Comparative Analysis Based On The Poor And The Yield Of The Arch Models

This paper briefly recalled the development experience of, the GARCH model and range. Engle originally proposed ARCH models which can successfully simulate the time-varying variance model, variance and conditional variance to separate and let the conditional variance as a function of past errors and changes, so as to solve the heteroscedasticity problem provides a new way. Engle's students Bollerslev (1986) developed the theory of Engle's, metioned the generalized autoregressive conditional heteroskedasticity (GARCH) model based on Autoregressive Conditional Heteroskedasticity model, and overcame some shortages of ARCH models, such as needing a lot of parameters and so on. Later, taking into account the yield on the securities will be dependent on its volatility, there were scholars who applied GARCH model to the mean equation proposed GARCH-M model, GARCH-M model implies a rate of yield before and after the sequence of correlation. Nelson (1991) proposed the exponential GARCH model (EGARCH model). The model embodies the principles of positive and negative return on assets of the non-symmetrical effect. Since then, ARCH model formed the systematic ethnic formation, and became of the important role for finance research. Parkinson (1980) study pointed out that it with the traditional measurement of volatility of financial assets compared with the use of range will have better results, range contains more information, so certainly than the rate of return more effectively characterize volatility. Later, Taiwan, China scholar Dr. Zhou Yutian Chou (2005) combines the GARCH model ideas and range presented dynamically grasp of financial asset price volatility in Range Condition Autoregressive model (CARR), Chinese mainland scholars in recent years also actively studies price volatility issues using range, and predictively analysis The Shanghai Composite Index gained good results. By their influence, this article will adopt range sequence and return sequences to use ARCH model and do a systematic comparative analysis.In this paper, after sampling, using Eviews software strictly verify non-stationary nature of the range and stock prices, and thus made a first-order differential smoothing operation obtained return sequences and range sequence. Then use ARCH models for return and range to conduct a comprehensive analysis and forecast the volatility, and find that it is not better to use range to forecast volatility than using return. On this basis, further found that range is not Granger cause volatility in the reasons.