LAUSR

China has become one of the most serious PM2.5-dominated air pollution country. Despite a great deal of research has focused on analysing the influence of social and economic driving forces of PM2.5 pollution in China, most research in existence either applying mean regression or failing to consider the spatial autocorrelation. Motivated by this, this paper utilizes a Bayesian hierarchical spatial quantile regression method to explore the effect of socioeconomic activity on PM2.5 air pollution. By introducing spatial random effects into the model, the spatial autocorrelations of residuals are significantly reduced. The empirical study demonstrated that the PM2.5 concentration levels were strongly correlated with total population, urbanization rate, industrialization level and energy efficiency at all quantiles. For upper quantiles, the impact of urbanization rate on the haze is the greatest among all the predictors, then followed by the total population; while for lower quantiles, industrialization has the greatest impact on the PM2.5 concentration. The impacts of energy efficiency in the lower 15% and upper 15% quantiles are higher compared to any of the other quantiles.