Power sharing institutions are currently being established to end war and lay the foundations for peace and democracy in a number of states emerging from civil war, such as Burundi, Rwanda, Sudan and Côte d’Ivoire. While a number of authors have contributed to develop theory on power sharing and have drawn attention to the strengths and weaknesses of such systems of governance (Lijphart, 1990; Lake and Rotchild, 1996; Snyder and Jervis, 1999; Paris, 2004 etc.), little empirical evidence has been gathered on how power sharing works on the ground. This thesis therefore sets out to assess power sharing as a strategy to manage conflict in a post-war state. The analysis takes a closer look at a state that is currently implementing a recent power sharing agreement, namely Burundi. Burundi has been chosen as a case study because it provides new insights on several levels, both theoretical and empirical. Firstly, Burundi is currently making its second attempt at power sharing, which allows for making a comparative case study between two power sharing agreements within a single context. Secondly, Burundi’s second power sharing agreement combines two different models of power sharing so far seen as opposites in the literature; the consociational model and the integrative model. This will improve our understanding of the consequences of mixing the two models. Finally, Burundi is an example of power sharing in a neo-patrimonial state. Whereas the mainstream analysis in the literature has been centred on power sharing in weak and failed states, the case study of Burundi will shift the focus to how power sharing fares in a neo-patrimonial state.