The Logics of Multi-Ethnic Coalitions : Power-Sharing, Party Systems, and Ethnic Conflict Management

Multi-ethnic coalition governments are a common prescription for societies affected by ethnic conflict. Yet ethnic power-sharing’s ability to deliver sustainable peace diverges widely between cases and across time. This dissertation asks: What explains variations in power-sharing’s ability to manage ethnic conflict? Building on pragmatist practice theory, the five essays that compose the dissertation develop an explanation that integrates and extends existing approaches – which focused on institutional design or power configurations – by shifting attention to what actors actually do when they share power. The dissertation argues that the distinct logics of ethnic conflict management that inform multi-ethnic coalitions explain their varying capacity to sustain peace. It introduces a previously underspecified type of multi-ethnic coalition: associational power-sharing, defined as multi-ethnic coalition governments in cross-ethnic party systems. Associational power-sharing manages ethnic conflict by reorienting political alliances away from ethnicity altogether, thereby reducing the risk of ethnic conflict through party-system-level guarantees against political exclusion. This logic of association contrasts with two other coalitional logics, each entailing distinct obligations, opportunities, and challenges: consociational power-sharing’s logic of pillarisation, which manages ethnic conflict by providing autonomous ethnic representation in a post-electoral coalition government; and centripetal power-sharing’s logic of moderation, which manages ethnic conflict by reinforcing ethnic moderates against extremists. Drawing on quantitative and qualitative evidence – including a novel dataset of multi-ethnic coalitions in Africa (1990–2020), comparative case studies of Northern Ireland and Burundi, and archival research and interviews in Mauritius – the dissertation shows that associational power-sharing is the most frequent type of ethnic power-sharing in Africa, is associated with longer periods of peace than its counterparts, and significantly decreases the probability of ethnic conflict onset. A key challenge, however, is that associational power-sharing is prone to tokenistic minority representation. The dissertation advances knowledge on constitutional design for ethnically divided societies by demonstrating that what matters is not only whether, but also how ethnic categories are included in government. It reframes the long-standing debate between consociationalism and centripetalism by recasting power-sharing types as demanding practices rather than rival models. It also highlights the central role of party system structure in stabilising ethnically divided societies. Overall, the dissertation adds associational power-sharing to the menu of ethnic conflict management practices available to conflict parties, peace mediators, and constitution-builders.