This dissertation consists of three chapters on the direct and indirect effects of tobacco control policy on risky health behaviors. Among the direct effects of tobacco taxes, I study cigarette and cigar use, while among the indirect effects, I study marijuana use and body weight.

In chapter one of my dissertation, I estimate the effect of tobacco taxes on youth cigarette, cigar, and marijuana use. Using data from the Youth Risk Behavioral Surveillance System (1999-2017) and a difference-in-differences approach, I exploit the variation in state cigarette and cigar tax over time to identify causal effects. I examine the cross-tax elasticities of cigarette and cigar use and find that, while an increase in cigar tax leads to a decrease in cigarette use, an increase in cigarette tax increases cigar use. I explain this asymmetry in cross-tax elasticity through a third product, marijuana, the most smoked product among youth. My findings imply that cigarettes and marijuana are substitutes, while cigars and marijuana are complements. Future anti-tobacco policies must take into account the high prevalence of marijuana and its role in tobacco use.

In the second chapter of my dissertation, I use quasi-experimental linear and nonlinear methods to reconcile the conflicting experimental and observational literature on smoking and body weight. I use the Behavior Risk Factor Surveillance System to study the relationship between cigarette taxes, smoking, and body weight. My findings suggest that specification choice and accounting for misreporting may be the missing links required to reconcile the two strands of research.

In the third chapter, using data from the Youth Risk Behavior Surveys, I document a diminishing marginal impact of cigarette tax rate on youth smoking using parametric and semi-parametric methods.