This paper analyses the economic growth impact of organised political violence. First, we identify the various manifestations of political violence (riots, coups and civil war) and their risk of occurrence by using a multinomial model. Second, we use predicted probabilities of aggregate violence and its three manifestations to identify their growth effects in an encompassing growth model. The results of Generalised Method of Moments dynamic panel regressions suggest that organised political violence, especially civil war, significantly lowers long-term economic growth. Moreover, unlike most previous studies, we also find ethnic fractionalisation to have a negative and direct effect on growth, though its effect is substantially ameliorated by the institutions specific to a non-factional democratic society. Third, we find that Sub-Saharan Africa (SSA) has been disproportionately impacted by civil war, which explains a substantial share of its economic decline, including the widening income gap relative to East Asia. Civil wars have also been very costly for SSA. For the case of Sudan, a typical large African country experiencing a long-duration conflict, war cost amounts to $46 billion (in 2000 fixed prices), which is roughly double the country's current stock of external debt. Fourth, we suggest that to break free from its conflict-underdevelopment trap, Africa needs to better manage its ethnic diversity and the way to do it is to develop inclusive, non-factional democracy. A democratic but factional polity will not do the trick and is only marginally better than authoritarian regimes.