This dissertation focuses on using empirical, or applied, microeconomics to answer questions of interest. The first two questions of interest are, whether EBT card reforms had unintended consequences, both good and bad, in the form of altered criminal behavior and food security. Chapters 1 and 2, respectively, address these questions using empirical techniques common to applied microeconomics. The third chapter is related in the sense that is also an exploration of a question of interest using empirical techniques common to applied microeconomics. In particular, the third chapter seeks to uncover which individual traits are associated with divorce in China, and how those characteristics have changed across generations. Moreover, I also seek to see if results regarding physical and mental health outcomes are replicable using Chinese data sources.

Chapter 1 considers EBT card reforms and the impact on black market food stamp sales and trafficking. By altering households' real income, reforms may have altered criminal behavior. Theory is ambiguous as diminished income may decrease alcohol and drug related crimes, but may also cause income motivated crimes such as robbery and burglary to rise. I exploit variation in the timing of EBT adoption in California to identify the effect of EBT on arrests. Despite representing a small decrease in income, I find a significant, though transitory, increase in criminal arrests resulting from EBT adoption. In particular, following a conversion to EBT benefits delivery, the average county experiences an additional 108 arrests or an increase of about 5%. This increase lasts about 6 months before fading out. This increase in crimes is most noticeable for crimes traditionally motivated by income shortfalls, such as burglary, larceny, prostitution and robbery. In the average county, income motivated criminal arrests rise by about 25 arrests or 9.7%. The increase is persistent for 6 months before fading out. I find little evidence that drug and alcohol related crimes diminished.

Chapter 2 investigates EBT card reforms in California's food stamp program, and the impact on food insecurity. Our hypothesis is that EBT cards will reduce food insecurity by reducing the food costs associated with loss and theft of benefits, as well as decreasing fraudulent sales of benefits. We use the California Health Interview Survey, and the roll-out of EBT card reforms across California counties, to conduct an event study. Our findings suggest no evidence for a decrease in food insecurity. We do, however, find evidence of a transitory increase in food insecurity immediately following EBT card reforms. Reforms increase the likelihood of food insecurity by about 2% for 1-2 months depending on the measure of food insecurity used. The result is distinguishable from zero, and robust to changes in specification, inclusion of controls, and measurement choices. We posit the increase was due to a less than perfectly smooth transition to the EBT card system.

Chapter 3 uses large, national surveys to investigate individual attributes associated with probability of divorce. In recent years China's divorce rate has risen rapidly. With this rapid rise has come a large number of potential explanations, both those grounded in economic theory, and those widely discussed in the public discourse. We investigate which individual attributes are associated with an increased probability of divorce, and to explain which explanations are not empirically substantiated. We find that Western attitudes and a sense of relative affluence are predictive of divorce. Furthermore, most popular explanations are not empirically confirmed. Finally, we find that previous results suggesting divorce is associated with worse mental health outcomes are applicable in China and not only Western nations.

Alcohol is both a dangerous drug and an economically significant good. State regulation of alcohol is motivated by an emphasis on public safety and efficient distribution of the product. The dual nature of alcohol policy in the post-Prohibition Era has been characterized by what can only be classified by anti-competitive regulations, including: policies that require distributors to charge the same "delivered" price to all retailers, regardless of actual delivery costs; policies that require alcohol distributors to charge uniform prices to all retailers; bans on volume discounts for beer and wine sales; franchise protection laws; policies granting wholesalers specific geographic territories; prohibitions on central warehousing of alcohol by retailers; policies that require distributors and producers to post their prices in advance and/or share those prices with rivals; and minimum mark-ups on sales from producers to wholesalers and from distributors to retailers, among others. Examination of state variation in regulation of alcohol markets is therefore an important area of applied microeconomic research.

The primary context for my analysis is liquor privatization in the State of Washington. On Nov. 8, 2011, Washington residents voted to end the state's nearly 80-year monopoly on the retail sale and distribution of liquor within the state. The first-order effect of the new private liquor market was higher prices - a mostly unanticipated outcome of new fees that were imposed as part of privatization on retailers and distributors.

