管理层的“弦外之音”,投资者能听得懂吗？——基于管理层语意的LSTM深度学习研究

财经论丛 ›› 2019, Vol. 35 ›› Issue (6): 63-72.

管理层的“弦外之音”,投资者能听得懂吗？——基于管理层语意的LSTM深度学习研究

杨七中¹, 马蓓丽²

1.南京财经大学会计学院,江苏南京 210023;
2.江苏理工学院商学院,江苏常州 213001

收稿日期:2018-07-16 出版日期:2019-06-10 发布日期:2019-06-13
作者简介:杨七中(1978-),男,江苏徐州人,南京财经大学会计学院副教授,博士;马蓓丽(1981-),女,江苏常州人,江苏理工学院商学院讲师。
基金资助:
江苏省社会科学基金一般项目(18EYB010);江苏省高校哲学社会科学研究重点项目(2018SJZDI088)

Can Chinese Investors Absorb the Meanings behind the Executives' Words?——LSTM Deep Learning Analysis Based on Management Meanings

YANG Qizhong¹, MA Beili²

1. School of Accountancy, Nanjing University of Finance and Economics, Nanjing 210023, China;
2.Business school, Jiangsu university of technology, Changzhou 213001, China

Received:2018-07-16 Online:2019-06-10 Published:2019-06-13

摘要/Abstract

摘要：

我国资本市场上的投资者能够听得懂管理层的“话里话、弦外音”吗？本文选用LSTM深度学习技术,对我国上市公司2010~2014年度业绩说明会上的管理层回复内容进行文本分析,研究管理层所表达的真正语意。结果发现,投资者能理解管理层的真实语意,资本市场对正面语意做出滞后显著的正向反应,而对负面语意做出更及时显著的负向反应;进一步基于Fama and French(1993)三因子模型研究管理层语意是否是投资者股票交易策略的因子,但未发现投资者把管理层语意作为交易策略因子的证据。本文的研究一方面表明业绩说明会这样的非财务信息披露渠道具有存在的价值并且验证了行为经济学领域的前景理论,另一方面LSTM深度学习技术能避免传统“词袋法”的缺陷,值得拓展到其他非财务信息文本分析领域,适当开展进一步应用。

关键词: 管理层语意, 文本分析, LSTM深度学习, 市场反应

Abstract:

Can Chinese capital market absorb the meanings behind the executives' words? Based on the management answering data at earnings communication conferences held by Chinese listed companies from 2010 to 2014, this paper applies the long-short term memory deep learning technique to analyze this issue. It finds that the investors can understand the real meanings transferred by the top management team, reacting positively to positive management meanings and negatively to negative meanings. This paper further investigates if management meaning can act as one of strategy trade factors by Fama-French three factors model. However, no evidence is found concerning this. The result indicates that for one thing, earnings communication conference is an efficient nonfinancial information disclosure channel, and for another, the LSTM technique, which can help avoid the drawback of “the bag of words” approach, deserves further develpment in the future.

Key words: Management Meanings, Textual Analysis, LSTM Deep Learning, Market Reaction

中图分类号:

F713.55

杨七中, 马蓓丽. 管理层的“弦外之音”,投资者能听得懂吗？——基于管理层语意的LSTM深度学习研究[J]. 财经论丛, 2019, 35(6): 63-72.

YANG Qizhong, MA Beili. Can Chinese Investors Absorb the Meanings behind the Executives' Words?——LSTM Deep Learning Analysis Based on Management Meanings[J]. , 2019, 35(6): 63-72.

参考文献

[1] 爱德华·霍尔.超越文化[M]. 北京:北京大学出版社, 2010.
[2] Tetlock P. C., Saar-Tsechansky M., Macskassy S. More than Words:Quantifying Language to Measure Firms' Fundamentals[J]. Journal of Finance, 2008, 63(3):1437-1467.
[3] Loughran T., Mcdonald B.When is ALiability not ALiability? Textual Analysis, Dictionaries and 10-Ks[J]. Journal of Finance, 2011, 66(1):35-65.
[4] 谢德仁,林乐.管理层语调能预示公司未来业绩吗?——基于我国上市公司年度业绩说明会的文本分析[J]. 会计研究,2015, (2):20-27.
[5] 林乐,谢德仁.投资者会听话听音吗?——基于管理层语调视角的实证研究[J]. 财经研究,2016, (7):28-39.
[6] 林乐,谢德仁.分析师荐股更新利用管理层语调吗?——基于业绩说明会的文本分析[J]. 管理世界,2017, (11):125-145.
[7] Jegadeesh, N., D.Wu.Word Power:A New Approach for Content Analysis.[J]. Journal of Financial Economics, 2013, 110:712-729.
[8] Price S. M., Doran J. S., Peterson D. R., et al. Earnings Conference Calls and Stock Returns:The Incremental Informativeness of Textual Tone[J]. Journal of Banking and Finance, 2012, 36(4):992-1011.
[9] Mayew W. J., Sethuraman M., Venkatachalam M. MD&A. Disclosure and the Firm's Ability to Continue as AGoing Concern[J]. The Accounting Review, 2018.
[10] Feng Li.Annual Report Readability, Current Earnings and Earnings Persistence[J]. Journal of Accounting and Economics, 2008, 45(2):221-247.
[11] Lawrence A.Individual Investors and Financial Disclosure[J]. Journal of Accounting and Economics, 2013, 56(1):130-147.
[12] Davis A. K., Ge W., Matsumoto D., Zhang J. L.The Effect of Manager-specific Optimism on The Tone of Earnings Conference Calls[J]. Review of Accounting Studies, 2015, 20(2):639-673.
[13] Purda L., Skillicorn D.Accounting Variables, Deception and A Bag of Words:Assessing The Tools of Fraud Detection[J]. Contemporary Accounting Research, 2015, 32(3):1193–1223.
[14] 彭红枫,赵海燕,周洋.借款陈述会影响借款成本和借款成功率吗？——基于网络借贷陈述的文本分析[J]. 金融研究,2016,(4):158-173.
[15] 朱朝晖,许文瀚.管理层语调是否配合了盈余管理行为[J]. 广东财经大学学报,2018,33(1):86-98.
[16] 孔翔宇,毕秀春,张曙光.财经新闻与股市预测——基于数据挖掘技术的实证分析[J]. 数理统计与管理,2016,35(2):215-224.
[17] 徐巍,陈冬华.自媒体披露的信息作用——来自新浪微博的实证证据[J]. 金融研究,2016,(3):157-173.
[18] 金德环,李岩.群体智慧:同伴观点与价值发现——来自社交媒体的经验证据[J]. 经济管理,2017,(12):157-173.
[19] 李增泉.激励机制与企业绩效——一项基于上市公司的实证研究[J]. 会计研究,2000,(1):24-30.
[20] Nicholas A.Bloom.Really Uncertain Business Cycles[J]. Economerica, 2018, 3(86):1031-1065.
[21] Leuz C., Wysocki P. D. The Economics of Disclosure and Financial Reporting Regulation: Evidence and Suggestions for Future Research[J]. Journal of Accounting Research, 2016, 54(2):525-622.