The Big Picture

« The Value(less) U.S. Dollar | Main | Saturday Afternoon Chart-Porn »

NFP WTF ?

Saturday, October 06, 2007 | 08:22 AM

in Data Analysis | Economy | Employment | Financial Press | Psychology/Sentiment

As mentioned yesterday, thanks to a dental crown install, I missed much of the fun (at least I had a steady supply of nitrous while in the chair). I saw the data, posted on it, and ran out the door.

Let's get into the details of what was by most traditional measures, a mediocre NFP report:

• Private sector employment growth was stagnant.

• Health care, and bars & restaurants were responsible for more than half the rise.

• Temp jobs, usually a good leading indicator of future job creation, fell 20,000.

• Government was responsible for nearly all of the upward revisions to the July and August;

• Manufacturing continues to lose lots of jobs quickly. (Negative 60,000 over the past two months).

• Retail hiring is trending downwards;

Having been out of the loop all day, I am somewhat perplexed over the subsequent reports about this NFP report. I cannot say I understand why people are outright gushing over this otherwise middling number:

The Jobs Machine (WSJ)

Job Growth Looks Rosier, Easing Recession Fears (NYT)

Jobs: A September Shocker (BusinessWeek)

Talk about the soft prejudice of low expectations: These headlines are over a mere 110,000 jobs. There was obviously relief that this wasn't even a worse number. But understand the details: This is in no stretch of the imagination a solid month of job creation. This was weak NFP growth, failing even to keep up with population growth.

To review: Any report under approximately ~150k month (subject to revisions) is weak. It means that job creation is failing to keep up with population growth. Thanks to the shrinking Labor pool, this has not had much of an impact on the Unemployment Rate, which ticked up a mere 0.1% yesterday.

Let's put these data points into some context average monthly job creation:

- In 2006 was 226,000 new jobs created per month
- In 2007, that number fell to 122,000;
- In Q3 2007, that number fell to 74,000.

Even more damning has been the Birth/Death adjustments. In the 1990s, these were relatively insignificant. In 2001-02, it was less than 10% of total NFP each year. After a 2001 change to the format and weighting (effective in 2003), the B/D contribution to NFP went up dramatically. From 2003-06, Birth Death adjustments contributed between 35-41% of the total non-farm payroll job creation.

Then there is 2007. If we take this year's B/D adjustment versus NFP, we get a B/D contribution that is in excess of 76%. (Trailing 12 months B/D versus 2007 NFP data annualized).

Strong? Shocker? Jobs machine?

Hardly . . .

UPDATE: October 7, 2007 7:32am

I (belatedly) see the NYT's Floyd Norris reached the same conclusion in yesterday's Times:

"JOB growth in the United States has slowed significantly over the last year, a fact that renews questions about the ability of the overall economy to withstand a slowdown in housing . . .

The job figures each month are based on a survey of employers, but the numbers released are not exactly what the survey shows. Instead, the Labor Department adds or subtracts jobs, based on estimates of how many people found work with newly created companies, less those who lost jobs at companies that went out of business. The figures come from what the department calls its birth-death model.

The figures are eventually corrected based on data on the number of people covered by unemployment insurance, which is more accurate but available on a delayed basis.

At turning points in the economy, the birth-death model figures can be considerably off. If the economy is strengthening, as it was in recent years, the result can be an underestimation of job growth. Indeed, a year ago the government said it needed to add 810,000 jobs for the year ended in March 2006. But over the following year, the estimates were too generous.

Over the last 12 months, most of the reported job growth has come from the birth-death model, not from the actual survey of employers. If the model overestimated employment by new companies, then the figures for that period are likely to be revised down a year from now. . . ."

Monthly Data Aside, Trend Is Weak
FLOYD NORRIS
NYT, October 6, 2007
http://www.nytimes.com/2007/10/06/business/06data.html

The video below is an example of a somewhat gushing review. (I am obviously not seeing what everyone else is)

Sources:

charts courtesy of CHRIS PUPLAVA

Historical Net Birth/Death Adjustments
BLS, Department of Labor Statistics
http://www.bls.gov/ces/cesbdhst.htm

CES Net Birth/Death Model
BLS, Department of Labor Statistics
http://www.bls.gov/web/cesbd.htm

The Jobs Machine
WSJ, October 6, 2007; Page A20
http://online.wsj.com/article/SB119162827290050813.html

Job Growth Looks Rosier, Easing Recession Fears
MICHAEL M. GRYNBAUM
NYT, October 6, 2007
http://www.nytimes.com/2007/10/06/business/06econ.html

Jobs: A September Shocker
Michael Englund
BusinessWeek, October 6, 2007
http://www.businessweek.com/investor/content/oct2007/pi2007105_921337.htm?

U.S. Economy Down, Not Out
Overall Health Is Seen In Face of Deceleration;
Job Gains Lift Optimism
KELLY EVANS
WSJ, October 6, 2007; Page A3
http://online.wsj.com/article/SB119158634981050085.html