针对金融市场条件收益存在的“有偏胖尾”分布与非对称波动性特征以及长记忆特征等典型事实特征,运用ARFIMA-FIGARCH-SKST模型等来测度股市动态风险,并通过规范的返回测试检验中的LRT和DQR方法实证考察了测度模型的可靠性。得到了一些非常有价值的实证结果:有无长记忆约束的非对称结构风险模型在中国大陆沪深股市动态风险测度能力上并无实质性差异;ARFIMA-FIAPARCH-SKST模型能够准确测度股市的动态风险;股票市场极端风险的测度尤其不能放弃非对称结构的这一约束条件。 Aiming at the typical fact features such as the distribution and asymmetric volatility of “fat tail” and the characteristics of long memory in financial market conditions, ARFIMA-FIGARCH-SKST model is used to measure the dynamic risk in the stock market. Through the standard return The LRT and DQR methods in the test test empirically examine the reliability of the measurement model. Some very valuable empirical results are obtained: the asymmetric structural risk model with or without long memory constraints has no substantial difference in the dynamic risk measurement ability of the Shanghai and Shenzhen stock markets in China; the ARFIMA-FIAPARCH-SKST model can accurately measure the stock market dynamics Risk; the measurement of extreme risk in the stock market can not, in particular, give up this constraint on asymmetric structures.