Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Toward success in combating malnutrition: An assessment of what works

F. James Levinson
Consultant, UN ACC Sub-committee on Nutrition

The series of case studies that follows emanates from a request by the Sub-committee on Nutrition of the United Nations Administrative Committee on Co-ordination (ACC/ SCN) to carry out an examination of what works in nutrition programmes in low-income countries. The idea was to look at successful nutrition initiatives, and at those that might have been successful under somewhat different sets of conditions. Such information, it was hoped, might assist potentially interested governments and assistance agencies in assessing whether expenditures on nutrition today have the potential of helping to meet this most important of basic human needs.

For this purpose a group of case studies are presented here covering successful, or nearly successful, activities, with a concluding section seeking to generalize from this experience. The concept of nutrition in this context is broadly defined to cover activities with a potentially direct effect on food intake and/or nutritional status. This encompasses a spectrum of activities from consumer food subsidies to community health programmes.

Selection of cases to study was-perhaps inevitably-a somewhat arbitrary process. There were, however, several basic guidelines used. The first was to include at least one activity in each of the following programme categories: food subsidies, food security, the incorporation of nutrition considerations in broader development pursuits, community health and nutrition, nutrition education, and food fortification.

A second criterion was one of size. Large is by no means necessarily beautiful in the area of nutrition; but, as there have been of late a number of outstanding publications concentrating on smaller, community-level activities (see, for example, refs. 1-3, and numerous issues of Les Carnets d'Enfance/Assignment Children), this paper seeks to complement those important documents by examining projects that cover larger physical areas and that raise somewhat different but equally important questions.

The second criterion led naturally to a third, a desire not to duplicate coverage of projects discussed in those publications. Accordingly, I have left out a number of activities that might otherwise well have been included, among them the Narangwal, Candelaria, and Companiganj projects, each with its important and particular sets of lessons.



In recent years there has been much discussion of food subsidies (subsidized food consumption) in low-income countries. In fact, there are few development issues that have elicited as much controversy.

Food subsidy systems using food rationing, food stamps, or government-controlled and subsidized open market prices currently operate in every country of South Asia plus Burma, Thailand, Indonesia, Egypt, Mali, Mauritius, Morocco, Tanzania, Zambia, and Colombia.

Critics of subsidized consumption systems argue that they are enormously expensive and utilize funds that otherwise could be invested. In Egypt in 1975 these subsidies represented 21 per cent of the annual government expenditure, and in Sri Lanka the figure has been as high as 25 per cent. Critics also argue that such programmes have high administrative opportunity costs and may adversely affect domestic agricultural production. The most active opponent of food subsidies has been the International Monetary Fund, which has applied considerable pressure on many governments to reduce or eliminate such subsidies. The IMF (with several other development banks and international assistance agencies) has taken the position that food subsidies are expensive "political" gestures that must be replaced by sound "economic" policies.

Supporters of subsidized food consumption urge that they are by no means wasteful political gestures but rather represent potentially powerful instruments to improve the distribution of income, to increase food consumption, and to reduce malnutrition among low-income groups. In Pakistan, for example, the subsidy resulted in a real income increment of 9 per cent to low-income recipients and provided up to 14 per cent of recommended caloric intakes for this group (4).

TABLE 1. Apparent Per Capita Daily Caloric Intake from Rice in Sri Lanka, by Income Group, 1969-1970


Intake by Income Group*

0 - 200 200-399 400-599 600-799 800-899 1,000+ All groups
Calories from free-ration rice 430 424 403 393 368 323 417
Calories from additional rice 416 515 591 593 617 597 500
Total caloric intake 2,064 2,272 2,436 2,512 2,540 2,641 2,264
Free-ration calories as percentage of total intake








* Groups classified by income in rupees per month.
Source: ref. 5.

The advocates argue that these food subsidies represent the cornerstone of overall basic needs programmes and that, if the meeting of basic human needs constitutes a primary goal of development, these subsidies constitute a crucial element of development strategy and should be maintained until the need for them has been eliminated. Finally, advocates suggest that food subsidies, as part of a basic needs package, may contribute to a reduction in the rate of population growth.

This section and the next seek to examine these food subsidy systems in Sri Lanka and Bangladesh. Both are poor relative to other Third World countries. Average per capita GNP is roughly US$150 in Sri Lanka and US$110 in Bangladesh.

The Programme

As in many countries, food rationing in Sri Lanka was initiated during the Second World War, when the supply of food imports was interrupted. Subsidized rationing then was continued as a means of protecting lower-income groups against inflation. The subsidy went through various phases with different political regimes. During most of the early 1970s, however, each citizen over one year of age was entitled to a free ration of two pounds of rice per week, with an equal amount provided at a subsidized rate. The food subsidy also included wheat flour, sugar, and a range of less important items.

