基于生成对抗网络的混频资产定价研究

doi:10.16381/j.cnki.issn1003-207x.2022.2261

Abstract

Abstract:

Asset pricing explores the potential and latent factors that drive the prices of and returns on real and financial assets. In the big data era， asset pricing is influenced by many factors， such as high frequency market information， low frequency firm characteristic information and macroeconomic information. In this sense， typical mixed frequency data issue then arises. In addition， test assets play a critical role in the process of estimating pricing kernels. However， artificially constructing test assets to examine and optimize asset pricing models can not fully identify weaknesses in model pricing. How to efficiently capture useful pricing factors from multi-source heterogeneous information and achieve accurate asset pricing is valuable to practitioners， regulators， and academic researchers alike. To this end， the concept of mixed frequency stochastic discount factor （MF-SDF） is proposed through introducing a mixed data sampling （MIDAS） approach into the stochastic discount factor （SDF） theory. MF-SDF is used as a criterion function to construct a generative adversarial network （GAN） which solves the asset pricing issue under the mixed frequency data environment. Taken together， the proposed MIDAS-SDF-GAN model is able to handle mixed frequency data via the MIDAS approach and model the complex nonlinear pattern hidden in high-dimensional pricing factors through deep learning. It can employ the generative adversarial game mechanism to construct test assets， guiding itself to learn pricing patterns and continuously improve the pricing power in a data-driven way. All Chinese A-stock listed firms from January 2000 to December 2020 are chosen as our sample data and construct a large mixed frequency dataset containing 48 high frequency market variables， 153 low frequency firm characteristic variables and 11 low frequency macroeconomic variables. The data come from the CSMAR database. Then， the performance of the MIDAS-SDF-GAN model in asset pricing is empirically investigated. The results show that： 1） The MIDAS-SDF-GAN model is able to construct test assets in a data-driven way and guide itself to learn the pricing mechanism. It outperforms the other four competing models in both stock return prediction and investment profitability. 2） The MIDAS-SDF-GAN model is able to fully exploit information in mixed frequency data. It anchors the pricing basis through low frequency firm characteristic information and macroeconomic information， and optimizes the pricing accuracy through high frequency market information. These two types of information work together to improve the pricing effect. The study is an example of applying deep learning methods to the field of asset pricing. It is of great significance to solve the problems encountered in the process of asset pricing with the help of advanced artificial intelligence technology under the framework of financial theory.

Key words: asset pricing, stochastic discount factor, mixed data sampling （MIDAS）, generative adversarial network （GAN）, test assets

CLC Number:

F832

Qifa Xu,Zezhou Wang,Cuixia Jiang. Research on Mixed Frequency Asset Pricing Based on Generative Adversarial Network[J]. Chinese Journal of Management Science, 2024, 32(11): 53-64.

Figures/Tables 12

模型	$R t o t a l 2$	EV
MIDAS-SDF-GAN	0.18	0.19
GAN	0.15	0.15
DL-MIDAS	0.12	0.13
ANN	0.14	0.14
LSDF	0.10	0.10

投资组合	$α$					EV
投资组合	MIDAS-SDF-GAN	GAN	DL-MIDAS	ANN	LSDF	MIDAS-SDF-GAN	GAN	DL-MIDAS	ANN	LSDF
bm单排序投资组合	0.0475	0.0477	0.0500	0.0507	0.0501	0.8072	0.7845	0.7798	0.7591	0.7821
mom36m单排序投资组合	0.0450	0.0524	0.0532	0.0482	0.0607	0.9306	0.8783	0.9091	0.9208	0.8769
volatility单排序投资组合	0.0537	0.0595	0.0577	0.0637	0.0623	0.8010	0.7244	0.7548	0.7059	0.7230
lev单排序投资组合	0.0674	0.0892	0.0797	0.0891	0.0847	0.7943	0.5387	0.6693	0.5897	0.5821
bm-mve双排序投资组合	0.0333	0.0441	0.0397	0.0428	0.0428	0.9078	0.7940	0.8535	0.8241	0.8113
mom36m-chmom双排序投资组合	0.0333	0.0441	0.0397	0.0428	0.0428	0.9192	0.8745	0.8987	0.9119	0.8693
I/A单排序投资组合	0.0358	0.0528	0.0453	0.0480	0.0613	0.9442	0.8780	0.8437	0.9184	0.8744
ROE单排序投资组合	0.0382	0.0526	0.0447	0.0480	0.0612	0.9492	0.8775	0.9274	0.9184	0.8743
I/A-ROE双排序投资组合	0.0447	0.0532	0.0545	0.0490	0.0620	0.9255	0.8776	0.9066	0.9078	0.8713

