
期权隐含信息提取及对股票市场的影响: 基于机器学习的视角
Extraction of Implied Information from Options and Its Impact on the Stock Market: A Machine Learning Perspective
如何提取期权隐含信息并研究其对标的股票收益的影响一直是学界和业界关心的问题之一. 现有研究主要依赖某个单一维度: 在值程度(Moneyness)或者期限结构(Maturity), 来提取期权隐含信息, 如隐含波动率、隐含偏度或者隐含尾部风险等. 如何在两个维度上同时提取隐含信息, 并且如何从众多信息中提取共同信息因子是本文的研究重点. 为解决上述问题, 本文使用了主成分分析结合机器学习的方法, 从期权波动率曲面中提取隐含信息, 并检验其对标的股票收益率的可预测性. 区别于传统方法, PCA-LASSO可以捕捉期权隐含信息的时变性, 同时提炼出不同类型信息的共同驱动因子, 因此对股票收益率具有更好的预测能力.
How to extract implied information from options and study its impact on the returns of underlying stocks has always been a concern in both academia and industry. Existing research primarily relies on a single dimension: Either moneyness or maturity structure, to extract implied information, such as implied volatility, implied skewness, or implied tail risk, etc. How to extract implied information simultaneously from both dimensions, and how to extract common information factors from numerous pieces of information, are the focal points of this study. To address these issues, this paper utilizes a method combining principal component analysis with machine learning to extract implied information from the options volatility surface, and tests its predictability on the returns of the underlying stocks. Unlike traditional methods, PCA-LASSO can capture the time-varying nature of implied option information, while also refining common driving factors of different types of information, thus providing better predictive power for stock returns.
股指期权 / 期权隐含信息 / 主成分分析 / 机器学习 {{custom_keyword}} /
index option / option implied information / principle component analysis / machine learning {{custom_keyword}} /
表1 描述性统计 |
观测值 | 均值 | 标准差 | 最小值 | 最大值 | 偏度 | 峰度 | |
263 | 0.33 | 4.33 | -18.63 | 10.23 | -0.84 | 1.67 | |
13728 | 0.13 | 7.45 | 0.05 | 0.68 | 1.46 | 3.64 | |
20514 | 0.08 | 2.89 | -0.13 | 0.21 | 0.84 | 5.82 | |
20514 | 0.06 | 2.19 | -0.09 | 0.19 | 0.89 | 6.07 | |
21040 | 0.07 | 2.91 | -0.19 | 0.25 | 0.82 | 7.99 | |
12624 | 0.02 | 2.83 | -0.16 | 0.19 | 0.66 | 6.09 | |
21040 | 0.01 | 3.09 | -0.19 | 0.24 | 0.73 | 6.29 |
表2 样本外预测结果 |
MSFE-adj | |||
PCA-LASSO | 4.57*** | 3.13 | 0.01 |
PCA-EN | 4.56*** | 3.10 | 0.01 |
PCA-LR | 1.20* | 1.79 | 0.07 |
注: *, **, ***分别表示10%, 5%, 1% 的统计显著性.后表同 |
表3 不同预测方法收益预测结果的比较 |
MSFE-adj | ||||
PC1 | -1.69 | 0.68 | 0.25 | 6.26 |
PC2 | -3.62 | 0.65 | 0.26 | 8.19 |
PC3 | -2.32 | 0.87 | 0.19 | 6.89 |
PC4 | -1.00 | 1.07 | 0.14 | 5.57 |
PC5 | -0.59 | 1.13 | 0.13 | 5.16 |
IC | 1.55 | 1.21 | 0.11 | 3.57 |
Mean | 1.72*** | 2.89 | 0.01 | 2.85 |
Trim mean | 1.21*** | 1.84 | 0.04 | 3.36 |
Median | 1.19** | 1.88 | 0.03 | 3.38 |
PLS | -0.54 | 0.85 | 0.20 | 5.11 |
表4 与其它预测变量的比较 |
(1) | (2) | (3) MSFE-adj | (4) | (5) | (6) | (7) | (8) MSFE-adj | (9) | (10) |
DP | -0.05* | 1.28 | 0.10 | 4.62 | DFY | -0.04 | -0.23 | 0.59 | 4.61 |
DY | -0.37* | 1.48 | 0.07 | 4.94 | DFR | -0.01 | 0.31 | 0.38 | 4.58 |
EP | -1.88 | 0.58 | 0.28 | 6.45 | INFL | -0.09 | 0.00 | 0.50 | 4.66 |
DE | -2.04 | -1.88 | 0.97 | 6.61 | -1.53 | -0.25 | 0.40 | 6.10 | |
BM | -1.74 | 0.50 | 0.31 | 6.31 | -1.05 | -0.01 | 0.50 | 5.62 | |
SVAR | 0.32 | 0.95 | 0.17 | 4.25 | Ⅳ Skew | -0.73* | -1.28 | 0.10 | 5.30 |
NTIS | -0.91 | 0.23 | 0.41 | 5.48 | Ⅳ Slope | -0.63 | -0.68 | 0.25 | 5.20 |
TLB | -0.01 | 1.34 | 0.09 | 4.58 | VIX | 0.42 | 0.97 | 0.17 | 4.15 |
LTY | -1.17 | 1.17 | 0.12 | 5.74 | VRP | 1.74 | 1.10 | 0.14 | 2.83 |
LTR | -0.08 | 0.84 | 0.20 | 4.65 | TR | 1.42 | 0.93 | 0.18 | 3.15 |
TMS | 0.06 | 0.99 | 0.16 | 4.51 | TRP | 1.76* | 1.27 | 0.10 | 2.81 |
表5 预测包含性检验 |
(1) 预测变量 | (2) | (3) 预测变量 | (4) | (5) 预测变量 | (6) |
DP | 0.00 | TMS | 0.00 | TRP | 0.08 |
DY | 0.00 | DFY | 0.03 | PC1 | 0.00 |
EP | 0.00 | DFR | 0.00 | PC2 | 0.00 |
DE | 0.02 | INFL | 0.00 | PC3 | 0.00 |
BM | 0.00 | 0.03 | PC4 | 0.00 | |
SVAR | 0.03 | 0.03 | PC5 | 0.00 | |
NTIS | 0.00 | Ⅳ Skew | 0.00 | PLS | 0.00 |
TLB | 0.00 | Ⅳ Slope | 0.00 | Mean | 0.07 |
LTY | 0.00 | VIX | 0.00 | Median | 0.03 |
LTR | 0.00 | TR | 0.03 | Trim Mean | 0.00 |
表6 资产配置结果 |
面板A: | 面板B: | 面板C: | ||||||
SR | SR | SR | ||||||
PCA-LASSO | 6.19 | 0.79 | 4.93 | 0.72 | 3.27 | 0.57 | ||
PCA-EN | 6.16 | 0.76 | 4.49 | 0.73 | 3.05 | 0.55 |
表7 经济状态关联分析 |
(1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | (9) | (10) | |
RU | MU | FU | ADSI | CFNAI | SRP | KCFSI | IPG | PG | GDPG | |
面板 A: MUF1 预测结果 | ||||||||||
-0.32*** | -0.31** | -0.23** | 0.29*** | 0.24** | -0.23** | -0.33** | 0.2 | 0.16** | 0.26** | |
[-2.60] | [-2.57] | [-1.97] | [2.80] | [2.40] | [-2.10] | [-2.10] | [1.10] | [1.96] | [2.01] | |
10.63 | 9.94 | 5.65 | 8.43 | 6.15 | 5.82 | 10.98 | 6.44 | 0.52 | 10.10 | |
面板 B: MUF2 预测结果 | ||||||||||
-0.16 | -0.13 | -0.06 | 0.15 | 0.11 | -0.16 | -0.17 | -0.03 | -0.07 | 0.25*** | |
[-1.20] | [-1.00] | [-0.41] | [1.50] | [1.10] | [-1.60] | [-1.10] | [-0.28] | [-0.68] | [3.20] | |
2.61 | 1.93 | 0.34 | 2.24 | 1.17 | 2.43 | 2.74 | 0.07 | 0.63 | 6.72 | |
面板 C: MUF3 预测结果 | ||||||||||
0.13*** | 0.11** | 0.16*** | -0.06 | -0.06 | 0.16* | 0.10 | -0.03 | 0.01 | -0.06 | |
[2.80] | [2.00] | [3.10] | [-0.72] | [-0.83] | [1.70] | [1.60] | [-0.29] | [0.41] | [-0.62] | |
1.65 | 1.35 | 2.69 | 0.38 | 0.42 | 2.60 | 0.98 | 0.08 | 0.01 | 0.37 |
陈坚, 张轶凡, 洪集民, 期权隐含尾部风险及其对股票收益率的预测[J]. 管理科学学报, 2019, 22 (10): 72- 81.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
Guan Y, Dy J, (2009). Sparse Probabilistic Principal Component Analysis[C]// Proceedings of the Twelth International Conference on Artificial Intelligence and Statistics. PMLR: 185–192.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
Han Y, Liu F, Tang X, (2020). The Information Content of the Implied Volatility Surface: Can Option Prices Predict Jumps?[R]. Working Paper.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
Rapach D, Zhou G, (2019). Sparse Macro Factors[R]. SSRN Working Paper.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
/
〈 |
|
〉 |