Do changes in education levels explain trends in smoking prevalence ? Evidence from France.

Summary This paper analyzes the relationship between education and smoking in France over the last century, using time series of prices and sales, several surveys of French educational qualifications, and retrospective individual data on smoking careers. A reduction in per capita consumption and a decline in the price elasticity of tobacco are observed in the second half of the 20th century. The increase in the average level of education could explain this trend because, on the one hand, education can improve the understanding of public health messages, which is a productive efficiency argument traditionally put forward in health economics, and, on the other hand, because the increase in the level of education is associated with an increase in the standard of living and therefore in the potential costs (in terms of welfare losses) of early death. The authors try to test the second explanation by assuming that individual well-being is relative, and by using the position of individuals in the distribution of diplomas specific to their cohort and their gender as a proxy for their relative social position. After controlling for price changes and unobserved heterogeneities, they find that this proxie is positively correlated with age of smoking initiation and negatively correlated with duration of smoking, whereas education has little effect on duration. Moreover, if the productive efficiency argument is valid, the most educated should have more price-elastic behavior after 1976, when information campaigns and tax increases were correlated. This prediction is invalidated for the elasticity of age of initiation, but is validated for the elasticity of duration of consumption. However, the latter result can also be interpreted in terms of opportunity costs: over the long term, differences in opportunity costs translate into higher consumption and therefore greater dependence on tobacco and, ultimately, less responsiveness to prices. Thus, despite the massive increase in education, the less educated remain less responsive to anti-smoking policies primarily because they have fewer incentives to quit. As a result, the reduction in aggregate sales is primarily due to reduced consumption among the affluent.