtyfold over the income spectrum while rice consumption only doubles. These studies also indicate that the food consumption patterns of the poorest deciles are highly sensitive to any changes in income or prices except as these affect rice intake, a dietery given for all families. 2.3 Food consumption for middle- and upper-income groups improved or remained constant between 1969 and 1981/82, but that for the poorest three deciles declined steadily. (Caloric consumption per Adult Equivalent Unit, or AEU, for the poorest decile dropped from 2,156 in 1969/70 to 1,566 in 1981/82 reflecting, in part the 1977 elimination of food subsidies.)6/ Over the same period, the percentage of the population not achieving calorie adequacy rose from 40 to 55? and the percentage of the population consuming less than 1,400 calories per day increased from 4 to 15?. Food intake of the poor probably increased somewhat in the construction and tea boom years which immediately followed as the Government's liberalization policies took hold. The 1985-86 Labor Force and Socio Economic Survey found that the percentage of households not meeting caloric requirements dropped from 55? (1980-1) to 492 and that caloric intake of the poorest decile improved marginally.71 Preliminary analysis of the 1986-87 CFS indicates continued improvement in the caloric intake of the poorest decile. Although comparison of expenditure deciles in the older surveys and income deciles in the more recent ones is problematic, data from 1986-7 would appear to indicate that food consumption levels for some lov incume groups have improved, perhaps to 1969-70 levels, in the case of lowest income estate workers, and certainly have not worsened. There are no 1986-7 figures comparable to those in the 1500-1600 calories per adult equivalent range for the poorest decile which were recorded in 1981-2 (these after a very sharp increase in prices in early 1980). Instead, 1986-7 figures are in the 1800-2000 range for the three poorest urban deciles as shown in 5/ See Annex A, Table 1. 6/ See Annex A, Table 2. 7/ Unfortunately, the available analysis in this survey and the 1986-87 Consumer Finance Survey is not by total expenditure deciles but by less reliable (and probably significantly under-reported) income groups.  - 8 - Table 1.3 below, nearly 2000 for the poorest rural decile, and nutritionally adequate on the estates. (All of these AEU figures would have been up to around 170 calories lower in the absence of food stamps at 1987 entitlement levels. See Annex ., Tables 7-9.) At the same time, however, despite some restructuring of the food stamp program in 1986, these groups (poor urban dwellers, landless laborer families, some poor estate workers and recent migrant families and the households of other underemployed workers in the unorganized sector) have been the disproportionate victims of rising underemployment, deteriorating real wages, steadily eroding value of the food stamps until 1989 and declining economic growth since 1985 and thus remain vulnerable in terms of household food security. Table 1.8: DAILY CALORIE CONSUMPTION PER ADULT EQUIVALENT UNIT Urban Rural Estte 1969-70 /I 1981-82 pj 1988-87 /b 1969-70 /k 1981-82 /k 1966-87 /k 1969-70 /I 1981-82 /1 184-e7 /k 1,848 1,521 1,886 2,188 1,670 1,998 2,209 1,617 2,216 2,192 1,771 1,796 2,482 2,082 2,148 2,691 2,180 2,405 2,861 1,982 1,982 2,760 2,826 2,242 8,009 2,684 2,608 ,I Expenditure Data. /b Income Data. Source: Annex A, Table 2, 1988-87 data based on Consumer Finance Survey 2.4 During 1969-82, the poorest decile increased its total nominal expenditures (a more efficient proxy for income in these surveys) by just over 300o (although this translated into an increase of only SL Rs 75 per person per year). Over the same period, the Colombo consumer price index (CPI) calculated by the Department of Census and Statistics registered a 3842 increase and the cost of a food basket traditionally consumed by low-income households rose by just over 400X. More than half of those increases occurred between 1977-82 while liberalization of the economy was beginning to take shape. The difference of around 33? batween price and income (total expenditure) increases, 1969-82, explains in large part the decreased caloric intake of the poor over that period. However, between 1982-87, the Colombo CPI rose by 57?, while food prices registered a 54? increase, thereby modestly reversing the previous relationship between them. 2.5 Since 1985, the combination of sluggish employment, drought and the disruption caused by communal violence makes it unlikely that real incomes have increased significantly. Annex A, Table 3 indicates that wages for low- skill construction workers have just kept up with food price rises, while agricultural wages fell well behind in 1987. Even in the case of agricultural laborers, real wage rates are unlikely to have increased significantly despite increased paddy production, given high rural unemployment (roughly 15X) and the high percentage of rural wage earners willing to undertake additional work when available.  -9- 2.6 Additionally, since the poor spend a larger proportion of income on food, they are disproportionately affected by seasonal price changes which are masked in annual consumer price indices. Seasonal increases of 502 in the open market price of rice and even greater fluctuations in the prices of other foods have not been unusual in recent years. (In 1986, for example, the open market price of samba rice fluctuated from SL Rs 8 per kilogram in June to SL Rs 12 in January. In 1985, the price of gram rose from SL Rs 30 in January to SL Rs 44 in September, while over the same period the price of coconut oil dropped from SL Rs 19 to SL Rs 9.) There has been little evidence, however, of major regional shortages or regional price differentials. Behavioral Factors 2.7 While household food insecurity represents a principal determinant of malnutrition in Sri Lanka, it is complicated by behavioral patterns, particularly of vulnerable population groups. The nutritional status of a 12-month-old child of a poor family is not likely to be low simply because of family income level (the marginal cost of feeding a young child is low, particularly in rural areas), but rather because of a combination of time and resource constraints aggravated by certain long-held but sometimes deleterious feeding practices. 2.8 Inappropriate dietary behavior occurs throughout the reproductive and growth cycle. Inadequate energy and nutrient intake during pregnancy, particularly the last trimester, is common in low-income households and arises primarily from inappropriate food distribution within the family. Household food insecurity results in priority-to those perceived as being economically most active. However, the concept of economic loss incurred by wasteful reproduction--high mortality in low-birth-weight infants resulting from maternal malnutrition--is nct reflected in existing cultural practices. The failure to adhere to iron and folate supplementation regimes results in low hemoglobin values that further affect fetal growth and increase the hazards associated with obstetrical hemorrhage. After delivery, maternal malnutrition adversely affects the quantity and, if severe, the quality of breast milk, with direct consequences for infant health and survival. 2.9 Initial lactation practices are crucial in delivering the optimal dose of maternal antibody to the newborn through colostrum. The Dry Zone study revealed that 30? of mothers did not feed colostrum, of whom half thought it was harmful. The wide availability of powdered milk in retail outlets and cooperative stores throughout the country (although controlled by a Marketing Code) has resulted in the introduction of such milk feeds at an early age, in part to compensate for actual or perceived lack of breast milk; this in turn reduces the frequency of suckling and the volume of breast milk. This practice leads, paradoxically, to extended reliance on purchased milk as the dietary staple with postponement of the introduction of semisolid foods to as late as one year. As the price of powdered milk is high (SL Rs 35/400 g) in relation to food stamp values (SL Rs 50/child), overdilution is used to extend the supply, thus reducing energy and protein intake. Where working mothers delegate child care to older siblings, the high frequency of feeding required with low-density rice gruels is often neglected, while poor food hygiene results in more frequent intestinal infections. These early feeding practices result in a progressive deficit of energy and nutrients from 4  - 10 - months onwards. Longitudinal growth recording shows a marked leveling off in weight gain from 6 months, with subsequent reduction in height increments. From this position, it is difficult for the child to regain its initial growth momentum, though it may continue to grow adequately at a lower trajectory. The delay in solid food supplementation and inadequate caloric density also may increase malnutrition and susceptibility to disease. An effective nutrition education program could help counter these problems. Infectious Diseases 2.10 Finally, infectious disease and other health factors invariably will have a significant effect on nutritional status in Sri Lanka as elsewhere. Food intake is reduced, metabolism is elevated, absorption is limited, and losses of energy and nutrients can be high. These effects are compouILded when, as is often the case, feeding is not actively maintained dur!4:g the infectious process. The frequency and severity of infectious disease is itself a function of poor environmental sanitation and inadequate food hygiene which result in high levels of intestinal parasitism as well as diarrheal disease. 2.11 Sri Lanka is making substantial progress against the six immunizable childhood diseases. The Expanded Program of Immunization has achieved high coverage levels throughout the country for DPT, BCG and OPV. Measles vaccine was introduced recently and already covers about two thirds of the target group. The recent DHS results showed that health cards were held for 822 of children between 12-36 monthst for these the following coverage levels were recordeds BCG--992, completed DPT--93Z, completed OPV--93Z and measles- -68.5Z. These remarkably high levels of coverage have resulted in effective control of these six diseases and major reductions are recorded in health statistics reports (see Table 2.2). Table 2.1: FREQUENCY OF IMMUNIZABLE DISEASES--1986 (per 100,000 population) Disease 1980 1986 Polio 1.8 0.1 Diphtheria 0.3 0.0 Pertussis 3.7 1.0 Tetanus 8.4 2.1 Tetanus neonatal 81.8 7.0 Tuberculosis /a 42.1 40.9 /a Affects all ages, 0-14 year rate: 5.1/100,000 for 1986. Source: 1986 Annual Health Bulletin, Ministry of Health, Colonibo. 2.12 Major problems exist, however, in respect of the nonimmunizable diseases, particularly acute respiratory infections, diarrheal disease,  - 11 - malaria and intestinal parasitism. All of these conditions have profound effects on the nutritional status of children which in turn influences the frequency and severity of infectious episodes. 2.13 The combination of slow fetal growth, fetal malnutrition and immatu- rity appear to comprise the leading specific cause of childhood hospital deaths in Sri Lanka. Acute respiratory infections are the leading specific cause of hospital death in children <5 years. They represent between 10-302 of child admissions and 30-352 of out-patient attendances. LBW children are at particular risk of death from this cause. Despite heavy emphasis on control through oral rehydration, diarrheal disease is the third leading specific cause of hospital death in children under 5 years. Of all reported diarrheal deaths, 48.5Z are in children <5 years, of which 46Z, or 22? of the total, are in children <1 year. It represents 102 of all hospital admissions of which 502 are for under-fives. At any one time, between 20-30? of pediatric beds are occupied for dehydration. 2.14 Morbidity patterns have remained relatively stable in recent years as gauged by hospital data, the principal source. Respiratory infections are the principal specific cause of hospital admissions both generally and for pre- school children. Diarrhea ranks second for pre-school children and third, just after accidents and injuries, for the general population. Malaria is the fourth leading overall cause of hospital admission. The rate of hospitaliza- tion for malaria rose from 4371100,000 population in 1985, to 830 in 1986 according to Ministry of Health statistics. Service data show that the percentage of positive blood slides rose from 10? in 1985 to 28? in 1986; falciparum positive slides rose from 12 to 5.6? over the same period. Intes- tinal parasitism is acknowledged to be highly prevalent and special studies have reported overall rates for soil-transmitted nematodes (ascaris, hookworm and trichuris) between 50-90Z. Children hospitalized for severe malnutrition have parasite rates 2-3 times that of nonmalnourished admissions. The continuation of these mortality and morbidity levels and patterns indicates that Sri Lanka needs to do more on the disease prevention and control front since all these problems are amenable to public health measures. B. Most Seriously Affected Groups 2.15 Analysis of the 1979-82 national nutrition survey provided a geographic and socioeconomic map of malnutrition in Sri Lanka. Wasting was most widespread in Kurunegala and Puttulam districts, while stunting was most prevalent in Kandy and Nuwara Eliya. Combined wasting and stunting was highest in Kandy and Nuwara Eliya as well as Amparai, Anuradhapura, Batticaloa, Polonnaruwa and Trincomalee districts. 2.16 Calorie intake patterns differed from anthropometric findings. The greatest calorie deficiencies were in Matale and Badulla, where around one third of all households suffered more than a 302 intake deficit against estab- lished norms. In Hambantota and Moneragala districts combined, and in Galle and Matara districts combined, around 30? of all households suffered a similar calorie shortfall. Only in the northern districts of Jaffna, Vavuniya, Mannar and Mullaitivu were less than 11? of all households so affected. Higher prevalence of malnutrition in areas of higher caloric intake imply the presence of high rates of infection, intrafamily food distribution problems on both.  - 12 - 2.17 Agricultural and animal husbandry worker households appeared to have the lowest calorie intakes, mainly on estates and rural farms. Laborer fami- lies appeared to be almost as badly.off from a nutritional perspective; those in urban areas and the urban unemployed are at particularly high risk. Seri- ous regional differences prevent the postulation of a universally well-defined relationship between land-holding and nutritional risk, but small holdings and landlessness, as expected, seem to be associated with poor nutritional performance. 2.18 Biological factors associated with reproduction, growth, development and physical work increase the demand for energy and nutrients and determine the form in which they can most effectively be assimilated. In addition, environmental conditions and behavioral factors associated with low income and poverty increase the risk of infectious disease which further compounds economic and biological influences. As a result, low-income pregnant women, lactating mothers and preschool children are the most severely affected. 2.19 Protein-energy malnutrition is frequently accompanied by iron, folate and Vitamin A deficiency. Seasonal effects also operate. The high energy requirements of planting and harvesting in unmechanized agriculture increase demands on all economically active members of the rural family. Thus women reaching the third trimester of pregnancy at planting or harvesting are at additional risk of malnutrition and of producing a low-birth-weight infant with seriously diminished chances of survival. Other seasonal factors such as diarrhea incidence and malaria prevalence add further risk. Iodine deficiency, being primarily determined by geography, affects a wider range of socioeconomic levels although reproductive women carry additional risk by reason of increased requirements. Districts of highest prevalence are Kandy, Nuwera Eliya, Kalutara, Ratnapura and Kegalle. 