 |
Social media (Lawrence K. Ho, Los Angeles Times / March 14, 2012) |
Computer programs that scour posts on the Internet are helping institutional stock traders spot comments about companies that could point their way up — or down.
By Walter Hamilton, Los Angeles Times
8:46 PM PDT, March 30, 2012
Professional traders are developing computer programs that scour Internet posts in search of the next stock market darling. Their technology analyzes everything uttered about a company — good or bad, racy or mundane — and spits out trading recommendations.
The logic is that the popular sentiment expressed by millions of people on sites such as Twitter and Facebook yields investment clues that can't be gleaned from financial statements or analyst reports.
Complaints about a new product can foreshadow a drop in sales and, hence, in a stock. Internet posts also can showcase investor emotions, such as when panic might cause people to irrationally dump a stock.
Given the frenzy over social media, it was only a matter of time before hedge funds and other Wall Street trading shops tried to cash in. Unlike individual investors who look for slow and steady returns, these traders thirst for any sort of an edge in a lightning-fast market.
"Any time you can predict the future, there's money to be made," said Richard Peterson, whose Santa Monica-based MarketPsych sells social-media data to hedge funds and banks.
Sound far-fetched? Plenty of skeptics, including a fair number on Wall Street, question whether any of this actually works. And even if it does, it would be Wall Street professionals who benefit rather than people actually using the sites.
For example, on Jan. 18 Peterson's computers recommended buying semiconductor maker Spreadtrum Communications Inc. based on comments on a Twitter-like site for investors. The shares had fallen sharply and several investors said they might buy the stock if a fledgling recovery could hold.
"So tempting to jump back in," one investor wrote.
Peterson's computers read that as a bullish sign for anyone who could jump in first. A trader buying 1,000 shares at the next day's opening would have made more than $500 as the shares rose nearly 4% over next five days.
Doubters scoff at the idea of trading based on unfiltered rants in the blogosphere.
It reminds John Buckingham, chief investment officer at Al Frank Asset Management Inc. in Aliso Viejo, of the late-1990s dot-com bubble. He'd mention stocks on CNBC, he said, and even those he recommended selling would surge.
"The stock chart would go up [on the screen] and people would buy it," Buckingham said. "Investors should focus on quality companies, not what somebody is saying in a chat room or on a social-media site."
Even social-media practitioners acknowledge how unorthodox their strategy is.
"I openly accept that there are a lot of eye-rolls out there," said Randy Saaf, co-founder of AlphaGenius Technologies in Los Angeles. "There's a long path ahead of us to get people to shift away from what we've been taught by Warren Buffett."
Peterson admits his method doesn't always work — a recommendation to buy water company Heckmann Corp. on Jan. 26 backfired when the stock fell 11% by the time the sell signal came a few days later.
But he views social-media investing as a promising strategy in an industry that has long looked for secret formulas to trading stocks. MarketPsych, AlphaGenius and other companies sell customized data feeds to hedge funds, high-frequency traders and other professionals.
A handful of studies have shown a correlation between sentiment on social-media sites and the performance of stock prices.
Johan Bollen, a professor at Indiana University, found that Twitter posts correctly predicted the direction of the Dow Jones industrial average 86.7% of the time.
His study analyzed almost 10 million tweets by 2.7 million users over a nearly 10-month period in 2008. It found that users' moods one day presaged the direction of the Dow Jones industrial average three days later.
"There's a lot of promise," Bollen said. "It's really tantalizing to be able to tap into the global zeitgeist."
Another study found a high correlation between a stock's popularity on Facebook and its performance. Arthur O'Connor, a Pace University doctoral student, examined 30 consumer stocks for the year that ended in June 2011. The stocks of companies with growing fan bases tended to outperform.
Some Wall Street veterans see the potential.
Sam Stovall, chief investment strategist at Standard & Poor's Corp., likens it to bring-your-children-to-work day when his firm's retail-stock analysts pepper their co-workers' children with questions about their consumer favorites.
"You want to be the first on Wall Street to know what's hot in the local junior high," Stovall said. "The feeling is that if you can't talk to young adults directly, maybe you can eavesdrop through social media."
No one has more faith in the strategy than Peterson.
The Stanford-trained psychiatrist used to see patients to earn enough money to pay programmers to write his social-media code. The bookcase in his office brims with titles such as "Nerds on Wall Street" and "Handbook of Emotional Regulation."
The 39-year-old, who was at the office one day recently in green khakis, worn hiking shoes and rimless glasses, says his computers sift through 2 million posts, tweets, blog posts and news articles a day, analyzing the psychological context of each.
He started a hedge fund during the depths of the global financial crisis in late 2008. It surged 40% in nine months. But after he tweaked his formula to improve it, the fund lost 1% a month for the next 15 months, and Peterson closed it at the end of 2010. He says he's fixed it and recently began managing a small amount of money for friends and family.
There's a lot of money to be made interpreting the moods of other investors, he said.
"It's huge," Peterson said. "The biggest barrier is convincing people that it matters."
walter.hamilton@latimes.com
Copyright © 2012, Los Angeles Times
