Different arrangements of decentralized forest management have been promoted as alternatives to centralized and top down approaches to halt tropical deforestation and forest degradation. Ethiopia is one of the countries piloting one of these approaches. To inform future programs and projects it is essential to learn from existing pilots and experiences. This paper analyses five of the pilot participatory forest management (PFM) programs undertaken in Ethiopia. The study is based on the Forest User Group (FUG) members' analyses of the programs using selected outcome variables: forest income, change in forest conditions, forest ownership feelings and effectiveness of FUGs as forest managing institutions. These variables were assessed at three points in time-before the introduction of PFM, during the project implementation and after the projects ended. Data were collected using group discussions, key informant interviews and transect walks through the PFM forests. The results show that in all of the five cases the state of the forest is perceived to have improved with the introduction of PFM, and in four of the cases the improvement was maintained after projects ended. Regulated access to the forests following introduction of PFM was not perceived to have affected forest income negatively. There are, however, serious concerns about the institutional effectiveness of the FUGs after projects ended, and this may affect the success of the PFM approach in the longer term.