WDP-49 4 J E World Bank Discussion Papers Improving Nutnrtion in India Policies and Programs and Their Impact K. Subbarao ,'A RECENT WORLD BANK DISCUSSION PAPERS No. 1. Public Enterprises in Sub-Saharan Africa. John R. Nellis No. 2. Raising School Quality in Developing Countries: What Investments Boost Learning? Bruce Filler No. 3. A System for Eva'luatinq the Performance of Government-Invested Enterprises in the Republic of Korea. Young C. Park No. 4. Country Commitment to Development Projects. Richard Heaver and Arturo Israel No. 5. Public Expenditure in Latin America: Effects on Poverty. Guy P. Pfeffermann No. 6. Community Participation in Development Projects: The World Bank Experience. Samuel Paul No. 7. International Financial Flows to Brazil since the Late 1960s: An Analysis of Debt Expansion and Payments Problems. Paulo Nogueira Batista, Jr. No. 8. Macroeconomic Policies, Debt Accumulation, and Adjustment in Brazil, 1965-84. Celso L. Martone No. 9. The Safe Motherhcood Initiative: Proposals for Action. Barbara Herz and Anthony R. Measharn TAlso available in French (9F) and Spanish (9S)I No. 10. Improving Urban Employment and Labor Productivity. Friedrich Kahnerl: No. 11. Divestiture in Developing Countries. Elliot Berg and Mary M. Shirley No. 12. Economic Growth and the Returns to Investment. Dennis Anderson No. 13. Institutional Development and Technical Assistance in Macroeconomic FPolicy Formulat:ion: A Case Study of Toqo. Sven B. Kjelistrom and Ayite-Fily d'Almeida No. 14. Managing Economic Policy Change: Institutional Dimensions. Geoffrey Lamb No. 15. Dairy Development and Milk Cooperatives: The Effects of a Dairy Project in India. George Mergos and Roger Slade No. 16. Macroeconomic Policies and Adjustment in Yugoslavia: Some Counterfactual Simulations. Fahrettin Yagci and Steven Kamin No. 17. Private Enterprise in Africa: Creating a Better Environment. Keith Marsden and Tferese Belot No. 18. Rural Water Sup1 and Sanitation: Time for a Change. Anthony A. Churchill, with the assistance ot David de Ferranti, Robert Roche, Carolyn Tager, Alan A. Walters, and Anthony Yazer No. 19. The Public Revenue and Economic Policy in African Countries: An Overview of IssLes and Policy Options. Dennis Anderson No. 22. Demographic Trends in China from 1950 to 1982. Kenneth Hill No. 23. Food Import Dependence in Somalia: Magnitude, Causes, and Policy Options. Y. Hc;sein Farzin No. 24. The Relationship of External Debt and Growth: Sudan's Experience, 1975-1984. Y. HDssein 1Farzin No. 25. The Poor and the Floorest: Some Interim Findings. Michael Lipton No. 26. Road Transport Taxation in Developing Countries: The Design of User Charges and "aKes for Tunisia. David Newbery, Gordon Hughes, William D.O. Paterson, and Esra Bennathan No. 27. Trade and Industrial Policies in the Developing Countries of East Asiai. Amarendri 13hattacharya and Johannes F. Linn No. 28. Agricultural Trade! Protectionism in Japan: A Survey. Delbert A. Fitchett No. 29. Multisector Framework for Analysis of Stabilization and Structural Adjustment Polcies: The Case of Morocco. Abel M. Mateus and others No. 30. Improving the Quality of Textbooks in China. Barbara W. Searle and Michael Mertaigh with Anthony Read and Philip Cohen (Continued cn the insidl back covE . - 49K z World Bank Discussion Papers Improving Nutrition in Idia Policies and Programs and Their Impact K. Subbarao The World Bank Washington, D.C. Copyright C( 1989 The World Bank 1818 H Street, N.W. Washington, D.C. 2043:3, U.S.A. All riglhts reserved Manufactured in the United States of America Secondl printing May 1990 Discussion Papers are not formal publications of the World Bank. They present preliminary and unrpolished results of country analysis or research that is circulated to encourage discussion and comment; citation and the use of such a paper should take account of its provisional character. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Any maps that accompany the text have been prepared solely for the convenience of readers; the designations and presentation of material in them do not imply the expression of any opinion whatsoever on the part of the World Bank, its affiliates, or its Board or member courntries concerning the legal status of any country, territory, city, or area or of the authorities thereof or concerning the delimitation of its boundaries or its national affiliation. Because of the inforrmality and to present the results of research with the least possible delay, the typescript has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to Director, Publications Department at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normailly give permission promptly and, when the reproduction is for noncommercial purposes, without askinLg a fee. Permission to photocopy portions for classroom use is not required, though notification of such use having been made will be appreciated. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list and indexes of subjects, authors, and countries and regions; it is of value principally to libraries and institutional purchasers. The latest edition of each of these is available free of charge from Publications Sales Unit, Department F, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Plablications, The World Bank, 66, avenue d'Iena, 75116 Paris, France. K. Subbarao, a professor of economics at the Institute of Economic Growth in New Delhi, is a consultant to the World Bank. Library of Congress Cataloging-in-Publication Data Subbarao, K. IMproving nutrition in India. (W:rld Bank discussion papers ; 49) Bibliography: p. 1. Nutrition--Government policy--India. 2. Malnutri- tion--India. I. Title. II. Series. TX360.I4S83 1989 363.8'0954 89-5556 ISBN 0-8213-1192-1 iii ABSTRACT This paper attempts a disaggregated analysis of existing data sets pertaining to nutrition in India and identifies the priority geographical areas, social and age groups for possible nutrition interventions; analyzes the effectiveness of public responses to the nutrition problem as reflected in the distribution of expenditures, and the impact of one major direct and one indirect nutrition intervention; and discusses the outstanding nutritional issues. The broad conclusions are that there is substantial empirical evidence to suggest wide variations in the distribution of malnutrition across regions, age and social groups and by gender; that the direct nutrition expenditures remained modest in the past, and that the distribution of even the modest expenditures has not at all been sensitive to the observed variations in the intensity of malnutrition; and that there appears to be considerable scope for improving the productivity of expenditures by strengthening and reorienting the existing programs and also lby reducing the prevailing mismatch between expenditures and what we know about the distribution of need. Overall, the paper argues the need for striking a proper balance between needs, potential demand and available resources. This paper was written as a background paper for the 1989 Country Economic Memorandum for India, when the author was working in the Country Operations Division of the Bank at their Resident Mission in New Delhi. ACKNOWLEDGEMENT I wish to thank Sajitha Bashir for painstaking assistance and unfailing intellectual involvement in this research project. During the early' stages of this study, I have benefitted from discussions with C. Gopalan, N. Prahlad Rao, R. Radhakrishna, G. Sastry and Urmil Sharma. For many constructive comments and suggestions on earlier drafts, I am indebted tc Bela Belassa, Alan Berg, Pe?ter Berman, Nat Colletta, David Greene, James Hanscn, Sawon Hong, Richard He!aver, Samuel Lieberman, Michael Lipton, Dan Ritchic and Richard Skolnik. Successive drafts of this paper were cheerfully and competenitly typed by J'asvir Bhasin. I am solely responsible for the viesWs expressed in this paper. Table of Contents Page No. I Introduction 1 II Distribution of Malnutrition by Region, Social Status, Age and Sex ............................ 3 a) Household-level average nutrient intake by region and social status .................... 3 b) Incidence of malnutrition among children in different regions and social groups ......... 12 c) Distribution of Malnutrition by Sex ............ 21 III Public Policy .................................. 24 a) Size and Distribution of Direct Nutrition Expenditures ................................... 26 b) Expenditures and Sustainability of interven- tions: A case study of Andhra Pradesh .......... 34 IV The Effectiveness of Major Direct and Indirect Nutrition Interventions ........................ 38 a) Integrated Child Development Services (ICDS). 38 b) Public Distribution System ..................... 66 V ConcLusions and Implications for Policy ........ 74 References Appendices List of Tables Table No. :?ave Hlo. 1. State-vise calorie intake (Kcal/cu) 1975-80 Average (Ruiral) by Social Class 4 2. Calorie Consumption of the ultra-poor and the Average: Mean and Growth rate 1975-80 7 3. Summary Nutritional Status of Households in Rural Bihar by Social Class (% Distribution).... 9 4. Intake of E:nergy as % of RDA in ICRISAT villages (Lean and Surplus Seasons, 1976-77).... 10 5. Mean Energy Intakes in Rural and Urban India, 1987 (per Cu per day) .12 6. Nutritional Status of Children (1-5 years) in the selected states by weight-for-age (%) 13 7. Summary Nutritional Status by Age (% Distribution) in Bihar Villages 15 8. Percentage of Malnourished (Gr. III + IV) children in IC'DS Projects by Caste Status, 1981. 16 9. Percentage of Malnourished (Gr. III + IV) children (0-36) months of Scheduled Castes and Tribes in ICDS Projects of location, 1981 17 10. Regional Profile of Malnutrition among Children 19 11. Calorie Inadequacy among Adult Males and Females (Percent of Population) .23 12. State-vise D)istribution of Nutrition Expendi- tures ............................................ 28 13. Outlay and Expenditure - Seventh Plan - 1985-90 Nutrition .30 14. Statevise Distribution of Children fed under various feeding programs....................... 15. Percentage of Severely Malnourished children in ICDS Project Areas (Base line) in 1984 and corresponding State averages (1982) ..2 16. State-vise Position of ICDS 1986-87 ............ 43 17. Impact of ICDS on Nutritional Status Children in grades III & IV malnutrition 47 18. Distribution of ICDS Project Areas by changes in the StaLtUS of malnourished (Gr. III + IV) children: 1979-80/80-81 over 1984-85 .48 19. Characteristics of some ICDS Project Areas Deterioration .50 20. Supplementary feeding under ICDS by Social Class, 1981 ......... : ........................ 52 21. Staff Position - Paramedical Functionaries in ICDS PHCs reported as on 31st March, 1987 54 22. Staff Position - Health Functionaries - Medical Officers ICDS PHCs reported as on 31st March, 1987 55 22. ICDS: Recruitment & Training of Personnel in Central Sector Projects: 1987 57 23. PDS Supplies and Population below Poverty line. 70 I. Introduction 1.1 Since India's independence, the country's major development effort has been to improve the living standards of its poor. Substantial progress has been made on a number of fronts, the most fundamental being a large increase in life expectancy. This is due not only to a reduced incidence of number of major diseases, but also to better food availability and distribution, of which the most obvious manifestation has been the complete elimination of famines in times of even severe droughts such as the last year's. While, over several decades, the poverty incidence has not shown a clear trend, the Seventh Plan reports that the latest data indicate a fall from 48.3% in 1977 to 37.4% in 1983. Any discussion of India's still pervasive malnutrition problem must begin by recognizing these achievements. 1.2 The nutrition status of India's population and the methodological issues pertaining to its measurement have been the centre of controversy and much debate in India.1 There has, however, not been a systematic delineation of the distribution of nutrition levels in India by region, social status, age and sex. Since the policy debate on nutrition interventions in India has so far been conducted with a data base fairly aggregated at all-India level or state-level, there has also not been a critical assessment of the diverse state-level responses to the nutrition problem. This paper aims to fill these gaps. tThe principaL contributors to the debate were Sukhatme (1982) and Dandekar (1982). For a good, recent review of this debate, see Kakwani (1986 a). Depending upon the methodology chosen, the extent of malnutrition is estimated at 50 to 25 percent of population. Most earlier studies estimate the extent of malnutrition by taking the "household" as a unit of observation. Kakwani's estimate based on individuals, rather than households, came close to 50% for the year 1971-72. In this paper, we followed the usual 2400 calorie norm for adults and the Gomez classification and the Indian Academy of Paedetricians (IAP) classification for children. These have been defined at relevant places. It may be noted that relative distribution by region, age, sex and social status attempted in this paper will not be affected by the choice of the methodology for measuring malnutrition. 2 The paper begins with a review of diverse data sets. The disaggregated analysis identifies the priority geographical areas and social groups for nutrition interventions (Part II). This is followed by an ara:.ysis of the various public responses to the nutrition problem, done on the basis o:. th(e following criteria: ( a) whether or not the distribution of expenditures corresponded to needs, and whether or not the interventions are slustainable :in the long run; (Part III); and (b) whether or not the major interventions in operation for a sufficiently long time have been successful in reaching the identified malnourished groups/regions (Part IV). Most of the analysis concentrates OnL two programs: the Integrated Child Development Services (ICDS), by far the most extensive dlirect nutrit:ion intervention attempted in India since Independence; 2 and the Pufblic Distribution System (PDS), an indirect approach for impr:cving nutrition that involves large Central and state budget expenditures and whose importance in policy clebates has grown following the drought of 1987. PooI- reliability of existing data base makes it difficult to assess fully the expenditure effectiveness of these programs in terms of physical performlance. Hence, physical performance data are suitably supplemented by a qualit:ative assessment of the chosen interventions. The last section contains a summolry of the main findings, and a discussion of the outstanding nutritional issues What need to be resolved in the near future. 21 am aware that ICDS has objectives other than nutrition such as pre-schoot, and my description of the program as a "mntrition intervention" is an approximation. A more detailed description of the program is given in Section IV. 3 I. Distribution of Nutritional status by Region. Social Status. Aae and Sex. To know that malnutrition is widespread and severe in India hardly requires a detailed analysis of data. But to know the appropriate means for reducing it, and to know where and when the means will be effective, requires a disaggregated analysis of the data. This section attempts such an analysis. (a)Household-level average nutrient intake, by region and social status 2.1 The data generated by NNMB, after due correction so as to render them comparable over time and across regions,3 are very useful for estimating the mean intakes of nutrients for various classes of households in different parts of rural India. One problem with NNMB data is they are somewhat dated - the last published survey is for 1982; the disaggregated data are available only up to 1980-81.4 The results of the combined NSS-NNMB survey of 1983 have not yet been released. We used the disaggregated NNMB data sets for the years 1975-1980. Of the ten states for which NNMB data are available, Tamil Nadu experimented with a number of direct nutrition interventions, the most important being the Chief Minister's Nutritious Meal Program (CMNMP) for school children, since 1982, and Tamil Nadu Integrated Nutrition Project for pregnant and lactating women and children of 6-36 months age in 9 districts since 1982-84. In 1983, Andhra Pradesh introduced a major indirect intervention (i.e., Rs. 2 a kilo rice scheme). The 3The representativeness of the NNMB Sample is ensured by a rigorous multistage sampling procedure evolved by Dr. C. R. Rao, the then Director of the Indian Statistical Institute. The major criticism of NNMB data has been that it has suddenly switched from Harvard scale of normalcy to the Hyderabad (Indian) scale of normalcy since 1976. The data examined in this paper are derived using the some scale of normalcy over the entire period 1975-80. For an excellent summary of the sampling procedure, and the particular merits of this data source for drawing inferences regarding the regional distribution of malnutrition, see Srikantia (1988). 4It must be noted that most other data in India are also dated. * Nationat Nutrition Monitoring Bureau 4 relative nutrition situation among the remaining states may not have changed much since 1980. The drought of 1987 probably worsened the nutrition situation of the ultra poor (the bottom two deciles) - but this cannot be empirically demonstrated at this time, given the time lags in the release of nutrition data. We used NNMB data because they are still very relevant: not only not much has changed in many states but also the social-class disaggregation is rich enough. to desel:vi art analysis. 2.2 The calorie intake (Kcal/cu) data averaged for all households for the 1975-80 period for ten states are presented in Table 1, along with the Table 1 State-wise calorie intake (Kcal/cu) 1975-80 Average (Rural) by Social Class Calorie Intake of Calorie intake of Scheduled Castes Scheduled Tril:e3 as percent of as percent of State State Recommended State Recommendled States Average Average Daily Allowance Average Daily All.owance (2400) (2L01)) Karnataka 2837 86.6 102.4 SS Andhra Pradesh 2517 96.3 101.0 SS - Orissa 2324 94.1 91.0 96.3 93.3 Tamil Nadu 2293 88.7 84.7 SS - Maharashtra 2286 85.6 31.5 98.1 93.5 West Bengal 2227 98.7 91.5 92.2 85.5 Gujarat 2211 98.0 90.3 92.4 85.1 Madhya Pradesh 2160 92.5 83.3 89.4 80.5 Uttar Pradesh 2123 97.2 86.0 SS - Kerala 1942 94.1 67.4 SS SS: Small Samjple Source: Computed from disaggregated NNMB data. 5 averages for the scheduled castes and tribes.5 Communities belonging to these castes and tribes (so called because these castes and tribes were listed in a Schedule of the Indian Constitution) have historically suffered from social and economic deprivation. Prior to Independence, some of these castes were regarded as "untouchables". Although untouchability has long since been abolished, these castes and tribes continue to be at the bottom of the social hierarchy in India, and as such, deserve a separate treatment from a nutritional perspective. Two points deserve to be stressed. First, the average calorie intake was lower than the recommended allowance of 2,400 Kcal/cu in all states except Karnataka and Andhra Pradesh. Second, the average Kcal/cu conceals the fact that in all the states the Kcal/cu of the scheduled castes and tribes was lower than the state average which itself was lower than the recommended allowance. In particular, the intakes of scheduled castes in Kerala, Maharashtra, Tami'L Nadu, Madhya Pradesh and Uttar Pradesh, and tribes in Madhya Pradesh, Gujarat and West Bengal, were alarmingly low. Considering that the decline in the poverty ratio was the lowest among these social groups (see C.H.H. Rao), their nutrition probably has not improved much since 1980/81. 2.3 The relatively high average intake for Karnataka, which has a high poverty ratio, and a substantial area in India's arid (dry) zone, remains a puzzle. The very low calorie average for Kerala shown in the NNMB data will remain a puzzle especially because the infant mortality rate and the death rate are the lowest in this state. Recent evidence, however, suggests the 5In this paper, we did not analyse the time trends averaged for all-India or for states, since our focus is on a disaggregated analysis. These time trends averaged for all sample households within each state, and for all ten states together, have been analysed by NNMB themselves, who noted a steady decli-e in the incidence of malnutrition between 1974 and 1982. Their findings, although disputed by many, are still in concordance with the poverty incidence data from the 1977 and 1983 National Sample Surveys. 6 prevalence of high levels of malnutrition and morbidity in this state despite a reduction in infantinortality (see Panikar and Sonam). Kerala's very pocr agricultural performance (registering a negative rate of growth recently) and very high Eree market retail rice price (40% higher than all-India rice pi-Ice) could also have contri'buted to the observed low calorie consumption in Ke:-ala. 2.4 Disaggregated NNMB data are also available according to per capita income (Rs. per day) groups. From these data, we estimated the trends in calorie initake for the 'ultra poor', (roughly corresponding to 65% of the poverty-level income) and compared with the average for all income groups. 6 These are shown in TabLe 2, along with the relevant socio-economic indicators. Our interest is primarlly in knowing whether or not the ultra poor's nutr:i.ent intake has improved; and if so which states registered a better performance.7 2.5 Data pooled for all the states (see Table2) shows no trend increase in the calorie-intake of the ultra-poor during 1975-80, although for all income groups, shows a modest trend increase.8 Again, Karnataka and Andhira Pradesh show high average for the ultra poor, but no trend growth.9 By ccn:rast, Kerala's ultra poor have a low average calorie intake but this has shown cl substantial and significant trend increase during 1975-80, which was higher than average. It is possible that Kerala's public distribution systen (see P.1S. George) may have contributed to the observed positive trend growth in nutient 6Since the data correspond to discrete income groups, the cut-off points for the ultra poor are not exactly 65%. 