股市涨跌预测与量化投资策略:基于时变矩成分分析

doi:10.16381/j.cnki.issn1003-207x.2020.02.001

中国管理科学 ›› 2020, Vol. 28 ›› Issue (2): 1-12.doi: 10.16381/j.cnki.issn1003-207x.2020.02.001

• 论文 • 下一篇

股市涨跌预测与量化投资策略:基于时变矩成分分析

鲁万波¹, 黄光麟¹, Kris Boudt²

1. 西南财经大学统计学院, 四川成都 611130;
2. 比利时布鲁塞尔自由大学, 布鲁塞尔 1050

收稿日期:2018-11-21 修回日期:2019-03-20 出版日期:2020-02-20 发布日期:2020-03-03
通讯作者: 鲁万波(1977-),男(汉族),贵州贵阳人,西南财经大学统计学院,教授,博士生导师,研究方向:金融计量、风险管理,E-mail:luwb@swufe.edu.cn. E-mail:luwb@swufe.edu.cn
基金资助:
国家自然科学基金资助面上项目（71771187）；国家自然科学基金国际合作交流项目（7191101155）；教育部"新世纪优秀人才支持计划"项目（NCET-13-0961）；西南财经大学"光华英才"百人计划；中央高校基本科研业务费专项资金项目（JBK190602，JBK1907201850）

Stock Market Rise-Fall Forecast and Quantitative Investment Strategy: Based on Time Varying MCA

LU Wan-bo¹, HUANG Guang-lin¹, Kris Boudt²

1. School of Statistics, Southwestern University of Finance and Economics, Chengdu 611130, China;
2. Solvay Business School, Vrije Universiteit Brussel, Brussel 1050, Belgium

Received:2018-11-21 Revised:2019-03-20 Online:2020-02-20 Published:2020-03-03

摘要/Abstract

摘要： 利用时变矩成分分析提取高阶矩吸收率，改进基于阈值的高阶矩因子个数选择方法，提出基于元素值的联合矩成分分析权重设定，构造了一种基于股票市场高阶矩相关结构的量化投资策略。研究表明，基于矩成分吸收率的投资策略能够对股市涨跌做出有效预测，对于股市的重大系统风险尤为敏感，在熊市中也有良好表现；基于单因子吸收率、累积三因子吸收率和赫芬达尔吸收率的高阶矩投资策略优于二阶矩吸收率投资策略，而三阶矩单因子吸收率投资策略最优，基于元素值权重的投资策略优于基于元素个数权重的投资策略；量化投资策略具有参数稳健性，且可通过优化高阶矩的时变结构对投资效果进行优化。

关键词: 时变矩成分分析, 吸收率, 高阶矩, 因子结构, 量化投资策略

Abstract: The research on the correlation of financial assets is of great significance to capital risk management, portfolio selection and financial supervision. Due to the return of financial assets are fat-tail and skewed, using higher order moment correlation structure is more reasonable for financial market modeling. Moment Component Analysis (MCA), which is an expansion of PCA, is a new way to study the higher order correlation structure of financial market. The Time-Varying MCA is proposed to extract the high order co-moment absorption ratios in order to predict the systematic rise-fall of the stock market. Meanwhile, a method of selecting the number of high-order moment factors based on the threshold is improved in this paper, this method can fit the marginal distribution of different assets. On the other hand, the weights of the Joint-MCA based on the element values is proposed to improve the performance of the Joint-MCA. Finally, a quantitative investment strategy based on the MCA absorption ratios is constructed in stock market. This new method is illustrated with 52 component stocks of the CSI300 Index. The empirical study shows that the investment strategies based on the MCA absorption ratios can make effective predictions for the rise-fall of the stock market, particularly sensitive to systematic risks of the stock market, while also have good performance in the bear market. The investment strategy of the single factor absorption ratios, the Herfindahl absorption ratios and the accumulated three factor absorption ratios all have better performance than the investment strategy of the PCA absorption ratios, and the single factor absorption ratio of third order co-moment has the best performance. In addition, the weight based on element values outperformsthan the weight based on the number of elements. The robust analysis shows that the investment strategies have parameter-robustness and the investment performance can be improved by optimizing the time-varying structure of higher-order co-moments. To a large extent, the empirical results of this paper demonstrate the advantages of the Time-Varying MCA methods, which can predict the rise-fall of the stock market more efficiently rather than PCA, and give investors a better reference for asset allocation.

Key words: moment component analysis, absorption ratio, higher-order moments, factor structure, quantitative investment strategy

中图分类号:

F830.91

鲁万波, 黄光麟, Kris Boudt. 股市涨跌预测与量化投资策略:基于时变矩成分分析[J]. 中国管理科学, 2020, 28(2): 1-12.

LU Wan-bo, HUANG Guang-lin, Kris Boudt. Stock Market Rise-Fall Forecast and Quantitative Investment Strategy: Based on Time Varying MCA[J]. Chinese Journal of Management Science, 2020, 28(2): 1-12.

