Philosophiae Doctor - PhD / Aid coordination in the health sector is known to be challenging in general, but even more in post-conflict settings, due to the multiplicity of actors of development, to the sense of emergency in providing health services, combined with the so-called weak institutional capacities‘ at local level, resulting from the conflict. This study sought to analyze broad determinants of aid coordination using the example of HRH policies in Burundi, during the post-conflict period. Burundi is a country in Central Africa, which experienced cyclic ethnic conflicts since its independence in 1962, the last conflict being the longest (1993-2006).Determinants of coordination were analyzed using the policy-analysis triangle (Gilson et Walt), using data from documents and semi-structured interviews, conducted in 2009 and in 2011, at national, provincial and facility-levels. A conceptual framework, combining organizational and social sciences theories, was devised in order to assess the organizational power of MoH, the one supposed to act as coordinator in the health sector. Findings showed a lack of coordination due to post-conflict specific context, to competition over scarce resources between both donor and recipient organizations and to an insufficiently incentivized and complex coordination process in practical. Most importantly, this research demonstrated the crucial role of post-conflict habitus and mistrust in the behavior of MoH and their influence on organizational power, and, in turn on their capacity to coordinate and exert an appropriate leadership. These findings, together with the growing body of literature on organizational sociology and collective trust, point at the crucial need to rebuild some of the wounded collective trust and organizational leadership in Burundi and in other fragile states.

This dissertation examines the following research question: Which types of electoral rules chosen in post-conflict states best promote peace? And are those effects conditional upon other factors? I argue that the effects are conditional upon the types of political parties that exist in the post-conflict environment. Although this explanation is contrary to scholars that speak of political parties as products of the electoral system, political parties often predate the choice of electoral system. Especially in post-conflict states, political parties play an important role in the negotiation process and hence in the design of the electoral rules. I argue that the effects of electoral rules on peace duration are mitigated by the degree to which a party system is broad (nonexclusive) or narrow (exclusive). I develop a theoretical model that led to three hypotheses focusing on the independent role that political parties play in mitigating the effects of electoral rules on peace duration. To test these hypotheses, I use the Cox proportional hazard model on 57 post-conflict states from 1990 to 2009 and had competitive elections. The empirical results show support for the main argument of this study. First, the findings show that electoral rules alone do not increase or decrease the risk of civil war outbreak, yet when interacting with the degree to which political parties are broad or narrow, there is a significant effect on the outbreak of civil war. Second, the results show that post-conflict states with party centered electoral systems (closed list PR system) are less likely to have an outbreak of civil war when more seats in the parliament are controlled by broad-based parties. In addition, I conduct a comparative case study analysis of two post-conflict states, Angola (1975-1992) and Mozambique (1975-1994), using the most similar systems (MSS) research design.