Statistical models of civil war onset are often unsupportive of a link between measures of cultural demography and conflict. This study suggests that this is in part because most studies fail to account for what factors make demographic cleavages salient, such as policies of exclusion and repression against growing minorities that are threatening to incumbent regimes. A comparison of Côte d’Ivoire and Ghana is used to shed light on this process. Based on a state of the art statistical model of civil war onset, the countries had strikingly similar conflict risk in the early 2000s, but conflict only erupted only in Côte d’Ivoire in 2002. An important factor to explain this is the exclusion and repression in the Ivorian case, spurred by a perceived increase in the northern Muslim population, vs the more accommodative policy in neighboring Ghana. Implementing lessons from this study could improve future statistical models of civil war.