Beneficiary Assessment for Independent Evaluation of Anticipatory Action Pilot in Somalia (August 2021)

Executive Summary

Following widespread and severe droughts in Somalia in 2010-2011 and 2016-2017, the UN Office for the Coordination of Humanitarian Affairs (UNOCHA), Humanitarian Country Team, the World Bank and other partners developed the Anticipatory Action (AA) Framework. Through this framework, the Central Emergency Response Fund (CERF) releases funding against a pre-developed Anticipatory Action Plan to help mitigate projected life-threatening humanitarian impact and protect vulnerable people under worsening conditions. The pilot of this framework was launched in 2019, combining three pre-agreed components: forecast and triggers; anticipatory actions; and finance. In this way, the pilot establishes when and on what basis the action will be triggered for a specific event and how much funding will be allocated to a particular agency, as well as what activities the funding will be used for.

The assessment involved assessing beneficiary experiences and drawing lessons from four anticipatory action interventions conducted by three UN agencies. The reviewed anticipatory actions include two projects implemented by the World Food Programme (WFP), one focusing on nutrition and the other focusing on cash transfer. The reviewed anticipatory actions also include a water and sanitation project implemented by the International Organization for Migration (IOM) as well as an agricultural project targeting pastoralist communities implemented by the Food and Agriculture Organization (FAO). The key findings of the beneficiary assessments are:

• The Anticipatory Action is a well-suited approach given the high predictability of shocks like drought and floods. The pattern of crises in Somalia defines the basis of crisis prediction in the country. Several of the recent crises in Somalia are recurring and follow the seasonal calendar. While floods tend to occur once or twice every year during the rainy Gu and Deyr seasons, droughts tend to occur once every two or three years during the dry Jilaal season.
  Pest infestations can also be predicted by virtue of their association with the rainy seasons. Therefore, the predictability of these crises is high and can be anticipated.

• AA is most effective when linked to a specific shock and specific shock indicator. The AA framework was developed specifically for drought and was used for different and multiple shocks, which is not ideal. Not all crises warrant the same type of response, hence the anticipatory actions should link to specific shocks and should not try to cover multiple shocks with a single framework. Food security can be triggered by many factors, hence it might not be the best trigger to use for a multi-sectoral response. Desert locust control is very time sensitive as the intervention needs to occur prior the exponential development of the locust population. Provision of cash, health or nutrition services to targeted vulnerable households who have exhausted all their coping mechanisms is also time bound as it can be a matter of life-saving issues. However, water points are less time sensitive since they are expected to be operational even during crises (if constructed/rehabilitated to adequate minimum standards and in ways that could mitigate shocks).

• Beneficiaries, based on their livelihood are able to predict their needs in advance, clarifying the demand and the opportunity to develop further AA. The predictability of crises is very high amongst local groups such as farmers and pastoralists, whose livelihoods depend on natural resources and the seasons. The survey respondents reported that they knew they would face a crisis between 5 and 15 weeks in advance. The community could be used to monitor shocks and crises as part of a community-based early warning (EW) mechanism. Greater participation of communities is also expected to improve satisfaction with project interventions which was reported to be low.

• Prioritizing actions that help expand the risk management options of the most vulnerable, especially IDPs, women, etc. have the potential for greater impact. The most vulnerable groups under the AA interventions are IDPs and Women and are the most represented amongst IOM and WFP beneficiaries. As the WFP interventions are significantly geared toward lifesaving, the timing of intervention is more important than it is for IOM and FAO interventions. The WFP intervention is targeting urban poor, mostly IDPs unemployed or casual workers. Therefore, the ability of WFP beneficiaries to adapt is very limited. The importance of timely (early) intervention in emergency interventions for vulnerable groups is evident in the fact that only IOM beneficiaries reported in a high proportion that they would have done things differently had they received earlier assistance. Moreover, the beneficiaries of the WFP cash transfer project reported in a relatively higher proportion that the earlier cash transfer made a difference in their ability to cope with the crisis. It is important to note that not only should the profile of beneficiaries be used to determine the necessity of early intervention, but it should also be used to define indicators and as such, there should be a set of indicators within similar livelihood groups. For example, timely intervention should aim at decreasing the loss of livestock, or the selling of productive assets.

