when NPA took over distributions in November 1997.  Discussions with villagers and NPA staff

detailed two major problems with applying these categories (apart from the important question of

who within the village decides which people should be in which group):

  Firstly, the categories had been understood to apply to individuals rather than households,

with no account taken of support from family members or even (for group 1) relative wealth.

 Unable to work  had been taken to refer mainly to the old and handicapped, and pregnant

women. Thus an elderly infirm person living in a relatively well off household supported by

able bodied family members would be considered eligible for food aid,  while an extremely

poor family with no income and many children, but with no old or sick family members,  would

be excluded.

  This issue highlights a disconnect in the concept of `vulnerability' between food aid planners

and recipients, which was echoed in discussions in both Uganda and Kenya.  The village

informants clearly saw the old, sick, orphans and widows as the `vulnerable' members of the

community (i.e. individuals who rely on the support of others), while the outsiders see relative

household poverty (assets, income, and other resources which enable people to `cope' in

difficult times) as central to famine vulnerability.  Selecting households by relative wealth within

communities is not only difficult and liable to create conflicts, but seems to go against people's

perception of who the beneficiaries ought to be.

Free distribution had been almost the only mechanism used in Dodoma.  Although people had

been registered in the standard three categories including  able to work ,  food for work schemes

were not available in the area.  During the government relief distribution in November 1997, the

District Disaster Prevention Committee in Dodoma Rural had allocated some of the food to village

schools to keep up attendance (having observed that children were missing school to go and

collect wild foods from the bush).  NPA commented that the District would have liked to continue

this with the later WFP distributions, but that the WFP distribution plan / rules did not allow it.

Dodoma is one of the drought prone areas identified by WFP for Quick Action Projects in both

School Feeding and Food for Work [refs 44 and 45], as 

post

 emergency measures.  The general

issues of using these distribution mechanisms 

during 

emergencies are discussed in section 2.2.

1.2.d.

Focus area 2:  Arusha Region

Arusha Region, which borders Kenya in the north, is more complex and varied in terms of food

security: it includes relatively rich urban and trading areas as well as widely varying farming

conditions, and pastoralist populations (particularly in Ngorongoro).  The importance of

disaggregating assessments of people's economic environment and coping ability in the face of

drought or crop failure, at least to District level, is clear in this context.

In Arusha, the study team focused mainly on the government led process of needs assessment

and targeting at Stage 2 (i.e. area targeting within the Region) and the system of information

provision from the Region for national decision makers.  As far as the team could tell from

discussions in Dar Es Salaam, the system used in Arusha was followed throughout the country:

Arusha is discussed here as an example of that system and not as an exception.  The section is

largely based on very helpful and open discussions with Regional and District officers, who had

clearly found the management of food aid problematic and were interested in discussing

alternative approaches for the future.

A number of analytical flaws in the needs assessment process are shown in the  Report on

Targeting Drought Affected Villages  issued by the Regional Commissioner's Office following the

joint government NGO village targeting exercise in November 1997 [ref 1].  Firstly, the overall

balance sheet calculations for the Region are made by multiplying the total population of each

A 12