2.20 Sri Lanka's nutrition and health conditions consistently have been worse in the estate sector, except for the remarkably low levels of acute child malnutrition reported in the 1987 DHS. Rural populations have better nutrition levels than estates, in general, but worse than urban populations. Within urban groups, Colombo has higher nutrition levels than other cities. 2.21 Females are marginally more malnourished than males but differences are much less pronounced than elsewhere in Asia. Malnutrition in women, including anemia, is most pronounced during lactation and pregnsncy, when its greatest impact is on the fetus during the last trimester of gestation. 2.22 Stunting increases progressively with age in childhood; acute malnu- trition is significantly higher following weaning in the second year of life. Child nutrition is in transition; severe manifestations have been relatively well-controlled but conditions remain suboptimal.  - 13 - III. NUTRITION IN SRI LANKA'S FOOD AND AGRICULTURE SYSTEM A. Dietary Patterns 3.1 The most basic food items in the Sri Lankan diet are rice and coconut (including coconut oil) which also are che most important domestic agricul- tural commodities. These, along with bread (most important in urban areas), wheat flour (more important among Tamil workers on the estates) and sugar, constitute 802 of calories in the Sri Lankan diet. (Rice itself contributes roughly 502 and coconut products roughly 202 on average.) For the poorest three deciles, these commodities plus yams and cassava constitute almost 902 of the calories consumed. 3.2 Of the major foods consumed, Sri Lanka imports wheat, sugar and a small proportion of its rice. The country is approaching self-sufficiency in rice but from the standpoint of effective demand rather than nutritional need. Looking at consumption trends over time, up to 1985-86 there has been in urban areas an increase in meat and dairy products among all income groups. In rural areas, there was a slight increase in rice consumption up to 1981. By 1986-87, rice consumption increased in absolute terms for the poorest deciles, but decreased somewhat as a percentage of total calories reflecting continued dietary diversification.8/ In the estate areas, the rice-wheat balance has been shifting in favor of rice although wheat, used in the making of chapattis and dosas mainly by the Tamil population, still represents 172 of total calories. Sugar consumption has risen sharply. Finally, bread is now consumed in some amount by 902 of families in the country. 3.3 Unlike many other low-income countries, there are in Sri Lanka no major food staples disproportionately consumed by low-income groups which might lend themselves to policy intervention addressing nutritional needs of the poor. Rice is the major component of the diet for every income group, and virtually all of the subsidiary food crops, as in the case of rice, are consumed in greater quantities by upper-income groups. Accordingly, the types of income-specific elasticity analysis for a range of food staples which has helped identify intervention points in other countries may be less useful in Sri Lanka. B. The Role of Agriculture 3.4 Paddy production increased by close to 72 p.a., 1959-85, before slowing,9/ but production in the export-oriented tree crop sector--tea, rubber and ck onut--has been growing only fractionally. Tea and rubber production has actually been declining.101 3.5 The agricultural sector has been a relatively efficient absorber of labor in the economy with a growth rate in agricultural employment of 1.52 8/ See Annex A, Tables 4 and 7. 9/ See Annex A, Table 5. lo/ With the recent drought, coconut production also declined.  - 14 - p.a. between 1963 and 1980, although agriculture as a percentage of total employment has been slowly declining (from 52.92 to 45.92, 1963-86).11/ This has not been sufficient, however, to solve the country's chronic un- and underemployment problems. During that period, unemployment fluctuated between 12 and 252, and now stands at an estimated 18-20Z with around 25Z of the population below the poverty line. 3.6 With the change of government in 1977, government expenditures in agriculture increased from less than 102 of total expenditures to over 20X. The greatest expenditure has been heavy capital investment in irrigation, most_y in the Mahaweli region; also a subsidy on fertilizer (612 of which is used in the paddy sector), producer price supports and credit. Rice Development 3.7 In terms of policy attention and resource allocation, the govern- ment's primary emphasis within the agricultural sector has been on paddy production. Since independence, self-sufficiency in rice has been a priority national goal, and paddy farmers have been an important political constituency of each government. The paddy sector has benefited from a range of policy measures including irrigation, price support and subsidized fertilizer. There also is an impressive agricultural research infrastructure (some locally developed varieties are equal to the best imported varieties) and the bulk of agricultural extension outside of the tree sector is devoted to paddy. All of this has been justified as a way to achieve: (a) higher incomes for paddy farmers; (b) increased self-sufficiency in rice; and (c) improvements in the balance of payments (i.e. elimination of remaining rice imports). However, paddy research has focussed more on increasing production rather than on the efficiency and costs of rice production and Sri Lanka rem-ins a relatively high-cost paddy producer, particularly when quality is taken into account. Subsidiary Food Crops 3.8 Successive governments in Sri Lanka have had fairly well-defined strategies in the area of paddy production, but this has not been the case with the so-called subsidiary food crops (coarse grains, grain legumes, oil- seeds, chillies, onions, yams and tubers). These crops, produced in highland areas of the wet zone (southwestern part of the country) under the so-called chena (slash-and-burn) system of agriculture, have registered significant production increases over the past two decades, but mainly as a result of area expansion. There has been virtually no increase in yields, a reflection of relatively little policy and resource attention devoted to these crops. 3.9 To the extent that policy attention has been focused on subsidiary food crops, it has not necessarily been contsistent. Domestic production of these crops was protected by import restrictions which were lifted in the late 1970s. In addition, producers of these crops have received neither priority in the provision of credit nor new seed varieties. Moreover, increasing amounts of unirrigated highland land appropriate for subsidiary food crops have been converted into paddy land. With more serious recent discussion of 11/ Source: Thorbecke, E. and Svegnar, J. 'Effects of Macroeconomic Policies on Agricultural Performance in Sri Lanka, 1960-82.0 OECD, Paris, July 1985.  - 15 _ agricultural diversification, the government has sought to stimulate produc- tion by instituting support prices for many cf these crops. In at least one instance (maize in 1987), however, the support price lad to large-scale procurement by the government which it then was unable to sell as feed because of the presence of lower-priced feed imports. Food Production and Marketina Patterns 3.10 The importance of subsidiary food crops to the low-income population is not so much in their consumption but rather in their marketed production which is concentrated on smaller farms. The smallest quartile of landhold- ings, in fact, is located almost entirely in highland areas where such produc- tion is concentrated. Accordingly, the burden of minimal policy and program- matic attention to these crops has fallen mainly on small and marginal farmers, for whom increased production probably would have a positive income effect despite still unresolved questions of how to market significantly higher quantities at home and abroad. 3.11 The case of rice in Sri Lanka is different. Paddy proauction takes place on both large and small holdings, but the major effect of the rice economy on the poorest deciles of the population is through rice consumption rather than income generation. This has important nutritional and public policy implications requiring some examination. 3.12 Of land-holding families (over 80Z of households in rural areas), between 42 and 541 of holdings (depending on the data source used) are less than one acre. This percentage is increasing as the average size of holding erodes steadily over time. (The average size of a holding in 1982 was 402 smaller than in 1946). According to reliable information, the larger the holding, the larger the proportion of that holding that is likely to be devoted to paddy production. (Yet even among the larger landowners with holdings exceeding 2 acres, only 402 of land on the average is in paddy production, indicating more existing crop diversity than is sometimes recog- nized. Less than 22 of all holdings are exclusively paddy land.) 3.13 As expected, the larger the landholding the larger the proportion of the marketed crop. (Among the poorest quartile of paddy producers, only 8Z sell any paddy.) The strong correlation between size of landholding and the percentage of household income emanating from production on that land is unusual.12/ Among small landholders, who invariably have off-farm employment, less than 52 of reported income (or, more often, imputed income) is the result of their own agricultural production. The comparable figure for large land- owners is almost 5O0. Only a quarter of the smallest landowners report agri- culture as their primary occupation. Accordingly, incomes of these small producers are likely to be proportionately less affected by agricultural pricing (see paras. 3.15-17, below). 3.14 Clearly most households grow food primarily for their own consump- tion. Analysis indicates, however, that most families which sell rice at the time of harvest (because of the need for cash or because of inadequate on-farm 12/ See Annex A, Table 6.  - 16 _ storage capacity) will buy back later in the year at least part of the quan- tity sold, and at substantially higher prices. Fully 84t 13, of rural house- holds in the country are net consumers of rice (purchasing more rice than they sell). Only when annual paddy production reaches 90 bushels (requiring around 2.5 acres of paddy) is a household virtually certain to be a net seller. Imnact of Prices 3.15 From the above, two tentative conclusions with respect to the effects of rice prices are possible. These do not take Into account any indirect effects of increased rlce production on overall rural employment and wages or on production or consumption levels of other crops. (a) Assuming no significavt production decline, lower rice prices will directly benefit the &ajority (two thirds) of rural households which do not produce paddy as well as the urban and estate population. (b) Of rice-producing households, on the foregoing assumptions, those with smaller holdings who are likely to be at substantial nutritional risk and purchase more rice than they sell (and also tend to have their major employment outside of agriculture) are likely to be net beneficiaries from lower rice prices. 3.16 Analysis estimates that a 201 decline in rice prices would lead to roughly a 1OX Increase in caloric intake for the poorest decile.l41 This boost in caloric Intake would, in turn, reduce the percentage of poorest decile households consuming less than 802 of caloric needs (using the FAOIVHO Recom- mended Dietary Intake figures) from 751 to 571. For the second poorest decile, the percentage falling below the 80S level would be reduced by half (from 48 to 241). These calculations assume zero or minimal effects on productionlconsumption levels of other goods. 3.17 Conversely, among paddy-producing households, a 201 increase in rice prices would result In caloric intake losses for all but the quartile with the largest holdings, under the above substitution assumption. For the majotity of paddy-producing households, increased income from higher producer prices will not compensate for the higher prices they have to pay as consumers.151 Higher prices thus result In a net transfer of income from smaller to larger producers. Non-paddy producing consumers in the lower income deciles would suffer s4$ilar declines in rice intake. In geueral, increased rice prices will tend to hurt the poor even when they are paddy producers. Procurement and Distribution Policies 3.18 Government procurement and food distribution policies also impact on nutrition by affecting both supply and demand. Producer price support for 13/ Source: Sahn, David. PFood Consumption Patterns and Parameters in Sri Lankas The Census and Contrcl of Malnutrition. Draft 1. International Food Policy Research Institute, Washington, D.C., June 1985. 14/ Sahn, OD ct 15/ Ibid.  - 17 - rice was initiated in 1948 by the Department of Agrarian Services, then trans- ferred in 1971 to the Paddy Marketing Board (PMB) with actual procurement carried out by Multipurpose Cooperative Societies. Of total rice produced in the country (1.8 million tons or roughly 100 million bushels in 1987), only about 5? was procured by the government. This compares with procuLement as high as 40 to 50? during the peak of the rice-rationing program in the 1970s. The bulk of rice consumption in the country now is dependent on market sup- plies provided by private traders. 3.19 The PMB support price for paddy, presently SL Rs 80 per bushel, is between 10 and 15? lower than the average market price which, in turn, has averaged 15? above the border price, 1983-87. Accordingly, the PMI serves as the buyer of last resort. Its limited purchases are concentrated at the peak of the season when farm-gate prices are at their lowest. As a percentage of total ptoduction, this procurement is less than 52 of what it was at the peak of the rice rationing program. In addition to providing this floor price, the government (through the Food Commissioner's Department) also has the capacity to ensure ceiling prices by injecting supplies onto the market during periods of supply shortfalls and rising prices. With increased domestic rice supply in recent years, however, there has been less frequent need for this ceiling price imposition. Resolving the question of why the market price of rice remains high despite producer subsidies requires analysis of: (a) the extent to which government price intervention poliVies may have distorted the market and (b) what would happen to paddy production if the government permitted market forces to determine prices and output. Processing 3.20 The PMB arranges milling through its own mills and the private sector and then provides the rice to the Food Commissioner's Department which, in turn, makes this rice and imported rice available to the cooperatives for distribution including the food stamp program and to authorized dealers. 3.21 The Sri Lankan rice milling industry at present is a low-technology industry producing low-quality rice with high wastage. Of 27 mills operated by the PMB, 13 are over 20 years old, and 40Z of overall operating capacity is currently disfunctional. Among other problems, most mills lack precleaners and mechanical dryers. Most of the.roughly 750 private mills also are old and in need of modernization, a process which the mills have resisted for reasons which are not entirely clear. One of the earlier factors was uncertainty about government policy, while generally unsettled cor.!itions have dampened the recent investment climate. Consideration is being given to incentives, including the waiving of import duties, to permit necessary renovation. If savings are significant in terms of loss reductions and could lower the market price of rice and reduce foreign exchange requirements, moderrization may well be justified, particularly in the private sector which accounts for most milling. However, labor-displacing effects of modern processing technologies also would have to be taken into account. 