Research On VaR Of Chinese Stock Market Based On High Frequency Data

The academic and financial regulatory authorities always focus on the riskmeasurement of financial market. Most of traditional risk measurements are usingGARCH model or SV model based on low-frequency intraday data. Although thesemodels can measure the volatility of time series well, because of the stock markettransaction frequently in every day, low-frequency data model will loss a large numberof daily important information. The current study shows that, the traditional GARCHmodels cannot be directly used to estimate high-frequency volatility. It is significant toestablish an efficiency risk measurement model based on high-frequency data. So thatwe can provide an effective theory method and policy recommendations for risk controlof financial institutions and regulatory authorities.Based on previous research on realized high-frequency volatility,this papercompares realized volatility, realized bipower volatility and weighted realized volatility.Lighting of long memory of WRBV, establish ARFIMA-WRBV-VaR model to measurethe risk of Chinese stock market. At the same time, we compare it with GARCH modelbased on daily-frequency data.The result shows the following facts: The predictive efficiency based onARFIMA-WRBV-VaR model is better than EGARCH-VaR model. The realizedhigh-frequency volatility has several features, such as, jump point, calendar effects, andlong memory. Those features are affected by the microscopic structure of the market,especially the asymmetry information and investor psychology. According to these, weput forward some policy suggestions for risk supervision, such as, improve theinformation disclosure mechanism and enhance the quality of investors.