David Tuckett and many others write,
The prototype createsv what is called a relative sentiment shift (RSS) time series. This calculates
changes, in any text database across time, in the number of words related to the category of excitement relative
to the number of words in the category of anxiety, adjusting for the number of words in the articles, etc.
Results suggest that this form of analysis has strong potential for improving our understanding of what is happening in the economy and where policy action might be required.
One digital data source is the Reuters News Archive, which spans 1996 to 2013 and contains over 14 million text documents. Figure 2 shows a relative sentiment shift time series generated from all texts originating in the United States (dashed curve) plotted against US GDP (solid curve). The sharp drop in GDP in the recession of 2008-9 is evident. It is equally evident that relative sentiment series begins to fall well in advance of the decline in GDP.
You might want to read the whole article, which treats a number of methodological issues, not always in ways that I agree. My thoughts on this particular example:
1. For some reason, I am reminded of the way that animals can sense bad weather coming.
2. Perhaps it is not surprising the the participants in the economy sense that things are bad before the economic statistics reflect that. And note that policy makers receive economics statistics with a lag.
3. This sort of analysis does not tell us anything about what is causing anxiety to rise relative to excitement, or what to do about it.
In a sense this isn’t that new. There was a lot of talk of irrational exuberance, plenty of discussion of the housing bubble, and now a decent recognition of stock valuations, but you aren’t allowed to find your chair until the music actually stops.
I can remember exactly when I noticed that the economic downturn wasn’t as catastrophic in my neighborhood as had been assumed: first week in May 2009, I couldn’t find a parking spot at the post office at 10pm because of all the new bars that had just opened.