Welcome to the Age of DIY Financial Analysis
Personal finance websites such as Seeking Alpha are popular platforms for peer-based investment advice. According to a recent study, these sites are also able to predict the stock market. An exhaustive textual analysis of Seeking Alpha content showed that the views expressed in the articles were strongly associated with stock returns on the day of publication and the period after; for example, when the fraction of negative words increased by 1 percent, abnormal returns on the day the article was published were 0.289 percent lower. The effect was stronger for articles that received more attention and for companies likely to be neglected by traditional media and professional analysts, says researcher Byoung-Hyoun Hwang of the Krannert School of Management at Purdue University. Hwang reported the study’s findings at a Queen’s School of Business conference.
If you’re looking for reliable stock-picking advice to pad your investment portfolio, you may prefer the views of professional analysts over those of amateurs blogging their views. But those so-called amateurs may be on to something: it appears peer-based investment advisers in the online world are giving the pros a run for their money. Surprisingly, they are also having a marked effect on stock prices.
Studies have shown that consumer-generated reviews that now rule the e-commerce world play an important role in shaping purchasing decisions across a wide range of products and services. Byoung-Hyoun Hwang, of the Krannert School of Management at Purdue University, has been researching the extent to which this phenomenon holds true in financial markets. He reported his findings at a behavioural research conference at Queen’s School of Business.
Hwang and his colleagues Hailiang Chen, Prabuddha De, and Yu (Jeffrey) Hu based their study on Seeking Alpha, the most popular personal finance website, with more than 5,000 contributing authors and 130,000 commenters. Articles on Seeking Alpha are generally written by investors who reveal their personal approach to stock picking and portfolio management; they are required to disclose their identity and holdings on the stocks they write about.
The researchers conducted a textual analysis of all articles published by Seeking Alpha between 2005 and 2012, focusing on the frequency of negative words used in each article. Previous research has shown that the frequency of negative words captures the tone of the article being studied.
Hwang and his colleagues found that the fraction of negative words (number of negative words divided by the total number of words) contained in Seeking Alpha articles negatively correlated with the corresponding company's stock returns. “The association is statistically significant and economically meaningful,” they note in their paper. Specifically, when the fraction of negative words increased by 1 percent, abnormal returns on the day the article was published were 0.289 percent lower.
The same was true for future stock returns: Seeking Alpha articles with a 1 percent higher fraction of negative words were associated with 0.379 percent lower abnormal stock returns over the ensuing three months, even after controlling for the company’s stock market performance on the day the article was published.
Hwang said they found that the association between abnormal returns and the views expressed on Seeking Alpha was stronger for articles that received more attention and for companies given short shrift by traditional media and professional investment research firms.
But do these articles really contain meaningful information on company fundamentals or are they merely inciting impressionable and inexperienced investors to call their brokers? To find out, the researchers looked at earnings surprises (the difference between reported earnings-per-share and the average of financial analysts’ most recent EPS forecasts).
“These findings point to the importance of social media as both a source of peer-based advice and a channel through which views become reflected in stock prices”
“If opinions expressed through Seeking Alpha were unrelated to firms’ fundamentals, or if information were spurious and already fully incorporated by financial analysts into their reported EPS forecasts, then no association should be observed between subsequent scaled earnings surprise and the measure of social-media views,” the researchers explain. In fact, they found that the fraction of negative words in articles published prior to the earnings announcement strongly predicted subsequent scaled earnings surprises.
“These findings point to the importance of social media as both a source of peer-based advice and a channel through which views become reflected in stock prices.”
What motivates people to share their investment intel with others? As Hwang and his colleagues point out in their paper, recent theories suggest that investors want to publicize investment ideas to receive constructive feedback, gain access to new ideas, and attract additional capital. (There may even be compensation involved: recently, Seeking Alpha has begun to pay contributors whose articles are deemed particularly insightful.)
Finance-oriented social-media websites are only too happy to provide that platform. They do not carry the baggage of financial analysts, who have inherent conflicts of interest and few incentives to cover smaller, less popular companies. On the other hand, the Wild West nature and lack of regulation of the internet make investment blogs easy to “game” by stock manipulators. With their growing popularity and the high stakes at play, calls for greater online regulation may be heard in the years ahead.
— Alan Morantz