Admiral Janeway’s new U3

Originally published at The Benfell Blog. You can comment here or there.

This is a very different version of this chart than you may have seen before. It is based on different underlying data.

When I went to do my usual thing with the unemployment data this morning, I found that the data file the Bureau of Labor Statistics has made available in the past was no longer there. Instead, casting around, I found monthly data going back to 1948 in multiple forms.1

So things are going to be different—and I think better—from here on out. But it remains important to be intelligent with this data, so here’s what I’ve done.

  • I am not using seasonally adjusted data. When landlords start “seasonally adjusting” their rents, when bankers start “seasonally adjusting” their mortgates, when people start “seasonally adjusting” their needs for food, and when bill collectors start “seasonally adjusting” their bills, I will consider accepting “seasonal adjustments” as legitimate. Until then, people who need work still need work, and “seasonal adjustments” belong on the bookshelf along side Alice in Wonderland. This is one factor making this chart considerably messier than in the past.

  • I am using monthly data that the BLS has made more easily available. In combination with the lack of seasonal adjustment and with the compression of time since 1948 across the chart, this makes rates appears as more of a band than a line. I have a sneaking hunch this is closer to the truth.

  • My calculation methods remain the same, but because I’m using non-seasonally adjusted data, my calculation of the BLS U3 comes to 9.79 percent rather than the more widely reported 9.0 percent.2

  • As before, I assume that given a reasonable opportunity, the highest labor force participation rate to date is a more honest reflection of the proportion of the population that is available to work. In order to preserve precision (the BLS rounds the number off at three significant digits), I recalculate this from the supplied population and labor force sizes. I use this number in calculating Admiral Janeway’s U3 (named for my cat, who is in turn facetiously named for the Star Trek character) and in the unemployment rate that counts people who “want a job now” but whom the BLS excludes from the labor force. When conditions improve for the middle and working classes such that they no longer have as great a need to work, this assumption will deserve reconsideration. But as of now, we’re still headed in the opposite direction.3

  • Also as before, these statistics do not consider people who are working fewer hours than they need or would prefer. Classically this means part-time workers who want or need full-time work. Nor does it consider people who are working at jobs below their skill levels because they have not found work for which their qualifications are appropriate. Notably, Gallup currently shows an underemployment rate of 19.2 percent (and an unemployment rate of 9.9 percent).4

  • I also found a statistic the BLS has apparently been collecting since 1994 of people it excludes from the labor force but who “want a job now.” I add it to the number of people the BLS counts as unemployed (and in the labor force) and divide it by the labor force size I calculate as described above. The relationship between this measure and others is best illustrated in a version of the above chart that covers the period since 1994:

  • Finally, I do not know what adjustments the BLS may have made to its supposedly “unadjusted” data. But not all adjustments are seasonal adjustments and given the dichotomy with “seasonally adjusted” data, it is unclear to me which side these other adjustments fall—I distrust them all. And when economists try to claim that they’ve been through this, that economists have reached a consensus on these adjustments,5 it is important to remember that economics is not only the “dismal science” but is also in a contest with political science for recognition as the most ideological “science.”

With all this, Admiral Janeway’s U3, at 15.34 percent, indicates by far the worst unemployment picture since 1948. The previous peak was in January 2010, at 15.16 percent.

I am disappointed by the “want a job now” statistic (which counts people who want a job but whom BLS excludes from the labor force), which I add to the number of people the BLS counts as unemployed. Until November 2003, it was consistently higher than Admiral Janeway’s U3. Since then, the record is more mixed. It still does occasionally exceed Admiral Janeway’s U3, most recently in July 2009. But even it shows a high proportion, 13.28 percent, of people who want work. But it was higher in January (13.80 percent) and February (13.68 percent) 2010.

As for the BLS U3, it is worth considering that its labor force participation rate is sinking towards levels last seen in early 1984. At 63.90 percent, it falls between the levels of April (63.73 percent) and May (64.30 percent) 1984. The trough in 1984 was in January at 63.25 percent; 1983 saw levels that were worse. A low labor force participation rate is a major factor holding down the headline U3.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.