The Beijing-Tianjin-Hebei region has been plagued by atmospheric pollution for many years. Adjustment of industrial structure can harmonize the relationship between economic development and environmental protection. Studying the effect of industrial restructuring on air pollution has become an urgent task in reduction of regional air pollutant emissions. At present, the research on the impact of industrial structure adjustment on air pollution mainly adopts three methods:EKC curve verification, time series analysis, and spatial econometric method. This research selects time series data of Beijing, Tianjin and Hebei's industrial structure and certain air-pollutant emissions during 1999 to 2015 to establish the vector autoregressive (VAR) models for the secondary industry proportion, sulfur dioxide and industrial smoke (dust) emissions, then completes the impulse responses function (IRF) analysis. The findings are that the impact of industrial structure adjustment on different pollutants in different regions are various in terms of direction, degree and duration. In terms of the direction, industrial restructuring cannot reduce emissions of all pollutants in all provinces. In terms of the degree, SO₂ emissions are greatly affected and industrial dust emissions are less affected; emissions in Beijing and Hebei are greatly affected, while the impact in Tianjin is small. In terms of the duration, SO₂ can be affected for a long time, but the impact on industrial dust can't last long. In Beijing and Tianjin, the positive impact can last for at least 4 years, while Hebei has only a period not longer than 2~3 years. Therefore, to ensure a high efficiency of industrial restructuring policies, Beijing-Tianjin-Hebei should take a three-year reference cycle to adjust the policies. Besides, different control methods should be adopted for different pollutants. And air pollutant emission reduction work should match the local reality.

Beijing-Tianjin-Hebei  /  industrial structure adjustment  /  air pollutant emissions  /  vector autoregression model  /  impulse response function