Research

Instrumental variables built from thermal inversions are increasingly used to estimate the causal effect of air pollution on health, but researchers have largely overlooked how instrument construction choices affect validity and the reliability of the resulting estimates. I systematically evaluate how construction choices affect instrument validity and the reliability of causal estimates, using county-month-level U.S. data from 2010 to 2023 to study the effect of fine particulate matter (PM2.5) on cardiovascular mortality. I identify the construction that best matches actual ground-level conditions and is most robust to potential violations of the exclusion restriction, and find that a one-unit increase in PM2.5 raises cardiovascular mortality by 3.8 percent. Results are robust to a range of alternative specifications and pass placebo tests. A separate test finds no evidence of mortality harvesting. Effects are larger among Black, female, elderly, low-income, and rural populations. These results suggest that current air quality standards may understate the cardiovascular mortality costs of PM2.5 exposure, and that instrument construction matters for how these damages are estimated.

Estimates the causal impacts of smoking on mental health and functional limitations using BRFSS data and cigarette tax variation. Addresses smoking status misreporting via a bivariate probit two-step model, showing that corrected estimates substantially exceed standard estimates.

Many effects of pollution exposure on health and education outcomes of children have been identified, but little is known about education effects on preschool-aged children. We estimate the effect of particulate matter air pollution on preschool attendance using restricted administrative data from the state of Georgia. We use thermal inversions, weather phenomena that trap pollutants, as an instrumental variable for pollution. A one-unit increase in the county-week average ambient particulate pollution level decreases attendance by around 2 percentage points. Effects seem to be larger for racial minorities and for children with working parents. [NBER Working Paper No. w35367]

Uses staggered difference-in-differences over 2010–2019 to estimate whether targeted regulations on abortion providers impose broader financial costs on households beyond direct healthcare access.