Smoking and wage rates - Evidence from Japanese panel data

This paper analyzes smoking behavior and its effects on wage rate. It is widely known that smoking has negative health effects and to discourage the practice, governments frequently implement policies such as tax increases and restrictions on public smoking. To evaluate these policies, it is necessary to understand the economic benefits and costs of reducing the number of smokers. We conduct a panel data analysis while controlling for unobserved heterogeneity and show that smoking is not the main factor causing observed wage differentials between smokers and non-smokers in Japan. As for the relationship between smoking and wages, we find that male smokers receive lower hourly wages than male nonsmokers. Smoking behavior generally depends on environmental, congenital, or social factors. These factors, however, do not only affect smoking behavior but also the wages of these individuals. To isolate the effect of smoking on wages, we control for individual heterogeneity using panel data to verify the existence of such an effect. Using tax change as an instrumental variable in the fixed effects model and several robustness checks, the results show that smoking has no statistically significant effect on wage rate. This suggests that smoking does not directly affect wages; rather, unobserved individual heterogeneity (other factors influencing both smoking and wages) leads to wage differences between smokers and nonsmokers. Nevertheless, smoking can affect wages in the long run through health problems. The nine-year panel data used in this research, however, is insufficient to capture the long-term indirect health effects. Thus, future research should consider analyzing indirect health effects using panel data spread across a wider time frame.