Dave Altig’s December 5 blog entry, “The recession in pictorial context,” elicited an array of interesting commentary. In this recessionary period, the questions that arose are relevant to today’s current economic state and it is worthwhile taking a blog to explore these issues in greater depth.
Comments
I might suggest that considering the 1981–82 recession by itself, rather than in combination with its 1980 little brother, somewhat understates the "badness" of that period.
The 81–82 recession graphs include data from 12 months before the first day of the recession to 12 months after the last day of the recession, therefore the dates range from 7/1/1980 to 11/1/1983. The 1980 recession ended in July of 1980, so these time periods overlap each other by one month. Including this month does not “understate the ‘badness’ of the period” because the calculations are based solely on the recessionary period being examined. In other words, the time span has no effect in this case. The first day of the recession is set to one. The months before and after are normalized to the first day of the recession. If we were to remove the 1980 recession date and run the time span 10 months before the 1981 recession to 10 months after, the graph does not change in anyway apart from a shorter time span. This is because the data is scaled to 7/1/1981, the beginning of the recession. You can see this by looking at the 12-month and 10-month graphs, below.
Are the unemployment rate comparisons U-3 across the various recessions? If so, I would be keen to see a U-6 comparison as well, I suspect it would be...illuminating.
The graphs used were U-3 comparisons. To clarify, the U-3 employment rate is the official employment rate used by the Bureau of Labor Statistics and is the most commonly used measure. The employment measures range from U-1 to U-6, U-1 being the most restrictive and U-6, the most broad. The U-3 rate is defined as: “Total unemployed as a percentage of the civilian labor force” (BLS), while the U-6 is defined as “Total unemployed plus all marginally attached workers, plus total employed part time for economic reasons as a percentage of the civilian labor force plus all marginally attached workers” (BLS). U-6 data only ranges back to 1994, therefore only the 2001 recessionary period can be examined.
The following graphs are U-6 comparisons—the first graph compares the two unemployment measures, and the second is comparing the 2001 recession and the current levels (normalized to 3/1/2001, the first day of the recession).
Are these apples to apples comparisons? Are the numbers from the prior recessions the numbers that were reported in "real time" or are they "final/revised" numbers?
The unemployment rate does not go through revisions. It is, however, benchmarked annually in March. Payroll employment is revised as additional information becomes available. All of the payroll employment graphs in the December 5 posting contained revised data with the exception of November. To see how these numbers are revised over time, here’s a comparison of these same graphs with real time (unrevised) data and the revised numbers (from December 5 blog).
By Courtney Nosal, economic research analyst in the Atlanta Fed’s research department