The Big Picture

« Bracing for NFP Day | Main | Job Loss Trend Dismal »

Reviewing the NFP Data

Friday, May 02, 2008 | 09:58 AM

Courtesy CNBC

Fist off, the numbers: BLS reported Payroll employment dropped 20,000, far below the consensus estimates. Unemployment fell to 5.0%.

If you believe the headline numbers are an accurate reflection of the current US employment situation, then you can stop reading this, and head over to this article. Everyone else should feel free to keep reading.

Let's consider a few items:

• Private payrolls have fallen for five straight months. Weakness in the goods-producing sector is intensifying;

• Employees working part time jobs is +306k this month to 5.2 million. This increase is either because a) Hours have been curtailed; or B)They cannot find full-time employment. Note that if your hours get cut back, you do not show up in the Unemployment Rate or any layoff data.

• As noted earlier, the Birth/Death model was a major distortion. (in several months, we will get the revisions). Lets look at how the B/D has changed from April 2007 (+262) to April 2008 (+267):

+45k construction jobs v 37k April 2007
+8k jobs were added in financial activities versus 1k last April.
+72k in professional/business services versus 48k last April.
+83k in leisure/hospitality (95k last April).

I am certain that some country on some planet in our galaxy is adding more jobs in construction and finance versus one year ago, but it ain't the USA on planet Earth, that's for sure.

Remember, the B/D adjustment is not a one for one addition, it goes into the total employment measure (not just the increase) then gets seasonally adjusted.

• The primary payroll improvement was in the service-providing sector. Ian Shepherdson of High Frequency Economics suggests this is noise, as the the seasonal factor was 81,000 bigger than in April 07. (Easter seasonal problems?)

• The overall trend is increasing weakness in job creation.

UPDATE: May 2, 2008 10:38

A brief explanation of the B/D adjustment: The changes in the Birth Death model were designed to capture new job creation that BLS was missing at the early stages of a recovery. However, the improvements in accuracy at that part of the cycle seem to be wildly offset by a decrease in accuracy in the latter parts of the cycle -- i.e., early turns in employment pre-recession.

Sources:
Employment Situation Summary
THE EMPLOYMENT SITUATION: APRIL 2008
http://www.bls.gov/news.release/empsit.nr0.htm
Download April_2008.pdf

CES Net Birth/Death Model
http://www.bls.gov/web/cesbd.htm