Numerous studies have demonstrated a negative relationship between civil war and national level per capita income. However, the specific theories or mechanisms that have been offered as
explanations for this relationship vary considerably, and many theoretical arguments offered – ranging from the role of state strength to opportunity costs for potential insurgents – are not
necessarily consistent with one another. Moreover, poverty and inequality are rarely uniformly distributed within countries, and conflicts are often confined to atypical areas of a country that may
not be adequately represented by national level averages or aggregates. We show that findings from existing studies at the country level are ambiguous and lack the ability to distinguish between different income distributions and rival interpretations. We consider propositions on the relationship between poverty, inequality, and conflict at the level of smaller geographical units entail different mechanisms and implications, and explore these hypotheses empirically using geographically disaggregated data on income and conflict.