Analyzing Google search trends could predict recessions faster: study

TORONTO – A new report suggests that analyzing Google search trends could predict a recession more quickly than traditional GDP data, which are available only after a two-month lag.

The report by the C.D. Howe Institute says this kind of real-time data mining could allow policymakers to respond more quickly to weakening economic conditions.

Report author Greg Tkacz says when many people plug the same economic terms into Google, it can provide a clue about changing conditions, such as an impending recession.

Following this logic, Tkacz analyzed the usage of the search terms “recession” and “jobs” prior to the 2008-2009 economic downturn.

He found that analyzing search trends could have predicted the recession three months prior to its onset.

However, Tkacz cautions that since Google search trends are only available as far back as 2004 — a short timespan as far as business cycles go — the study should be viewed as illustrative rather than conclusive.