Social media bots and stock markets

• Author: Rui Fan,Oleksandr Talavera,Vu Tran
• DOI: http://doi.org/10.1111/eufm.12245
• Published date: 01 June 2020
• Date: 01 June 2020

Eur Financ Manag. 2020;26:753–777. wileyonlinelibrary.com/journal/eufm © 2019 John Wiley & Sons Ltd.

|

753

DOI: 10.1111/eufm.12245

ORIGINAL ARTICLE

Social media bots and stock markets

Rui Fan

1

|

Oleksandr Talavera

2

|

Vu Tran

1

1

School of Management, Swansea

University, Swansea, UK

2

Birmingham Business School, University

of Birmingham, Birmingham, UK

Correspondence

Vu Tran, School of Management,

Swansea University, Fabian Way,

Swansea SA1 8EN, UK.

Email: v.tran@swansea.ac.uk

Abstract

This study examines the link between information

spread by social media bots and stock trading. Based

on a large sample of tweets mentioning 55 companies in

the FTSE 100 composites, we find significant relations

between bot tweets and stock returns, volatility, and

trading volume at both daily and intraday levels. These

results are also confirmed by an event study of stock

response following abnormal increases in the volume of

tweets. The findings are robust to various specifications,

including controlling for traditional news channel,

alternative measures of volatility, information flows in

pretrading hours, and different measures of sentiment.

KEYWORDS

computational linguistics, investor sentiment, noise traders, social

media bots, text classification

JEL CLASSIFICATION

G12; G14; L86

1

|

INTRODUCTION

Recent years have witnessed the increasing importance of social media platforms as alternative

information sources. Given the enormous volume and the rapid speed of information

transmission, social media provides a more comprehensive real‐time news database compared

to traditional media channels (e.g., Zhang, Fuehres, & Gloor, 2011). However, this large amount

of instant information can also contain potential noise that might mislead readers. These

concerns are more critical given the recent rise of social bots, cybots, and social media farms

EUROPEAN

FINANCIAL MANAGEMENT

We thank John Doukas (the Editor) and an anonymous referee for their valuable suggestions. We also would like to

thank participants of the Swansea research seminar, 2018 Royal Economic Society Annual Conference, and Financial

Management Association 2019 European Conference for helpful comments and suggestions. Any remaining errors are

our own. The standard disclaimer applies.

(e.g., Ferrara, Varol, Davis, Menczer, & Flammini, 2016), which could be weaponized to

disseminate fake news and manipulate stock markets (Forbes, 2017).

Indeed, the influence of social media bot activities on stock markets is not negligible. In our

data, there are cases in which a sudden increase in the volume of automated (bot) tweets is

associated with significant changes in stock returns. For example, on 1 May 2017, there was an

upsurge in the number of ‘bot’tweets with positive sentiment for Pearson from 18 to 4,349. On

the same day, the Pearson stock price increased by 1.01%, compared to a 0.63% rise of the FTSE

100. Another case is dated on 29 March 2017. The number of bot‐created tweets containing the

keyword ‘Barclays’increased from 5 to 14,668 (all with negative sentiment) and Barclays stock

price decreased by 0.35%. In the following week, Barclays’shares lost over 5%. These concerns

and stylized observations lead to a question of whether there is an empirically justified link

between information spread by automated social media accounts and stock markets.

Our paper is related to a few strands of literature. First, this paper contributes to the recent

literature on social media and the stock market. Multiple studies have suggested that stock market

participation and Twitter use are positively correlated (e.g., Bonaparte & Kumar, 2013). Tweets can

also be used to forecast aggregate market indexes and individual stock performance (e.g., Sprenger,

Sandner, Tumasjan, & Welpe, 2014a; Sprenger, Tumasjan, Sandner, & Welpe, 2014b; Zhang et al.,

2011). A few papers investigate the link between social media and stock market manipulation (e.g.,

Nasseri, Tucker, & De Cesare, 2015; Renault, 2017b). However, these studies do not directly

consider the fact that not all messages posted on social media are created by humans. Some Twitter

users are bots, automated computer algorithms that are designed to pump intended information

into public domains. Given that Twitter bots’tweets are autonomously created and spread, they can

potentially contain helpful information, noise, and even unreliable information.

1

Thus, it is

reasonable to expect differences in the effects of bot tweets and human tweets on stock markets.

This has not yet been considered in the existing literature.

In addition, most existing literature investigates the influence of information spread in professional

investing platforms and/or social media on daily stock prices (e.g., Rakowski, Shirley, & Stark, 2018;

Sprenger, Tumasjan, et al., 2014b). However, the use of daily data might not always capture the

feature of swift information flows in social media. There are a few recent studies (e.g., Behrendt &

Schmidt, 2018; Renault, 2017a) examining the impacts of social media information on intraday stock

prices. Renault (2017a) finds that online investor sentiment derived from messages posted on

StockTwits can help predict intraday stock returns, while Behrendt and Schmidt (2018) do not find

economically meaningful co‐movement between intraday volatility and stock‐related Twitter

information. Our study complements and contributes to this strand of literature in two ways. To

the best of our knowledge, this work is the first attempt to investigate the link between stock markets

and social media content created by automated accounts. Further, the unique dataset allows us to

account for the near‐instant information flows in social media that might have a different effect on

the stock market versus information from other sources that is typically spread at lower speed.

Second, this study makes a multidisciplinary contribution to the field of social media bots,

politics, and the stock market. Gorodnichenko, Pham, and Talavera (2018) detect spillover

effects from bots to human activities on social media during political events such as the 2016

Brexit Referendum and 2016 US Presidential Election. However, it is difficult to evaluate the

relationship between bots and political outcomes because the latter are not observable in real

1

Several studies have found evidence for the disproportional role of Twitter bots in spreading low‐credibility content (e.g., Shao et al., 2018). Additionally, some

studies (e.g., Kogan, Moskowitz, & Niessner, 2018) have examined the impacts of fake news on stock markets, while our paper focuses on the information

spread by social media bots rather than fake information in general.

754

|

EUROPEAN

FINANCIAL MANAGEMENT

FAN ET AL.