Assessing the human and environmental impacts of biofuels requires unraveling the connection between international trade, on one hand, and local land-use and social change, on the other, while accounting for cross-scalar linkages between and within social and environmental systems. We propose a disaggregated approach to model how macro shocks shape rural households’ decisions, and how these decisions integrate onto aggregate supply and land use patterns. The approach, built on an agent-based model of rural Mexico, is used to explore the impacts of ethanol-driven US corn price increases. Our estimate of a 5.7% expansion in corn area by 2008 and wide variation across regions corresponds fairly well with ex post reports. Estimates from alternative models exceed ours by up to 200%. Corn land expanded between 1.6% in the southeast and 16% in the northwest. A 3% increase in agricultural value added nevertheless did not promote rural development, whether measured in terms of total rural value added or income. Direct and indirect (multiplier) effects on rural incomes were limited. Rural households experienced a 0.02% increase in real income, while absentee (non-rural) landholders’ income increased 3.9%. Our approach highlights the crucial role of local market conditions and interactions among microeconomic actors in shaping biofuels’ impacts via local feedback mechanisms. It suggests that subsistence activities might keep deforestation pressures in check in some developing areas while precluding the rural population from benefiting. A disaggregated approach should help integrate future research on land-use change and economics.