基于GARCH模型的期望损失ES非参数核估计

Nonparametric Kernel Estimation of Expected Shortfall Based on The GARCH Model

风险度量ES最新的非参数估计方法,不依赖于分布假设,但不能动态反应金融时间序列的风险.针对金融时间序列的波动,结合GARCH模型进行期望损失ES的非参数核估计,得到随市场波动而动态变化的ES预测.通过数值模拟和对近两年的上证指数实证分析验证了该方法能准确而有效的反映市场风险.

The latest nonparametric estimation method of risk measure ES（Expected Short- fall） does not depen on distributional assumptions, but it can not reflect dynamically the risk of financial time series . In view of fluctuations of financial time series, combining ES non- parametric kernel estimation of GARCH model, get the ES 15rediction that dynamic changes along with the market fluctuation. Empirical analysis of the Shanghai Composite Index for nearly two years verifies the method can reflect the market risk accurately and effectively. By numerical simulation and empirical analysis of the Shanghai Composite Index in nearly two years, it is showed that the method can accurately and effectively reflect the market risk.