NFP Payroll: Stickin’ with the Under

Its the first Friday of the month, and you know what that means: Nonfarm Payroll data for the prior month. With the consensus expectations at a very high 250k, why wouldn’t I once again take "the Under?"

Indeed, this month presents several additional reasons to do so, beyond my mere lack of faith in the forecasts of economists:

• Layoffs for January 2006 were 12 percent higher than the January 2005 level.  Announced layoffs were above 100,000, and came from Ford, Mattel, GM, Mitsubishi, Time Warner, Electronic Arts, BF Goodrich, and others. When you are laying off, you are not likely to be hiring;

• The Birth/Death adjustment in January tends to be quite negative, as companies dissolve after years’ and quarter’s end. This statistical adjustment has become quite significant, with the Birth/Death adjustment accounting for well over a third (~37%) of all newly created jobs since the recession ended.

• New unemployment claims were low in January, and many economists use them as a significant part of their NFP projections. But as MS Howell’s Brian Reynolds notes:

"Many economic forecasters who form the consensus rely heavily on the claims data when making their guesses (and they are nothing more than guesses) about what the payroll data may look like.

This is understandable, but often wrong. For one thing, the two series measure two different things. Friday’s payroll number attempts to measure the net number of new jobs created, while the claims data only attempts to measure half of that equation, the number of people who are filing for unemployment after being let go.

In addition, the two series are calculated entirely differently, and there can be timing differences. So, while the two often correlate in the long run, there is often no correlation between them on a month-to-month basis."   (emphasis added).

As far as the markets are concerned, I would expect a large range that would not be too disruptive after yesterday’s 2 billion plus share sell off. Anything between 150k to 250k implies modest growth. A number that’s too low supports the bear argument that the economy is already starting to cool, while a number that’s too high (300k plus) might stoke fears that the Fed’s tightening cycle has much further to go.      

>

Final caveat: I have long ago stated that the BLS does not monkey around with the numbers. Sure, the model is biased to the upside (just as the CPI model is biased to the downside). But thats just the politics of governing, and both parties are guilty of this.

However, consider that over the past 4 years, the Birth Death model has subtracted an average of 308k jobs each January (2002-2005 = -239k, -391k, -321, -280).

Given this very reliable pattern, I would cry foul if the B/D number was wildly off its prior trend. Its something worth watching.