Social Network And Macroeconomic Uncertainty Spillover ——Base On Microblog Data And Exponential Random Graph Model

The matter of macroeconomic uncertainty has gradually aggravated in recent years,a growing number of scholars start to focus on the regional macroeconomic uncertainty and the spillover effect of macroeconomic uncertainty between regions.we consider the computing method of uncertainty spillover at the national level to measure and analyze the uncertainty spillover among provinces in China.Further considering the influence factors of uncertainty spillover,this paper innovatively applies the provincial social network data to macroeconomic research.This paper uses microblog user data to measure the social relationship between provinces,constructs the social association network between provinces,and explores whether the social association between provinces has a certain impact on its uncertainty spillover,so as to better analyze the spillover mechanism of macroeconomic uncertainty between provinces in China.Firstly,we crawls and processes the microblog data.Referring to the way of building social connection index by Facebook,we built inter provincial social connection network and makes exploration and analysis.The results show that in general,China's provinces are most closely related to themselves,while other provinces pay more attention to areas like Beijing and Shanghai.More importantly,regions with relatively backward economic development are more likely to be associated with regions with better economic conditions than themselves,and neighboring provinces usually have higher degree of social association.In the construction and analysis of China's macroeconomic uncertainty spillover network,hierarchical dynamic factor model and generalized variance decomposition are mainly used.Through the analysis,it is found that the phenomenon of cross provincial transmission of macroeconomic uncertainty is significant.At the same time,the dynamic analysis shows that the spillover in the whole country and provinces has big event effect.In addition,there are also great differences in uncertainty spillover and spillover between regions.The province's uncertainty spillover is relatively stable,while the spillover effect fluctuates greatly.Finally,based on the construction of the above two networks,we use the exponential random graph model to explore the factors that affect uncertainty spillover network.we choose the social association network as the covariate network and add it into the exponential random graph model to study the impact of social association network on the uncertainty spillover effect between provinces at different times.The results indicate that the inter provincial social association has a notable effect on its uncertainty spillover.Further robustness test shows that the estimation coefficients of the provincial association network of each region model are significant,and the results of the model are robust.