Analysis of the Volatility of SCI 300 Based on the Conditional Auto-Regressive Range Model

In 2005, Chou suggested an approach to conduct the volatility analysis by use of the Auto-regressive Range Model, yet the model was blame for its deficiencies found in the distribution assumptions of logarithmic mean value. By selection of different distributions, improvements were made to the model. By contrast, modeling was made of the volatility rate of the logarithmic return rate by selection of the GARCH Model and by use of Bayesian the parameters of the two models were subjected to a parameter analysis and an empirical analysis was done by taking advantage of the SCI300 index data; and a parameter simulation was also done to the two models by WinBugs. By means of the analysis of the parameter estimates it was concluded that the Auto-Regressive Model is of a better performance in capturing the short-term effects while the GARCH Model is better at capturing the long-term effects.