This thesis tries to explore the evolution of the political violence in Colombia in light of consociational theory. From 1958 – 1978 1948 the Colombian Liberal and Conservative elites shared all political power equally between then in an effort to restore stability, and initially they were highly successful. However, the exclusionary nature of such regimes hindered other political expressions to be voiced, and portions of the population were effectively excluded from all political institutions. The excluded sectors of society were organised in several guerrilla organizations, and by the late eighties the political violence reached unprecedented levels. Several of the guerrillas entered into peace-negotiation which resulted in the rewriting of the Colombian constitution. The new Constitution of 1991 sought to bring an end to the violence by changing the institutional democracy from being an exclusionary consociational system to an inclusive and participatory adversarial system. In the thesis I ask why the violence increased after the new political system was introduced, and analyse the changes in light of the four characteristics of the consosicational democracy; executive coalitions, proportionality, mutual vet, and segmental autonomy. My finding were that most of the institutional changes in relational to the first three characteristics had little if any effect. However, segmental autonomy has never been part of the Colombia consociational democracy, and hen this was introduced through a process of political, administrative, and fiscal decentralization. I argue that the increase in violence can best be understood as a consequence of introducing the dimension of segmental autonomy, and that Lijphart’s only viable solution for divided societies has proven not to be viable for Colombia.