The educational expansion experienced by cohorts born after World War II in France is used to estimate the effect of schooling on the decisions to start and quit smoking. Educational expansion was driven by several reforms of the school system, and generated important changes in the distribution of schooling, especially for cohorts born between the 40s and the 70s. However, the share of those individuals who graduated with top university degrees and from the "Grandes Ecoles" remained stable throughout most of the period. We compare changes in smoking between these individuals (the control group) and the rest of the population (the treatment group), and between cohorts, to identify the effect of schooling on smoking. This difference-in-difference strategy provides evidence that, for women born between 1945 and 1965, education has had a negative effect on the decision to start smoking, and a negative impact on smoking duration. Results for men from the same birth cohorts are imprecise, but also suggest negative effects of education on smoking, albeit smaller than those observed for women.