Please wait a minute...
浙江大学学报(理学版)  2022, Vol. 49 Issue (1): 66-75    DOI: 10.3785/j.issn.1008-9497.2022.01.010
数学与计算机科学     
沪深300指数波动率和VaR预测研究——基于投资者情绪的HAR-RV GAS模型
沈银芳(),严鑫
浙江财经大学 数据科学学院,浙江 杭州 310018
Research on volatility and VaR prediction of Shanghai and Shenzhen 300 index
Yinfang SHEN(),Xin YAN
School of Data Sciences,Zhejiang University of Finance & Economics,Hangzhou 310018,China
 全文: PDF(1004 KB)   HTML( 7 )
摘要:

基于广义自回归得分 (generalized autoregressive score,GAS)和已实现波动率异质自回归 (heterogeneous autoregressive of realized volatility,HAR-RV)模型,引入投资者情绪因素,构建了HAR-RV GAS和HAR -RV-SENT GAS波动率模型,旨在预测沪深300指数波动率和风险价值(value at risk,VaR)度量。用自相关函数曲线和高级预测能力(superior predictive ability,SPA)检验方法,分别解析了模型样本内拟合和样本外波动率的预测能力。运用VaR预测序列图和平均分位数损失函数,实证分析了VaR预测效果。结果表明,HAR-RV-SENT GAS模型在波动率和VaR预测能力上效果较佳,HAR-RV GAS模型次之。研究可为金融投资者和风险管理者提供理论参考。

关键词: 波动率VaRHAR-RV GAS模型投资者情绪已实现波动率    
Abstract:

Based on the generalized autoregressive score (GAS) and heterogeneous autoregressive of realized volatility (HAR-RV) models, introduce investor sentiment factor, this paper constructs HAR-RV GAS and HAR -RV SENT GAS volatility models, aim to predict the superior predictive ability (SPA) test, analyze the ability of in-sample data fitting and out of sample volatility prediction respectively.Through VAR prediction sequence diagram and average quantile loss function, the effect of VaR prediction is analyzed empirically. The results show that HAR-RV-SENT GAS model is the best in volatility and VaR prediction ability, followed by HAR -RV GAS model. This study can provide theoretical reference for financial investors and risk managers.

Key words: volatility    VaR    HAR-RV GAS model    investor sentiment    realized volatility
收稿日期: 2020-12-03 出版日期: 2022-01-18
CLC:  F 830.91  
基金资助: 浙江省哲学社会科学规划课题(17NDJC171);浙江省统计科学研究基地项目(21TJJD07);浙江省一流学科A类(浙江财经大学统计学)规划项目(Z0111116008/008)
作者简介: 沈银芳(1978—),ORCID:https://orcid.org/0000-0003-2080-6530,女,博士,副教授,主要从事金融时间序列分析研究,E-mail:fsilver@163.com.
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
沈银芳
严鑫

引用本文:

沈银芳,严鑫. 沪深300指数波动率和VaR预测研究——基于投资者情绪的HAR-RV GAS模型[J]. 浙江大学学报(理学版), 2022, 49(1): 66-75.

Yinfang SHEN,Xin YAN. Research on volatility and VaR prediction of Shanghai and Shenzhen 300 index. Journal of Zhejiang University (Science Edition), 2022, 49(1): 66-75.

链接本文:

https://www.zjujournals.com/sci/CN/10.3785/j.issn.1008-9497.2022.01.010        https://www.zjujournals.com/sci/CN/Y2022/V49/I1/66

