高频数据波动率建模——基于厚尾分布的Realized GARCH模型

"厚尾现象"是金融时间序列分布的一个普遍特征,本文将RealizedGARCH模型推广到容纳厚尾分布的情形,并将杠杆函数的幂次放松为待估参数。结果显示,使用Skewed-t分布的模型能够较好地反映收益率序列的厚尾和偏峰性质,放松的幂次参数可以给出更贴合数据的"信息冲击曲线"。引入厚尾分布亦可用改进Realized GARCH模型对实现测度的预测,其中使用标准t分布的模型给出的预测精度最高。

High Frequency Volatility Modeling

Fat-tail is a stylized feature of financial time series.This paper extents Realized GARCH model proposed to allow fat-tail distribution and data determined power parameter in leverage function.Results show that skewed-t distribution based model can fit the fat-tail and skewness features of data.Also,data determined leverage function will deliver more flexible news impact curve.Fat-tail based model especially student’s-t distribution outperformed standard Realized GARCH model in terms of forecasting realized measure.