This dissertation consists of three separate chapters on topics in health economics. Using data from the Behavioral Risk Factor Surveillance System (BRFSS) from 2000 to 2010 the first chapter, entitled "The Effectiveness of Cigarette Taxes on Older Adult Smokers: Evidence from Recent State Tax Increases," investigates the impact of cigarette excise taxes on the smoking participation across the age distribution of older adult smokers. I find that a one dollar increase in the excise tax reduces smoking participation by 2.5 percent for adults age 30 to 74. An investigation of the age distribution finds that results are sensitive to how older adult age groups are defined. The effect of one dollar excise tax increase is largest for a 45 to 49 year old group, reducing smoking participation by over 7 percent. The effect of taxation is more widespread by age for women and low-income smokers. Additionally, women and high-income smokers reduce their smoking participation because of taxation earlier in life. This paper establishes that the impact of taxation on smoking participation for older smokers is diminished at higher tax and price levels.

The second chapter, entitled "The Effect of Alcohol Access on Mental Well-Being: Evidence from the Minimum Legal Drinking Age," uses data from the 2008-2012 National Survey on Drug Use and Health to examine the relationship between alcohol consumption and mental well-being. A regression discontinuity framework at the minimum legal drinking age is used to identify the causal effect of alcohol consumption on short-term self-reported mental well-being. My empirical analysis finds some conservative evidence of a positive, causal effect of alcohol consumption on mental well-being. I find that alcohol use jumps by 20.7% at the age of 21. At the same threshold there is a 9.8% decrease in "feeling hopeless." The implied instrumental variable estimate finds that a 10% increase in the percent of alcohol use is associated with a 5.6% decrease in feeling hopeless. This result is fairly robust to specification and bandwidth. There is less evidence of an effect for "feeling down on yourself, no good or worthless" or no evidence of an effect for "feeling nervous." I also explore the differences by gender and race.

The third chapter, entitled "The Substitutability of Alcohol and Marijuana: Where there is Smoke, is there Fire?," uses data from the 2002-2007 National Survey on Drug Use and Health to investigate recent evidence from a regression discontinuity framework that alcohol and marijuana are substitutes for young adults. The central assumption underlying this method is that the model correctly specifies the smooth function of the forcing variable, in this case, age. I consider a wide variety of parametric and nonparametric models to test the robustness of the discontinuous effect found for marijuana use at the minimum legal drinking age (MLDA). The recent finding that alcohol and marijuana are substitutes is sensitive to specification choice for the whole sample. Regardless of the specification there is no evidence of a significant change in marijuana use by men, while the substitution effect for women is robust. I corroborate and investigate the gender difference using data from the National Longitudinal Survey of Youth 1997 Cohort (NLSY97). I find that the reduction in marijuana use at the MLDA by women is heterogeneous by education and race. There is also evidence of a complementary relationship between alcohol and marijuana use for parts of the male sample.