Trading on a World of Sentiment

Investors are consulting software that analyzes millions of online comments about companies to help them decide when to buyand sellstocks

Dr. Richard Peterson, a psychiatrist by training, knows first-hand that emotions often rule investors. After the financial crash of 2008, he invested in stocks such as Citigroup (C), when others were selling. His now-closed MarketPsy Long-Short Fund went on to outperform the Standard & Poor's 500-stock index from September 2008 through the end of last year.

"In the end, 75 percent of the overall strategy was based on sentiment," says Peterson, managing partner of the fund's parent MarketPsych Capital. The fund, which began with $1 million, had a 28 percent return from Sept. 2, 2008, through Dec. 31, 2010. The S&P 500 lost 1.6 percent over the same period.

Peterson was able to do this by using software that his company developed over seven years that looks at online financial news, financial social media, and corporate interviews. It quantifies 400 types of sentiments and topics from optimism and anger and management changes or product releases for 6,000 U.S. stocks and exchange-traded funds. He now sells that software to hedge funds and other financial firms.

Going Long on Swine Flu Fears

In recent years, more investors have begun scouring the Web for subjective information about stocks. For Peterson's hedge fund, it meant analyzing prevailing sentiment and buying some stocks that investors were generally pessimistic about. For example, the World Health Organization declared a public health emergency over H1N1 swine flu on Apr. 25, 2009. Investors on American Airlines' (AMR) online stock message boards signaled their concern about possible reductions in airline travel. Two days later, the stock slumped 13 percent. On Apr. 30, 2009, the MarketPsy hedge fund bought AMR at $4.81 a share. It sold the position six days later for $5.95 a share, a 24 percent gain.

MarketPsych, like other quantitative investment firms, uses algorithms and computerized models to trade stocks. Typically, those models have included income statements, market-related data, prices, volume and earnings forecasts, which computers then crunch to help inform trading decisions. Over the years, as quantitative investing became more popular, the returns associated with those conventional data points have slowly eroded, says Rochester Cahan, quantitative equity strategist at Deutsche Bank (DBK).

"In the past two to three years, it has led a big drive among quants to find data that's less scrutinized," Cahan says. It spurred some hedge fund managers to look at more subjective data such as sentiment in news stories and to figure out a way to put that information into their trading models, he says.

Institutions Appreciate Fast Insights

MarketPsych faces competition from other companies that sell software and services to monitor sentiment about stocks on the Web, including Ravenpack, Thomson Reuters (TRI), and Bloomberg Businessweek's parent company, Bloomberg LP. RavenPack's business has quadrupled in the past year, and includes some of the world's top hedge funds, says Chief Executive Officer Armando Gonzalez, who declines to name any customers. "The institutional clients are starting to understand the value of news and having it faster than anybody else to detect the things you don't expect," he says. Gonzalez predicts the market will grow "significantly" in the next five years.

Just as a greater number of mainstream investors are starting to see the benefit of examining online news for sentiment, some hedge funds are already looking at other sources, such as transcripts of CEO interviews on CNBC, says Deutsche Bank's Cahan. Some firms are parsing disclosures in regulatory filings or analyzing stock discussions on blogging site Twitter, trying to glean sentiment from the way sentences are phrased. "It opens up a whole new world of data that investors have never used," he says.

Derwent Capital Markets, a family owned hedge fund, was scheduled to launch trading in February in a fund that would follow posts on Twitter and track sentiment on stocks. While the Derwent Absolute Return Fund has been delayed, the company said it may begin in April with about $39 million under management.

Road Show in Madoff's Shadow

In 2009, Peterson and his colleagues at MarketPsych went on a road show to New York to market their fund. The trip took place at the height of the Bernard Madoff Ponzi scheme scandal, a bad time to raise money for a hedge fund, Peterson says. "People would look at our trade list, when we bought Citigroup at 97¢, and say, `that could have gone bankrupt,'" Peterson remembers, even though he sold Citigroup at $1.20 a share. "We were taking advantage of an overreaction in the market." Ultimately, the economic climate proved too difficult to raise money for the fund, and he shelved it at the end of last year.

Peterson says that since he's no longer managing a hedge fund, he has more time to look at how sentiment affects stock returns. MarketPsych examined the daily performance of the top eight tech stocks over 13 years. It looked at factors such as how innovative a company is thought to be. Peterson says those that had the highest innovation perceptions—the top 14 percent—outperformed the average group by 10 basis points a day, which translated to a 25 percent annual excess return over the S&P 500.

Whereas in 2009, Peterson saw a great deal of overreaction in the market, now he says he sees underreaction. Apple (AAPL) is doing so well that some investors think it won't last, he says, though he remains bullish on the iPhone and iPad maker. He also likes Qualcomm (QCOM) because the company makes smartphone chips, and mobile devices are becoming more important to consumers. "Apple and Qualcomm are at the top of our list," he says. "They will probably outperform."

Before it's here, it's on the Bloomberg Terminal.