7We are aware of the problem arising out of short time span. Fortunately the initial and terminal years were "normal" years. It must be stressed that we are using the trend growth rates essentially for cross-sectional (inter-state) comparisons. 8See also footnote 5. 9It may be noted that NNMB's findings of high average calorie consumption by the ultra poor in Karnataka is contradicted by NSS data for the bottom three deciles for the same state for comparable time period (See Gupta, 1987). 7 intake of the ultra poor, partly mitigating the adverse effects of the high rural free market prices of rice noted above. Table 2 Calorie Consumption of the untra-poor and the Average: Mean and Growth Rate 1975-80, along with state-wise socio-economic indicators State Category Mean Growth Per capita Total Popu- Literacy Urban Net Sown Calorie rate income Popu- lation rate Popu- Area Intake 1975- at current lation den- 1981 lation Irrigated 1975-80 80 prices 1981 sity 1981 1983/84 in Rs 1983 (mill) 1981 (M) (M) (%) Pooled Ultra-poor 2255 Nil (9 states) Average 2352 0.8* Kerala Untra-poor 1723 3.3* Average 1942 1.6 ns 1761 25.5 655 70.4 18.7 11.9 Tamil Ultra-poor 2052 2.0* Nadu Average 2293 1.7* 1827 48.4 372 46.8 32.9 42.9 Karna- Ultra-poor 2401 Nil taka Average 2517 Nil 1957 37.1 194 38.5 28.9 14.3 Andhra Ultra-poor 2756 Nil Pradesh Average 2837 Nil 1955 53.5 195 29.9 23.3 32.0 Maha- Ultra-poor 2229 1.1* rashtra Average 2286 Nil 3032 62.8 204 47.2 35.0 10.5 Gujarat Ultra-poor 2117 1.8* Average 2211 1.5* 2795 34.1 174 43.7 32.2 22.3 Madhya Ultra-poor 1972 Nil Pradesh Average 2160 1.7 ns 1646 52.2 118 27.9 20.3 14.0 West Ultra-poor 2227 1.3 ns Bengal Average 2277 -1.7 ns 2231 54.6 615 40.9 26.5 33.0 Uttar Ultra-poor 2065 -2.6* Pradesh Average 2123 -1.7* 1567 110.9 377 27.2 17.9 57.4 * Statistically significant ns Not significant Source: Computed from disaggregated NNMB data sets. 2.6 Apart from Kerala, in Tamil Nadu the ultra-poor's calorie intake also registered a significant trend increase. The average for all income groups also suggests a trend increase in Tamil Nadu. Both Maharashtra and Gujarat have registered a trend increase for the ultra-poor. In the states of Madhya Pradesh, West Bengal and Uttar Pradesh, the ultra-poor did not register any trend growth in calorie intake. In Uttar Pradesh, the ultra-poor and the average all showed a 8 statistically significant negative growth over the period 1975-80. 2.7 The differences in calorie intake across the states reflect differences in such factors as food lprice levels, the extent of access to publicly distributed food, differences in foocd habits and quality of food consumed, the general ec:onomic performance of the state, etc. Quite clearly there appear to be forces op-rating in the states of Kerala, Tamil Nadu, Maharashtra and Gujarat contributing so an improvement in the calorie intake of the ultra poor, although their level is 1bwest imn Kerala. The present stat:e of knowledge does not enable one to systematical:Ly analys,e the factors contributing to the improvement of nutritional status of the ultra poor Iln these states. 10 One can only hazard a guess. The first two states are wel;l known for their twin programs of noon meals and an effective public distribution system; Maharashtra for her food--wage linked Employment Guarantee Scheme, and Gujarat for her more than average economic performance. 2.8 Recent studies have shown that rural poverty is heavily concentrated in the arid zone and in the eastern states and among the scheduled castes.(C.H.E. Rao, 2R. cit.) An assessment of the nutrition situation in Bihar can be done wit'i data on a sample of rural Bihar households in twelve villages spread over all the six regions of thLe state for the year 1981 drawn from recent Working Papers oF1 ILO, While there were inter-v:lllage differences in average calorie intakes, the study noted that there was little evidence of higher intakes in the more advanced viliages, 10At the Level of aggregation of states, the number of observations are too small for statistical anal isis. ALso, some of the states (e.g., Uttar Pradesh) are very large, with considerable intra-state varia:i)n. District-wise calorie-intakes for the ultra poor and average are available for 1980-81, a normal year. An attempt was made to explain intetr-district variation in the calorie-intake (Appendix A). The uLtra-p,)o-'s calorie consumption was positively correlated with the per capita food production of the district, and the eiel of development of the district. Data for more districts and overtime are needed for more reliable analysiis of inter-district variation in caloi-ie intakes. 9 Social-class differences in calorie intakes appeared more prominent than either regional or seasonal differences. According to this source, nearly a half of the landless agricultural labour households, the majority of whom belong to the scheduled castes, suffered from varying degrees of malnutrition. (See Table 3). The sharp decline in the percentage of acutely malnourished between the landless labor Table 3 Summary Nutritional Status of Households in Rural Bihar by Social Class (% Distribution) * Class Normal Wasted Stunted Acute N Agricultural Labor 43.7 31.0 15.1 10.2 270 Agricultural Labor tied 45.6 28.6 14.5 11.3 103 Poor-Middle peasant 57.6 27.9 9.3 5.2 73 Middle peasant 75.4 17.1 4.9 2.0 90 Big peasant 57.3 31.1 8.3 3.3 243 Landlord 70.6 20.6 8.4 0.3 164 Non-Agricultural/ No Activity 45.2 31.4 11.8 11.6 49 All 53.5 28.4 11.3 6.8 992 * The norms used by the authors for defining "normal" "stunted", etc are as follows: Height for age Weight for age over 85% over 85% "normal" less than 85% over 85% "stunted" over 85% less than 85% "wasted" less than 85% less than 85% "acute" Source: P. H. Prasad. et al, "The Pattern of Poverty in Bihar" (World Employment Programme Research), Working Paper No. 152. class and small peasant class suggests the vulnerability of the former class to malnutrition. The substantial variations in calorie intake between socio-economic groups suggest that these are not be due to mere statistical fluctuations, but reflect differences in the groups' command over resources. 10 2.9 For the arid zone in Maharashtra and Andhra Pradesh, ICRISAT data provide information on nutrient intake in four villages in lean season and in surplu;s season for the mid-1970s. Thes(e data are retabulated and presented in Table 4. Four points Table 4 Intake of Energy as % of Recommended Daily Allowance (RDA) in ICRISAT villages in Lean and Surplus Seasons, 1976-77 Nutrient Intake in the Arid Zone: Percentage of Sample Consuming Less than 50% of RDA Farm-size Age Group Aurepalle Dokur Sholapur Akola group (in years) ------------- ------------- -------------- ----- ------- Lean Surplus Lean Surplus Lean Surplus Lean Suirplus Landless 1-3 46.2 29.4 42.9 26.7 76.5 44.4 27.'1 40.0 Laborers 4-6 40.7 34.8 0.0 8.7 77.5 20.0 0.0 10.2 and small 7-12 20.0 15.8 7.7 0.0 87.0 7.1 0.0 7.8 farmers 13-18 Males 0.0 10.0 0.0 9.1 88.9 22.2 0.0 2.4 13-18 Females 7.7 5.3 0.0 5.3 90.2 6.7 0.0 6.3 >18 Males 7.3 15.6 0.0 6.6 87.1 4.0 1.61 2.6 >18 Females 23.9 15.9 2.9 2.9 95.2 10.4 2.0 3.0 Medium 1L-3 14.3 30.0 0.0 25.0 75.0 30.8 20.0 36.B farmers 4-6 33.3 11.1 0.0 0.0 78.1 7.7 0. C 4.5 7-12 8.3 0.0 0.0 8.3 95.5 0.0 0.0 10.3 13-18 Males 20.0 0.0 0.0 0.0 94.4 0.0 O.C 0.0 13-18 Females 0.0 0.0 0.0 0.0 95.8 0.0 O.C 0.0 >18 Males 11.5 7.1 22.2 0.0 89.8 0.0 o.C 1.4 >18 Females 4.2 0.0 0.0 3.2 94.9 2.9 0.0C 4.3 Large 1-3 100.0 0.0 40.0 25.0 78.3 28.6 22.2 44.8 farmers 4-6 0.0 0.0 0.0 13.0 58.8 0.0 0.0 23.3 7-12 17.4 0.0 0.0 14.8 81.8 0.10 3.6 6.8 1:3-18 Males 20.0 0.0 0.0 3.8 91.5 0.0 0.0 14.3 13-18 Female 10.0 10.0 0.0 0.0 85.4 0.0 0.0 7.7 >L8 Males 19.4 5.3 0.0 0.0 83.6 0.0l 2.9 9.5 >18 Females 7.7 0.0 0.0 6.3 90.7 2.6 0.0 2.9 Source: Data retabulated from ICRISAT Research Bulletin No. 7, 1984. deserve to be noted. First, as in Bihar, the percentage of the energy-de:lizient individuals among landless labor and small farm households in all age groatp;; was * International Crops Research Institute for Semi-Arid Tropics. 11 quite high. Second, the proportion of energy-deficient persons varied across villages as it did in Bihar (the highest being in Sholapur, Maharashtra), and across age groups, with a high proportion among children in the age group 1-3 of landless labour households in all the villages. Third, the percentage of energy-deficient people was lower in the surplus season. Fourth, a high proportion (ranging from 25 to 45 percent) of children in 1-3 age group in all villages and among all social groups, rich as well as poor, and in both lean and surplus seasons, were having energy intakes of less than 50% of recommended daily allowance."1 The last aspect is discussed further in the next section, which focuses on malnutrition among children. 2.10 We have so far examined the location of malnutrition among the rural households. It would be useful to compare these with the nutritional status of social groups in urban India. A summary position of rural and urban India is presented in Table 5. Three points deserve to be noted. First, the urban groups in general were no better off than rural groups. Second, the energy intakes of urban slum dwellers were no better than those of rural landless and Harijans. Third, the rural rich (10 acres and above) appeared to have higher intakes than the urban high income groups. There could be many reasons for this difference: the easier accessibility of food to the rural rich, higher calorie-consciousness of the urban rich, etc. t1The very high numbers for the Lean season in Sholapur and for 1-3 age children of large farmers in Aurepalle are very puzzling, and as such undoubtedly raise questions regarding reliability of this data source - the only source available for the arid zone. 12 Table 5 Mean Energy Intakes in Rural and Urban India, 1987 (per Cu per day) Urban 1 /a 2 /b Rural HIG 2000-3085 /c 2600 10 acre Land Holding 2375-3100 2000-2675 /d 2420 5-10 acre 2100-2860 Land Holding Landless 1865-2310 MIG 1880-2715 2365 LIG 1760-2665 2230 IL 1900-2510 2245 SL 1760-2290 2010 E'arijans 1600-2460 ======-- -- --- - - = = _ _ ==_= = _ = = = = _ = __ _= = === = = :- /a Range oi- means for various cities. /b Pooled average for the group. Zc Means of: various cities /d Trivandrum, Bangalore excluded. Legend: HIG = High incorme group MIG = Medium income group LIG = Low income group IL = IndustriaL Labour SL = Slum DwelLers Source: NNMP, as reported in Kamala S. Jaya Rao (1985) b) Incidence of malnutrition among children in different Regions and social groups. 2.11 Information on the nutritional status of children in the age giD2ip 1-5 years old is also available from the surveys carried out by NNMB. While thcsie data are available at the state-level, unfortunately the social--class disaggreget:on of malnourished children is not available. State-wise data are presented in Table 6^ These data show substantial improvements in the nutritional status of childrea onLy in two states, viz, Kerala and West Bengal. In all other states, the ahana2>t over time were marginal. As for inter-state differences, severe malnutritio>: amoag children seemed higher in Andhra Pradesh, Gujarat, Karnataka, Madhya PE^z.deslh, Maharashtra and Orissa. But these are state-level averages with their )liviou s 13 limitations. Yet, can we learn anything more from these state-level average figures? 2.12 In Figure 1,* the calorie-protein adequacy of all household members is juxtoposed against those averaged for children 1-4 years of age, for ten major states which NNMB surveys covered. In a number of states, the percentages of children with adequate calorie-protein intakes were much lower than the corresponding percentages for households. The map also shows that even in the agriculturally (and otherwise) prosperous states like Andhra Pradesh, nearly a half of the children were suffering from calorie-protein inadequacy; the proportion was much higher in the poor eastern states of Orissa, West Bengal, Table 6 Nutritional Status of Children (1-5 years) in the selected states by weight-for-age (percentage) La Normal Mild Moderate Severe State 1975-79 1982 1975-79 1982 1975-79 1982 1975-79 1982 Kerala 16.4 31.8 45.3 49.3 33.5 17.4 4.8 1.5 Uttar Pradesh 18.8 - 48.3 - 26.8 - 6.1 - Tamil Nadu 14.6 16.1 43.6 44.1 35.3 34.6 6.5 5.2 West Bengal 10.6 21.3 43.0 55.7 39.6 23.0 6.6 0.0 Karnataka 9.9 13.8 43.0 43.4 39.5 37.2 7,7 5.6 Andhra Pradesh 13.3 12.6 42.2 43.0 36.7 38.5 7.8 5.9 Orissa 15.3 13.0 42.1 35.8 33.3 42.3 9.3 8.9 Gujarat 9.8 11.7 37.8 28.7 43.0 44.4 9.4 15.2 Maharashtra 9.3 13.6 37.4 38.8 42.0 40.7 11.3 6.9 Madhya Pradesh 12.0 - 36.3 - 40.2 - 11.5 - ,a Gomez classification, as shown below: < 60% weight for age: Severe 61-75% weight for age: Moderate 76-90% weight iEor age: Mild > 90% weight for age: Normal Source: National Nuatrition Monitoring Bureau as quoted in UNICEF (1983) and Uttar Pradesh. Economic development of a state does not seem to protect children It may be noted that NNMEB's figure pubLished in 1980 gave data for chiLdren 1-4 years, whereas in later years data were released for children 1-5 years. 14 CA,LORIE-PROTEIH ADEQUACY OF HOUSEHOLDS - INDIA Source: UNNB Reporl (19801 Hlh', iydefabod, India (~~~~~~ ) NoE7 QzDL;buL /-yV/iX297eJ- UTTAR I) S/TAPRADESH f U '9zw ~~~~MADHYA F t * / TAMIL X0 -50 . -3ATA PR OT AEIJkLAELECY 16Ii'ncli ic nsliiirto of Nutritior.. TaTriobo7 HyderdDH PE, Indio. as given zin Anderson (1987) 15 from malnutrition to the same extent as it does for adults. In the arid zone (excluding Rajasthan for which we have no information) nearly 60% of the children suffered from calorie-protein inadequacy. Of the remaining states, sample survey data for 1981-83 on Bihar's children (see Table 7) showed that nearly a half of the children aged less than 54 months suffered from a nutritional deficiency; the proportion rose sharply to 78% for children of the age group 6-18 months. Although the nutritional status of 53.5% of all members of the households was normal, only 22.8% of the children in the age group 6-18 months was normal; acute malnutrition was also concentrated in this age group. Table 7 Summary Nutritional Status by Age (% Distribution) in Bihar Villages* Age Normal Wasted Stunted Acute N 6 Months but less than 18 22.8 42.5 22.8 11.9 41 18 Months but less than 54 36.7 33.1 19.3 10.9 248 54 Months but less than 114 52.5 31.0 9.9 6.6 419 114 Months or more 76.3 17.9 3.7 2.2 279 All 53.5 28.4 11.3 6.8 992 * For definitions of "normal", "wasted", etc., See footnote to Table 3. Source: P. H. Prasad, et.al., "The Pattern of Poverty in Bihar" (World Employment Programme Research) Working Paper No. 152. 2.13 Because the data depicted in Figure 1 are state-wise averages, it is not possible to infer about the nutrition status of children of specific social groups. For example, within large states such as Madhya Pradesh and Andhra Pradesh, there is a significant concentration of tribals in some districts; in Maharashtra the proportion of urban poor is high in the western districts. These categories would need a separate (disaggregated) treatment at the district level. Similarly, the arid zone within Karnataka, Maharashtra and Andhra Pradesh also deserves a separate treatment. Fortunately, the data on the third and fourth-degree malnutrition among 16 children collected by the ICDS monitoring authorities and the data on ICRISAT villages reported earlier (Table 2) can be used for this purpose. 2.14 The ICDS monitoring authorities have collected information on the nutritional status of children by caste status and by location. These are available for 1981 at the leve:L of aggregation of a state for 15 states, ard at a district-levrel for 1984-85, and separately for children 0-3 years and 0-l. years. These are retabulated and presented in Tables 8, 9 and 10. Table 8 Percentage of Malnourished (Gr. III + IV) children in ICDS Projects by Caste Status, 1981 La Average Scheduled Castes Scheduled Tribes Sn. 0-36 0-72 0-36 0-72 0-36 0-72 No. State months months months months moniths monl:hs 1 ANDHRA PRADESH 9.6 8.6 10.4 8.3 9.9 7.5 2 BIHAR 31.8 31.7 39.5 40.9 -- - 3 GUJARAT 7.3 6.2 6.0 3.9 11.7 - 4 HARYANA 4.6 3.5 5 HIMACHAL PRADESH 5.3 4.3 7.0 9.3 6 KARNATAKA 8.8 8.3 10,1 8.5 5.0 2.5 7 KERALA 7.7 7.8 11.0 10.2 17.5 15 6 8 MADHYA PRADESH - - - - 24.3 12 7 9 MAHARASHTR.A 15.8 13.3 16.7 14.8 23.7 20 7 10 ORISSA 16.7 13.0 19.0 16.8 11 PUNJAB 8.6 8.2 13.9 12.3 12 RAJASTHAN 8.2 8.7 17.3 12.1 8.1 7.6 13 TAMIL NADIJ 8.1 6.4 10.1 7.1 14 UTTAR PRADESH :L3.1 10.5 17.1 13.2 16.3 13.4 15 WEST BENGAL :L9.9 17.3 26.5 21.3 17.0 12.1 la ICDS authorities follow the Indian Academy of Paediatrics (IAP) classific:aAion, as shown below: < 50% weight for age: Grade IV malnutrition 51-60% weight for age: Grade III malnutrition 61-70% weight for age: GradeII malnutrition 71-80% weight for age: Grade I malnutrition > 80% Normal Source: Compiled from Child in India. A Statistical Profile, Ministry of Welfare, Government of India. * Integrated ChiLd DeveLopment Services. 17 Table 9 Percentage of Malnourished (Gr. III + IV) children* (0-36) months of Scheduled Castes and Tribes in Rural/Uraban/Tribal ICDS Projects, 1981 (%) Sn. No. State Average Rural La Tribal ab Urban /c 1. Andhra Pradesh 9.6 11.0 10.8 10.3 2. Himachal Pradesh 5.3 8.1 12.4 26.3 3. Maharashtra 15.8 27.3 14.3 16.4 4. Uttar Pradesh 13.1 30.8 13.9 20.8 5. West Bengal 19.9 16.2 33.8 /a and /c Children of Scheduled castes Lb children of scheduled tribes * IAP classification (defined in footnote to Table 8). Source: Compiled from Child in India. A Statistical Profile, Ministry of Welfare, Government of India. Care was taken to see that the ICDS Projects for which information was tabulated for 1980-81 were those which became operational during the period 1978-80, so that the data are comparable across the states. Before the findings are discussed, it is worth noting that the ICDS data may be expected to be more accurate in reflecting the differences in the geographical location of malnutrition across the states/projects than in capturing the age-specific levels of malnutrition in a given area. This is because, ICDS has so far been less successful in reaching the 18 under-threes than three-to-six-year-olds,12 so that the levels for the former may have been under-estimated. Bearing this in mind the following findings ara worth stressing. First, the proportion of malnourished children among the scheduleti ^astes and scheduled tribes was higher than the average in most places (see Tabl.e 8). Second, a comparison of the data for 0-36 months and 0-72 months shows that Niith few exceptions the percentages of malnourished children were higher for the grcup 0-36 months even among the scheduled castes and tribes. Third, the location - riral or urban or tri'bal - does not seem to significantly alter the above finding, as can 'be seen from Table 9 wherein the break up is presented for the five states fo:- which data are available. For example, the children of scheduled caste's and tr:ites in all locations have either equal to, or higher than average percentage of c1lildren malnourished. Fourth, in terms of percentage of malnourished children of I:ribals in Andhra Pradesh, Kerala, Uttar Pradesh, West Bengal and Madhya Pradesh an,d those of scheduled castes in Bihar, Orissa, Rajasthan, West Bengal and Uttar :'.radesh suffered fromi very high levels of severe (third and fourth degree) malnutri:ion. 2.15 More recent data for 1984 (see Table 10) only confirm the above findings. For this year, we have tabulated the percentage of malnolurished ctildreni for those ICDS project/district areas which became operational 10-12 months 'DEfore the date of data collection, so that these numbers could well serve as control group depicting the nutritional status of children in these districts/areas for the period 1983-84. A basic limitation of this data, however, is that further disaggre;,,ation for scheduled castes is not available, but the tribal situation can be guessel from the location of the projects. The disaggregated project-wise information is s~iown 12See Part IV of this paper for a discussion on this aspect. 