Issue 38 - Article 4

Preparedness for community-driven responses to disasters in Kenya: lessons from a mixed response to drought in 2006

July 31, 2007
Nik Bredholt and Matt Wingate, CAFOD Horn & East Africa Office

CAFOD responded to the drought in Kenya during 2006, but the response came late. This article considers ways in which better preparedness and greater and more timely involvement with drought-affected communities could have improved the response, to save lives and support livelihoods. Evidence of impending drought in Kenya was available from at least early 2005. Credible early warning information, delivered through recognised and well-resourced regional and national structures, spoke of successive rain failures, depleting pastures and worsening human and animal health. Despite this, a discernible, collective humanitarian response only got underway after another rains failure in late 2005, followed by a declaration of national emergency at the end of the year. By March 2006, when most international agencies, the UN and the government were getting up to speed, acute malnutrition rates were well above the emergency threshold, 3.5 million people were said to be affected, livestock were dying in large numbers and there were severe water shortages.

For CAFOD, and many organisations like it, the response was late, despite early warnings, including from affected communities themselves. Why? The answer lies in the specific needs of the humanitarian system, to get the right information at the right time, and in the interaction with affected local communities through which this information is made available.

Information that triggers a humanitarian response

Like others, CAFOD requires evidence to launch a humanitarian response: indicators of trends sufficiently bad to justify releasing resources, launching an appeal or scaling up staff. CAFOD’s local church partners were saying that there was a problem, but the information was largely anecdotal and was not credible enough.

There is a dichotomy here. On the one hand, professional, large-scale early-warning systems lack the flexibility or programme linkages to trigger responses at the local level. On the other, local actors speak with communities on the ground, but lack the systems or capacity to get their information heard. This quandary is echoed in wider evaluations of early warning systems. The UN recently flagged the need for greater emphasis on what it called ‘people orientated early warning’, and the Tsunami Evaluation Coalition (TEC) highlighted the critical role of local people and local organisations in ringing alarm bells when natural disasters occur.

Considerable efforts are now being made to consider the links between early warning information and prompt humanitarian response; in the Greater Horn of Africa, for instance, the Integrated Food Security and Humanitarian Phase Classification system developed by the Food Security Analysis Unit for Somalia is gaining wider recognition. Yet there is little evidence that consistent and adequate consideration is being given to the fundamental role of local actors and communities. The key problem is that local actors lack technical capacity, when measured by usual international humanitarian benchmarks. For CAFOD and other partner-driven organisations, its methodology is rooted in a belief in the potential of local organisations to reduce poverty and bring about sustainable change. Their reason for existing is in large part to help draw out that potential. The challenge for CAFOD and like-minded organisations, therefore, is not only to support the identification of roles for local actors in early warning and response, but also to consider how to strengthen capacity to meet those responsibilities.

Local capacity limitations

Capacity limitations go beyond early warning and problem identification. CAFOD responded to the latest drought in Kenya through a programme in seven districts, working with six partners and through about 50 rural health facilities. The implementers in this case were largely local church and development offices, with a history of response rooted in the Catholic Church’s charitable mandate to assist those in need. That mandate has endowed many of these partners with tremendous strengths, but has also given them technical weaknesses. The historic role in service provision and the clear sustainability of these institutions, which have existed for decades, means that there is no incentive for them to withdraw, and too often the response has been general food aid. This has done little to strengthen the community’s capacity to cope, nor has it challenged the belief held by some in the humanitarian community that local actors, and particularly faith-based organisations, are seldom adequately equipped to deliver technically sound humanitarian responses.

When CAFOD sought government funds on behalf of its local partners for a supplementary feeding programme, the reply to the application highlighted concerns around the technical capacity of such partners. Could they really deliver humanitarian programmes to international standards? The immediate priority after all is saving lives, and risks should not be taken in such matters. CAFOD proceeded with a response anyway, using its own resources to deliver a nutrition programme during the first half of 2006. With limited additional staff capacity and training restricted largely to district-level officers and senior staff in partner organisations, both field staff and beneficiaries struggled to see the distinction between a supplementary feeding programme and their general food ration. The consequent frustration did little to encourage partners or communities that a so-called ‘technically sound’ nutrition response was worthwhile. In many locations the programme reverted to the partner’s default course of general food distribution.

