Assessing the Impacts of Armed Conflict and Climate-Related Disasters on Vulnerability: a Machine Learning Approach

Abstract:

Armed conflicts have been associated with a variety of detrimental impacts on human security and development, and represent a crucial vector of vulnerability to climate hazards. The burgeoning literature on climate security has highlighted that climate hazards may indirectly increase conflict risk in vulnerable locations. However, solid knowledge of the role of armed conflicts in shaping vulnerability to climate hazards, especially in complex crises where conflicts and climate-related disasters compound, is still lacking. This study fills the gap by using a Machine Learning framework to study how armed conflicts, climate-related disasters, and the combination thereof, impact countries’ vulnerability to subsequent climate hazards. The paper uses global, time-varying data on climate-related disasters, armed conflict, and vulnerability for 189 countries between 1995 and 2019, combining information. We find that the combination of armed conflicts and climate-related disasters is associated with increased predicted vulnerability to climate hazards, and that these impacts are time persistent, non-linear and extend beyond strictly economic losses.

Authors:

Mariagrazia D’Angeli and Paola Vesco