This paper investigates the empirical role of violent conflicts for the causal effect of democracy on economic growth. Exploiting within-country variation to identify the effect of democratization during the “Third Wave”, we find evidence that the effect of democratization is weaker than reported previously once one accounts for the incidence of conflict, while the incidence of conflict itself significantly reduces growth. The results show in turn that permanent democratic transitions significantly reduce the incidence and onset of conflict, which suggests that part of the positive growth effect of democratization arises because democratization reduces conflict incidence. When accounting for the role of violence during democratization, we find evidence that peaceful transitions to democracy have a significant positive effect on growth that is even larger than reported in the previous literature, while violent transitions to democracy have no, or even negative, effects on economic growth.