This dissertation studies three topics in the economics of health policy. In Chapter 1, I study how avape ban in the US would impact cigarette smoking and life expectancy. In Chapter 2, I study the unequal geographic distribution of primary care physicians. In Chapter 3, we study the mechanisms that drive a nutrition program's spillover effects onto nonparticipating households. In Chapter 1, I study how a vape ban in the US would impact cigarette smoking and life expectancy. Bans on vape sales have risen in popularity across the world despite ongoing debate about their public health implications. Vapes may help existing smokers quit smoking cigarettes, which are more harmful, but their availability may also encourage young non-smokers to start vaping, become addicted, and potentially transition to smoking. I study whether a vape ban in the US would improve life expectancy, accounting for three key considerations: substitution to cigarettes, differences in vape and cigarette addictiveness, and greater health harms from product use in older age. I first document three facts using the Population Assessment of Tobacco and Health: (i) older vapers are more likely than younger vapers to smoke next year, (ii) vapers become less addicted than smokers, and (iii) vapers are more likely than abstainers to smoke next year, consistent with either a "gateway" effect or unobserved correlated preferences for vapes and cigarettes. I then develop and estimate a dynamic panel data model of cigarette and vape demand that incorporates cigarette and vape addiction, age and cohort effects, unobserved preferences, and changes in prices. Finally, I simulate life-cycle smoking and vaping decisions with and without a vape ban, and analyze the implications for life expectancy. I find that a vape ban would decrease average life expectancy in the current US population unless vapes were at least 60% as harmful to life expectancy as cigarettes. Drawing on existing research suggesting that vapes are 25% as harmful as cigarettes, a vape ban would decrease life by a total of 11 million life-years, and harm current adults while benefiting current youth. This is because older vapers are more likely to substitute to cigarettes, vapers who switch to smoking may smoke for longer than they would have vaped, and increased smoking persists into later stages of life when smoking is more harmful. In Chapter 2, I study the unequal geographic distribution of primary care physicians, a source of inequality in primary care access and health outcomes. One driver of physician location decisions is the distance to their medical school. Nearly 50 new medical schools have opened over the last two decades and may increase primary care access in the surrounding areas. I study primary care physician's location choices between small geographic markets for primary care and consider how a recently opened medical school in California would affect inequality in primary care access in the state. Utilizing a comprehensive dataset of nearly 200,000 primary care physicians' practice locations in 2016, I document that (i) there is substantial variation in physician allocation across and within states, highlighting the importance of local market analysis, (ii) physicians tend to locate near their medical school, and (iii) new medical schools increase the total number of students from that state, suggesting that new schools increase the fraction of physicians with a preference for that location. I then develop and estimate an entry model where physicians choose between counties in California and other states. Physicians trade off the distance from their medical school, county amenities, and local competition. I find that while medical school proximity influences location choices, local amenities and market competition are more decisive factors. A recently opened medical school in an underserved California county would have a limited impact in reducing inequality in primary care access within the state. In Chapter 3, we study the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). WIC is a large US government program that provides infant formula to low income households and serves 50% of US infants. Each state contracts with a single manufacturer that provides infant formula for all participants in that state. However, the WIC supplier may become more attractive to consumers who do not participate in the program if their products receive more or better shelf space in retail stores, consumers interpret the WIC label in retail stores as a signal of government endorsement, or hospitals provide the WIC supplier's formula to all new mothers. We study the mechanisms through which the WIC program impacts purchasing behavior for consumers who are not eligible for the program. Using household-level data and the timing of WIC contract changes across states, we first provide empirical evidence that the manufacturer that wins the WIC contract increases its share among ineligible households by 30 percentage points. Next, to identify the retailer effect, we leverage the fact that WIC does not permit online ordering by comparing household purchases made online and offline and studying how these purchases change when a new WIC supplier replaces the incumbent in a state. To identify the hospital effect, we leverage variation in hospital provision of formula samples across states and the distance from consumers to hospitals who do and do not provide infant formula samples. We find that the in-store retailer environment drives the spillover effect more than hospital-provided formula samples. We also show that among non-WIC households, those who are lower-income are most affected by the spillover effect.