Empirical work relating trade liberalization and income distribution has identified an important anomaly. The Stolper-Samuelson theorem predicts that trade liberalization will shift income toward a country's abundant factor. For developing countries, this suggests liberalization will principally benefit the abundant unskilled labor. Yet extensive empirical studies have identified many cases with a contrary result. This paper develops a simple theoretical explanation for this anomaly. It shows that countries which are labor abundant in a global sense may see wages decline with liberalization if they are capital abundant in a local sense. The current absence of empirical work that would allow us to identify the relevant local abundance implies that virtually all assertions regarding anticipated distributional consequences of trade liberalization are without foundation. There may likewise be important implications for industrialized countries that border developing countries undertaking trade liberalization, particularly in regard to the incentives for migration.