The first chapter uses store-level data on liquor sales in Oregon to estimate the effect of higher prices in Washington on the tendency for consumers to cross the border to buy cheaper liquor. I calculate the impact of ad-valorem taxes on tax avoidance using variation in Oregon stores' proximity to the Washington-Oregon (WA-OR) border. The empirical methodology benefits from 1) detailed location data on Oregon liquor stores and border crossings, 2) uniform pricing of liquor in Oregon, and 3) detailed sales for each of the 250 stores licensed to sell liquor in Oregon. I find that the stores closest to the WA-OR border experienced an additional 21 percent increase in sales relative to interior stores.

The second chapter examines the effect of liquor store entry on crime in the City of Seattle following liquor privatization. Previous research has linked the commission of violent crime and alcohol, although the causal relationship is ambiguous. I take advantage of precisely geocoded 911 police response data to measure the impact that store entry has on crime. Little evidence is found for an increase in the weekly number of calls for service for potentially "alcohol-related offenses", such as assaults, robbery, nuisance violations, social disorder more broadly defined, and driving while under the influence. Event study analysis and spatial analysis at the census block group level confirm the main results.

In the third chapter, I examine the outcome of Washington's public auction of the retail liquor rights associated with its former state-run stores. Privatization mandated that Washington auction off these rights ahead of privatization. Overall, I find winning bidders paid 40 percent more for the retail rights to stores closest to Oregon and Idaho border crossings. This outcome, as well as the fact that the State of Washington made $31 million from the public auction, suggests significant optimism on the part of entering firms. However, higher liquor prices from new fees have significantly reduced the value of owning a liquor store in Washington under privatization.

To my knowledge there exists no systematic study of the economic impact of state regulation of alcohol. I believe there is scope for research which identifies the components which influence consumer and firm behavior in alcohol markets. Broadly speaking, such insights would be beneficial in evaluating public policies for a wide range of decision-making environments. From a social welfare perspective, the highly regulated alcohol industry in the U.S. has been shown to generate inefficiency, welfare loss, and higher consumer prices.

This dissertation explores health care utilization and provision for the elderly. My research aims to answer questions about the health effects of specific types of care, government and insurance policies, and major life events.

The first chapter, entitled "Retirement and Health: Evidence from England," investigates the effects of retirement on individual health. With many governments actively considering incentivizing longer working lives by raising the minimum age for pension programs, I examine the discontinuity at the State Pension Age in order to find the causal effect of retirement on health behaviors, health outcomes, health care utilization, and mortality. I focus on England due to a clearly-defined pension age and lack of confounding government programs, and use a full range of novel datasets, including census records, large nationwide surveys, and mortality records. I find that retirement substantially improves well-being, self-reported health, and decreases the proportion of people reporting long-term illnesses and disabilities. However, there is no evidence of an immediate effect on cognitive ability, health behaviors, or health care utilization. Consistent with this, I further show that there is no effect of retirement on mortality. Finally, I show that individuals' show signs of lower stress, with changes to both health markers and day-to-day activities. My estimates are generally more precise than those found in previous literature, and this paper is the first to examine a full range of health-related outcomes with administrative and survey data in a unified context.

The second chapter investigates the effect of skilled nursing facilities. Here, I leverage a Medicare policy that induces a sharp change in the probability of being discharged to a skilled nursing facility (SNF) in order to generate causal estimates of the health effects of this type of care. Despite being a significant source of health care spending and of Medicare's annual budget, SNF care has been understudied, with systematic differences in individuals that enter into a SNF. The identification of this paper uses Medicare's requirement of a three-day stay in order to cover SNF care. Using inpatient and emergency department records, I find that those that narrowly made the cutoff experience a substantial increase in the probability of being discharged to a SNF, and a subsequent decrease in readmission probability.

The third chapter explores the long-term effects of month of birth on self-reported health and morality rates. I use data from the Census of England and Wales and death certificates from the Office of National Statistics to examine these effects, focusing on a large sample of individuals born between 1940 and 1960. I find that individuals born in fall and early winter months are less likely to report being in poor health and have better mortality rates, particularly compared to those born in late spring and summer.

