Much of the debate over potential wage changes under a North American Free Trade Agreement (NAFTA) reflects views about the links between output prices and factor prices as described in the Stolper-Samuelson theorem. But the Stolper-Samuelson theorem does not fully describe the likely labour market effects of NAFTA because it includes the assumption that factors do not migrate. There are two forces at work that will affect US-Mexican wages under NAFTA; (a) indirect links between prices and wages as described in the Stolper-Samuelson theorem, and (b) direct effects of migration on labour supplies in the two countries. We use an 11-sector computable general equilibrium model of Mexico and the USA with both price changes and migration to determine which wage effect dominates following trade liberalization. We find that migration effects generally dominate Stolper-Samuelson effects on wages. Empirically, Stolper-Samuelson effects are very small.