An Empirical Study On The Volatility Spillovers Across Financial Institutions

In recent years,the correlation of financial market volatility has become a way to measure financial risk,which provides an important basis for risk management.In order to further study the fluctuation correlation and the influence degree between financial institutions in China,and figure out which institution can lead the role under different marketing conditions,this paper studies the volatility spillovers among financial institutions.In this paper,the non-parametric approach proposed by Anderson et al.(2005)is used to calculate the realized volatility of financial institutions using high frequency data,and then based on the spillover indices proposed by Diebold and Yilmaz(2012)to measure the fluctuation spillovers between financial institutions.The basis of the spillovers is the Generalized Forecast Error Variance Decomposition in the vector autoregressive model.By calculation,It allows us to examine not only the overall level of volatility spillovers among different institutions in the system,but also the direction of fluctuation and the degree of mutual influence.This paper introduces the idea of lasso punishment in the vector autoregressive model fitting,and adds the Elastic Net penalty term.In addition,the paper uses the segmentation fitting model through the sliding window,so that we can get a clearer understanding of the dynamic changes of the fluctuation spillover level among financial institutions.Finally,based on the theory of complex network,this paper constructs a weighted complex network model by using the spillover between financial institutions,and systematically examines the topological characteristics and potential laws of financial networks.The institutions studied in this paper are 35 A-share listed companies in different financial sectors.By processing and analyzing the 5-minute high-frequency data from December 16,2011 to May 31,2016.We have the following conclusion:First,on average,the volatility spillover of the whole financial system is in a high level,among with a strong linkage and interaction;Second,for the entire financial system,when the market is in a turbulent state,including the bull market and stock market crash,the overall level of the market spill higher,and the securities sector spillover effect is significantly enhanced.When the market returns to a normal level,the overall level of volatility is low,but the role of the banks is improved.