基于GJR—ALaplace方法的开放式基金风险研究

根据我国开放式基金收益率序列的尖峰、厚尾、有偏和波动时变的特征，引入非对gg,Laplace分布对收益率序列进行刻画和拟合。构建度量基金风险的动态GJR—Asymmetric～Laplace模型，在非对称Laplace分布、Laplace分布和正态分布三种分布假设下测算VaR，并做返回检验。选取12只开放式基金在2007．01．04～2009．12．31期间的日累计净值数据做实证研究。实证表明：除了基金大成债券外，其余11只基金显著通过假设，符合非对称Laplace分布，相rELaplace分布和正态分布来说，非对称Laplace分布能更好地拟合基金收益率序列。正态分布假设下风险度量值通过检验的基金数显著少于Laplace分布假设，而Laplace分布下通过检验的基金数亦少于非对称Laplace分布，可知非对称hplace分布假设下得出的基金动态风险值更为有效。

Study on VaR of Open-ended Fund based on GJR-ALaplace Model

The paper introduces the Asymmetric Laplace distribution to discrible and fit the return series accaccording to the characteristic of sharp kurtosis, heavy tails, skew-hess and time-varying volatility of the return series of China＇s Open-ended. And sets up the GJR-Asymmetrie-Laplace model to estimates the dynamic VaR of the funds under the assume of Asymmetric Laplace distribution, Laplace distribution and Normal distribution, and makes the back-testing. The empirical reseach selected 12 Open-ended funds＂ accumulative net value in the period 2007.01.04-2009.12.31. The result shows： The remaining 11 funds pass the assuming significantly except for the DaCheng Bonds, accord to the Asymmetric Laplace distribution. It means the Asymmetric Laplace distribution can fit the characteristic of the return series better compare to the Laplace distribution and the Normal distribution. And the quantity of the funds which the VaR pass the back-testing based on the Normal distribution less then the Laplace distribution, also the quantity of the funds which the VaR pass the back-testing based on the Asymmetric Laplace distribution. It shows that measure of VaR using the model under the assume of Asymmetric Laplace distribution are more valid.