Ballot box and tinder box : can electoral engineering save multiethnic democracy?

Abstract

The objective of this dissertation is to systematize the existing hypotheses in the electoral engineering literature and to test them in a set of selected case studies in order to answer a central question: does the electoral system affect the structure of political parties in ethnically divided societies and if so how? The academic debate on electoral design for divided societies has focused on the impact of institutional choices on ultimate conflict outcomes. The findings of previous studies have been generally inconclusive, while the lack of sub-national data on ethnic composition and voting patterns has made it difficult to examine mechanisms regarding the role of demographics. To approach the problem from a different angle, I propose a research design focusing on the intermediate link from electoral institutions to the ethnic structure of the party system. For the empirical portion of my work, I chose to conduct a structured historical comparison of four societies which implemented major electoral reforms: Turkey, Northern Ireland, Guyana, and Sri Lanka. Based on the study of these cases, I am arguing that politicians and voters have not responded to electoral incentives in the ways predicted by existent theories, and that no clear relationship can be observed between the electoral system's proportionality, the heterogeneity of electoral constituencies, and the number of parties or the types of ethnic appeals they make to voters. These findings indicate that the hopes placed in electoral system design for divided societies are unwarranted and that attention among political scientists and policymakers should shift to other peace-building approaches.