Commissioners to identify the needs of smallholders in their area. For example, in 1992/93,
national criteria were established, to register smallholders whose crops were severely affected,
female headed-households without sources of income, elderly or handicapped heads of household
and chronically sick or destitute households. The registration process was carried out in advance
and districts were phased in according to perception of need. The overall amount of relief food
necessary was identified at the national level, and DCs had to distribute the amount of food they
were allocated as best they could. It should be noted that the national criteria used would identify
not only those affected directly by the drought but also some of the chronically food insecure
whether or not they were directly drought affected. This illustrates one of the difficulties, both
politically and practically, of targeting in response to a specific shock when there is a high level
of chronic food insecurity in normal years.

 Once FEWS got underway, efforts were made, in conjunction with WFP, to develop a
more objective, statistical basis for relief targeting. The first step was to disaggregate crop
estimates to the lowest possible level, the Extension Planning Area (EPA). Then all crops were
translated in terms of per capital daily kilocalories produced. This was then compared to the
historical average for the EPA, to give a measure of dislocation. In 1994/95, following the
previous year's poor harvest, donors and NGOs put together a monitoring system which collected
information on child nutrition (Middle Upper Arm Circumference measures), consumption and
grain and livestock prices. These were combined to give a ranking of the most vulnerable EPAs
in the country. Tonnages of food required were compiled using EPA population figures, on the
assumption that not more than 85percent of an EPA would qualify for relief. The food available
was then allocated according EPA ranking and the logistic ability of an EPA to receive the
tonnage allocated. The system operated fairly smoothly, but there was a tendency for DCs to
redirect some of the food to EPAs which had not been included in the distribution. It is often
politically difficult to exclude areas from relief operations, particularly if some of the excluded
population have clearly been affected.

 In 1995/96, there was a more modest relief operation. This was targeted by EPA, initially
using the Kilocalorie analysis from FEWS, which was replaced by a targeting system based on
the results of a Rapid Food Security Assessment carried out by the Malawi Red Cross. This
appraised the situation of the 78 most frequently targeted EPAs in 1994/95. The targeting
indicator was developed from information on remaining on-farm stocks and the availability of food
from income generating activities and coping strategies. EPAs were then identified as either food
surplus or deficit. Deficit EPAs were targeted for relief food. This exercise was carried out on
a monthly basis. Unlike previous relief programs, three channels were used to distribute food:
distribution through schools; vulnerable group feeding where rations were distributed to people
registered at MCHs (mother-child health clinics); and FFW programs.

 An evaluation of the effectiveness of these channels indicated that VGF distribution was
more effective in targeting the most vulnerable households, whereas school distribution reached
more households, often from a wider socio-economic range of households. Sometimes the poorest
children were excluded because they did not attend school on a regular basis and therefore were