Using National Longitudinal Survey of Youth data, the authors examine the effect of smoking on wages. Their analysis controls for differences in individual characteristics that may be correlated with both smoking and wages, including unobservable person-specific characteristics that are constant over time, and unobservable characteristics that are constant within a family. Estimates from alternative specifications indicate that smoking reduced wages by roughly 4-8%. Empirical tests of three potential explanations for this finding yield no conclusive results.