The December 2022 issue of IEEE Spectrum is here!

Close bar
Google Searches About Politics Predict the Stock Market
Illustration: Randi Klett; Images: iStockphoto

The number of Google searches related to business and politics can help predict falls in the stock market, researchers at the University of Warwick, in England, say.

Scientists have recently begun investigating what people look for on Google and Wikipedia to help forecast the future. For instance, prior research has shown the rate at which people look up information about the flu helps predict the spread of the disease

In recent work, "we found evidence that data on Google searches for financially related words and views of financially related pages on Wikipedia could have provided early warning signs of stock market moves," says Suzy Moat, a data scientist at the University of Warwick. "However, the financial markets constitute a large, complex system, which influences and is influenced by many different aspects of modern society. We therefore wondered if searches for other topics might also provide insight into subsequent stock market moves."

First, the researchers used a computer program to scan all the articles in the English version of Wikipedia and identify 55 lists of words that frequently appeared together — for instance, the word "debt" was more often found in articles relating to finance than fruit. They then used the online service Amazon Mechanical Turk  to label which topics—such as medicine, entertainment, or travel— each of these list of words represented. (Mechanical Turk is actually a group of humans who are paid to perform specific tasks.) Next, they used Google Trends to investigate how frequently Google users searched for the 30 most common words of each topic from 2004 to 2012. Finally, they compared this data with stock market movements during the same period.

Predicting stock movements a week in advance using this strategy may no longer be possible.

The scientists found increases in Google searches related to business were not the only ones that preceded stock market stumbles a week in advance: "We also found a similar relationship for groups of words relating to politics," says Chester Curme, a data scientist at Boston University who collaborated with Moat. They suggest rises in searches for politics and business may constitute early signs of concern about the state of society or the economy, which might lead to a reduction of confidence in the value of stocks.

Curme adds that their strategy "readily extends to languages other than English, so that similar analyses could be carried out in different countries and markets." However, the researchers did note the relationships between searches and the stock market have weakened in recent years, perhaps reflecting the rising incorporation of Web data into automated trading strategies and suggesting that predicting stock movements a week in advance using this strategy "may no longer be possible," Curme says. "We aren't pitching a 'get rich quick' scheme."

The scientists detailed their findings online this week in the journal Proceedings of the National Academy of Sciences.

The Conversation (0)

Why Functional Programming Should Be the Future of Software Development

It’s hard to learn, but your code will produce fewer nasty surprises

11 min read
A plate of spaghetti made from code
Shira Inbar

You’d expectthe longest and most costly phase in the lifecycle of a software product to be the initial development of the system, when all those great features are first imagined and then created. In fact, the hardest part comes later, during the maintenance phase. That’s when programmers pay the price for the shortcuts they took during development.

So why did they take shortcuts? Maybe they didn’t realize that they were cutting any corners. Only when their code was deployed and exercised by a lot of users did its hidden flaws come to light. And maybe the developers were rushed. Time-to-market pressures would almost guarantee that their software will contain more bugs than it would otherwise.

Keep Reading ↓Show less