Taking the politics out of resource allocation: the Kenya experience

By Jeremy Shoham

Jeremy Shoham is co-director of the ENN and editor of Field Exchange. Over the past two years he has been working periodically as a consultant for the Government of Kenya, DfID and WFP Kenya helping to strengthen methods of targeting emergency food aid.

For many decades, the main basis for emergency food aid geographic targeting decisions taken by the Government of Kenya (GoK) has been district requests. These requests are made to the Famine Allocation Relief Committee (FRC) or the Permanent Secretary in the Relief and Rehabilitation Department of the Government. Requests were made by the District Social Dimensions of Development Committee (DSDDC) or in the case of the Arid Land districts¹ the District Steering Group (DSG). The system has never been transparent and has been critically subject to political lobbying and media pressure. It is widely acknowledged that politicians at various levels have perceived the provision of emergency food aid as a resource through which political advantage can be gained and have therefore often ignored objective criteria of need.

One key factor which has contributed to the capacity of political lobbyists to influence the geographic targeting decision-making process, has been the absence within the FRC of a framework for analysing changes in food security. This has meant that available data have not been used effectively to inform geographic targeting decisions.

An outcome of the 'district request' led system of geographic targeting is that GoK food rations have been allocated to many districts with some receiving assistance on an almost continuous basis irrespective of crop performance or food availability (see figures 1 and 2). This has partly contributed to the very small size of rations actually allocated through the GoK emergency food distribution system with the result that extremely food insecure households have had very limited benefit from the GoK programme despite the large amounts of money spent. Over the 16 month period of the 1996/7 GoK emergency response more than 200,000 metric tons of maize were purchased and allocated (7% of national maize consumption). Yet in spite of this enormous expenditure by the GoK, per capita rations of as little as 2kg of maize per month were frequently recorded. The GoK therefore received very little recognition or praise for its on-going emergency relief programme and indeed on many occasions attracted criticisms from external agencies for the small quantities of emergency food aid which beneficiaries received.

The targeting system also did not make a distinction between the chronically poor or destitute and emergency affected populations. Although it is difficult to estimate the numbers of poor and destitute people who have been targeted by the GoK emergency relief programme in the past, it is clear that the chronically poor/destitute have been considered as legitimate beneficiaries of emergency food aid even when there has been no emergency event. Sources in the Ministry of Agriculture have indicated that even in good years, when maize is being exported, there has been in excess of one million recipients of GoK emergency food aid. Analysis of the data on GoK maize allocations shows that in some instances when there has been excellent harvests the number of districts receiving food aid has remained more or less the same (see figure 3).

Political influence within districts

Geographic targeting within districts has also been highly politicised. Decisions about the amount of food to be given to each division are taken by members of the DSDDC or DSG. The process of decision-making bears many similarities to that at national level.

Political influence on intra-district food allocations has occurred for many reasons:

1. The District Commissioner has been the final arbiter of resource allocation and is subject to a high degree of political lobbying.
2. The district committees (DSDDCs and DSGs) comprised 'politicians' as well as technical staff from line ministries and NGOs (in the case of DSGs).
3. The district committees lacked a clear set of objectives for emergency food aid and had no analytical framework for using the food security information to which they have access.
4. There was very little normative comparison of data and no systematic use of baseline information from previous years for comparison.
5. There was lack of clarity over whether food aid was to be used to meet the needs of the chronically poor or of those affected by acute emergency events.
6. There were many data gaps e.g. on livelihood systems, baseline information on key indicators like prices or malnutrition data from MCH clinics.

As a result it was often extremely difficult to omit any location or sub-location from food distributions. For example, in Machakos district all 226 sublocations received emergency food aid between 1998- 2000.

Following earlier trips to Kenya connected with emergency food aid planning and provision I was asked to return in January 2000 on behalf of DfID and WFP to review the method of GoK emergency food aid targeting and to make recommendations for improvement. Part of my terms of reference involved looking at targeting within districts.

My key recommendations (made to the Kenya Food Security Steering Group comprising WFP, GoK and INGOs) for improving targeting at district level were as follows:

District committees require a clear set of guidelines on the objectives of emergency food aid and means of targeting food aid commodities.

Masai pastoralist on his way to the market to sell a goat.

These guidelines need to specify that:

1. emergency food aid should only be provided for those livelihood groups affected by a recent shock to their food system and that the long-term poor should not be recipients of emergency food aid
2. in order to identify and target those affected by a recent emergency event a normative and livelihood type analysis must be conducted, i.e. determine how the indicator levels (prices, crop production, pasture, etc) compare to normal and what the levels demonstrate in terms of access to food and income for specific livelihood groups.
3. emergency food aid also has a role in protecting livelihoods and preventing malnutrition and mortality, e.g. it can be used to prevent people resorting to survival strategies that undermine longterm viability such as distress sale of key livestock, etc.

