Assessing the Effectiveness of Scrubber Installation on Air Pollution Emissions Reductions Among Coal-Fired Power Plants: Application of Statistical Methods for Causal Inference

Abstract

The 1990 amendment to the Clean Air Act implemented a cap-and-trade system that required electricity-generating power plants to dramatically reduce Sulfur Dioxide (SO2) and Nitrogen Oxide (NOx) emissions. Plants impacted by this legislation had a variety of compliance options, including decreasing factory operation, purchasing carbon credits, installing scrubbers, and changing fuel inputs. Using data from 1997-2012 of 995 coal-burning power plants, we examine the effectiveness of scrubber installation in reducing SO2 and NOx emissions. Specifically, we employ two methods—a propensity score algorithm and a matching algorithm—to estimate: 1) the causal effect of scrubber installation prior 1997 on the emissions during 1997; and 2) the causal effect of scrubber installation at any time during the period 1997-2012 on emissions two months following scrubber installation. Using a propensity score method, we found that pre-1997 SO2 scrubbers reduced 1997 SO2 emissions by 68% (95% CI 58% to 76%), and pre-1997 NOx scrubbers reduced 1997 NOx emissions by 28% (16%, 38%). Additionally, installing SO2 and NOx scrubbers at any time during the period 1997-2012 reduces SO2 and NOx emissions by 89% (88%, 90%) and 21% (19%, 24%) two months following installation, respectively. These final two results are corroborated by a matching algorithm, which finds scrubbers cause SO2 and NOx emissions decline by 88% (87%, 89%) and by 20%. (17%, 22%) two months following installation, respectively.