Research On Financial Risk Measure CVaR And Its Application In China’s Securities Market

Risk is the eternal topic of concern in the financial sphere. Prevention andmanagement of financial risk has become the core of the entire economic security. Riskmeasurement is the quantitative analysis and evaluation of risk, which is an important partof the risk management. Based on the emerging financial risk measurement-ConditionalValue at Risk as the main research object, we introduces its basic principle and method ofcalculation. Meanwhile, we combined with the daily yield sequence of Shanghai compositeindex and Shenzhen component index and their high frequency yield sequence of fiveminutes, to measure the risk of China’s securities market. Employing the GARCH modelswith errors follow normal distribution、student t distribution and GED distribution toestimate the volatility of the low frequency yield sequence. Modeling the realized volatilitycalculated on five minutes returns by the HAR-RV model and MIDAS model. We expandthe HAR-RV model into a HAR-RV-GARCH model based on the regression results, andobtain a very good prediction effect. After getting the predicted values of the conditionalvariance, we calculate the risk measure values of VaR and CVaR sequence underdifferent confidence levels through R programming. We evaluate the validity ofVaR under different models through BACKTESTING. At the same time, in order toexplore the superiority of CVaR, we make a comparison analysis between theCVaR sequence and the VaR sequence as well as the actual loss sequence. Ultimately, weselect the optimal model of securities market’s risk measurement. The final conclusion is:CVaR calculated with low and high frequency volatility models has higher measuringaccuracy thanVaR, which illustrate that CVaR is a more effective risk measurement toolthan VaR; CVaR of high frequency data is more robust and effective than that of lowfrequency data, and CVaR calculated with HAR-RV-GARCH model is the most effective.