A standard, multiregion general equilibrium (GE) model is developed and contrasted with typical partial equilibrium (PE) models of agricultural trade for two trade policy reform experiments. In the case of reforms affecting both food and nonfood sectors, the PE model has difficulty predicting changes in patterns of food production and trade. When the shock is sector-specific, however, PE models perform very well. In this case, the major benefit of GE analysis is its ability to draw the link between agricultural and nonagricultural interests in trade policy.