Empirical studies on the earnings effects of tobacco use have found significant wage penalties attached to smoking. We produce evidence that suggests that these estimates are significantly upward biased. The bias arises from a general failure in the literature to control for the past smoking...