We study the relationship between inequality and inter-groups conflicts (riots), focussing on social inequality. Disadvantaged societal groups experience discrimination and thus have limited access to some social and labour resources like education or employment. First, we experimentally investigate whether social inequality is a driving force of inter-group conflicts. Second, we investigate the factors that make preferences for riot translate into actions. Riots require coordination. Our experiment consists of a two-stage game. First, subjects play a proportional rent-seeking game to share a prize. Social inequality is modelled exogenously by attributing to some subjects (the advantaged group) a larger share of the price than other subjects (the disadvantaged group) for the same amount of effort. In a second stage players can coordinate with the other members of their group to reduce (“burn”) the other group members' payoff. Treatments differ in the degree of social inequality set between the two groups. We observe frequent social conflicts, where, as expected, disadvantaged groups riot more than advantaged groups. Surprisingly, the frequency of riots decreases with the degree of inequality. A control treatment allows us to identify resignation as the driving force behind this phenomenon.