Previous PageTable Of ContentsNext Page


Experiences in Targeting Nutrition

Targeted relief food aid


Ethiopia has a long history of food aid relief activities. The methods used to target these relief interventions have changed substantially over time. Targeting guidelines were most recently revised in 1993, outlining key principles for the implementation of relief activities. The principles, in part, give priority for food distribution to the areas where lives and livelihoods are most threatened. In such areas, relief food aid is targeted to the most needy, with no free distribution of food to the able-bodied population. The latter only have access to relief food through their participation in public works, which are an integral part of the country's social safety net.

Eligible recipients of free relief food must meet the following criteria. They must be: a) without relatives who are able and bound by custom to provide support and b )over 60 years of age, infirm, blind, crippled or mentally incapacitated, c) lactating or pregnant women, d) young children without any family support, e) required to attend young children and incapacitated adults constantly, or f) women who have given birth recently (who by custom are unable to appear in public). Beneficiaries are screened by a local council consisting of seniors and locally elected officials.

Recipients of employment-based relief food must be adults over 14 years of age, who are fit to work and who are members of the community in which the public works activities are being implemented. The policy states that all able-bodied members of a household may be permitted to work, but it recognizes the likely need to ration work in times of particular stress. Workdays may be rationed within households depending on household size, although no explicit policy is expressed in this regard.

Household-level targeting relies mainly on administrative and community-based targeting methods. Actual targeting criteria are similar in most locations, but their interpretation is complex. Households are generally judged, not according to any single indicator, but according to a subjective perception of need, which implicitly takes several factors into account. Although apparently open to anyone, food-for-work opportunities are frequently rationed among households according to criteria established by the communities themselves. In one community, for example, those with no assets to sell and no land were given priority. Within communities, there is widespread rejection of the idea of using low wages as a way of self-targeting food-for-work to only the most needy. In many areas, there is a tendency to spread relief food supplies over a large number of communities and households, without much regard for needs assessment. Biases in the allocation of work based on favouritism are also a common problem and reduce targeting effectiveness.

In Ethiopia, regional targeting is an important concern. One analysis suggests that food availability varies more across local administrative units, weredas, than within those weredas, implying that targeting at the wereda level would be more effective than household-level targeting. In theory, it should be possible to allocate relief food supplies in proportion to local estimates of need. In practice, decision-making is more difficult, since supplies are usually inadequate to meet estimated local needs fully. Furthermore, the severity of conditions across locations is not fully reflected in estimates of the number of households in need. Relief food allocations from central and regional authorities to the wereda level are often criticized for being too heavily based on estimates of local food production, without adequately considering broader and, admittedly, more complex socio-economic factors that determine household food access.

The use of geographic, rather than household-level, targeting methods would seem to have greater potential. At present, food aid is received by 42 percent of weredas, mainly those concentrated in two chronically food-deficit regions. However, there is some evidence that food aid is not being well targeted on a regional basis. Among the weredas receiving free food aid, 48 percent of households are considered to have inadequate calorie availability, roughly the same proportion as exists among households in weredas that do not receive free food aid. In addition, the proportion of food-deficit households among the weredas that receive food aid through food-for-work programmes is about the same as it is for weredas that do not. This system of wereda-level targeting resulted in slightly more than 45 percent coverage of all food-deficit households in the country. By contrast, if the relief food aid were targeted only to the weredas that are most in need, even a random allocation at the household level would imply that about 50 percent of all participating households would be food-deficit households. Within each wereda, the food-deficit conditions among households tend to be very similar, and this practically eliminates the need for household-level targeting, suggesting that geographic targeting would be more appropriate.


In response to serious problems associated with the purchase and distribution of maize as emergency relief, a new community-based system has been designed and is currently being implemented by the Government of Kenya. The problems with the previous system included: a) a lack of targeting, resulting in beneficiary households usually receiving very small monthly rations (as little as 2 kg); b) insufficient allocation of funds for transport, resulting in a significant portion of maize supplies being used to finance transport costs; and c) considerable misappropriation of emergency maize supplies. In addition, a parallel system operated by a number of international non-governmental organizations (NGOs) distributed international food aid during the 1997 drought and the 1998 floods. It was recognized that this system was not sustainable. Late in 1999, a new and unified system was designed with the participation of the Government of Kenya, donors, NGOs and UN agencies.

The main objectives of the new system, which pools government and non-government emergency food aid, are to maximize the humanitarian impact of the relief food aid and lower the administrative costs of the food aid programme. Specifically, the new system aims at improving geographic targeting of the most affected districts and divisions, and relieving the administrative burden on provincial and district governments by appointing a lead agency to coordinate and implement relief operations for each district. The system puts a strong emphasis on community participation and women's involvement in targeting the most vulnerable households in the community and in managing the distribution of emergency food aid. A District Steering Group (DSG), which is approved by the Office of the President and the World Food Programme (WFP), allocates the district's emergency food supplies to different divisions. Relief committees allocate food at the community level.

The system was first implemented in the district of Turkana in December 1999 in response to a severe food security crisis caused by drought, which has since worsened. World Vision was appointed as the lead agency. In June 2000, it became necessary to target 75 percent of the population in the district. The districts of Marsabit, Moyale and Mandera were targeted under the new system in March 2000. Sub-DSGs , made up of relief committee chairpersons, chiefs, counsellors and important elders, allocate food from the division at the community level in Mandera.

To date, experience of the new system has shown the importance of DSG's role as a coordinating body and the need to allocate emergency food aid on the basis of solid information about the food insecurity situation in different parts of the district. The success of the system depends on the sensitization of communities and local governments, and on the capacity of the lead agency to take into account local tribal and cultural factors. There have been instances in which local leaders or politicians have attempted to interfere with food distributions, but strong support from the government and DSG has made such attempts ineffectual. The targeting of eligible households may be a difficult task for relief committees in communities where there is a low percentage of households to be targeted, and where sensitization of the whole community is required. It has been relatively easy to resolve differences in the estimates of food aid required that were made at the national level and by the DSG. Feedback from beneficiaries about the functioning of the new system has generally been positive.

Targeted food stamp programmes


The Honduran Bono de Madre Jefe de Familia Programme relies on teachers to conduct simple means tests. At the beginning of the school year, primary school teachers in participating states are required to identify students from households headed by women and with incomes below a set level. In addition, at schools where the results of the annual nutrition survey show high malnutrition, all first graders are eligible for the programme. To obtain information on income levels, the teachers interview the students' mothers during home visits or at the school. Teachers spend about three days at the beginning of the school year identifying the beneficiaries of the programme. About 13 000 teachers are involved in the seven departments where the programme is operating. The programme benefits about 125 000 students per year in grades 1 to 3. Food stamps are also delivered through health centres in Honduras for the benefit of children under five years of age and for pregnant and lactating women, who are identified through nutritional surveillance.


In the Jamaican Food Stamp Programme, social workers from the Ministry of Labour, Welfare and Sport visit each candidate household and fill in a short form recording the candidate's address, household conditions and household income. The home visit allows the social worker to verify whether visible living conditions are in accordance with the level of income that the family reports. The characteristics of the dwelling or the ownership of durable goods are not formally used in the eligibility verification. During the early years of the programme, a single income threshold for eligibility was applied. Recently, a two-tiered threshold has been adopted, one tier for single-person households and another for larger households. About 150 full-time field workers, who work on a two-month cycle, administer the Jamaican programme. The first month of the cycle is devoted to identifying the beneficiaries, and the second to distributing the food stamps. The programme has about 300 000 beneficiaries. For half of these, eligibility is established through a means test, while the other half participate in the maternal and child health part of the programme, for which there is no means test to establish eligibility.


The Venezuelan Food Stamp Programmes are targeted to all children in selected primary schools. The choice of school is left to the education authorities at the state level, although schools in low-income urban or rural zones are supposed to be selected. Rather than providing set guidelines or standards for defining a needy school, officials are left to use their "expert opinions" regarding the level of need at each school. The percentage of schools that receive food stamps varies from 60 percent of those in the poorest third of the states (as defined by the national poverty map) to 40 percent of schools in the wealthiest third of the states. A committee made up of representatives from the Ministry of Education, parent-teacher associations, neighbourhood associations and the school certifies the lists of beneficiaries.

Targeted school feeding programmes


In Bangladesh's Food-for-Education Programme, eligible beneficiaries are selected from poor households, which are identified as being: a) landless, or owning less than 0.5 acres (0.2 hectares) of land; b) headed by someone who works as a day labourer; c) headed by a woman; or d) headed by someone engaged in low-paid work (such as fishers, potters and cobblers). Children enrolled in the programme must attend 85 percent of classes each month. If a household has only one primary school-age child, it is entitled to 15 kg of wheat per month. If a household has more than one primary school-age child, and sends them all to school, it is entitled to 30 kg of wheat per month. As a result of the programme, school attendance increased from 63 percent in 1993 to 77.6 percent in 1994, and drop-out rates declined from 18.5 to 10.9 percent. Leakage to the non-poor has been estimated at only 6.5 percent of total benefits.


The objective of Chile's School Feeding Programme is to provide social and food assistance to low-income children attending public or private schools. The main goals of the programme are to promote school attendance and improve academic performance by providing free meals to schoolchildren. Food assistance consists of one of the following combinations: only breakfast, breakfast and lunch, or lunch and afternoon snack. The meals are distributed approximately 180 days per year. Children in boarding schools receive all their meals through the programme. The dietary energy content of the meals provided depends on the vulnerability index of the school.

The programme is targeted at the school level, but individual data on children are used to rank schools according to a vulnerability index. Some children in participating schools do not take part in the programme, apparently to avoid the stigma associated with participation, while others arrive early to receive their meals (for the same reason).

The programme's targeting method has changed substantially over time. Until 1980, school teachers selected programme participants according to their subjective assessment of each child's socio-economic situation. However, that method was determined to be too subjective to ensure an equitable distribution of programme resources. Between 1980 and 1982, the CAS socio-economic index system was used to identify beneficiaries, but it was decided that the variables included in the CAS were not appropriate for classifying poor households in rural areas. In addition, there were also logistic problems in applying this instrument, making it difficult to obtain up-to-date CAS ratings for all of the families that requested programme participation.

In 1983, a school census was implemented to identify children in need of food. The census was conducted annually until 1985 and included information on weight-for-height, height-for-age, CAS index, teacher's assessment of need, the distance between the child's home and the school, location of the school, and the child's age. Given evidence that the food needs of first graders were representative of the overall need at the primary school level, only this group was included in data collection efforts. In 1985, a qualitative evaluation determined that the targeting strategy was not achieving its objectives fully, and a new approach was tested. Programme benefits were defined at the county level rather than the individual school level, and local county authorities were responsible for determining the number and types of meals to be offered in each school. Within each school, the teachers again selected the children who were to receive the benefit.

In 1986, another evaluation indicated that 13 percent of children in the upper-income quintiles were receiving benefits. To control this leakage, the targeting strategy was revised again. A weighted index was defined which included the following variables: mother's educational level, percentage of high-aged children in first grade, prevalence of height deficits, teachers' criteria of urgent need, and repetition rates. Information on these variables was collected from first graders to determine a school-level rating, which was then used as a basis for allocating the number of meals per school. A new statistical approach was applied in 1990 to identify and weigh different targeting indicators for predicting school needs. Different models were developed in 1993 for urban and rural primary schools, as well as a model for the allocation of meals to secondary schools.

The process of data collection is gradual and without pressure. Information is analysed in the second semester and is used for targeting of the following year's programme. By the end of the year, the process of ranking the schools on the vulnerability index is completed, in time for that information to be integrated easily into the budgeting process for the next year.

Results confirm that the programme definitively reaches the poor. More than 80 percent of its beneficiaries at the primary school level belong to the lowest income quintiles. Targeting costs are estimated to be remarkably low: US$36 772, or less than 0.05 percent of the total programme budget. The programme provides a strong incentive for parents to send their children to school: 58 percent of the children in rural areas completed primary education in 1990, compared with only 40 percent in 1986.


For many years, all schools were included in Costa Rica's School Lunch Programme. During that period, an evaluation found that 62 percent of the benefits went to children of the poorest 40 percent of households, which suggests a substantial leakage rate. More recently, three levels of benefits have been specified, and the amount of lunch subsidy per child that a school receives is determined by the size of the school and the poverty rating of the area that the school serves, in accordance with the Planning Ministry's poverty map. Combining information on school size with that from the poverty map establishes three priority levels for the distribution of lunches. Priority 1 schools are those in the poorest strata, according to the poverty map, and schools with only one teacher and fewer than 100 students. Priority 2 schools include those in the intermediate poverty strata as well as those in the higher strata, but with between only 100 to 500 students. Priority 3 schools include all schools in the most well-off strata and with more than 500 students. This is an example of clear geographic targeting, without additional individual targeting


The Nutrition Programme in Jamaica is meant to serve schools in poor areas. The selection of schools is based solely on the Ministry of Education's informal knowledge of which schools are located in poor areas. The programme involves the daily delivery of food products from central bakeries, so schools with access to good roads tend to be served better than those in more remote areas. Nevertheless, 72 percent of benefits go to the poorest 40 percent of the population.


The school feeding programme in Uruguay dates back to the early 1900s, when it was started in rural schools and later extended to include urban schools. In 1983, the administration of the food services passed to the Council of Primary Education, and the purpose of the programme was then defined as supporting nutritional improvement among schoolchildren. In September 1991, the goals set for 1995-2000 were expanding programme coverage and improving service delivery. The programme provides different food services, depending on the nutritional needs defined for different schoolchildren: lunch; breakfast and/or lunch, plus a snack; breakfast, lunch and dinner, plus a snack or a glass of milk. In 1994, the programme reached 37 365 children in 173 schools (31.6 percent of the total) in Montevideo, while the remaining 68.4 percent did not receive any food service. In 1995, the programme reached 128 661 children in rural schools (85 percent), mostly by providing lunch.

The programme has no clearly defined target population and its nutritional and educational impacts were not known until an evaluation was undertaken in 1996-1997. The evaluation set out to examine, among other things, the outcome of the programme's targeting schemes. As far as the programme's nutritional impact is concerned, the evaluation of schools in Montevideo concluded that: i) participating children did not generally improve their nutritional status over time, compared with non-participating children; ii) overall, participating children had a worse nutritional status than non-participating children; and iii) among participating children, the greater their dependency on the school feeding programme for daily food intake, the more likely that their nutritional status was to worsen. Since part of the evaluation is based on a cross-sectional analysis, those results can be interpreted to indicate that the overall targeting schemes were successful in reaching the schoolchildren at greatest risk of nutritional deficiency.

Programme targeting is done administratively at two levels: a) selection of schools to receive a food service; and b) selection of schoolchildren to participate in the meal or snack service of schools that offer such a service. The evaluation found that the median percentage of children with deficient height-for-age was 20 percent among schools with a food service, and 14 percent among schools without one. Among the primary schools in Montevideo, 31.6 percent offered a food service, reaching 18 percent of primary school students. When schools were stratified according to an index of school deficiency (Indice de Escuelas Carenciadas), there was a clear tendency for the percentage of schools with a food service (and the programme's student coverage) to increase with the school deficiency score: 76 percent of schools with a score greater than 75, accounting for 46 percent of the student body, versus 0 percent of schools with a score less than or equal to 40. Neighbourhoods where primary schools were located were also stratified according to a poverty index of deficiency in basic needs (Indice de Necesidades Basicas Insatisfechas) based on housing conditions and the education and occupation of heads of household. Some 55 percent of schools in neighbourhoods with a score greater than 40 offered a food service, covering 33 percent of the student body, compared with 7 percent of schools in neighbourhoods with a score less than or equal to 20, reaching 2 percent of the student body.

The intra-school selection of students who are eligible to participate in the food service is the responsibility of the school principal. No specific and operational criteria were established. The school principal's subjective assessment of children in need of complementary food determines actual participation. In some cases, the principal is assisted in the selection process by the inspector of food services and/or a social worker. Elements that are taken into account by school principals include: children who appear to be at nutritional risk, particularly those referred by health centres; the family situation of the child; and the child's school performance.

The results of the analysis showed that, among schools in low-income neighbourhoods, participation in the food service was perfectly random (E = 0.03), when taking nutritional status (height-for-age) as the criterion for programme participation. Among schools in better-off neighbourhoods, food service access was actually less likely to be directed to students with poor nutritional status (E = -0.27). When the family's NBI score is taken as the criterion for student access to the food service, the targeting results were somewhat better (E = 0.38 among schools in low-income neighbourhoods, and E = 0.14 among schools in better-off neighbourhoods). This suggests that the principals' selection of students for participation in the food service may be more influenced by consideration of the family's socio-economic conditions than of the child's nutritional risk.

The overall conclusion was that targeting of the school feeding programme involved a significant degree of leakage and undercoverage in terms of reaching primary schoolchildren at nutritional risk, probably owing more to the intra-school selection process than to the school selection process.

Targeted integrated nutrition programmes


In Bangladesh, an NGO called BRAC has been carrying out a targeted nutrition project in the Muktagacha thana (subdistrict) since 1993. The main objective of the project is to improve the health and nutritional status of children, adolescent girls and pregnant and lactating women in this area. The project entails nutritional surveillance, nutrition and health education, food supplementation for the faltering and severely malnourished, and the referral of the most severe and complicated cases to government health facilities. In addition to basic health and nutrition services, the project also provides adolescent education, along with skills development and credit for women's income-generation in a community participation format.

Supplementation is targeted to the severely malnourished and those whose growth is found, through surveillance activities, to be faltering. Specific targeting criteria were determined by the project staff's detailed review of other targeting methods employed in similar activities, such as the Tamil Nadu Integrated Nutrition Programme.

The selected targeting criteria are:

Although detailed data on targeting effectiveness are not available, the project reports minimal leakage of project benefits, along with some degree of undercoverage. The latter is caused by the project's inability to overcome a variety of social and personal barriers to participation, as well as unspecified lapses in project implementation. Coverage of the surveillance portion of the activity has increased over time, and 90 percent of the children in participating communities were under surveillance in 1995. The percentage that needed and received supplementation rose from 26 percent in 1993 to 46 percent in 1994, falling back to 32 percent in 1995. An important concern for the project is whether the exit criteria are appropriate. For example, underweight children who show a gain in weight following supplementation but remain below the standard weight are currently required to exit from the supplementation programme. The project is assessing alternative mechanisms to rehabilitate these children and ensure that their families are given special attention.

Targeted supplementary feeding programmes


The National Supplementary Feeding Programme (PNAC) in Chile is the longest running nutrition intervention in the country and has the broadest level of coverage. Food supplements are distributed through public clinics on a monthly basis as an integral part of the primary health care (PHC) system. The types and amounts of food items supplied vary according to the beneficiary's age and nutritional status. At present, 1.2 million children under six years of age and 200 000 pregnant women participate annually. This corresponds to approximately 80 percent of the national population of infants under two years and 70 percent of pre-schoolers and pregnant and nursing mothers. In 1996, about 30 000 tonnes of food products were distributed at a total cost of approximately US$60 million, or roughly 7 percent of the Ministry of Health's total budget.

The basic PNAC supplement is provided to beneficiaries who are nutritionally normal and show no evidence of protein-energy malnutrition (defined as weight gain greater than 75 percent of that expected for their age, based on World Health Organization [WHO] standards). The programme was enhanced in 1983 to provide extra food supplements to infants and children identified as being at risk because of poor growth and development, or through socio-economic indicators. Food amounts are substantially greater in the enhanced programme rations and include additional quantities of rice to benefit all household members, in order to prevent intra-household sharing of foods destined for infants.

Women. In 1989, nutritional surveillance began to include maternal anthropometry. Until 1987, eligibility for participation in PNAC was determined according to a standard weight gain during pregnancy developed by the University of Chile. The Rosso-Mardones standard, based on ideal weight-for-height and adjusted for gestation age and baseline weight, was introduced. Currently, about 80 percent of the total amount of food supplements received by pregnant women are assigned to those classified as being at risk according to these standards.

Children. From 1983 to 1994, children with poor growth were defined at programme entry by one of the following criteria: a) weight-for-height less than -1 SD of the WHO reference norm; or b) weight-for-height greater than -1 SD, but a weight gain of less than 75 percent of that expected for age in two consecutive health controls; or c) those under two years of age and weighing less than 50 percent of expected weight-for-age. The criteria for discharge required a weight gain that is over the norm in three consecutive health check-ups. In 1990, 17 percent of all children in the PNAC programme were classified as being at "biomedical risk", and received 52 percent of the total amount of food distributed. These children were followed more closely and had more frequent health check-ups.

A review of the programme in 1992 indicated that the nutritional benefits of the enhanced programme were marginal for most children and that there was no relationship between the length of participation in the enhanced programme and nutritional status on entry. In fact, on average, children who had a weight-for-height of less than -1 SD of the WHO reference norm on entry had similar permanence in the programme as those who had a weight-for-height greater than or equal to the median. Only 14 percent of the children entering the enhanced programme were even mildly wasted (weight-for-height less than -1 SD). The rest had normal or elevated weight-for-height. The weight gain criteria were also demonstrated to be extremely sensitive but of little specificity in determining who needed extra food. The review concluded that the enhanced programme tended to target children who were not really at risk of malnutrition. In addition, the criteria for discharge were determined to be inadequate, since a large proportion of the children kept on the programme gained no real benefit in terms of growth.

As a result of this review, targeting methods have been revised and currently follow international guidelines: malnourished children are defined as those with a weight-for-age that is less than -2 SD of the WHO reference norms for children under two years of age. For children over two years of age, a weight-for-height of less than -2 SD defines malnutrition. At-risk children are those defined as having a weight-for-age between -1 and -2 SD at under two years of age, while for those over two years of age, a weight-for-height between -1 and -2 SD defines being at risk. The programme has also modified the criteria used for discharge. Currently, at-risk children have a maximum permanence of six months per year. Malnourished children have to demonstrate, in three consecutive health evaluations, a weight-for-height that is above -1 SD (WHO reference norm) for discharge. These modifications, based on solid information and technical criteria, have improved targeting and cost-effectiveness.

Because the PNAC relies on information gathered during regular health check-ups, there is no additional administrative burden involved in targeting. Given the good results of targeting, it is clear that an efficient clinic-based surveillance system, which can keep track of the levels of benefit according to age and nutritional status, is vital to ensure rational use of limited resources. Implicit geographic targeting also occurs with the PNAC, since there are more public clinics in poor than upper-income areas. The PNAC is, in part, self-targeting since many upper-middle and upper-income families do not participate because of the long waiting time to obtain the supplement. Until 1994, participation in the two upper quintiles was relatively high, at 60 percent and 23 percent, respectively. Although income level is not a specific targeting criterion, evidence suggests that more stringent targeting requirements might still be appropriate.


The CEN-CENA nutrition centres provide: a) day care with feeding, nutrition education, growth monitoring and early childhood education; b) a take-home milk programme; and c) a take-home family food basket. Targeting is based on geographic location, expert referrals, the employment status of mothers and psycho-social risk scores derived from the use of a simple form. For all programmes, children are granted entry according to the following priorities: a) those who are malnourished, maltreated or abandoned, those referred by an institution or those with risk scores of at least 40; b) those whose mothers work outside the home and who have risk scores of 50 or more; c) those whose mothers do not work outside the home and have risk scores of 60 or more. Day care centres also require that children live within a radius of 1 km. Malnourished children are identified through local health centres. Those admitted as a result of a risk assessment are usually evaluated at the request of family members. Workers in clinics and schools may also refer children for a risk assessment.


The Mother-Child Programme (PROM) operates only in the three poorest regions of the country, and consists of growth monitoring, nutrition education and supplementary rations to pregnant and lactating women and children under three years of age who are determined to be nutritionally at-risk. Pregnant women are considered at risk if: a) they have an upper-arm circumference of less than 23.5 cm; b) they were pregnant in the preceding year; c) their last child died during the first year after birth; or d) they are over 35 years of age. Children are considered at risk if they have second- or third-degree low weight-for-age, or if they have not gained weight for two consecutive months. Evaluations are made on-site by the health worker.


Targeting of the Nutrition and Feeding Programme for High-Risk Families is conducted through health workers' on-site evaluations and extension visits. For each child evaluated under the programme, the health worker in the participating health post fills out a form that evaluates ten risk factors, including both socio-economic and biomedical factors. Children under two years of age who are less than 90 percent of normal weight-for-age automatically qualify for participation. Otherwise, the health care worker ranks children according to the number of risk factors indicated in their responses. The remaining rations are allocated in order of priorities. The allocation of rations to each health post is based on available information on the population and facility-specific measurements of nutritional status. Implementation of this system has not been very effective. Some 89 percent of the assessed households were found to be eligible for benefits, mainly based on an evaluation of socio-economic indicators with limited assessment of biomedical risk factors. In practice, extension visits are very rare.

Targeted food subsidies


Egypt has a large and relatively complex food subsidy system dating back to 1941, when the Egyptian Government introduced rationing in order to provide consumers with such necessities as oil, sugar, tea and kerosene at low prices. The food subsidy system is part of a consistent policy that has been followed over the last 50 years to assure general access by the population to basic needs. During the 1950s and 1960s the total cost of the system remained small, but during the 1970s costs increased significantly in response to increases in world wheat prices. By 1980, food subsidies had expanded to include 18 food items and to account for 17 percent of total government expenditure. At that time, virtually the entire Egyptian population had access to subsidized foods through the use of ration cards.

Since the early 1980s, the Egyptian Government has been cutting back on public expenditures on all subsidies, including food. To reduce the costs of food subsidies, the government has used a variety of strategies, including increasing the price of subsidized food commodities, reducing the number of ration book holders and reducing both the number and the quantity of subsidized food items available to consumers. These strategies have been implemented slowly and gradually, enabling consumers to adjust more easily to the changes and avoiding any immediate hardship which could be caused by abrupt changes in food subsidies.

Since 1995, only four food items - bread, wheat flour, sugar and oil - have been subsidized in Egypt, accounting for less than 6 percent of total government expenditure. Brown bread (baladi) and wheat flour are sold to all Egyptians without restriction. Sugar and oil are distributed to consumers on a monthly quota basis through ration cards which are divided into two categories, with income as the defining criterion: green cards provide a higher rate of subsidy for lower-income households, and red cards are for higher-income households. Within the ration card system, the amount of subsidized oil and sugar that can be purchased monthly is based on family size (per capita rations). All but approximately one-fifth of Egyptian households hold some kind of ration card.

Bakers receive a daily quota of subsidized wheat flour and are required to produce a specified number of loaves of baladi bread of a specified weight per kilogram of flour. Production monitoring is carried out to ensure that these standards are met and to guard against leakage; there is a system of fines for infractions of the rules.

Through these reforms the government has been able to cut the total cost of food subsidies from LE 2 918 million in 1980-1981 to LE 865 million in 1994-1995. In 1994-1995 bread and wheat subsidies accounted for 60 percent of the total budget for food subsidies. In addition, the existence of other types of non-subsidized bread has led to demand from the non-needy population, reducing the beneficiary ratio cost from 92 percent to only 67 percent. It is worth noting that the subsidized brown baladi bread (82 to 90 percent extraction rate) is nutritionally superior to the non-subsidized bread made from fine or very fine wheat (76 and 72 percent extraction rates, respectively) since it is higher in fibre, vitamin and mineral content and has an appealing taste.

The current food subsidy system has generally been effective in providing certain basic items for food security for the poor through a combination of self-targeting (for baladi bread and wheat) and income-based targeting for oil and sugar. However, the absolute cost of food subsidies remains high. In an effort to continue improving the programme's effectiveness in reaching all of the most needy population and to improve its overall cost-effectiveness, the government is currently conducting a thorough review of the food subsidy programme.


Subsidizing of basic food commodities is a long-running consumer price policy in the Islamic Republic of Iran. A major objective of this policy is to maintain low food prices to consumers. Subsidized commodities include wheat flour and bread, sugar, rice, milk products, meat, tea and edible oil.

The Iranian food subsidy programme covers all rural and urban socio-economic classes. This non-targeted subsidy programme has little administrative cost, but carries substantial fiscal costs that must be borne by the government. For instance, in 1997, the cost of the food subsidy amounted to almost 6 percent of gross domestic product (GDP) at current prices; about 64 percent of it went to urban consumers and the rest to rural people.

Most of the consumer food subsidy budget (75 percent in 1997) has been spent on wheat flour to keep the price of bread to consumers low. In order to accomplish this, the government purchases the wheat crop from local farmers, based on a guaranteed price which is revised every year by the National Economic Council (NEC), and sells it to the bakeries at a much lower price. For example, in 1997, the government paid 480 rials/kg of wheat to the local farmer and provided it to the bakeries at the price of 40 rials/kg (less than one-tenth of the price paid to producers).

Because of the high fiscal costs of the general food price subsidies, the government is now considering a more efficient way of subsidizing foods through self-targeting.


Since 1970, the Tunisian Government has subsidized the consumption of basic foodstuffs. By the 1980s, the universal subsidy had turned out to be very costly, accounting for 4 percent of GDP and 10 percent of total government expenditure. The huge and rising programme costs, combined with substantial leakage to the non-poor, made an overhaul of the universal subsidy system an urgent priority. The government introduced a reform programme which included: 1) improving the targeting intervention towards the poor; 2) gradually adjusting prices to reduce and eliminate subsidies on certain products (such as animal feed) progressively; and 3) reducing unnecessary production and distribution costs for subsidized products.

