How Robust is Comparative Advantage?

This paper reviews the theoretical development of the concept of comparative advantage, starting with the two‐good model of Ricardo and the two‐good extension and reinterpretation by Haberler. In both, the presence of comparative advantage provides the scope for countries to gain from trade by specializing, and the pattern of that trade is explained by the pattern of comparative advantage. These strong results of the two‐good model can be extended under certain circumstances to multiple goods and countries, but under more general assumptions such strong results no longer are assured. Instead one can derive much weaker results, usually in the form of correlations between comparative advantage and trade, and these weaker results hold in a much wider variety of circumstances. The paper examines those assumptions that permit such generalizations, but then also examines when those assumptions are most likely to fail, and what happens as a result.