19 in Table 10, along with the area in which the project was located (tribal, rural or urban). Table 10 REGIONAL PROFILE OF MALNUTRITION AMONG CHILDREN Percentage Severely of Malnourished (grades III & IV) children* in ICDS Project Areas during 1984-85 /a PERCENT CHILD MALNOURISHED MORTALITY DISTRICT RATE /c CHARACTERISTICS TYPE OF ------------- ---------- --------------------------- Sn DISTRICT PROJECT 0/3 3-6 0-2 0-5 SC + ST Id AVA le IRR/f No STATE NAME lb POP 1 2 3 4 5 6 7 -------------------------------------------------------------------------__--__----------------- X Rs. X 1 ANDHRA PRADESH Guntur Rural 8.80 11.30 99 142.00 13.00 317.00 39.00 Vizianagaram Tribal 22.20 32.50 169 196.00 19.00 61.00 30.00 2 KARNATAKA Dharwar Rural 15.40 12.90 124 182.00 15.00 307.00 6.00 Mysore Rural 17.20 10.80 109 149.00 24.00 270.00 17.00 3 TAMIL NADU Ranyakumari Urban 5.70 4.60 70 80.00 5.00 161.00 52.00 Salem Urban 5.70 0.30 74 101.00 20.00 245.00 22.00 4 KERALA Mallapuram Rural 1.20 4.30 66 104.00 9.00 355.00 57.00 Palghat Rural 0.70 4.00 74 116.00 19.00 421.00 57.00 5 MADHYA PRADESH Jabalpur Rural 34.40 16.90 220 265.00 29.00 152.00 5.00 Dhar Tribal 26.00 17.30 138 174.00 59.00 350.00 7.00 6 UTTAR PRADESH Pratapgarh Rural 14.60 13.00 163 211.00 22.00 222.00 28.00 Kheri Rural 6.10 3.10 154 187.00 27.00 587.00 10.00 7 RAJASTHAN Bharatpur Rural 12.40 n.a. 214 252.00 24.00 369.00 20.00 Banswara Tribal 14.20 0.70 170 169.00 25.00 235.00 7.00 8 MAHARASHTRA YaystMal Rural 22.30 25.10 177 227.00 26.00 264.00 2.00 Nanded Urban 12.00 10.20 87 107.00 21.00 289.00 3.00 9 GUJARAT Valsad Tribal 4.60 3.30 91 96.00 57.00 192.00 13.00 Baroda Tribal 7.30 5.60 115 129.00 31.00 206.00 21.00 10 ORISSA Cuttack Rural 3.20 2.80 196 204.00 21.00 271.00 32.00 11 WEST BENGAL Nadia Rural 21.30 12.20 115 149.00 28.00 268.00 33.00 Bankura Tribal 6.50 8.20 77 89.00 40.00 427.00 33.00 12 HARYANA Hissar Rural 6.70 4.20 120 146.00 22.00 901.00 64.00 Bhivani Rural 5.50 0.70 114 129.00 18.00 558.00 20.00 /a The data in columns 1 and 2 are taken from ICDS Project-wise Progress Report for 1984185, for those Projects which which were sanctioned during 1983/84. As such, these figures reflect the nutrition sitution atiabout the commencement of ICDS in these districts. lb R = Rural; T = Tribal; U = Urban. to District-wise Child Mortality Rates for 1981 for the age groups less than two and less than five years were made available for the first time in the occasional Paper No. 5, 1988, Office of Registrat General, Ministry fo Home Affairs. £d Percentage of scheduled caste and scheduled tribe population to total population of the State. e4 Agricultural value-added per capita. If Irrigation ratio. * IAP classification (defined in footnote to Table 8). Since the data are disaggregated and available at the district level, it is possible to conduct tests of their reliability with other sources of data at the district level.. For example, child mortality rates for the age groups 0-2 and 0-5 at the district level were recently released by the office of the Registrar-General of Census. In order to test the reliability of the numbers on the 20 percentage of children severely malnourished collected by the ICDS monitoring authroities, we estimated the correlation between these and the ch:ild mortalil:y rates for the corresponding 23 districts. A statistically significant pcsitive correlation of 0.57 between columns 1 and 3 suggests that the ICDS base line c.ata can be relied upon for assessing geographical v7ariations in malnutrition. 2.16 What can we infer from these data? The percentage of children s verely malnourished are indeed very high for triba:L Andhra Pradesh, Bihar, Rajastha[1 and Madhya Pradesh, and also in rural Karnataka, Madhya Pradesh, Uttar Pradesh, Maharashtra and West Bengal. Many factors may have contributed to the observed inter-project variation in the incidence of malnutrition. It is tempting to examine whether the variation across the project/districts is at all related to the level of development across the 23 districts in which the projects were located, as: one would expect. Although no data are available on the districts' income, a receIt study found that the irrigation ratio serves as an excellent proxy for the general development of a district (See C.H.H. Rao, S.K. Ray and K. Subbarao). The following regression equation with t:he percentage of malnourished children as the deper.dent variable (Y), and irrigation ratio (IR) as the :independent variable, confirmEd the expected negative relationship between the level of malnutrition and developmzr.t with 2. high level of statistical significance: * ~~~~2 Y = 18.29 - 0.25 IR R = C1.29 (- 2.96) * Significant at 1 percent level. The above finding reinforces the general impression that the project/district; withl a preponderancea of malnourished scheduled caste and tribal children are also :Ile districts that are less developed. 13 StatisticaLLy significant at 1X LeveL. 21 2.17 The ICRISAT data for the extremely backward arid zone (for 1976), though dated, also confirms the incidence of severe malnutrition among children in the age group 1-3 among all classes, (see table 4).t4 In general, the incidence of energy deficiency was higher among the children of landless labourers and small farmers. In addition to this finding, the ICRISAT data are also useful in segregating the effects of season on malnutrition. As one moves from lean to surplus season, the percentage of sample individuals consuming less than 50% of recommended energy intakes falls in almost all age groups, but the extent of decline was smallest for children, whose proportions, though lower in surplus than in the lean season, still continue to be very high in an absolute sense even in the surplus season. (c) Distribution of Malnutrition by Sex 2.18 It is well-known that in traditional, poor households in rural India food is first served to male members, and children and women eat last. This has given rise to the notion that male members get the lion's share of family's food allocation, and the women get the crumbs. Systematic evidence on gender discrimination in food and nutrient intake is hard to come by. There are serious data limitations. The common method of testing for discrimination would be to compare the food intakes of males and females after adjusting for their requirements. This procedure involves several conceptural and practical difficulties. An alternative method would be to observe certain anthropometric measurements by gender.'5 While there is some evidence of sex bias in intrafamily distribution of 14See also footnote 11 on page 16, for probLems with this data base. 1 Sen puts greater emphasis on this. He argues that it is more important to Look at the consequences of disparity in food intake rather than in comparing food intake itself. 22 food, the evidence is by no means conclusive. The NXMB have calculated the calor:e adequacy status in each age-sex group. It has also collected anthropometr:Lc measurements on these individuals. The NNMB data 1975-82 showed consistent:y that both sexes suffer to the same extent from calorie inadequacy (calorie irtake 2 standard dev:Lations or more below the average), although the figures were .n fact slightly less for females (see Table 11). The NNMB data (not presented here) also suggest that even in the pre-school age group, where the incidence of severe p;rotein - energy-malnutrition is definitely higher among the girls, calorie intake is nct comparatively lower among the girls. According to NNMB, the greater incid.nce of PEM in girls is obviously not due to lower food intakes, but probably due to g;enera.l neglect of the health of girl children. 16 However, in some cases, adequate caloric intakes may be accompanied by excessively low protein intakes. 2.19 Sen and Sengupt:a measured malnutrition of children in two villages i.n West Bengal by comparing their actual weights with expected weights in relation to age. They noted systematic sex bias against girl children which was reflected both in greater undernourishment of various degrees, and also in the lower growth dynamics of girls vis--a-vis boys. In times of famines or general food shortage, the very young and old of both sexes were allowed to go with inadequate food. Kakwani (>986- b) later developed an alternative statistical procedure which allows for mea.suring the statistical significance in the observed bias against girl children. Kalvani's results confirm Sen-Sengupta findings only in one of the two villaLges, but RI not support their other conclusions relating intervillage differences in the se. bias. Kakwani also examined sex bias by age groups on data collected during an econon.ically distressed year (floods in West Bengal during 1979), and showed the prevale:ice of 16Kakwani (1986-b) conducts a series of statistical tests on NNMB data and arrives at simiLar concLuAicns. 23 strong sex bias among younger children (up to four years of age). The reasons why families discriminate against younger girls during economic distress are not clear. Table 11 Calorie Inadequacy* Among Adult Males and Females (Percent of Population) 1975-1979 1979 1980 States Males Females Males Females Males Females Kerala 60.8 50.9 60.6 54.6 81.8 58.5 Madhya Pradesh 48.4 28.8 63.3 55.0 - - West Bengal 45.7 38.4 53.1 54.6 35.3 30.4 Orissa 42.6 24.0 39.6 22.1 39.3 20.3 Maharashtra 40.3 27.9 44.0 36.9 - - Uttar Pradesh 36.7 32.2 28.8 29.5 38.4 25.8 Andhra Pradesh 35.6 18.5 22.9 7.7 35.1 24.1 Gujarat 35.2 27.3 24.2 17.2 29.3 20.9 Tamil Nadu 34.8 25.4 15.7 16.7 41.4 36.1 Karnataka 18.8 10.4 19.8 7.9 11.0 10.3 * Intakes two standard deviations or more below the mean. Source: NNMB data, as reported in Kamala S. Jaya Rao, "Undernutrition Among Adult Indian Males", NFI Bulletin, July 1984. 2.20 ICRISAT's research on the arid zone in Central India suggests that while children as a whole appeared to get low shares of energy, regression analysis and t-testing for statistical significance led to the conclusion: "There is no significant difference between boys and girls in the intake of 10 nutrients" (Walker, et. al.). These findings along with Kakwani's results and NNMB findings suggest that the available nutritional evidence on sex bias does not appear to be conclusive. However, there appear to be distinct regional patterns. For example, Harriss ,noted that discrimination in energy intakes within the household seemed to be greater in the north than in the south India. Also, in the north, the intra-household distribution of food appeared least fair for the very young and very old females, and probably also for women in greatest need associated with pregnancy and lactation. 24 2.21 SummarY: Thus, the NNMB data alongside the ICDS and other data sources, examined at a disagregated level, are helpful in delineating the location of severe malnutrition across regions, age and social groups more thoroughly than state-LeveL averages. In terms of households' average nutritional status, severe malnutritioni is more prevalent in the! eastern states, and Uttar Pradesh, Madhya Pradeb;h and Kerala. In terms of social groups, the nutritional status of scheduled castes and tribes was substantially Lower than the recommended minimum in most states. As for age groups, the incidence of acute malnutrition is definitely high among children especially in the age group 0-3 in almost all states; this was even higher .in the tribal tracts. This last finding is important, given the Bank's experier'ce in reducing the incidence of severe malnutrition among young children. While the evidence on gender discrimination is not conclusive, regional figures suggest some discriminatiotn in north India, especially against very young girls. III Public Policy: Expenditure Levels and Sustainability 3.1 Broadly speaking state and central governments have been following two approaches to fill nutrit:Lon gaps. The first approach is direct feeding prc,grams for children and women. The second approach is aimed at providing food security at the household level by supplying a portion of their foodgrain needs at lower t:han market prices through a network of fair price shops. 3.2 The first approach dominated public policy during the 1950's and 1960's 25 when supplementary feeding programs of various kinds were introduced. In 1959 the Applied Nutrition Program was initiated for pre-school children and pregnant women. This was followed by school lunch programs for improving school attendance as well as child nutrition. These school lunch programs were most popular in Andhra Pradesh, Tamil Nadu and Kerala. A modicum of geographical targetting was introduced under the Special Nutrition Program in 1970-71 which aimed at providing food to children below six years, and to expectant and nursing mothers living in urban slums, tribal and backward areas. The Integrated Child Development Services (ICDS) introduced in 1975 reflects the culmination of various health and nutrition action programs focussed on vulnerable groups. (See section IV for a full description.) During the mid 1970s, in addition to nutrition interventions, health-related interventions such as prophylaxis against Vitamin A deficiency, nutritional anemia and endemic goitre were also introduced. 3.3 The second approach - indirect intervention via, a Public Distribution System (PDS) to improve the households' access to food - has occupied a central place in public policy since mid-1970's. Initially the role of PDS was largely confined to providing cheap food to major cities. The progress in extending PDS to the vulnerable households especially in the rural areas was slow. Kerala was the first state to cover the entire state with easily accessible public distribution outlets for staple foodgrains. The entire vulnerable population of Tamil Nadu and Gujarat (rural and urban) has been covered by the PDS network effectively since 1980. Andhra Pradesh and Karnataka have extended PDS in 1983 and 1986 respectively. In the remaining states, PDS has continued to be confined largely to the urban areas, so that industrial workers and marginals in urban conglommerates such as Bombay and Calcutta were effectively served by PDS outlets over the last two decades. But the 26 rural poor in these states by and large remained outside the purview of PDS.1; 3.4 The purpose of this section and the next is to assess t'he effectiveness of these public responses; to the malnutrition problem. This is done with reflerence to three criteria: (a) w'hether or not public expenditures were in conformit:y with regional and social class nutritional needs identified in the previous sectioris; (b) whether or not the various programs introduced were based on considerations of long-run sustainability, without which no improvement in nutritional status can be expected; and tc) whether or not the impact studies confirm realisation of the objectives of the programts. It may be noted that these norms are to some extent inter-related. For example, if expenditures do not conform to the needs, the overall impact is bound to be disappointing. 'ret we treated these three aspects separately because it is empirically convenient t:o do so. This part deals with the first and the second aspects; the impact studies; are given a detailed treatment in Part IV. The issues arising from all the four Farts of the paper are woven together in Part V. (a) Size and Distribution of Direct Nutrition Expenditures 3.5 Direct expenditure on nutrition support for children is incurred on three. major items: (a) mid-day meals, (b) special nutrition program, and (c) ICDS. Durlng l986,'87, the states spenit about Rs. 3110 million on the first two, while the Centre spent about Rs. 1150 million on ICDS. It is worth noting that the Central Government's cx?enditur.e during 1986/87, the first year of the Seventh Plan, is substantially higher th,.n the 17Unfortunatety the data base does not permit computation of precise numbers of population covercd by Location (ruraL/urban) and by state. 27 Sixth Plan average. Total expenditures incurred by the States and the Centre on direct nutrition support amounted to 1.86 and 0.5% of State and Central Plan outlays respectively, even after accounting for the increased allocations for ICDS. Together, the scale of expenditure during 1986/87 amounted to no more than 1.09% of plan outlays of the Centre and the States, or 0.16% the nation's GDP. While commitment of additional resources during the Seventh Plan is undoubtedly a significant welcome development, the resources spent on nutrition continue to be rather modest. The main concern of this section is: how have these modest expenditures been distributed across the states and how productive this spending has been in alleviating malnutrition. 3.6 It was shown in the foregoing sections that children of scheduled castes and tribes in Madhya Pradesh, Rajasthan, Andhra Pradesh, Orissa, Bihar, West Bengal and Uttar Pradesh suffer more heavily from malnutrition. Let us now examine how far the allocation of resources responded to the observed regional variations in needs. We compute the state-wise real expenditure per child during the Sixth Plan and the first year of the Seventh Plan (see table 12). Direct nutrition expenditures per child in all the above mentioned states are extremely small - almost nil as in Andhra Pradesh and Rajasthan. By contrast the expenditures on direct nutrition programs (supplementary feeding) have been high in the three southern states (Tamil Nadu, Kerala, Karnataka) and Gujarat. In these states, the need as measured by calorie intakes is not arguably the greatest. It is possible that high expenditures in the past may have led to high calorie intakes (lower needs), although it is difficult to test this hypothesis owing to lack of state-wise, time series data on levels of malnutrition with appropriate control groups. Nonetheless, there appears to be no recent significant change in expenditures in the states with high levels 28 of malnutrition. Even the expenditures during the first year of the Sevenlth FPLan did not produce any substantial increase in the states with low calorie intakes.16 Table 12 State-wise Distribution of Nutrition Expenditures Annual real expen- Annual Annual real expen- diture asvailable per Plan diture per child child uader pove.ty lir.e Sixth Plan (1985/86) ----------------- -----------_________._- annual average expend- Sixth 1985186 Sixth 1985186 Sn. at '80-81 prices iture Plan Plan No. States Its. million Rs. million Rs. Rs. Rs. 1 2 3 4 5 6 1 ANDHRA P:RADESH 1899.00 28.00 99.00 1.50 259.00 3.80 2 ASSAM 92.00 81.00 10.00 9.10 49.00 42.60 3 BIHAR 157.00 377.00 6.00 14.10 11.00 26.90 4 GUJARAT 205.00 3232.00 16.00 258.60 63.00 1010.00 5 HARYANA 72.00 92.00 14.00 18.40 85.00 102.20 6 HIMACHAL PRADESH 53.00 37.00 38.00 26.44 270.00 185.00 7 JAMMU & KASHMIR 22.00 84.00 10.00 38.20 59.00 210.00 8 KARNATAKi 380.00 1224.00 28.00 88.70 74.00 240.00 9 KERALA 252.00 337.00 27.00 35.90 95.00 124.80 10 MADHYA PRADESH 314.00 224.00 14.00 10.00 29.00 20.60 11 MAHARASHTRA 438.00 374.00 20.00 10.70 54.00 46.20 12 ORISSA 101.00 196.00 10.00 19.40 22.00 43.60 13 PUNJAB 29.00 148.00 5.00 26.00 36.00 185.00 14 RAJASTHAN 59.00 68.00 4.00 5.90 11.00 13.30 15 TAMIL NADU 1213.00 3947.00 74.00 242.20 180.00 589.10 16 UTTAR PRADESH 222.00 534.00 5.00 12.80 11.00 26.70 17 WEST BENGAL 352.00 231.00 16.00 10.30 38.00 24.80 3.7 The second question is, even if some states, such as Tamil 4..du anc Gujarat, are allocating a,dequiate resources to the nutrition sector, are these 1:esources actually beirg spent on 1:he social groups identified as acutely malnourisheO? As is well known, in both Tamil Nadu and Gujarat school feeding constituates the :rincipal nutrition intervention. Considering that the incidence of acute malnutritio:. Ls more among the under-threes, this scheme is serving children of age groups :hat are identified as less vulnerable. The productivrity of these expenditures wciLd have increased many-fold if they were directed strictly towards malnourished social groups. This is illustrated in table 12, which computes the annual real expenditures I;er: child under two assumptions: i(a) there is virtually no targetting of direct r.ul:rition 18This asiect is discussed further in 3.8. 29 expenditures (which is by and large the case in most states), (cols. 3 and 4) anrd (b) these expenditures are confined at least towards children under the poverty line, though not strictly towards malnourished children (cols. 5 and 6). Since all children are covered under the programs in Tamil Nadu and Gujarat, these two states spent about Rs. 250/- per annum per child during the first year of the Seventh Plan. If efforts were made to even modestly orient these programs towards children of families under the poverty line, (col. 5 and 6), a substantially higher nutrition support per poor child would be poss:Lble; alternatively more needy children could be covered. For example, if nutrition support were strictly confined to poor children, in eight states (Gujarat, Himachal Pradesh, Haryana, Jammu and Kashmir, Karnataka, Kerala, Punjab and Tamil Nadu), a nutrition support of more than Rs. 100 per child per annum could have been effected. But in reality feeding programs in some states resorted to coverage of all children, so that their impact on nutrition continued to be small, thus lowering the productivity of even modest expenditures. 3.8 The third question is: are there any visible signs of change in the Seventh Plan as compared to the Sixth Plan. If so, which states are attempting to change their efforts? And in which direction? The state-wise distribution of state's expenditures on nutrition for the Seventh Plan for 17 major states is presented in table 13, alongside planned outlays and actual expenditures for 1985/86, the first year of the Seventh Plan. Actual direct expenditure on nutrition as a percent of plan outlays ranged from 11% in Andhra Pradesh to 183% in Himachal Pradesh. One of the reasons for the observed wide disparity between intentions (planned outlays) and actual expenditures is that some state governments such as Orissa and Himachal Pradesh found their initial allocations insufficient for the needs. But for most state governments, actual expenditures; fell short of outlays largely because of paucity of resources. 30 In 11 out o:f 17 states, actual expenditure during 1985/86 was less than 15% of the Seventh Plan outlay (se,e. col. 6). This would mean that in real terms,19 there is a strong possibility of shortfalls in direct nutrition expenditures for the PLan as a whole. Considering that the aggregate outlays for nutrition continue to 1te modest, such shortfa.lls in actual expenditures would imply decline in the "revealed" Priority to nutrition in plan expenditures. This is especially so for Andhra Pradesh, Assani., Haryana, Rajasthan and Wfest Bengal. Table 13 Outlay and 'Expenditure - Seventh Plan - 1985-90 - Nutrition (Rs. million) Seventh 1985-86 Col. 4 as a Plan ----------------------- percent of Sn 1985-90 Outlay Expendi- ----------------- No. States Target ture Col. 3 Col. . 1 2 3 4 5 6 1 ANDHRA PRADESH 5360.00 372.00 41.00 11.02 0.76 2 ASSAM 2000.00 120.00 120.00 100.00 6.00 3 BIHAR 3500.00 558.00 556.00 99.64 15.89 4 GUJARAT 59550.00 11272.00 4760.00 42.23 7.99 5 HARYANA 2794.00 429.00 136.00 31.70 4.87 6 HIMACHAL PRADESH 282.00 30.00 55.00 183.33 19.50 7 JAMMU & KASHMIR 755.00 150.00 124.00 82.67 16.42 8 KARNATAKA 11000.00 2445.00 1803.00 73.74 16.39 9 KERALA 4000.00 538.00 497.00 92.38 12.43 10 MADHYA PRADESH 3389.00 608.00 330.00 54.28 9.74 11 MAHARASHTRA 5000.00 944.00 551.00 58.37 11.02 12 ORISSA 1600.00 200.00 288.00 144.00 18.00 13 PUNJAB 1650.00 252.00 218.00 86.51 13.21 14 RAJASTHAN 1596.00 108.00 100.00 92.59 6.27 15 TAMIL NADU 54000.00 10337.00 5814.00 56.24 10.77 16 UTTAR PRADESH 4470.00 843.00 787.00 93.36 17.61 17 WEST BENGAL 5000.00 870.00 340.00 39.08 6.80 165946.00 3007,6.00 16520.00 1341.14 193.67 Source: Average Computed. 