Detecting Disasters and Disaster Recovery in Southeast Asia: Findings from Space

Abstract

This paper builds on the group of studies that have used satellite night light data to measure the impact of natural disasters on economic development. It examines the dynamics of disasters using monthly subnational-level data for Southeast Asian countries covering the period 2012–2018. Four issues are considered. First, the paper determines the proportion of disasters that can be detected through an exogenous structural break in the night light luminosity data. Second, it identifies the factors that are associated with detection. Third, it uses the luminosity data to examine how consistent disaster recovery patterns are with theoretical scenarios. Finally, it identifies the factors associated with different recovery patterns. Findings suggest that the probability of disasters being detected using night light data is positively associated with disaster intensity. Recoveries from many, but not all, disasters do not result in a significant change in luminosity from its predicted trend. Although some disasters result in luminosity returning to a permanently lower growth path, it is more common for it to return to a higher-than-predicted trend.