Photo: Wikimedia Commons
You may not recall this, but back in November, there was a pretty big mini-controversy after it was revealed that the strong October jobs report only looked good because of a seasonal adjustment.The headline number was 151K new jobs, but sceptics (like John Mauldin, John Williams of Shadowstats, Alan Abelson from Barron’s, Stephanie Pomboy of MacroMavens, and BI’s Henry Blodget) suggested that by changing the seasonal benchmark, the BLS has conjured an extra 100K jobs out of thin air.
That incident comes to mind now because on Friday we got another gigantic disparity between the headline jobs number and the seasonally adjusted number.
Get ready for this: On a non-seasonally adjusted number, payrolls jumped +376K (!) in June vs. the mere 18K that was reported. You can see the full dataset here. So the seasonal adjustment brought down payrolls by 358K.
But that’s the beginning, because remember the charge of the sceptics is that the government fiddled with the seasonal adjustments back in October to make the number look better.
So let’s go back to June of 2010 to see how things stack up: The headline number was a loss of 192K jobs due to the census (data here), but on a non-seasonally adjusted basis, jobs GREW by 107K (data here). So the seasonal adjustment brought the level down by 299K.
So think about that, if the seasonal adjustment this time had merely been what it was last June, -299K, we would have had another 59K jobs in June (because this time it was -358K), meaning Friday’s number would have been a far more respectable +77K. That’s not amazing, but it certainly suggests that all of the sceptics from last October (Mauldin, etc.) should at least be pointing out how this number wasn’t nearly as bad as the headline suggests.
At a minimum, everyone should drop the seasonal adjustment conspiracy theories that suggest the government only adjusts up, and never down.
See also: More charts from this Friday’s report >