How Does Actual Unemployment and the Perceived Risk of Joblessness Affect Smoking Behavior? Gender and Intra-family Effects

Considerable attention recently has been devoted to the effect of economic fluctuations on individual health. This paper examines how one specific aspect of a change in economic status – unemployment – affects smoking behavior. We characterize an individual’s unemployment experience in three different ways: whether an individual is currently unemployed; whether an individual had a recent unemployment spell; and whether an individual is at risk for potential unemployment in the future. Risk for potential unemployment in the future is reflected by the current unemployment rate in an individual’s state of residence.

Unlike most of existing studies, we focus on the change in smoking behavior as an outcome rather than on a static measure of smoking status. We measure a positive or favorable change in smoking behavior in two ways: (1) a reduction in cigarette consumption (among those who do and do not quit) and (2) quitting cigarette smoking. We measure a negative (unfavorable) change in smoking behavior in two ways as well: (1) by increased cigarette consumption (including individuals those who relapsed and those who continued smoking) and (2) by a relapse into smoking cigarettes.

The analysis is based on the 1999-2011 waves of the Panel Study of Income Dynamics. We concentrate on a sample of adults ages 25-60 years old who currently participate in the labor market and who are present in at least two consecutive waves of the PSID. We find that about 10% of former smokers relapse and nearly a fifth of current smokers quit smoking between two consecutive waves. Depending on specific smoking behavior outcome the sample varies between 4,000 to 8,000 person-year observations. About 5% of individuals are classified as being currently unemployed in the sample, with additional 5% having recent unemployment experience. Since one’s own unemployment as well as spousal unemployment may affect an individual’s smoking behavior, and since smoking behavior of men and women may be affected differentially, we estimate models for men and women separately and include variables reflecting both own and spousal unemployment experience into our econometric models.

We find that women’s smoking behavior is likely to be impacted by both their own and spousal unemployment experience. Women who are currently unemployed are about 7 percentage points less likely to quit and about 6 percentage points more likely to increase their cigarette consumption than are currently employed women. Women who have a currently unemployed husband are about 5 percentage points less likely to quit and about 9 percentage point more likely to increase cigarette consumption. We find that the recent unemployment experience of a spouse negatively affects the possibility of either spouse decreasing cigarette consumption or quitting cigarettes altogether. Men seem to be less affected by their own unemployment experience than women. However, spouses of currently unemployed women are about 7 percentage points less likely to quit smoking and about 7 percentage points more likely to increase cigarette consumption. We do not find evidence that a higher risk of future unemployment changes the smoking behavior of either men or women.