Selective siting or strategic cleaning : comparing Chinese ambient pollution monitoring data to remote sensing of air pollution

How well do ambient pollution monitoring data represent the true air quality in China? This paper evaluates two potential strategic reactions of local regulators in reducing monitor-based pollution reading without actually improving overall the air quality: selective siting, where monitors are placed in clean spots, and strategic cleaning, where areas near the monitors are targeted for more intense pollution control. Comparing monitoring data to satellite-derived estimates, we find no evidence of the former, largely because of the fact that the location of monitoring stations is determined mostly by the central authority. However, we find that monitors are gradually becoming the locally cleaner spot as the surrounding areas got cleaned up first. Further analysis suggests that this behavior is in response to a massive rollout of real-time air quality monitoring and disclosure program, which makes previously rampant data tampering practices virtually impossible at the local level. In sum, strategic cleaning could account for 5% of the total observed reduction in PM2.5 from 2011 to 2017 in China.