As seen in table 1, the subsidy provided a major increment to the caloric value of the diet, with major benefits going to low-income groups. For those earning less than 400 rupees (SRs 400 = approximately US$27) per month (80 per cent of the population), this subsidy constituted 20 per cent of their total caloric intake and 14 to 15 per cent of their incomes (6, p. 33). The food subsidy, coupled with similar subsidies in education, health, and transportation, and a progressive tax structure, resulted in major redistribution of income. Between 1965 and 1973 the lowest docile of the population increased its percentage of the nation's total disposable income from 3.5 to 6.1 per cent, while the highest decile suffered a reduction from 31 to 22 per cent. (Paul Isenman, in a particularly provocative paper on Sri Lanka, makes the point that the severe political constraints to substantial redistribution of income or assets in most countries appear to operate less stringently when food, rather than money or land, is to be distributed [7] .)

Although nutritional deficiencies continued to exist in Sri Lanka, the public food distribution system over those years appears to have had the effect of avoiding the highly skewed distribution of food intake that characterizes many low-income countries. Data from the 1969/70 socio-economic survey indicate that only 25 per cent of the population consumed less than the 2,200-calorie-per-day standard set for that country, and that only 5 per cent consumed less than 1,900 calories (7). By comparison, in Bangladesh, a country with a similar per capita GNP and much higher levels of foreign assistance even on a per capita basis, fully 35 per cent of the population were found to consume less than 1,800 calories per day (8). Sri Lanka's comparatively balanced food distribution, in combination with a broad-based health care system. and a major anti-malaria campaign, resulted in a reduction in the infant mortality rate from an estimated 142 in 1946 (about the same as in Bangladesh today) to 46 in 1973 (about the same as present day Yugoslavia).

TABLE 2. Effect of Inflation on the Cost (in Rupees) of Commodities in the Sri Lanka Food-Stamp Programme from 1979 to 1981

Monthly Stamp Allowance September 1979 July 1981
Price per unit* Total cost Price per unit* Total cost
rice (8 lb) 1.50 12.00 3 00 24 00
flour (21/2 lb) 1.30 3.25 2.75 6.88
total   15.25 30.88  
Child, 8-12 yrs        
rice and flour   (as for adult)    
sugar (2 lb) 3.00 6.00 8.25 16.50
total   21.25 47.38  
Child, under 8 yrs        
rice and flour   (as for adult)    
sugar (1/2 lb) 3.00 1.50 8.25 4.12
Lakspray (1 lb) 8.50 8.50 15.00 15.00
total   25.25 50.00  
Kerosene, 7 bottles 10.58 9.26 17.65 15.44

* Rupees per pound for foodstuffs; rupees per gallon for kerosene (1 gallon = 8 bottles).
Source: ref. 10.

In 1978 an attempt was made to target the programme better and reduce costs by restricting ration system eligibility to households earning less than SRs 300 (US$20) per month (slightly higher for larger families). This reduced eligible beneficiaries by roughly 50 per cent. Then a year later the ration system was eliminated, along with most of the price subsidies, and replaced by a food-stamp programme with the same eligibility. This action was part of a larger package urged by the IMF, which included reduction or elimination of all subsidies, government decontrol in many sectors, and the encouragement of foreign private investment. This programme further reduced costs and in 1980 represented only 7 per cent of total government expenditures (as opposed to the high of 25 per cent under the ration system).

At the outset, the benefits of the food-stamp programme per beneficiary were roughly equivalent to those under the previous ration system. However, after two years of the programme, rapid inflation had reduced the purchasing power of the stamps to roughly half of their original worth, as indicated in table 2. In addition, the Government at this time eliminated the subsidies that had been placed over the years on imported rice, sugar, and wheat flour. As a result, and in spite of a vigorous employment-generating strategy, recent surveys carried out by the Food and Nutrition Policy Planning Division of the Ministry of Health indicate that the nutritional and health status of lower socio-economic groups has deteriorated in several districts of the country (10). Other recent data indicate that the country's impressive redistribution of income in the 1960s and early 1970s has been reversed, and those gains largely eliminated.

Primary criticism of the food-stamp programme within the Government, however, relates to its high cost, even at these reduced levels, and to abuses in programme operations. (As many as 30 per cent of present recipients may be above the economic cut-off point, while an estimated 10 per cent of eligible families have not been included.) As a result, the Government is currently considering replacement of the food-stamp programme with some form of welfare payment to needy families.

Lessons from the Sri Lanka Experience

Any assessment of the effect of the public food distribution programmes in Sri Lanka on nutritional well-being, at least from the late 1950s to the early 1970s, must recognize these programmes in the context of a broader development strategy that sought, with considerable commitment, to meet the basic needs of its low-income population. The strategy included land reform, progressive income taxes and labour and wage laws, subsidies on primary education, broad-based health care, and cheap transport.

