基于马尔科夫状态转换和跳跃的高频波动率模型预测  

The Forecasting Performance of the High-frequency Volatility Models with the Markov-switching Regime and Jump

作　　者：马锋[1] 魏宇[1] 黄登仕[1] 夏泽安[2] 

机构地区：[1]西南交通大学经济管理学院,四川成都610031 [2]西南财经大学金融智能与金融工程四川省重点实验室,四川成都611130

出　　处：《系统工程》2016年第1期10-16,共7页Systems Engineering

基　　金：国家自然科学基金资助项目(71372109;71371157;71401077);高等学校博士学科点专项科研基金资助课题(20120184110020);四川省科技青年基金资助项目(15QNJJ0032);西南交通大学研究生创新实验实践项目(YC201405118)

摘　　要：以HAR-RV、HAR-RV-J、HAR-RV-CJ和HAR-RV-TCJ模型为基础,结合马尔科夫状态转换机制,构建了四种新的马尔科夫状态转换高频波动率模型。同时,用沪深300股指期货的高频数据,运用滚动时间窗的样本外预测方法和信度设定检验(Model Confidence Set,MCS),分析了四种新模型对未来波动率的预测能力。实证结果表明,总体上,四种新的马尔科夫状态转换模型比原有波动率模型具有更好的预测表现;在众多波动率模型里,MS-HAR-RV-TCJ模型具有更高的预测精度。然而,实务界常用的低频波动率模型(如GARCH等)的预测能力表现得并不突出。Based on the four basic high-frequency models：HAR-RV,HAR-RV-J,HAR-RV-CJ and HAR-RV-TCJ and combined the Markov-switching regime,we firstly propose four new Markov-switching regime high-frequency volatility models.Taking 5-minute high frequency data of the CSI 300 index futures contracts for example,we apply the out-of-sample rolling time window forecasting combined with Model Confidence Set which is proved superior to SPA test,to explore the forecasting performance of the new models.The empirical results show that the Markov-switching regime models have better performance in forecasting in all.Moreover,the MS-HAR-RV-TCJ model is the best model among models we have discussed in this paper.However,the GARCH-types which are popular in financial academe and practice,perform worst for volatility predicting of CSI300 index futures.

关 键 词：马尔科夫状态转换机制 HAR-RV 跳跃 MCS检验 

分 类 号：F830[经济管理—金融学]