The heterogeneous effects of socioeconomic determinants on PM2.5 concentrations using a two-step panel quantile regression

The haze pollution caused by high PM2.5 concentrations has adverse health effects worldwide, especially in rapidly developing China. As meteorological conditions are uncontrollable, this study aims to investigate how anthropogenic factors affect the PM2.5 concentration under high, medium and low emission levels. The distribution of socioeconomic variables is often non-normal, with important information hidden in the tail. By using balanced panel data of 273 Chinese cities from 2010 to 2016, two-step panel quantile regression is adopted to examine the cross-quantile heterogeneity of seven socioeconomic variables: economic growth, industrial structure, urbanization, foreign direct investment (FDI), population density, public transportation and energy consumption. The empirical results show that the relationships of PM2.5 concentration with economic growth, urbanization, industrialization and FDI are heterogeneous. Compared with other variables, population density exerts the greatest positive effect on PM2.5 pollution across all quantile cities. Moreover, the impact of GDP per capita on PM2.5 concentration in the lower 25th quantile cities is stronger than those in the 25th-50th, 50th-75th and upper 75th quantile cities. The effects of FDI in the upper 75th and lower 25th quantile cities are higher than those in the 25th-50th and 50th-75th quantile cities, which supports the "pollution haven" hypothesis. The impact of industrial structure on PM2.5 concentration in the upper 75th quantile cities is larger than those in the 0-25th, 25th-50th, and 50th-75th quantile cities. The heterogeneous effects of these socioeconomic determinants could assist policymakers in implementing differentiated policies that fit cities with different levels of air pollution.