参考文献

[1] Cont R. Empirical properties of asset returns:Stylized facts and satistical issues[J]. Quantitative Finance, 2001, 1(2):223-236.
[2] Dittmar R. Nonlinear pricing kernels, kurtosis preference, and evidence from the cross section of equity returns[J]. Journal of Finance, 2002, 57(1):369-403.
[3] Mitton T, Vorkink K. Equilibrium underdiversifi-cation and the preference for skewness[J]. Review ofFinancial Studies, 2007, 20(4):1255-1288.
[4] Jondeau E, Rockinger M. Optimal portfolio allocation under higher moments[J]. European Financial Management, 2006, 12(1):29-55.
[5] Martellini L, Ziemann V. Improved estimates of higher-order comoments and implications for portfolio selection[J]. Review of Financial Studies, 2010, 23(4):1467-1502.
[6] Boudt K, Lu Wanbo, Peeters B. Higher order comoments of multifactor models and asset allocation[J]. Finance Research Letters, 2015, 13:225-233.
[7] 蒋翠侠,许启发,张世英.金融市场条件高阶矩风险与动态组合投资[J].中国管理科学,2007(1):27-33.
[8] 王鹏,王建琼,魏宇.自回归条件方差-偏度-峰度:一个新的模型[J].管理科学学报,2009,12(5):121-129.
[9] 黄卓,李超.动态金融高阶矩建模:基于Generalized-t分布和Gram-Charlier展开分布的比较研究[J].中国管理科学,2015,23(10):11-18.
[10] 方立兵,曾勇.股市收益率高阶矩风险的产生机制检验[J].中国管理科学,2016,24(4):27-36.
[11] Ang A, Chen J. Asymmetric correlations of equity portfolios[J]. Journal of Financial Economics, 2002, 63(3):443-494.
[12] Jondeau E, Jurczenko E, Rockinger M. Moment component analysis:An illustration with inter-national stock markets[J].Journal of Business and Economic Statistics,2018, 36(4):576-598.
[13] Zheng Yu. Trajectory data mining:An overview[J]. ACM Transactions on Intelligent Systems and Techno-logy (TIST), 2015, 6(3):29.
[14] Billio M, Getmansky M, Lo A W, et al. Econometric measures of connectedness and systemic risk in the finance and insurance sectors[J]. Journal of Financial Economics, 2012, 104(3):535-559.
[15] Kritzman M, Li Y, Page S, et al. Principal components as a measure of systemic risk[J]. Journal of Portfolio Management, 2011, 37(4):112-126.
[16] De Lathauwer L, De Moor B, Vandewalle J. On the best rank-1 and rank-(R1,R2,…,RN) approxim-ation of higher-order tensors[J]. SIAM Journal on Matrix Analysis and Applications, 2000, 21(4):1324-1342.
[17] Tucker L R. Some mathematical notes on three-mode factor analysis[J]. Psychometrika, 1966, 31(3):279-311.
[18] Stekhoven D J, Bühlmann P. Miss forest-non-parametric missing value imputation for mixed-type data[J]. Bioinformatics, 2012, 28(1):112-118.
[19] Man G. Competition and the growth of nations:International evidence from Bayesian model average-ng[J]. Economic Modelling, 2015, 51:491-501.
[20] Johnstone I M. On the distribution of the largest eigenvalue in principal components analysis[J]. Annals of Statistics, 2001, 29(2):295-327.
[21] Bai Jushan, Ng S. Determining the number of factors in approximate factor models[J]. Econome-trica, 2010, 70(1):191-221.
[22] Onatski A. Determining the number of factors from empirical distribution of eigenvalues[J]. Review of Economics and Statistics, 2010, 92(4):1004-1016.
[23] Ahn S C, Horenstein A R. Eigenvalue ratio test for the number of factors[J]. Econometrica, 2013, 81(3):1203-1227.
[24] Theodossiou P. Financial data and the skewed generalized T distribution[J]. Management Science, 1998, 44(12):1650-1661.

股市涨跌预测与量化投资策略:基于时变矩成分分析

Stock Market Rise-Fall Forecast and Quantitative Investment Strategy: Based on Time Varying MCA

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 6

Metrics

本文评价

推荐阅读 10

[1]	朱鹏飞, 唐勇, 钟莉. 基于小波-高阶矩模型的投资组合策略——以国际原油市场为例[J]. 中国管理科学, 2020, 28(10): 24-35.
[2]	郭范勇, 潘和平. 基于β系数优化的动态投资组合策略研究[J]. 中国管理科学, 2019, 27(7): 1-10.
[3]	方立兵, 曾勇. 股市收益率高阶矩风险的产生机制检验[J]. 中国管理科学, 2016, 24(4): 27-36.
[4]	黄卓, 李超. 动态金融高阶矩建模:基于Generalized-t分布和Gram-Charlier展开分布的比较研究[J]. 中国管理科学, 2015, 23(10): 11-18.
[5]	蒋翠侠, 张世英. 金融高阶矩风险溢出效应研究[J]. 中国管理科学, 2009, 17(1): 17-28.
[6]	蒋翠侠, 许启发, 张世英. 金融市场条件高阶矩风险与动态组合投资[J]. 中国管理科学, 2007, 15(1): 27-33.