The Unemployment Severity Index

The BLS collects and publishes a massive data collection.  No single researcher can appreciate all of it, but there is something there for nearly everyone.

There is a problem that can come from too much data.  Journalists must make choices about what to report.  How does one determine the real story?

Background:  Measures of Unemployment

Last week we applauded the effort of Barry Ritholtz at The Big Picture, one of our featured sources, to broaden media coverage of unemployment data.  Barry sought to give a lot of publicity to the broadest possible measure of unemployment, a rate much higher than U3, the traditional measure.

We agreed in concept, and proposed that he add his strong voice in proposing that the media also report the BLS measure of those unemployed for more than fifteen weeks.  This measure, while lower in absolute terms, has (like the other measures) moved higher during the current economic weakness.

We were disappointed that Barry did not embrace this proposal. He clarified his position in comments on our post, suggesting as follows:

Does U1 help reach that goal? I guess it does — if you believe that
the U3 measure of employment is OVERSTATING unemployment (which I

To begin with, we realize that the distinctions involved make the eyes of the average reader glaze over.  That makes our task more difficult.  It also makes the reward greater for the (very) few readers who grasp the entire issue.

So we should all understand that Barry uses his own conception of the state of national unemployment as a starting point, and then wants the media to choose a measure which conforms.  He is concerned that the media reports have not grasped the full extent of current economic distress.

We wonder?  Really?  If anything, the general public and the media have overplayed economic problems for several years.  The percentage of the public perceiving negative economic conditions or recession has significantly exceeded the ratio of professional economists.  The bad economy story is "man bites dog."

At "A Dash" we recommend something quite different:  Use data to form your conclusions.  Be willing to change your opinions as you get new data.  It is an old-fashioned idea, but it works.

(We are also disappointed that he did not respond in his comment to our point that he had made a flat-out mistake in asserting that the BLS had changed the definition of the U3 series.)

A Fresh Approach to the Unemployment Question

We were delighted to hear from an old friend and colleague, Marty Finkler.  Marty has the highest and purest of motives — teaching young people what they really need to know, without bias or preconceived conclusions.  We listen to him carefully, and so should you.

Marty wrote to comment on our unemployment rate article.  First, he agreed with the idea of informing the public about the wide range of measures.  He also highlighted an approach, unemployment severity, which he attributes to a decades-old book by Charles Baird.  (Readers please inform us about good citations.  We know that the ECRI considers unemployment severity).

The unemployment severity index is calculated as follows:

One takes the standard U3 measure of unemployment, multiplies by the mean duration of unemployment (in weeks) and multiplies by 5 to convert to days.  The result is the average number of days someone in the labor force is unemployed during a given year.  Marty's chart, based upon BLS data, is shown below.

Labor hardship
One can readily see that by this measure, the unemployment severity is not at an unusually high level.  While it reflects the economic weakness shown by other data, it is nothing like the peaks from prior times of economic distress.


We thank Marty for his observations.  He has shared some other ideas, so readers can look forward to more information from a great source.

Secondarily, but more important to some, is the idea that one can profit from understanding data.  If many are engaged in making things seem worse than they really are —the conclusion is easy.

Opportunities abound.

You may also like