Community-based forestry (CBF) is often cited as a way for improving livelihood of local communities while conserving biodiversity and mitigating climate change. However, empirical assessments of CBF outcomes are still lacking, especially for their biophysical conditions. This study investigates the extent to which a type of CBF practices in Indonesia, Community Forest (HKM), managed to maintain forest cover. We applied a propensity scoring approach to empirically measure the rates of deforestation between 2007 and 2016 for Conservation Forests and Protection Forests with and without HKM concessions. Our finding is that HKMs are less effective than Conservation Forests (e.g. National Parks) in reducing forest cover loss, but more effective than other similar forests without CBF management. This is a promising starting point for expanding CBF in Indonesia. We recommend: 1) explicit consideration of biophysical characteristics of CBF locations in designating future sites; 2) providing site-specific financial and technical supports for local communities; 3) utilizing remotely sensed data and propensity scoring for monitoring conservation outcomes.