Using longitudinal data from the China Health and Nutrition Survey, we examine how individual wages change in line with the share of college graduates in a given province. The individual fixed effect model shows that the external returns to education in China appear to be zero. We estimate an instrumental variables fixed effects model where share of college graduates is instrumented by the number of universities with special status and find positive external returns to education of about 10 per cent to 14 per cent. We also find that the returns are affected by individual heterogeneity. While negligible returns are found for urban, women, and high-educated workers, the returns are positive and statistically significant for rural, men, and low-educated workers. This finding provides the motivation for increasing education investment in rural China and targeting it more toward poorly educated workers.