earticle

영어

In recent years, financial risk has become the attention focus of the economic entities around the world. How to effectively control financial risks has become a hotspot in financial community. Currently, the popular financial risk management method is Value at Risk (VaR), which has more practical values and reference values compared with the traditional financial risk management models. Taking the new composite index of Shanghai Stock Exchange (SSE) as sample, this paper calculated the VaR of GARCH, EGARCH and PARCHES models as well as their corresponding GARCH-M, EGARCH -M and PARCH-M models, analyzed the applicability of the model at different confidence levels under different distributional assumptions and evaluated the models by loss function test method. The results show that: 1) the VaR under t distribution is greatly overestimated, so t distribution does not apply to Chinese stock market; 2) the generalized error distribution describes the market risk more accurately than the normal distribution and the VaR calculated in PARCH (1,1) model under generalized error distribution is the best; 3)the VaR-PARCH (1,1)-GED model is more suitable to measure the investment risk of Chinese stock market.