Can industrial agglomeration promote pollution agglomeration? Evidence from China

After industrial agglomeration has developed to a certain extent, its negative externality will appear gradually, and the problem of environmental pollution will become serious. In this study, a comprehensive index of pollution emission is calculated by the improved entropy method, and the pollution agglomeration level is obtained through the geographical concentration. The degree of industrial agglomeration is estimated by the location quotient index. The impact of industrial agglomeration on pollution agglomeration is investigated by using the model of Geographically and Temporally Weighted Regression (GTWR). The main results are as follows. (1) There is spatial autocorrelation between pollution agglomeration and industrial agglomeration in China. (2) At the national level, industrial agglomeration increases the pollution agglomeration. (3) At the provincial level, industrial agglomeration has a positive effect on the pollution agglomeration, but the degrees of influence in different regions are different. The positive effect in the northern region (such as Inner Mongolia, Heilongjiang, and Jilin) and the eastern region (such as Beijing, Tianjin, and Liaoning) is the strongest, whereas the effect in the western region (such as Sichuan, Guizhou, and Guangxi) is on the low side. (4) The relationship between industrial agglomeration and pollution agglomeration is relatively stable from the time perspective, and it may be hard to weaken the positive effect over a short time. To improve the environmental quality, this study puts forward some suggestions, such as optimising high-tech industrial agglomeration and strengthening the regional linkage. •This paper integrates the temporal and spatial effects into the model.•Industrial agglomeration is the most important factor.•The impact of industrial agglomeration has spatial heterogeneity.•The fluctuation of regression parameters in different regions is different.•The regression parameters in adjacent areas have little difference.