The authors argue that the academic literature, both qualitative and quantitative, has mislabeled most episodes of large-scale violence in Africa as civil war; these episodes better fit their concept of regional war complexes. The paper seeks to highlight the fundamental flaws in the conception of civil war in the econometric literature and their implications for econometric specification and estimation, problems that this literature is inherently incapable of rectifying. The authors advocate the comparative study of regional war complexes in Africa based on historical narratives.