Unemployment, insurance and smoking

Abstract

Using recent state-level data from the United States, this article examines new influences on cigarette demand. In particular, we uniquely focus on the effects of unemployment and health insurance coverage on smoking. Results show that higher cigarette prices, a lack of health insurance and restrictions on smoking at home, all lead to reduced smoking. On the other hand, literacy, income, unemployment, workplace smoking restrictions, smokeless tobacco taxes and tobacco-producing states do not seem to have an appreciable impact. The magnitude of the price elasticity of demand is greater than that found in the pre-MSA era. Policy implications are discussed.

Acknowledgements

Comments of a referee and research support from the Katie School of Insurance are gratefully acknowledged. The views expressed and remaining errors are mine.

Notes

¹ See Viscusi (1999) for a discussion of the private and social costs of smoking. Also see Kenkel (2000).

² A more basic measure of literacy might be a person's ability to read and write.

³ The number of observations in this case is smaller due to missing data and due to some states reporting their smokeless tobacco taxes in terms of cents, rather than percentages.

⁴ Details are available upon request.

⁵ However, one should bear in mind that our elasticity estimate is based on simple linear demand specifications. It is possible that alternate specifications might yield somewhat different results (Gallet and List, 2003).

⁶ A technical reason for the lack of statistical significance of the education variable might be that education and income are likely to be positively correlated.

⁷ Some studies have found that greater unemployment leads to a reduction in smoking (Ruhm, 2000).

⁸ Appropriate caution, however, should be exercised in interpreting our results with respect to health insurance since the nature of insurance coverage might vary qualitatively. For instance, health insurance might be state provided (public coverage) or it may be private. Furthermore, private coverage may be obtained as part of a group (e.g. employees of a corporation) or individually (e.g. small business owners). Moral hazard and adverse selection problems associated with smoking might be different in each case. Unfortunately, the level of aggregation in our data does not allow us to consider these qualitative differences.

⁹ Another recent development that has been difficult to track (and study) is the increasing cigarette sales via the Internet. These sales might have contributed to the reduction in the influence of tobacco-producing states.

¹⁰ In comparing the relative effects of NoInsure and UN, it is quite possible that some of the unemployment might be short term, in which case the unemployed individuals might have health insurance. Conversely, some of the employed individuals are likely to lack health coverage. This might partly be the reason behind the weak statistical significance of the unemployment coefficient.