Abstract: ARCH model, GARCH model and GARCH-M model were introduced, and the character of ARCH-type model was analyzed. Shanghai stock market composite price index was regarded as the main study object, and the time vary of return rate on Shanghai stock market composite price index was analyzed by using the ARCH model and the statistic software Eviews. The empirical conclusion indicates that GARCH(1,1) model can fit the volatility of return rate in Shanghai stock market, such as volatility clustering, long-memory character and so on; GARCH(1,1)－M can also describe the relation between risks and return in stock markets.

Key words: ARCH model ,  Volatility clustering , Long-memory character , Return rate

摘要： 介绍ARCH 模型、GARCH 模型和GARCH—M 模型,分析ARCH 类模型的特点。然后以上证综指
日收益率作为研究对象,采用ARCH 类模型结合Eviews统计软件对上证综指日收益率的时变性进行实证分析。实
证结果表明,GARCH(1,1)模型能较好地拟合上海股市收益率的波动特征,如波动聚集性、长记忆性等;GARCH(1,
1)-M 模型也能很好地拟合股市中风险与收益率之间的关系。

关键词: ARCH 类模型 , 波动聚集性 , 长记忆性 , 收益率