Sentiment in the Fed Beige Book Plummets
We are starting to get reports on what the U.S. economy is looking like after the shutdowns to battle the COVID-19 pandemic. One source is the Federal Reserve's Beige Book. The Beige Book collects narrative information about business about conditions around the country. In April the report was not pretty.
The chart above shows a sentiment score for each Beige Book Report from March 2008 to 2020. The score was computed by assigning each word in each report to positive,negative, or neutral sentiment. Then I simply counted up the number of positive minus the number of negative and divided by the number of negatives. A score of 0 would have exactly half positive and half negative words.
I have been running this analysis for each report and in recent years this score has been positive through March 2020. It plummeted in April.
You can also see the negative sentiment by counting up the times words like "strong" show up versus words like "weak". Again, in recent months strong has far outweighed "weak" as businesses report things like "strong demand". That has changed in April.
If we want to know a bit more about what is in the report, but would like to not read it ourselves. We can use the term frequency and inverse document frequency analysis to gauge what terms are important in each report. The td-idf statistic will decrease the weight on very common words and increase the weight on words that only appear in a few documents. In essence, we extract what’s special about each report.
The highest tf-idf terms in 2020 April report relative to previous Aprils 2008-2019:
- covid
- pandemic
- cornavirus
- essential
- outbreak
- shut
- nonessential
- furloughed
- plummeted
- distancing
In chart form.
A couple years back I shared R code to do this analysis and make charts like these. The charts above followed the analysis laid out in this post: https://lenkiefer.com/2018/07/29/beige-ian-statistics/