This thesis consists of three projects based on the application of Regression Discontinuity Design (RDD) analysis to questions of public economics. Chapter 1 evaluates the effects of English Language Learner programs on student achievement. There is a significant and well-documented academic achievement gap between native English speakers and English language learners (ELLs) in US public schools. This gap is particularly large in California, the US state with the largest population of ELLs. As a result of being classified as having limited English proficiency, ELLs follow a different track through the primary and secondary public school system in the U.S., according to federal requirements and implemented at the state level. The effect of this differential treatment of non-native English speakers is ambiguous, and the division of students into separate tracks based on English lacks support in existing literature. I use longitudinal microdata from a large northern California school district to estimate the effects of ELL programs on student outcomes. Students in California are categorized as ELLs based on performance on a standardized assessment of English language proficiency, the California English Language Development Test (CELDT). Students classified as are eligible for enrollment in a bilingual programs. For primary school students, this is the Alternative Language Acquisition (ALA) program. I exploit discontinuity in the probability of enrollment in ALA at the threshold CELDT score for English proficiency. I find that schools do not observe the rule for assignment to ALA based on CELDT score.

Chapter 2 evaluates whether consumers respond to information about product quality in the context of automobile safety. Consumer choice over new vehicles is a function of multiple vehicle attributes, including price, fuel efficiency and safety. However, because vehicle safety is often correlated with other characteristics of vehicle quality, estimating consumer preferences for safety over other attributes is empirically difficult. Using a federal program in the United States that provides public safety ratings for new passenger vehicles, I exploit discontinuity in the assignment of 5-star vehicle safety ratings in continuous probability of injury measurements calculated based on crash test performance. I evaluate whether new vehicle models that just miss a star threshold on the National Highway Transportation and Safety Administration's New Car Assessment Program's 5-Star Rating scale experience lower national sales volumes relative to vehicles that just exceed the ratings threshold.

Chapter 3 evaluates the accuracy of the New Car Assessment Program (NCAP) safety ratings in predicting real-world crash outcomes. Regulatory policies such as safety standards for seat belts or airbags are aimed at improving vehicle safety and reducing occupant injury and fatality rates, but may distort driver and occupant behavior. Vehicle safety ratings can provide standardized, transparent and comparative information to buyers. However, the value of any safety rating to consumers depends on the accuracy of the rating regime. In the context of transportation policy, this means real-world loss of life, injury and property damage on a national scale. In the U.S., the National Highway and Transportation Administration (NHTSA) evaluates the safety of all new vehicles sold in the United States and publishes these safety ratings via the New Car Assessment Program (NCAP), a program which is emulated internationally. Using a novel dataset with the continuous underlying running variable, probability of injury, used to calculate the 5-Star NCAP safety rating seen by consumers, I evaluate whether U.S. NCAP safety ratings accurately predict real-world crash outcomes in terms of vehicular damage, personal injury and loss of life. Matching NCAP rating with crash report data from the U.S. Fatality Analysis Reporting System and Texas' Crash Records Information System, I find minimal correlation between NCAP rating and real-world crash outcomes.

In summary, this dissertation applies Regression Discontinuity Design analysis to public education policy in the context of the CELDT and to regulatory and transportation policy in the context of NCAP. Successful implementation of these policy regimes depends on accuracy of classification, whether of students or vehicles. Failures of evaluation and enforcement result in distortionary effects with real-world effects on educational attainment and motor vehicle safety. Regression Discontinuity Design provides a powerful tool for evaluating the true efficacy of public policies.

This dissertation studies the effect of education policies on the behavior of individuals—particularly the criminal behavior of juveniles.

