NFP: The Trend is Not Your Friend

Another month ticks by, and once again it is time for everyone’s favorite economic data point/recency effect proof, the Employment Situation Report.

Consensus is for a -524,000 job loss, out of a range of -750,000 to -450,000; On U3 Unemployment Rate – the consensus is for a jump from 7.2% to 7.5%,  with a 7.2% to 7.6% range.

This data release is the most over-analyzed, least important economic report for investors. And yet each month, we go through this charade that the we know the precise number of jobs lost and what it means to equities. Traders eagerly await the data and respond to it with enthusiasm. To many, NFP is the single most important economic data point there is.

The polite word for this is Poppycock.

Here is the statistical reality: We have a working employment force of more than 140 million people, and we do not really have the ability (yet) to measure the change in realtime of a few 100,000 people each month. Instead, we get a reasonable approximation of what those changes were, a number that is neither precise (similar results) nor accurate (closeness to reality).

But it is, as the expression goes, close enough for government work.

Regardless of the mathematical accuracy of this measurement, here is the grim reality that we do know with both precision and accuracy: We are coming off a post-recession cycle were job creation was anemic. 2002-2007 was the weakest employment creation cycle in the post-war era.

Second, we also know that the jobs number today will be bad. As we have long exhorted, any single data point is a meaningless statistic as a standalone number, and what actually matters is the overall trend of the past 12 months. That we also know with grim certainty: It has been bad for quite a long time, and is getting even worse. When it comes to jobs, the trend is not your friend.

Are we likely to see a million plus loss? Possible but doubtful.

Any chance we see a jobs gain? All but impossible.

My range is 600-700k, but if its better or worse it is irrelevant. A loss of 450k jobs is not a good thing, and a bigger loss of 850k jobs is not that much more significant to the overall economy.

In a work force of 143 million, is there a significant difference, given what we know about how the sausage BLS data gets made, between -450k and -750k?  That’s really a margin of error difference when we are discussing changes in group of this size (143 million).


One other factor is the unemployment rate. I expect U-3 to inch closer to 7.6-7.8%, but watch the U-6 level (found here: Table A-12. Alternative measures of labor underutilization). It has soared from 8.7% in December 2007 to 13.5% in December 2008. We could see a number with a 14 handle (i.e., 14.5%), eventually on its way towards 15%. Befor this recession is over, I expect U6 will be closer to 20% than 15%.

Two other factors to note:

1) We get the Revisions in the Establishment Survey Data this morning.

With the release of January 2009 data on February 6, 2009, the Current Employment Statistics survey will introduce revisions to the nonfarm payroll employment, hours, and earnings data to reflect the annual benchmark adjustments for March 2008 and updated seasonal adjustment factors.  Not seasonally adjusted data beginning with April 2007 and seasonally adjusted data beginning with January 2004 are subject to revision.

2) We also are treated to the Changes in the Household Survey :

Effective with the release of data for January 2009, revisions will be introduced into the population controls for the household survey.  These changes reflect the routine annual updating of inter-censal population estimates by the U.S. Census Bureau.

We will find out if these are significant updates or not at 8:30 am.

Print Friendly, PDF & Email

What's been said:

Discussions found on the web:

Posted Under