The effect of habitual smoking on labour productivity

Abstract

This article investigates the effect of habitual smoking on full-time employees' hourly wages, which represent one of the social costs of smoking. Because the decision whether to smoke is affected by the cigarette tax level and by various socioeconomic factors, it is appropriate that we treat smoking indicators as endogenous in econometric analyses. To control this endogeneity, I use the levels of state and federal cigarette taxes per package and family attributes as instrumental variables for habitual smoking. According to the estimation results, there is no difference in wages between smokers and nonsmokers of both genders after appropriately controlling for the endogeneity of smoking. This result is different from the results of most previous studies, most of which suffer from several methodological problems. I also found that cigarette taxes have a strong impact on smoking participation for both genders. However, it is also found that cigarette taxes do not make smokers without rationality quit smoking.

Acknowledgements

I thank Yuichi Furukawa, Masayoshi Hayashi, Hidehiko Ichimura, Masako Ii, Yasushi Iwamoto, Ryo Kanbayashi, Daiji Kawaguchi, Miki Kohara, Tsunao Okumura, Makoto Saito, Motohiro Sato and the participants in seminars at Yokohama National University, Osaka University, and at the Institute of Statistical Research for their helpful comments and suggestions. This study is financially supported by research fellowships from the Japan Society for the Promotion of Science for Young Scientists (No. 17-10149), the Grant-in-Aid for Young Scientists Start-up (No. 19830087) and the Grant-in-Aid for Scientific Research B (No. 20330062).

Notes

¹In fact, the instruments used in this article have significant effects on smoking in the probit estimation. See Table A2 in Appendix.

² Summary statistics of the subsample are presented in Table A1 in Appendix.