Comparison For Volatility Models Of Yield On Different Distributing

The risk of finance has been paid attention by more and more people.The risk can be measured by variance in statistics.The volatility of yield has been predicted by many models such as ARCH models and stochastic volatility models.In this paper,I compare GARCH model with stochastic volatility model in different distributions,and get some good results which are useful in our practice.Through the software such as matlab and winbugs,I estimate the parameters of SV model using the MCMC method and the parameters of GARCH model.I get the conclusions that the SV model with the residuals submitted to mixed normal distribution is a model which is not good as other models, while the SV model with the residuals submitted to normal distribution is better than GARCH models.The distribution of yield submit to stable distribution other than normal distribution because its kurtosis is not three.