Navigation Path: Simon Search Home » S2M Marketing Blog » s2m Marketing Blog » Google Domestic Trends Predicts Unemployment

Google Domestic Trends Predicts Unemployment

by Geoff Simon on September 7, 2009

in Industry News

According to a post by Google’s Chief Economist , Google Domestic Trends has been used to mimic variables such as initial claims of unemployment, among other things to predict short economic trends.  Given that one of the biggest leading indicators of economic activity is number of people who file for unemployment this could be modeled and predicted using Google Domestic Trends.

According to Research by Macro-economists Robert Gordon and James Hamilton, “…in each of the last six recessions, the recovery began within 8 weeks of the peak in new unemployment claims.”

Google applied the methodology outlined in the above research to predict initial unemployment claims using past values in a time-series then applied the Google Trends data to see how much it improved the forecast.  Mr. Varian and his team found over a 15% reduction in mean absolute error in predictions one week out.  Detailed information can be found in this PDF.

unemploymentThe point of the story is that initial claims have been declining from their peak and the Google Trend data predicts further declines.  This is one Google trend everyone hopes continues.

Technorati Tags: , , , ,

  • I wonder if this prediction model has been accurate since this post was written?
  • Unemployment claims down 26,000 according to ETA Press Release this morning, good news for the economy, good news for these Google predictions.
blog comments powered by Disqus

Previous post:

Next post: