Field Exchange Mar 2003: Supplementary feeding in Zimbabwe

Summary of research1

During the periods 1992-3 and 1995-6, a child supplementary feeding programme (CSFP) was used in Zimbabwe to combat child malnutrition associated with times of drought. Evaluations in the past have concluded that the CSFP was effective in preventing an increase in malnutrition among children under five years, especially during the 1992-3 period. However, such evaluations made only cursory use of available household survey data. A recently published article based on a more detailed analysis of household surveys presents evidence that contradicts previous findings.

The objectives of Zimbabwe's CSFP included feeding vulnerable children, so as to maintain or improve the nutritional status of children under five in drought affected areas and to assist in averting starvation-associated child deaths. The CSFP aimed to provide children aged 6-59 months in all target areas with a nutritious meal that met 40% of daily energy requirements and 88% of daily protein needs. The official daily ration consisted of 66g of maize meal, 20g of groundnuts, 20g of beans and 12 ml of cooking oil. The programme comprised of wet feeding for five days a week, with local mothers providing the labour and water for cooking and cleaning up. Government and NGO officials managed the distribution and storage of CSFP rations.

Four nation-wide household surveys had previously been carried out in Zimbabwe, one occurring at the beginning or towards the end of both programme periods (1992-3 and 1995-6). Three of the national surveys were implemented by government ministries with support from UNICEF, while the fourth was a poverty assessment study. Re-analysis of these surveys set out to establish the following:

  • the level of coverage of children under five in drought affected areas during the two programme periods

  • the extent to which the CSFP reached children from poor and nutritionally vulnerable households, and how did this coverage compare to coverage of children from better-off households

  • the proportion of malnourished children who actually got fed by the CSFP.

In order to check whether children from poor households got CSFP benefits, several poverty indices were developed, based on income poverty, consumption poverty and asset poverty. Assessment was questionnairebased, although specific content varied from one survey to another. Using mid upper arm circumference (the anthropometric index available in three of the four household surveys), the proportion of currently malnourished children receiving supplementary feeding was calculated.

CSFP coverage

In establishing the CSFP coverage level (table 1), re-analysis observed that:

  • CSFP coverage was lower earlier in the life cycle of feeding programme and rose towards the end.

  • Coverage was higher and more evenly spread in the 1992-3 period- the worst of the two droughts.

  • CSFP coverage was patchy, with between one-quarter and two-fifths of areas without feeding programmes.

  • Even within those areas covered, large numbers of children were not receiving supplementary feeding.

On the positive side, the regularity of supplementary feeding of beneficiary children appeared high.

Table 1 CSFP coverage of children under five according to household surveys, 1992-3 and 1995-6

Date of survey
Proportion of under fives in CSFP for all rural areas
Proportion of under-fives in CSFP where CSFP operational
October 1992
March 1993
Late 1995
March 1996

Targeting vulnerable children

With the exception of the March 1993 survey findings, half or more of under-fives from poor and nutritionally vulnerable households were not enrolled in supplementary feeding. Coverage of children from these households was lower in 1995-6 compared to similar periods in 1992-3. Even in areas where CSFP was operational, large numbers of children from poor and nutritionally vulnerable households did not receive supplementary feeding. It is not clear whether under-fives from poor households found it easier or harder to access supplementary feeding than did children from non-poor households - the evidence is mixed. What is clear is that the probability of children receiving or not receiving supplementary feeding was not particularly different between poor and non-poor households.

CSFP impact on nutritional status

The areas where CSFP was operational were not necessarily where levels of malnutrition were highest. Record keeping at feeding points on children's nutritional status before, during and after supplementary feeding was extremely haphazard. Large numbers of malnourished children were not in supplementary feeding, even in areas where CSFPs were operational. Less than half of malnourished children surveyed early in the life-cycle of both programmes were in supplementary feeding. Even at the height of programme coverage in March 1993, one-quarter of children were not receiving feeding in operational areas. Furthermore there was some evidence that malnourished children were less likely to be in the programme than other children. This could mean that the programme was effective or that malnourished children had less access. However, the reasons why children, even malnourished or poorer ones, did not attend remains obscure. One possible explanation put forward has been the opportunity cost to mothers.


A comprehensive evaluation of the CSFP impact on nutrition in Zimbabwe is not possible. No data were collected at feeding points and no information is available on the regularity or quantity with which individual children were fed. Furthermore, it is inappropriate to move, as the original evaluators did, straight into the analysis of impact indicators, while bypassing important questions about how many children got fed and what their socio-economic and nutritional status was.

The re-analysis does show that the 1992-3 CSFP should be distinguished from its successor in 1995-6. Overall, programme coverage was higher in 1992-3, as was the proportion of malnourished children who received supplementary feeding, especially in the latter part of the programme's life cycle. However, the success of the 1992-3 feeding programme was at best a qualified one. Even during peak coverage around March 1993, almost one-third of children, and one-quarter of all malnourished children, were without supplementary feeding.

Children who were enrolled in the programme tended to get five meals a week and amongst these, there is some evidence that feeding did provide nutritional benefits.

In essence, evidence from the household surveys casts considerable doubt on previous evaluations, in particular claims that CSFP played a major part in controlling malnutrition rates and preventing famine in Zimbabwe in 1992-3 and, especially, in 1995-6.


1 Munro L (2002).Zimbabwe's child supplementary feeding programme: a re-assessment using household survey data.Disasters, 2002, 26 (3), pp 242-261