首页|基于混频数据抽样的已实现EGARCH模型的波动率预测

基于混频数据抽样的已实现EGARCH模型的波动率预测

扫码查看
该文以沪深 300 股指期货高频数据为样本,在Realized EGARCH模型的基础上引入了混频数据抽样(MIDAS)结构与时变波动,构建了基于偏t分布的SMA-Realized EGARCH MIDAS模型,该模型提高了模型捕捉长记忆性的能力,更好地刻画了模型的时变波动性.通过滚动时间窗的方法对模型进行VaR预测与后验测试,采用MCS检验评估各模型在不同测度下的波动率预测能力.研究结果显示:相比于传统的Realized GARCH模型、Realized EGARCH 模型和 Realized EGARCH MIDAS 模型,本文提出的 SMA-Realized EGARCH MIDAS模型具有更好的样本拟合效果与样本外波动率预测精度.
The Realized EGARCH Model Volatility Prediction Based on Mixed Data Sampling
In this paper,the high-frequency data of CSI 300 stock index futures are taken as samples.Based on the Realized EGARCH model,the structure of mixed-frequency data sampling(MIDAS)and time-varying fluctuation are introduced,and the SMA-Realized EGARCH MIDAS model based on skewed t distribution is constructed,which improves the ability of the model to capture the long-term memory and better describes the time-varying volatility of the model.VaR prediction and posterior test are conducted on the models by rolling time window method,and MCS test is used to evaluate the volatility prediction ability of each model under different measures.The results show that compared with the traditional Realized GARCH model,the Realized EGARCH model and the Realized EGARCH MIDAS model,the SMA-Realized EGARCH MIDAS model proposed in this paper has better sample fitting effect and prediction accuracy of out-of-sample volatility.

mixed-frequency data samplingtime-varying fluctuationSMA-Realized EGARCH MIDAS modelpos-terior testMCS Test

苏小囡、张蕾、邢钰、徐鸣一

展开 >

南京审计大学数学学院,江苏 南京 211815

南京审计大学金融工程重点实验室,江苏 南京 211815

南京审计大学统计与数据科学学院,江苏 南京 211815

南京审计大学金融学院,江苏 南京 211815

展开 >

混频数据抽样 时变波动 SMA-Realized EGARCH MIDAS模型 后验测试 MCS检验

江苏省高等学校哲学社会科学研究一般课题江苏省高等学校哲学社会科学研究一般课题江苏省研究生科研与实践创新计划江苏省金融工程重点实验室开放基金江苏省金融工程重点实验室开放基金

2021SJA03622022SJYB0365KYCX21_1882NSK2021-13NSK2021-15

2024

江西师范大学学报(自然科学版)
江西师范大学

江西师范大学学报(自然科学版)

CSTPCD北大核心
影响因子:0.538
ISSN:1000-5862
年,卷(期):2024.48(1)
  • 24