变量TurntPERtUpdowntCCItVolumetTurnt-1PERt-1Updownt-1CCIt-1Volumet-1
SENT10.2370.020-0.058 60.9890.2480.2480.029 40.033 40.9900.259
表1  初始情绪指数 SENT1与代理变量之间的相关性
图1  收益率、已实现波动率与投资者情绪指数的时序图
序列均值标准差偏度峰度J-B统计量
收益率0.000 2250.011 4-0.2056.772394.831***
已实现波动率0.000 1270.000 1895.82349.83873 988.955***
投资者情绪指数0.000 0116.2430.3435.13221.513***
表2  收益率、已实现波动率和投资者情绪指数的描述性统计
指标GPH检验Q(5)Q(10)Q(20)
日收益率0.202 2143.1***146.3***153.6***
已实现波动率0.396 9269.7***354.3***436.8***
投资者情绪指数0.498 42496.9***4 292.6***6 329.4***
表3  日收益率、已实现波动率和投资者情绪指数序列的长记忆性检验
图2  不同模型下自相关函数理论值与实际值
图3  模型的波动率拟合结果
模型ωA1B1C1似然函数值
GAS0.866 20.059 90.993 2-658.764
HAR-RV GAS0.510 10.041 80.836 20.163 6-655.740
HAR-RV-SENT GAS0.533 00.049 50.837 60.162 8-651.437
表4  GAS、HAR-RV GAS和HAR-RV-SENT GAS模型的参数估计结果
损失函数GAS模型HAR-RV GAS模型HAR-RV-SENT GAS模型
SPAlSPAcSPAuSPAlSPAcSPAuSPAlSPAcSPAu
L1(MSE)0.003 50.014 50.794 50.051 00.056 00.063 50.152 00.357 50.864 0
L2(MAE)0.004 00.005 00.807 00.051 00.056 00.063 50.152 00.357 50.864 0
L3(HMSE)0.014 50.019 50.765 00.160 00.161 50.167 50.174 00.175 00.177 0
L4(HMAE)0.070 50.077 00.087 00.000 50.060 50.793 00.089 50.199 50.211 5
L5(QLIKE)000.785 00.006 50.017 00.019 00.053 50.155 50.157 0
表5  GAS、HAR-RV GAS和HAR-RV-SENT GAS模型的SPA检验结果
图4  模型的波动率预测值
图5  置信水平为95%时沪深300指数的VaR预测结果
图6  置信水平为99%时沪深300指数的VaR预测结果

