Panel Paper:
The Environmental and Health Impacts of Transboundary Air Pollution from China to the U.S
*Names in bold indicate Presenter
In this study, we leverage a natural experiment using the exogenous variation in the timing of major events in China which strongly influence China's pollution concentrations. For example, the timing of Chinese New Year (CNY), a major Chinese holiday where many industries close for the week, varies each calendar year since it is based on the lunar new year. In which case, the timing of these events could be considered "as good as random," since they are unlikely correlated to factors that affect pollution trends in the U.S.. More specifically, we examine the impacts on pollution and health in California (CA), home to 38 million people, because in this region, winds generally come from the West or the Pacific Ocean, while weather on the East Coast is more influenced by winds from the Gulf of Mexico. We also focus on the pollutant, particulate matter (PM), since it is often used as an indicator of urban air pollution.
First, we find statistically significant relationships among daily lagged pollution and major events that reduce or increase pollution in China on air quality in CA. For example, the literature suggests the long range transport (LRT) of air pollution between East Asia and North America can take between 5 days and 2 weeks. In our study, we find if CNY occurred in China 2 weeks ago, today's CA PM2.5 decreased by 20%. Second, we obtain reduced form estimates of the effects of these events in China on health in CA, where for the latter we use restricted daily data on emergency department (ED) and patient discharge data (PDD) visits for respiratory and heart-related diseases between 2005 and 2012. We find that several policies in China intended to reduce pollution levels in preparation for the 2008 Beijing Olympics were associated with a 7 to 9% decrease in average daily ED visits for respiratory disease in CA. We also conduct several falsification tests and robustness checks. First, our reduced-form health estimates are robust to accounting for possibly confounding variables, including zip code fixed effects (FEs), seasonality and time-varying unobservables at the county-level. Second, we observe impacts on the number of injuries in CA as a falsification test and find no significant relationship, as expected. These results have important implications regarding international environmental cooperative policies for transboundary air pollution, which is not only a problem for the U.S., but globally.