Learning to Avoid Pollution: Public Health Bene ts from Beijing's Air Quality Index Forecast

Joseph E. Aldy,Jing Cao,Nan Zhong
2019-01-01
Abstract:An extensive literature has documented the adverse morbidity and mortality impacts of exposure to ne particulate matter, ozone, and other air pollutants. In addition to emission-reduction policies, governments around the world have provided information to the public to facilitate behavioral changes to limit their exposure to ambient air pollution. Given its high levels of air pollution, the City of Beijing began publishing online day-ahead pollution forecasts in 2012. We evaluate how these day-ahead pollution forecasts in Beijing enabled averting behavior by its residents. We present evidence that the forecast of a poor air quality day (conditional on realized air quality) contributed to an increase in search activity for mask and air pollution on the Baidu search engine. This suggests that individuals respond to pollution forecasts to seek out ways of mitigating their exposure to air pollution. We then exploit errors in the pollution forecast to identify the impacts of forecasting air pollution again conditional on realized air quality on health outcomes, in particular, ambulance call rates for cardiovascular risks. We nd that the day-ahead pollution forecasts contribute to a statistically signi cant 5 percent decline in cardiovascular ambulance calls. We assess the robustness of this nding with a series of placebo analyses that focus on *This paper has been prepared for the for the Harvard Center for Risk Analysis Risk Assessment, Economic Evaluation, and Decisions workshop, September 26-27 2019. Aldy is a liated with the Harvard Kennedy School, Resources for the Future, National Bureau of Economic Research, and Center for Strategic and International Studies, joseph_aldy@hks.harvard.edu, Ofer 403, Mailbox 114, Harvard Kennedy School, 79 JFK Street, Cambridge, MA 02138; Cao is a liated with Tsinghua University, jingcao2007@gmail.com; and Zhong is a liated with Xiamen University, nanzhong@xmu.edu.cn. We thank Blake Barr for excellent research assistance and Aldy thanks the Taubman Center for State and Local Government for nancial support.
What problem does this paper attempt to address?