Risk Measurement Research Based On Quantile GARCH Models In China Stock Market

Driven by globalization and liberalization of the financial sector, the global financial will have to bear collective responsible and loss caused. When a financial crisis happened in one country, the global economy will suffer serious economic losses due to the impact of a chain reaction. The domestic existing literature is mainly based on the traditional quantile GARCH model to measure financial risk. In this context, the use of quantile GARCH expand research on financial risk measure is particularly important, through the accurate quantification of risk levels and reflecting the extent of the loss of return on assets, which makes investors in a timely manner to avoid market risks and reduce losses, but also contribute to the relevant regulatory authorities make response measures for the financial market volatility and ensure good market to operate effectively.The traditional quantile GARCH regression model is only starting from the market risk itself, and it does not involve other factors on the impact of market risk.This paper is based on the traditional quantile GARCH regression model theory,taking into account the existence of trading liquidity in the financial market and non expected trading volume on the field, the author come up with Liquidity-Adjusted QGARCH model and QGARCH volume model. Through the selected Shanghai Composite and Shenzhen Component Index Return QGARCH, LA-QGARCH and QGARCH-V three risk measurement model are established in order to compare different models measurement accuracy and stability at different points at the site of the Shanghai and Shenzhen stock exchanges, which explored QGARCH family-related differences in the prediction accuracy of the model in different markets at different sub-sites model. The results are that LA-QGARCH model and QGARCH-V model are at the same time very good performance in the 5% quantile on the prediction of the risk of Shanghai stock market, while in the 1% and 2.5%sub-site the traditional QGARCH of VaR forecast has the absolute advantage, which accuracy degree is much higher than other models, but the former is lower in high quantile;for the Shenzhen Component Index on the Shenzhen stock trading market,LA-QGARCH and QGARCH-V two models risk prediction accuracy is better in the1% and 2.5% low quantile, while in high quantile VaR prediction accuracy of traditional QGARCH model is relatively high.In summary, although LA-QGARCH model based on liquidity indicators andquantile GARCH-V model based on non expected trading volume is not the best in all sub-sites, but from the stock index return rate and VaR sequence comparison chart we can see that the constructed model is relatively stable, which provides a new idea for improving risk measurement model.