3.22 Milling of imported wheat in the country is done in Trincomalee by Prima Flour Mills of Singapore. Its current 20-year contract with the Gov- erinent of Sri Lanka involves the milling and provision of 74? extraction white flour in exchange for the fiber and Vitamin B-rich outer portions of the  - 18 - wheat kernels which are then exported by Prima. Much of the wheat is provided on concessional terms by the United States, Canada (which has imposed restric- tions on export of the by-products) and, to a lesser extent, Australia. Nutrition Effects 3.23 The agriculture sector influences the nutrition of Sri Lanka house- holds in ways which are important but difficult to quantify with precision. First, rice prices directly affect consumption by low-income groups. Second, production and prices of subsidiary food crops affect the incomes of poor farmers. The foregoing points to the need to find ways of increasing the efficiency of rice production; lowering the cost of production can have the twin benefits of lowering the market price of rice without hurting producer incomes. A stronger push for development of subsidiary food crops could raise incomes and purchasing power of small producers while increasing the supply of low-cost calories and animal feeds. 3.24 The choice of post-harvest technology also has collateral impact on food consumption by the poor both as consumers benefitting from efficiencies or losing out if mechanized processes displace labor without creation of compensatory job opportunities. In any case, these production and consumption effects are likely to be more important than the impact of components in agriculture projects which seek to address nutritional deficiencies.16/ Nevertheless, projects which seek, for example, to increase the proportion of green leafy vegetables in home gardens to boost iron intake and decrease the incidence of iron-deficiency anemia can have a positive effect if well-imple- mented. Perhaps more importantly, they can link agricultural extension services to the needs of low-income families. IV. NUTRITION PLANNING AND PROGRAMMING 4.1 Sri Lanka is one of relatively few developing countries explicitly to try to integrate nutrition into national development plans and to set up several Government mechanisms ior that purpose. Sri Lanka has had some success in articulating both important features of a national nutrition policy and selected measures to carry it out. However, the links between policy and strategy elements need strengthening and efforts to translate strategy into effective programs have been uneven. A. Institutional Arrangements 4.2 The apex political body for nutrition concerns is the National Health Council (NHC), currently chaired by.the President, which consists of the Cabinet. Set up in 1980, the NHC's main mandate is to consider policy and 16/ Sahn's analysis, discussed earlier, for example, while extremely helpful in sorting out the rice production/consumption dynamic, does not attempt to include the indirect effect of rice price changes on employment and wagen or the production and substitution affects necessary for such quantification.  - 19 - strategy issues relating to health. It meets an average of at least twice monthly and nutrition is frequently a major agenda topic. When he was Prime Minister, the President emphasized the need to develop tregeted intervention programs against a perceived deterioration in nutrition conditions. A National Health Development Committee presided over by the Secretary of Health acts as NHC secretariat. It consists of the secretaries of Agriculture, Ayurveda, Education and heads of relevant departments. Other mechanisms to address nutrition issues also have operated in Sri Lanka but have yet to be reactivated under the present Government. A National Food and Nutrition Coordinating Committee (NFNCC), chaired by the Minister of Agricultural Development and Research, was set up to coordinate responsibility for food and nutrition policy but later was supplanted by a number of technical committees chaired by senior administrators. These included a standing committee on nutrition considerations in development sectors and committees for thriposha, food stamps and nutrition education and communications. A National Agricultural Pricing and Food Policy Committee had oversight responsibility to the Cabinet for food policy concerns including availability, imports, pricing, procurement, distribution and buffer stocks. Chaired by the Cabinet Secretary, membership comprised the Secretary to the President and the secretaries of Agricultural Development and Research, Finance and Planning, Food and Cooperatives and Plan Implementation (MOPI). 4.3 Until recently, explicit responsibility for promoting the inclusion of nutrition considerations into food and other national planning efforts, interministerial nutrition coordination and the incorporation of nutrition components into ministerial action programs rested with the Food and Nutrition Policy Planning Division (FNPPD) of MOPI. However, FNPPD recently was dismantled through MOPI incorporation into the new Ministry of Policy Planning and Implementation (MPPI), leaving a vacuum which needs to be filled because of FNPPD's important and unique role. Over its 12 years of operation, FNPPD took the lead in setting the national nutrition agenda by performing or commissioning key analytic tasks and translating the results into policy and strategy frameworks. It also promoted the establishment of district food and nutrition committees staffed with full-time MOPI personnel, to develop, plan and fund local nutrition initiatives. Chaired by either the District Minister or the Government Agent (GA) as his nominee, the district committees met quarterly. A similar structure operated at the sub-district (divisional) level, chaired by the Assistant GA, and was to be established in around 6,700 village councils (gramodava mandalavas). Where they were tried out, the committees were unevenly effective, reflecting the varying skill and commitment levels by local FNPPD staff and higher-level officials' perception of district priorities. FNPPD also serviced the technical committees developed by NFNCC. Moreover, FNPPD managed a diverse portfolio of nutrition activities including the kola kenda program, nutrition education and communications as well as the forthcoming National Nutrition Survey. FNPPD thus served as both a focal point for national and district nutrition planning, programming and advocacy and the Government's prime technical resource on nutrition. 4.4 However, as :ndicated in paras. 4.7-4.25 below, the previous combina- tion of institutional arrangements was both too complex and not particularly effective in getting key ministries-and other actors to agree on and carry out a coherent national nutrition strategy. Recognizing the need to simplify and  - 20 - unify that intersectoral responsibility at a high political level, the Administrative Reforms Committee (ARC) recommended establishment of a National Nutrition Policy Council to be chaired by the Prime Minister. The Council would be the focal point for formulation and monitoring of national nutrition policy. It would consist of representatives of all concerned ministries, special interest groups and--essential under administrative decentralization- -the Provincial Councils. A Nutrition Policy Bureau, presumably the successor to FNPPD, would service the Council from the Prime Minister's Office. If broadened to include similar responsibilities for national nutrition strategy, the ARC recommendation is a promising approach to the thorny problem of intersectoral nutrition coordination. However, the Government may decide that NuC should continue to set and monitor the national nutrition agenda. If so, a separate technical division is needed to help NHC address nutrition issues in sectors other than health and to carry out technical analysis for the Council. MPPI would be a logical home for such a division. The Government also needs to carry out two important analytic tasks for nutrition strategy planning. First it should complete the analysis of the 1989 national nutrition survey, which will provide the first reliable assessment of widespread nutrition trends over the last ten years. Second, it should analyze food expenditure and income data from the 1986-7 CFS. Both combined will provide a key basis for defining optional strategies to address nutrition needs of food-insecure households. B. Nutrition Objectives. Policies and Strategies Obiectives 4.5 Sri Lanka's nutrition objectives were enunciated in both an FNPPD memorandum on national nutritional status introduced in July, 1980, by the President and approved by the Cabinet, and a second memorandum issued by the President two years later. Those objectives are to reduce: (a) mortality and nutrition-related morbidity among children under four years of age (b) the incidence of chronic and/or acute protein-energy under-nutrition among infants and pre-schoolers, with special emphasis on districtsl sectors where incidence is higher than the national average (c) the prevalence of maternal under-nutrition, thereby reducing the number of low birth weight babies and enhancing the mother's capacity to breast-feed, provide better child care and actively participate in family and community welfare activities, and (d) the incidence of nutritional anemia among infants, pre-schoolers, pregnant women and women of child-bearing age. 4.6 The Cabinet memo also identified six equally ambitious planning or intervention areas as the main paths to those nutrition goals: (a) formulation of a food and agriculture strategy based on food produc- tion, consumption and nutrition needs  - 21 - (b) strong emphasis on income distribution in all development programs and activities to sustain low income groups and improve their share of real income (c) maximization of essential food distribution schemes and nutrition services to..target groups..until development (enables them] to meet their specific needs (d) continuous review and monitoring of the entire pricing structure..to promote the production of low-cost food staples and the development of cheap nutrition-oriented food technologies to sustain..low-income groups (e) the inclusion of suitable nutrition components in all development planning policies as a pre-requisite..to maintain satisfactory nutri- tional levels and (f) strengthened intersectoral coordination at central and local govern- ment levels through appropriate infrastructure development so that an integrated nutrition package involving population, food production, nutrition, sanitation and primary health services could be better implemented. Nutrition and Related Food and Agriculture Strategy Development 4.7 In late 1983, FNPPD wove a number of important nutrition-related findings into a status report and a set of short, medium and longer-term strategy recommendations. The report drew on the earlier memos to Cabinet, comparative results of national 1979-82 district nutrition surveys of pre- school children against 1975/76 survey data, evaluations of major intervention programs including food stamps and thriposha, and an analysis of changes in food availability, as well as per capita consumption and wages in relation to calorie needs. The report stated that 1980/81 calorie consumption of the lowest three income deciles had fallen below 1969/70 levels although overall per capita calorie consumption had recovered after dipping in the mid '70s. The percentage of households with inadequate caloric intake also had risen significantly between 1969/70 and 1980/81: Table 4.1: SRI LANKA: Z OF HOUSEHOLDS WITH INADEQUATE CALORIE INTAKE, 1969/70 AND 1980/81 Sector 1969/70 1980/81 2 change Rural 32 43 34 Urban 30 50 67 Estate 20 33 65 Source: Nutrition Strategy. Ministry of Plan Implementation, Colombo, 1984.  - 22 - The report pointed out that apolicy measures aimed at improving..socio- economic conditions of the masses should not ignore the short-term need for targeting resources...to specific segments of the population considered to be..malnourished or at risk or belonging to the real poverty groups." 4.8 The resulting short-term strategy called for: (a) enhancement of food and income supplementation through -- higher wage rates or higher agricultural prices to producers -- increased value of food stamps, indexed to prices and based on per capita rather than household incomes, with higher income cut-offs, administrative improvements and more careful targeting -- stabilization of open market food prices to protect the poor against inflation, and -- review and monitoring of food aid to ensure proper beneficiary coverage, nutritional adequacy and management (b) expansion of supplementary feeding programs, particularly -- thriposha. with emphasis on greater outreach, coverage and delivery system efficiency, and -- school biscuits, especially for those in th.e first three years of school (c) promotional campaigns for breast-feeding, sound weaning practices and use of low-cost nutritious foods, and (d) expansion of maternal and child health programs, including clinics, special field programs for pre-schoolers such as early detection of nutrition deficiencies, on-site feeding, immunization and treatment of minor ailments through voluntary and paramedical personnel. 4.9 The main medium and longer-term strategy recommendations were: (a) development of an agriculture and food strategy for supply of low- cost food staples to low-income groups, with special attention to production of low-cost protein rich crops against nutritionally- inferior, higher-cost commodities (b) research and development towards production of low-cost weaning and complementary infant foods based on indigenous raw materials and development of low-cost food technologies (c) exploration of the feasibility of fortifying suitable staples with soya, iodine, minerals, iron and Vitamin A (d) increased investment in development of social infrastructure such as water supply, housing and sanitation  - 23 - (e) development of a national nutrition education policy for implementa- tion through the formal education system, standardized training programs for planners, educators, administrators and extension workers and expanded interpersonal and mass media programs coordina- ted by FNPPD and planned and executed by appropriate agencies including the Ministry of Information and Broadcasting. 4.10 In 1984, the National Planning Division of the Ministry of Finance and Planning drafted a National Agricu."ture, Food and Nutrition Strategy (NAFNS) drawing on the work of FNPPD and seven interministerial task forces: crop agriculture, smallholder tea and rubber, coconut, fisheries, livestock, land and water resources development and nutrition. The explicit nutrition content of the strategy focussed on two main themes: expansion of agricultural employment and food supplementation for vulnerable populations, principally through the food stamps program. However, other aspects of the draft strategy also could have positive nutrition implications. Improved marketing could benefit lower-income consumers to some extent. Research could lead to increased yields of manioc, corn and other subsidiary field crops in areas where rice production is uncertain. Efforts to lower the production cost of rice could benefit low-income groups. Conversely, the recommendation for continued subsidies to stimulate sugar production could work against the poorest deciles, for whom sugar is a significant calorie source, if domestic prices thereby remained well above border prices. 4.11 Later that year, MOPI issued a national nutrition strategy document which again suggested that the food consumption situation had deteriorated between 1969/70 and 1980/81 for lower income groups. Income distribution had worsened, around 55? of households failed to meet calorie adequacy against 40Z earlier, and the budgeted family food share had risen to around 702 of house- hold income against 52Z ten years earlier. It identified Badulla, Hambantota, Kandy, Moneragala and Nuwara Eliya as highest priority districts for nutrition intervention, based on the 1979-82 national survey, with Galle, Kegalle, Matara and Ratnapura grouped as the next highest. The strategy repeated earlier emphasis on generating income for the poor through agriculture, parti- cularly in Galle, Kegalle,Matara and Ratnapura, but effectively also in other districts such as Badulla, Hambantota, Kurunegala, Matale, Moneragala and Puttalam. Where agriculture would be a less effective instrument (Anuradhapura, Amparai, Batticaloa, Colombo, Gampaha, Kalutara, Jaffna, Kandy, Mannar, Mullaitivu, Polonnaruwa and Trincomalee), food supplementation would be the intervention of choice. National production of legumes, pulses and soybean should be encouraged to supplement and, over the long term, partially substitute for rice. The strategy also called for efficiency improvements in the food stamps program to reduce leakago, while ensuring maintenance of the real value of the program for those at greatest nutritional risk. 4.12 During this same period, IFPRI (see para 5.6) contributed to the strategy dialogue by collaborating with FNPPD to analyze Sri Lankan food consumption patterns and the workings of the food stamp program. The main conclusions of that extensive and careful analysis 17/ confirmed the nutritionally precarious and worsening situation reported by FNPPD. In conducting its food consumption work, IFPRI paid attention to particularly at- risk group--the "ultra-poor", families who spend more than 80Z of income for 17/ Sahn. op. cit.  - 24 - food but achieve less than 80 calorie adequacy by accepted international standards.18/ Around an estimated eight percent of Sri Lanka households, or a conservatively estimated 1.3 million persons on the basis of present population figures, would be in that group. At least an additional two million moderately poor persons suffer serious calorie shortfalls on the basis of IFPRI's estimates, also mainly in the lowest three expenditure deciles. According to IFPRI's analysis, the families of two occupational groups, comprising over half the households in the lowest two income deciles and 302 of all households, are at greatest risk. They are non-classified laborers, who make up 30? of the households in the lowest two income deciles, and 12? of all households, and agricutbural and animal husbandry workers, who are the principal earners in over 202 of households in the lowest two deciles and around 17? of Sri Lanka's total households. 4.13 IFPRI's main conclusions were that a combination of growth emphasiz- ing redistribution to the lowest income deciles and moderation in rice prices was key to eliminating malnutrition in Sri Lanka. It also recommended indexing the value of food stamps to either rice prices or cost-of-living changes in view of the sensitivity of the poor to even modest price and income shifts. Regarding other food commodities, IFPRI recommended considering ways to raise economically viable production and subsidize or lower the consumer cost of foods eaten mainly by the poor, e.a. yams, cassava or other roots, tubers and coarse grains. The report pointed out that most agricultural research had focussed on increasing rice production at the expense of other commodities and, for example, had failed to explore adequately the potential role of coarse grains such as corn, millet and sorghum as rice substitutes and income sources on marginal lands for the poor. 4.14 In February, 1985, the NFNCC held a special meeting to discuss the NAFNS and its individual strategy papers. A number of recommendations emerged from that discussion which did not significantly reflect either the MOPI strategy paper or IFPRI's findings, but later found their way into an NAFNS action plan (see para. 4.23), indicating a lack of effective coordination between the NFNCC and FNPPD. 4.15 The MOPI nutrition strategy document, IFPRI's reports and the NAFNS all agreed that both targetting improvements and maintenance of the real value of the food stamps program were necessary. However, on the agriculture front, differences in approach on key agricultural issues were apparent. These included relative emphasis on calorie, vs. protein consumption, higher producer vs. lower consumer prices for rice and sugar and the role of subsi- diary food crops. The next stage of the process--translating strategies into action plans and programs--also continued to reflect the differences rather than effective coordination among the key institutional actors. National Nutrition Action Plan 4.16 In May, 1986, FNPPD in collaboration with relevant agencies prepared a comprehensive food and nutrition action plan based on the 1984 NAFNS (see para. 4.10) and MOPI (see para. 4.11) nutrition strategy documents, as well as 18/ See Annex B.  - 25 - IF,4RI's findings. The plan effectively reviewed the nutrition situation and its policy implications in the context of each of six key sectors, and proposed short, medium and longer-term actions to meet the nutrition and related socio-economic needs of target groups over the next 10 years. The six sectors were: food, trade and shipping; agriculture and livestock; fisheries; health: social services and plantations; education and communication, and (NGOs). The principal objectives of the action plan were to reduce: - chronic undernutrition by at least 7.5? in the short- to-medium term (3-5 years) and by a further 5-7.5X in the medium- to-long term (5-1u years); - acute undernutrition by 5? in the short-to-medium term and by a further 2.52 in the long term; - iron deficiency by 20-30? particularly among children, mothers and the work force over the next 15 years; and - goiter incidence in the major goiter pockets to around 2.5?. Other objectives were to contain levels of Vitamin A and B deficiencies to the present level of under 2? and to increase the purchasing power of the poor. The national steering committee for the plan was chaired by the Secretary, MOPI. The committee and support staff included senior representatives of Finance and Planning, Agriculture, Ayurveda, Education, Fisheries, Health, Mahaweli, Rural Industrial Development, Social Services, Trade, the plantation sector, the Central Bank and non-governmental agencies as well as nutrition and communication experts. 4.17 The action plan called for reorientation of sectoral and subsectoral plans in each concerned ministry to take account of consumption and health factors affecting nutritional status. It recommended general actions cutting across individual development sectors as well as specific measures within each. Overall recommendations over the next five years were made at both district and national levels. The plan recommended strengthening district food and nutrition planning by upgraded staffing and training, increased plan- ning budgets and better coordination with population and health infrastruc- ture. Nationally the plan urged greater use of available human resources in line ministries for nutrition education, adoption of a lead role by the Women's Affairs ministry in regard to the role of women in nutrition and improved coordination within the NFNCC and between it and the NHC. The most significant longer-term overall recommendation was for the establishment of nutrition considerations, including incorporation of a specific and budgeted nutrition subsector with projects and activities, into all relevant sectoral planning programs, and the establishment of nutrition cells in line ministries for that purpose. The plan also recommended that FNPPD take responsibility for setting up, managing and coordinating a nutrition surveillance system to be implemented in coordination with agencies including the Ministry of Health, the Department of Census and Statistics, the Central Bank and the Registrar- General's Department.  - 26 - 4.18 Among the proposed agricultural initiatives were two with particu- larly important potential benefits to low-income groups: improving the effi- ciency of rice production through better post-harvest technology (which the NAFNS had recommended) and developing the subsidiary food crops sector, with particular emphasis on low-cost starchy staples (e.2. yams, cassava) and coarse grains, pulses, legumes and oilseeds. In addition to improving profit- ability, improved rice processing could result in savings to consumers. Particularly in vulnerable areas of Moneragala, northern Badulla, Batticaloa and Amparai, subsidiary food crops including maize and cassava are important cash and seasonal food crops, respectively, for both low-income producers and consumers. 4.19 Other noteworthy agricultural recommendations were to reorient research priorities to the nutrition needs of the lower income deciles, to explore food fortification, including soy-fortified flour as a possible food stamp commodity, and to promote home Sardens. On the marketing side, the possibility of using Agrarian Service Centers to sell a select basket of cheap nutritious foods for food stamps was suggest-d. To carry out nutrition exten- sion and related activities more effectively, reorganization and diversifica- tion of the Home Economics Extension Division of the Ministry of Agriculture also was recommended. 4.20 The action plan called over the next five years for the health sector to "accept a major responsibility' regarding prevention of malnutrition through promotion of family health services and expanded nutrition education, emphasizing improved worker-client interaction; household rehabilitation of malncurished children through home visiting and monitoring, promotion of improved weaning practices, and nutrition manpower development. Over the longer-term, the establishment of a national nutrition surveillance system for at-risk groups was proposed along with health-related nutrition research and defining a nutrition role for the ayurvedic sector, particularly in regard to growth monitoring, supplementary feeding, and immunization. 4.21 Regarding broader aspects of nutrition education, trair.ing and com- munication, the report recommended a number of important measures over the next five years. The proposed agenda included expansion of staff nutrition training for all institutions engaged in community nutrition education, par- ticularly in the agriculture, health and NGO sectors; upgrading of nutrition education in the school and university system, and nutrition sensitization of concerned officials, particularly at district and village levels. It also urged each sector to develop combined mass media and interpersonal communica- tion programs consistent with its technical responsibilities under appropriate overall coordination and technical supervision. 4.22 Finally, the plan suggested closer government-NGO linkages in respect of food, nutrition and health in addition to inclusion of NGOs on sectoral planning committees in each public agency concerned with nutrition planning and development. NGOs, on the other hand, should aim to link up with and supplement government nutrition efforts through sharing resources, personnel and experiences. 4.23 Complementing the plan, which would represent a sound medium-term approach to Sri Lanka's nutritio;; problems, was a set of preliminary project  - 27 - proposals with estimated budgets totalling around Rs 100 million over five years for proposed inclusion in the 1987 Public Investment Program. Several of the proposals would appear to contribute only indirectly to the main goals of the action plan, but they pointed to the emergence of useful awareness of nutrition as a program concern in sectors other than health. NAFNS Implementation Plan 4.24 One month after FNPPD preparation of the national nutrition action plan, the Ministry of Finance and Planning published a separate NAFNS imple- mentation plan of prioritized but unbudgeted proposals confined to the food and agriculture sectors which contained several with potentially important nutrition implications. The highest-priority proposals with greatest nutri- tion potential were: rice research to develop varieties and production tech- nologies for different agro-ecological environments; implementation of policy initiatives, production and marketing programs to promote other field crops; development of screening and indexing mechanisms for the food stamp program, and incorporation of nutrition considerations into district development plans. However, that implementation plan did not speak to the question of how best to orient the food and agriculture initiatives to the needs of low-income consumers or discuss an implementation timeframe. Implementation Status of Nutrition and NAFtIS Plans 4.25 Translating these generally soundly-oriented planning documents into active programs has been disappointingly slow because of fiscal constraints combined with line ministry difficulties in setting operational objectives and formulating detailed operational plans and budgets. Moreover, the disintegra- tion of the NFNCC (see para. 4.2) has closed off a Cabinet-level political forum analogous to the NHC to reach consensus on major policy, strategy and program aspects of nutrition from a food and agriculture perspective. The absence of a unifying focal point to harmonize the ideas of individual agencies has led to sometimes conflicting institutional approaches. Establishment of an active National Nutrition Policy Council and National Nutrition Bureau (see para. 4.4) to develop consensus and push implementation might help resolve that problem. V. NUTRITION COSTS AND INTERVENTIONS A. Nutrition Intervention Programs 5.1 Sri Lanka's largest publicly-funded nutrition intervention programs consist mainly of targeted food interventions at the household level, on-site pLeschool and school feeding and nutrition education. MCH and other collateral programs also address nutrition issues. 5.2 The two household food interventions--food stamps and thriposha- -account for around two thirds of total nutrition expenditure in Sri Lanka. Food stamps are targeted to families by income and represent essentially a family income transfer. Thriposha, a widely-acceptable take-home supplement is targeted to malnourished preschoolers and selected women in pregnancy or lactation.  - 28 - Food Stamps 5.3 The food stamp program was introduced in 1979 as part of the economic reforms package of the new government. The stamps replaced a rice rationing/ subsidy program which had operated since before independence and is widely thought to have contributed to some of Sri Lanka's favorable social indica- tors, including a substantial reduction in levels of acute child malnutrition. The objective of the food stamps program was to protect low-income consumers from the adverse impact of removing food subsidies which had accounted for as much as 152 of total government expenditure and 52 of GNP in the mid-70's. It was also felt that the rationing and subsidy programs had been, on balance, a relatively ineffective and costly way of targetting benefits to the poorest income deciles. 5.4 Food stamps are redeemable at cooperative stores for basic foods, such as rice, at market prices. As originally designed, eligibility was con- fined to households with declared monthly incomes of less than Rs 300. The monthly value of the stamp entitlement varied with beneficiary age. Those under eight years of age received Rs 25; those between eight and 12 years received Rs 20 and those older than 12 years received Rs 15. Food stamp households also received kerosene stamps to counter rising fuel costs, the monthly value of which rose from an initial Rs 9.50 to around Rs 22 by 1987. 