No man is an Island: A spatially explicit approach to measure development resilience.

• This study develops a spatially explicit approach to measure development resilience through capturing human interactions and the impact of surrounding systems. • The proposed method significantly and robustly improves overall targeting accuracy, especially at aggregated level and for households experiencing drastic well-being shifts. • From a policy perspective, incorporating spatial elements into measuring resilience has the potential to improve resource allocation efficiency for development and humanitarian interventions. • The method chosen for measuring development resilience implies a policy tradeoff between over-coverage and leakage. Building resilience is paramount in the development and humanitarian agenda, yet existing resilience measures "are all, at best, imperfect, and at worst, deeply flawed" (Upton, Constenla-Villoslada and Barrett, 2022). This study contributes to improving the conceptualization and the measurement of resilience. Conceptually, it builds on the notion of systemic resilience. Operationally, it explicitly models the spatial structure underlying human-system interactions and locational effects. To do so, a generalization of the conditional moments of well-being approach (Cissé and Barrett, 2018) is proposed. Our method has relatively low data requirements and is reasonably straightforward to implement. It largely preserves the desirable in-sample features of the parent method while significantly increasing the targeting accuracy. From a policy perspective, incorporating spatial elements into measuring resilience has the potential to improve resource allocation efficiency for development and humanitarian interventions. It also emphasizes the importance of synergizing community-based interventions to complement household-level ones in building resilience. [ABSTRACT FROM AUTHOR]