Social media has already changed the way people find new music, TV shows, and cat pictures, but distilling your friends’ sense of taste is a natural function of gathering all their opinions. Using social media for decisions that normally rely on years of experience and expertise seems less reliable.

Crowdfunding Wisdom Of Crowds

But a recent study from the City University of Hong Kong, Purdue University and Georgia Institute of Technology argues that social media is able to consistently beat the market, as long as it has certain characteristics in place (h/t Yuliya Chernova at The Wall Street Journal).

In their paper Wisdom of Crowds: The value of stock opinions transmitted through social media, authors Hailiang Chen, Prabuddha De, Yu (Jeffrey) Hu, and Byoung-Hyoun Hwang looked at a articles written on SeekingAlpha and measured the position of the article by counting the number of negative words it used (an absence of negative words signaling a bullish stance, for example) as well as the negative words in the comments.

On average, someone investing according to the stance of the combined articles and comments and holding positions for three months would consistently beat the market, the study found. This finding held even when a number of variables were controlled for (analyst recommendations, upgrade/downgrades, earnings surprises, and negative word ratios in traditional media).

Study finds that wisdom of crowds also reflects earnings surprises

The first reaction to this is that SeekingAlpha might be moving the market instead of predicting it, so the authors also regressed for the ratio of negative words to earnings surprise. The SeekingAlpha community probably has enough capital to impact stock prices, for small and mid-caps at the very least, but there’s no reasonable way for it to affect a company’s fundamentals. Sure enough, the study found that the aggregate opinion of SeekingAlpha articles and comments had a strong correlation with earnings surprises.

“We find that the fraction of negative words in SA articles and comments strongly predict subsequent scaled earnings surprises,” says the study. “The earnings-surprise predictability suggests that the opinions expressed in SA articles and comments indeed provide value-relevant information (beyond that provided by financial analysts).”

Social media can motivate more research on a stock

What’s even more interesting is that when comments and articles have different tones, it is the comments that are better predictors of future returns.

“Our evidence implies that followers disagree with authors more when the authors’ articles have been inconsistent,” says the study. “For these historically inconsistent authors, our evidence also suggests that in instances where the tone of comments is in disagreement with the tone of the underlying article, it is the tone of the comments that, statistically speaking, more reliably predicts subsequent stock market performance.”

We already know that financial professionals increasing use social media as a source for investment ideas, but they don’t just put their money down because someone suggested it. Instead, they read about an interesting stock and then do more research on their own. Agreement between articles and comments, which are the highest correlated with market performance, happens when the authors’ idea mesh with the independent research of the readers.

In other words, it could be that sites like SeekingAlpha are useful because they prompt independent research that wouldn’t have otherwise taken place.