The dissertation focuses on a specific form of forest decentralization, participatory forest management (PFM). The underlying premise of PFM is that sustainable forest management is most likely to occur when local communities manage local forests, and when they get access to direct benefits from participating in forest management. However, the outcomes of PFM have so far been reported as “mixed” and where success is reported, it usually relates to the forest condition rather than to improving local livelihoods. The key research questions in this PhD study are what have been the impacts of PFM on livelihood, forest governance and forest conditions in Ethiopia? The study approaches these questions by disentangling outcomes that can be attributed to PFM rather than other factors through quasi-experimental designs. The significance of the study lays in its holistic assessment of the theoretically expected outcomes of PFM. In the four articles that form the thesis, the study argues that the PFM programme in Ethiopia contributes to forest conservation compared to other types of management regimes. However, conservation is also challenged mainly by lack of support from the authorities to forest user groups. Though various pilot projects have contributed valuable experiences on the performance of PFM in Ethiopia, the programme is currently being scaled up to the national level without taking these into account. Indeed, the PFM up-scaling programme remains based on the discretion of the individual donors and NGOs, with a model where only subsistence level incentives are made available to forest user group members. The study confirms the theoretical claim that rules imposed from above are not followed, and uniquely shows that commercialization of timber and forest conservation can go side by side in decentralized forest management.