Can Google Predict Economic Trends?

The promise of big data is that it will help us reveal secrets we might not otherwise be able to discern from our own experience or information sources. But more specifically, it's allure is in predicting future behavior or trends rather than simply combing over the past.

Some research has revealed that Twitter and other social media postings might be revelatory. There is still some skepticism about this due, in part, to lingering suspicions about the value of social media but also to the usual array of methodological and data validity issues.

Google, as a larger, older source of data wealth is another obvious candidate for this sort of analysis. There are fewer concerns because the sample size is larger and potentially suffers from fewer of the biases potentially inherent in social.

This potential may have been realized, as the following article explains, when Google search trends for the several months prior to the recession are examined. Had policy makers had access to the data and were willing to act on it, counter-cyclical stimuli might have been implemented to offset some of the subsequent recessionary impact.

There are two problems with this view: the first is that having the data is only the first step; convincing those in a position to believe it is another, and far more daunting, task. Secondly, this assumes that the machinery of government will move expeditiously enough to generate a measurable counterbalance. Given the recent experience in the US and Europe, both seem somewhat quixotic: people are not inclined to believe that which they do not want to or that which flies in the face of their own self-interest. And disagreements about policy are often more emotional than rational. That said, it is becoming apparent that humans have more predictive information than ever before. Whether they choose to use it wisely is another matter. JL

Karen Johnson reports in the Wall Street Journal:

What did Google know ahead of the global economic recession, and when did it know it?
A lot, a new Canadian research report suggests, and earlier than most recession forecasters.
The world’s dominant search engine could have predicted the global recession up to three months before its onset, based on the rising use of the search terms “recession” and “jobs,” according to Greg Tkacz, associate professor and economics chairman at St. Francis Xavier University in Antigonish, Nova Scotia.
Many of the world’s official economic indicators are released with a time delay or at infrequent intervals, or require significant revisions after their release, Mr. Tkacz said in a report released by the C.D. Howe Institute, a Toronto-based think tank.  But Google Trends search data offer more timely glimpses into the economic well-being of hundreds of millions of people, he said.
Google search-trend data, available free from the search giant, are aggregated on a weekly basis, making them well-suited for speedy analysis, Mr. Tkacz told Canada Real Time. And that matters. The sooner policymakers know about a recession, the sooner they can implement counter-cyclical policies to dampen its impact and, potentially, hasten a recovery.
Mr. Tkacz said that could have shortened Canada’s most recent recession, which began in the third quarter of 2008, and ended in the third quarter of 2009. Economists define a recession as two or more consecutive quarters of negative economic growth.
In October and November 2008, Canadian policymakers, not yet knowing that the quarter’s gross domestic product would show a contraction, said they believed the country would avert a recession, Mr. Tkacz said. That November, the Canadian government issued a fall economic update which contained no new stimulus measures.
“Had they had a better measure of the pulse of the economy in August or September of 2008, they could have readied the kinds of stimulus measures needed to help offset the negative impact of the recession that was in the pipeline,” Mr. Tkacz said.
In January 2009, the Canadian government, saying, “these are extraordinary times for the global economy and they call for extraordinary global measures,” released a recession-targeting budget that included 40 billion Canadian dollars ($38.2 billion) in stimulus spending and C$20 billion in personal income tax cuts.
To be sure, Google search trends are far-from-perfect reflections of an economy. The records, which extend back only to 2004, cover just one major recession, and that limits the scope for analysis. And search terms can mislead–as they did in October  2011, when searches for the word “jobs” spiked in the wake of Apple Inc. co-founder and chief executive Steve Jobs’ death, Mr. Tkacz said. But Google trends nonetheless “can potentially quantify what had previously been non-quantifiable,” he said, such as consumer confidence sentiment, another telling macroeconomic factor, and can be broken down by country, region or city.
“This is kind of like looking at economic data under a microscope, which is something we have always dreamed of doing in economics,” Mr. Tkacz said.