沪深股市风险度量中半参数ES模型的实证检验

Semi-parametric ES Models in Measuring the Risk of Shanghai and Shenzhen Stock Markets

目前度量预期不足(Expected Shortfall, ES)的风险技术大多基于参数模型,其建模过程避免不了对收益的分布类型做出假定,但这些分布往往与现实相悖。为此,介绍两种重要半参数模型,即CARE模型和CARES模型,并应用我国2007-2016年上证综合指数与深证成分指数的相关数据评估模型优劣。结果表明:CARES模型与CARE模型在度量我国股市风险中都具有较好的效果,但两者比较,CARES模型明显优于CARE模型。因此,CARES模型能作为我国股市风险度量工具中的一个重要补充。

Most risk measurement techniques for expected shortfall are based on parametric models, which doesn’t avoid distribution hypothesizes for asset returns. But these assumptions usually contradict with the reality. In view of this, two main semi-parametric models for measuring ES are proposed, they are CARE and CARES models. By applying these models to Shanghai Composite Index and Shenzhen Component Index sample data from 2007 to 2016, the results show that both the two kinds of models perform well, but the CARES model does better than the CARE model. Therefore, the CARES model can be regarded as an important supplement for the risk measurement yield in China’s stock markets.