Civil wars are intricate social, political and psychological phenomena. However, economics can offer analytical insights which are useful alongside the more conventional approach of case-studies. Indeed, the policy conclusions drawn from economic analysis sometimes cast doubt on conventional advice. The use of economic theory and statistical evidence help to guard against excessive generalization from individual civil wars that inevitably suffer from both a surfeit of possible explanations and advocacy. Rigorous empirical study of civil war requires a precise definition of an imprecise and poorly observed phenomenon, a process that provides considerable room for legitimate disagreement. Hence, we begin by discussing the choices made in constructing the major data sets that describe the duration and severity of civil wars. Ideological, religious or ethnic differences are conventionally regarded as the causes of civil war. Economic theory explains civil war in the framework of incentives and constraints rather than ideologies or identities. This framework enables economists to analyse the distinctive feature of civil war: the emergence and persistence of a rebel army: some conditions make rebellion both more attractive and more feasible than others. Consistent with this emphasis on incentives and constraints, statistical studies suggest that economic characteristics, notably the level, growth and structure of income, are important influences on the risk of war. In addition to the explanation of the initiation and duration of civil wars, economic methods can also generate estimates of their costs and consequences. This is an essential step towards the cost-benefit analysis of policy interventions.