波动率度量模型的评价方法:拟合优度和平滑性

Evaluating method of volatility models:Goodness-of-fit and smoothness

提出了一种评价波动率模型的方法,与文献中已有准则有较大不同.该方法兼顾了模型拟合新息平方序列的能力和模型的平滑性,而避免了经典的评价方法只考虑模型拟合能力的缺陷.该方法有利于投资者挑选出交易成本相对较低和风险对冲能力相对较强的波动率模型.实证例子是估计中国股票市场的行业时变风险:对三大类波动率模型进行了评价,并指出评价方法或标准的选择直接影响金融风险估计或预测的评价结果.

This paper proposes one method to evaluate the volatility models, which differs a lot from the meth- ods mentioned in the literatures. This method combines the goodness-of-fit and the smoothness of models, and avoids the defect of the traditional assessing methods which only take the goodness-of-fit into consideration. This method may help investors to select the volatility models which reduce the transaction costs and increase the risk hedging powers to the highest degree. The empirical example discussed in this paper is to estimate the industry time-varying risk of Chinese stock markets. Three classes of volatility models were evaluated, which indicates that the choice of evaluation methods or criteria directly affects the evaluation results of the financial risk estimation and prediction.