This paper examines the link between absolute and relative poverty and the location of civil war events. Drawing on the ACLED dataset, which breaks internal conflicts down to individual events at the local level, this paper takes a disaggregated approach to the study of conflict. The conflict data are linked with geographically referenced socioeconomic data from the Liberian Demographic and Health Survey (DHS) conducted in 1986. With geographical units (grid cells) of approximately 76 km2 as the units of analysis, we test how both absolute and relative welfare levels affect the presence and number of conflict events in
Liberia during the civil war in 1989–2002. We include a number of geographically disaggregated control variables, such as neighboring conflict events, distance to national
borders, distance to the capital, population density, diamond deposit, and ethnic affiliations. Negative binomial regression results indicate that civil war events predominantly happen in the relatively richer provinces. We argue that this finding reflects the relative strength of the rebel group(s).