A Closer Look at Tobacco Policy and Physical Activity

Maintaining a healthy weight, engaging in regular and vigorous physical activity and not smoking are widely regarded as the most effective pathways to good health and improved life expectancy.  However, it is not clear – either theoretically or empirically -- how individuals make decisions regarding these potentially complementary and compensating behaviors.  Numerous studies have empirically investigated spillover effects on obesity from policies to reduce smoking, whereas far less is known about the possible effects on other compensatory behaviors, such as exercise.  Exercise has a potentially complicated relationship with smoking, as both behaviors may help manage weight but have conflicting effects on health, both now and in the future.  To our knowledge, only two studies have investigated this relationship and have arrived at opposite conclusions.  Given the importance of physical activity to long term health, understanding the effects of tobacco policies on exercise behavior is critical to evaluating their efficacy in improving overall health.

Our research makes four contributions in an attempt to close this gap.  First, we extend the forward-looking conceptual framework of longevity and health investments formulated by Gary Becker to incorporate the effects of smoking and physical activity.  This framework reveals the ambiguous effects of tobacco policies on exercise and helps explain other empirical regularities frequently observed.  As proposed by Becker and others, there should be complementarity in disease prevention activities due to a “competing risk of death” effect; however, smoking may also separately affect the demand for physical activity through its role in weight management and its effects on the ability to exercise.  Second, our empirical analyses provide a direct test of Becker’s framework, which posits that an exogenous decrease in mortality from a given disease (for instance, smoking-related lung cancer as higher taxes induce smokers to quit smoking) would increase the benefits of investment in general health (for instance, physical activity).  Third, by using post-2000 data in our analyses, our estimates provide a critical update to past research. The 2000s saw much greater variation in and levels of cigarette taxes than in previous years, which helps identify empirically their effects on physical activity and makes a behavioral response more likely than in past years when such taxes were relatively low.  Finally, we use data from both the American Time Use Survey (ATUS) (2003-2012) and the Behavioral Risk Factor Surveillance System (BRFSS) (2000-2012) for a more comprehensive empirical approach.  The ATUS can measure more precisely both the duration and intensity of physical activity than the BRFSS used in past research; it can also distinguish if the physical activity came from work behavior, deliberate exercise or other types of exertion.  The ATUS further allows measures of total physical exertion, which we also examine in order to assess whether shifts in recreational exercise per se are being counteracted or reinforced by shifts in other activities. The post-2000 BRFSS, while containing cruder measures of exercise, provides a complementary analysis that can take account of past smoking behavior, weight status and health, facilitating a tighter link to our conceptual framework.