• Targeting and tailoring the package of interventions to key vulnerable groups is key to meeting their needs. Beneficiaries under the reviewed FAO project have some production capacity at household levels (e.g. farmers, pastoralist, business owners, etc) and are able to generate income on the basis of investment, labour and risk management. However, unemployed profiles with no production capacity or low-level casual workers such as IDPs could in many cases be fully dependent on emergency support during crises. Majority of such beneficiaries were found to be under the reviewed WFP and IOM projects. As the WFP beneficiaries have no production capacities and low capacities in dealing with shocks, they require programming closer to emergency response compare to FAO beneficiaries. The survey conducted as part of this assessment shows that 47% and 55% of the beneficiaries of the nutrition and cash transfer WFP projects reported earlier response respectively. Meanwhile, 36% and 30% of FAO and IOM beneficiaries respectively reported early response.

• There is a need to consider nuances and advantages of different targeting approaches. The projects assessed different targeting criteria; the IOM and FAO projects defined targeting both at community and household levels. Targeting at community level compared to targeting at household level improves value for money as the number of beneficiaries is higher for targeting at community level than for targeting at household level. The VFM of the IOM project was found to be goof due to the higher level of sustainability of the intervention on infrastructures. Although, the issue on desert locusts addressed by FAO is very much time sensitive and while the cascading positive effects of this activity can be exponential, the intervention is a one off. Water points and other infrastructures should be appreciated for their long term effects and their ability to provide support to the whole community over different shocks.

• The assistance provided should have been provided earlier to ensure higher beneficiary satisfaction. In general, the appreciation of the AA interventions was found to be lower than expected among beneficiaries responding to the survey. During this survey, the average score given by beneficiaries respondents was between 4.6 and 5.1 when prompted how they would recommend the intervention to a friend on a scale of 0 to 10 (10 being the highest). The specific reasons for this low scoring are unclear but it should be noted issues of recalling information could have influenced their opinion. Furthermore, the low scoring could be related to ownership, consultation, timing, mismatch between beneficiary expectations and assistance provided, etc. This highlights the need to consider the process and quality of interventions along the timing of response, but also the level of expectations of beneficiaries. This would require considering options better appreciated by the target communities and following processes warranting high quality and timely responses. This point should be nuanced with the fact that 82% of respondents reported that humanitarian assistance received in general made it easier to handle their issue (48% agree and 38% somehow agree) and that 70% of respondents reported that these specific interventions have improved the quality of their life.

Based on the findings of the assessment, the following recommendations are given for AA interventions:

1. Consider categorizing beneficiaries based on the level of vulnerability and separating those in need of life-saving interventions and those in need of interventions addressing longer term vulnerabilities.

2. The anticipatory action should build upon Early Warning (anticipation, crises pattern) and resilience programming (complementarity) in Somalia. This would contribute to more effective interventions and reduced levels of vulnerability.

3. Increase targeting at community level on infrastructures as these will benefit all community members and will remain operational for future shocks increasing de facto the sustainability of the intervention but also increasing the readiness of these infrastructures for the next crises.

4. Along the institutional early warning, develop community-based monitoring based on community information. Predictability of crises can built on the long term analysis and sophisticated data management from actors involved in EW but there are also opportunities to better listen to the communities, notably through social network monitoring. An example of tweets analysis has been provided in this report.

5. Some interventions were implemented at a very slow pace but are still relevant within the overall cycle of crises. Therefore, thinking beyond the response to the crisis, anticipatory should look at the cycle of recurrent crisis and should provide responses addressing vulnerabilities met by specific groups at a specific time of vulnerability.

6. Build regular target community consultation into the anticipatory action framework and ensure that this is also reflected in agency projects. This is essential for ensuring the selection of the groups that are most in need of assistance as well as beneficiary satisfaction and appreciation of the project.