Modeling Subnational Risk of Acute Malnutrition in Conflict-Affected Settings

• In Event: Poster Session: Experimental Research & Qualitative and Multi-Method Research
  In Poster Session: Poster Session: Qualitative and Multi-Method Research

Abstract

Nigeria, Somalia, South Sudan and Yemen are currently facing famine conditions - the first recorded instance of four countries experiencing such declared crises simultaneously. All these countries have been prone to violent conflict and remain susceptible to climate shocks, contributing to greater vulnerability to famine. The gravity of these situations serves as inspiration for devising effective means to better anticipate crises and guide appropriate responses. This paper presents initial results of empirical analyses focusing on the cases of Karamoja, Uganda and West Pokot, Kenya. An evidence-driven computational model is used to assess the reliability of leading indicators in predicting trajectories of susceptibility to acute malnutrition of children at different levels of spatial and temporal granularity. Household and expert surveys administered in both contexts shape the design of the model, which is then seeded and calibrated with data on individuals, households and their environment, as well as validated with out-of-sample tests. A final step uses the model to explore counterfactual scenarios of different interventions, including allocations of humanitarian and development assistance.

Authors

• Ravi Bhavnani, The Graduate Institute, Geneva

• Karsten Donnay, University of Konstanz

• Mirko Reul, Graduate Institute Geneva