19At 1985/86 prices. 31 3.9 Across the states, there are also interesting sharp "turn-abouts" from the Sixth Plan to the firsit year of the Seventh Plan. A classic case of such a sharp change is that of Andhra Pradesh, where direct nutrition expenditure per annum fell from Rs. 1899 million to Rs. 28 million (see cols. 1 and 2, Table 12). Gujarat, Karnataka, and Tamil Nadu represent the opposite extreme: a sharp rise in direct nutrition expenditures per annum owing essentially to introduction and/or expansion of school feeding programs. While resource-deficient states such as Bihar were unable to introduce direct nutrition support programs despite their acute need, states that are relatively better-off, such as Gujarat, did not hesitate to introduce massive feeding programs. 3.10 The mismatch between regional needs (incidence of malnutrition) and expenditures noted above is reflected in physical performance. For example, regional asymmetry in the coverage of feeding programs is clearly evident from Table 14. Different state governments are running different types of feeding programs, often many types in the same state. An effort has been made to arrive at an aggregate figure of total number of children fed (avoiding double-counting), for each of the 17 large states. This is shown in column 4. It is instructive to compare this column with the next four columns showing the number of all children (column 5); poor children in the age group 0-9 (column 6); moderately and severe:Ly malnourished children (Col. 7) and severely malnourished children (Col. 8).20 For each state column 9 gives total children fed as a percent of all children in the age group (i.e. assuming zero targetting); column 10 gives total children fed as a percent of children of families below the poverty line (i.e. assuming some targetting towards the poor though not strictly towards the malnourished); column 11 presents the percentage if the feeding were confined to all the "moderately and severely" malnourished 2OWe used the NNME estimates for the malnourished children for the ten states for which data are available for 1982. Table 14 Statevise Distribution of Children fed under various Feeding Programs In '000s ___________________________________________________________________________________________________________________________________________________ NUMBER OF COL 4 AS PERCENT OF SN STATES SNP /a MDM /b OTHERS /c TOTAL ALL CHILDREN POOR CHILD- MALNOURISHED Col. 5 Col. 6 Col. 7 Col. 8 NO 0-9 REN 0-9 CHILDREN /d A B 1 2 3 4 5 6 7 8 9 10 11 12 …__ _ - - - - -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1 ANDHRA PRADESH 908 11 919 12600 4590 5594 743 7.29 20,02 16.43 >100O 2 ASSAM 512 95 12 619 7000 1650 8.84 37.52 3 BIHAR 690 7 697 18600 9210 3.75 7.57 4 GUJARAT 1151 3000 43 4194 8800 2140 5245 1338 47.66 195.98 79.96 >100% 5 HARYANA 284 6 290 3500 550 8.29 52.73 6 HIMACHAL PRADESH 75 2 77 1000 140 7.70 55.00 7 JAMMU & KASHMIR 102 1 103 1600 260 6.44 39.62 8 KARNATAKA 884 11 895 9300 3260 3980 520 9.62 27.45 22.49 >100% 9 KERALA 552 950 8 1510 6000 1610 1134 90 25.17 93.79 133.16 >100O 10 MADHYA PRADESH 1048 1200 11 2259 15400 7120 7962 1771 14.67 31.73 28.37 >100% 11 MAHARASHTRA 954 2000 40 2994 14600 5080 6950 1007 20.51 58,94 43.08 >100% 12 ORISSA 1314 184 11 1509 7100 3040 3635 632 21.25 49.64 41.51 >100% 13 PUBJAB 159 6 165 3800 520 4.34 31.73 14 RAJASTHAN 563 5 568 10200 3500 5.57 16.23 15 TAMIL NADU 300 6400 6 6706 10000 3960 3980 520 67.06 169.34 168.49 >100% 16 UTTAR PRADESH 682 27 709 29000 13140 9541 1769 2.44 5.40 7.43 >10OX 17 WEST BENGAL 1250 3200 11 4461 15500 6080 3565 - 28.78 73.37 125.13 ----------------------------------------------------------------__-----------__-------------------------------------------------- 11428 17029 218 28675 174000 65850 16.48 43.55 La Special Nutrition Programs, run both under ICDS and outside ICDS. /b Mid-day meal program for school children. Ic Various other feeding programs under the ampices of Indian Council Welfare, Bharitiya Adimajoti Sewak Somgli, etc. Ld In arriving at the numbers in these columns, we used the NNMB sample survey data for the year 1982 A - Moderately and severely malbourished children. B - Severely malnourished ehildren 33 children; and column 12 if the feeding were confined only for the severely malnourished children. 3.11 Even assuming targetting to poor children (col. 10), which is by and large not the case in reality, the coverage was extremely poor in states with the greatest need such as Andhra Pradesh, Bihar, Rajasthan, Uttar Pradesh.21 However, if the feeding were to be strictly confined to all the moderately and severely malnourished children, then the prevailing scale of the program is reasonably adequate in Gujarat, and is more than adequate in Kerala, Tamil Nadu and West Bengal (col. 11). In other words, in these four states the nutrition expenditures currently being incurred are adequate for combating malnutrition if only these were directed only at the malnourished children. If the feeding were to be directed only to "severely malnourished" children, the existing resources would appear to provide complete coverage, except for Uttar Pradesh (col. 12) and probably Bihar. 3.12 These considerations suggest the potential benefits from reform in supplementary feeding programs currently being administered. As at present, regional asymmetry in the coverage is striking: four states in which only 26% of the poor children of the country are located (Gujarat, Maharashtra, Tamil Nadu and West Bengal) accounted for as much as 64% of the country's children fed under the various feeding programs (including ICDS) .22 21The data on state-wise number of poor and malnourished children reflecting "needs" belong to early 1980s, so that they represent the position after netting out the influence of other interventions such as EGS in Maharashtra. 22More or less similar regional asymmetry prevails with respect to the Public Distribution System, which is discussed later. 34 (b) Expenditures and Sustainabilitv of interventions: A case study of Andhra Pradesh 3.13 For any significant positive impact on the nutritional status of populaticon to be felt, it is important to (a) design the intervention carefully so as to maximis:e nutritional impact on the nmalnourished, (b) minimise the budgetary costs so that the cthosen intervention can remain in force for a sufficiently long time. If programs are nont designed carefutlly and expenditures are allowed to escalate, sustainability cannot be ensured. Inadequate care to targetting and costing of schemes has led 1;0 hasty introduction of schemes and quick reversals. A classic case in recent times :Ls that of Andhra Pradesh, which experimented with both mid-day meals and public distributLcon system during the period 1980-85. This section deals with this case study. 3.14 An,dhra Pradesh was the only state to introduce and quickly withdraw a noon meals program during the Sixth Plan. The dramatic year to year changes in budgetary expenditures on nutrition are shown below.23 Nutrition Expenditure in Andhra Pradesh Years Total direct Nutri- of which Rs. Million tion Support /a mid-day meals col. 2 as (Rs. Million) (Rs. Million) % of Col. 1 1 2 3 1980-81 31.2 3.4 10.9 1981-82 27.9 3.1 11.1 1982-83 299.5 292.4 97.6 1983-84 378.8 366.0 96.6 1984-85 244.4 210.2 86.0 1985-86 4.1 Nil 1986-87 22.9 Nil La This does not include expenditure on PDS. 231 am indebted to R. Radhakrishna for interaction in writing of this section. 35 During 1980-81 and 1981-82, the government was feeding Ready to Eat (RTE) food at the rate of 60 grams per child as a mid-day meal to 535,000 school children of grates I to V (primary classes) in five backward districts. During the same period, CARE food was supplied at 13,495 centres benefitting 1.36 million children mainly belonging to the weaker sections, incurring an expenditure of Rs. 11.5 million. During 1982-83, the program of noon-meal (outside CARE purview) was extended to all school children in the grades I to V in all the 23 districts throughout the state, so that the expenditure rose from Rs 3.1 million in 1981-82 to Rs. 292.4 million in 1983. This was done by the new government (Telugu Desam) which came into power at this time. During the following year the scheme was further extended to children in classes VI and VII; the CARE food was suitably supplemented to bring it to the level of food supplied under the noon-meal program; the education department became the nodal agency for monitoring and a senior officer of Indian Administrative Service was placed in charge of the scheme exclusively. Expenditure further rose to Rs. 366.0 million during 1983-84 - roughly 5% of the annual plan outlay of the state and 97% of the total direct nutrition expenditure. After a detailed review of the scheme, efforts were made towards targetting, and the beneficiaries were confined to children of scheduled castes, tribes and backward classes, and children of parents whose annual income did not exceed Rs. 3000. Government supplied noon-meals was now entirely withdrawn in the five CARE districts. These measures brought down the annual expenditures to Rs. 210.2 million during 1984/85. Noon-meal program was totally given up during the Seventh Plan. Indeed, expenditure on all direct nutrition support (including ICDS) fell sharply to Rs. 22.9 million during 1986/87. 3.14 The new government also wanted to honour the electoral promise of providing rice to the vulnerable population at Rs. 2 a kilo. This is a scheme aimed at improving the poor households' overall access to foodgrains. Subsidy cost of this scheme, which was 36 untargetted, rose sharply pari Rassu with the subsidy on noon-meal scheme. During :L983-84, both programs together exceeded 12% of the annual plan outlay. Clearly the gouerl:rent could not sustain two welfare programs when each accounted for substantial budgetary explenditure, The government withdrew the noon-meal scheme, but continued with the subsid:sed food distribution scheme.24 The rationale for preferring one scheme over: the othir is not: entirely clear, although one reason given was that noon meals scheme couli not, be administered eff.iciently owlng to lack of advance preparation. 3.15 What lessons can one draw from the Andhra Pradesh experience? First, noon-meal scheme was introduced without a prior careful analysis of its t).dge1:arSj implications and administrative feasibility for ensuring its long-run sustainabil::ty. I such an analysis had been done, then the budgetary implications might have provicietl a pushil to more careful. targetting. Second, Andhra Pradesh experience also suggesi:s tha-1: targetting (e.g. restricting access to vulnerable groups) is difficult: where children o:i the poor and the near-poor are numerous in certain locations. Third, the suddeu rise iln expenditure on this scheme during the Sixth Plan must have brought about adjustmen.s (cuts: in Plan allocations on other schemes, and must have depleted the resources nl.eied for long-term development, the consequences of which are difficult to assess. Thou~:,h beyonca the scope of thisi paper, it is important to consider the macro-economic: impacts i:cf sudder. swings in plan allocations following the introduction of an expensive we'Lfare sche:nE. This is because policy simulation models using all-India data do suggest that the nust impact of such untargetted and expensive welfare schemes on the poor could as well be re!ressive, both because of the inflationary potential of such. schemes, and also because of adverse 24This schene is discussed further in Part IV. 37 effects on employment/incomes consequent upon cut backs in development expenditure. (de Janvry and Subbarao 1987). 3.16 The foregoing suggests that noon-meal schemes may not be the right answer to the problem of malnutrition in India, whatever be their other advantages such as improvement in school attendance.25 Such schemes are usually addressed to older children wh.ile malnutrition is more serious among the age group 0-3; even for the older children they often substitute rather than supplement a meal. The relevance of noon meals scheme appears particularly dubious for Andhra Pradesh. We have already seen that except for tribal blocks, the average nutrition level in Andhra Pradesh is better than in other states, so that a scheme such as a well-targetted PDS would have pushed the average up, while effective outreach of ICDS would have improved nutrition in tribal areas. To this extent, the choice of PDS (Rs. 2 a kilo scheme) could be justified on a Eriori grounds, even if its wide coverage is unjustified. The Andhra Pradesh experience also suggests how important it is to pay attention to nuts and bolts issues such as alternative means of restricting access to the malnourished groups, appropriate costing of alternatives, etc. Only a careful analysis can determine the right mix of welfare interventions that are sustainable in the long run. 25Past experience suggests that while attendance improved with school feeding, the programs by and large failed to attract youncg girls belonging to bottom deciles (especially of landless labor) who have to fetch water, fodder for the cattle, and look after their younger brothers and sisters, while their parents are out in the fields. 38 IV Public Policy: The Effectiveness of ICDS and PDS 4.1 In the following sections, the public response to the malnutrition problem is futher assessed with reference to the effectiveness of one direct interveiit ion - the ICDS - and one major indirect intervention, viz., the Public Distribution System. (a) Integrated Child DeveloRment Services (ICDS) 4.2 Program DescriRtLon: ICDS is a country-wide program for child development. It adopts a holistic approach to improve both pre-natal and post-natal environmenit of the child. As such, mothers are also included in the program's outreach. The program aims to achieve four objectives: (i) to improve, the nutritional status of children 0-6 years via supplementary feeding to "sele:ted" beneficiaries; (ii) Ito encourage school enrollment via early pre-s':bool stimulation program for children 3-6 years old; (iii) to enhance the mothler's awareness via health arLd nutrition education, and (iv) to coordinate with hi-alth departments to ensure delivery of the required health inputs inclnding immunisation to children and mothers, so that both morbidity and mortality 7ates decline over time. A package of six services are thus delivered through Iu,ES: - health check-up - immuaisation - referral services - suppLementary nutrition - non-formal education - nutrition and health education to mothers 4.3 This program is Centrally-sponsored but administered by the ,tate 39 governments. The responsibility for financing ICDS is shared between the Government of India and the States. The Government of India provides support for the establishment and maintenance of infrastructure including (a) salaries of all ICDS staff (b) basic equipment for Anganwadi centres,26 (c) indoor play materials for children, (d) petrol and oil expenses (e) Anganwadi centre rent, where applicable; and (f) medicine and medical kits. The state government meets the expenses for supplementary food for all Anganwadi Centres. Some state govenments also have started some ICDS Anganwadis on their own, for which they meet the expenses for all the inputs including establishment and infrastructure maintenance. There are 167 state-sponsored anganwadi projects in operation as of 1985/86. 4.4 The Government of India intends to extend the program to cover 50 % of all poor children below the age six years (approximately 23.7 million by the end of the Seventh Plan (1985-90). The estimated cost of expansion is approximately US$ 1000 million (at September 1985 exchange rate), of which US$ 400 million will be provided by the Government of India, and the rest by the State governments. These figures do not include the cost of health inputs - staff, drugs, referral services - which are provided to ICDS through the Primary Health Care System. 4.5 The expansion of ICDS is basically a sound policy response because (a) it is aimed at children and women among whom malnutrition is heavily concentrated; and (b) the projects have been set up generally in areas of acute malnutrition (such as tribal areas). During the Seventh Plan period, all tribal areas, and 26"Anganwadi" is a community centre where children and mothers assemble to receive the services of ICDS outlined above. 40 rural areas with a concentration of scheduled caste population, and urban slums are expected to receive priority in the selection of projects. ICDS thus adcpts both "program" targetting inasmuch as the beneficiaries for supplementary feiading are expected to be chosen on the criterion of need, and "area" targetting to the extent the projects are located in regions of acute need. 4.6 Once an ICDS project is made operational, the Anganwadi_ Centre (AWC) becomes the focal point for the delivery of services. The responsibilitils of the Anganwadi Worker (AWW) include non-formal pre-school education for the older children, organisatiorL of supplementary feeding after identifying eliL;ible children and mothers, imparting health and nutrition education to wl:>men; educating parents by home visits; eliciting community support and participal:ion; primary health care for children and mothers and referring the neiedy to mer.ical personnel (under the health department); liasing with local organizations like Mahila Majdals;27 and finally maintaining records and furnishing pro,,;ress reports. AWW receives supervision and guidance from a Supervisor who norliLally looks after 20 AWWs. A full time Child Development Project Officer is in clLarge of each ICDS proj3ect. The health infrastructure in ICDS areas is providld by one Auxiliary Nurse Mid-wife (ANM), one Lady Health visitor (to supervisc the work of four ANMs) and one doctor per block. All these health services arv the responsibility of the Health Department. ICDS is thus a multi-departmintal endeavour, and its success critically hinges on coordination between the two departments; at the AWW level. Its success also depends to a great extent on the content and quality of training imparted to various functionaries iin the prov!ram. 27"Mahila Mandals" is a local women's organization. 41 4.7 Finally, a mechanism for monitoring is built into the ICDS program. A regular flow of information on the services delivered is ensured via progress reports which are consolidated at a Central Cell at the All India Institute of ror Medical Sciences, New Delhi. The monitoring system has/understandable reasons, addressed itself only to those services which can be quantified, and not to the other qualitative aspects such as nutrition education of mothers etc. which are also important components of the program. * 4.8 Program Coverage: Once a Block is selected for ICDS, theoretically all villages in that Block become eligible for services at a ratio of one AWC per 1000 people in rural and urban areas, and 700 in tribal areas. Although conceptually the program is targetted towards the severely malnourished, inasmuch as the selection of beneficiaries for supplementary feeding is (theoretically) to be done after nutritional screening, in practice this is rarely done.28 This objective of reaching the malnourished is sought to be achieved by judicious location of projects. In other words, there is less of "program targetting" in practice, but more of "area targetting". Let us consider whether the latter is being done efficiently. In order to do so, we have compared the base-line data on the percent of severely malnourished children reported in some specific ICDS proijects, with the state-level estimates of malnourished children reported in NNMB surveys (see table 15). 28See Gopalan (1988), p. 66. * A Block is an administrative unit smalLer than a district. There are 5,143 BLocks in India contained within 432 districts. 42 Table 15 Percentage of Severely Malnourished children in ICriS Project Areas (Base Line) in 1984 and corresponding State averages (1982) PERCENT MAlNOURISHED /b STATE AVERAGES c ICDS PROJECT DETAILS ----------------------- --------------- SN ------------------------------------ 0/3 3-6 (1-5 YEARS) NO. STATE DISTRICT R/T/U la 1 2 3 1 ANDHRA PRADESH Guntur Rural 8.80 11.30 5.9 Vizianagaram Tribal 22.20 32.50 2 KARNATAFA Dharwar Rural 15.40 12.90 5.6 Mysore Rural 17.20 10.80 3 TAMIL NADU Kanyakamari Urban 5.70 4.60 5.2 Salem Urban 5.70 0.30 4 KERALA Mallapiram Rural 1.20 4.30 1.5 Palghat Rural 0.70 4.00 5 MADHYA PRADESH Jabalplir Rural 34.40 16.90 11.5 Dhar Tribal 26.00 17.30 6 UTTAR PRADESH Pratapgarh Rural 14.60 13.00 6.1 Kheri Rural 6.10 3.10 7 RAJASTHAN Bharatpur Rural 12.40 0.00 N. A. Banswara Tribal 14.20 0.70 8 MAHARASHTRA Yaystmal Rural 22.30 25.10 6.9 Nanded Urban 12.00 10.20 9 GUJARAT Valsad Tribal 4.60 3.30 15.2 Baroda Tribal 7.30 5.60 10 ORISSA Cuttac]c Rural 3.20 2.80 8.9 11 WEST BENGAL Nadia Rural 21.30 12.20 6.6 Bankura Tribal 6.50 8.20 12 HARYANA Hissar Rural 6.70 4.20 N. A. Bhiwan:L Rural 5.50 0.70 /a R = Rural, T = Tribal, U = Urban. /b Grades III & IV malnutr:Ltion, IAP classification, i.e.,< 60% weight for age. le Gomez classification of severe malnutrition, i.e. < 60% of weight for age. Source: Col. 1 & 2, same as for Table 11, Col. 3 : NNM3 The levels of: malnutrition reported in ICDS projects are much higher than th:! stat:e averages in almost all projects. Clearly "area" targetting as reflected in the selection of project areas appears to be reasonably efficient except in Guja:S>at and Orissa. 