The first chapter of this dissertation, "School Entry and Criminal Behavior", examines the effect of an education policy that affects when children are allowed to begin school on their likelihood of committing crime later in life. Children whose birthdays fall just before the school entry cutoff are the youngest in their cohort for their entire education. A large literature documents that, among other things, the youngest students in a classroom have lower test scores, are more likely to be held back, and are more likely to be diagnosed with ADHD. This paper investigates if being the youngest in a cohort has an impact on an individual's propensity to commit crime. I use records for over 4 million arrests spanning a 20-year period in California to assess if being the youngest in a school cohort increases the likelihood of being arrested at any point between the early teen and young adult years. Overall, I find no effect on the probability of arrest for serious crimes and no persistent effect for less serious crimes. However, the youngest students in a cohort have a higher risk of arrest for certain offenses at age 14, corresponding to the age at which they would transition to high school. This may reflect the influence of school setting, peer group composition, and monitoring standards on behavior and the probability of arrest.

The second chapter of this dissertation, "School Quality, Student Performance, and Behavior", examines the effect of an education policy that places low-performing schools on probation on the school's academic progress and student misbehavior. School districts and state education boards across the U.S. have implemented school accountability systems in order to identify which schools may need to be improved. I use an accountability system implemented by Chicago Public Schools (CPS) which placed schools on probation if they failed to reach a certain performance level based on a combination of test scores, attendance, and student growth. Schools that are placed on probation and do not show improvement may face sanctions that include principal removal, staff turnaround, or school closure. I find that schools that receive low enough performance ratings to be placed on probation in one year show slightly more improvement in their performance in the following year relative to schools that just miss being placed on probation. That is, schools that received low performance ratings increased their rating in the next year by 3 more percentage points than schools that were not placed on probation. However, this difference was generally not large enough or sustained for long enough for many schools to get off of probation. Given the nature of these schools, this is potentially a resource issue as many of these schools are in poor, urban settings. From a behavior standpoint, I do not find evidence that receiving poor performance ratings has any effect on changes in student misconduct across the probation threshold.

The third chapter of this dissertation, "The Effect of School Year Length on Juvenile Crime", examines the effect of a shortened school year on juvenile crime. As a result of the Great Recession during the late 2000s and early 2010s, many public sector employees were subject to mandatory work furloughs due to budgetary shortfalls. In order to alleviate the effects of the recession, public schools in California instituted work furloughs, which in effect decreased the number of days students would attend school. Specifically, schools were allowed to shorten their school year by five to twelve days. Given that a shorter school year implies that students will be in school less often, there may be changes in their behavior stemming from the incapacitation effect of attending school. Literature has shown that juveniles are less likely to commit property crime while in school due to a higher probability of detection and more likely to commit violent crime due higher concentrations of juveniles on school grounds. This paper examines the effect of a shortened school year on juvenile criminal behavior. I find that decreasing the amount of school days in the school year has no significant impact on juvenile arrests. This result may be explained by the overall decrease in juvenile crime over the past 20 years, as opposed to school year length changes having no effect on juvenile criminal behavior.

This dissertation uses reduced form techniques to causally answer questions of direct importance in the fields of health economics, underage alcohol enforcement, and education. The first chapter tests whether Randomized Control Trial evidence showing a procedure to be ineffective substantially changes doctors’ treatment decisions. Leveraging differential timing of publication of clinical trials that show a currently used procedure is ineffective or potentially harmful, I implement an event-study design to estimate how doctors respond to said evidence. I find that the use of the procedures in question only declines modestly by 10% within two years of publication and 30% within four years. Furthermore, focusing on a subset of publications with more definitive findings, I find an effect size of similar magnitude. The slow adoption of new evidence is similar between privately and government insured patients, and there is only weak evidence that non-profit hospitals abandon procedures at a higher rate than medical-school affiliated and for-profit ones. That medical procedures are still commonly performed long after the publication of evidence revealing they are ineffective or harmful suggests the need for greater integration between research and practice.

The second chapter examines the impact of Minor Decoy (MD) citations (a law enforcement strategy targeting liquor license holders rather than underage consumers) in curbing underage alcohol-related crimes. Leveraging spatial variation in when and where citations occur, I use an event-study analysis to study the impact of these citations on arrest rates per 10,000 people for 18-20 year olds, 21-24 year olds and all adults 18 and over. I find suggestive evidence that there is an increase in overall alcohol-related arrests on the day of citation across age groups, mostly driven by liquor law violations. All other arrest categories show no detectable impact of MD citations in curbing underage alcohol-related crime. This suggests that the presence of police in licensed establishments greatly increases likelihood of arrest across all age groups, but does not provide any evidence that the MD programs curb underage alcohol-related arrests. This may be due to the scope of the treatment itself rather than a true null effect.

