Are We What We Search?

Ancient Greece had the Oracle at Delphi. The Shang dynasty had oracle bones. Contemporary America has Google.

Earlier this month, Lawrence Summers, President Obama’s top economic adviser, unveiled a new class of tea leaf to gauge the direction of the American economy: Google searches. The number of queries for “Great Depression,” which surged earlier in the year, had declined sharply, Mr. Summers noted. Economic anxiety is abating. The economy is probably turning the corner.

It was not the first time Google was invoked to show us the way. The company has a tool to track the path of the flu virus by looking at geographic trends in Internet queries for related terms. A study by Google researchers suggested search patterns could be used to track everything from home sales to the popularity of tourist destinations, and add to the accuracy of forecasts for new-home starts and car sales.

Polling has long been used by marketers and politicians to detect shifts in fashion and policy. Surveys of consumer confidence play a major role in economic forecasting. Economists have also tried to marshal collective wisdom with “prediction markets,” where people’s bets on all sorts of things are distilled into a remarkably accurate forecast. (Punters on, by the way, think there is a 75 percent chance that the economy will grow in the third quarter, but only a 12 percent chance that it grew in the second.)

Still, there is something new and mysterious about the high-tech oracle. Polls are a useful indicator of people’s tastes because they ask people what they like. Prediction markets require players to forecast. Google searches offer a roundabout impression of the world: what people want, what they fear, what they expect. It’s harder to tell what they mean.

What was going on when searches for terms related to suicide jumped fivefold last November? Was it the recession? Were people looking for techniques? Maybe it was the proximity of Christmas.

The number of searches for terms about depression in the economics category peaked in February, but so did searches about recovery. Anxiety about a Great Depression has abated since then, as Mr. Summers suggested. So has interest in the term “dead cat bounce.” But in June and July people became very interested in a “double dip” recession.

For all the uncertainty about the meaning of Google’s statistics, using them to help make the case that the economy is turning the corner might be a shrewd move. We might not know exactly what they signify, but they can provide any economic forecast with a populist cover.

Eduardo Porter, New York Times


Full article: