The Empirical Research Of Financial Volatility Models Based On Fat-Tailed Distribution

Volatility has been playing an important role in portfolio risk management and option hedging strategy. In modern financial theory, volatility is widely used to represent risk and is always measured as the variance of return. Variance is supposed to be constant in the traditional econometric model. But with the development of new research, the hypothesis is proved to be improper. Many empirical researches in financial data showed that variance is changeable with time varying. ARCH model had been applied in the research of volatility of stock market.In this paper, I first give the outline of the significance of the volatility research and the present situation of domestic and foreign research. Then the typical characters of volatility of the financial market price and the common models to describe the volatility are regarded. At the last two parts of the paper we regarded Shenzhen stock composite price index as the main study object, and try to describe the volatility character with the statistic software EViews5. 0. And the main results are the following: excess kurtosis and heteroskedasticity of the series data, asymmetric effect of the series data volatility. 12 models which follow Normal distribution and two fat-tailed distributions are used to carry on our empirical research.Through the comparison of parameter estimating, diagnosing and forecasting precision of each model, the GARCH(1,1)-GED model is proved to be the best one to simulate the volatility character of the return series of Shenzhen stock composite price index.