Two of the most notable trends in public health over the past 30 yr are the reductions in smoking rates and the rapid increase in obesity rates. Based on data from the National Health Interview Survey, the smoking rate among adult Americans was roughly 37% in the early 1970s and is 21% today, a reduction of more than 40%. In contrast, the adult obesity rate increased 135% from 14% in the early 1970s to 32.2% in 2004 National Health and Nutrition Examination Survey [NHANES] 71-04; Figure 1). Because the medical literature has shown that smoking cessation is associated with weight gain, these trends have given rise to a puzzling question: Have efforts to reduce smoking rates in the United States inadvertently caused increases in rates of obesity?
There is substantial evidence that, at least in the short term, smoking and obesity are not independent. Klesges and Shumaker (1992) found that reductions in smoking and/or quitting altogether is associated with increases in dietary intake as well as metabolic changes that promote weight gain. As a result, weight gain is often an unwanted side effect of quitting smoking. The 1990 Surgeon General's Report, The Health Benefits of Smoking Cessation, concluded that the average weight gain after smoking cessation was about 5 pounds or 2.3 kg (U.S. Department of Health and Human Services 1990). However, there is additional evidence that this increase may be short-lived. Mizoue et al. (1998) found that, given enough time, the weight of former smokers inevitably returns to that of never smokers. As a result, the long-term impact on weight of efforts to reduce smoking rates may be negligible.
[FIGURE 1 OMITTED]
To date, three articles have attempted to quantify what proportion of the increase in obesity rates over the past several decades is caused by efforts to reduce smoking. Chou, Grossman, and Saffer (2004) used repeated cross-sections from 1984 to 1999 Behavioral Risk Factor Surveillance System (BRFSS) data to estimate the impact of several state-level measures on body mass index (BMI) and on the probability of being obese, including increases in state-level cigarette prices (Figure 1). The results of this study suggest that increases in cigarette prices may be responsible for as much as 20% of the increase in BMI and trends in obesity rates over this period. This finding is used to argue that rising obesity rates are an unintended consequence of the success of the antismoking movement.
Gruber and Frakes (2006) extended the models presented in Chou, Grossman, and Saffer (2004). Using the same data set, they examined the relationship between cigarette taxes and body weight and argue that cigarette taxes are a more appropriate measure than prices because price changes may be driven by market factors that affect the rates of smoking and eating. Their specification also includes several other differences, including controls for state and year effects. Gruber and Frakes found that cigarette taxes are negatively associated with body weight, suggesting that increases in cigarette taxes result in reductions in BMI. This result is in direct contrast with the findings of Chou, Grossman, and Saffer. However, not even the authors find their result plausible (p. 196).
Flegal (2007) presents yet another analysis of the link between smoking and obesity. Using NHANES data, the author models the contemporaneous relationship between smoking and weight and then reestimates the current prevalence of obesity under the assumption that past smoking levels had remained unchanged. Using this approach, she concludes that decreases in the prevalence of cigarette smoking had only a small effect (less than 1 percentage point) on the overall prevalence of obesity.
In this article, we revisit the analyses presented by Chou, Grossman, and Saffer and Gruber and Frakes to explain the dichotomy of results and provide further evidence …