Xiaohui Guo

This study presents the first causal evidence on how U.S. pro-joint physical custody laws (PJPCLs) affect mothers' mental health. Exploiting the staggered adoption of state reforms during the 2010s, we apply the Gardner (2022) difference-indifferences estimator to multiple large-scale survey and administrative datasets. Our first-stage analysis provides descriptive evidence that court-ordered joint physical custody increases by 17.8 percent following the reforms. Using the 2010-2024 Behavioral Risk Factor Surveillance System (BRFSS), we find that PJPCLs increase the number of poor mental health days among mothers aged 18-44 by 12.6 percent, with the effects emerging gradually over the post-reform period. Our estimates are robust to inference procedures designed for a small number of treated clusters and are similar under triple-difference specifications comparing mothers to childless women within the same state. We also find little evidence that the estimated mental health effects are driven by selection into parenthood or interstate migration. Additional analyses reveal that mothers increase childcare time and reduce labor supply, alongside declines in marital dissolution and intimate partner violence. These patterns are consistent with heavier childcare constraints and weakened intra-household bargaining power. Overall, these results suggest that reforms intended to equalize parental involvement impose unintended psychological costs on the parent who had traditionally been the primary caregiver.

This study is the first to estimate the effects of unemployment insurance (UI) benefit contractions on mental health. We leverage a unique quasi‐natural experiment: early termination of the Federal Pandemic Unemployment Compensation (FPUC) and Pandemic Unemployment Assistance (PUA) programs across more than 20 states during the summer of 2021. Using high‐frequency data from the Census Bureau's Household Pulse Survey (HPS), our difference‐in‐differences estimates suggest that early withdrawal from FPUC and PUA increases the probability of reporting anxiety symptoms among working‐age adults without a college degree, a group particularly vulnerable to unexpected income shocks. The adverse effects are more pronounced among racial and ethnic minorities, older individuals, and women. However, the increase in anxiety symptoms is short‐lived, persisting for about 1 month after the benefit cessation. Finally, we find evidence that financial hardship is a likely mechanism underlying these findings.

This study provides the first estimates of the mental health effects of state age verification laws (AVLs), which require pornographic websites to verify that users are over 18. In response, the largest adult-content platform exited nearly all AVL states. Using Google Trends data, we document disruptions to habitual browsing: AVLs are associated with sharp declines in search interest for the dominant compliant platform, accompanied by substantial substitution towards noncompliant sites and increased searches for circumvention tools. We then leverage individual-level data from the Household Pulse Survey and find robust evidence that AVLs increase self-reported anxiety symptoms among adult men, with stronger impacts for younger, single, and non-white individuals. Event-study estimates indicate that these effects are short-lived, consistent with psychological frictions from disrupted status quo. In contrast, we find minimal mental health effects among women. Our findings suggest that online content regulation aimed at protecting children may have unintended consequences for the broader population.

This is the first study to estimate the impact of sports gambling legalization on food sufficiency using the staggered implementation of state laws between 2021-2023. By analyzing Google searches for online sportsbooks and legal wagers, we show that interest in sports gambling increases sharply following legalization. Applying an imputation-based difference-in-differences design to biweekly Household Pulse Survey (HPS) data, we find that legalized sports gambling reduces household food sufficiency by 2.1 percent among working-age adults without a college degree, which translates to a 10.5 percent decline among active bettors. This effect is cyclical, persisting for three-to-five months during NFL seasons. Additionally, the declines in food sufficiency are larger among households with adults aged 25-44 and racial/ethnic minorities. Using additional data from the HPS, Behavioral Risk Factor Surveillance System and Current Population Survey, our exploration of mechanisms suggests that the negative effects on food sufficiency operate through financial distress rather than changes in mental health or labor supply. A back-of-the-envelope calculation indicates that legalization of sports gambling in nine states resulted in an additional 284,000 food-insufficient households and $130.2 million in excess healthcare expenditures annually.Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

The U.S. paid‐sick‐leave (PSL) mandates allow private‐sector workers to take time off for their own or family members' illnesses. However, the impacts of these laws on workers' care seeking and provision activities are relatively understudied. We address this gap by exploiting cross‐state variation in mandate enactment between 2012 and 2019. Using data from the American Time Use Survey, our difference‐in‐differences estimates show little evidence that PSL mandates affect the probability of spending time on overall care provision at the population level. However, we find a positive association between the mandates and care provided to adults by workers in industries with larger exposure to the policies. Our results provide empirical support for the potential of paid leave policies to help workers who otherwise have limited access to the benefit reconcile workplace responsibilities and caregiving.

