Adjusted Jobless Claims Suggest Recession Has Ended

The September employment report was released Friday, and the graph above of Initial Jobless Claims as a Percent of the Labor Force (1974-2009) has been updated to reflect the September labor force of 154,006,000 and the September average for initial unemployment claims (559,625 for the 4-week moving weekly average). This measure of initial jobless claims, adjusted for the size of the U.S. labor force, shows that jobless claims peaked during this recession above the levels of the last two recessions (1990-1991 and 2001), but were never anywhere close to the levels of the previous three recessions in the mid-1970s and early 1980s (see chart above).

In other words, this recession was worse than the last two, but not nearly as severe as the previous three, using this adjusted measure of jobless claims. Additionally, the sharp .059% reduction in adjusted jobless claims from the March 2009 high of 0.4226% to 0.3634% in September follows the same pattern of 0.05%-0.06% reductions in adjusted claims at the end of the 2001 recession (a .052% reduction from .3318% in October 2001 to February 2002) and at the end of the 1990-1991 recession (a .058% reduction from .3915% in March 1991 to .3327% in July 1991).

Finally, the current level of 0.3634% for jobless claims as a share of the September labor force is above the level at the end of the 2001 recession, but is very close to the levels that marked the ends of the four previous recessions (see dashed blue line in graph above).

See a very similar analysis here from Scott Grannis, who alternatively calculates jobless claims as a percent of payrolls with the exact same graphical pattern presented here using jobless claims as a percent of the labor force (slightly different denominator, but same numerator, and same story).

Comments

[Nick36:]

Whether the unemployment is still increasing or not depends on the number of people getting hired and not just on the number of people getting fired. And that's something your graph doesn't show.

If there are more people finding jobs than loosing jobs. Then indeed the recession might be over, even with 500,000 initial claims every week.

But if the balance is tilted the other way, and more people are loosing jobs than finding jobs. Then the recession most likely is continuing, at least as far as the working people and the consumers are concerned. And 70% of the US economy depends on US consumer spending.

If you look at correlations in numbers without sound logical reasoning. Then you can find all kinds of meaningless correlations and trends that can easily deceive you into believing something that isn't true.

Oct 04 08:45 AM

[Jim Hawthorne:]

Dr. Spin! Please!
The deeper one examines the employment picture, the worse it gets!
(See: www.ritholtz.com/blog/.../ )
There certainly is meaningful debate ongoing as to whether or not the recession has ended. But to attempt to use the Initial Jobless Claims as support for the recession's end at this point is truly silly!
Just we can't have an 'earning-less' recovery, so too is a 'job-less' recovery and "Alice in Wonderland" fantasy!

Oct 04 08:55 AM

[Jan Paul:]

Thanks to Karl Denninger, digging in the U.S. Gov. employment numbers, he uncovered an interesting statistic. First, here is the link to the site he found the statistic in.

www.bls.gov/news.relea...

Go to page 11 (PDF file) and the "HOUSEHOLD DATA Table A-1. Employment status of the civilian population by sex and age"

Now drop down to "Employed" For Aug and Sept numbers. You will find that the "unadjusted numbers equate to a drop of about 995,000 employed ( Aug. 140,074,000 Sept. 139,079,000), not the 263,000 that the "seasonally adjusted" number (further to the right on that line) that was reported.

Here is how they define "seasonally adjusted."

quote
Most seasonally adjusted series are independently
adjusted in both the household and establishment surveys.
However, the adjusted series for many major estimates, such
as total payroll employment, employment in most
supersectors, total employment, and unemployment are
computed by aggregating independently adjusted component
series.
======================

Now, dropped out, retired, drawing SSI could all be reasons the number employed might not match unemployed but, that is still a very big number.

Or it could be we just get a "public number" that is suspect.

Oct 04 10:28 AM

[Old Trader:]

Unfortunately, imho, we're a ways from being out of the woods. Jan Paul makes good points above.

Oct 04 12:21 PM

[jerrydd:]

Notice 4 of the 6 peaks were oil shocks!

When will we learn we need to be independent of imported oil in 5 yrs.

For investing oil will probably drop some until probably spring as it will go up when the world economy to $150/bbl or whatever it takes for us to go back into recession.

One solution would be build some composite thus lightweight, aero 3wheel, 2 front, car powered either by gas or electric, would get 75-125mpg or 3x's that electric equivalent. There really is the tech there if anyone would think outside the box. These can cost so little that driving them is cheaper than just the gas in a car will.

Another is switch semi's, other trucks to NG Converting would save enough in fuel, now at about $.50/gal gasoline equivalent. Had I had a truck shipping line I would convert to NG a long time ago. Buying new ones cost no more, even less because the engines are far cheaper than diesels.

And remember we 'corrected' the employment number since the early yrs of this graph. And real present unemployment, U6 is 17% now which is a more realistic and how we use to before the 70's.

Oct 04 06:47 PM

[YaddaMinski:]

Only companies are not hiring. So if 200,000 are fired each month now and new workers are being added every month, unemployment on the U-3 measure will probably hit 11%. And want if this drags demand down to cause a double dip?

Oct 04 08:19 PM