Disaster Shocks, Forced Displacement and Banking Stability

This study examines how natural disaster severity and disaster-related internal displacement are associated with banking stability in an annual country panel covering 2008-2021. EM-DAT records are linked to IDMC displacement data, World Bank banking indicators and macroeconomic and governance controls. Country and year fixed effects are estimated with standard errors clustered by country. The main Z-score sample contains 825 country-year observations from 114 countries, including years with no recorded natural disaster. Reported disaster mortality has no statistically discernible standalone association with the Z-score. Its interaction with displacement is negative in the level specification and remains negative when the Z-score is expressed in logarithms. Among five alternative disaster measures, only the mortality interaction in the level model remains significant after Holm adjustment. The interaction weakens when the sample is restricted to years with complete mortality reporting. Weather-related and flood-only specifications do not reproduce the main association. A representative nonperforming-loan model yields no statistically discernible relationship. The evidence suggests a conditional link between reported mortality, displacement and banking buffers, with conclusions limited by reporting gaps, sample selection and the observational design.