4.9 It is; not proposed to discuss the relative merits and demerits of "1:rogram targetting" and "area targetting". However, one particular aspect of ICDS deserves discussion. There is considerable time lag between sanction of a 43 Table 16 Statewise Position of ICDS 1986-87 No. of Projects sanctioned as on 31.3.87 --------------------- Operational Col. 4 Total State Central projects as as a S.no. States Sector Sector on 31.3.87 of 1 1 2 3 4 5 1 ANDHRA PRADESH 99 9 90 38 38.38 2 ASSAM 51 Nil 51 37 72.55 3 BIHAR 142 Nil 142 107 75.35 4 GUJARAT 74 12 62 69 93.24 5 HARYANA 75 40 35 40 53.33 6 HIMACHAL PRADESH 21 Nil 21 14 66.67 7 JANMU AND KASHMIR 41 18 23 22 53.66 8 KARNATAKA 94 30 64 86 91.49 9 KERALA 73 23 50 69 94.52 10 MADHYA PRADESH 130 Nil 130 90 69.23 11 MAHARASHTRA 105 Nil 105 101 96.19 12 ORISSA 84 Nil 84 52 61.90 13 PUNJAB 43 Nil 43 34 79.07 14 RAJASTHAN 76 13 63 54 71.05 15 TAMIL NADU 65 Nil 65 39 60.00 16 UTTAR PRADESH 202 8 194 165 81.68 17 WEST BENGAL 130 15 115 91 70.00 Source: Central Teclnical Committee, All India Institute of Medical Sciences, Annual Report on ICDS, 1986-87. project and its becoming operational, sometimes up to seven years. The distribution of ICDS projects is given in Table 16. As already noted, some state governments (especially Gujarat, Haryana, Kerala, Karnataka and West Bengal) have initiated ICDS projects on their own without Central assistance, (see col. 2) in addition to the Centrally-sponsored projects (col. 3). There is considerable inter-state variation in the operational projects as a percent of sanctioned projects (see col. 5). Though paucity of finances could be advanced as an explanation for the states not 44 operationalising sancticned projects, it does not appear to be the main facto-r. This is because, as already noted, international agencies (WFP and CARE) hare been supplying food thereby partly mitigating the resource constraint problem of the states. For example, nearly 100% of the beneficiaries in ICDS projects in CUJarat, Madhya Pradesh, Orissa, Kerala and Rajasthan received either CARE or WFP food; in the remaining states the proportion varied between 55 to 75 percent of benefic iaries. Therefore, the reasons for the states' reluctance to make sanctioned projects operational are not entirely clear. One reason could be that some states may l3.ready be operating other direct or indirect nutrition support programs. For examrl.e, n the southern states, whiLe Karnataka and Kerala made almost all sanctioned p-cjec1ts operational, Tamil Nadu and Andhra Pradesh did so to a much lesser extent, pre3t.mably because in thiese two states mid-day-meals and Rs. 2 a kilo scheme already erere in operation, accounting for a substantial share of the budget,, We have already noted that there is a mismatch between the regional distribution of children fed under all programs and the distribution of severely malnourished chi:Ldren. This is Further accenltuated by the long delays in some states in operationalising sanctionied ICEIS projects . 4.10 ProLam Impact Studies: The number of studies of ICDS - largely ado: :tLng a "before and after" approach - is indeed impressive. In fact, ICDS has been claimed as an "over-researched" program. Some 800 surveys and several post-gradust4 and doctoral dissertations are reported to have studied various aspects of ICDS, Most of the surveys were conducted by All-India Institute of Medical Sciences (AIIM',) in collaboration with other medical teaching institutions in the country. Abs tracts of nearly 203 of such studies have recently been put together by the N t_..ona1 Institute of Public Cooperation and Child Development. (NIPCCD, uindated). These 45 studies contribute to our understanding of the working of the program and also help in identifying the major deficiencies and gaps. Yet a common problem of all these micro-studies is, as pointed out by Sharma, (1987), that they are "limited to small geographical areas and cultural regions, making generalizations difficult. Besides, many methodological limitations were also observed in these studies. The samples, by and large, were small and not representative. The conceptual understanding of researchers varied a great deal with respect to different components of ICDS. The same component, therefore, was tackled differently depending on their background and experience". These studies provided piece-meal information on individual components of the program and have not attempted a systematic analysis of the stock of inputs vis-a-vis the outputs. In brief, "the integrated package has been studied in a disintegrated manner." (Sharma, op. cit.) In some studies the provision of inputs and coverage is confused with output or qualitative changes in the beneficiary group. It is hardly surprising that the findings of different studies are often conflicting. 4.11 In addition to the individual studies by medical researchers mentioned above, * AIIMS, which is the monitoring cell for ICDS, also brings out Annual Surveys giving project-wise data on ICDS.29 These are available for the years 1978, 1979, 1982-84, and 1984-85. An important limitation of this data base is that it is unclear whether the population base in the pre-intervention period is comparable with that of the post-intervention year. This is because the nutritional status is assessed only for the children/women who reported at the AWCs, and not for those who did not. Consequently, if the base for the pre-intervention year came from a different population than that of the post intervention year, the derived percentages would These are the basis for the findings reported in the two oft-quoted papers by Tandon (1983 and 1i984). * ALL India Institute of Medical Sciences. 46 not be comparable. (Scr:Lmshaw). So far AIIMS has not conducted longitudinal studies on the nutritional status of beneficiaries and non-beneficiaries inl the same project area. Definitive conclulsions regarding the impact of ICDS can be drawn onily when control groups of reasonable sample size in different regions are used for comparison. 4.11 The specific limitations of the data base used by ICDS researchers have been stressed at length, not with a view to questioning the utility of the program but to underscore the point that the physical performance data on ICDS are nct yet reliable enough to attempt a full cost effectiveness analysis. Further, prscisely because of the weaknesses of the data base, an overall assessment of ICDS based on physical periFormance alone is not adequate; it must be supplemented by a qualitative assessment as well. This has not been attempted before; so it is hoped that th:Ls modest attempt in this direction will be helpful at least in raising the right set of questions on ICDS, though not in providing definitive answers. 4.12 Maior Findings of Evaluations The most striking aspect of the r*.searc.h findings is the wide var:Lation in the impact of ICDS by component, region, ;.ge an.d caste status of households. For example, many of the studies show a sign:lficant improvement in the nutritional and immunisation status and a progressive dec:ine in *" IMR.- But other studies for other areas indicate that there has been no sign:i.fLcant or only marginal improvement. Also, even within a district, the extent of impact has varied significantly. No systematic attempts have been made to explain variations in impact across regions, age groups and social groups, in compariscn with 0The most recent study is that of NFI (1988). * Infant mortality rate. 47 control groups of adequate sample size. The largest ICDS evaluation study, based on the data collected by the AIIMS monitoring cell, was conducted by Tandon (undated Mimeo) and this showed that the prevalance of severe malnutrition (grades III & IV) declined, (see table 17). Table 17 Impact of ICDS on Nutritional Status Children in grades III & IV malnutriton/a (%) 1 Non-ICDS 19 (Baseline 1976) 2 Non-ICDS 11 (Repeat Survey 1983) 3 ICDS (1983) 8 4 IMR ICDS Projects 88/1000 (1982-83) 5 National average (1981) 110/1000 La IAP classification. Source: Tandon (mimeo). Compared to the baseline figures, ICDS projects do show lower levels of malnutrition. However, the number of observations, means and standard deviations are not available so that the statistical significance of the differences cannot be obtained.3' Moreover, this gives an aggregative picture for all projects in the country. Program performance has not been analysed in terms of changes in different components over time in the same project area, and across regions in a given year, with appropriate control groups. 4.13 Analysis based on changes in the program performance in the same project area over time, using the same project-wise data as was used by Tandon, leaves one with much less confidence than the above mentioned findings. From the project reports released by the Monitoring Cell, data for 39 ICDS projects are available on a 31However, statistical validity of the observed temporal changes have been questioned by others. See, e.g., Sharma, Urmil. I,Mimeo). 48 continuous basis for the years 1978, 1979, 1982-84, and 1984-85. All these projects began operation roughly at the same time. The data are tabulated below under three categories (a) those showing definite improvement in nutritonal status; (b) those showing no definite tren,l, i.e. at first the percentage of malnourished falling, but then again rising; and (c) those showing definite deterioration. These a:-e also cross-classified according to location, viz, "Rural", "Urban" and "Tribal". I's can be seen from Table 18, clefinite improvement was reported in less than half cf the Projects whereas in nearly a quarter of the projects there was an actual deterioration in nutritional status. It is possible to argue that the situation in some of these projects could have gotten even worse without ICDS Projects. The purpose of this exercise is not to accurately measure ICDS impact - t,'is is impossible without project-specific control groups - but to underscore the erormous inter-project variation in impact over time which is not apparent in aggregative exercises such as those of Tandon 2p. cit. Table 18 Distribution of ICDS Project Areas by changes in the status of malnourished (Gr. III + IV)/a children: 1979-80/80-81 over 1984-85. RURAL URBAN TRIBAL TOTAL Definite 5 6 5 16 Improvement Ambiguous 7 4 2 13 Definite 7 - 3 10 Deterioration 19 10 10 39 /a IAP classi.fication. Source: Compiled from Integrated Child Development Services, Project wise Data:., for the years 1979/80 to 1984/85.. Information over time on the same project area was available only for 39 project areas out of 87 projects covered with Reports :.rom 1978/79 to 1984/85. 4.14 The location of the ten projects showing deterioration is presented in Table 49 19 alongside other characteristics. The three tribal districts are extremely backward, with an .Lrrigation ratio less than 10%, very low levels of overall development as reflected in CMIE indices, and a high proportion of SC & ST population. Of the 7 rural projects, 4 are located in relatively backward districts. It is interesting that 3 rural projects show deterioration despite being located in relatively developed districts. Fortunately, we have independent information on one of the three projects, viz. Jawan, Aligarh. The dismal performance of this project was confirmed by S. N. Sinha et. al who also noted that high morbidity in this highly irrigated district was due to the incidence of malaria. We do not have clues on the poor performance of the remaining two projects located in developed regions. 4.15 The above finding of poor nutritional impact in a significant number of projects based on indirect evidence is confirmed by a recent study done at NIPPCD32 based on direct evidence collected from 26 projects spread over ten states. For example, this study pointed out significant inter-project variation in the number of feeding days ranging from 30 to 300 across projects. Clearly no improvement in nutritional status could be expected if the feeding component were neglected. The precise numbers of projects showing no improvement and the reasons for poor performance will become clearer only when the monitoring cell collects longitudinal information on the nutritional status of children (both beneficiaries and eligible non-beneficiaries) in the same project areas, along with the socio-economic backgrounds of the households. It would then be possible to assess the performance of different components in the same project (abstracting from regional effects), and that of each component across the regions. Till such data base becomes available, 32NIPCCO, (1988) Status of Nutrition Component of ICDS, Mimeo. * National Institute for Public Cooperation in Child Development. 50 any overall assessment including our own must remain tentative. Table 19 Characteristics of some ICDS :Project Areas showing deterioration CMIE* Gross Lite- Proportion of sche- Index irrigated rary duled caste ar.d all India area as % rate tribe populati.on to = 100 of Gross total populatio:a Block, District, ------------- cropped % % State, Location State District area Utnoor, Adilabad Andhra Pradesh, Tribal. 107 66 8 13.6 35 Rama, Jhabna Madhya Pradesh, Tribal. 62 40 3 7.2 86 Basia, Ranchi Bihar, Tribal 51 71 4 24 61 Billawar, Kathera Jammu & Kashmir, Rural 100 55 21 29 23 Kanakapur, B.angalore, Karnataka, Rural - - 16 31 18 Vengara, Mal,appuram Kerala, Rural 126 72 58 60 9 Jawan, Aligarh Uttar Pradesh, Rural 72 76 76 28 23 Shankargarh, Allahabad Uttar Pradesh, Rural 72 65 21 21 25 Kalayat, Jind Haryana, Ruiral 166 145 70 22 19 Raipur Rani, Ambala Haryana, Rutral 166 183 29 36 23 * Index of Infrastructural Development computed by the Centre for Monitoring of the Indian Economy. 4.16 Let us now turn to specific aspects. The most important deficiency ncted by many observers is the low coverage of children under three years of age, ar.d even lower coverage of pregnant and lactating women. In this context, Mary AndErson's (1987) detailed and painstaking research in two districts of Gujarat and Maharzahtra deserves special mention. According to this study, the coverage of essential w.alth services for pregnant women (such as tetanus-toxoid immunisation) was very 'Lcw al: 20%, coverage of DPT for children in the age group 1-3 years was less than 101; and nearly a half of the severely malnourished under-threes were not reached by * Vaccine against diptheria, pertussis and tetanus. 51 supplementary nutrition. In brief, this study showed that both health and nutrition coverage of women and children under three was far from satisfactory. Considering that the incidence of malnutrition is highest among them, reaching them inadequately is a major shortcoming of ICDS.33 4.17 Such coverage problems mean that ICDS fails to prevent many preventable child deaths, as illustrated by an interesting study of deaths in 8 AWs in Medak District, Andhra Pradesh. (Reddy and Reddy, 1986). This year-long study was conducted in 1985. The study recorded that there were 47 child deaths; nearly 1/3 of the deaths took place before the first birthday. The study showed that nearly 25 out of 47 deaths were preventable via immunisation and/or ORT* Clearly, if ORT and immnunisation were systematically practiced in the Anganwadis, the mortality rate could have fallen. 4.18 As for supplementary feeding, two points are of interest. First, most studies pointed out that proportionately much higher percentage of 3-6 years old are covered under suppplementary nutrition than 0-3 year olds. Sharma (2p. cit.) noted that the difference in coverage across age groups could be attributed to the inability of the 0-3s to come to AWs on their own. The need for someone to accompany them imposes a constraint on their coverage. Second, almost all the Schedule Caste and Schedule Tribes attending A;Ts seem to be receiving supplementary nutrition, as can be seen from Table 20. To this extent, the "area targetting" implicit in the program is 33The failure to reach pregnant women and children below three years of age was noted by scholars nost sympathetic to the ICDS program. See, for example NFI (1988) p. 43. * Oral Rehydration Therapy. 52 Table 20 Supplementary feeding under ICDS by Social class, 1981 Proportion of Proportion of ---------------------------------------------------------------__------------__---------------------------.------.. SC Children ST Children receiving supple- receiving supjlAe- SC Child- SC Children mentary nutrition ST Child- ST Clhildren mentary nutri icn Sn ren in the in ICDS to total children ren in the in ICDS to total children No. States State Projects in ICDS projects State Projects in ICDS projects 1 2 3 I 4 S 6 1 ANDHRA PRADESH 14.9 19.4 18.4 5.9 27.7 29 2 ASSAM N.A. 11.7 9.6 |N.A. 36.8 34.2 3 BIHAR 14.5 21L.6 26.9 9 8.3 23 30.7 4 GUJARAT 7.2 .1 5.5 14.2 35.2 30 HARYANA 19.1 18.9 19.2 N.A. 0 N.A 6 HIMACHAL PRADESH 24.6 27.7 30.6 4.6 9.3 11.6 7 JAM4U & KASHMIR 8.3 10.7 6.9 N.A. 0 0 8 KARNATAKA 15.1 25S.7 25.3 4.9 1.1 1.4 9 KERAIA 10 lC.5 11.5 1 6.4 6.5 10 MADHYA PRADESE 14.1 18.4 28.5 23 68 55.8 11 MAHARASHTRA 7.1 38.2 41 9.2 23.9 26.2 12 ORISSA 14.7 16.3 14.1 22.4 26.6 29.4 13 PUNJAB 26.9 45.3 27 N.A. 0.9 0 14 RAJASTHAN 17 20.6 22.2 12.2 39.1 35.7 1S TAMIL NADU 18.4 28.5 37.1 1.1 6.3 7.5 16 UTTAR PRADESH 21.2 32.9 42.5 0.2 10.4 7 17 WEST BENGAL 22 18.4 18.4 5.6 13.8 17.9 Note: SC - Schedule Caste ST - Schedule Tribe Source: Comuted from Child in Indis 53 rendering it accessible to the scheduled castes provided they attend the AWs. But at the same time, as shown by Mary Ann Anderson (1987), many children in need especially in the age-groups 0-3, both scheduled castes and others, do not get fed because they do not attend the AWs. Sharma's (1987) study which is based on a sample covering many states, also concluded that while more than 60% of the beneficiaries of supplementary feeding were from households below the poverty line, the 0-3 was low and even showed some decline especially in rural and tribal blocks. Third, an extensive review of earlier research on supplementary feeding in ICDS has pointed out that the duration and regularity of feeding varied a great deal and was frequently less than the prescribed 300 days; the study also argued the case for a review of the screening, selection and coverage of eligible children for feeding so that children most in need get the supplement. (Khanna and Ohri, 1986). A recent study by NIPPCD (1988) confirms these earlier findings. It is interesting that food supplies were erratic even in a state like Punjab where food is otherwise plentiful.34 However, the NFI study on ICDS argues that even though the present feeding program leaves much to be desired, and "at best may only have a marginal impact on the child's nutrition," it is serving as a "bait" for inducing mothers and children to AWs where they can be offered other imRortant services of the package as well. 4.19 The "other important services" offered by ICDS have also been evaluated by many scholars, the most recent comprehensive studies being those of NFI (1988), and Sharma (1987). Both these studies were based on large, comprehensive samples drawn from many states. The NFI study pointed out that weak links with the health infrastructure persisted in most locations. This is supported by indirect evidence 34See Report on the Evaluation of ICOS in Punjab, Government of Punjab. 54 on the distribution of medical personnel given in Table 21 and 22. The ni:mber of Medical Officers, Lady Health Visitors and Auxiliari, Nurse-Midwives actualltr posted as a percent of sanctioned strength varied a great deal across the states. For example, LHVs posted as a percent of sanctioned strength was low in Haryana, flirnachal Pradesh, Madhya Pradesh, Orissa, Rajasthan and West Bengal; and that of ANT's in Table 21 Staff Position - Paramedical Functionaries in ICDS PHCs reported as on 31st March, 1987 No. of LHV L H V A N Ms FHCs Sanc- Posted -------------------------- Studied tioned ---------- Sanc- Col 6 Sn. No. % of tioned Posted as % No States Col 2 No. of ccl 5 1 2 3 4 5 6 7 1 ANDHRA PRADESH 18 118 96 81.4 495 375 75.8 2 ASSAM 7 43 37 86 290 279 96.2 3 BIHAR 32 110 90 81.8 573 534 93.2 4 GUJARAT 45 320 287 89.7 1491 1357 91 5 HARYANA 14 108 61 56.5 657 569 86.6 6 HIMACHAL PRADESH 9 38 27 71.1 270 174 64.4 7 JAMMU AND KASHMIR 8 7 0 0 104 52 50 8 KARNATAKA 168 614 546 88.9 3634 3211 88.4 9 KERALA 71 298 224 75.2 1503 1278 85 10 MADHYA PRADESH 44 350 249 71.1 1564 1105 70.7 11 MAHARASHTRA 84 1365 1141 83.6 3932 3571 90.8 12 ORISSA 39 205 155 75.6 887 784 88.4 13 PUNJAB 22 132 116 87.9 592 499 84.3 14 RAJASTHAN 36 203 151 74.4 1197 995 83.1 15 TAMIL NADIJ 14 10 9 90 176 166 94.3 16 UTTAR PRADESH 80 425 398 93.6 2189 1902 86.9 17 WEST BENGAiL 48 783 510 65.1 1909 1376 72.1 739 5129 4097 1271.9 21463 18227 1401.2 Source: Central Technical Committee, All India Institute of Medical Sciences, Annual Report on ICDS, 1986-87. 55 Himachal Pradesh, Jammu & Kashmir, Madhya Pradesh and West Bengal. The percentage of trained medical officers to total posted to look after ICDS activities was very low in most states (col. 7, table 22). Thus, the evidence suggests that health links appeared weak at all layers of health infrastructure. The referral system was poor; "the AWW neither kept satisfactory records nor did they feel responsible to forward identified cases" (Sharma, p.102). According to the same author, "the poor performance on survival indicators points out that the health status of children in the rural blocks requires attention. The situation could be attributed to under-utilisation or non-availability of health services." (Sharma, p. 74). Table 22 Staff Position - Health Functionaries - Medical Officers ICDS PHCs reported as on 31st March, 1987 No. of MOs MOs Posted MOs Trained PHCs sanc- -------------- ------------------- S.No. States Studied tioned No. % No. % 1 2 3 4 5 7 1 ANDHRA PRADESH 22 81 62 76.5 23 37.1 2 ASSAM 14 94 76 80.9 39 51.3 3 BIHAR 25 100 68 68 37 54.4 4 GUJARAT 39 144 114 79.2 58 50.9 5 HARYANA 18 108 79 73.1 39 49.4 6 HIMACHAL PRADESH 9 69 58 84.1 29 50.0 7 JAMMU AND KASHMIR 10 80 72 90 64 88.9 8 KARNATAKA 131 895 773 86.4 592 76.6 9 KERALA 65 314 299 95.2 187 62.5 10 MADHYA PRADESH 36 197 145 73.6 103 71.0 11 MAHARASHTRA 49 475 370 77.9 227 61.4 12 ORISSA 36 184 151 82.1 99 65.6 13 PUNJAB 20 297 264 88.9 162 61.4 14 RAJASTHAN 32 263 210 79.8 168 80.0 15 TAMIL NADU 35 38 37 97.4 32 86.5 16 UTTAR PRADESH 55 215 167 77.7 90 53.9 17 WEST BENGAL 41 208 178 85.6 84 47.2 Source: Central Technical Committee, All India Institute of Medical Sciences, Annual Report on ICDS, 1986-87. 56 Mothers' pressence at AWs was not exploited for imparting nutrition education. 