The short-term dynamics of conflict-driven displacement: Bayesian modeling of disaggregated data from Somalia

Understanding the short-run dynamics of conflict and forced displacement is crucial for the design of effective policy responses, yet quantitative analyses in this realm are sparse. This is primarily due to the scarcity of high-frequency displacement data and methodological challenges arising when modeling imperfect data collected in conflict zones. Addressing both issues, we develop a Bayesian panel regression model to assess the short-term impact of conflict on displacement in Somalia, utilizing weekly panel data that encompasses eight million displacements and 19,000 conflict events from 2017 to 2023. Results suggest a rapid and nonlinear displacement response postconflict, with significant heterogeneity in effects dependent on the nature of conflict events. In a displacement forecasting exercise, our model outperforms standard benchmarks, underscoring its potential for informing decision-makers in crisis scenarios.