Identifying Spatiotemporal Effects of Climate on Civil Conflict
Authors
LIN, YIJYUN
Issue Date
2020
Type
Dissertation
Language
Keywords
climate change , conflict , democracy , machine learning , vulnerability
Alternative Title
Abstract
How do changes in climatic conditions and disaster patterns affect the persistence of civil unrest across countries over time? Existing studies postulate that changing climate conditions will exacerbate various social conflicts through their impacts on degraded environmental and economic conditions, which is further conditional on political institutions. Nevertheless, there are two major pitfalls in the existing studies. First, vulnerability as a major underlying mechanism has been used as an umbrella term or been presumed. Using vulnerability as an umbrella term has a detrimental effect on climate-conflict theory-building because it prevents scholars from deriving testable empirical implications for relevant concepts. Second, previous research has pinpointed the importance of political institutions in moderating impacts of climate on conflict, but the literature says little about what aspects of political institutions might aggravate or alleviate vulnerability to climate in ways that are simmering or amplifying civil strife. Using the structural causal approach and machine learning methods, this dissertation improves the identification of the mediation effect of vulnerability and the moderation effect of political institutions on the climate-conflict relationship. The important mechanisms and implications revealed by this study are twofold. First, this dissertation finds that the impacts of extreme climatic events are more important in shaping local vulnerability than that of annual weather variations, and that adaptive capacity is more important than economic sensitivity in mitigating local vulnerability. Annual weather variations (i.e., the slow-moving mechanics) have a significant impact on cumulative conflict hazards, whereas extreme climatic events (i.e., the fast-moving drivers) fuel onset of a new conflict. In the presence of socio-psychological vulnerabilities, an increase in annual weather variations can boil new conflicts. Second, the state capacity is more important than democracy in exacerbating a country’s vulnerability to climate, and the degree of executive bribery especially plays a crucial role in moderating the impacts of vulnerability to climate on civil conflict. However, of different aspects of democracy, freedom of academic and cultural expression has the most important moderating effect on conflict. What is striking is the role of socio-psychological vulnerability in transmitting the impacts of extreme climate and weather variations on civil conflict. Mainstream conflict theory has shown that institutional and economic conditions are the most important factors determining conflicts even though socio-psychological factors are meaningful contexts. However, this present study shows that socio-psychological vulnerability is more important than institutional and economic conditions in shaping civil conflict.