The thesis focuses on local level responses to recurrent small disasters and conflict in Afghanistan, Indonesia and the Philippines. It critically reflects on Community-Based Disaster Risk Reduction (CBDRR) approaches to understand the gap between CBDRR policy and actual outcomes. It considers the multi-level institutions through with meaning and implementation of CBDRR policy are negotiated and transformed, from the conceptual policy design stage until the arena where decisions on risk solutions and resource allocation are made. Disasters and conflict are both understood as the product of a cumulative set of institutional arrangements and policy decisions over a long period of time. Vice versa, disasters and conflict affect institutional arrangements and re-order power relations. Interventions like CBDRR are not isolated, distinct entities, but are very much embedded in a context of particular institutional arrangements, which constrain or enable local actors to advance their risk-solutions. Through CBDRR interventions actors defend and mobilize around CBDRR practices that are meaningful to them, or resist institutions and practices that carry meanings they find disagreeable. This results in the manifold manifestations of CBDRR practices and outcomes. The research concludes that there is no such thing as the CBDRR approach. Instead, there are different processes through which local NGOs, civil society organizations, funding agencies and government agencies arrive at a specific framing of local realities and their responses in the context they live and work. These are related to their histories, current state - civil society relationships, and their mandate on how they legitimize their interventions. These actors either underscore the politics of their interventions or rather de-politicize them. From the experiences of this research it is plausible to conclude that when one ignores to view CBDRR interventions in a political and institutional manner, the outcomes of the interventions are likely to reproduce the status quo and are not supporting the vulnerable populations. The implication for humanitarian aid agencies is to include an institutional and political analysis in risk and vulnerability assessments to explain people’ vulnerability. This is crucial for strategizing actions and to engage in the political arena of disaster risk reduction with the aim to create safe and resilient communities. Rather than simply aiming for isolated village-level project objectives, CBDRR interventions have to think ahead of results to be achieved at district and even national level.