This study is the first to estimate the effects of unemployment insurance (UI) benefit contractions on mental health. We leverage a unique quasi-natural experiment: early termination of the Federal Pandemic Unemployment Compensation (FPUC) and Pandemic Unemployment Assistance (PUA) programs in over 20 states during the summer of 2021. Using high-frequency data from the Census Bureau’s Household Pulse Survey (HPS), our difference-in-differences estimates suggest that early withdrawal from FPUC and PUA increases the probability of reporting anxiety symptoms among working-age adults without a college degree, a group particularly vulnerable to negative income shocks. The adverse effects are more pronounced among racial and ethnic minorities, older individuals, and women. However, the increases in anxiety symptoms are transitory and disappear two months after the benefit cessation. Results from a supplementary analysis using Google Trends data also show a similar short-term increase in searches for anxiety. Finally, we find evidence that financial hardship is a likely mechanism underlying these findings.

This is the first U.S. study on the causal relationship between unemployment insurance (UI) expansions during the Great Recession and the use of prescribed opioids. Using annual county‐level prescription data from 2006 to 2013 and plausibly exogenous variation in state and federal policies, we estimate that a one standard deviation increase in UI generosity ($10,800) reduces per‐capita prescription opioid use by 1.9%. We supplement our main analysis with a border discontinuity design and find a similar effect. We also find no evidence of pre‐trends in the outcome. Finally, additional analyses suggest that increased health care utilization after exposure to higher UI generosity could be a plausible explanation for the declines in opioid prescriptions.

We determine how enrollment in the Children's Health Insurance Program (CHIP) over an extended period of time affects medical and dental care use, health status and academic achievement. In contrast to prior research, which focuses on the program's effects during infancy and early childhood, we examine CHIP enrollment among elementary and middle school students. Using the 1999-2007 panels of the Early Childhood Longitudinal Study, Kindergarten Class and an instrumental variables model to address selection bias, we find that an additional year of CHIP enrollment increases the regular use of routine medical care by 16 percent, but has no detectible effects on overall parent-reported health status, obesity or test scores in reading and math.

We study the relationship between macroeconomic conditions and health in the United States using data from the Behavioral Risk Factor Surveillance System between 2004 and 2017. Unlike many existing studies that use state or county as the level of aggregation, our analysis uses a sample of metropolitan and micropolitan statistical areas. Our results suggest strong associations between worsening macroeconomic conditions and reduced access to care and health insurance coverage. While we do not detect any robust relationships between business cycles and health outcomes in the general population, we do find consistent evidence of worse self‐reported health during economic downturns among minorities and less‐educated individuals. In addition, there is some suggestive evidence of countercyclicality of healthier lifestyle choices. However, the findings for health behavior outcomes are not robust to adjusting for multiple hypothesis testing.

Using a sample of New York City restaurants, we examine the relationship between a wine's bottle margin and whether the restaurant offers that same wine by the glass. We find that restaurants offer less expensive wines by the glass but set higher margins on these bottles than for similar wines offered only in bottles. Overall, offering wine by the glass is associated with a 5.0% increase in the bottle price and a 12.2% increase in the bottle margin. We find similar results for retail and wholesale markups of wine bottles. Our results offer evidence that settles a theoretical ambiguity in the menu-pricing literature (Anderson and Dana, 2009) about whether to raise or lower the price of a high-quantity package when introducing a low-quantity package of a good, as it applies to restaurant wine pricing. (JEL Classifications: L11, L83)