The government adopted a policy to promote self-targeting using quality grading, which involves examining household expenditure data to determine whether there are significant differences in consumption across income groups. Thus, if the poor consume a different basket of goods than wealthier consumers do, the poor's basket can be selected for subsidization in order to focus on foods (inferior) that are unattractive to wealthier consumers because there are higher-quality alternatives. The "superior goods" approach allows other varieties or qualities to be available on the market at higher costs for those who can afford them, leading to a shift in demand from wealthier households who then consume less of the subsidized products. With such types of self-targeting strategies, subsidized food commodities are still available to all, but they are selected specifically to discourage the rich from consuming them.

By 1993, the Tunisian self-targeting reform programme had significantly reduced overall costs and government expenses. The reform has also been effective on equity grounds: under the universal subsidy programme more absolute benefits were transferred to the rich than the poor, while the poor benefited more from food subsidies than the rich through self-targeting.

Several lessons emerge from the Tunisian example:

Summary of self-targeting efforts in Tunisia




Durum wheat products

  • Maintain subsidies on semolina (continuous)
  • Eliminate direct subsidies on couscous and pasta (1993)

Bread wheat products

  • Replace subsidies on gros pain with subsidies on a single subsidized bread (pain unique) made from a distinct high-extraction rateflour (1995-1996)
  • Eliminate subsidies on baguettes (1993)
  • Eliminate subsidy on pastry flour (1991)
  • Liberalize production of other breads that attract the demand of high- income consumers (1991-present)

Cooking oils

  • Eliminate olive oil from subsidized huile de melange resulting in the creation of a subsidized generic grain oil (1989)
  • Shift subsidy towards generic oil, sold mostly to the poor in small quantities from large drums (en vrac), and away from bottled oil which is mainly purchased by the rich (oil type differentials:since 1989; packaging differentials:proposed)
  • Liberalize imports of pure grain oils packaged in glass bottles and sold at cost to attract high-income consumers and reduce consumption of subsidized oils (1992)
  • Create "fixed mixed" cooking oil with 40% olive oil and 60% grain oil to diversify range of products on market and reduce unit subsidies per litre (1993)
  • Olive oil also serves as a "superior" good (continuous)


  • Shift subsidy towards less-refined brown sugar which is not preferred by upper-income groups (1990)
  • Eliminate subsidies on cube sugar, which is consumed virtually exclusively by the rich (1991)


  • Shift subsidies to pasteurized-Reconstituted milk (least-preferred type of milk) packaged in cheaper cartons that are less attractive to wealthy consumers, such as berlingot cartons and coussin milk pouches (1991-1994)
  • Apply a small, temporary promotional subsidy to fresh, locally produced sterilized milk (a luxury) which is consumed primarily by the rich (1992)

Targeted food fortification programmes


Pakistan has tried to correct a very serious vitamin A deficiency problem through food fortification. Banaspati ghee, a form of hydrogenated oil, was selected as a food vehicle for vitamin A fortification, since it is widely used for cooking purposes. To narrow the gap between the increasing demand for hydrogenated oil and its limited supply, the government adopted measures to create incentives for manufacturers to increase production, while fortifying it with vitamin A (33 iu/g or 9.9 ug/g). The government made the vitamin A fortification of ghee and oil compulsory through legislation.

While the adopted measures resulted in increased production, little attention was given to quality control and assurance. This led to a relatively large amount of oil being produced, but with a low level of vitamin A fortification and, in some cases, without being fortified. An assessment in 1993 of the vitamin A level of different brands of Banaspati ghee, vegetable oil and margarine revealed that none of these products contained the level of vitamin A specified by the Pakistan Standard Institute. Only 40 percent of the analysed samples contained half of the recommended level as their maximum level, while the remaining 60 percent had even lower levels of vitamin A content.

This is an example of self-targeting through the selection of the food vehicle to be fortified. A food that is very much part of the daily diet of low-income groups, as is the case with ghee in Pakistan, will benefit the most vulnerable groups, but may also contribute to prevention of micronutrient deficiency in other groups, depending on their consumption patterns of the fortified group. Here, the concern is more with adequate coverage of the vulnerable groups than with leakage to less vulnerable groups.

Previous PageTable Of ContentsNext Page