Early studies of the grievance-conflict nexus suggested that economic grievances make civil conflict more likely. Follow-up empirical studies that employed objective measures of grievance at the national level often failed to find a relationship between grievances and conflict, while more recent studies that disaggregate these measures to the subnational level often find support for a grievance-conflict relationship. In this paper, we argue that regardless of the level of analysis of previous work, nearly all suffer from the limitation that they employ objective indicators as measures of grievance, while the theoretical arguments for a grievance-conflict relationship require that economic hardship be subjectively perceived by actors on the ground. Bringing together subnationally aggregated public opinion data from cross-national surveys with a variety of demographic and other population-level information, we use advances in multilevel regression with poststratification to construct aggregate measures of subjective grievances among the populations of subnational units. Employing these measures, we show that subjective measures of grievance predict conflict in the way suggested by previous subnational researchers, largely washing away the effects of objective measures when both are considered simultaneously. This research provides a starting point and an exemplary framework for the study of factors related to conflict that are rooted in actors’ subjective perceptions of the political, economic, and social context.