Nithi), Coast (Taita Taveta) and North Eastern (Mandera) Provinces  [ibid].  The EMOP expansion

in September reduced the scale of this operation, targeting the expanded school feeding more

narrowly on 151,292 children in drought affected pockets of Kitui and Makueni for the third term

(September to November).

In addition to the demographic targeting effect of school feeding (i.e. ensuring that food aid is

received by a nutritionally vulnerable age group, primary and pre primary children), WFP Kenya

has noted two major advantages of  expanded school feeding as an emergency mechanism:

lower than usual risk of diversion; and

  limited additional planning and preparation inputs (since the institutional arrangements are

already in place for the regular program).

Its main weakness as a targeting mechanism, is that it may not  reach the poorest and most

vulnerable children. The SCF Food Economy Assessment of Kitui in July 1997 found that:

 According to key informants, many of the children of poor households do not go to

school or if they do, they drop out early on. The main reason stated is that many parents

of Kitui, as with many other areas of Kenya, cannot afford the accompanying costs  

uniform, textbooks, and parent/ teacher association fees.   Key informants reported that

the `poor' and `very poor' can generally only afford to send one child to school, thus the

SFP only contributed a small percentage of the family food needs. As the `middle' and

`rich' can afford to send their children to school, and have smaller families, the

contribution of SFP to their household intake is greater  

[SCF (UK) 1997b, ref 131]

This is not universal, however (as shown by the Ugandan example cited in appendix 1.2.a., and

the comments of the DEO in Garissa mentioned in section 3.2.d.):  it depends on who goes to

school in a particular community, and why.  SCF's conclusion from this observation in Kitui was

that distribution mechanisms to target the poorest were needed 

in addition to

 school feeding.

By the time the 

El Nino  floods

 hit, in October 1997, the drought relief operation was entering a

process of phasing out, with both GoK and WFP distributions planned to continue at a diminishing

rate of coverage until the end of the year.   The new disaster was unexpected in its severity and

unpredictable in its development (no one knew when the rain would stop or how bad conditions

would get).  Many areas were cut off from road access (and therefore from normal market supplies

of food) for months.  The food security and livelihood impacts were not only in the short term loss

of crops, food stocks, seeds, and so on, but in a protracted health crisis for both people and

livestock due to the rise in water borne diseases and vectors, and the unaccustomed cold and wet

conditions in normally arid areas.   Livestock losses in some of the pastoralist areas were

devastating.  The development of the disaster is reflected in WFP's planning figures for food

beneficiaries, which actually grew from 390,000 in the immediate aftermath [EMOP 5969,

December 1997] to 539,600 in the first expansion [EMOP 5969/01, March 1998] and then to

587,480 in the third phase [EMOP 5969/02, July 1998].

The flood damage to roads,  which hampered the distribution of relief and necessitated air lifts of

food and medicines, also hampered the flow of information for targeting decisions.  Many areas

were unreachable. The UNICEF/ UNDP Flood Assessment Mission in March, for example, was

only able to assess malnutrition among children close to helicopter landing sites.

In this situation, the main targeting question was the quantity of food to be distributed in each

affected District: this question was broken down into the arithmetic of  

how many people?

multiplied by

 what ration rate?. 

 There was no real disagreement about which Districts needed help

in the flood disaster: however, targeting 

within

 the Districts, and differential ration rates 

between

the Districts, were points of contention.  Targeting of the neediest within communities was not an

issue,  partly because action was urgent and beneficiary selection would have been too time 

consuming, and partly because vulnerability in this kind of sudden disaster was not necessarily

A 37