5.5 'While targeted to the poorest, the new program emerged as income supplementation for a much larger population, totalling an estimated 6.8 mil- lion. IFPRI analysis as part of a research project funded by USAID 19/ indicated that the poorest and second income quintiles received only 382 and 282, respectively, of total food stamp outlays. In addition, as many as 102 of eligible persons were not included in the program. IFPRI concluded that limiting eligibility to the poorest income quintile would not have seriously affected consumption levels of less-poor consumers who also received food stamps. (That may no longer be true as a result of deteriorating real incomes in a substantial population segment.) Inflation compounded the effects of mistargeting. The actual value of the unindexed stamps declined by almost two-thirds between 1979 and 1987. 5.6 In 1985-6 the program was restructured to increase the value of the stamps for the poor without increased government expenditure. In addition to shifting implementation responsibility to the Social Services department, principal changes were: opening the rolls to most of the eligible but non- participating households, raising the eligibility cut-off point to households reporting under Rs 700 monthly income and weighting family assistance levels by income. Under the new arrangements, households with reported monthly incomes of less than Rs 300 were eligible to receive stamps for at least five members. Households with reported monthly incomes of Rs 300-399, Rs 400-599, and Rs 600-700, were eligible for stamps for four, three and two members, respectively. 5.7 Several other measures also were introduceu to improve targetting of new entrants. A series of district Poor Relief Committees was set up in 1987. 19/ Edirisinghe, Neville, The Food Stamp Scheme in Sri Lanka. International Food Policy Research Institute, Washington, D.C. October 1986.  - 29 - Each consists of four officials and a nominated member who screen out unqualified applicants. Those whose prelipiinary applications are approved provide further information to be checked by a government Committee member through personal visit. The names of household heads finally approved by the Committee are publicly listed as a further safeguard against wrongful inclusion or exclusion. As a result of these changes, the total number of per&ons receiving food stamps rose by around 102 to 7.6 million--472 of the total population--while the cost of the program remained at around Rs 1.6 billion, through reduced benefits to those at the upper end of income eligibility. Under those arrangements, stamps reached around 1.7 million children under seven years of age, 801,000 children aged 8-12 years, and around 5.1 million adults. To what extent the restructuring particularly benefitted the most severely deprived is unclear, but they appear to have at least shared in the new enrollment opportunity. Unfortunately, the shift in entry criteria was not accompanied by elimination of unqualified existing beneficiaries. Moreover, in December, 1988, the Cabinet decided to add to the rolls all those on the waiting list and in January, 1989, decided to double the value of the stamps. This combination of changes raised the cost of the program by 125Z and increased the number of beneficiaries to over eight million. 5.8 Food stamps now provide the equivalent purchasing power of fewer than 1,000 calories per day to a low-income family of five. According to IPPRI analysis of 1981-82 data the additional food stamp income raised calorie consumption by 122 and 61 in the lowest two income quintiles, respectively. The analysis also concluded that because of leakages to higher-income recipients, who should have been excluded from the program, food stamps for the lowest income quintile have cost the government "150 percent more than what it cost the recipient households to purchase the given amount of calories.' It also reported that "food stamp incomes appeared to increase calorie consumption of preschool-aged children in the lowest quintile by 5.4 percent; all other members in the same households increased their calorie consumption by nearly 10 percent due to food stamps.' 5.9 The current purchasing power of food stamps should be sufficient to undergird the nutritional needs of most food-insecure households, although the program as a whole is inefficient because of substantial overenrollment. Analysis of the 1986-7 CFS indicates a daily family calorie shortfall of 1,268 and 680 in the lowest urban and rural income deciles, respectively, excluding food stamps. The next higher deciles, and the estate sector, are in less precarious conditions. At current purchasing power and levels of food stamps, only families in the lowest two urban income deciles still would need further support to reach minimally acceptable consumption levels of 2,000 calories per adult-day. As shown in Annex A, Table 10, the annual cost of a perfectly targeted food stamp program meeting raising the consumption threshhold to 2,000 calories per person-day for all food-insecure Sri Lanka families would be less than RS 1,000 million--less than one-third the currently budgeted outlay. Thriposha 5.10 Thriposha is distributed to up to 580,000 pregnant and nursing women and malnourished children under five years of age. A 50-gram serving contains  - 30- 186 calories and 13 grams of protein, along with vitamins and minerals. The ration is estimated as adequate to meet around 402 of the protein and an average of around 152 of the daily calorie needs of an under-three child. Child beneficiaries are identified mainly through monthly weighing at MCH facilities, where pregnant and nursing women are also enrolled by medical personnel on the basis of clinical signs of anemia or breast-feeding problems. 5.11 Thriposha is a formulated, pre-cooked food consisting of locally- grown and processed corn and soya combined with imported non-fat dried milk (NFDM) donated by the US Government under Public Law 480. A vitamin-mineral premix is part of the formula but has been unevenly available over the past year. The Ministry of Health (MH) finances local commodity and operating costs of the program and distributes thriposha. Around 500,000 beneficiaries are expected to receive the supplement monthly on a take-home basis through health centers. Some 60,000 estate and 20,000 voluntary agency (largely through Sarvodava) beneficiaries are expected to receive on-site supplementa- tion. The number of enrolled beneficiaries actually collecting thriposha tends to vary somewhat by month, according to program records. However, parti- cipation began to increase steadily in late 1987 and rose to around 92Z of those enrolled in the program by February, 1988, the last month for which data were available to the mission. Around 59Z of the beneficiaries were children aged 13-59 months; 162 were children aged 7-12 months; 14Z were pregnant and 10X were lactating women, and the other 12 was hospital cases. That coverage totals around 472 of all children aged 7-12 months, 202 of those aged 13-59 months, and 322 of women in the last six months of pregnancy or first six month3 of lactation. 5.12 CARE, tha US voluntary agency, initiated the program in 1972 and has played a major collaborative role in its design and implementation, including the establishment and oversight of thriposha production facilities with AID- funded technical assistance. CARE also assists in program monitoring to assure accountability for distribution to service delivery points and to bene- ficiary families. Discussions currently are under way regarding: (a) the planned phaseout 'jy 1991 of AID commodity assistance to the program, (b) how best to implemert a 1987 Cabinet decision to double current thriDosha produc- tion capacity and (c) prospects for privatizing production rather than contin- uing it under Government auspices. 5.13 Two evaluations of the program have highlighted both its accomplish- ments and shortcomings. One, conducted under AID contract, found through longitudinal analysis of beneficiaries that "the package of services [including thriposha] provided through MCH clinics improved nutritional status" although it was not possible to segregate the impact of thriposha alone. It went on to conclude that thrinosha targetting was generally effec- tive: four of five child beneficiaries were either malnourished or at serious risk. It also observed that the program had the important spin-off effect of increasing family contact with the public health system. 5.14 The AID-sponsored evaluation suggested operational and management improvements to the program. These included better compliance with and monitoring of targetting procedures, commercialization of thriposha production and more involvement of mothers in growth monitoring by weighing their own children and filling out growth charts. It also observed that sharing of  -31 - thriposha among family members took place, possibly diluting its nutrLtional impact. Reduction or termination of the program might be difficult since so many poor beneficiary families had become dependent on thriposha as an entitlement. The other evaluation, published by FNPPD, also noted deficien- cies in the beneficiary screening procers and problems of thriposha leakage to ineligible family members. It observed that nutrition/health monitoring and financial record-keeping needed considerable improvement at the field level and made several recommendations regarding logistics of thriposha suipply to MCH facilities. 5.15 A subsequent AID-government workshop to review thriposha recommended government clarification as to whether the program was meant as supplemental feeding for general nutrition improvement or a nutrition safety net for mal- nourished preschoolers and mothers. It also cited ineligible beneficiaries, the lack of adequately-linked nutrition education and the need to incorporate thriposha more effectively into a comprehensive preventive care program. 5.16 MH since then has taken steps to improve the logistics aspects of the program, but fundamental problems inherent in most take-home supplementation efforts persist. The quality of growth monitoring is variable; both the regu- larity and accuracy of the process are uncertain. In some areas monthly weigh- ing is by village volunteers, each responsible for around 25 families. Super- vision is constrained by other demands on health worker time and performance quality has not been formally assessed. These same time pressures also affect the regularity and reliability of weighing by health workers. In theory, thriposha eligibility is determined by the subdivisional Medical Officer of Health. In practice, the task can devolve to the peripheral health worker who has little incentive to screen beneficiaries on nutritional criteria since thriPosha supply often exceeds demand. Sufficient and clearly allocated time are required for effective performance of counselling and educational aspects of the growth monitoring process but frequently receive low priority against competing tasks. 5.17 Additionally the nutritional response to thriposha is not closely monitored and criteria for terminating supplementation are vague. Conse- quently, little motivation exists for health staff involved with the program to ensure the achievemen. of normal child growth patterns. 5.18 The economic cost of producing and packing a 750g packet of thriposha is currently estimated at Rs 23.05, including the imputed value of donated commodities. Those unit costs are projected to decline as cheaper local commodities replace expensive NFDM. However, the financial costs of th_iposha production are already artificially high because of the requirement to buy maize from the Paddy Marketing Board at a negotiated rather than the open market price and a support price for domestic soy which is well above border levels. The Government should consider liberalizing proctiement arrangements to lower the production cost of thriposha. 5.19 Doubling local production of thriposha would keep availability constant when AID's pre-processed input phases out. Current annual beneficiary coverage is around 405,000 under-five children and 128,000 pregnant and nursing women. If properly targeted, the available thriposha supply could supplement all children aged 6-59 months who a:e likely to be  - 32 - either acutely malnourished or both stunted and wasted, with enough left over to cover an additional 1OZ of that population whose growth may otherwise be faltering. It also could provide supplementation to the 25? of pregnant women at high risk of low-birth weight babies and 252 of new mothers who may need supplementation during lactation. On-site Feeding 5.20 FNPPD and the Ministry of Indigenous Medicine for some years collaborated in an effort to popularize kola kenda, a nutritious gruel, as a supplement for children. The product usually consists of mung bean, rice, the juice and sometimes mashed pulp of local varieties of green leaves (e.g. kotukola, mukunvenna, hatavariva. penella and batu) and, when affordable. coconut milk. The gruel varies in content, consistency and nutrient value, depending on its local composition. However, a six-ounce cup is said to contain around 180-200 calories, some micronutrients aMd a proportional amount of high-quality protein. The value of the raw materials, including fuel, varies, but mission calculations in the field averaged Rs 1.20 per serving. The importance of kola kenda is that it is a community-based intervention. The government provides cooking utensils worth Rs 2,500 to each facility where the community agrees to provide the raw materials and cook the gruel ence or twice a week. Sarvodava provides kola kenda up to_five times weekly to around 29,000 beneficiaries at over 1,000 pre-schools; FNPPD reports that the supple- ment also is avpilable at more than 1,200 schools in 23 districts. The use of an indigenously-designed supplement through extension community participation has considerable potential merits. However, wide variations in composition of the supplement and its sporadic delivery generally hamper the value of kola kenda as a nutrition intervention. 5.21 The new Government recently announced a national free lunch program for all of Sri Lanka's estimated 4,000,000 school children. The program is to provide a 500-calorie cooked meal for 180 school days annually at a daily budgeted cost of Rs 3 per beneficiary. Provincial Councils are to implement the program with assistance from the ministries of Education; Agriculture, Food and Cooperatives; Health and Women's Affairs; Social Welfare, and Public Administration, Provincial Councils and Home Affairs. Commodities are to be procured by officials at each school. Community v-:unteers are expected to htelp prepare the meals. An initial Rs 2 billion has been budgeted for the program for calendar 1989 on the implicit assumption that annualize costs will run around Rs 2.6 billion. However, it is more likely Lhat the real cost of the program will be substantially higher. The cost of calories now averages aro. d Rs 7 per thousand for average diets. Administrative costs, either as sup.'.ementary staff payments or calculated as the opportunity cost of time, will add a minimum 20? to the program. Combined with inflation, these factors make it likely that the total cost of the program will approach Rs 4.55 per student-day, or around Rs 3.3 billion for the first year alone. School feeding may promote better school performance and reduced absenteeism, but as a nutrition intervention is of lower priority than efforts to counter pre- school malnourishment and miconutrient disorders, which particularly during the formative first three years of life have a more profound and permanent impact on child growth and development.  - 33 - Nutrition Education 5.22 Relatx. ly small amounts of nutrition education take place through a number of agencies. These include activities funded both centrally and at district levels through FNPPD. the interesting work of the Farm Women's Agricultural Extension program, some Ministry of EducAtion programs and some as part of MCH efforts. The Secretary, MOPI, chairs a special sub-committee to monitor and coordinate nutrition education and communication programs whose membership includes representation from FNPPD, the media and the ministries of Agriculture, Education, Health and State. However, a national set of nutri- tion education objectives and a strategy to carry out them out has yet to be developed. 5.23 FNPPD's nutrition education programs have centered on seminars and workshops for officials and community leaders and the creation of public nutrition awareness. Emphasis has been on the use of radio, the press and television; preparation of nutrition education readers and other publications for schools, and tne publication of nutrition periodicals for the general public. However, in the absence of clear objectives or a strategy, the impact of these efforts is likbly to be limited. The World Food Programme (WFP) has provided food commodities to compensate pregnant and nursing women, parents of malnourished children ai.. community leadezs for the opportunity costs of attending a pilot FNPPD nutrition education program in seven districts. The nutrition education activities ranged in duration from one to 26 days, the project included mobile health clinics to immunize children and reached around 79,000 adults and 166,000 children. Expansion and refinement of that program with WFP assistance is now under consideration based on an evaluation of the pilot phase, which is underway. 5.24 At the district level, applied nutrition programs in around 24 of Sri Lanka's 25 districts have included nutrition education workshops and promotion of home gardens. However, the allotted funds have been modest, running around Rs 2-3 million yearly from the decentralized budget. The 1988 allocation rose to Rs 10 million but local absorptive capacity limited expenditures to around Rs 4 million, including costs of kola kenda, salt iodization, promotion of home gardens and nutrition education seminars. 5.25 With FAO and then UNICEF support, since 1976 the Farm Women's Agricultural Extension (FWAE) program of the Department of Agriculture has sought to work with rural women to promote better community health and nutri- tion through household food production, improved management and organization of home activities and better planning and preparation of nutritious meals. A network of 33 female subject matter specialists carries out class and house- hold instruction in home gardening; home, food and nutrition management and income-generating activities such as bee-keeping, floriculture and handi- crafts. The current UNICEF-aided program, begun in 1984, covers 6,860 families in eight of the 22 districts where FWAE already operates as an effort to increase the income and health and nutrition standards of ten of the poorest families in each village. The program is to expand to all of Sri Lanka's 25 districts coverage over the next four years as more FWAE staff get recruited and trained. Although not yet formally evaluated, the combination of extension and health and nutrition education, provided on a woman-to-woman basis, is an attractive approach which, while labor intensive for FWAE, could prove t3 be cost-effective.  - 34 - Health-Related Interventions 5.26 Widespread iron and folate deficiency problems particularly affecting pregnant and nursing women and probably preschool children (see para 1.11) are addressed by distribution of iron supplements and messages regarding appro- priate dietary practices through peripheral health workers. However, most observers agree that compliance is uneven and the iron deficiency problem has failed to show significant improvement in recent years. Ministry of Health guidelines recommend 200 mg of iron thrice daily, which some women tolerate poorly. Small-scale testing by a Sri Lanka voluntary agency indicates that 60 mg daily is effective in raising hemoglobin levels and well-tolerated by women when administered with a meal, a finding which suggests the need to consider adjusting current guidelines. 5.27 In early 1987 the Cabinet approved remedial action against iodine deficiency disorders through a pilot project for salt iodation in Kalutara district, where the sale of unfortified salt has been banned. The National Salt Corporation produces and distributes half-kilo packets mainly through cooperative stores at a retail price of Rs 2.75 per bulk kilo against Rs 1.75 for common salt. The Sri Lanka Standards Institute maintains qualLty contrcl; the Medical Research Institute (MRI) is responsible for baseline and follow-up surveys to determine impact. However, the program operates on quite a limited scale. Annual distribution of around 20 metric tons is enough to meet the yearly needs of only around 4,000 beneficiaries. It is estimated that by the year 2,000, when Sri Lanka's population is likely to be around 20 million, the country's total salt consumption would be around 100,000 tons. According to an international fortification expert, in current prices the investment cost of physical facilities and equipment to fortify that salt would be US$5.8 million and the annual operating cost of fortification including chemicals, labor, power, overheads and depreciation would run around US$3.3 million, or around US$0.17 (Rs 5) per capita.20/ 5.28 However, lack of effective demand is likely to be more of a long-term constraint to salt iodation than technical or managerial issues. Low-income households tend to prefer coarse or even rock salt, which is cheaper and absorbs less moisture than finely-granulated varieties. Elimination of tradi- tional salt varieties from the market would be difficult until a combination of education, availability and pricing generated sufficient demand for the iodized product. In the interim, the government could consider an optional approach to goiter control: injectible iodine for married women in high-risk areas who intend to have more children. At a cost of US$0.25 per dose, the iodized oil confers three years of protection for both the woman and her fetus. Delivery would be feasible through the Health ministry's extensive infrastructure. Assuming that around 300,000 women would be an upper bound for the program, annual delivery costs would run less than Rs 1 million. 5.29 Vitamin A deficiency is believed relatively low in Sri Lanka at present. However, it has been high in the past and areas and groups with higher prevalence persist. Combined with continuing high morbidity from respiratory and intestinal infections, this evidence suggests that fairly widespread subclinical deficiency may exist. Vitamin A supplementation 20/ Communication from S. Venkatesh Mannar to IBRD staff, March 1988.  - 35 - currently takes place sporadically where the frequency of clinical signs exceeds WHO-defined thresholds; there are no plans for general pre-school prophylaxis. Vitamin A is not administered routinely to measles cases although proven effective in reducing complications. Experience in Indonesia indicates that Vitamin A prophylaxis in pre-school populations with marginal nutrition and health status is effective in reducing mortality and morbidity, particularly from diarrheal disease and acute respiratory infections. Sri Lanka might consider the benefits of mass vitamin A prophylaxis for pre-school children on that basis. Based on current international costs for high-doze oral Vitamin A, the mission estimates that a national pre-school program would cost around Rs 2 million. 5.30 In addition to addressing micronutrient deficiency problems, MCH interventions can improve the biological utilization of available nutrients. They affect nutrition status by improving weight gain during pregnancy, promoting breast-feeding and safe weaning practices and by effective control of non-immunizable infectious diseases through oral rehydration, prompt treat- ment of acute respiratory infections and deparasitization. Malaria prophylaxis in pregnancy and effective household control measures for children also would contribute. Against these needs, health sector performance has been mixed. Sri Lanka has promulgated and enforced a comprehensive code regu- lating infant food formulas and breast milk substitutes. However, low birth weigh remains a key factor affecting infant nutrition status. Control and treatment are considered routine aspects of MCH care, but available morbidity data indicate that diarrhoea, respiratory morbidity, intestinal parasites and malaria remain as nagging problems which the health system needs to address more vigorously. 5.31 The use of oral rehydration therapy (ORT) is the preferred control measure until environmental sanitation measures eliminate diarrhea as a public health problem. However, the 1987 DHS and Dry Zone studies indicate that almost twice as many families know about than actually use ORT. Further strengthening of promotional and educational programs against diarrhoea probably are needed, perhaps on a campaign basis. 5.32 At present, acute respiratory infections are treated with antibiotics only by medical staff. Primary health programs in other countries have successfully lowered qiortality by providing antibiotic cover to small children at lower levels of care through paraprofessionals. Oral co-trimoxazole can be administered easily to small children at the village level by auxiliary personnel. Routine deworming of children with mebendazole, has proven effec- tive elsewhere and is worth Sri Lanka consideration in areas of particularly high infestation. However, the per-beneficiary incremental costs, at around US$7.01 per year, would be too high for mass national administration. The present approach to treatment of presumptive malaria fever with chloroquine might be strengthened by continual or seasonal chemoprophylaxis with pyrimethamine/dapsone, as in the Gambia. Other Initiatives 5.33 Spurred by the apparent deterioration in nutrition status, the Cabinet in late 1987 approved the Prime Minister's suggestion that Sri Lanka develop a short-term project to improve nutrition standards of children aged  - 36 - 6-36 months with assistance from donors including the World Bank. FNPPD examined the matrix of existing interventions to see whether they could be modified to meet the Cabinet's objectives but concluded that it would be difficult quickly to rationalize the food stamps or thriposha programs to bring about desired changes. For example, needed improvements in thriposha targeting, nutrition education and management would produce only medium-term results. 5.34 In response to the Cabinet mandate, FNPPD formulated a pilot project for 'nutrition villages" in Sri Lanka. This effort would blend selected features of nutrition programs in other developing countries of Asia to improve the status of the ultra-poor through a combination of poverty allevia- tion, employment 2nd nutrition interventions. The estimated cost of the program over five years would be around Rs 1,600 million, equivalent to one 7ear's food stamps outlay. It would aim at generating adequate and self- sustaining incomes for 189,000 poverty-stricken frmilies comprising around 1.1 million beneficiaries, while improving the nutritional status of preschool children and pregnant and nursing women. 5.35 Beneficiaries would be selected first through growth monitoring to identify households with moderately or severely malnourished or micronutrient deficient preschool children and pregnant or nursing women. Economic criteria would then be applied to screen households further, including consideration of food stamp recipient or landless households or those without any economic activity. Seventy-five households would be considered a "nutrition village" from an organizational and managerial perspective. A panel headed by the Assistant Government Agent would make final selection of beneficiaries and 'villages" . 5.36 The components of the project would consist of a Rs 2,500 family subsidy over two years to each participating household, food supplementation, environmental sanitation and nutrition education. The subsidy would be based on criteria and principles to be worked out at the district level by the Government Agent and concerned agencies; it would be expected to result in self-sustaining income generation and employment for one family member. During this period, eligible preschoolers and women wouid receive on-site food supplementation five days a week at a budgeted cost, respectively, of Rs 3.50 and Rs 5 per daily ration. The suggested dietary combinations are mixtures of cereals (preferably rice) and pulses, kola kenda, egg and margarine-smeared bread and milk and a bun or biscuits. Environmental sanitation would be improved through a latrine for each household and three community wells per 'village'. Intensive nutrition education would take place to improve family feeding and related practices. 5.37 This proposal is now being pilot-tested in 28 villages where, for the most part, surveys have taken place and, in some, food supplementation has already gotten under way. An impact evaluation is scheduled for mid-1989. Other Donor Activity 5.38 In addition to donor support referred to above, several local and foreign voluntary and government agencies also are active in smaller-scale nutrition interventions. These include Redd Barna, a Norwegian voluntary  -37 - agency undertaking food production, supplementation, health and nutrition education programs in several parts of the country, including the north; both the UK and US. Save the Children foundations, and the local Saukyadhana movement. B. Costs 5.39 Sri Lanka's nutrition spending has tripled since 1987. At around Re 6.2 billion (US$188 million at current exchange rates), Sri Lanka's total nutrition spending constitutes around 9Z of projected total and around 12% of projected recurrent Government expenditure for 1989. Around US$109 million equivalent, or 602 of those nutrition outlays, are for a single Government subsidy program--food stamps. A second program--school lunches--accounts for one-third of currently budgeted nutrition costs. Another household food supplementation program--Government distribution of thriposha for mothers and children with CARE assistance--accounts for a further 42 of overall nutrition outlays. Other Government, voluntary agency and official foreign aid programs account for the residual 3Z. Until the recent jump in food stamp costs and the introduction of the school lunch program, annual nutrition expenditures varied by around a nominal 62 in recent years, principally because: (a) food stamps accounted for a steady Rs 1.6 billion annually, (b) other existing programs remained relatively constant and (c) the few aew nutrition initiatives involved minimal spending, particularly in the face of overall budgetary constraints. VI. ISSUES AND OPTIONS: AN ACTION AGENDA 6.1 Sri Lanka's nutrition problems are essentially twofold. The first is how to deal with chronic nutrition insecurity at the household level, affect- ing the lowest three urban and lowest rural income deciles. The second is how to ensure adequate nutrient availability and utilization by specific target groups, particularly young children and pregnant and nursing mothers. The latter problem exacerbates the already inadequate availability of food to poor households and is responsible for a 252 rate of low birth weight, child nutri- tion deficits particularly during weaning in the second year of life and subsequent high rates of childhood stunting. Nutrition policies, strategies and programs should be directed specifically to both the household and vulner- able group problems. 