Increasing use of biofuels increases the demand for agricultural land. Credible empirical evidence supports the common-sense judgment that this will lead to the conversion of forests and other habitats to generate more cropland, particularly in the tropics, where land conversion is cheapest. However, when analyzing the effects of biofuels on land use, governments frequently use a particular class of economic models, including the popular “GTAP” model, to justify a finding that biofuels will cause little additional land conversion. We argue that the GTAP model does not provide a credible scientific basis for this conclusion because it lacks an econometric basis for its economic parameters, generates physically impossible results by a wide margin, and incorporates several unsupported assumptions that guarantee little land use change, such as constraints on international trade and a failure to account for unmanaged forests.