Abstract
Mexico's attempt to deregulate its sorghum market in 1990 was unsuccessful because policy analysts failed to consider the seasonal and spatial dimensions of trade. In this paper I present a detailed, seasonal, and spatially dissaggregated nonlinear programming model of the Mexican sorghum market calibrated with survey data of storage and transport costs in Mexico, and compare the results obtained with models that ignore these characteristics. The comparison demonstrates that the failure to incorporate these costs in trade models may lead to misleading conclusions and inappropriate policies.