We test the longstanding hypothesis that ethnic groups organized around “segmentary lineages” are more prone to conflict. Ethnographic accounts suggest that in such societies, which are characterized by strong allegiances to distant relatives, individuals are obligated to come to the aid of fellow lineage members when they become involved in conflicts. As a consequence, small disagreements often escalate into larger‐scale conflicts involving many individuals. We test for a link between segmentary lineage organization and conflict across ethnic groups in sub‐Saharan Africa. Using a number of estimation strategies, including a regression discontinuity design at ethnic boundaries, we find that segmentary lineage societies experience more conflicts, and particularly ones that are retaliatory, long in duration, and large in scale.

Lack of trust in state institutions, often due to poor service provision, is a pervasive problem in many developing countries. If this increases reliance on non-state actors for crucial services, the resulting self-reinforcing cycle can further weaken the state. This paper examines whether such a cycle can be disrupted. We focus on dispute resolution in rural Punjab, Pakistan. We find that providing information about reduced delays in state courts leads to citizens reporting higher willingness to use state courts and to greater fund allocations to the state in two lab-in-the-field games designed to measure trust in state and non-state actors in a high-stakes setting. More interestingly, we find indirect effects on non-state actors. After receiving state positive information, respondents report lower likelihood of using non-state institutions and reduce funds allocated to them in field games. Furthermore, we find similar direct and indirect effects on a battery of questions concerning people's beliefs about these actors, including a decreased allegiance to the non-state actor. We rationalize these results with a model of motivated reasoning whereby reduced usage of non-state institutions makes people less likely to hold positive views about them. These results indicate that, despite substantial distrust of the state in Pakistan, credible new information can change beliefs and behavior. The feedback loop between state ineffectiveness and the legitimacy of non-state actors may be reversible.