The final chapter studies the interaction between for-profit college profitability and partisan elections. Identification based on policy announcements is hindered by market anticipation, whereas elections provide well-quantified shocks to the policy environment. For-profit college stocks experienced large and immediate abnormal returns after the 2012 and 2016 presidential elections, but little change after presidential elections or midterm elections before 2016. Private student loan stocks have been sensitive to presidential and congressional results over the last four election cycles. The pattern of estimates is consistent with an important role for recent gainful employment rules, greater data availability, and the expansion of direct federal loans. The effects are largest for colleges with poor debt-to-earnings ratios and high veteran enrollment rates, but abnormal returns are evident across nearly all firms, suggesting that federal policies pose a threat to the profitability and viability of a significant fraction of the industry.

This dissertation uses statistical causal inference methods to answer causal questions in the fields of health economics and economics of culture and institutions. The first chapter is an application of examiner designs in health economics. Despite efforts to reduce emergency department (ED) care and transition patients to alternative settings, there is limited evidence on its impact on patient outcomes. This chapter studies patients that call into a nurse advice line that are directed toward ED care in a rater instrumental variables design. Marginal patients are 5.3 percentage points more likely to be admitted as an inpatient within 3 days since triage and 3.7 percentage points more likely to have a second ED visit within 4 to 30 days. This chapter also shows increased outpatient utilization with no differences in short-term mortality, indicating efforts to divert patients to less acute settings are likely justified.

The second chapter examines the origins of the cross-nurse difference in average ED visit rates of similar patients quasi-randomly assigned to telephone triage nurses. Medical practitioners often make substantially different choices for similar patients. This chapter examines how variation in telephone triage practice style across nurses affect downstream patient healthcare utilization. While a triage decision-support tool standardizes the telephone triage process across nurses, the triage nurses can still exercise discretion through two potential margins: (i) overriding triage recommendations and (ii) intensifying verbal communication to ensure patient compliance with triage disposition. I construct two nurse practice measures to quantify each nurse’s average ED recommendation and verbal communication tendencies, exploiting quasi-random assignment of calls to nurses within call centers. My reduced-form estimates suggest that reassigning a call to a nurse with a higher ED recommendation and a longer call duration tendency increases the patient’s probability of seeking in-person medical attention.

The third chapter investigates the causal effects of superstitious beliefs that discriminate against women born in the years of Goat in the Chinese zodiac. Translating the Goat-year superstition into a simple framework of partner search under the zodiac discrimination, I consider two empirical strategies to test for the presence of marriage discrimination: a regression discontinuity (RD) design and a difference-in-differences (DID) method. The RD design identifies the superstition effects if women born just before and after the Goat-new year had, on average, the same unobserved characteristics. However, I find that marriage and other socioeconomic outcomes exhibit persistent seasonal patterns, suggesting that the Goat and adjacent birth cohorts are not comparable without controlling for the unobserved differences. Our DID method controls for the unobserved differences across cohorts around the RD threshold and identifies the causal effects of the Goat-year superstition. Applying the DID method to the 1979 Goat-year women in a 1% sample of the 2000 Chinese Population Census, I find no statistical evidence of the superstition effects on marriage and other outcomes.

This dissertation focuses on using causal inference techniques to assess the impact that public policies have on health outcomes, and the behavioral response of consumers and firms to consumer information. In two of the three chapters of this dissertation, I rely on quasi-experimental variation, namely regression discontinuity designs, to find causal estimates of the determinants of health and the behavioral responses to public health policies and ratings systems.

