The Kalobeyei refugee settlement in Turkana County in Kenya was established in 2016 with the aim of transitioning refugee assistance from a traditional aid-based model to one based on the principles of supporting host communities, offering self-reliance and promoting a development-based approach to assistance. The settlement is situated 3.5 kilometers from the Kakuma refugee camps and is home to 37,500 refugees from South Sudan (71 percent), Ethiopia (13 percent) and Burundi (9 percent) with smaller numbers from DRC, Uganda, and Sudan.
This article examines whether the policies and programs implemented in Kalobeyei are actually different from those implemented in Kakuma, and whether there are different self-reliance outcomes between recently arrived refugees in Kalobeyei and Kakuma, one year after the creation of the settlement. The authors employ a mixed-method approach that combines conceptual reasoning, quantitative research and qualitative research. The authors conceptualize self-reliance as a process through which self-reliance inputs (a combination of enabling factors and aid programs) lead to self-reliance outcomes (socio-economic outcomes and autonomy).
- About 15 months after the first arrivals in Kalobeyei, self-reliance-enabling factors are limited in both Kalobeyei and Kakuma, across nearly all indicators: environment, assets, access to networks, access to markets and access to public goods. In a few areas, Kakuma actually performs slightly better (reflecting the age of the settlement).
- Kalobeyei is not a pure ‘self-reliance’ model and Kakuma is not a pure ‘aid’ model; but they occupy contrasting positions on that spectrum. Kalobeyei has a number of policy features that distinguish it from Kakuma camps: it uses a cash-assistance program, which has been used on a limited scale in Kakuma; and it promotes dry land agriculture through the promotion of ‘kitchen gardens’.
- Refugees in Kalobeyei have slightly better self-reliance outcomes (in terms of nutritional outcomes and greater perceived autonomy) than Kakuma. This may be due to observed differences in the aid model, but the authors do not test this hypothesis empirically.