Friday, November 11, 2016

September JOLTS report: this cycle diverges from the last


 - by New Deal democrat

The JOLTS Survey, like the Labor Market Conditions Index and the Senior Loan Officer Survey that I discussed yesterday, has great promise, but like those surveys, does not have a long enough real-time history to use with full confidence.

That shortcoming is on display in the September report, released earlier this week. 

The below graph shows hires (blue), job openings (red), and quits (green, right scale) from the inception of the series nearly 20 years ago:



In the one and only complete cycle since the series began, hires and quits peaked first, while openings continued to increase until shortly before the onset of the 2008 recession.  During this cycle, hires have gone sideways for over a year, while openings continued to rise until recently.  Meanwhile quits just spiked to a new high. 

Another way to look at this is YoY.  Here's the YoY% change in from 2004-07:



Here is the YoY% change in the last several years:



In the last cycle, YoY job openings held up until nearly the end.  This time around, it appears at least for now that openings may be turning before quits.

In other words, unless job openings make new highs in the months ahead, this cycle will not match the last one.  We just don't have enough history with the JOLTS series to know which resolution is more likely.

One related point is that, generally, hiring slows down before firing increases shortly before the expansion ends.  Here it is, comparing the quarterly average of the m/m% change in employment growth vs. the quarterly average of new jobless claims (right scale, inverted):



Note that employment growth (which is the net of hiring over firing) decelerates markedly before layoffs begin to increase (shown as a downturn in the graph).

JOLTS decomposes this into hiring (blue) vs. firing (red, right scale)(averaged quarterly):



We can see in the 2002-07 cycle that hiring decelerated before firing turned up. It appears that the same thing is happening now.

So the bulk of the evidence from JOLTS is that we continue to have late cycle deceleration, but with no imminent downturn.  But take this with an extra grain of salt.