置信

水平/%

QL(HAR-RV GAS模型)/QL(GAS模型)QL(HAR-RV-SENT GAS模型)/QL(GAS模型)
90.00.870.86
95.00.860.85
97.50.860.84
99.00.850.84
99.50.850.84
99.750.850.84
表6  在不同置信水平下HAR-RV GAS和HAR-RV-SENT GAS模型与GAS模型的平均分位数损失值
1 罗樱,于欣,刘康生.安全第一标准下基于VaR度量的投资组合优化问题[J].浙江大学学报(理学版),2006,33(5):503-506. DOI:10.3321/j.issn:1008-9497.2006.05.006
LUO Y,YU X,LIU K S.The problem of portfolio optimization based on safety-first criteria and the measure of VaR[J].Journal of Zhejiang University(Science Edition),2006,33(5):503-506. DOI:10. 3321/j.issn:1008-9497.2006.05.006
doi: 10. 3321/j.issn:1008-9497.2006.05.006
2 CREAL D,KOOPMAN S J,LUCAS A.Generalized autoregressive score models with applications[J].Journal of Applied Econometrics,2013,28(5):777-795. DOI:10.1002/jae.1279
doi: 10.1002/jae.1279
3 王天一,黄卓.Realized GAS-GARCH及其在VaR预测中的应用[J].管理科学学报,2015,18(5):79-86. DOI:10.3969/j.issn.1007-9807.2015.05.008
WANG T Y,HUANG Z.Realized GAS-GARCH model and its application in value-at-risk forecast[J].Journal of Management Sciences in China,2015,18(5):79-86. DOI:10.3969/j.issn.1007-9807.2015. 05.008
doi: 10.3969/j.issn.1007-9807.2015. 05.008
4 沈根祥,邹欣悦.已实现波动 GAS-HEAVY 模型及其实证研究[J].中国管理科学,2019,27(1):1-10. DOI:10.16381/j.cnki.issn1003-207x.2019.01.001
SHEN G X,ZOU X Y.GAS-HEAVY model for realized measures of volatility and returns[J]. Chinese Journal of Management Science,2019,27(1):1-10. DOI:10.16381/j.cnki.issn1003-207x.2019.01.001
doi: 10.16381/j.cnki.issn1003-207x.2019.01.001
5 NANI A,GAMOUDI I,GHOURABI M E.Value-at-risk estimation by LS-SVR and FS-LS-SVR based on GAS model[J].Journal of Applied Statistics,2019,46(12):2237-2253. DOI:10.1080/02664763.2019.1584161
doi: 10.1080/02664763.2019.1584161
6 PATTON A J,ZIEGEL J F,CHEN R.Dynamic semiparametric models for expected shortfall (and value-at-risk)[J].Journal of Econometrics,2019,211(2):388-413. DOI:10.1016/j.jeconom.2018. 10.008
doi: 10.1016/j.jeconom.2018. 10.008
7 宋加山,蒋坤良,周学伟.基于GAS-混合Copula模型的不同行业系统性风险研究[J].统计与信息论坛,2020,35(5):52-60. DOI:10.3969/j.issn.1007-3116. 2020.05.007
SONG J S,JIANG K L,ZHOU X W.Research on the impact of government intervention on China's high-quality economic development:Spatial econometric model based on panel data[J].Statistical & Information Forum,2020,35(5):52-60. DOI:10. 3969/j.issn.1007-3116.2020.05.007
doi: 10. 3969/j.issn.1007-3116.2020.05.007
8 ANDERSEN T G,BOLLERSLEV T,DIEBOLD F X,et al.The distribution of realized stock return volatility[J].Journal of Financial Economics,2001,61(1):43-76. DOI:10.1016/S0304-405X(01)00055-1
doi: 10.1016/S0304-405X(01)00055-1
9 邵锡栋,殷炼乾.基于实现极差和实现波动率的中国金融市场风险测度研究[J].金融研究,2008(6):109-121.
SHAO X D,YIN L Q.The study on financial market risk measures in China based on realized range and realized volatility[J].Journal of Financial Research,2008(6):109-121.
10 MAHEU J M,MCCURDY T H.Do high-frequency measures of volatility improve forecasts of return distributions[J].Journal of Econometrics,2011,160(1):69-76. DOI:10.1016/j.jeconom.2010.03.016
doi: 10.1016/j.jeconom.2010.03.016
11 CORSI F,FUSARI N,VECCHIA D L.Realizing smiles:Options pricing with realized volatility[J].Journal of Financial Economics,2013,107(2):284-304. DOI:10.1016/j.jfineco.2012.08.015
doi: 10.1016/j.jfineco.2012.08.015
12 王春峰,郭华,房振明,等.高频视角下考虑收益非对称性结构的VaR和ES风险测度[J].系统工程,2013,31(2):23-29.
WANG C F,GUO H,FANG Z M,et al.Risk management of Chinese stock market with asymmetric structure of return in the perspective of high-frequency data[J]. Systems Engineering,2013,31(2):23-29.
13 刘晓倩,王健,吴广.基于高频数据HAR-CVX模型的沪深300指数的预测研究[J].中国管理科学,2017,25(6):1-10. DOI:10.16381/j.cnki.issn1003-207x.2017.06.001
