The report for October is out. At first glance it’s ugly – 10.2% is up 0.4% from the prior month, and that is a LOT of increase.
When I looked in more depth, however, I started becoming troubled. Or maybe ‘encouraged’ is a better word. There are what seem to be some strange elements buried underneath.
Take for example the difference between seasonally adjusted and not seasonally adjusted (SA and NSA, respectively) numbers. The NSA did NOT CHANGE from September to October, and in fact there was an increase in total employment according to the NSA numbers. There are similar indicators in SA vs NSA numbers in employment to population ratios.
It’s important to remember that SA is a tweak to the numbers. It’s built to try and minimize misleading spikes in the data. It has a major problem, however, in that the adjustment is based on historical trend and expectations. As a result when the trends reverse there’s a period of a few (2-5) months where the indications of SA are contrary to the indications of NSA.
One month could be a demonstration of why we use SA. On the other hand, I’m noticing the difference between SA and NSA is very large. As in – according to NSA the unemployment rate is 9.5%, or 0.7% lower than the SA number. In this report that’s a large spread. It makes the possibility of a reversal more likely – not guaranteed by any means as this spread has happened in the past, but it’s also quite common to see in trend reversals.
If it’s a reversal we’ll get a similar peculiarity in the next couple of months’ reports. Keep an eye on it, but don’t bet on it yet.