Natural Disaster and Conflict Dynamics in Civil War

With ongoing climate change and rising frequency of extreme weather events, understanding the impact of natural disasters on conflict becomes increasingly urgent. This thesis explores under what conditions and how natural disasters affect conflict dynamics during civil war. Across three papers, I argue that the impact of natural disasters is not unconditional but conditional on the strategic environment of civil war, which changes the vulnerabilities, incentives, and expectations of armed actors. I develop theory and empirical evidence for three strategic environments: (1) the spatial configuration of territorial control, (2) the role of third-party commercial actors, and (3) armed actors’ anticipation of future disasters. This thesis contributes to the climate–conflict literature by demonstrating conditions under which natural disasters affect conflict dynamics. It also contributes to broader theories of conflict escalation and de-escalation, particularly debates on the impact of exogenous negative shocks and commitment problems on conflict. Furthermore, it introduces a new typology of spatial configurations of territorial control, develops a new measure of rainfall-related disasters, and includes a previously under-explored actor – a third-party commercial actor – into the study of civil war under sudden and negative shocks. It also develops a new empirical design to estimate the causal impact of disaster anticipation on conflict.

In the first paper, The Impact of Natural Disasters on Ongoing Civil War: The Role of Spatial Configuration of Territorial Control, I examine how the pre-disaster territorial configurations moderate post-disaster conflict dynamics by introducing a typology of spatial configurations of territorial control and developing a satellite-based, rainfall-related disaster measure. Using the ongoing civil war between the New People’s Army (NPA) and the Philippine government as an empirical context, I demonstrate theoretically and empirically that, depending on the type of pre-disaster territorial configurations, disasters trigger different armed actors’ strategies and thus lead to different outcomes regarding shifts of territorial control and battle-related violence. The second paper, Buying Stability After Sudden and Negative Shocks: How Extractive Companies Dampen Post-Disaster Conflict, analyzes how third-party extractive companies moderate post-disaster conflict dynamics. Using fine-grained geospatial data on different types of natural disasters and resource extraction sites in Myanmar, I show that the presence of companies operating capital-intensive resource extraction sites dampens post-disaster violence, motivated by their incentives to stabilize local conditions in order to restore operations and revenue quickly. In support of my argument, I also demonstrate empirically that this mitigating effect fades as extraction licenses approach expiration. In the third paper, Anticipation, Commitment Problems and Civil War Violence: Evidence from Normal vs. Flash Droughts, coauthored with Dr. Mahmoud Osman (NASA), we examine the impact of anticipation of upcoming droughts on conflict, using the fine-grained geospatial data on droughts and conflict across Sub-Saharan Africa. We develop a credible design that isolates the causal effect of anticipation by contrasting two climatic shocks with comparable agro-economic impact but different levels of predictability: anticipated normal droughts and unanticipated flash droughts. The empirical results support the anticipation-based commitment problem by revealing that violence increases before normal drought but not before flash drought. Taken together, this thesis highlights the conditionality of the impact of natural disasters and has actionable policy implications for anticipating where and when increasingly extreme weather events are most likely to have destabilizing effects.