Risk Measurement And Volatility Analysis Of Stock Market Based On GARCH Model

We set up a GARCH model, which can be used to quantify market risk. Based on this model, we designed a risk management system in stock investment, which can solve the nonintuitive risk quantify and over relying on investor’s personal experiences problems. The system can also solve the compliance lag problem well in traditional risk management.After analyzed the time series of Shanghai Composite Index, and quantified the volatility with GARCH model, we found heteroskedasticity in it, which makes it possible to quantify the volatility with GARCH model. Then the model ARMA(1,1)-GARCH(1,1)-GED is set up to describe the volatility well. Based on this model, the daily VaR on different confidence levels is estimated which can be used to measure the daily risk. After comparison of the failure rate of VaR based on this model with failure rate of traditional parametric VaR based on Normal Distribution and GED. We found the GED Parametric VaR model will underestimate risk at any confidence level. The Normal Parametric VaR will underestimate risk at high confidence level, and overestimate risk at low confidence level. However, the GARCH VaR works better than them at different confidence levels.After analyzing the volatility of Shanghai Composite Index, Growth Enterprise Index and some typical industry index such as nonferrous metal, Steel and Compute. The risk management system can analyze these the risks in these industries with these models directly.