Why are some countries more prone to social violence than others? Despite the fact that annual deaths due to homicides worldwide outnumber those due to organized armed conflict by a factor of roughly 3 to 1, this question has received very little attention from conflict and development specialists in recent years. As a modest first step in addressing this gap in the literature we draw together insights from the conflict and criminology literatures to develop a model of social violence that accounts for both political-institutional and socio-economic factors. While there is an extensive literature on the socio-economic determinants of social violence, there are only a handful of studies that consider the significance of political-institutional arrangements. Using cross-country estimates of homicides produced by the World Health Organization as an indicator of social violence, we test our model using OLS regression analysis for a sample of more than 120 countries. We find that countries with ‘hybrid’ political orders experience higher rates of social violence than those with strong autocratic or strong democratic regimes, and that weakly institutionalized democracies are particularly violent. We also find robust associations between indicators of poverty, inequality and ethnic diversity and social violence. These results indicate that social and political violence share some common underlying causes. We conclude by suggesting that the apparent global decline in organized armed conflict and the concomitant rise in social violence in recent decades may be linked to world urbanization and the ‘third wave’ of democratization in the global South, although further research is required to confirm this hypothesis.