Spurious Regressions and Panel IV Estimation: Revisiting the Causes of Conflict

Abstract

Development and humanitarian agencies have rapidly embraced the concept of resilience since the 2008 global financial and food price crises. We report the results of a formal scoping review of the literature on development resilience over the ensuing period. The review identifies the theoretical and methodological underpinnings and empirical applications of resilience as the concept has been applied to individual or household well-being in low-and middle-income countries. From 9,558 search records spanning 2008-20, 301 studies met our pre-registered inclusion criteria. Among these, we identify three broad conceptualizations employed – resilience as capacity, as a normative condition, or as return to equilibrium – and explain how the resulting variation in framing leads to marked differences in empirical methods and findings. We study in greater depth a set of 45 studies that met five key criteria for empirical studies of resilience. The larger, more established, qualitative empirical literature yields insights suggestive that the concept of resilience can add value. The quantitative literature is thinner and divided among methods that limit cross-study comparability of findings. Overall, we find that development resilience remains inconsistently theorized and reliant on methods that have not been adequately reconciled to identify which tools are best suited to which questions. Despite much published evidence, most findings concentrate on just a few countries and natural shocks, and rely on cross-sectional data at just one scale of analysis. The result is a dearth of generalizable evidence, especially of rigorous impact evaluations, to guide whether or how agencies might build resilience among target populations.