This paper develops and implements a novel test of the Stolper-Samuelson theorem. We use nationally-representative matched employer-employee panel data from 1997 through 2015 to study the effect of the rise in China’s exports on French worker earnings. Our version of the Stolper-Samuelson theorem states that there is a negative correlation between occupation exposure to Chinese competition and change in worker earnings. First, we document substantial heterogeneity in trade adjustment across occupations. Then, consistent with the Stolper-Samuelson prediction, we show that workers initially employed in occupations more intensively used in hard-hit industries experience larger declines in earnings. We also show that workers tend to move out of hard-hit industries, but they tend to remain in their initial occupation.