The Study Of Systemic Risk Measurement And Its Spillover Effects In Financial Industry

With the acceleration of financial innovation and the acceleration of financial mixed business,business transactions among financial institutions have become more frequent,financial innovation products have become more diversified,and the multi-level financial market has been gradually improved.At the same time,the financial industry faces risks,in the interaction of a variety of factors,risk points hidden,complex,sudden,which is prone to the problem of risk cross-infection,further structural imbalance problems,and finally a high probability of systemic phoenix risk.Risk measurement and prevention of risk contagion across industries have been the subject of extensive academic attention.Based on this,this thesis investigates the measurement of systemic risk and spillover effects of different industries of financial institutions using Realized GARCH model based on high frequency data and DY spillover index method.This thesis measures industry systemic risk and risk contagion by selecting intraday high-frequency trading data of listed companies in the financial sectors of banks,securities,insurance and funds from 2018 to 2022.The specific empirical process is as follows: first,the sample data were initially analyzed and the ARCH effects were eliminated by parameter fitting based on Realized GARCH models with three different distributions,further the value at risk(Va R)of each industry was predicted by rolling forecast method,and the reliability and accuracy of the risk measure results of the model were tested by unconditional coverage test and conditional coverage test.It also confirms the validity of the high-frequency data model(Realized GARCH model),which provides more intra-day information,for risk measurement,making the risk measurement more accurate.Second,the optimal Copula function for each financial industry is screened,and the nonlinear dependence between the financial industry and the financial system is calculated,on which the dynamic conditional value at risk(Co Va R)is inscribed.Finally,the DY spillover index method is used to construct the risk spillover matrix of financial markets to analyze the degree of risk spillover and dynamic spillover effects.Based on the theoretical basis and empirical operations,the analysis in this thesis leads to several conclusions: First,descriptive statistics and heteroskedasticity tests on the sample data reveal that the data have characteristics such as spikes and thick tails and heteroskedasticity,indicating the rationality of establishing GARCH class models.Second,analyzing from the model fitting perspective,the Realized GARCH model with t distribution outperforms the models based on normal and skewed t distribution from the perspective of AIC,BIC criterion and log-likelihood function values,indicating that the t distribution is the best fit for return volatility.Third,from the perspective of calculating the value-at-risk,the Va R prediction with a fixed window length sliding one step forward to calculate the out of sample tail level of 5% is robust,banks have a better risk prevention ability than other industries,the securities industry has the worst risk defense ability,and the securities industry has the greatest risk contagion effect on the entire financial market when extreme situations occur in the entire financial system.Fourth,analyzed from the perspective of risk spillover,the volatility spillover index indicates greater risk transfer within each sector than between sectors,suggesting a greater lagged impact of the sector itself.The securities and insurance industries are net transmitters of risk and have a strong influence on other industries.From the dynamic spillover effect graph,it is found that the spillover index generally changes with the occurrence of financial extreme events.