This study explores determinants of the wage penalty borne by smokers. The authors reconstruct individual smoking histories by pooling PSID (Panel Study of Income Dynamics) data for 1986-2001. They find no wage gap between former smokers and those who had never smoked, but statistically significant wage gaps between smokers who would continue smoking and three other groups: those who would later quit smoking, those who had quit smoking already, and those who never smoked. The wage penalty for smoking, observed in the 1986 cross-section, is largely driven by those who would continue smoking over the years 1986-2001. These results suggest that the smoker/nonsmoker wage differential observed at any given time may be driven by a non-causal explanation rather than by smoking per se. For example, persistent smokers may be characterized by myopia that leads to reduced investment in health capital and firm-specific or other human capital. [PUBLICATION ABSTRACT]