Contemporary conflict research usually measures the influence of ethnicity on conflict by capturing ethnic constellations as country-based indices, such as ethnic fractionalization or polarization. However, such aggregated measures are likely to conceal the actual operation of actor-specific mechanisms. In this article, therefore, we introduce a disaggregated model that measures ethnic groups’ access to power. We do so by disaggregating both ethnicity and conflict to the level of explicitly geo-coded center-periphery dyads. This procedure allows us to measure the power balance between politically excluded ethnic groups and dominant actors in terms of group sizes, distances between the center and the periphery, and the roughness of the latter’s terrain. We rely on geographic information systems (GIS) to compute demographic and ethno-geographic variables. The dyadic analysis enables us to show that exclusion of powerful ethnic minorities increases the likelihood of conflict considerably. In addition, we show that the risk of conflict is positively associated with the extent of rough terrain in the peripheral group’s home region and its distance from the political center.