LAUSR

Based on the panel data of 280 prefecture-level cities in China from 2014 to 2018, we construct the digital divide index from three aspects including "access divide", "use divide" and "efficiency divide", using Thiel index to measure income inequality and study the impact of the digital divide on income inequality between different regions. We apply the two-stage spatial DID combined IV model. In the first stage, we introduce the "Internet + government affairs" policy to construct the spatial DID model, so as to obtain the estimated value of the digital divide index. When the estimated value is substituted into the two-stage regression as an instrumental variable, it is found that every 1unit increase in the digital divide will widen the income divide by 0.134 units. This conclusion is still robust under the replacement of Theil index and change of weights. The results of heterogeneity analysis show that digital divide has a more significant impact on the income divide in eastern China in terms of regions. It is found after distinguishing different types of digital divide that the other two types of digital divides have the most significant impact than income divide, 0.034 units higher than that of the traditional. Finally, we introduce the mechanism variables from three levels of "access divide", "use divide" and "efficiency divide", thus verifying the effect of information asymmetry, effect of human capital differentiation and effect of delaying industrial upgrade. Then we put forward the following policy recommendations: in order to reduce the deterioration effect of digital divide on income distribution, it is necessary to improve infrastructure, enhance the degree of digitalization of human capital and optimize the industrial structure.