There is a consensus that civil wars entail enormous economic costs, but reliable estimates are lacking, due to the endogenous relationship between violence and socio-economic conditions. This paper measures the economic consequences of civil wars with the synthetic control method, identifying appropriate counterfactuals for assessing the national-level economic impact of civil war in a sample of 20 countries.
The study finds that the average annual loss of GDP per capita is 17.5 percent. The estimates of annual losses are used to study the determinants of war destructiveness, focusing on the effects of ethnic heterogeneity. Building on an emerging literature on the relationships between ethnicity, trust, economic outcomes, and conflict, the paper argues that civil war erodes interethnic trust, and highly fractionalized societies pay an especially high “price” as they rely heavily on interethnic business relations. It also finds a consistent positive effect of ethnic fractionalization economic war-induced loss.
- Civil wars kill, main, and destroy on a large scale. While this basic assertion is beyond dispute, reliably estimating the economic costs of civil war has been challenging due to omitted variable and reverse causality problems. As war-torn countries are inherently different from peaceful ones and civil war violence is both a cause and a consequence of dismal economic conditions, comparisons between conflict-ridden countries and peaceful neighbors (or countries at comparable levels of development) and between pre-war and wartime economic performance are not necessarily useful.
- By adopting the synthetic control method, the study explicitly address the counterfactual question of what a country’s GDP per capita would have been had it not experienced armed conflict. This method allows for the construction of artificial comparison units for each war-torn country and thus obtain more reliable counterfactuals than any real-world country. This enables the analysis of a considerable number of cases with consistent criteria and statistical rigor, while also paying attention to the heterogeneity of war’s impact in each case.
- Estimates of GDP loss are used to explore the determinants of the significant variation in war’s destructiveness across the 20 countries in the sample, focusing on the consequences of ethnic heterogeneity. Building on recent findings on the relationships between ethnicity, trust, economic outcomes, and violent conflict, the paper finds that civil war erodes interethnic trust and that highly fractionalized societies experience especially serious losses, as they rely heavily on interethnic business relations.
- Civil wars exact a heavy toll on all affected countries, but for some the economic effects are especially dire. A nuanced understanding of this variation and its causes represents an important stepping stone to effective conflict prevention and post-conflict development policies.