In mid-2006, CAFOD made a second attempt to deliver a supplementary feeding programme. Using three of the same partners, the programme considered lessons from the first phase, the most prominent of which were the lack of technical capacity and insufficient personnel to allow the scale up of such a programme. This time, however, the programme invested significantly in additional staff at rural health facilities, additional resources for transportation for monitoring and outreach and technical training and accompaniment, not only with office-based staff, but also at the community level through nurses and community health workers.

The huge difference between the first and second phase of programming has led to sometimes obvious but important insights for future programmes with local partners.

  • Most notably, it is clear that signs of drought are seen earliest at the local level. Whilst macro-level indicators, such as nutrition rates, grain reserves, national livestock prices or depleting water tables, are all valuable in highlighting the extent of a crisis, climate-dependent households feel the impact of a pending drought many months earlier, and are already discussing it among themselves and with local organisations.
  • For local actors already on the ground, there are no inherent capacity constraints that cannot be overcome to produce quick and ‘technically sound’ responses. However, support is needed, and it should come at appropriate times in the disaster cycle.

Locally relevant information and community participation

Communities have an important part to play in humanitarian action. A mechanism must be found to engage locally rooted health facilities, and their outreach services, in both the collection and dissemination of early warning information, with an emphasis on sharing information directly with disaster-prone local communities.

This extensive and established resource is hugely under utilised. With a little effort and support, rural health facilities could enhance their quantitative data and the utility of their anecdotal information to generate local responses to emerging drought, as it manifests itself at the village and even household levels. This demands the active involvement of community members, local nurses and volunteer mobilisers, and could thus serve the dual purpose of developing the evidence base, whilst also increasing the risk reduction capacity of communities, enabling them to make early and appropriate livelihoods decisions in the face of drought or other disasters.

In a recent CAFOD assessment of four pastoralist districts, Marsabit, Isiolo, Kitui and Mwingi, communities time and again spoke of threats to their traditional coping mechanisms thanks to a decade of poor rains, and petitioned tirelessly for knowledge and training to respond better to drought. This proposition is in contrast to the situation CAFOD and its partners currently face in marginalised parts of Kenya. Early warning information is generated for external responders, not for those affected by a threat. It is macro-level and one-way in its flow. In Kenya, although the most prominent system – the Arid Lands Resource Management Project (see http://www.aridland.go.kea) – is at the district level, it has little power to act until the information has passed through national structures. On the return journey, there is no evidence that information actually reaches the communities from where it came, and so what is created plays little or no role in their decision-making processes in relation to drought. At a time when participation and community-driven response are increasingly upheld by humanitarian actors, the structure of early warning in the region serves largely to reaffirm the dependency of the communities we strive to support.


Although some additional hardware is required, the challenge in creating an information collection mechanism through local structures lies, not primarily in equipment or infrastructure, but in capacity. Stakeholders should strive to help establish information formats that include anecdotal information, whilst also taking into account the vital statistical data that should contribute to humanitarian action, strengthening the accessibility of established systems and enhancing them wherever possible. Through a programme that contributes to the long-term management of malnutrition through rural health networks, a complementary structure could effectively be established using the same staff and volunteers to provide early warning information to local actors in a format that enables them to generate a response.

Meanwhile, parallel support must be given to address the capacity constraints most local agencies face. Genuine emergency preparedness must take place away from periods of humanitarian crisis. Since valued local actors are rarely just emergency response organisations, such capacity-building should be integrated with longer-term activities, and should recognise the impact of disasters on the livelihoods activities these actors also undertake. In particular, emergency preparedness should include contingency planning and technical specialisation, supporting partners to read the signs that local early warning is generating and establishing clear-cut response processes according to the evidence immediately – and locally – available.

Nik Bredholt ( and Matt Wingate work in CAFOD’s Horn & East Africa Office.


Comments are available for logged in members only.