Applied policy research often examines interventions that differ in design and whose effects on downstream outcomes are uncertain in direction and magnitude, complicating welfare evaluation. We develop a transparent framework to synthesize evidence across heterogeneous policy instruments that operate through a shared intermediary outcome but lack a common treatment scale. We apply this approach to youth e-cigarette regulations and their impacts on youth smoking using data from a systematic review. We consider evidence from 11 natural experiments included in a recent systematic review, which span e-cigarette minimum legal sale age laws (MLSAs), e-cigarette taxes, and flavored e-cigarette restrictions. We standardize these heterogeneous policies based on their estimated effects on e-cigarette use and estimate their downstream effect on youth cigarette smoking using fixed and random effects meta-analysis. The evidence overall supports the conclusion that e-cigarette availability reduces youth cigarette smoking. More broadly, our framework provides a practical approach for synthesizing policy evidence when interventions share a common objective but are not directly comparable on a single treatment scale.

• Working paper available here.

We estimate the causal effects of state hearing aid coverage mandates on labor market outcomes, public benefit receipt, and well-being measures among adults ages 45–64 using the Current Population Survey Annual Social and Economic Supplement from 2003 through 2023. We exploit staggered variation in private insurance and Medicaid mandate adoption and repeal across three policy regimes, using a cohort-based synthetic difference-in-differences estimation approach. Among Medicaid-enrolled adults, mandate adoption is associated with a 6.4 percentage point increase in labor force participation and a 7.2 percentage point increase in employment. These effects are qualitatively large, representing relative increases of 18% and 23%, respectively. Effects of private insurance mandate adoption are broadly null on the extensive margin, with the exception of an estimated 1.35 percentage point increase in full-time work conditional on employment. Medicaid repeal effects are directionally consistent with adoption. Effects on public benefit receipt and well-being are generally null or mixed, and event studies are suggestive of differential pre-trends. The large Medicaid adoption effects are consistent with hearing aid coverage improving communicative function among a highly health-limited population, though the magnitudes warrant further investigation.

• Pre-analysis plan here.

• Presented at American Society for Health Economists 2025 Annual Meeting

The Earned Income Tax Credit (EITC) may improve population health and reduce racial health disparities, but its health effects may vary with the labor market contexts in which it is implemented. We analyze data on physical and mental health among non-Hispanic Black and White single mothers without a college degree using 1993–2022 waves of the Behavioral Risk Factor Surveillance System. We use a continuous treatment difference-in-differences approach to estimate the effects of state EITC expansions on health and assess whether effects differ by maternal race and the degree of racial inequality in state employment rates. Refundable state EITC expansions were associated with significant reductions in fair or poor overall health, frequent poor physical health, and frequent psychological distress among single mothers, with no corresponding effects among married mothers. Further, stratification by the Black-White employment rate ratio indicated that non-Hispanic Black single mothers experienced disproportionately larger reductions in psychological distress than non-Hispanic White single mothers, with the largest differences observed in states with greater racial labor market equality. These findings suggest that refundable state EITC expansions may reduce racial disparities in psychological distress among single mothers and that the labor market contexts in which work-based safety-net policies are implemented may shape their health equity effects.

We study the health consequences of redistributing political power through the 1975 extension of the Voting Rights Act, which eliminated barriers to voting for previously disenfranchised nonwhite populations. The intervention led to broad declines in under-five mortality but sharply contrasting effects in other age groups: mortality fell among non-white children, younger adults, and older women, yet rose among whites and older non-white men. These differences cannot be reconciled by changes in population composition or material conditions. Instead, we present evidence suggesting psychosocial stress and retaliatory responses arising from perceived status threat as key mechanisms.

• NBER working paper here.

Youth and adolescents have faced an unprecedented confluence of stressors to their mental health in the wake of the COVID-19 pandemic. This study draws on individual-level electronic health records from a large and geographically diverse cohort of primary care patients to examine how the COVID-19 induced labor market shock of Spring 2020 affected the prevalence of common mental health conditions among youth ages 12–17. We leverage variation in the impact of the pandemic-induced recession on county-level, sector-specific employment to estimate a series of generalized difference-in-differences models to compare outcomes across counties with more versus less exposure to the pandemic economic shock. Despite an overall decrease in the use of primary care among our study sample, we find that the frequency of primary care visits with positive diagnosis for mental health conditions increased following the pandemic onset. We find significant increases in the probability of positive mental health diagnoses among patients in counties with larger pre-pandemic shares of sectors that were most vulnerable to pandemic-induced contractions. We further find that this overall effect is primarily driven by visits for ADHD, anxiety, and affective mood disorders. Triple-differences estimates suggest larger increases were experienced by patients with pre-existing mental health diagnoses, patients with non-private insurance coverage, and among White patients. 

