Beyond the Snapshot: A Survival-Based Systems Framework for Measuring Resilient Impact in Global Development.

Development projects often fail to produce lasting change once external funding and support disappear. This is particularly visible in agricultural sectors where initial gains in productivity frequently evaporate within a few seasons. The persistence of this pattern suggests that our current evaluation methods are limited: we tend to measure success as a static outcome rather than a dynamic system property. This paper introduces Resilient Impact Systems Analysis (RISA), a framework designed to determine whether development impact can survive real-world pressures. Unlike traditional evaluations that focus on point-in-time estimates, RISA integrates four key dimensions: core causal effects, behavioural persistence, exposure to external risks, and the capacity of local institutions. The framework is built on a survival-based mathematical model. By applying exponential decay to adoption rates, RISA moves beyond simple snapshots to provide predictive insights. A central feature of this model is the introduction of the Impact Half-Life (T½ = ln(2)/k), which identifies the specific point in time when systemic friction and adoption decay reduce the initial benefits by half. This metric identifies the specific point in time when systemic friction and adoption decay reduce the initial benefits by half. This provides a clear benchmark for durability, allowing for more honest comparisons between different interventions. Through an illustrative application in agriculture using longitudinal adoption data and post-exit follow-up surveys, the paper shows how projects deemed successful by traditional metrics may be highly fragile. RISA provides a practical tool for designers and donors to move from short-term wins toward investments that are truly anchored in the systems they aim to change.