模型	$α$	EV
MIDAS-SDF-GAN	0.0489	0.7934
GAN	0.0514	0.7514
DL-MIDAS	0.0521	0.7683
ANN	0.0554	0.7356
LSDF	0.0524	0.7537

References 24

1	Cochrane J H. Asset pricing theory： Revised edition［M］. Princeton： Princeton University Press， 2005.
2	Lettau M， Pelger M. Estimating latent asset-pricing factors［J］. Journal of Econometrics， 2020， 218（1）： 1-31.
3	Freyberger J， Neuhierl A， Weber M. Dissecting characteristics nonparametrically［J］. The Review of Financial Studies， 2020， 33（5）： 2326-2377.
4	Kozak S， Nagel S， Santosh S. Shrinking the cross-section［J］. Journal of Financial Economics， 2020， 135（2）： 271-292.
5	Kelly B， Pruitt S， Su Y N. Characteristics are covariances： A unified model of risk and return［J］. Journal of Financial Economics， 2019， 134（3）： 501-524.
6	Gu S H， Kelly B， Xiu D C. Empirical asset pricing via machine learning［J］. The Review of Financial Studies， 2020， 33（5）： 2223-2273.
7	Orimoloye L O， Sung M C， Ma T J， et al. Comparing the effectiveness of deep feedforward neural networks and shallow architectures for predicting stock price indices［J］.Expert Systems with Applications，2020，139： 112828.
8	Wu W B， Chen J Q， Yang Z B， et al. A cross-sectional machine learning approach for hedge fund return prediction and selection［J］. Management Science， 2020， 67（7）： 4577-4601.
9	姜富伟，马甜，张宏伟. 高风险低收益？基于机器学习的动态CAPM模型解释［J］. 管理科学学报， 2021， 24（1）： 109-126.
	Jiang F W， Ma T， Zhang H W. High risk low return？ Explanation from machine learning based conditional CAPM model［J］. Journal of Management Science and Engineering， 2021， 24（1）： 109-126.
10	Gu S H， Kelly B， Xiu D C. Autoencoder asset pricing models［J］. Journal of Econometrics， 2021， 222（1， Part B）： 429-450.
11	Fama E F， French K R. International tests of a five-factor asset pricing model［J］. Journal of Financial Economics， 2017， 123（3）： 441-463.
12	Green J， Hand J R M， Zhang X F. The Characteristics that provide independent information about average U.S. monthly stock returns［J］. The Review of Financial Studies， 2017， 30（12）： 4389-4436.
13	Fama E F， French K R. Common risk factors in the returns on stocks and bonds［J］. Journal of Financial Economics， 1993， 33（1）： 3-56.
14	Nagel S， Singleton K J. Estimation and evaluation of conditional asset pricing models［J］. The Journal of Finance， 2011， 66（3）： 873-909.
15	Hansen L P， Jagannathan R. Assessing specification errors in stochastic discount factor models［J］. The Journal of Finance， 1997， 52（2）： 557-590.
16	Goodfellow I， Pouget-Abadie J， Mirza M， et al. Generative adversarial nets［J］. Advances in Neural Information Processing Systems， 2014， 27： 2672-2680.
17	Chen L Y， Pelger M， Zhu J. Deep learning in asset pricing［J］.Management Science， 2024，70（2）：714-750.
18	Ghysels E， Santa-Clara P， Valkanov R. Predicting volatility： Getting the most out of return data sampled at different frequencies［J］. Journal of Econometrics， 2006， 131（1）： 59-95.
19	许启发，卓杏轩，蒋翠侠. 反向有约束混频数据模型的市场化利率预测［J］. 管理科学学报， 2019， 22（10）： 55-71.
	Xu Q F， Zhuo X X， Jiang C X. Predicting market interest rates via reverse restricted MIDAS model［J］. Journal of Management Science and Engineering， 2019， 22（10）： 55-71.
20	谭德凯，田利辉.黄金是股票市场的“避险天堂”吗？——基于动态条件相关混频数据抽样模型［J］. 中国管理科学， 2022， 30（10）： 1-12.
	Tan D K， Tian L H. Is gold a safe haven of the stock market？Based on dynamic conditional correlation mixed data sampling model［J］. Chinese Journal of Management Science， 2022， 30（10）： 1-12.
21	夏婷，闻岳春. 经济不确定性是股市波动的因子吗？——基于GARCH-MIDAS模型的分析［J］. 中国管理科学， 2018， 26（12）： 1-11.
	Xia T， Wen Y C. Does economic uncertainty matter for stock market volatility？ An analysis based on GARCH-MIDAS［J］. Chinese Journal of Management Science， 2018， 26（12）： 1-11.
22	Yang C P， Zhang R G. Does mixed-frequency investor sentiment impact stock returns？ Based on the empirical study of MIDAS regression model［J］. Applied Economics， 2014， 46（9）： 966-972.
23	Leippold M， Wang Q， Zhou W Y. Machine learning in the Chinese stock market［J］. Journal of Financial Economics， 2022， 145（2）： 64-82.
24	张鹏，党世力，黄梅雨. 基于机器学习预测股票收益率的两步骤M-SV投资组合优化［J］.中国管理科学， 2022，31（12）：96-106.
	Zhang P， Dang S L， Huang M Y. Two-stage mean semi-variance portfolio optimization with stock return prediction using machine learning［J］. Chinese Journal of Management Science， 2022，31（12）：96-106.