4.20 Child-related nutrition schemes are both person-intensive and resource-initensive. The responsibility of the functionaries is to see that their tasks matchL the needs of children and their mothers. Since young :hildren (especially below three years of age) are in the fold of their mothers, inter;rentions aimed at them simply cannot succeed unless there is total identity cf the functionaries with the mothers. This is possible only when AWWs are able to elicit community participation, This usually implies (a) involvement of the commlxnity in achieving ICDS goals via participation in decision-making at regular workir.g group meetings of women and ATWs; (b) contribution of the community in cash, kind, irta-.:power and organisational help; (c) involvement of the community in assessment of Le:aefits from the program and most important of all, (d) regular home visits by AWs to achieve (a) to (c) above. ICDS evaluations are unanimous on one point: the p:^ogram failed to elicit community participation as described above in most locatiol,s. The most recent evaluation by Sharma (op. cit.) pointed out that since hea>th and nutrition activities were rarely conducted in women's working groups, thnie wias hardly any c:hance for mothers' involvement. As noted by NFI study, the ::ask of eliciting community participation is admittedly difficult in villages stratifLed by social barriers and economic status. This is because there is no rallying cr nodal point in the villages where communities, especially of the deprived sections, can be effectively reached. So far these communities have regarded ICDS as a govi!ynment service to be used in emergency, and not as their own program. Special effo:-ts are clearly needed for converting the prevailing status of ICDS as a "government :-entre" into that of a "community centre". 57 4.21 Indirect evidence about the quality of ICDS services can also be gained from statistics on the trained personnel, and on the supervisory ratios. There appear to be significant inter-state differences in the distribution of trained workers and supervisors. The percent Table 23 ICDS: Recruitment & Training of Personnel in Central Sector Projects: 1987 SI STATE / U. T. PERCENT PERCENT PERCENT SUPER- AWWs NO CDPOs SUPERVISOR AWWs VISOR PER TRAINED TRAINED TRAINED PER CDPO SUPER- (%) (%) (%) VISOR 1 2 3 4 5 1 ANDHRA PRADESH 78.46 82.21 80.93 5.02 26.48 2 ASSAM 82.05 78.60 86.93 5.51 19.35 3 BIHAR 86.73 82.50 85.41 5.13 24.34 4 GUJARAT 85.42 79.28 83.18 11.06 16.07 5 HARYANA 75.00 97.56 90.09 4.39 28.62 6 HIMACHAL PRADESH. 66.67 60.00 81.53 3.57 25.33 7 JAMMU & KASHMIR 85.71 90.63 87.67 2.29 37.00 8 KARNATAXA 95.56 88.97 88.87 8.87 20.47 9 KERALA 80.43 90.00 94.66 4.13 27.82 10 MADHYA PRADESH 92.00 53.27 81.92 6.93 22.70 11 MAHARASHTRA 82.14 76.95 74.27 7.80 21.81 12 ORISSA 100.00 88.64 87.03 4.63 19.68 13 PUNJAB 80.65 78.24 87.95 5.48 21.19 14 RAJASTHAN 30.61 67.90 76.05 5.53 24.22 15 TAMIL NADU 59.18 87.50 74.60 4.08 22.43 16 UTTAR PRADESH 98.78 72.10 90.16 6.51 20.63 17 WEST BENGAL 83.15 98.39 92.03 6.99 19.45 Source: Computed. of Child Development Project Officers (CDPO) trained varied from a low 31% in Rajasthan to 100% in Orissa; those of supervisors from 53% in Madhya Pradesh to 98% in West Bengal; and AWUs from 74% to 94% (see table 23). The NFI study also pointed out that even though the training manuals prepared were good, the actual training carried out in 200 centres spread over the country was apparently of uneven quality. In six out of seventeen states the number of supervisors per CDPO exceeded the norm 58 of about 5 anid in 9 the number of AWWs per supervisor exceeded the norm of about 210. The NFI stud,y M. cit. noted that the supervisor's visits to tribal areas were more infrequent than to other areas which was particularily unfortunate, because the AWWls in tribal areas were found to be much less competent and hence needed intensive supervision.35 4.22 The government and the donor agencies supporting ICDS now recognize the need to take a fresh look at training needs and capacity. In particular, four shortcomings have been Ldentified. First, the existing capacity is inadequlate t.o meet the demands of the system and as a result the backlog of trainees has been growing. Second, the training syllabus and materials are much too standaritsed and do not allow for variations to suit tribal, rural and urban areas. There have practically been no attempts to create awareness among the trainees aboat the location-specific disease patterns which is possible only when the traLning infrastructure is sufficiently decentralised and taken to the bloclk-level. Tlird, most evaluations have shown that the training of middle-level personnel (supervis;ors) has been the weakest.36 Fourth, the existing provision is totally inadequa:e for retraining of all types of functionaries. On the whole, lack of supervisory Ltput'i and rigidity in training syllabus and materials may also have contributed I:c the observed inter-project differences in the impact of the program. 35The problem of training and supervision are not peculiar to ICOS. It is encountered in other pctety alleviation programs as well. Therefore, there appears to be a case for training a cadre of workers to idesitfy poverty groups ztnd get through l:hem, and sensitize them to their problems. In this context, it is wo-th noting that much of the success of tenancy reform in West Bengal (Operation Barga) could be traced t: ;he orientation camps of beaurocrats with beneficiaries. 36For example, the WFP report concluded that the supervisors on the whole did not seem to be ir a position to help their AWWs. Similiar views were expressed by CARE officials in informal discussions. Te IlFI study OD. cit. noted that the Supervisors themselves were not receivinq adequate support from Child Develcpmant Project officers. Similar findings were noted by Sharma OD. cit. 59 4.23 To sum up, though on the aggregate ICDS impact on nutritional status was shown to be "satisfactory" by the monitoring cell, there appear to be significant inter-project differences in its impact. While systematic analysis of the factors accounting for these inter-project differences must await project-wise longitudinal data on beneficiaries and eligible non-beneficiaries, rural and tribal projects seem to be performing much worse than urban projects. Health linkages, referral services, mother-counselling and nutrition education, coverage of under-threes, and community participation appear to be unsatisfactory in most projects. This raises the issue whether the observed infirmities of ICDS are only a passing phase, likely to disappear with more effective implementation, or whether there are any deficiencies in the program design that are limiting its effectiveness even in the administratively better-endowed states. The NFI study op. cit. noted that the Supervisors themselves were not receiving adequate support from Child Development Project officers. Similar findings were noted by Sharma op. cit. 4.24 The most popular view best articulated in the NFI study is that ICDS is a well-conceived program reflecting a bold and innovative approach; many of its observed ills can be removed by effective implementation; and the observed failure in eliciting community participation is not peculiar to ICDS, but evident in other welfare programs such as IRDP. Probably, meaningful lessons in areas such as community participation and mother counselling can be learnt, not by comparing ICDS with dissimilar welfare programs like IRDP, but with similar nutrition and health interventions such as the Tamil Nadu Integrated Nutrition Project (TINP). It must be stressed that the two programs differ in their objectives as well as coverage and as such a comparison of the two programs' inputs and outputs is problematic. I am * Integrated Rural Development Program. 60 not attempting this. Instead, a comparison of some of the qualitative components of the two programs and the job description of the principal worker is attempted, so that the factors underlying the observed differences in such aspects as cormunuity participation can be better understood. 4.25 TINP covers 9000 villages in rural Tamil Nadu with a population of over ten million. Each Community Nutrition Centre is staffed with a Community Nut.rit:Lon Worker (CNW) and a helper. The services provided include weighing of all children under three years, public health interventions for both mothers and 6-36 morth old children that are similar to those in ICDS, and a carefully targetted supplfmentary feeding for those children under three years of age who are severely malno:urished or are not gaining weight. Nutrition education of mothers is an important compDnent, but there is no pre-scbool activity. A limited number of pregnant ant l.ctating mothers also receive supplementary feeding. The project was designed with a heavy monitoring a1nd evaluation component including base line, mid-term and te=minal Surveys. Mid-term evaluations showed that severe malnutrition rates of 15 1:o 20% fell to 8 to 9% over a four-year period range. In the project areias the dec.l:ne in severe malnutrition ranged 32 to 50% compared to 12% in the control block;. The category of normal children rose.by 20% in the project areas compared to 5% i.n the control blocks. Unlike in ICDS, progressive reduction in the incidence of severe malnutrition was noted in every project area. The findings of the final evaLtation are consistent with the mid-term review; the project is effective in cont::clling malnutrition in the post-weaning period and in improving the nutritional st.tus of pre-school children. 4.26 The three qualitative differences between the ICDS and TINP relate to (a) 61 the coverage of beneficiaries and consequential job description of the principal workers, (b) involvement of mothers and the general levels of community participation; and (c) recruitment and supervision of workers. These are discussed below with a view to gaining insights into differences in the productivity of the two programs. 4.27 As already noted, ICDS is aimed at all children 0-6 years of age, and pregnant and lactating women, whereas TINP serves only children 6-36 months of age, and mothers. This difference manifests itself in the job description of AWWs under ICDS and CNWs under TINP, The responsibilities of the former are manifold: non-formal preschool for the older children; supplementary feeding; imparting health and nutrition education to mothers; home visits for ensuring health care for infants and pregnant women; referral services for the needy children and liason with para-medical and medical personnel and other local organisations like Mahila Mandals; eliciting community participation; and maintaining records. According to a recent study, nearly five hours are spent by a majority of AWWs under ICDS on three activities: pre-school, feeding and record-keeping. By contrast a CNW's activities are limited: she first conducts a survey of the entire village, weights all children 6-36 months of age, and identifies children of the target group. For these children growth charts are plotted (one is retained with the mother and one at the Centre). Needy children are given supplementary feeding by CNW, and each child's progress is watched. Besides weighing and feeding, CNW's other job include ORT, administering vitamin A, organising mother's working groups and eliciting community participation in all projects activities such as weighting of children, feeding growth-chart maintenance, etc. Thus the whole job description of CNW is centred around pregnant women, mothers and children of 6-36 months age. She is not burdened with preschool 62 activity, and older children (3 + age) are not covered under the program. 4.28 The basic purpose of the above comparison of the two programs is to E.ddrass the question: how can the observed deficiencies of ICDS - the relative fa:Llure to reach 0-3s and pregnant and lactating women, the poor health and nutrition edlucation, and the neglect of home visits and community participation - be improved in :lhe near future? One of the factors contributing to unsatisfactory performance oE ICDS appears to be the overloaded job description of AWWs. The bias towarcDs older children seems to be a natural consequence of its broader objectives - nob: simply a problem of management or implementation or lack of supervision, although these problems do seem to plague the program especially in tribal and rural areas. AWWs preference for the easier job of handling 3-6 year old children, and neglect of under-threes, is probably rational under the circumstances. This is not to miinimise the significant service of pre-school being offered by ICDS. Pre-school offers sizable benefits to the children, their parents and communities as a whol.e. It provides child care, preparation for formal schooling especially fo) the disadvantaged, and girl children among them; and has linkages into the formal school system that encourage school attendance. It also attracts long-term partic;Lation of all mothers with the anganwadi, as they receive a real benefit everyday (unlilce TINP, where only those in "need" receive such a benefit). Thus pre-schoo. is eLn important service under ]CDS. The issue that merits a debate is: can a AWW p>erform efficiently all the nutrition-health related functions of all children unde:c six years of age and mothers, and also run a pre-school, especially in regions wLth a high proportion of program clients to child population. 4.29 If the broader objectives of ICDS are to be maintained intact, it appears 63 desirable to recruit; two AWUs, one for pre-school, feeding, and health referrals for the older children, and another to look after under-threes, pregnant and lactating women, nutrition education and home visits. This of course is likely to raise the unit costs of the program. Therefore, to begin with it appears desirable to experiment with two-AWW model at least in those projects where the percentage of malnourished children (i.e. program clients) are likely to be higher than the average for all AWWs. Assuming that this is the case in atleast 25% of the Projects, how much does 2-AWW model cost the exchequer? 4.30 It may be useful to recall that the total direct expenditure on nutrition support by both Central and State Governments amounted to only Rs. 4262.5 million37 during 1986-87. Of this, the state governments met Rs. 3220 million, or 75% of the total expenditure. The amount budgeted for 1987-88 is Rs. 4950 million. It is also worth noting that state governemnts' expenditures on direct nutrition support amounted to 1.86% of the state plan budgets during 1986/87, while the Central government's expenditure amounted to only 0.5% of the central budget. Thus, as of now, direct expenditure on nutrition support especially by the Centre, appears to be rather modest, notwithstanding the improved allocation for ICDS during the Seventh Plan. 4.31 At present, the average annual expenditure per AW works out to Rs. 7000. A non-matriculate Air is paid Rs. 200 a month, and a matriculate Rs. 250 a month. Assuming an average of Rs 225 per month, an additional AWW would cost Rs. 2700 per annum. With a second worker, the coverage of children under three and mothers may 37This includes Central and State expenditures on Special Nutrition Program, Mid-day-meals, and ICDS. 64 be expected to improve, so that expenditure on nutrition supplementation may also go up. Assuming an addiLtional Rs. 1000 per annum on nutrition supplementation due to better coverage, the two AWW model may cost about Rs. 10,700 per annum, or Rs. 11,000 per annum allowing for possible rise in overall costs in the recent period. The incremental cost per AW would thus work out to Rs. 4,000 per annum. Lssuming that at least 40,000 centres (out of 160000 AWs in operation) would need an add'itional worker, the total annual incremental cost would be Rs. 160 million. This is likely to increase Centre's ICES annual outlay by Rs. 120 million, i.e. from the ,tirrent level of Rs. 1470 million (as indicated in the Annual Plan, 1987-88), to R!;. 15iL0 million, or by about 9%. The states share of expenditure will increase by Rs. 40 million, assusming no further contribution of free food by the donors. It needs to be stressed that the total direct expenditure on nutrition support by the Centre, as a per cent of total Central plan outlay, would still be modest at: less than ).6%. Besides, there are ways of further minimising costs. For instance, adolescent g~irls in villages, who are also malnourished, can be enrolled into the program for assisting AWWs. In return she could receive supplementary food and certain esselitial mothering skills. (Gopalan). In this context, it would be interesting to learn from Tamil Nadu where two AWW system has just been introduced in some districts. 4.32 Quite apart from strengthening selected Centres, there also appears to be significant scope for improving the productivity of AWs throughout the countr7y via better training and supervision. In this context, TINP's experience is worth mentioning. TINP's supervision and training systems and retraining of peripheral workers helped to produce a highly motivated and skilled work force. While the training components of ICDS have been researched and found to be gen!rally unsatisfactory, (NFI, 1988), the factors that contributed to successful supervi;i,ion 65 and training in TINP deserve to be studied further, so that these can be replicated elsewhere. However, particular qualitative differences relating to selection and training of CNWs under TINP may be noted, since this not only endured their stake in the village environment but also contributed to the relative ease with which community participation could be elicited. A CNW under TINP had to be a married woman settled in the village with two healthy children, educated up to VII grade, preferably belonging to the scheduled caste community. Efforts were made to locate one such woman in each village, acceptable to the village community. Such women were then persuaded to volunteer for CNW jobs which carried very low wages. A CNW first gets training at the block level for 60 days. The training includes creating of an awareness in location-specific child disease and treatment procedures. In TINP blocks, there are hardly any CNWs who do not belong to the village. In fact most of them have deep roots in the village. This facilitated social mobilisation and raised the level of knowledge and awareness of entire villages on issues such as child and maternal health, community hygiene, etc. It is possible that for AWW under ICDS the recruitment procedures may have been similar in principle. But in practice the recruitment procedures for AWWs under ICDS appear to be much less stringent, and the training inputs less rigorous and location-specific. According to Sharma, nearly a half of AWWs under ICDS did not belong to the local village; this proportion was nearly 75% in tribal blocks (Sharma, p. 24). The same study also confirmed that nearly 38 80% of AWWs in the sample AWs belonged to upper castes especially in rural blocks. Both 3See Sharma, 9p. cit. One reason could be that the renumeration for AWW was set at a higher level than for TINP workers, including the upper caste women to corner the jobs. Sharma also noted that there was no statistically significant association between "caste status" and community participation. This finding is somewhat questionable, because the extent of community particiaption in general in all spheres of ICDS was found to be dismal in the sample villages. Moreover, caste status of AWWs seemed to affect the intake of clients into the program. Thus in rural blocks where the AWWs recruitment was dominated by upper casted, only 24% of the beneficiary households belonged to scheduled castes. Given the Indian cultural context, it is somewhat naive to expect women of upper castes in the villages to look after the welfare of the children of the scheudled castes. 66 in recruitment and training of AWWs there appears to be considerable si:cpe for improvement .39 4.33 Finally, one of TINP's great merits has been successful community-based growth monitoring using mother-retained growth charts and supported by itmediate real interventions for needy children. The basic purpose of growth-monitoring in TINP has been for early detection of growth faltering, backed by immediate referral services. At the same t:ime, mother-retained growth charts also resulted in great:er involvement of mothers in the program, and in their nutrition education. 0 4.34 To sum up, ICDS is an important watershed in the evolution of public ;'olicy towards child and maternal malnutrition in India. With the inclusion of welfare of women and children in the government's 20-point program, ICDS has come to st:ay as a model for future integrated child development activities. An objiective assessment of the various impact studies and social components of the program suggest thLat the program as it is now in operation is tilted in favour of older children at the expense of under-threes among whom the incidence of malnutrition is highest:. The coverage of pregnant women also appears to be unsatisfactory. Accurate assinssmerLt is handicapped because ol the absence of project-specific control groups in ]:ost of the studies. The available studies suggest that the impact on nutritional status varies a great deal across projects, rural and tribal projects faring much wor,n e than 39That all is not well with AWW recuitment is also clear from recent press reports. See for ex.:mple, Indian Express, (1988). 