In the context of such a strategy and level of commitment, the Sri Lanka food subsidies appeared to have considerable value in avoiding the high degree of malnutrition and malnutrition-related deaths from which most countries with average per capita incomes of less than US$200 suffer (7). In the absence of such broad-based commitment, food subsidies can have negligible or even negative effects on the well-being of the poor, as indicated in the following section on Bangladesh.

The major stumbling block from the first years of the programme through the present has been budgetary. This poses the question of whether the same substantial benefits -the avoidance of serious malnutrition and the sharp reduction in infant mortality-could have accrued at lower cost.

At the outside, reference should at least be made to the possibility posed by James Grant, Executive Director of UNICEF, and others that, if these reductions in malnutrition and mortality and the associated substantial fall in the birth rate were in part the result of a large-scale food subsidy programme, the cost-even 25 per cent of a national budget-is not necessarily too high.

But, even putting aside this possibility, there almost certainly were means by which a different set of decisions (at least in the early years before expectations were conditioned and political battle lines set) might have produced these results at a much reduced cost.

1. The early decision to provide some portion of food free meant that, by definition, programme costs would continually increase with inflation. If a political pledge were necessary, the more practical one might have been to a ration price at a fixed percentage below the open market.

2. For most of the life of the programme, high costs also resulted from its universal coverage, a luxury that no country can easily afford. The two means of reducing beneficiary size are: a. imposition of an administratively imposed means test, as eventually was done in Sri Lanka; b. utilization of lower-status food commodities, inherently less desirable to upper and middle-income groups, as opposed to rice, the world's highest-status food grain (see the following section on the use of this principle in Bangladesh).

3. The range of commodities offered at subsidized prices in Sri Lanka probably was unnecessarily large, given that food purchase savings just as easily could have been concentrated in one or two commodities. The Government further added to operational and administrative costs by offering larger numbers of food stamps to young children than to adults, theoretically to permit the purchase of more expensive condensed milk. In fact, as indicated by the results of the recent programme evaluation (10), milk purchase was found to be unrelated to the number or age of children in the family. More often, milk was purchased with food stamps, given the ease of re-selling it on the market. Sale proceeds then generally were used to purchase rice, which, strangely, was often less expensive on the open market than in the government-controlled co-operatives where food stamps were redeemed.

4. If programme costs still were found to be unacceptably high, operation of the programme could have been restricted to those months of the year when food needs are greater (the traditional hungry seasons). This would add a seasonal targeting dimension to income-group targeting.



As suggested in section I, one potentially valuable means of targeting beneficiaries and reducing costs in a food subsidy programme is through the utilization of lower-status food commodities. This was found to be the case in Pakistan, where the use of flour-milled whole wheat (nutritionally valuable but of low status) in that country's ration system had the effect of targeting offtakes toward the lowest-income groups, without a politically and administratively problematic means test.

The need for such targeting is particularly acute in Bangladesh, where the ration system has traditionally been oriented towards the politically pivotal urban middle class, rather than towards the basic needs of the poor. The Bangladesh ration system, which provides rice and wheat at roughly two-thirds of open-market prices, reaches less than 25 per cent of the country's population and less than 15 per cent of the rural population. Fully 70 per cent of food distributed through the ration system goes to urban consumers, yet only 27 per cent of the poorest and most malnourished urban families have access to the system.

The Sorghum Experiment

In 1978 a group of concerned individuals from the Bangladesh Government's Ministry of Food, the Institute of Nutrition and Food Science at the University of Dacca, and the United States Agency for International Development began meeting to consider whether it might be possible to bias the offtake to the country's ration system toward lower-income groups through changes in the commodities offered.

The foods found to exhibit desirable characteristics in Bangladesh-i.e., low price, a low or negative income elasticity of demand among middle- and upper-income groups but a relatively high elasticity of demand among the poor, and relatively high nutritional value-were the coarse grains (the millets and sorghum). These crops were also desirable because they are produced easily and inexpensively on land not suited for rice and wheat production. Irrigation is not necessary, as the coarse grains are drought-resistant. At present, almost half a million acres are under coarse-grain production in Bangladesh.

To examine the viability of this concept, a study was initiated in the urban ration system in Dacca and in the rural ration systems in Faridpur and Jessore districts in the southwestern part of the country. The coarse grain selected for the test, sorghum, was introduced into 20 ration shops scattered throughout Dacca city, and 103 rural ration shops in Faridpur and Jessore, principally from August to October 1978. A price of 1.25 take (Tk 1.25 = approximately US$0.08) per kilogram, about half the price of rice and five-eighths the price of wheat, was set by the Government. Sorghum was offered to consumers either in addition to their regular rice and/or wheat rations, or as an equal volume substitute for all or part of the regular ration.

