Comments on: Year-over-year Non-farm Payrolls

By: David Merkel

David Merkel — Wed, 28 May 2008 20:36:27 +0000

Vermont — things are weakening now. No doubt, but the question is how much and how long?

By: Vermont Trader..

Vermont Trader.. — Thu, 08 May 2008 01:36:48 +0000

Just looking at your graph it is obvious that year/year job growth is weaking. This is consistant with an economy that is getting weaker and a pattern seen prior to every reccession since the 70’s.

So the data is bearish no matter what the B/D adjustment is.

By: David Merkel

David Merkel — Wed, 07 May 2008 01:49:14 +0000

Yes, it is a better way to view it. It would be useful if the BLS published the job creation and destruction series. Do they?

By: Jeff

Jeff — Wed, 07 May 2008 01:41:12 +0000

Good question on the seasonality. I am comparing to the QCEW data which has its own seasonal adjustment.

We might both try to write something on seasonal adjusting. People really want to compare one month to another, so the concept is sound. As you point out here, there may be a question about what is seasonal.

The main point I am making about your chart (and everyone else’s of course) is that you are taking the B/D adjustment as the numerator for the NET job change — a highly variable number which can be small or even negative.

It seems more appropriate to view the B/D adjustment as part of the overall process of estimating job growth from business births, which run at 2.5 million per month. Viewed this way, it is a small fraction of a relatively stable number. Doesn’t that seem a better way to view it?

(There are also, of course, about 2.5 million job losses from business deaths.)

By: David Merkel

David Merkel — Tue, 06 May 2008 19:52:37 +0000

“Eventually we get actual numbers. Since the introduction of the B/D adjustment there has been no quarter where the BLS forecast would have been more accurate without the adjustment. It is something to think about.”

Interesting. Is that added accuracy with the seasonal adjustment, or without it?

There has to be something like the B/D adjustment… don’t get me wrong. I am interested in whether it is the right overall size, and whether the seasonalization is being done right.

Thanks for posting. This one takes a while, and I am not sure I have it right yet.

By: Jeff

Jeff — Tue, 06 May 2008 19:31:59 +0000

Thanks for the mention, David. It happens that I am working on this topic, trying to take a different approach that may provide more clarity.

Meanwhile, it would help if everyone would keep in mind the number of new jobs actually created each month — about 2.5 million or about 30 million per year. The B/D adjustment is only a small part of this.

Eventually we get actual numbers. Since the introduction of the B/D adjustment there has been no quarter where the BLS forecast would have been more accurate without the adjustment. It is something to think about.