UC Berkeley

This dissertation contains three empirical studies that explore the health impacts of two dimensions of environmental quality, outdoor air pollution and extreme ambient temperature. All three studies rely on high-resolution, publicly available data on air pollution concentrations and weather, as well as restricted-access administrative data on individual-level mortality outcomes and hospital utilization. Chapters 1 and 3 explore the potential for health insurance to mitigate health damages from air pollution and temperature, respectively, while Chapter 2 incorporates recent advances in air particle dispersion modeling to assess whether health benefits projected under policies to reduce power plant emissions actually manifest in practice. The first two chapters generate causal estimates of the effects of poor air quality on health outcomes, while the third primarily provides suggestive evidence that public health insurance may play a role in driving heterogeneity in the relationship between extreme temperature and health.

In the first chapter, I investigate the potential for health insurance to mitigate adverse health impacts of air pollution in the United States -- a country that lacks universal health care and continues to bear a large health burden from air pollution every year. To explore this issue, I focus on California as a case study, relying on exogenous variation in pollution exposure derived from wildfire smoke and a policy-induced increase in public health insurance coverage that varied in its regional intensity. Estimates for the state as a whole imply a 3\% reduction in the health burden of air pollution for each one percentage point increase in the publicly insured rate. Back-of-the-envelope calculations based on these estimates imply a lower bound of \$250 million per year in newly-measured benefits from California's ten percentage point Medicaid expansion under the Affordable Care Act. These results suggest that extending health insurance may be one way to address the increasing threat of wildfire smoke exposure in the western U.S., and potentially air pollution more broadly.

The second chapter, co-authored with Meredith Fowlie and Edward Rubin, focuses on the role of co-benefits in benefit-cost analyses of air pollution reduction policies. An environmental regulation generates pollution `co-benefits' when it indirectly induces reductions in pollutants that are not the target of the authorizing legislation. In recent EPA rulemakings, quantified pollution co-benefits have exceeded directly targeted benefits. This increased prominence has invited increased scrutiny. We aim to advance the theory and inform the practice of air pollution co-benefits accounting, focusing in particular on the health impacts of reduced exposure to small particulate matter (PM$_{2.5}$). We integrate a model of air particle dispersion into an empirical analysis of how reductions in power plant emissions impact local air quality downwind. We highlight the potential role of interactions with existing policies, documenting suggestive evidence that changes in power plant emissions have had differential impacts across areas that are more or less constrained by pre-existing air quality regulations. Future work will incorporate more precise modeling of pollution trajectories to explore these complex relationships, and associated health outcomes in greater detail.

In the third and final chapter, I return to the role of health insurance as a potential mediator of environmental health damages, in this case focusing on extreme temperature rather than air pollution. While the relationship between ambient temperature and human health has been studied at length by researchers across disciplines, this study is one of the first to consider the impacts of temperature on morbidity, rather than mortality, in a standard econometric framework. Relying on administrative hospital data from the state of Texas, I find that both extreme cold and extreme hot temperatures are associated with higher health care utilization, relative to a mild day. I also find that public health insurance may drive heterogeneity in the temperature-health relationship. Specifically, I estimate that health care utilization for temperature-related illness is relatively lower among adults insured by Medicaid, whereas the opposite holds for children. One potential explanation for this pattern is that the positive health benefits of public insurance strengthen over time, and that the lower relative benefit for children simply reflects their fewer number of potential years in the program. While these Medicaid-related estimates are suggestive rather than causal, this study provides novel evidence that income (which is correlated with Medicaid enrollment) may drive important heterogeneity in the relationship between climate and human health.