We use a Chinese longitudinal survey to examine the association between beauty and individual labor market earnings. We estimate Mincerian earning equations with both cross sectional and longitudinal samples using parental characteristics as instrumental variables for beauty. We include the hazard rate to control for selection into occupations where beauty is perceived as being productive in the cross sectional specification. In the longitudinal specification, we add the inverse Mills ratio in the earning equation and carry out a test of selection bias. We get similar results in both specifications. Estimating over the full sample, we find that beauty is associated with higher earnings for females, that its exogeneity is rejected using the appropriate Hausman test, and that there is no evidence for selection into occupations where beauty is important. Our preferred point estimates indicate that the impact on hourly labor earnings of a one standard deviation increase in beauty is equivalent to a seven year increase in the number of years of schooling for the cross sectional model. For the longitudinal model, a one standard deviation increase in beauty increases women’s annual income by 40% and increase men’s annual income by more than 20%.