Forecasting Armed Conflict in a Complex System of Instability

• In Event: What We Can Know about the Future of War and Its Humanitarian Impacts?

Abstract

Models of armed conflict with features built from past and present armed conflict produce accurate forecasts at disaggregate spatial and temporal levels. Yet, armed conflict does not exist in isolation. Rather, it affects and is affected by many other factors. Jointly, these components form a system with strong dependencies and complex relationships. This research presents a novel approach to modeling this type of system with a goal to improve forecasts of armed conflict and other destabilizing factors, specifically displacement and human movement, food and water insecurity, and social protest. First, each component of the system is modeled in a way to forecast that individual component. These models, by themselves, are similar to conflict forecasting models in the existing literature. Next, they are combined to jointly forecast each component in the system. Results show differences in improvements for forecasts of individual components, with varied relations among them. The goal of this research is to improve forecasts and understanding of these sources of instability by modeling the complex system within which they exist. Furthermore, this is a step toward modeling how other societal threats, such as climate change and extreme weather events, will impact instability.

Author

• Vito D'Orazio, West Virginia University