Despite their complexity, existing policy evaluation methods ignore many features of the real world that are pertinent for welfare analysis of trade policy. The main limitation of these technics is that they are static, which means they ignore important dynamic consequences of trade liberalization. This paper develops dynamic tools that overcome many of these weaknesses. I apply these technics to the North American Free Trade Agreement (NAFTA). My analysis suggests that while the steady state gains from NAFTA are significant, the transitional costs associated with moving to the liberalized steady state are relatively large, so that on net the trade policy produces modest welfare gains for North America.