Article dans une revue: Post-war expansion of education in France transformed the distribution of schooling for the cohorts born between the 1940s and the 1970s. However, throughout this expansion the proportion with the highest levels of qualifications remained stable, providing a natural control group. We evaluate the impact of schooling on smoking, for the beneficiaries of the post-war expansion, by comparing changes in their outcomes across birth cohorts with changes within the control group. We uncover robust evidence that educational expansion contributed to a decline in smoking prevalence of 2.9 points of percentage for men and 3.2 points for women at the turn of the 21st century. Our results also suggest that the persistence of the schooling-smoking gradient is better explained by differences in the education-related opportunity costs of smoking than by differences in information about smoking dangers.