We analyze two puzzling results released by the California Air Resource Board who recently revises its land use changes and greenhouse gas emissions induced by biofuels. First the absolute reduction in the US average soya biodiesel estimate is much greater than the reduction in the US average corn ethanol one. Second the EU canola biodiesel estimate is twice the US canola biodiesel one. We find that these two puzzling results are mostly explained by some weak initial economic data. In both cases, the underestimation of the oilmeal production values biases upwards the carbon emission estimates. We then recall that any economic analysis is only worth the quality of the supporting data. The current focus on unobserved elasticity values to assess biofuel impacts is not sufficient.