Pax Populi or Casus Belli? On the Conflict Resolution Potential of Self-Determination Referendums

Abstract

This dissertation investigates the conflict resolution potential of self-determination referendums. Over the past few decades, the use of referendums in the context of conflicts over secession and autonomy has proliferated remarkably. Self-determination referendums are also increasingly advocated by the international community, from Bosnia to Northern Ireland, East Timor, and South Sudan. However, very little is known about their ability to resolve conflicts over self-determination peacefully. Several perspectives can be found in the existing literature. While some see self-determination referendums positively, others see them as prone to incite violent conflict, and still others argue that self-determination referendums are likely to contribute to peaceful conflict resolution under some but not other conditions. To date, very little systematic empirical evidence exists to support either of these views. In this dissertation I develop an argument that the conflict resolution potential of self-determination referendums depends on whether their terms have previously been agreed by the two main parties to separatist conflicts, states and self-determination movements. I argue that mutually agreed self-determination referendums are likely to create a positive dynamic and increase chances for peace. Several reasons are made out, all generally related to the high legitimacy associated with agreed self-determination referendums. First, they are likely to foster perceptions of fair decision-making. Second, they may contribute to a reversal of hostile images. Third, they may lead to referendum-related coalitions that are willing to support their outcome. Fourth, they may sometimes push forward a peace process that would otherwise be blocked. And finally, they may increase the durability of settlements and alleviate commitment problems. By contrast, I argue that if self-determination referendums are unilaterally invoked by a state or a self-determination movement, they become more likely to inflame tensions than to reduce them. The legitimacy of unilateral self-determination referendums is often contested. Unilateral referendums are thus unlikely to have any of the beneficial consequences associated with agreed referendums. Rather, they are likely to entrench grievances, to generate reputation costs, and to reduce the bargaining range available for a negotiated settlement. Thus, unilateral self-determination referendums are likely to increase the risk of separatist armed conflict. The hypothesized effects of agreed and unilateral self-determination referendums are evaluated through a series of statistical tests. The main challenge this presents is the endogeneity of agreed and unilateral self-determination referendums to conflict processes. Finding agreement on a self-determination referendum often requires a substantial willingness to compromise, whereas this willingness is typically lacking in the case of unilateral self-determination referendums. Thus, while agreed referendums tend to emerge in rather peaceful and benign contexts, unilateral referendums tend to emerge in situations with an already significant ex-ante risk of separatist armed conflict. To counter the emanating threats to causal inference, I employ multiple regression in an effort to partial out the causal effects of agreed and unilateral self-determination referendums. The list of covariates is carefully assembled based on a separate analysis of the determinants of agreed and unilateral self-determination referendums. Relying on new data on self-determination referendums and noncolonial self-determination disputes in European and Asian countries, I find strong support for the argument that prior agreement on the terms of self-determination referendums shapes their conflict resolution potential. In line with expectations, I find that agreed self-determination referendums decrease the probability of new outbreaks of separatist armed conflict while increasing the probability that ongoing separatist armed conflicts come to an end. Also in line with expectations, I find that unilaterally initiated self-determination referendums increase the risk of new separatist armed conflicts and, where violence is already ongoing, the risk that separatist armed conflict continues. An extensive sensitivity analysis reveals that most results are robust to a great number of alternative measurement and specification choices, including fixed effects estimation, as well as to the possibility of hidden bias due to omitted confounders. The only partial exception emerges for the effect of agreed self-determination referendums on outbreaks of new separatist armed conflicts. Overall, the findings of this study suggest that self-determination referendums have value for conflict resolution, but only in situations where agreement can be reached between the key stakeholders on their terms.