Slowing the reduction, or increasing the accumulation, of organic carbon stored in biomass and soils has been suggested as a potentially rapid and cost-effective method to reduce the rate of atmospheric carbon increase 1 . The costs of mitigating climate change by increasing ecosystem carbon relative to the baseline or business-as-usual scenario has been quantified in numerous studies, but results have been contradictory, as both methodological issues and substance differences cause variability 2 . Here we show, based on 77 standardized face-to-face interviews of local experts with the best possible knowledge of local land-use economics and sociopolitical context in ten landscapes around the globe, that the estimated cost of increasing ecosystem carbon varied vastly and was perceived to be 16-27 times cheaper in two Indonesian landscapes dominated by peatlands compared with the average of the eight other landscapes. Hence, if reducing emissions from deforestation and forest degradation (REDD+) and other land-use mitigation efforts are to be distributed evenly across forested countries, for example, for the sake of international equity, their overall effectiveness would be dramatically lower than for a cost-minimizing distribution.