Using Google Trends to Predict a Recession

January 23, 2008

My brother had an interesting idea… why not use what people are searching for to see how the economy is doing? I did a little bit of research, and came up with the following incredibly unscientific experiment.

1) Put in various words into Google Trends associated with an economic depression, limit the area to the United States in the last twelve months, and see how many people have been searching for those words.

So what were the results?


It shouldn’t be much of a surprise to anybody who has been paying attention to the news lately… more and more people are searching for recession-related words. Let’s take a look at the specific results.

Keyword: Recession

Google Trends - Keyword: “Recession” - Last 12 Months

It’s pretty easy to see here that more and more people are thinking about a recession recently. I don’t know if it’s possible for the evidence to be any more clear. Since talking about a recession is often a self-fulfilling prophecy, since people stop spending when they feel less secure in the economy.

Keyword: “Economic Depression”

Google Trends - Keyword: “Economic Depression” - Last 12 Months

Please note that I used “economic depression” instead of “depression” since “depression” by itself tended to bring up a bunch of pages about clinical depression. Here, again, the evidence is quite clear. While for most of the last year, there weren’t even enough searches for Google to keep track of, in the last few months they have been shooting upwards.

Keyword: “Find Job”

Google Trends - Keyword: “Find Job” - Last Twelve Months

I apologize for the awkwardness of the phrase. It was difficult to come up with something in this topic that Google had enough information to track. Anyways, you can see the bump at the end of the graph. While not nearly as striking as the first two graphs, there is definitely an increased interest in finding jobs lately.

Keyword: “Foreclosed”

Google Trends - Keyword: “Foreclosed” - Last 12 Months

Another indicator of economic malaise, searches for “foreclosed” - while remaining strong all year - have definitely had an uptick recently. Interestingly, if we look at subregions, Florida is the state in the US which is searching the most on this topic. The top three cities for this search term are all in Florida.

In conclusion, the results of my absolutely unscientific, not at all back-tested, probably irrelevant experiment is that if we’re not heading into a recession, more and more people think we are. Only time will tell if they’re right!

Share this article... These icons link to social bookmarking sites where readers can share and discover new web pages.
  • Reddit
  • Digg
  • del.icio.us
  • StumbleUpon
  • Fark
  • MisterWong
  • Furl

Comments

3 Responses to “Using Google Trends to Predict a Recession”

  1. Gabe on January 28th, 2008 3:28 pm

    This is a neat article and a neat idea. I noticed though that if you step back and look at “All Years”, instead of merely the last “12 months” patterns become evident.

    For example, search “Loan” or “Bankruptcy”. I actually came away being less worried. A lot of the “spikes” seem to happen every year at nearly the same time.

    “Recession” is the exception, but it could be argued that the surge is due purely to the media, in this regard. Key indicators of a recession, for example “Jobs” or “Unemployment”, seem to be very comparable to previous years.

    I enjoyed checking this out though. Thanks for the idea!

  2. A change in general sentiment about debt marks the start of deflation | Trendinews.com - Comment, learn, and share with Trendirama authors. You can be an author too! on January 28th, 2008 6:43 pm

    […] Using Google Trends to Predict a Recession […]

  3. admin on January 29th, 2008 2:37 pm

    Gabe-
    Thanks for the extra research. I decided not to do “all years” because I figured the growth in the number of Google searches might overshadow the actual change in sentiment. However, looking over the results, they’re not as bad as I thought and you’re right - things aren’t as bad as they looked with my original data.

    Very interesting results, though, and I would like to go a bit further and backtest it if I get a chance. It’s all about finding the time to do it, though!

    Thanks again for the input!

    Bill

Got something to say?