Since its inception over ten years ago, REDD+ (Reducing Emissions from Deforestation and forest Degradation plus enhancing carbon stocks) has been piloted through more than 300 subnational initiatives across the tropics. Initiative implementers are applying REDD+ intervention packages that in customized ways combine disincentives (e.g. restrictions on forest access or conversion) and incentives (conditional or non-conditional) to allegedly achieve better protection of forests. Through CIFOR’s Global Comparative Study on REDD+, we evaluated the forest conservation and community well-being impacts of 22 subnational REDD+ initiatives in Brazil, Peru, Cameroon, Tanzania, Indonesia and Vietnam. We combined socioeconomic surveys in 150 communities and nearly 4,000 households (including control groups) in 2010-2012 (pre-intervention) and 2013-2014 (post-intervention), with an analysis of Global Forest Change data (2000-2015), to assess how different interventions affected changes in income, assets, perceived well-being and forest cover at the community level. We found a reduction in forest cover loss at half of the REDD+ sites, with better conservation outcomes in places where there was higher treatment intensity of disincentives. Both intervention types had neutral impacts on income and assets, but higher exposure to disincentives led to decreases in overall perceived well-being, except when balanced out by incentives. Although disincentives may be an effective REDD+ instrument for conserving forests, other measures are clearly needed to safeguard and enhance community well-being. This critically needed empirical evidence on REDD+ performance on the ground is relevant to policy makers and practitioners interested in developing REDD+ strategies that can provide both conservation and livelihood benefits.