Three Essays on the Effects of Tobacco Control Policies on Smoking Outcomes

This study explores the effects of tobacco control policies on smoking outcomes and the differences between self-reported measures of smoking outcomes and serum cotinine levels, a biomarker of recent nicotine exposure. The first paper uses quantile regression to estimate whether adult smokers' responses to tobacco control policies change across the distribution of smoking levels. I find that reductions in cigarette smoking from increases in cigarette excise taxes and cigarette prices are concentrated among the heaviest smokers. However, using serum cotinine levels, I find little evidence that cigarette excise taxes or prices affect smokers' intake of nicotine at any smoking level. I directly test whether the amount of nicotine smokers ingest from each cigarette is affected by tobacco control policies, and in fact I find evidence that the heaviest smokers consume more nicotine from each cigarette in response to higher cigarette prices. The second paper estimates the effects of tobacco control policies on non-smoking workers' exposure to secondhand smoke at their jobs. I find that smoke-free air laws reduce secondhand smoke exposure at work and these reductions translate into reduced overall nicotine exposure. I find some evidence that this reduction in nicotine exposure comes from reductions in secondhand exposure at work and evidence that smoke-free air laws reduce secondhand smoke exposure through other pathways as well. The third paper examines how adolescent smokers change their smoking behavior in response to tobacco control policies. I find that higher cigarette excise taxes are associated with reduced nicotine intake among adolescent smokers and these reductions are robust to controls for antismoking sentiment and other youth tobacco laws. In fact, I find some evidence that more stringent tobacco control policies actually lead to reductions in the amount of nicotine that adolescent smokers ingest from each cigarette. Lastly, I check whether misreports of smoking status are related to tobacco control policies, which could bias the coefficients in models where the dependent variable is self-reported smoking status. I find some evidence that measures of youth access laws are related to misreports.