Horrible continuing claims data speak to decelerating job growth

Real quick here.

I saw the jobless claims data today and thought they looked pretty good. But, I decided to go back to my 2008-2009 practice of sifting through the raw data for both initial claims and continuing claims. The numbers weren’t nearly as good. And, continuing claims look downright ugly.

The missing jobs data dive

Here’s the chart I want to focus on first.

Source: St. Louis Fed

What you’re seeing here are two very sizable upticks in continuing claims. The first began in early 2015 from a very bullish level, with in excess of -500,000 fewer people continuing to claim unemployment insurance in early 2015 versus the year ago level.  By May 2016, the number crested at just over -50,000. The end result was a substantial mid-cycle slowdown but no recession.

The second uptick began from a less bullish level simply because the expansion had been so long. We still had over -300,000 fewer people filing continuing unemployment insurance claims in September 2018 than a year prior. But now, just six months later, that number has shrunk to barely 100,000. This is a huge uptick. And it suggestive of job market weakness.

No one is talking about this. When I looked at the raw unadjusted numbers, the movement is the same, if more precipitous. The peak differential was the week ending December 15, with an average -244,469 fewer claims than the year ago level in December 2017. The last number was -133,875. And that’s up in three weeks from -205,015.

Should we be concerned?

This makes me nervous – especially because initial claims data have been weak. The 4-week average initial claims data are still above year ago levels. And the six-month comparisons look worse, with the numbers up over 15,000 since September. As long as we stay in the 230,000 range for jobless claims, we shouldn’t be overly concerned. But, if we start drifting above that level, it’s a sign that things are going pear-shaped and that we need to be watching for other signs of weakness.