Identifying within a group is an essential human need of belonging through which people often view their sense of self and how they perceive the rest of the world. Identity cleavages may therefore be considered an existential threat which can provoke strong responses, and therefore identity has frequently been at the core of research on inequality and conflict. However, group identity is not a universal concept, so its definition and social significance may vary across societies and time. What remains understudied in the literature is the impact of social salience of identity on conflict. This research seeks to analyze how the degree to which people identify within a group can affect conflict, as well as how religious diversity and socioeconomic divisions within a group can mediate the chance of ethnic violence. When tribal identity is important relative to national identity, a precarious political balance must be struck between ethnic groups where power-sharing is essential difficult to implement. Politics that stems from ethnic identity creates an environment that exacerbates the differences between ethnic identities and can increase intergroup distrust, which can eventually lead to conflict. A stronger sense of shared identity may also provide the potential for grievances to be experienced collectively. Therefore, for groups whose ethnic identity is more important relative to national identity or which are religiously homogeneous, we expect there to be a higher likelihood of conflict onset between groups, as groups that are ethnically homogeneous may be more inclined to mobilize than heterogeneous groups. Economic homogeneity within a group on the other hand is associated to opportunities of the group, and therefore is assumed to have a different effect. Groups that are economically homogeneous are less likely to enter into conflict because the costs for elites to provide the necessary material and/or purposive incentives to overcome collective action are higher than groups that are economically heterogeneous (groups with vertical inequality). To test these hypotheses, we conduct an empirical analysis on a dataset of sub-Saharan African countries in the period 1990-2014. National household data from the Demographic and Health Surveys (DHS) and the Afrobarometer public opinion surveys are used to compose indicators of socioeconomic inequalities between groups, religious diversity, vertical inequality and the salience of identity, as well as national context factors including the effects of regime change (coups, elections and other transitions) on the incidence of violence. Conflict data is taken from the Uppsala Conflict Data Program (UCDP) for 158 intrastate conflicts in that period. Multilevel regression is run on 5340 ethnic dyad year observations constructed from within-country combinations of 130 groups.
