Following the release of today’s unemployment report, a bunch of folks jumped up and down about the Birth-Death model — a way of adjusting for the fact that new businesses start and stop and leave the survey.
David Rosenberg even sent out a blast, saying:
The headline nonfarm payroll figure came in above expectations at -345,000 in May — the consensus was looking for something closer to -525,000. The markets are treating this as yet another in the line-up of ‘green shoots’ because the decline was less severe than it was in April (-504,000), March (-652,000), February (-681,000) and January (-741,000). However, let’s not forget that the fairy tale Birth-Death model from the Bureau of Labour Statistics (BLS) added 220,000 to the headline — so adjusting for that, we would have actually seen a 565,000 headline job decline. At least initially, this skew to the data is being readily dismissed.
Fine, but here’s the thing: The Birth-Death model is nothing new. It exists in every single unemployment survey that the BLS puts out. So unless you have reason to believe that the BLS jimmied around with the model this time in order to get a desirable result (and obviously some people will think that), then it’s a little bit specious to then, all of the sudden, just pluck that number out and deny that this was a (relatively) decent report, especially compared to the last one.
It’s kind of like when people (and we totally plead guilty to this) say something like: The REAL unemployment rate is 16% based on the broader measure of people undermployed or not looking for work. Yes, this is bad, but it mainly works as a rhetorical technique because when people think of unemployment they have that single-digit in their mind, and that’s what they’re ultimately comparing it too.