40Growth-monitoring has been attacked by nutritionists in India as a "waste of time" (Gopalan). We do not wish to enter into the debate, except to say that it would indeed by a waste of time if used on y for screening children for supplementary feeding. However, under TINP, contrary to popular misconceptions. qrcwth monitoring via mother-retained charts is used primarily for early detection of faltering, for mot:hers' involvement and cooperation, anc for their education. TINP Community Nutrition Workers have been trained to use charts as one of the important vehicles for imparting health and nutrition education to mothers. Even casual visitors to TINP blocks cannot fail to notice the enthusiasm with which mothers respond to the use of charts which are carefully preserved by them. 67 others. The program as it is now in operation seems to suffer from weaknesses with respect to recruitment, training and supervision of the core workers. However, the concept of providing an integrated package cf health and nutrition services exemplified in the ICDS has undoubtedly immense potential for reducing malnutrition in India. If appropriate changes in the program design are effected, and the recruitment, training and supervision procedures of the core workers are overhauled, some of the observed shortcomings of ICDS can be removed, and its productivity improved. The program can also benefit from incorporation of some aspects of the TINP approach such as training and supervision, mother counselling and nutrition education, communilty participation, etc. In some projects with high percentage of program clients, t:wo-worker model appears to offer potential for improving ICDS performance. This is likely to enhance costs, but it must be stressed that the overall direct nutrition support expenditure amounted to only 1.1% of total plan outlay of the Centre and the States during 1986-87. (b) Public Distribution System 4.35 It is now recognized that despite the improvement in the national food supply position, poor households especially in the less developed states are unable to improve their per capita foodgrain consumption levels essentially due to lack of purchasing power.41 The existing distribution of income and wealth and the regional concentration of production activity may be such as to leave large segments of society without the necessary purchasing power to buy staple foods in quantities that 4 For example, the per capita monthly foodgrain consumtion (at 1970-71 prices) for the lowest three deciles in Bihar 1983/84 ws (Rs. 15.58) less than half of that of Punjab. It was lso low in Andhra Pradesh, Madhya Pradesh, Karnataka, Kerala, Orissa, Maharashtra and Tamil Nadu (i.e. less than Rs. 18). In general, per capita food consumption of the lowest three deciles was low in the agriculturally lagging states, and in the chronically food-deficit states. (See Gupta, 1987). 68 are deemed necessary from a nutritional point of view. In the medium run, P:S has been envisaged as an important instrument to ensure access of essential coml:o5lities like foodgrains to vulnerable households especially in rural areas where tehe poor have to contend with low incomes as well as high weather-induced variability of incomes. In the long-run, the solution to this problem would be an improveunnt in the income/employment situation of these households and, interventions ouch as food-for-works have been developed to improve their income/employment sitiation. * Such interventions to improve and stabilise income/employment leveils (e.g., EGS ln Maharashtra), those aimed directly at vulnerable individuals within vulnierable households (child and maternal nutrition programs) and those aimed at vulterabl.e households as a whole (PDS) should be regarded as complementary to each other, rather than as alternatives. 4.36 Theoretically, PDS, by providing foodgrains at lower than market pri:e to the vulnerable households, can play a significant role in the medium-run in maintaining, if not in improving, the nutritiorLal levels of these households, as is demonstrated by the experience of Kerala and Gujarat.42 The central role of ID; has come to be recognised even more following the drought of 1987. The state gover:mntents were quick at expanding food-for-works programs in the affected regions, and ir. those areas where a well-established and properly supervised distribution system esisted to carry foodgrains to the work sites, the effectiveness of the employment pr:)grams increased. Clearly, a more widespread rural coverage of PDS to targetted popui.ation would enhance the outreach and effectiveness of income-generating wrorks pro,.;rams. Other advantages of PDS have also been recently recognised. In an extremely del:ailed 42The experience of Kerala and Gujarat are well documented by George (1985). * EmpLoyment Guarantee Scheme. 69 study of PDS in Andhra Pradesh, Venugopal noted that the newly introduced Rs. 2 a kilo scheme, by ensuring access to grain in the offseason at a low price, reduced the dependence of labourers on landlords for grain loans. As a consequence, the institution of "attached labourers" - an onerous contractual arrangement in existence for over a century - is on the decline. Moreover, the labourers are also taking advantages of seasonal hike in the wage rate during the peak season and demanding cash wages. Clearly, an effective PDS appears to generate other favourable spin-offs. 4.37 Recognising the above-mentioned advantages, serious attempts to develop PDS as an effective instrument for providing nutrition support to poor households have been proposed in the Seventh Plan. However, like other nutrition interventions, PDS coverage and extension to rural areas depended upon the state's own initiative and resources. This is because the distribution of food supplies from the Centre is not adequate to maintain PDS round the year. Therefore, states interested in improving the coverage must purchase additional grains within the state, thereby invariably incurring a subsidy expenditure. Further, the distribution of Central supplies have not always been in correspondence with state-wise needs. In actual practice, barring the states of Kerala, Andhra Pradesh, Tamil Nadu and Gujarat, PDS coverage of rural areas is reported to be weak. State-wise distribution of total PDS supplies is shown in table 24. An important limitation of this data is that rural/urban breakup of quantities distributed is not available. Further, these data show only the Centre's 70 Tabie 24 PDS Supplies and Population Below Poverty Line (Thousand Tonnes) Distribution of PDS Supplies to Percent Distribution of State Govt. and PDS Supplies to State Direct to Falr Govt. and Direct Sales Price Shops to Fair Price Shops X Diatribtuion ----------------------- ------------------------ of Population Rice Wheat Total Rice Wheat Total below Poverty States Foodgrains Foodgralns line Andhra Pradesh 991 103 1094 16.70 3.23 11.99 7.6 Assam 221 141 362 3.72 4.42 3.97 1.5 Bihar 84 271 355 1.42 8.49 3.89 13.5 Gujarat 92 18 110 1.55 0.56 1.21 3.2 Haryana 9 53 68 0 .15 1.66 0.75 0.8 Himachal Pradesh 31 23 54 0.52 0.72 0.59 0.2 Jammu & Kashmir 121 77 198 0.04 2.41 2.17 0.4 Karnataka 232 74 306 3.91 2.32 3.36 5.1 Kerala 1332 146 1468 22.28 4.58 16.09 2.6 Madhya Pradesh 167 46 213 2.81 1.44 2.33 9.4 Maharashtra 271 314 585 4.57 9.84 6.41 8.6 Orissa 43 146 189 0.7. 4.58 2.07 4.4 Punjab 1 15 116 0.02 0.47 0.18 0.9 Rajasthan 5 10 15 0.08 0.31 0.16 4.7 Tamil Nadu 271 95 366 4.57 2.98 4.01 7.4 Uttar Pradesh 168 138 306 2.81 4.33 3.35 19.6 West Bengal 731 669 1400 12.32 20.97 15.34 8.3 Other States aad Union TerritorLes 1174 852 2026 19.78 26.70 22.20 1.6 All India 5934 3191 9125 100.0 100.0 100.0 100.0 Note: PDS data belongs to the year 1984: distribution of people below poverty line is for the year 1983/184. Sources: Bulletin of Food Statistics, 1985-86: and the Planning Commiasion. direct contribution to the states, but do not include the state's own procLrement efforts - alithough the onily major state effort is in Andhra Pradesh. From the tabLe it is clear that states that account for a substantial share of all-India ?cverl:y population like Bihar, Uttar Pradesh, Madhya Pradesh and Rajasthan account for on2Ly a small share of the PDS supplies. A substantial share of supplies is going to the big cities. For example, Union Territories (Delhi in particular) accounted for 22% of the supplies, but their share in the poverty population was only 1.6% TIiat the regional distribution of PDS supplies is in general unrelated to the incid!nze of both rural and urban poverty is much too obvious to need futher analysis The pattern of mismatch between supplies and needs under PDS is much the same as ii'or the direct programs. Considered along with ICDS, the broad conclusion is gross iriequity 71 in the allocation of resources for nutrition. 4.38 PDS as it has operated so far in India is also expensive because wherever it was introduced recently, no serious efforts were made to limit its access to the vulnerable groups. Not only have costs increased, but effectiveness has been reduced. For example, in Andhra Pradesh, where the coverage is set wide, PDS met only 34% of the minimum requirements of rice and 31% of oil for the poorest segments (green card holders).43 Recent reports on the functioning of the scheme suggest that the government could ensure a supply of only 1.7 million tonnes. If the entire ration quota to only green card holders were to be supplied, 2.5 million tonnes would be required. Most green card holders could not obtain even a small share of their quota of ration. Clearly the government is experiencing difficulties in ensuring supplies; restricting its access to green card holders would substantially ease supply problem while serving the poor better. As at present, it is reported that the free market price of rice (where the poor now buy more than two-thirds of their requirements) rose sharply in the recent period. Thus, the scheme is becoming ineffective in serving the poor. 4.39 As of today, the Central government has been incurring a subsidy cost of Rs. 7500 million per annum on PDS. Yet vulnerable groups in many states do not have any access to PDS supplies, while the poor as well as the non-poor have access to it especially in the urban areas. The role of PDS during normal times and in drought years as a cost-effective instrument of food supply can be realised only if its 43This is based on the assumption that a poor household's (with a family size of 5) consumption of staple foodgrains approximate to 60 kg/per month. Given that a poor family's household size may be expected to be larger than 5, this estimate must be regarded as a lower bound. For an overall assessment of the scheme, see Venugopal (1987), and Radhakrishna and Indrakant (1987). 72 access is restricted to the hard-core poor. Assuming this can be done in princilple, how much would it cost the government? 4.40 We begin by assuming that a household with five members would require a minimum of 65 kg of rice per month to maintain bare minimum nlutritional Level," Following NSS 1977, average expenditure of the poor per annum per household works out to Rs. 3578 at 1987 prices. According to the same NSS data, 60% of this expenditure was on foodgrains. A poor household would be spending about R3. 245C per annum on foodgrains or buying 58 Kg per month assuming rice price at Rs 3.50/kg which was the prevailing retail rice price during 1987.45 This leaves a shortfall of 7 kg per month, or Ps. 294 shortfall in income/expenditure to reach the desired level of 65 kg. per month. 4.41 Assuming a minimum annual income support of Rs. 100/per household vi.l pubLic works programs (NREP and RLEGP), PDS would still have to effect a net income 1.1ansfer of Rs. 200/- per annum per household if the observed nutritional gap is to be filled. For inducing poor households to avail of PDS, we consider that a minimum diffl:rential of Rs. 1 between the market price and the issue price under PDS; or an isst.e price 30% lower than the former which ruled at Rs. 3.50 during 1987, would be esse2ltial. In other words, PDS would have to supply 200 kg per annum per household for effecting an annual income transfer of Rs. 200. For 58 million poor households in the colmtry, the annual ioodgrain requirements would thus work out to be 11.6 million tonnes. 44The estimates contained in this para are drawn from C.H.H. Rao, S. K. Ray and K. Subbarao (.968). 45We are using rice only for illustrated purposes. Other grains (millets) could be used. Su:ip ying interior grains which figure prominently in the budget of the poor has been advocated for ensuring a .rertain amount of self-targetting, but this suggestion has so far not been put into practice. 73 At present the government is distributing 10 million tonnes and incurring a subsidy cost of Rs. 7500 million or Rs. 750 per tonne. Assuming such perfect targeting the scheme would cost onily an extra Rs. 1200 million and an additional 1.6 million tonnes. 4.42 The above exercise is intended to show that with roughly the same expenditure as is being incurred now, it would be possible to ensure food security to all critically vulnerable households, provided the access of PDS is limited to them. Experience gained so far indicates that perfect targetting of food subsidy is operationally difficult especially in areas where the poor and the near-poor are too numerous, although in some states the coverage may have been set wide deliberately. There is undoubtedly a need for experimenting with self-targetting methods such as supply of inferior goods. The huge subsidy cost for a PDS with unlimited access has no economic justification whatever. In this context, the experience with a more targetted PDS in Kerala, Gujarat and Tamil Nadu is very important. 4.43 Kerala introduced PDS in rural and urban areas as early as 1964. Because Kerala is a chronically food deficit state, the PDS also serves the non-poor, but their access is very limited to 10 to 15% of requirements, whereas nearly two-thirds of the total grain requirements of the poor are met from PDS. Eiigibility to ration is determined on the basis of multiple criteria, one of which is land holding size. The evidence is that the distribution has favoured lower income groups (George, Kumar). The near-universal coverage led to the population developing a stake in the efficient functioning of the system and reduced leakages. The fact that private trade was never abolished, and a dual market always enabled the rich to buy from the open market, eased the pressure on resources for running the scheme. It needs to 74 be stressed that the Kerala Government has been able to recover part of the operating costs of the scheme from its consumers by charging higher prices from non-poor consumers. George (1985) noted substantial positive nutritional impact owing to a well-managed and targetted PDS in both Kerala and Gujarat. In ihe recent past in Gujarat, PDS coverage is reported to have been further restricted tovards the bottom two deciles f'or whom the ration price was further reduced. In Tamil Nadu, of the 11.8 million ration cards in operation, 8.6 million belonged to poor families with less than Rs. 500/- monthly income (Annadurai). Thuis, Kerala, Gujarat and Tamil Nadu represent relatively successful cases of eifective targetting. By contrast, in Andhra Pradesh, although the coverage was extended to the rural areas, nearly all rural households were covered. Gradually the scheme's coverage was tightened, yet nearly the bottom 70% of the population is covered under the 46 scheme now, so that the critically poor are able to get a small share of :ation quota as serious supply bottlenecks have emerged. An analytical study base:L on a dual market in rice showed that the policy improved the welfare levels of the poor (Radhakrishna and Indraknmt). However, the same study also pointed out thai if the coverage had been restricted to the bottom 40% of the population, the welfal:e gain to the poor would have sutbstantially increased, and the supply situjation wotld have become more manageable, ensuring the scheme's long run sustainability in a cost-effective manner. 4.44 The experience of the above states suggests that it is possible to re!;trict: access of PDS to the poor; that it is probably easier to administer the sch(.me if the coverage is restricted when the scheme is introduced, rather than Later that innovative methods for self-targetting do exist; that restricting access cE the 46It is reported that the government is now attempting to restrict the access to the population l:elow the poverty line. 75 scheme to the poor is absolutely essential since costs are directly related to coverage, as well as the ration quota and the ration price; and that a well-managed PDS does result in im,provement in the nutritional status of vulnerable households. The available knowledge of PDS in states limiting access to the poor could be usefully collated so as to help reorient PDS to serve the poor in the other states, especially in the rural areas in the eastern and central states. PDS would then become an important national indirect nutrition support program. Unless such restructuring of PDS towards the poor is attempted, which seems feasible with existing resources and foodgrains, the present food subsidy cost would continue to be largely unproductive. V. Conclusions and ImDlications for Policy 5.1 A clear empirical data base on various facets of malnutrition is essential for the evolution of a sound nutrition policy in India. Such knowledge is also needed for the Bank and other donors for supporting future nutrition interventions. This paper contains a detailed and disaggregated analysis of the existing data base on malnutrition in India and thus provides a firmer empirical basis for the evolution of nutrition strategics and policies. It also offers a general assessement of policy responses - of both Central and state governments - which is further supplemented by a detailed assessment of the effectiveness of major nutrition interventions tried in different states. This section summarises the main findings, and concludes with a few implications for nutrition policy planning in India. 5.2 There is substantial empirical evidence to suggest wide variations 76 in the intensity of malnutrition across regions, age and social groups and by gender. In terms of average levels of nutrient intake, severe malnutrition is prevalent in Bihar, Uttar Pradesh, Madhya Pradesh, Kerala, West Bengal, Orissa and Ra-a,than. The nutrient intake of the ultra-poor seems to be improving in Gujarat, Kerala, Maharashtra, Tamil Nadu and West Bengal. However, the levels of nutrient :ntak,es 0,,f scheduled castes and tribes are invariably lower than the average for all groups, and are way below the recommended levels in all states and locations - rural as, well as urban. The incidence of malnutrition is particularly severe amiong the t:vibals. A much higher percentage of children in the age group 0-3 years stuffer from severe malnutrition than others, in all locations, both developed and undeveloped. I'eneral economic development, and favourable seasonal changes and better cro0p conditicals seem to protect c.hildren fromi malnutrition to a far lesser extent than adults. While evidence on gender discrimination in food intake is not conclusive, regional p;..tterr.s do suggest the prevalence of discrimination in north India, especially againi:t very young female children. 5,3 Direct expenditure on nutrition support as a per cent o:f GDP remained modest at 0.16% during 1986/87. Even as a percent of Central and stat:e plan outlays, nutrition claimeBd only 1.09% of resources, the states sharing a li.gher burden than thLe Centre. An improved allocation of resources by the Centre fcc ICDS during the Serenth Plan unidoubtedly constitutes the most significant: developn r!t in the recent past, but the overall resources committed to nutrition still ::emaini modest. 5,4 The distribution of even the modest resources allocate.l for nutrltion has not at all been sensitive to the observed regional variations in the. 77 incidence of malnutrition in India. Regrettably direct nutrition expenditures per child were pitifully low precisely in the very states that account for a larger share of the nation's severely malnourished children. There appears to be considerable scope for improving the productivity of even these weagre expbysditi rs by' (-t such expenditures to the critically malnourished age and social groups. Because nutrition is a state subject, the states have the major responsibility, the Centre playing only a subsidiary role. With the launching of ICDS, the Centre's role has expanded, but regional asymmetry persists nevertheless. While the Centre's financial support for initial overheads, staff and training is enabling the expansion of ICDS projects rapidly, the duration of feeding in the operationalised projects which has a direct bearing on improving the nutritional status of children and women, and for which resources have to come from the states, is unsatisfactory especially in the states with the greatest need such as Bihar. Interestingly, the mismatch between regional allocations and program effectiveness persisted not only in direct nutrition support programs but. also in the public distribution system (PDS) - India's largest indirect nutrition support program. This mismatch in resource allocation found its reflection in physical performance, to the extent children fed under various feeding programs did not correspond with the regional location of number of severely malnourished children. Nor did PDS supplies match with th, distribution of population under the poverty line. On the whole, the broad conclulsion is that the existing resources and programs for nutrition support are poorly matched with what we know about the distribution of need and, therefore, there is an) urgent need to strike a proper balance between needs, potential demand ard available resources. 5.5 Presumably disappointed by the inadequate support from the Centre, 78 some of the states have allocated substantial resources to school. feeding schemes. But even among the states that have allocated sizable resource for nutrition, ithe outreach of programs was extended to age/social groups not necessarily identifLed as acutely malnourished, while excluding or poorly touching some of the critically malnourished age groups. In other words, in some states current expenditur.es might be adequate for major impact on the problem if they were allocated exclusi%7ely for the malnouriLshed children and women. In other states, the resources allocated to the nutrition sector as a whole are so inadequate that there is every poss.i.ility of shortfalls occurring in the Seventh Plan outlays. 