6.2 Over the long run, income and employment gains will drive durable increases in food consumption by Sri Lanka's poor, if focussed on those specific groups. The following observations are based on the assumption that such efforts will be intensified. However, they also need to be complemented by concerted and coherent medium-term approaches. In some cases, this can be accomplished by redirecting efforts, in others by intensifying what is now under way and in still others by new combinations of existing interventions. However, it would be inappropriate and misleading to assume that nutrition programs could meet the country's long-term nutrition needs in the absence of focussed income growth and employment generation.  -38- A. Obiectives 6.3 A manifest need exists to review, revise, prioritize and quantify existing nutrition objectives (see para. 4.5) and to phase and sequence them into realistic time horizons. Reductions of 251 in stunting among children under the age of three years, for example, while ambitious probably could be achieved in at least parts of Sri Lanka but would take a minimum of five years of steady effort, as would a similar reduction in the number of low birth- weight babies. Significant reductions in stunting for older pre-school children would take somewhat longer to emerge. A 30? reduction in the inci- dence of maternal iron and folate deficiency and substantial reductions in manifest IDD also should be possible over five years if given sufficient priority. Programs with the potential to affect such changes are already in existence but need considerable strengthening (see paras. 5.4-5.34). 6.4 The government may also wish to consider setting specific minimum food intake goals as an intermediate contribution to both the above objectives and to the improved health and nutrition of the most vulnerable segments of the population. Increasing daily consumption by the poorest 202 of the popu- lation, for example, to an adult per capita average of 2,000 calories would help provide a safety net against continued nutritional deprivation until their incomes rose to acceptable levels. B. Policies and Strategies 6.5 The Government's general nutrition intentions can be inferred from an array of available statements and documentation reflecting an abiding politi- cal and administrative concern about the problem. However, existing policies and strategies are necessary but generally not sufficient instruments to guide program development. Those affecting food supply and distribution pay only marginal attention to consumption issues. Health strategies have had mixed success in preventing growth failure and reducing iron and folate and other micro-nutrient deficits and rates of nonimmunizable infectious diseases. Food and Agriculture 6.6 In the food and agriculture sectors, two criteria should be cor,sid- ered in designing economically viable nutrition-oriented policies and strate- gies. First, they should seek disproportionately to benefit the poor in terms of food consumption and/or income increases. Second, they should aim to increase the availability of lower-cost nutrients generally in short supply. 6.7 Against that backdrop, three areas require further immediate develop- ment. First, efforts to lower rice prices to consumers through increased productivity and improved post-harvest technology also are needed. Second, steps to reduce sugar prices approximately to world levels should be consi- dered but the implications of such a move on the balance of payments and on government revenues should first be studied. Sugar is an important source of calories to the poorest deciles and the present price structure taxes low- income groups disproportionately as a source of general revenue. Third, policies and strategies to increase the production and marketing of subsidiary and food crops such as corn or manioc should be developed and implemented, in order to raise incomes and promote diversification including development of the domestic feed market. Successful diversification, however, would require  - 39 - improved research and extension services and would not necessarily succeed unless demand and relative agricultural output prices provided adequate incentives. Both the field crops and rice interventions were earlier identified by FNPPD as key development areas (see para. 4.17). 6.8 The argument favoring lower rice prices derives partly from opinion that government subsidy of paddy production may have inhibited development of both the tree crop subsector and other food crops by discouraging farmers from diversifying to other crops. Productivity increases leading to a reduction in financial incentives to paddy production (e.g. credit, fertilizer and producer price supports would tend to lower rice prices to consumers and thereby alter the wheat/rice price ratio. This might make rice relatively more attractive to consumers and reduce the country's dependence on imported wheat. This argument is reinforced by concerns that economically unattractive options for surplus disposal could arise if Sri Lanka produced more rice than needed for domestic consumption and tried to enter the export market with a relatively high-cost, lower-quality commodity. Advocates of diversification also argue that subsidiary crop development would be more labor-intensive and would gene- rate greater employment, particularly if marketability of these field crops could be increased through promotion of dietary change. 6.9 However, there also is an argument on nutritional grounds that a reduction in rice prices would benefit all but the largest paddy producers. Farmers' net returns on paddy production have been decreasing gradually in recent years for several reasons with the result that some less productive lands have been taken out of production. (The profitability of paddy produc- tion varies considerably by region and generally is less profitable in the wet zone and marginal areas than in the irrigated dry zone.) In addition there has been discussion of reducing the existing bias toward paddy production by introducing effective irrigation charges, a possible land tax increase and expansion of agricultural services to include other crops. Moreover, fertilizer costs, which represent around 8-102 of the current costs of paddy production, are likely to increase both in world terms and through a reduction of the input subsidy. These factors also point to the need for productivity increases in the paddy sector. 6.10 Resistdnce to a decrease in the paddy procurement price (it was increased from Rs 70 to Rs 80 per bushel in August, 1988) rests on two argu- ments: production would decrease and rural incomes would therefore fall. Both common sense and recent analysis for the OECD 21/ support the importance of price as a determinant of paddy production. Given limited alternatives, lower price support by itself would not be likely to result in major reduction of paddy on productive land. Whether a reduction in profit margins would significantly affect production or the provision of employment by larger producers--the net sellers of rice--who account for an estimated 30-402 of annual yields, needs to be determined. However, there also are indications, which need verification that shifts in land use have taken place on more marginal rather than productive landholdings in the wet zone, as net returns have decreased since 1986. 21/ Thorbecke and Svegnar op. cit.  -40 6.11 Because the present market price for paddy is higher than the pro- curement price and public procurement is relatively low, there may not be an existing mechanism for the government to reduce rice prices substantially if it wished to do so. However, by curtailing further increases in the paddy procurement price, the Government could encourage more efficient and lower- cost production and let market forces benefit low-income consumers. Health 6.12 Continuing micronutrient deficiency problems, the uneven quality of thrigosha implementation and scant attention to nutrition education underscore the need to incorporate nutrition more fully into Sri Lanka's strategic agenda for primary health care. The most important single development would be for the Ministry of Health both to send clear signals to its regional managers that pre-school and maternal nutrition are a high-priority concern by assist- ing districts in the preparation of realistic annual plans and programs specifically to control anemia, IDD, subclinical Vitamin A deficiency and growth failure in small children, which would need to be monitored both locally and from Colombo. However, the health system would need to strength its central and regional management-and operations for these initiatives to be effective. Nutrition Education 6.13 Strategies to improve behavior directly affecting nutrition status are needed as a linking mechanism at the interface of nutrition and both food consumption and health. However, they are not currently in place, although FNPPD provides an institutional base for their development. One key strategic objective would be to improve calorie intake by pregnant and nursing women and early introduction and use of semi-solid and solid foods for weaning children. A second would be effective reduction in iron and folate deficiency in preg- nant women, which contributes to low birth weights. Another would be fully to institutionalize home management of diarrhea combined with action to combat other infectious diseases. The strategy to achieve those goals would involve a mix of both face-to-face and mass media efforts carried out by an intersec- toral combination of government and voluntary agencies. FNPPD in consultation with those agencies should take the lead in drafting the national nutrition education agenda for endorsement by the Cabinet. C. Institutional Arrangements 6.14 A broad array of instruments for nutrition advocacy and to promote coordination and improvement at policy, strategic and program levels have operated with uneven effectiveness. Differences between the NAFNS and FNPPD's food and agriculture priorities remain unresolved. The gap between FNPPD recommendations and actions to improve nutrition intervention programs remains unclosed. Establishment of the proposed National Nutrition Policy Council under the Prime Minister, serviced by a Nutrition Policy Bureau, or the re- establishment of a division like FNPPD in MPPI to service the NHC (see para. 4.4) could help ensure that nutrition has an appropriately high-level public focal point and that consistent and coherent nutrition policies, strategies and programs were developed, implemented and properly monitored.  - 41 - D. Programs 6.15 In addition to the strategic considerations above, certain program options also should be considered, particularly for the main existing inter- ventions. Food Stamps 6.16 As indicated above, the food stamp program has faced two fundamental problems since its inception. The first is the regularly eroding value of the unindexed stamps with inflation. The second is the poorly targeted, and hence unnecessarily costly, nature of the program. 6.17 In December 1988 the government addressed the first of these problems, at least temporarily, by doubling the value of the food stamps. (A low income family of five now receives stamps worth roughly Rs 200 per month rather than Rs 100). These higher food stamp values will, as indicated below, bring the average family of five in each income decile, except for the poorest two urban deciles, to consumption levels of at least 2,000 calories per person per day. 6.18 The problem of poor targeting, however, has not been addressed. Recent analysis of 1986-7 CFS data by staff of the National Planning Division indicates that during that period nearly one-third of households 22/ in the poorest quintile of the population did not receive food stamps. At the same time gross underreporting of incomes and lack of controls in most rural areas have led to a situation in which only 10 of the 25 districts reported less than 502 of their population receiving food stamps. Five districts actually reported more than 702 of their population receiving them. 6.19 Based on preliminary analysis of the 1986-7 CFS data, (see Annex A, Tables 7-9), it appears that effective targeting of the program could bring all income deciles to near adequacy in terms of caloric intake 231 at a fraction its current cost. 6.20 At present the only population groups not consuming, on average, 2,000 calories per day without food stamps are the three poorest urban deciles and the poorest rural decile. Of these only the two poorest urban deciles would continue to fall short with the now doubled food stamp values, these by roughly 90-135 calories per person per day or roughly Rs 100 per household per month. (See Annex A, Table 10). 6.21 The cost of reaching 'near adequate" caloric intake for each of these four deciles with a perfectly targeted food stamp program, as indicated in that table, is Rs 986 million or roughly 27? of the estimated CY 1989 food 22/ When households are ranked by total household income. 23/ "Near adequacy' here is defined as 2,000 calories per adult equivalency unit. It is important to note, however, that consumption for each decile is an average figure, and many families, by definition, will fall below it.  - 42 - stamp program. With detailed household financial information presently being collected in preparation for the JSP, it should be possible to carry out a well-targeted food stamp program. 6.22 The Government intends to replace food stamps with monthly consumption allowances under the JSP, which is intended to substantially eliminate poverty over a two-year period. However, the annual Rs 28,000 million estimated annual cost of the proposed JSP consumption allowance is more than 28 times the estimated cost of a well-targetted food stamp program. Moreover, even raising food consumption of the poorest quintile up to middle- class levels would cost less than an estimated 20Z of projected JSP consumption outlays. 6.23 Overall, two features would seem essential for an effective Sri Lanka targeting system. First, the system needs to be visibly insulated from the political process and consciously biased against upper-income participation as a way of focussing attention on the poorest deciles. Second, it should rely initially on reported expenditures rather than continually misreported incomes as more accurate determinants of calorie intake. The most important precondi- tion is, of course, a commitment to ensuring both inclusion of the poor and exclusion of the ineligible. Previous reluctance to drop higher-income bene- ficiaries may be difficult to overcome. 6.24 Rice Grading. At present, rice represents roughly 75Z of food stamp purchases. The quality of rice available through cooperative stores for purchase with food stamps is relatively high. If consumers were given the option of purchasing lower quality (but nutritionally equal or superior) rice at a lower price, the caloric value of the stamps could appreciate consider- ably at little public cost. Initial inquiries indicate that there may well be considerable interest among lower income food stamp recipients in such an option, which might involve making broken or undermilled raw rice, or both, available for food stamps at cooperative shops, or at least ensuring that cooperative stores have a constant and adequate supply of the cheapest rice varieties currently in the market. 6.25 The availability of cheaper human-grade rice brokens (used widely in fair price shops in India) may be too limited for widespread use in Sri Lanka but should be explored, since they sell at two-thirds to one-half the price of commonly-eaten rice. Additionally, with existing milling equipment, it might be possible to increase the supply of undermilled raw rice, already said to be available in large quantities and up to 252 less expensive (depending on the season) than rice presently available in the cooperative stores. If under- milled rice were purchased by lowest decile households, it would increase by 20Z the caloric value of the stamps and would provide half of the calories needed by this group to reach 8SZ adequacy. An alternative would be to make such rice available only for those 6 months of the year--October to March- -when rice prices are higher (by as much as 251) and lower quality rice less available on the open market. This would at least provide the low-income family with food stamps providing roughly constant caloric value throughout the year. If domestic brokers were successfully added to the system at an even lower price, benefits would be correspondingly higher.  - 43 - 6.26 Introduction of undermilled raw rice or similar lower quality varie- ties into the cooperative stores would not be without problems, chief among them being shorter shelf life. In addition, the acceptability of a particular variety, even among low income households, will not be uniform through the country. Moreover, rice millers may not consider it economic to screen for brokens and sell the better-quality rice at higher prices. Accordingly, care- ful testing will be necessary (a) to assess the feasibility of producing brokens as well as the shelf life of a range of lower quality rice varieties; and (b) determine--by actual placement of such rice in cooperative stores--the quantities that would ba purchased by each income group in the program in each area of the country and in each major season. Thriposha 6.27 This program appears to need operational improvemei.t. It is aimed at an appropriate target group and the supplement itself is of acceptable quality. However, correctible deficiencies have limited its value as a nutri- tion intervention (see paras. 5.16-5.21). Therefore, the mission recommends a thorough review of the prugram to improve growth monitoring, beneficiary selection and family use of thriosha as an intervention against child growth failure in accord with program intentions. The latter goal would require more effective nutrition education than presently exists. Kola Kenda 6.28 Principally organized by communities themselves, the kola kenda program (see para. 4.20) is more important from community involvement than nutrition perspectives. Consideration should be given to intensifying its use for at-risk children under three years of age and to ensuring that children receive the supplement frequently enough to obtain substantive nutrition benefit. School Feeding 6.29 This program represents a significant financial share of child- directed interventions. However, it reaches children whose problems frequently originate before birth and are essentially the product of early nutrition and health deficiencies. From a nutrition perspective, school feed- ing has lower priority and potential than successful efforts targeted to the most vulnerable younger groups, although it may be considered important on social or other grounds. However, in a resource-constrained environment, the government may wish to shift those resources to other interventions such as deworming, goiter control and nutrition education. Nutrition Villages 6.30 The ambitious nutrition village concept (see paras 5.35-5.39) is attractive as a multisectoral construct but extremely difficult from a plan- ning and implementation perspective. Proposed program inputs vary in scope, complexity and lead time for expected impact. Effective coordination among the concerned institutional actors also will require careful planning and monitoring to ensure that interventions take place on time and at essential levels. Carrying out the; pilot program in even 40 such villages, as currently  - 44 - anticipated, may strain implementation capability unless the inputs are care- fully phased and sequenced. Additionally, the new program may divert energies from essential and lower-cost improvements to existing delivery programs which could have more widespread and systemic impact. For these reasons, it is recommended that the government consider restructuring the nutrition villages concept in a more limited fashion and that considerable attention be paid to the detailed design of the implementation system and its management. An Optional Approach to Nutrition VillaRes 6.31 Restructuring the nutrition villages concept also would integrate key interventions directly at the village level through a limited service delivery package. Interventions -;ould include weight monitoring, infectious disease control, micronutrient supplementation, nutrition education and selective short-term on-site food supplementation for at-risk mothers and young children. The precise mix of services and their relative priorities would derive from careful initial analysis at the community level. The delivery mechanism could be a combination of non-government organizations with signifi- cant village outreach capacity such as Sarvodava. the national network of an estimated 42,000 health volunteers, organized through the Ministry of Health and communities which themselves might wish to organize such programs at the village level. 6.32 If administratively feasible, this approach might prove a cost-effec- tive alternative to present modes of nutrition aelivery through the health system. Consonant with principles of both decentralization and people-based development, it could constitute a core program around which to develop income-genereting activities at a later stage. Thriposha could be the supple- ment until communities were in a position to develop their own weaning mixtures. The broad outline of a protocol for testing this approach is Annex C to this report. E. Recommendations 6.33 In addition to calling for a review and reformulation of nutrition objectives and strategies, this report proposes a package of programs in five specific areas: (a) agricultural research, prices and production, (b) food interventions targeted to both low-income households and particularly high- risk groups, (c) a concerted attack on micronutrient deficiencies, (d) nutri- tion education and (e) a community-based nutrition and health program. 6.34 On the agriculture side, the mission recommends research in two key areas: rice productivity and post-harvest technology to lower the cost of production of rice and the development of improved production and marketing systems for subsidiary field crops. The present investment rate on agricul- tural research is estimated at about 0.71Z of the agricultural gross domestic product. A doubling of the research percentage has been recommended. However, even an ePrmarked 20Z increase in the present rate probably would permit reasonable levels of productivity-oriented research (see para. 6.7) on both rice and development of better production and marketing systems for other field crops. Since agriculture contributes around 28Z to GDP, the research increment would amount to less than Rs 80 million per year.  - 45 - 6.35 In regard to food interventions, the cost of rice grading (see paras. 6.20-6.22) would be negligible, as screening costs would be offset by higher market prices for the better-quality offtake. Based on 1986-7 CFS data, careful targeting could reduce food stamps expenditure to around Rs 1,000 million per year, thereby saving Rs 2,600 million annually. Restricting the school lunch program to children in Standards I and II would reduce the cost of that program to Rs 1,000 million annually as a saving of Rs 1,600 million per year. Improving the thriposha program (see para. 6.25) involves the better use of existing levels of investment and recurrent resources rather than new outlays. 6.36 The addition of routine vitamin A administration for young children (see para. 5.31) and iodized oil for goiter control (see para. 5.30) would cost around Rs 2 million and Rs 1 million, respectively. 6.37 Based on figures recently prepared for family planning communica- tions, a substantial national nutrition education effort (see para. 6.13) could be carried out for around an estimated Rs 24 million per year. The incremental annual costs of the proposed community intervention project (see paras. 6.29-6.30) scaled up to national level would be around an estimated Rs 140 million using existing thriposha production. 6.38 While the agricultural research recommendations might take 5-8 years to bear fruit, the other interventions could product results over a shorter time frame. Rice grading would take one year of experimentation followed, if successful, by 18-24 months to be nationally operational. The health inter- ventions would take 2-3 years to show significant impact. Nutrition education would work in a 3-5 year time frame. The community nutrition initiative would take around 7 years to reach full national coverage. 6.39 Table 6.1 below shows current estimated costs of Sri Lanka's main nutrition interventions and the financial implications of adopting the program recommendations in this report. It highlights the considerable savings that could be obtained through better targeting of the food stamp and school lunch programs, as well as the relatively small cost of agricultural research, nutrition education and micronutrient interventions which could have large pay-offs in terms of improved nutrition status. 6.40 In sum, the proposed program package would be cost-effective as well as managerially and technically feasible. Its total cost of under Rs 3,000 million would be less than 40? of current nutrition outlays. Effective implementation would permit Sri Lanka substantially to improve the nutrition standards of its most vulnerable groups at a sustainable cost of less than 2? of the recurrert Government budget.  - 46 - Table 6.1: MAJOR NUTRITION INTERVENTIONS: mISUWRECoWmmENDATIONS AND COST TO GOVERNMENT Estimated Current Mission Incremental Annual Total Annual Annual Costs Recommendation Cost or Savings Cost (In S.L. Re '000) Existin E F-ood Sta a8,800,000 Effectively retarget to (2,600,000) 1,000,000 lowest three urban and lowest rural deciles ThriDosha 200,000 Tilghten up worker perf*r- 200,000 mance; use growth moni- toring more effectively; defer program expansion in favor of increased efficiency School Lunch 2,600,000 Limit beneficiaries to (1,600,000) 1,000,000 Standards I and II Kole Kenda negligible Focus on under-three negigible negligible children; Increase frequency and regularity of use Nutrition Villages 820,000 (at full Restructure as fully (180,000) 140,000 national coverage) integrated community nutrition and health program New Npricuitural Research 400,000 Research on rice 80,000 480,000 productivity and subsidiary food crops Nutrition Education negligible Focused program to change 24,000 24,000 specific behavior patterns Micronutrient Inter- negligible Regular supplementation 8,000 8,000 ventions (vitamin A of target groups iodized oil) Rice Grading negligible Test use of cosmetically negllgible negligible Inferior varieties Totals 7.120.000 (4.273,000) 2.897,000 /a Includes agricultural research and cost of nutrition villages at full national coverage.  -47 - ANNEX A- Tables  -48 - ANNEX A Table 1 SRI LANKA NUTRITION REVIEW Average Calorie Price by Expenditure Group. 1980/81, 1989 Per capita expenditure Calorie Price (Re/Calorie x 1.000) (Decile) Urban Rural Estate All Island 80-81 1989 80-81 1989 80-81 1989 80-81 1989 1 2.31 5.82 2.10 5.29 2.14 5.39 2.14 5.39 2 2.54 6.40 2.22 5.59 2.14 5.39 2.24 5.64 3 2.62 6.60 2.33 5.87 2.34 5.90 2.37 5.97 4 2.65 6.68 2.39 6.02 2.38 6.00 2.42 6.10 5 2.90 7.31 2.43 6.12 2.42 6.10 2.51 6.33 6 2.85 7.18 2.50 6.30 2.33 5.87 2.54 6.40 7 2.97 7.48 2.58 6.50 2.56 6.45 2.65 6.68 8 3.28 8.27 2.76 6.96 2.46 6.20 2.84 7.16 9 3.52 8.87 2.86 7.21 2.83 7.13 3.03 7.64 10 4.45 11.21 3.70 9.32 3.58 9.02 4.01 10.11 Total 3.24 8.16 2.55 6.43 2.46 6.20 2.67 6.73 Source: Sahn, David, 'Food Consumption Patterns and Parameters, in Sri Lanka: The Causes and Control of Malnutrition," Draft 2. June 1985, IFPRI, using data from the 1980/81 Labor Force and Socio Economic Survey, and updated to 1989 prices using Colombo Consumer Food Price Index.  ANNEX A Table 2 SRI LANKA NUTRITION REVIEW Per Adult Equivalent Calorie Consumption by Expenditure Decile by Sector, 1969170. 1978/79, 1980/81. and 1981182 (calories/adult equivalent/day) Per capita 1691970 1978/79 1980/81 1981/82 expenditure All All All All decile Urban Rural Estate Island Urban Rural Estate Island Urban Rural Estate Island Urban Rural Estate Island 1 1,848 2,183 2,209 2,166 1,668 1,749 1,712 1,730 1,477 1,613 1,584 1,587 1,521 1,670 1,617 1,588 2 2,192 2,482 2,691 2,500 2,089 2,142 2,432 2,147 1,853 2,076 2,088 2,047 1,771 2,062 2,186 2,031 3 2,361 2,780 3,009 2,774 2,213 2,368 2,765 2,378 2,099 2,309 2,322 2,280 1,982 2,326 2,584 2,306 4 2,481 2,989 3,178 2,984 2,340 2,568 2,963 2,575 2,282 2,626 2,217 .2,526 2,317 2,674 2,836 2,662 5 2,812 3,116 3,449 3,134 2,480 2,781 3,298 2,783 2,376 2,898 2,864 2,681 2,486 2,778 3,047 2,7B8 8 2,771 3,252 3,808 3,279 2,586 2,978 3,616 2,983 2,617 2,914 3,314 2,890 2,624 3,009 3,377 2,983 7 2,884 3,322 3,864 3,391 2,797 3,070 3,822 3,118 2,678 3,211 3,272 8,124 2,793 3,202 2,748 3,175 8 3,010 3,666 4,090 3,888 3,037 3,369 3,908 3,353 2,816 8,339 3,936 3,288 3,092 3,620 4,084 3,494 9 3,210 3,728 4,302 3,798 3,317 3,663 4,758 3,690 3,082 3,765 4,289 3,627 3,261 3,866 4,548 3,760 10 3,625 4,194 6,042 4,316 3,689 3,797 4,800 3,762 3,872 4,248 4,106 3,877 3,560 4,158 4,394 3,905 Average 2,913 3,121 3,452 3,180 2,756 2,784 3,548 2,852 2,629 2,807 2,994 2,791 2,796 2,828 3,344 2,866 Source: Sahn, David. 'Changes In Living Standards of the Poar In Sri Lanka During a Period of Macroeconomic Restructuring'. World Development, Vol. 16, No. 6, 1987.  - 50 - ANNEX A Table 3 SRI LANKA NUTRITION REVIEW Indices of Wages and Food Prices. 1982-87 Construction Wages La Agricultural Wages Food Prices /c (1982 = 100) (1982 - 100) (1982 - 100) 1983 120.59 118.34 112.41 1984 130.42 128.47 132.77 1985 140.17 133.09 132.85 1986 148.23 141.62 142.40 1987 159.04 139.61 154.75 la Wages for an unskilled carpentry helper; Sources Central Bank of Sri Lanka. lb Wages for paddy harvesting (male). /c Source: Colombo Consumers Price Index.  ANN A Table 4 SRI LANMA NUTRITION REVIEW Comparison of Annual Per Capita Intake for Particular Foods 1980-81: 1985-86 - Rural Eszenditure Deciles Dl D2 ______DIO__AveraDe 198Q-81 1985-d l9W-81 1985- '18-1 83-dOf i980-81 1985-80 198u-ni1 0965-6 Rural Rice 155.4 106.5 201.2 162.1 218.4 194.3 354.1 364.3 258.8 251.9 (lbs) 200.0 209.0 218.0 Bread 17.7 26.1 23.1 36.1 31.5 43.3 63.3 82.2 35.3 56.6 (lbs) 35.3 34.1 36.7 Sugar 150.9 187.6 220.6 274.8 269.5 327.6 681.3 890.2 360.0 489.6 (oz) 376.9 398.5 410.6 Estates Rice 140.7 103.1 181.8 142.4 236.3 179.9 415.5 396.3 259.1 234.1 206.5 223.3 242.4 Urban Rice 124.8 105.5 157.1 157.6 178.9 182.3 223.9 243.1 201.8 204.6 153.4 150.3 163.3 Bread 46.7 44.8 59.9 61.1 56.8 72.4 105.5 117.0 77.3 90.0 73.4 62.3 67.4 Sugar 174.9 272.0 240.9 359.5 267.9 490.1 641.7 944.7 427.5 653.6 538.2 435.3 485.1 Sources: Labor 'orce and Socio Economic Surveys 1980-81 and 1985-86 (preliminary tables). Bureau of Census and Statistics.  -52 - ANNEX A Table 5 SRI LANKA NUTRITION REVIEW Sir Lanka. Paddy Production. Imports and Government Procurement 1959-82 (millions of bushels and percentages) Share of Share of paddy government GPS/PMB imports as procurement Production Imports procurement percentage of as precentage (million bu) (million bu) (million bu) production of production (1) (2) (3) (4) (5) 1959 36.4 37.9 16.6 104 46 1960 43.1 38.8 20.8 90 48 1961 43.1 36.3 22.2 84 51 1962 48.0 30.4 27.0 63 56 1963 49.2 40.2 25.7 82 52 1964 50.5 39.9 29.4 79 58 1965 36.3 39.6 23.1 109 64 1966 45.7 35.4 28.0 77 61 1967 54.9 27.4 15.4 50 28 1968 64.5 25.1 14.9 39 23 1969 65.9 19.0 13.7 29 23 1970 77.4 38.4 26.2 50 34 1971 66.9 24.4 32.7 36 49 1972 62.9 24.7 26.4 39 42 1973 62.9 19.1 13.7 30 36 1974 76.8 21.7 20.9 28 27 1975 55.3 33.0 11.6 60 21 1976 60.0 30.5 12.9 51 22 1977 80.4 39.0 24.6 49 31 1978 90.6 11.5 32.3 13 36 1979 91.9 15.2 25.9 17 28 1980 102.2 13.6 10.1 13 10 1981 106.8 11.2 4.7 10 4 1982 103.3 11.5 4.0 11 4 1983 124.2 9.0 16.2 7 13 1984 121.0 1.9 8.5 2 7 1985 133.0 13.4 5.1 10 4 1986 129.4 16.2 7.7 12 6 1987 106.4 7.5 3.2 7 3 Source: Agricultural Statistics of Sri Lanka (1981), Central Bank data and Food Commissioner. Thorbecke E. and Svegnar, J., "Effects of Macroeconomic Policies on Agricultural Performance in Sri Lanka, 1960-82, OECD July 1985.  T*-