Skip to main content
See all news

Scheller News


Yu Jeffrey Hu, associate professor of IT Management

Researchers Find Social Media Able to Predict Stock Returns

Researchers from three prestigious schools - Georgia Tech Scheller College of Business, Purdue University, and City University of Hong Kong - have released what they believe is the first study proving the ability of social media to predict future stock returns and earnings surprises.

Forthcoming in the leading journal The Review of Financial Studies, the study (“Wisdom of Crowds: The Value of Stock Opinions Transmitted through Social Media”), examines data from 2005-2012 available on social investing site Seeking Alpha and compares it to market data for that period.

Citing the study’s conclusion that peer opinions published online reliably predict positive or negative stock returns anywhere from one month to three years in advance, Scheller College IT Management Associate Professor Yu Jeffrey Hu says, “Traditionally the domain of professional forecasters, financial analysis is increasingly being performed and broadcast by investors themselves.”

The study’s researchers performed textual analysis of more than seven years’ worth of material posted to the web site Seeking Alpha. Seeking Alpha is a crowd-sourced investing site that allows contributors to publish research and ideas, which the community of users then ‘peer-reviews.’

The data included more than 97,000 articles and 459,679 comments. Articles were written by 6,500 authors and covered over 7,000 stocks. When compared to market data and articles from Dow Jones, analysis showed:

  • Articles on stock investing and community comments on the Seeking Alpha site predict stock returns over every time-frame examined: three months, six months, one year and three years. (This was not true of previous studies of the predictive value of short chatter messages posted on Internet message boards, which demonstrated no predictive value. Unlike previous sell-side research showing that financial analyst opinions are quickly incorporated into the market price, this study finds that the value relevant information on Seeking Alpha site affects the market price at a slower pace).
  • The Seeking Alpha user community successfully identified the predictive value of authors in real time. (When the community disagreed with authors, their opinions had predictive value. Authors who were historically accurate met with less or no community disagreement).
  • Community sentiment - either positive or negative - was more accurate in predicting future stock prices and earnings surprises than Seeking Alpha articles alone, sell-side analysis, or similar content from Dow Jones.

Hu and his fellow researchers concluded, “We find that the opinions revealed on this site (Seeking Alpha) strongly predict future stock returns and earnings surprises. The predictability holds even after controlling for the effect of traditional advice sources, such as financial analysts and newspaper articles. Together, our findings point to the usefulness of peer-based advice in financial markets.

“Social media outlets are unique in the sense that they enable direct and immediate interactions among users. These interactions, combined with the seeming intelligence of the ‘crowd,’ may be one of the primary reasons social media platforms are able to produce value-relevant content that is incremental to that revealed through traditional news channels,” Hu says.

Professor Hu worked on this independent study with Hailiang Chen of City University of Hong Kong, Prabuddha De and Byoung-Hyoun Hwang of Purdue University. The study was conducted independently from Seeking Alpha.

 

 

CONTACTS
Hope Wilson
Director of Communications
Brad Dixon
Assistant Director of Communications