Spatio-temporal boundary effects on pollution-health costs estimation: the case of PM2.5 pollution in Hong Kong

Abstract

In this study, we estimate PM2.5-caused health costs in Hong Kong and examine spatio-temporal boundary effects on the estimated results. During the period between 2012 and 2016, mean annual welfare loss from PM2.5 pollution is estimated to be US$1.5–1.8 billion or 0.5%–0.7% of Hong Kong’s gross domestic product. Premature deaths associated with chronic exposure are the most important health endpoint, accounting for >95% of the total costs. The estimated results are subject to large spatio-temporal boundary effects. On the one hand, disregarding cross-district heterogeneity in air quality and socioeconomic conditions leads to a downward bias of up to 13%, due to spatial correlations among PM2.5 levels, district population, and household incomes. On the other hand, neglecting intra-year variations in PM2.5 concentrations results in overestimation of up to 18%, due to nonlinearity in concentration–response relationships. The estimation bias from coarse analysis units likely further increases in national or global studies, given the magnitude of the spatio-temporal variations involved at these levels.