LIU X Q,WANG J,WU G.The forecasting analysis of HS300 index based on HAR-CVX model of high frequency data[J].Chinese Journal of Management Science,2017,25(6):1-10. DOI:10.16381/j.cnki.issn1003-207x.2017.06.001
doi: 10.16381/j.cnki.issn1003-207x.2017.06.001
14 于孝建,王秀花.基于混频已实现GARCH模型的波动预测与VaR度量[J].统计研究,2018,35(1):104-116. DOI:10.19343/j.cnki.11-1302/c.2018.01.011
YU X J,WANG X H.Mix frequency realized GARCH models:The forecast of volatility and measure of VaR[J].Statistical Research,2018,35(1):104-116. DOI:10.19343/j.cnki.11-1302/c.2018.01.011
doi: 10.19343/j.cnki.11-1302/c.2018.01.011
15 陈声利,李一军,关涛.波动预测建模与尾部风险测量方法[J].管理科学,2018,31(6):17-32. DOI:10. 3969/j.issn.1672-0334.2018.06.002
CHEN S L,LI Y J,GUAN T.Methodology of volatility forecasting modelling and tail risk measurement[J].Journal of Management Science,2018,31(6):17-32. DOI:10.3969/j.issn.1672-0334.2018.06.002
doi: 10.3969/j.issn.1672-0334.2018.06.002
16 BAKER M,WURGLER J.Investor sentiment and the cross-section of stock returns[J].The Journal of Finance,2006,61(4):1645-1680. DOI:10.1111/j.1540-6261.2006.00885.x
doi: 10.1111/j.1540-6261.2006.00885.x
17 SEO S W,KIM J S.The information content of option-implied information for volatility forecasting with investor sentiment[J].Journal of Banking & Finance,2015,50:106-120. DOI:10.1016/j.jbankfin. 2014.09.010
doi: 10.1016/j.jbankfin. 2014.09.010
18 DEBASISH M,SAUMYA R D.Sentiment and stock market volatility revisited:A time-frequency domain approach[J].Journal of Behavioral and Experimental Finance,2017,15:74-91. DOI:10. 1016/j.jbef.2017.07.009
doi: 10. 1016/j.jbef.2017.07.009
19 SCHNELLER D,HEIDEN S,HEIDEN M,et al.Home is where you know your volatility-local investor sentiment and stock market volatility[J].German Economic Review,2018,19(2):209-236. DOI:10. 1111/geer.12125
doi: 10. 1111/geer.12125
20 佟孟华,于建玲,朱芳燕.投资者情绪对中国上市金融机构系统性风险的影响[J].投资研究,2018,37(11):4-15.
TONG M H,YU J L,ZHU F Y.The influence of investor sentiment on the systematic risk of China's listed financial institutions[J].Review of Investment Research,2018,37(11):4-15.
21 CORSI F.A simple approximate long-memory model of realized volatility[J].Journal of Financial Econometrics,2009,7(2):174-196. DOI:10.1093/jjfinec/nbp001
doi: 10.1093/jjfinec/nbp001
22 HUANG Z,LIU H,WANG T Y.Modeling long memory volatility using realized measures of volatility:A realized HAR GARCH model[J].Economic Modelling,2016,52(B):812-821. DOI:10.1016/j.econmod.2015.10.018
doi: 10.1016/j.econmod.2015.10.018
23 YANG C,GONG X,ZHANG H W.Volatility forecasting of crude oil futures:The role of investor sentiment and leverage effect[J].Resources Policy,2019,61:548-563. DOI:10.1016/j.resourpol.2018. 05.012
doi: 10.1016/j.resourpol.2018. 05.012
24 HANSEN P R.A test for superior predictive ability[J].Journal of Business & Economic Statistics,2005,23(4):365-380. DOI:10.1198/0735001050000 00063
doi: 10.1198/0735001050000 00063
25 魏宇.中国股市波动的异方差模型及其SPA检验[J].系统工程理论与实践,2007,27(6):27-35. DOI:10.3321/j.issn:1000-6788.2007.06.005
WEI Y.Heteroskedastic volatility models in Chinese stock market and their SPA test[J].System Engineering-Theory & Practice,2007,27(6):27-35. DOI:10.3321/j.issn:1000-6788.2007.06.005
doi: 10.3321/j.issn:1000-6788.2007.06.005
[1] 耿莉,王长鹏. 基于稀疏一致图分解的鲁棒多视图聚类算法[J]. 浙江大学学报(理学版), 2023, 50(5): 569-579.
[2] 高振斌,梁兴碧. 基于VAR和EGARCH的投资者情绪对股市收益率的影响研究[J]. 浙江大学学报(理学版), 2023, 50(4): 434-441.
[3] 何兴玥,丁欢欢. 一类k-Hessian方程k-允许解的唯一性和收敛性[J]. 浙江大学学报(理学版), 2023, 50(4): 424-428.
[4] 赵小鹏,戴磊,曹小红. 有界线性算子及其函数的(R)性质[J]. 浙江大学学报(理学版), 2022, 49(3): 294-299.
[5] 任欢, 刘婷, 康俊锋, 潘宁, 李敏靓, 艾顺毅. 一种基于百度指数的城市日游客规模预测方法[J]. 浙江大学学报(理学版), 2020, 47(6): 753-761.
[6] 柳顺义, 覃忠美. 图的Aα特征多项式系数的一个注记[J]. 浙江大学学报(理学版), 2019, 46(4): 399-404.