Want to know the future? Try looking at what people are searching for using Google.
Records of Google searches have been shown to predict box-office revenues, the popularity of new songs and video games, tourism, even flu epidemics. Google and the Centers for Disease Control and Prevention have shown that spikes in searches for terms related to influenza can detect flu outbreaks one to two weeks faster than the CDC’s own reports from the field.
Now comes research by a pair of Notre Dame finance faculty members suggesting that investors can profit by keeping their eyes on Google search trends.
Assistant professors Zhi Da and Pengjie Gao found that Google’s publicly available Search Volume Index data can gauge market sentiment, predict product sales and even identify stocks that are attracting the interest of retail investors. Such interest appears to lead to those stocks rising in price in the near term. In the case of initial public offerings (IPOs), the extra search attention correlated with a 6 percent price bump on the first day of trading, they found.
Warning to eager investors, though: The effect doesn’t last. Da, Gao and research partner Joseph Engelberg of the Kenan-Flagler Business School at the University of North Carolina at Chapel Hill began their study, “In Search of Attention,” with a question: How can you tell which stocks are attracting the attention of individual investors before the stocks are actually purchased by such investors?
Past attempts to gauge attention have relied on indirect measures such as counting the number of headlines and news stories appearing about a company. The assumption was that if a company was in the news a lot, individual investors were hearing about it and acting on the news.
That’s not always true, though, because not everyone sees a key news story. In fact, in the Internet Age, the researchers say, so much information bombards the public daily that investors may be paying less attention to any given source than in the past. On this point, they quote economist and Nobel Laureate Herbert Simon, who said, “[A] wealth of information creates a poverty of attention. ...”
To find out which companies were catching the attention of individual investors, the finance professors looked at searches for a particular company’s stock-ticker symbol. This was done to screen out people who are searching for a company by name to find information about its products or services, or people who were searching for a particular kind of “apple” instead of Apple, for example.
Da and Gao believe Google searches track amateur investors’ interests well because professionals don’t use Google to find news on a company. They get up-to-the-second information from comprehensive business-news services such as Bloomberg or have algorithms scanning news and public reports for certain terms, which can trigger computerized trades.
The professors also wanted to test the theory of another group of researchers—that unsophisticated individual investors have a “buy” bias. The theory is that such amateurs already know what stocks they can sell (only the ones they own), but their options are almost unlimited when it comes to buying. That wide-open field sets them on their Google-enabled quests for news.
The other reason the “buy” bias exists, Gao says, is that unsophisticated investors rarely have margin accounts or understand how to “short sell” (a way to profit when you believe a stock is headed south).
All of this means an upsurge in stock-symbol searching should lead to a bidding-up of the stock’s price.
And that’s exactly what the finance professors found when they compared the search data against public records of trades by retail investors.
Spikes in search frequency for a stock symbol coincided with the stock climbing the following two weeks. However, the gains almost disappeared 12 months later. The professors believe that’s due to more-sophisticated investors and company insiders, mindful of company fundamentals, taking profits or capitalizing on short-sell opportunities before prices wilt back to rational levels.
The researchers say the Google effect is most pronounced with small-cap stocks because those are more popular with individual investors than with institutions.
They found a similar search-spike effect with initial public offerings. IPOs with high search volume the week before the first day of trading outperformed those with fewer searches by 6 percent that first day. As with the non-IPO stocks, though, the IPOs that benefitted from a Google boost at birth eventually flared out and experienced a long-term price reversal, they report.
A dramatic example, though not included in the study period, involves the April 2010 IPO for Rosetta Stone, maker of language-learning software. One week prior to the public offering, Google search volume for the company’s ticker symbol, RST, increased by 28 percent compared with the previous six-week moving average, Gao says. The initial offering price was $18 per share. It opened for trading at $23 on April 16, 2010, and closed that day at $25.12. That’s a first-day return of 40 percent. But a year later, Rosetta Stone’s share price stood at $13.35.
Gao says the purpose of the research was to find a reliable way to track the attention of individual investors, not spot money-making opportunities in the stock market. However, he says he knows of at least one asset-management firm that has tested “some variation” of the methodology. Although they didn’t publish their findings until October 2011 (The Journal of Finance), the finance professors earlier presented papers at academic conferences and to financial-asset management companies. The paper can be read online at http://www.nd.edu/~zda/Google.pdf
Da and Gao’s paper won the prestigious 2010 Crowel First Prize for outstanding research, which is awarded by the Quantitative Research Group at PanAgora Asset Management. The paper also won first place during an academic competition sponsored by the Chicago Quantitative Alliance.
The researchers have since expanded their work with Google search data. In one paper, they describe an index of investor gloom that they’ve developed. Their so-called Financial and Economic Attitudes Revealed by Search (FEARS) index is drawn from records of searches for terms like “recession,” “bankruptcy” and “inflation.” It’s been shown to accurately predict low returns on stocks in the near term and higher returns later on.
In another study, they found that the volume of searches for a company’s most popular product can be used to predict revenue surprises (good and bad) that become public knowledge when a company announces earnings.