As expected, given the predominance of middle-class consumers in the urban ration system, sorghum sales in Dacca were extremely low, rarely exceeding 5 per cent of the consumers in any given shop. However, in the rural ration system the results were quite different.

Rural consumers can be categorized by the type of ration card allocated to them. These categories, from the lowest, A, to the highest, D, are based on the land tax paid: A-card holders, generally landless, pay no land tax; other card holders pay Tk 3 to Tk 15 (US$0.20 to US$1.00) per acre per year. (These rates are themselves a commentary on the Government's inability, or unwillingness, to generate revenue from land-holders.) The group unrepresented in this breakdown is the so-called floating population, the migrants with no specific address and, therefore, in Bangladesh, ineligible to use the ration system at all. This floating population, inevitably the poorest of all, is estimated at 10 per cent of rural households.

Table 3 presents the results for Jessore District, from which the most complete data were collected. The table indicates not only the substantial percentage of ration shop users who purchased sorghum during these three months, but also the strong bias toward the lowest-income consumers, the A- and B card holders. Finally, it indicates that, with greater exposure, the use of sorghum by these low-income groups increases to a level of almost 70 per cent among the A-card holders, while sorghum consumption among the D-card holders falls to 1.5 per cent of those using the system.

TABLE 3. Households Using Ration System in Selected Areas of Jessore District and Percentage Purchasing Sorghum, by Economic Status,* August-October 1978

  Households Using Ration System Households Purchasing Sorghum (%)
August 2,641 1,801 944 443 47.9 49.6 18.8 7.5
September 3,248 2,437 616 305 65.1 52.6 13.8 1.3
October 3,049 2,348 570 327 68.9 54.6 16.0 1.5

* See text for explanation of economic-status categories by type of ration card held, A-D.

The table indicates clearly that the introduction of sorghum in the rural ration system had the potential to bias offtakes sharply toward the poor and increase the benefits this group could derive from the system. It seemed to be a natural means of addressing the astonishingly high rates of malnutrition in the country and, indirectly, the highly skewed distribution of income.

Demise of the Programme

In terms of nutritional criteria, the project certainly could be labelled successful. It reached the lowest-income families, among whom the most serious malnutrition is concentrated in Bangladesh. It provided these families with the opportunity to double their caloric intake at the same cost, or to maintain constant caloric levels at half the cost. The food used was nutritionally equivalent to the existing commodities and was consumed by children as well as adults. Even in the absence of an explicit assessment, it is almost certain that changes in nutritional intake would result from the increased food available and/or the increased savings, because low-income groups in Bangladesh spend up to 85 per cent of income increments on food. These increases in food intake among families consuming less than 80 per cent of their caloric needs and among children suffering from second- or third-degree malnutrition would, in turn, generally result in marked improvement in nutritional status.

In spite of this (and, ironically, in part because of it), the programme was short-lived. Because in 1978 sorghum production in Bangladesh was still relatively small and beyond the scope of government procurement programmes the decision was made to provide sorghum from the United States for the initial project, with the idea that increased domestic consumption and increased consciousness of sorghum as a food for people would, in turn, boost local production. This had, in fact, been the case with wheat, a commodity not traditionally produced in Bangladesh.

While the United States Department of Agriculture agreed reluctantly to provide initial shipments of sorghum, the traditionally powerful wheat lobby working within it quickly perceived the sorghum initiative as a threat to its growing wheat exports to Bangladesh. {Sorghum, never a major commodity under US Public Law 480, the Agricultural Trade Development and Assistance Act, has no such interest group to support its export abroad.}

The Bangladesh Government, meanwhile, expressed no reluctance to receive the sorghum on a purely grant basis. It did, however, make clear from the outset that sorghum was lowest on its list of priorities, and that it would prefer, at any point, to substitute additional quantities of wheat, or better yet, cotton or edible oil-all commodities disproportionately consumed by the urban middle class.

Soon, however, even this passive acceptance began to erode. Ration shop dealers, used to a substantial profit on government supplies, found that sorghum-the poor person's food-had no black-market value. Accordingly, over time, despite government directives, an increasingly large number of ration shop dealers simply neglected to pick up their sorghum allotments, thus creating movement and storage problems for the Government.

Despite empirical evidence that programme benefits were going almost exclusively to those in greatest need and were helping to meet the critical nutritional needs of these unfortunate families, the opposition of the USDA and the apathy of the Bangladesh Government proved decisive. Sorghum imports and distribution were quietly terminated in early 1980 without either a complaint or a word or explanation.


Contents - Previous - Next