Thursday, December 4, 2008

Today's initial unemployment claims report not totally unexpected!

From, Jobless claims in surprise decline on December 4:
The number of Americans filing new unemployment insurance claims decreased last week, but the number of people continuing to collect benefits hit a 26-year high.

The Labor Department reported Thursday that initial filings for state jobless benefits decreased by 21,000 to 509,000 for the week ended Nov. 29. That was down from a revised 530,000 claims in the prior week.

Economists were expecting jobless claims to increase to 540,000, according to a consensus compiled by
Well obviously nobody believed the Gallup Poll, This Week’s Jobless Claims Report May Surprise, written on December 2:
Contrary to the consensus estimate of economists' expectations, Gallup's hiring measure shows U.S. employees' perceptions of the job situation at their companies improved a little last week. This suggests that Thursday's Labor Department report of first-time unemployment claims for the week ending Nov. 29 may well have increased less than the 529,000 reported last week -- in the opposite direction of the consensus estimate increase to 540,000.
One week's worth of data certainly doesn't indicate a trend in initial claims. It is better to look at the four-week moving average, which is high but not 1982 high. This is an extremely weak labor market that will certainly put downward pressure on labor income, further restricting private consumption and aggregate growth.

To beat the 1982 record, the four-week moving average (currently 524,500) must exceed 674,250. Unfortunately, with the following headlines from the same website (CNNMoney) and all published within one day of each other, this labor market is likely to rival 1982:

United plans to furlough 1,088 workers
Credit Suisse cuts 5,300 jobs worldwide
Job cuts mount as year-end nears

Initial claims fell this week, but it is very unlikely that they aims have reached a peak. Claims will rise further.

Rebecca Wilder


  1. Rebecca,

    CR has a post today with the 4-wk moving average of initial claims as a % of covered employment. It looks similar to yours, but the 70's/80's peak level of % claims is roughly TWICE that of today.

    So by saying this labor market is likely to rival 1982 (in terms of fundamentals), the implication is that jobless claims will rise to over 900k. Ironically, no one is expecting this because we seem to be anchored to absolute (rather than %) historical claims levels.

    By the same token, if unemployment grows by four percentage points to reach the '82 peak, we would experience a loss of about 6m jobs. If this peak were reached in a year's time, it implies a 500k reduction in employment per month. Again, no one seems to be expecting these types of numbers, yet they are easy to "back out" if you believe that this recession will be worse than the 1982 recession.

    Finally, I would note that if the above is true, much more of the job loss will come from a surprisingly cyclical service sector. This jibes with the weak ISM services number we got this week, and the even weaker spread between ISM and ISM services (now negative)

  2. In response to david pearson, I'd like to add that the redefinitions on the unemployment survey in 1994 don't make them directly comparable.

  3. Irrational,

    You are right on the unemployment rate. For once I'd like to see an apples-to-apples chart of this series. Too bad when BLS re-jiggers things they don't continue to publish the old data so as to keep the series going.

  4. Did you all see this post:

    You are right, one must compare the size of the labor force to the level of unemployed. With the size of today's labor force, a 1982-size labor contraction implies 8.3% unemployment.

    People always state in levels, but one must account for the size of the labor force (the statistic that CR quoted is available from the DoL and called the insured unemployment rate at:



Note: Only a member of this blog may post a comment.