• Presented at American Society for Health Economists 2023 Annual Meeting (Slides), APPAM Fall Research Conference 2023, Southern Economic Association Annual Meeting

We study how the 2015 implementation of the California Earned Income Tax Credit (CalEITC) was associated with changes in behavioral risk factors for colorectal cancer – the second leading cause of cancer mortality in the United States and a disease characterized by large socioeconomic disparities. Due to the unique benefit design of the CalEITC, we employ a synthetic control approach with data from the Behavioral Risk Factor Surveillance System to examine whether and how the policy affected a wide range of modifiable risk factors underlying the growing burden of colorectal cancer among low-income populations. Relative to post-2015 trends in our empirically constructed ``synthetic California", we find significant reductions in the population shares who are obese, those reporting greater than 14 days of poor physical and mental health, and those reporting that cost interfered with care. We find null effects across six additional outcomes for which a suitable pre-treatment fit was uncovered. We find no outcomes for which there was an increase in risk. This study presents preliminary evidence of the role that targeted income supplementation policies may play in reducing colorectal cancer risk over the long-term. 

• View working paper here.

• Presented at the 2024 Annual Conference of the American Society of Health Economists (Presenter: Rong) 

Read more about this project and other initiatives of the UPSTREAM Center here.

Using an event study approach and data from the US Behavioral Risk Factor Surveillance System from 1993–2019, we examine the effect of state-level minimum wage increases on self-reported mental, physical, and general health status and alcohol use among a sample of young adults likely exposed to minimum wage hikes. Our event study approach avoids the pitfalls in recent two-way-fixed effects studies and controls for long-term compositional changes, such as education level, proportion of Hispanic workers and health status of Hispanics. Since minimum wages disproportionately affect minorities, we estimate effects separately by race/ethnicity. We find  precise null effects among non-Hispanic whites, and imprecise null effects among Hispanics. By contrast, our analysis finds that increased minimum wages improve the  health of non-Hispanic black workers, particularly improving self-rated general health, number of poor mental health days and the rate of mental distress. (Sponsored by the Robert Wood Johnson Foundation Policies for Action Research Hub at the University of California, Berkeley)

• Presented at Population Association of America Annual Meeting 2025 (Schess), UC Berkeley conference on Minimum Wages and Monopsony Power (Dow), and 2025 Annual Conference of the American Society of Health Economists (Lowenstein)

The Earned Income Tax Credit (EITC) is the largest anti-poverty program in the U.S. for working families, however its effects on income inequality by race are not well understood. Because the EITC is targeted toward low-wage workers, it has profound potential to promote equity by reducing race inequality in post-tax income. However, qualifying for the EITC requires participation in labor markets that are characterized by structural racism and inequality. Effects of the EITC on income inequality may differ across these labor market conditions in theoretically ambiguous and potentially complex ways due to EITC eligibility and schedule rules. Using the 1987-2021 March Annual Social and Economic Supplement to the Current Population Survey, we evaluate effects of EITC expansions on after-tax income inequality for non-Hispanic Black and White single mothers. We first test whether effects differ by race, and then evaluate whether differential effects by race vary according to the labor market conditions in a woman’s metropolitan statistical area (MSA)--particularly, the percentage of unemployed persons in the MSA and the degree of employment inequity measured as the Black-White employment ratio. Findings illuminate the conditions under which the EITC is most, and least, effective at reducing racial-ethnic inequality. 

• Presented at the 2024 Association for Public Policy and Management Fall Research Conference

• Presented at Population Association of America Annual Meeting 2024 (Strully) 

The recent rise in suicide across racial and ethnic groups defies easy explanation. This paper examines the role of “perceived economic threat” as a potential driver of suicide in an effort to better understand the trends and racial patterning of suicide among working-age adults in the United States. I build on existing research suggesting that the false perception of economic parity across racial and ethnic groups in the United States may serve as a distal driver of suicide by fueling fears of job loss and more proximal determinants of suicide such as shame and anxiety. To test this hypothesis, I estimating a series of panel fixed effects and long difference models in which I use group-specific employment rates to proxy for changes in economic opportunities at the local level. To mitigate potential bias due to differential labor supply effects, I estimate a series of instrumental variables models in which I instrument for changes in employment rates across racial and ethnic groups using a Bartik-style shift share instrument. In my preferred stacked first difference models, I find consistent evidence that once controlling for own-group employment, suicide rates among non-Hispanic White adults vary countercylically with employment rates among non-White segments of the labor force. This cross-group protective effect is especially strong among females, for whom I estimate large reductions in suicide by firearm and intentional overdose in response to employment rate increases among non-White adults. 

• Presented at American Society for Health Economists 2023 Annual Meeting (Slides) 

• Presented at the Interdisciplinary Association for Population Health Sciences Annual Conference (Poster)