The effectiveness of land-sharing and land-sparing approaches has been widely debated. Yet, few studies quantify the environmental and socio-economic outcomes of these approaches within a real-world landscape. Indonesia's plans to increase its palm oil production present an opportunity to investigate the potential environmental and socio-economic implications of the land-sharing and sparing approaches. We developed a computer model to simulate the expansion of oil palm agriculture in Sumatra, Indonesia, under four different scenarios distinguishable by the dominance of scheme smallholders or industrial estates: business-as-usual, BAU (25 : 75, scheme smallholders:industrial estates); high-yielding industry dominated, ESTATE (10 : 90); low-yielding smallholder dominated, SMALLHOLDER (40 : 60), high-yielding smallholder dominated, HYBRID (40 : 60; but with improved smallholder yields). Our results reveal several trade-offs associated with varying the proportion of scheme smallholders and productivity of oil palm plantations. The ESTATE scenario (reflecting land-sparing) resulted in lowest environmental costs in terms of forest conversion, greenhouse gas emissions, biodiversity losses and nitrogen fertilizer usage. Additionally, infrastructural development and tax revenues were highest under the land-sparing approach, though fewer jobs were created. The SMALLHOLDER scenario (indicating land-sharing) resulted in highest environmental costs in terms of forest conversion, carbon dioxide emissions and biodiversity losses but involved more households in oil palm agriculture and thus created more employment opportunities. The HYBRID scenario ranked second best in terms of both minimizing forest loss and job creation. However, the drawbacks of this approach included high nitrogen fertilizer consumption, lower infrastructural development and lower tax revenues.