The local impacts of coal and oil power plant retirements on air pollution and cardiorespiratory health in California: An application of generalized synthetic control method

Abstract

Background: This study capitalized on coal and oil facility retirements to quantify their potential effects on fine particulate matter (PM_2.5) concentrations and cardiorespiratory hospitalizations in affected areas using a generalized synthetic control method.

Methods: We identified 11 coal and oil facilities in California that retired between 2006 and 2013. We classified zip code tabulation areas (ZCTA) as exposed or unexposed to a facility retirement using emissions information, distance, and a dispersion model. We calculated weekly ZCTA-specific PM_2.5 concentrations based on previously estimated daily time-series PM_2.5 concentrations from an ensemble model, and weekly cardiorespiratory hospitalization rates based on hospitalization data collected by the California Department of Health Care Access and Information. We estimated the average differences in weekly average PM_2.5 concentrations and cardiorespiratory hospitalization rates in four weeks after each facility retirement between the exposed ZCTAs and the synthetic control using all unexposed ZCTAs (i.e., the average treatment effect among the treated [ATT]) and pooled ATTs using meta-analysis. We conducted sensitivity analyses to consider different classification schemes to distinguish exposed from unexposed ZCTAs, including aggregating outcomes with different time intervals and including a subset of facilities with reported retirement date confirmed via emission record.

Results: The pooled ATTs were 0.02 μg/m³ (95% confidence interval (CI): -0.25 to 0.29 μg/m³) and 0.34 per 10,000 person-weeks (95%CI: -0.08 to 0.75 per 10,000 person-weeks) following the facility closure for weekly PM_2.5 and cardiorespiratory hospitalization rates, respectively. Our inferences remained the same after conducting sensitivity analyses.

Conclusions: We demonstrated a novel approach to study the potential benefits associated with industrial facility retirements. The declining contribution of industrial emissions to ambient air pollution in California may explain our null findings. We encourage future research to replicate this work in regions with different industrial activities.

Keywords: Air pollution; Cardiorespiratory health; Industrial facility; Quasi-experimental methods.