How the choice of food security indicators affects the assessment of resilience—an example from northern Ethiopia

Abstract

Using longitudinal survey data from northern Ethiopia collected over 18 months, this study shows that conclusions about household food security are highly sensitive to measurement decisions. Especially important are 1) decisions about which food security indicators and cut-offs are chosen, and 2) whether analysis focuses on food security status at a given point in time or food security resilience over time. We define resilience as the probability that a household is truly above a chosen food security cut-off, given its underlying assets, demographic characteristics, and past food security status. Our study finds that different factors determine food security status and food security resilience. We also find that the drivers of resilience vary depending on whether food security is measured by Food Consumption Score (FCS) or the reduced Coping Strategies Index (rCSI). Literacy and livestock holdings are associated with both FCS status and FCS resilience, and the latter is also predicted by access to safe water and sanitation, the dependency ratio, and debt. In contrast, only previous rCSI scores predict current rCSI status, while marital status, literacy, livestock, and other forces matter for determining rCSI resilience. We also find that conclusions about food security resilience are sensitive to the cut-offs chosen to signify a food secure state.