Data from Google Trends could be useful for predicting the ups and downs of the Dow Jones Industrial Average, according to new research published in Scientific Reports.
The new information from study authors Tobias Preis, Suzy Moat and H. Eugene Stanley showed that investing based on certain finance-related trending search terms could yield much higher returns than the average portfolio held from 2004 until 2011.
The authors of the study took 98 mostly finance-related terms such as “portfolio” and “debt” and compared the search volume of the terms to the closing prices of the Dow. They found that searches for the more finance-focused terms declined before rises in the market average. When there was an uptick in terms such as “debt” and “crisis,” though, the market average was more likely to fall.
The research was a continuation of earlier studies that Preis published in 2010. At that time, he and other colleagues found that Google Trends data could not predict a stock price, even though there was a correlation between the number of Google searchers for a company’s name and the number of times its stock was bought and sold.
This time, Preis, Moat and Stanley used broader search terms like “stock market” to determine whether or not information from Google Trends was a helpful predictor.
To test their hypothesis on Google Trends, the team tried out an investment strategy based on search volume. If the volume of search terms increased, the researchers sold stock, and bought when financial search volumes fell.
“Debt” was the term most indicative of a profitable investment. A portfolio sold when the search term “debt” rose would have yielded a 326 percent return during the study, significantly higher than the 16 percent they would have earned from a broad stock market fund bought in 2004 and held until 2011.
More broad terms such as “color” or “wedding” did not yield as high returns as “debt.”
The information could indicate that traders that were more apprehensive about the market were more likely to spend more time researching its current state, said Suzy Moat, computational social scientist at University College London.
“Our best explanation at this point is that perhaps our results can be explained by the idea of loss aversion,” she told the E-Commerce Times. “Behavioral economists have shown that people are much more concerned about losing $5 than missing out on gaining $5. For traders, the decision of greatest consequence that they would make is to sell a stock at a lower price than its worth, so they’re going to spend more time researching before making that decision.”
Taking Advantage of Digital Possibilities
Google Trends is not the only online influence with the power to impact markets. Last week, a tweet from a hacked Associated Press account claimed explosions at the White House. Trading algorithms picked up on the news and sent the Dow plunging 145 points within minutes.
The market recovered quickly. However, the incident and the recent research about Google’s ability to predict trends show that brokerages and investors haven’t quite figured out how to most efficiently use their new wealth of predictive data, said Benjamin Woo, managing director of Neuralytix Inc.
“Ultimately, the difference between the next generation of technologies vs. the last generation is that we’re no longer talking about using technology to automate. Instead, we’re using technology to optimize,” he told the E-Commerce Times. “Our society does not suffer from a lack of data. It suffers from a lack of insight. So we’re applying old techniques — i.e. using summarization and survey and making conclusions based on hypotheses.”
Google is one of the leaders in making that data available, but the companies that can make the jump in optimizing that info are going to be the ones that can get ahead in industries such as finance, said Matthew Jackson, economics professor at Stanford.
“Having timely information about market psychology should be very useful in forecasting market movements, and Google Trends is making interesting use of data that is different from that available elsewhere,” he told the E-Commerce Times.
“The immense amounts of data that are becoming available have changed the way a number of businesses and markets operate,” Jackson said, “and it will be interesting to see how such information will ultimately be harnessed for the forecasting of things like trading behavior in financial markets.”