Several studies of civil war have concluded that economic inequality between individuals does not increase the risk of internal armed conflict. This is perhaps not so surprising. Even though an individual may feel frustrated if he is poor compared with other individuals in society, he will not start a rebellion on his own. Civil wars are organized group conflicts, not a matter of individuals randomly committing violence against each other. Hence, we should not neglect the group aspect of human well-being and conflict. Systematic inequalities that coincide with ethnic, religious, or geographical cleavages in a country are often referred to as horizontal inequalities (or inter-group inequalities). Case studies of particular countries as well as some statistical studies have found that such inequalities between identity groups tend to be associated with a higher risk of internal conflict. But the emergence of violent group mobilization in a country with sharp horizontal inequalities may depend on the characteristics of the political regime. For example, in an autocracy, grievances that stem from group inequalities are likely to be large and frequent, but state repression may prevent them from being openly expressed. This paper investigates the relationship between horizontal inequalities, political environment, and civil war in developing countries. Based on national survey data from 55 countries it calculates welfare inequalities between ethnic, religious, and regional groups for each country using indicators such as household assets and educational levels. All the inequality measures, particularly regional inequality, are positively associated with higher risks of conflict outbreak. And it seems that the conflict potential of regional inequality is stronger for pure democratic and intermediate regimes than for pure autocratic regimes. Institutional arrangements also seem to matter.