It is very important to specify carefully the exact target for the program. Is it drought-
affected households? Is it poor households? Or is it chronically food deficit households? The
commodity or service being delivered by the program may also be relevant to he choice of
targeting indicator. This is likely to be different if a program is delivering food as opposed to
health services.

 The ranking of EPAs in Malawi is very
sensitive to the choice of indicators, as
examination of the following maps shows.
Figure 1 shows the impact of the short rainfall RURAL HOUSEHOLD
 FOOD SECURITY INDEX
in 1994/95, in terms of the percentage FOOD SECURITY
variation of actual production of crops, j
measured in kilocalories, from the historical
average kilocalorie production, by EPA. The
dark areas showed significant production
shortfalls of up to 150 percent reductions.
The light areas were less affected, and some
had significant production increases. The
worst-affected areas are spread fairly evenly
throughout the country. In fact, a variation on
this was used to target the 1994/95 drought
relief programme. EPAs whose production
was less than 1800 kilocalories per person per --
day, and had suffered a fall in production over
the historical average, were targeted for '""" -
emergency relief. This identified areas where
the production shortfall took the EPA below
a minimum coping level.

 FEWS has been developing an
indicator of overall vulnerability, which is F e
based on a principal component analysis of
data series on food availability, cropping
patterns, health, education, infrastructure and demographic variables. This has not yet been
finalised, but Figure 2 shows its current status. This shows clusters of vulnerable EPAs along the
borders with Mozambique and Zambia, with better-off EPAs along the lakeshore and in the North
of the country. This gives a very different picture from the drought map, partly because the aim
is to identify chronic, rather transitory status, and also because of the inclusion of data on health
and education variables. This type of indicator would be useful for programs where there was a
strong community element, such as the MASAF program.

 If, however, the program under consideration provided some kind of resource transfer to
the household, such as food, then it might be more appropriate to rank EPAs on the basis of