Abstract
Using within-family variation from twins and siblings, I find that smokers earn approximately 16% less than nonsmokers. Possible explanations for this earning difference are addiction-related productivity declines and earning reductions from higher health insurance costs. To investigate further, I use variation in the provision of employer-supplied health insurance (ESHI) to examine the mechanism of whether the addiction or insurance component has a larger influence on earnings. While I generally observe a larger earning penalty for smokers with ESHI than smokers without ESHI, the earning difference is statistically indistinguishable from zero.