5.6 The two major nutrition support programs have had only Limited impact in combating malnutrition, despite good intentions. On balance, the rBelised results appeared to be far below the programs' potential. Various Eactors contributed to the poor performance of ICDS. Deficiencies have been nctBd .in practically every component of the program. The program as it is now in opiration is tilted in favour of older children at the expense of under-threes, amor.g whcom malnutrition is greatest. The nutritional impact varied a great deal across projects, with tribal andl rural locations faring much worse than others. Pei:vasive weaknesses also persistetd in recruitment, training and supervision of worlcers. Similarly, the full potential of PDS as an important Indirect nutrition support program is yet to be realised despite allocation of substantial resources -c the scheme. The experience of some states does underscore the potential of a well-managed and targetted PDS for improving the nutritional status of vulnerable households. For realising this potential cost-effectively, self-targetting approaches (such as supplying inferior grains/tapioca, etc.) must be i:xied, Effective PDS coverage can also contribute to improving the outreach of foott-for- 79 works (NREP, EGS) programs. As it is in operation in the country as a whole today, PDS is far from becoming an active indirect nutrition support program. However, with appropriate restructuring, it seems feasible to render PDS as an effective nutrition support program with existing resources and foodgrains. 5.7 What basic issues arise from the above findings? First, substantial progress has been made towards evolution of a nutrition policy in India, as embodied in programs like ICDS which has been drawn into the fold of the 20-point program. There is continuing emphasis on supplementary rehabilitative feeding, linked to health and other services. Notwithstanding these policy gains, there are still a few planning issues that need to be squarely addressed. These are: determining the quantity of needed resources to be raised; how these are to be used; and what sequencing of nutrition investments (priorities) is to be followed. An overall review of adequacy of current resources and its allocation across competing needs and regions may have to be combined with further analytical and empirical research on how to make the existing programs more cost-effective. In other words, there is an urgent need for integrating macro planning (resource allocation) issues with micro-level (program design) issues with the overall. objective of ensuring adequacy and efficient use of resources allocated to nutrition. Such an exercise is necessary because the experience reviewed in this paper suggests that in the past governments have often put into effect programs without providing for necessary organisational inputs which resulted in considerable waste of resources. The need for relative weighing of costs of alternative programs in relation to benefits, the organisational, administrative and training inputs needed for ea^h of the components of the programs, etc. - the nuts and bolts issues - can hardly be overemphasized. 80 5.8 Given the social-class and regional distribution of malnutrition in India, and given paucity of resources, another issue that arises is the need for adoption of a regional approach to nutrition planning, combining it with overall development policy framework. This is because the geographical areas where a sizable proportion of the country's malnourished social groups are located also hcppen to be extremely backward and slow-growing regions. For these social groups, nutr:ition planning and intervention alone may not be adequate; these must be intertwined with programs of general development, primary health care, and infrastructural impro'rement of these regions. The twro major interventions - the ICDS and PDS - though bas.cally good responses, do not yet add upto such an overall nutrition planming poll y. 5.9 A third issue, which follows from the second, is that sich a regional approach would also enable flexibility in program design according to the needs of different regions. At present, a common package of services i being uniformly sulpplied by ICDS throughout the country. This approach is based on the assumption that the demands of the populations of dissimilar states; like Kerala and Bihar are the same. Kerala has a much higher level of female literacy and health infrastructure, and may not need health and nutrition education to the same e:ttent as the population of Bihar or Uttar Pradesh. Presumably ICDS program dcB3S not inhibit such flexibility,; yet in practice the program is not being moulded al: the project level to accommodate such regional differences. As a conseiquence, I(DS has not yet become a truly demand-based program. A pre-requisite for such a demand -based nutrition planning is decentralisation in decision-making, and active involcment of the beneficiaries in program design, in the day-to-day operation of the prEl,ram, and also in the assessement of its impact. 81 5.10 Finally, the prerequisites for improving the nutrition situation in the country appear to be (a) an appropriate institutional back-up, and (b) measures to improve the data base on nutrition. Most of the problems identified in this paper (e.g. overall resource inadequacy and inefficient use of existing resources, program designing issues and cost effectiveness, the asymmetry between needs and resources, etc.) have emerged essentially because of a lack of a rallying point for advancing nutrition interests,, and for periodically reviewing existing policies and programs. The building up of such an institutional front is inevitably a slow process, for there are no immediate tangible gains either to society, policy makers or politicians.47 It is not, therefore, surprising that except in a few states, nutrition has not received the enthusiastic support from the politicians that other programs received. As such, it is all the more necessary for building such an institutional capacity especially at the local (village community and district) levels to mobilise and manage resources for nutrition efficiently. The government should take the lead in this process, and aupplement the efforts of voluntary institutions. Strengthening the data base on the nutrition situation (which is now eight years' old and poor) is a key step for not only designing interventions appropriately moulded to accommodate for regional variations, but also for keeping track of the quality of implementation of programs, and for assessing other program impacts. 47By contrast, witness the quick building up of pressure groups and institutional lobbying for advancing the farming interests in recent times. REFERENCES Anderson, Mary "Reaching the Unreached: Why is it imperative?" Ann (1987) USAID. Mimeo. Annadurai, C "The Public Distribution System in Tamil Nadu: An Interim Assessment" Bulletin Madras Development Seminar series, XVIII (1). Berg, A. "Nutrition Planning Is Alive and Well, Thank You", (1987) Food Policy, Vol. 12 (4). Malnutrition: What can be done?, Bathmore, The Johns (1987) Hopkins University Press. Dandekar, V. M."On Measurement of Undernutrition", Economic and (1982) Political Weekly, Vol. 17 (6) February 6. pp.203-212. de Janvry, Agricultural Price Policy and Income Distribution K. Subbarao(1986) in India, Oxford University Press, Delhi. Evenson, R. E. "Food Consumption, Nutrient Intake and Agricultural (1986) Production in India", USAID (Delhi) Occassional Paper No. 3 George, P. S. "Some Aspects of Procurement and Distribution of (1985) Foodgrains in India," IFFRI, Washington, Working Paper on Food Subsidies no. 1. Gopalan, C. (ed) Combating Undernutrition: Basic Issues and Practical (1988) Approaches, Nutrition Foundation of India, Special Publication Series 3. =____=___=__ Nutrition, Problems and Programmes in South Asia, (1987) WHO. Govt. of Punjab Report on ICDS in Punjab. (undated) Gupta, Saroj "Inter-State Variations in Food Consumption (1987) Nutritional Adequacy and Levels of Poverty", USAID, Delhi, Occassional Paper No. 4. Harriss, B (1986) "The Intra-family Distribution of Hunger in South Mimeo. Asia", London School of Hygiene and Tropical Medicine. Indian Express "Anganvadi Centres: Cesspools of Corruption", July 8, (1988) Delhi. Kakwani, Nanak "On Measuring Undernutrition", WIDER Working Paper 8, (1986 a) Helsinki. Kumar Shubh K. Impact of Subsidised Rice on Food Consumption and (1987) Nutrition in Kerala IFFRI, Research No. 5, Washington D. C. __________ "Is Sex Bias Significant? ', EIDER 'Torki>-g PEaper 9, (1986 b) Helsinki. Khanna, R. and "Review of Research in Supplementary Nutrition in Ohri, P. (1986) India", Paper presented at a Conference on ICDS, NIPCCD, Delhi, Lipton, M. (1983) "Poverty, Undernutrition and Hunger", World Bank Staf.f Working Paper No. 597. _______ "Landi Assets and Rural Poverty", World Bank Staff Working Paper no. 744. National Institute Abstracts of 200 Papers on ICDS. Mimeo. for Public Coopera- tion in Chi ld Development., (undated) "Impact of Therapeutic Nutrition on Rehabilitation of (1987) the Severely Malnourished Children in ICDS", Mimeo. _"Status of Nutrition Component of IODS, Mimeo. (1988) National Nutrition Monitoring Bureau, Annual Reports 1979 to 1986. Nutrition Foundation "]:ntegrated Child Development Services. A Study of of India (1988) Some Aspects of the System" Scientific Report No. 7. Panikar, P. G. K. Health Status of Kerala Centre for Devel-opment and Soman, C. R. Studies, Trivandrum. (1984) Prasad, P. H. "The Pattern of Poverty in Bihar", I.L.C). Population et. al (1986) and labour Policies Programme Working Paper No,. 152. Radhakrishna, R "Effects of Rice Market Interventions Policies in Indrekant, S India: The Case of Andhra Pradesh", Centre for (1987) Economic and Social Studies, Hyderabad. Rao, Kamala S.Jaya "Urban Nutrition in India, I & II", in Copalan, (1988) (1985) op. cit. Rao, C.H.H. "Changes in Rural Poverty in India", Dr. Rajendra (1985) Pradad Memorial Lecture,, Akola. Rao, C.H.H., Unstable Agriculture and Droughts: Implications for Ray, S.K. and Policy, Delhi: Vikas. eSubbarao, K. (1988) Rao, N. Prahlad "Evaluation of Nutrition Programs" Nutrition News, (1983) Volume 4 (4), July. Rao, N. Prahlad, "Health and Nutritional Status of Rural Communities et. al. (1987) in Jammu and Kashmir", National Institute of Nutrition, Mimeo. Rao, N. P. and "Nutrition Profile in India over a Decade", Mimeo, Sastry, J. G. NIN, (1985) Reddy, J and Birth and Death Study in ICDS Blocks in Andhra Reddy B (1986) Pradesh (Paper presented at NIPPCD Seminar on ICDS). Scrimshaw, N. S. "Some Observations on ICDS of India", M.I.T., Mimeo. Sen, A. K. "Family and Food: Sex Bias in Poverty" pp. 346-368 in (1984) Resources, Values and Develo2ment, Blackwell, London. Sen, A. K. (1987) "Hunger and Entitlements" WIDER, Helsinki. Sen, A. K. and "Malnutrition of Rural children and the Sex Bias" Sengupta, (1983) Economic & Political Weekly, Annual Number, May. Sharma, Adarsh Social ComDonents of ICDS, NIPPCD. (1987) Sharma, Urmil "An Evaluative Study of Nutritional Status in ICDS in (undated) India". Mimeo Sinha,S.N. et.al "Morbidity Pattern in 0-6 year children in Aligarh" (undated) abstracted in NIPPCD, (undated Mimeo). Srikantia, S. G. "The National Nutrition Monitoring Bureau", NFI (1988) Bulletin, Vol. 9 (2) April. Srinivasan, T. N. "Malnutrition: Some Measurement and Policy Issues" (1981) World Bank Staff Working Paper 373. Subbarao, K. "Interventions to Fill Nutrition Gaps at the (1987) Household-level: A Review of India's Experience", Paper presented at a Workshop on "Poverty in India: Research and Policy" at Oxford University, Queen Elizabeth House. Sukhatme (1982) "Measurement of Undernourishment" Economic and Political Weekly, December 11. Tandon, B. N. "A Coordinated Approach to Children's Health in (1983) India: Progress Report After Five Years", Lancet. _______ "Management of Severely Malnourished Children by (1984) Village Workers in ICDS in India", Journal of Troipical Pediatrics, October. _ "The Impact of ICDS in India", Mimeo. (undated) UNICEF (1983) ICDSz An Assessment Venugopal, K. R. "The Implication of a Public Distribution System for (1987) Poverty" Centre For Policy Research, New Delhi. Walker T. et.al "Nutrition and Health", ICRISAT Mimeo. (1987) World Food Program "Interim Evaluation of iFP Assisted ICDS Projects". (1987) APPENDIX A For the years 1978-79, 1979-80 and 1980-81, district-level data on nutrient intake have been released by NNMB. Of these three years, it may be noted that 1978-79 was a very good year (i. e. actual food production was considerably above the trend line), whereas 1979-80 was a drought year. That leaves only 1980-81 as a normal year. We have used district-wise data for this year. Ignoring Karnataka districts whose NNMB do not match with other evidence, we have infromation on 22 districts spread over 8 states. Our dependent variable is calorie intake of the ultra-poor, and the average calorie intake. We have tried a number independent variables including per capita district food production, level of development of the district as reflected by CMIE index of development, nominal wage rate, and wage rate in terms of rice price, percentage of scheduled caste and tribal population, literacy rate, etc. However, only two variables were statistically significant. These are: per capita food production, and the CMIE index of development. We are reporting the estimated equations below: Dependent variable: Calorie Intake of Ultra Poor Average Constant 1854.3 2073.2 CMIE 1.56' 1.21n5 (2.45) (1.28) PCGR 1.32* 1.43 (2.67) (1.95) R2 0.27 0.10 (Figures in brackets are t values). It is interesting that while the average calorie intake is influenced by per capita food consumption, the ultra-poor's intake is also significantly influenced by the level of development of the district of which irrigation is an important component. Recognising that the demand for labour may be expected to be high in the developed districts, this finding seems to make sense, given that the ultra-poor have only their labor to sell. DISTRIBUTORS OF WORLD BANK PUBLICATIONS ARGENTINA FINLAND KOREA, REPUBLIC OF SPAIN Carlos Hirsch, SRL Akateeminen Kirjakauppa Pan Korea Book Corporation Mundi-l'rensa Libros. S.A. Galeria Guernes P.O. 3Sx 125 P. O. Box 101, Kwangwhamun Castello 37 Florida 165. 4th Floor-Ofc. 453/465 SF-00101 Seoul 2B001 Madnd 1333 Buenos Aires Helsijki 10 KUWAIT SlRi LANKA AND THE MALDIVES AUSTRALIA, PAPUA NEWV GUINEA, FIJI, FRANCE MEMRB Lake House Bookshop SOLOMON ISLANDS, VANUATU, AND Woricl Bank Pubhcations P.O. Box 5465 P.O Box 244 WESTERN SAMOA 66, aventue d'lena MALAYSIA 100, Sir Chittampalam A. Gardiner %law,, tha Info-Line 75116 Paris University of Malaya Cooperative Bookshop, Colombo 2 Overseas Document Delivery GERMIANY, FEDERAL REPUBLIC OF Limited SWEDEN Box 506, GPO UNO-Verlag P. 0. Box 1127, Jalan Pantai Baru Fritzes Fackboksforetaj!.?t Svdney, NSW 2001 Poppelsdorfer Allee 55 Kuala Lumpur Regenngsgatan 12. Bo, 16:56 AUSTRIA D-530) Bonn I MEXICO S-103 27 Stockhotm Gerold and Co. GREECE INFOTEC SWITZERLAND A-1011 Wien KEMi Apartado Postal 22-86 Libraine Pavot Graben 31 24, Ipoodamou Street Col. PE/A Pobre 6, rue Grenus BAHRAIN Athens-11t,35 14060 Tlalpan, Mexico D.F. Case postal 381 MEMRB Information Services GUATEMALA MOROCCO Cli 1211 Geneva 11 P.O. Box 2750 Libreras Piedra Santa Societe d'Etudes Marketing Marocaine TANZANIA NManama Town 317 Centro Cultural Piedra Santa 2 Rue Moliere, Bd. d'Anfa OCford Universitv Pres' BANGLADESH 11 calle 6-50 zona I Casablanca P. O. Box 5299 Micro Industries Development Assistance Guatemala City NETHERLANDS Dar es Salaam Society (MIDAS) HONG KONG, MACAO InOr-Publikaties b.v. fnIAlLAND House 56, Road 7A Asia 2300 Ltd. P.O. Box 14 Central Department Stv e Dhanmondi RIArea 6 Fl., 146 Prince Edward Road, W, 7240 BA Lochem 306 Silom Road Dhaka 1209 Kowloon NEW ZEALAND Bangkok BELGIUM Hong Kong Hills Library and Information Service TRINIDAD & TOBAG I), ANnTiGLIA, Publications des Nations Unies HUNGARY Private Bag BARBUDA, BARBADCS., DONEINICA, Av. du Roi 202 Kultura New Market GRENADA, GUYANA, JAMAICA, 1060 Brussels P.O. Box 139 Auckland MONTSERRAT, ST. KITTS A.ND) NEVR., BRAZIL 1389 Budapest 62 NIGERIA ST. LUCIA, ST. VINCi N1 & Publicacoes Tecnicas Intemacionais Ltda. INDIA. University Press Limited GRENADINES Rua Peixoto Gomide, 209 Alied Publishers Private Ltd. Three Crowns Building Jericho Systematics Studies U. 01409 Sao Paulo. SP 751 Mount Road Private Mail Bag 5095 55 Eastem Main Road CANADA Madrami - 600 002 Ibadan Curepe Le Diffuseur Bran-h offices: NORWAY Trinidad. West Indies C.P. 85, 1501 Ampere Street 15 j. Heredia Marg Narvesen Information Center TURKEY Boucherville, Quebec Ballard Estate Bertrand Narvesens vei 2 Haset Kitapevi, A S. J4B 5E6 Bom bay - 400 038 P.O. Box 6125 Davutpasa Caddesi CHINA 13/14 Asaf All Road N-0602 Oslo 6 Sergekale Sokak 113 China Financial & Economic Publishing New Delhi - 110 002 OMAN Topkapi House MEMRB Information Services Istinbul 8. Da Fo Si Dong Jie 17 Ciittaranjan Avenue P O. Box 1613, Seeb Airport UGANDA Beijing Calcuitta - 700 072 Muscat Uganda Bookshop COLOMBIA Jayacleva Hostel Building PAKISTAN P.O. Box 7145 Enlace Ltda. 5th Main Road Gandhinagar Mirza Book Agency Kampala Carera 6 No. 51-21 Bangalore -560009 65, Shahrah-e-Quaid-t-Azam UNITED ARAB EMIRATE1 Bogota D.E. 3-5-1129 Kachiguda Cross Road P.O. Box No. 729 IEMRB Gulf Co, Apartado Aereo 4430 HydE!rabad - 500 027 Lahore 3 P 0. Box 6097 Cai., Valle PTarthana Flats, 2nd Floor PERU Sha0ah COSTA RICA Near Thakore Baug, Navrangpura Editorial Desarrollo SA UNITED KJNGDOM Libreria Trejos Ahmedabad - 380009 Apartado 3824 Microinfo Ltd. CaGle 11-13 Patiala House irima P. (. Box 3 Av. Fernandez Guell 16-A Ashok Marg PHILIPPINES Alton, Hampshire GU34 IP; San Jose Lucdkuow - 226 001 National Book Store England COTE D'IVOIRE INDONESIA 701 Rizal Avenue URUGUAY Centre d'Edition et de Diffus:ion Africaines Pt. Indira Limited Metro Manila Instituto Nacional del Li: rc (CEDA) ll. Sam Ratulangi 37 l?OLAND San lose 1116 04 B.P. 541 Jakarta Pusat ORPAiN Monitevideo Abidjan 04 Plateau P.O. Box 181 Palac Kultury i Nauki VENiEZUELA CYPRUS IRELAND 00-001 Warszawa libreria del Este MEMRB Information Services TDC Publishers PORTUGAL Aptdo. 60.337 P.O. Box 2098 12 Nort.h Frederick Street Livraria Portugal Caracas 1063-A Nicosia Dublin I Rua Do Carmo 70-74 YUGOSLAVIA DENMARK ISRAEL 1200 Lisbon Jugcslovenska Kniiga SamtundsLitteratur The Jemisalem Post SAUDI ARABIA, QATAR YU-11000 Belgrade Trg Ripustike Rosenoerns Alle 11 The Jermsalem Post Building janr Book Store ZIMBABWE DK-1970 Fredenksberg C. P.O. Box 81 P. 0. Box 3196 Longman Zimbabwe DOMINICAN REPUBLIC Romenri Jeeusalem 91000 Riyadh 11471 P.O box ST 125, Southeron Editora Taller, C. por A. ITALY SINGAPORE, TAIWAN, BURMA, BRUNEI Harare Restauracuon Licosa Commissionaria Sansoni SPA Information Publications Apdo. postal 2190 Via Lamarmora 45 Private, Ltd. Santo Domingo Casella Postale 552 02Pi6 st Fl., Pei-Fu Industrial EGYPT, ARAB REPUBLIC OF 50121 Florence Bldg., 24 New Industrial Road Al Ahram JAPAN Singapore Al Galaa Street Eastern Book Service SOUTH AFRICA Cairo 37-3, Hango 3-Chome. Bunkyo-ku 113 (Oford University Press Southern Africa The Middle East Observer Tokyo P.O. Box 1141 8 Chawarbi Street Cairo JORDAN Cape Town 8000 Jordan Center for Marketing Research EL SALVADOR P.O. Box. 3143 FusadcesJalAm n Edifico La Centro Americana fo. Piso Jabal Amman Apartado Postal 01-278 KENYA San Salvador 011I Africa Boiok Service (E.A.) Ltd. P. 0. BOX 45245 Nairobi Prices and credit terms vary from country to country. Consult your local distributor before placing an order. RECENT WORLD BANK DISCUSSION PAPERS (continued) No. 31. Small Farmers in South Asia: Their Characteristics, Productivity, and Efficiency. Inderjit 1ingh No. 32. Tenancy in South Asia. Inderjit Singh No. 33. Land and Labor in South Asia. Inderjit Singh No. 34. The World Bank's Lending for Adjustment: An Interim Report. Peter Nicholas No. 35. Global Trends in Real Exchange Rates. Adrian Wood No. 36. Income Distribution and Economic Development in Malawi: Some Historical Perspectives. Frederic L. Pryor No. 37. Income Distribution and Economic Development in Madagascar: Some Historical Perspectives. Frederic L. Pryor No. 38. Quality Controls of Traded Commodities and Services in Developing Countries. Simon Rottenberg an ruce an e No. 39. Livestock Production in North Africa and the Middle East: Problems and Perspectives. John C. Glenn LAlso available in French (39F)J No. 40. Nongovernmental Organizations and Local Development. Michael M. Cernea No. 41. Patterns of Development: 1950 to 1983. Moises Syrquin and Hollis Chenery No. 42. Voluntary Debt-Reduction Operations: Bolivia, Mexico, and Beyond... Ruben Lamdany No. 44. Adjustment Programs and Social Welfare. Elaine Zuckerman No. 45. Primary School Teachers' Salaries in Sub-Saharan Africa. Manuel Zymelman and Joseph DeStefano No. 46. Education and Its Relation to Economic Growth, Poverty, and Income Distribution: Past Evidence and Further Analysis. Jandhyala B.G. Tilak No. 47. International Macroeconomic Adjustment, 1987-1992. Robert E. King and Helena Tang No. 48. Contract Plans and Public Enterprise Performance. John Nellis [Also available in French (48F)] No. 49. Improving Nutrition in India: Policies and Programs and Their Impact. K. Subbarao The World Bank Headquarters European Office Tokyo Office 1818 H Street, N.W. 66, avenue d'1ena Kokusai Building Washington, D.C. 20433, U.S.A. 75116 Paris, France 1-1 Marunouchi 3-chome Telephone: (202) 477-1234 Telephone: (1) 40.69.30.00 Chiyoda-ku, Tokyo 100,Japan Facsimile: (202) 477-6391 Facsimile: (1) 47.20.19.66 Telephone: (3) 214-5001 Telex: WUI 64145 WORLDBANK Telex: 842-620628 Facsimile: (3) 214-3657 RCA 248423 WORLDBK Telex: 781-26838 Cable Address: INTBAFRAD WASHINGTONDC ISSN 0259-210X / ISBN 0-8213-1192-1 / DP0049