This paper builds on the disconnection between scientific evidence and policy assumptions about the temporal profile of land use change (LUC) emissions. Whereas natural scientists find evidence of a decreasing time profile of LUC emissions, European energy policy relies on a steady time profile. We investigate the consequences of using such a uniform (constant) time profile when assessing biofuel projects with cost-benefit analysis, a widespread economic tool for public project assessment. We show that the use of the uniform time profile distorts LUC emissions costs downwards (upwards) when carbon prices grow slower (faster) than the discount rate. We illustrate our results with the conversion of grassland to wheat cultivation for bioethanol production in France. Under current assumptions in public project assessment, we find a 70% overestimation of costs related to direct LUC emissions. We propose two tools to aid in decision-making and address the decision error. Finally, we provide contextual policy recommendations.