From the thought process that brought us “Can Wikipedia predict the next market crash?” comes a second question about the internet’s predictive powers: “Can Google predict recessions?”
Just as researchers at the Warwick Business School suggest monitoring how often people viewed certain Wiki pages as a way of predicting a coming economic downturn, the C.D. Howe Institute is experimenting with tracking Google searches and the growth of electronic payments variables (such as debit and credit card transactions) as a way of foretelling recessions.
The result? We could have predicted the 2008 2009 recession. At least that’s the argument Greg Tkacz makes in his report for C.D. Howe, Predicting Recessions in Real-Time: Mining Google Trends and Electronic Payments Data for Clues.
“New sources of electronically recorded data are both timely and reflect the real-time intentions of millions (or billions) of agents,” states Tkacz. “Most policymakers and economists failed to predict the last recession because of the lag in traditional economic data. I look at the search data in Google Trends, as well as electronic payments data, to see if the recession signals were there in real time. ”
Many official economic indicators are released with a time lag, released infrequently and often require revision, Tkacz explains. Not too long ago, Canadian empirical macroeconomic researchers would have to wait two months for the release of the monthly National Accounts in order to update their models and forecasts. But in just the last 10-to-20 years, technological innovations have resulted in vast amounts of other data being recorded electronically and stored. New data series are now being generated at a rate faster than analysts can study them.
Google is the search engine of choice for most, which means it’s a goldmine of data researchers can use to map trends based on group interests in issues from ecomics to health to politics. According to Tkacz, tracking the search terms “recession” and “jobs” could have predicted the last recession up to three months in advance of its onset.
Because Google query data are only available from 2004 (a short timespan in the context of business cycles), Tkacz says his study’s predictions aren’t necessarily usable; rather, they illustrate the potential uses of electronic data.
Until there is enough data to test Tkacz’s theory, we continue to rely on economists for their predictions.