模型	混频数据处理能力	生成对抗博弈机制	SDF理论架构	非线性定价结构
MIDAS-SDF-GAN	√	√	√	√
GAN	×	√	√	√
DL-MIDAS	√	×	√	√
ANN	×	×	×	√
LSDF	×	×	√	×

模型	SR	AR	MaxLoss	MD
MIDAS-SDF-GAN	1.1477	0.1153	0.0456	0.0081
GAN	0.5426	0.0438	0.0545	0.0307
DL-MIDAS	0.7813	0.0615	0.0172	0.0123
ANN	0.4604	0.0402	0.0572	0.0326
LSDF	0.5491	0.0305	0.0551	0.0662

[1]	Zezhou Wang, Qifa Xu, Cuixia Jiang. Can Interaction and Dissemination of Mixed-Frequency Information in Companies' Multi-Layer Relationship Networks Improve Asset Pricing? Asset Pricing Study Based on Graph Neural Networks [J]. Chinese Journal of Management Science, 2026, 34(4): 47-62.
[2]	Juanjuan Lin, Zhigang Huang, Yong Tang. Data Quality, Quantity and Data Asset Pricing: Based on the Perspective of Consumer Heterogeneity [J]. Chinese Journal of Management Science, 2025, 33(5): 88-98.
[3]	Xingyi Li, Zhongfei Li, Qiqian Li, Yujun Liu, Wenjin Tang. A Review of Research on Asset Return Prediction Based on Machine Learning [J]. Chinese Journal of Management Science, 2025, 33(1): 311-322.
[4]	Xuanming Ni,Tiantian Zheng,Huimin Zhao,Kangping Wu. Asset Pricing Based on the Optimal Idiosyncratic Return Factor [J]. Chinese Journal of Management Science, 2024, 32(8): 50-60.
[5]	CHEN Miao-xin, HUANG Zhen-wei. Long Memory Volatility and Cross-section Stock Returns: Empirical Research in Chinese Stock Market [J]. Chinese Journal of Management Science, 2023, 31(4): 1-10.
[6]	LU Jing, ZHANG Yin-ying. Idiosyncratic Volatility Puzzle and Its Estimation Model [J]. Chinese Journal of Management Science, 2022, 30(9): 36-48.
[7]	LIU Yan, ZHU Hong-quan. Individual Investor or Institutional Investor, Who Dominates Asset Pricing in Chinese A-share Stock Market? [J]. Chinese Journal of Management Science, 2018, 26(4): 120-130.
[8]	XU Yuan-dong. The Consistency of Logical Structure about Asset Pricing in BSV, DHS model and Asset Pricing under Ambiguity [J]. Chinese Journal of Management Science, 2017, 25(6): 22-31.
[9]	GONG Xu, WEN Feng-hua, HUANG Chuang-xia, YANG Xiao-guang. Downside Risk, Signed Jump Risk and Asset Pricing of Industry Portfolios [J]. Chinese Journal of Management Science, 2017, 25(10): 1-10.
[10]	LIU Wei-qi, XING Hong-wei, ZHANG Xin-dong. Investment Preference and the Idiosyncratic Volatility Puzzle——Evidence from China Stock Market [J]. Chinese Journal of Management Science, 2014, 22(8): 10-20.
[11]	ZHOU Fang, ZHANG Wei, ZHOU Bing. Capital Asset Pricing Model Based on Liquidity Risk [J]. Chinese Journal of Management Science, 2013, 21(5): 1-7.
[12]	WU Xin-yu, ZHOU Hai-lin, MA Chao-qun, WANG Shou-yang. Stochastic Discount Factor-Based Approach for Warrant Pricing [J]. Chinese Journal of Management Science, 2012, (4): 1-7.
[13]	ZHU Hong-quan, CHEN Lin, PAN Ning-ning. Industry, Local and Market Information, Who Dominates Price Movement in Chinese Stock Market? [J]. Chinese Journal of Management Science, 2011, 19(4): 1-8.
[14]	LUO Deng-yue, WANG Chun-feng, FANG Zhen-ming. An Empirical Research on the Relationship between Aggregate Liquidity and Asset Pricing in China Stock Market [J]. Chinese Journal of Management Science, 2007, 15(2): 33-38.
[15]	LU Jing, TANG Xiao-wo. The Empirical Study on Multi-Factor Pricing Model Based on Liquidity Risk [J]. Chinese Journal of Management Science, 2006, (5): 45-51.

Research on Mixed Frequency Asset Pricing Based on Generative Adversarial Network

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 12

References 24

Related Articles 15

Metrics

Comments

Recommended 0

模型	SR	AR	MaxLoss	MD
MIDAS-SDF-GAN	0.90	0.05	0.05	0.06
GAN	0.42	0.03	0.12	0.15
DL-MIDAS	0.62	0.03	0.06	0.14
ANN	0.32	0.02	0.15	0.25
LSDF	0.23	0.01	0.12	0.35