Forecasting Civil Wars: Theory and Structure in an Age of "Big Data" and Machine Learning. (November 2020)

• Record Type: Journal Article
• Authors: Blair, Robert A.; Sambanis, Nicholas
• Is Part Of: Journal of conflict resolution. Volume 64:Number 10(2020)
• Page Start: 1885
• Page End: 1915
• Publication Date: 2020-11
• Subjects: forecasting -- civil wars -- event data -- machine learning; War and society -- Periodicals; Peace -- Periodicals; Arbitration (International law) -- Periodicals; 327.16
• DOI: 10.1177/0022002720918923
• Languages: English
• ISSNs: 0022-0027

Abstract

Does theory contribute to forecasting accuracy? We use event data to show that a parsimonious model grounded in prominent theories of conflict escalation can forecast civil war onset with high accuracy and over shorter temporal windows than has generally been possible. Our forecasting model draws on "procedural" variables, building on insights from the contentious politics literature. We show that a procedural model outperforms more inductive, atheoretical alternatives and also outperforms models based on countries' structural characteristics, which previously dominated models of civil war onset. We find that process can substitute for structure over short forecasting windows. We also find a more direct connection between theory and forecasting than is sometimes assumed, though we suggest that future researchers treat the value-added of theory for prediction not as an assumption but rather as a hypothesis to test.