"A Torrential Ordeal: Understanding Health Damages of Flooding" (with Elaine Hill) [Draft Available Upon Request]
Climate change is likely to alter the intensity and spatial distribution of precipitation (Arnell and Gosling, 2014; Alfieri et al., 2017). Consequently, many places will face an increased risk of flooding, making it important to understand how extreme weather events impact welfare. As we show in this paper, a mother experiencing flooding during her pregnancy can lead to worse health at birth for her child. We achieve this by combining NCHS natality records with flood inundation maps produced from satellite images compiled in the global floods database. By combining these records, we show that, relative to children that would have experienced a flood after birth, the health of children whose mothers experienced a flood in their county of residence were .7% more likely to be born premature. We also find a non-linear relationship between the intensity of the flood and birth outcomes that provides direction for future work.
“Wildfire Smoke and Drug Overdoses” (with Elaine Hill, Konstantin Kunze, and Jamie Mullins ) [Draft Available Upon Request]
Wildfire smoke is an increasingly prominent source of fine particulate matter in the United States. There has been an established line of literature connecting poor air quality with declines in cognitive abilities, and recent work has shown that wildfire smoke exposure raises suicide risk, specifically in rural areas (\cite{molitor2023air}). There is little known about how smoke affects drug- and opioid-related fatal and nonfatal overdoses. This paper examines whether acute changes in wildfire smoke exposure have an effect on county-level drug and opioid overdose rates, and whether impacts differ by age, race, sex, education, or rurality. We assemble a monthly county-level panel for the United States from 2007-2023 that links NCHS mortality data on all drug and opioid-specific deaths with satellite data of wildfire smoke plumes from the Hazard Mapping System used in Molitor et al. We find that across all counties, an additional medium-or-thicker smoke day is associated with an increase in all drug deaths by 0.028 per million residents (0.16 percent of the mean). We also find evidence that smoke exposure elevates fatal overdose risk for younger individuals and some measures of nonfatal overdose risk among the publicly insured population (both Medicare and Medicaid).
"A Storm Brewing: Costs of Climate-Induced Pollution in Drinking Water" [Working Paper] (Revise and Resubmit, Journal of Health Economics).
This study uses Florida birth and water quality records (2013 – 2021) to estimate how tropical cyclones, which climate change may intensify, harm infant health by damaging drinking water infrastructure. Identification exploits the random timing and path of storms, comparing exposed and unexposed births from the same mother. I distinguish between residential and water source exposure to isolate water pollution as a key mechanism that mediates cyclones’ health impact. Source exposure increased coliform detection by 17\%, premature births by 6.4\%, and cost society over \$344 million annually. These findings are robust to common identification challenges and underscore climate change’s health risks.
"Are "Killer Trees'' Helpful for Understanding the Effects of the NOx Budget Program?" (with Lala Ma, Lucija Muehlenbachs, and Sagbo Anicet Hounton) [Working Paper] [Resources for the Future Working Paper 25-24] (Revise and Resubmit, Journal of Public Economics).
Forests play a largely unrecognized role in determining the effectiveness of air-quality regulations. We study the NOx Budget Trading Program, a cap-and-trade system covering 19 U.S. states designed to curb ground-level ozone. We show that evergreen forest cover, a major source of VOCs, can be used to determine where NOx reductions translate into the largest ozone reductions. The most heavily forested areas experienced ozone reductions four times larger than less-forested areas. However, we demonstrate that mortality reductions do not mirror these pollution reductions. For policy design, our findings illustrate how ignoring cross-pollutant interactions and heterogeneous marginal damages can lead regulators to mis-estimate benefits.
"Housing Supply Responses to Floodplain Regulation." (with Sarah Brown)