Search Climate Insights Directory

The study analyzes conflict rates in over 1,600 Cambodian communes to identify how war legacy, poverty, urbanity, and governance quality influence violence. Using semi-parametric Bayesian models, the authors find that many relationships are non-linear, with specific socio-economic thresholds triggering significant changes in conflict rates. Notably, governance improvements only significantly reduce violence at very high levels of service provision, and rapid population growth strongly correlates with increased domestic and land conflicts. Spatial effects are found to be the strongest predictors, suggesting that unobserved, spatially correlated factors—such as the demographic impact of genocide—play a critical role in human security.