中文摘要：

就成交量信息是否有助于预测股票市场的波动率这一问题,目前学术界有两种截然相反的观点存在。本文以中国股票市场代表性指数的代表性波动周期为例,对上述问题进行了实证研究。通过采用较以往研究更为严谨和稳健的样本外滚动时间窗预测法和高级预测能力检验法（Superiorpredictive ability,SPA）,本文得到的分析结论包括：（1）成交量信息对中国股票市场的波动过程有显著影响;（2）将成交量纳入GARCH族模型会导致条件方差方程中的波动持续性出现明显下降;（3）引入成交量作为附加解释变量的GARCH族模型并未表现出比一般GARCH族模型更优的波动率预测能力。最后对实证结果给出了理论解释。

英文摘要：

Take the most important index in Chinese stock market as sample, various GARCH models augmented by the addition of measures of volume are constructed and their forecasting performance are compared with the initial corresponding models using SPA test. The empirical results show that, GARCH models augmented by volume are inferior to initial ones in volatility forecasting performance. In company with conclusions on mature capital markets, this conclusion support the viewpoint that volatility models augmented by volume reduce persistence of volatility and forecasting ability of initial models accordingly.