In order to institutionalise these guidelines at district level my report recommended a process of sensitisation at district level through training.

key informant interview with Masai women in the Narok district.

These recommendations for Kenya were by no means new. Agencies like Oxfam had supported district teams in Wajir to adopt and successfully use a livelihood framework for targeting in previous years. Reports from other consultants had also contained similar recommendations².

Capacity building at district level

I returned to Kenya in Jan 2001 to assist in the district level sensitisation and training, this time working for the Relief and Rehabilitation Department in the Government. The work involved training district level rapid assessment teams³ in the principles of food economy analysis and to help them apply the methodology to the current drought situation and assist in analysing the results. Four regional training programmes were carried out simultaneously (involving three other trainers) covering 22 districts in Kenya that were requesting emergency food aid at the time. This training took place over a three week period.

The first week involved working with the district team to identify and locate major food economy groups in their respective districts and to attribute and quantify food and income sources for each food economy group in a normal year. The results were represented on pie charts as percent of total food and income coming from different sources. This was then followed by theoretical exercises demonstrating how by comparing current food security related information (e.g. market prices, milk availability, crop production), with baseline values it is possible to estimate and quantify current access to food and income sources and to determine whether there is a food deficit and need for emergency food aid provision. Teams then went back to their districts in the second week to collect food security information from District Head Quarters level and through key informant interviews at village level for each food economy group. They returned in the third week to analyse the data in relation to each food economy group, i.e. to assess food and income sources in relation to normality and to determine whether there was a food deficit for each food economy group. This was not a rigorous training in food economy analysis. Furthermore, the limited time for training meant that certain aspects of traditional food economy analysis were omitted, e.g. stratifying food economy groups into wealth groups. However, the training did provide district government, local NGO and WFP staff with a framework for analysing food security information. This was something that was previously lacking.

An interview with farmers from Taite-Taucte district following the 2001 short-rain failure.

The district teams felt that the assessment methods enabled them to present a coherent assessment of food security and to argue more confidently about intra-district food aid needs. The objectivity of the Food Economy assessment framework empowered district teams while disempowering politicians on the DSDDCs and DSGs. In short, the district teams embraced the training and worked very hard to use the tools to come up with credible estimates of food aid needs for their respective districts.

Although I was not involved in writing the final report that was presented to the KFSSG, I understand that WFP and GoK relied substantially on the district team findings to target food aid during 2001. This was the first time a standardised framework and methodology had been used by the GoK to quantify inter and intradistrict emergency food aid needs in Kenya. Furthermore, it was noted that the final report submitted to the KFSSG recommended emergency food aid for a lesser number of districts than the initial number requesting emergency food aid prior to the assessment by the district teams.

Conclusion

In conclusion, what seems to have occurred in Kenya is that a critical mass of dissatisfaction was reached with the politically driven system of emergency food aid targeting that had been operating for several decades. This resulted in co-ordinated pressure from key stakeholders for change. Changes have occurred at many levels (see article by Robin Wheeler in Field Exchange 12). One key development has been the adoption of a framework for analysing changes in food security (something which has hitherto been absent). This has provided a more objective and transparent method of assessing emergency food aid needs for population groups and as a consequence has reduced political influence on the process of geographic targeting. Capacity building has begun at national and district-decision making level. While this article is not championing food economy analysis as the only framework for food security assessment and estimating emergency food aid needs, it has aimed to highlight the contribution an accepted analytical framework can have in de-politicising emergency food aid decision-making. From being a country with one of the least rigorous methods of defining emergency food aid needs, Kenya's evolving methodology for targeting emergency food aid could now very well become a model for other countries in the region.

For further information regarding evolving targeting methodologies in Kenya contact either: Pippa Coutts at the Arid Lands Resource Management Project in the Office of the President, Nairobi. (Email: pippac@africaonline.co.ke) or Jeremy Shoham (email: jshoham@easynet.co.uk).

¹Arid Lands Resource Management Project

²Jaspars.S., 1998, 'discussion Paper. Food aid Strategy for the Kenya Emergency Floods Operation' presented at WFP/NGO/GoK meeting on 11th June. Jaspars S. 1998. 'Final Strategy for Food Distribution in Flood Affected Areas of Eastern, North Eastern Province, Kenya, 6th August

³Teams comprised staff from government (mainly district agricultural and veterinary officers), NGOs and WFP.