The first chapter examines the behavioral responses of patients and hospitals to public information disclosure about hospital quality. Specifically, I analyze the assignment of letter grades to hospitals by the non-profit consortium known as the Leapfrog Group. Letter grades are assigned based on sharp cutoffs of a continuous score, so I implement a regression discontinuity design which provides causal estimates of whether patients sort into hospitals with a better grade and if lower-rated hospitals improve quality relative to higher-rated hospitals in order to achieve a higher grade in the next period. Using inpatient discharge records from 7 states, I find evidence that lower-rated hospitals reduce their rate of healthcare-associated infections and patients sort into hospitals that receive the highest grade. These responses occur nearly entirely in competitive hospital markets where patients have more choice and hospitals have a higher incentive to improve, and in large hospitals, where hospitals have a higher capacity to improve.

The second chapter exploits two compulsory schooling reforms in England to find estimates of the causal effect of education on health. We use a regression discontinuity design to compare individuals that were just barely subject or not subject to the reforms, and examine the differences in their long term health outcomes. Using hospital records from the National Health Service of England, we do not find any causal effect of an additional year of high school education on mortality or morbidity.

The third chapter analyzes the effect of one of the largest public health interventions in the United States, the artificial fluoridation of community water systems, on dental health, general health, and education. In this chapter, I merge several surveys and administrative data sets to obtain a rich set of controls, a fixed effects model to examine the effect of fluoride on educational achievement, and non-parametric specifications to assess how fluoride affects dental health at different levels of health and fluoride levels. I find consistent evidence that fluoridated water has a significant impact on the prevention of dental cavities in baby teeth, but no robust effect on permanent teeth or on general health or educational attainment. I do find however that fluoridated water significantly increases mild or severe cases of fluorosis.

The main focus of this dissertation is on social issues affecting people's health and welfare, within both the developed world and developing world. Particularly, policy changes are employed as natural experiments, to identify the health and social impacts of policies and welfare. This dissertation consists of three chapters.

Chapter 1 analyzes the health effects of China's centralized home heating infrastructure. The primary health mechanism is that recipients of centralized heating do not need to use traditional fuels to heat their homes, and thus suffer from less exposure to indoor air pollution. A secondary mechanism is the increased quality of modern heating leads to improvement in health. I exploit spatial and time variation in the implementation of China's centralized heating program. Spatial variation takes the form of a discrete geographical cutoff between regions that receive centralized home heating, and those that do not. I compare areas near the geographical cutoff, before and after the provision of centralized heat using a difference-in-differences framework. My outcomes of interest are perinatal mortality and prevalence of low birth weight infants, since existing research suggests this sub-population is most vulnerable to indoor air pollution. I find that areas with centralized heating have a 0.13% decrease in perinatal deaths and a decrease of 1.17% in the proportion of low birth weight infants. This translates to about 7,000 fewer deaths and about 60,000 fewer low birth weight infants each year.

Chapter 2 investigates EBT card reforms in California's food stamp program, and the impact on food insecurity. Our hypothesis is that EBT cards will reduce food insecurity by reducing the food costs associated with loss and theft of benefits, as well as decreasing fraudulent sales of benefits. We use the California Health Interview Survey, and the roll-out of EBT card reforms across California counties, to conduct an event study. Our findings suggest no evidence for a decrease in food insecurity. We do, however, find evidence of a transitory increase in food insecurity immediately following EBT card reforms. Reforms increase the likelihood of food insecurity by about 2% for 1-2 months depending on the measure of food insecurity used. The result is distinguishable from zero, and robust to changes in specification, inclusion of controls, and measurement choices. We posit the increase was due to a less than perfectly smooth transition to the EBT card system.

Chapter 3 uses large, national surveys to investigate individual attributes associated with probability of divorce. In recent years China's divorce rate has risen rapidly. With this rapid rise has come a large number of potential explanations, both those grounded in economic theory, and those widely discussed in the public discourse. We investigate which individual attributes are associated with an increased probability of divorce, and to explain which explanations are not empirically substantiated. We find that Western attitudes and a sense of relative affluence are predictive of divorce. Furthermore, most popular explanations are not empirically confirmed. Finally, we find that previous results suggesting divorce is associated with worse mental health outcomes are applicable in China and not only Western nations.