Monday, November 21, 2016

Will early 2017 be Indian Summer for employment?


 - by New Deal democrat

I've been writing that the economy is in Indian Summer.  By that I mean, a spell of good economic data well after the mid-point of an expansion.  A variety of measures that tend to peak near the middle of the expansion did so at about year-end 2014.  We had a poor period early this year, but by Q3 the economic metrics had improved.

One such mid-cycle indicator that is particularly un-noisy is the YoY% change in employment:


As you can see it tends to be quite smooth, and tends to peak near the midpoint of economic expansions, except in those cases like the 1980s and 1990s when the Fed goes through 2 loosening and tightening cycles.

In this expansion YoY employment growth peaked at about 250,000 per month.  Recently employment gains have been averaging about 150,000 to 175,000 a month.  I suspect we may have a period of improvement over the next quarter or two, perhaps to over 200,000 per month.  Here's why.

First, while the relationship is noisy, real retail sales tend to lead employment.  Here is the YoY% change in real retail sales vs. employment for the last 25 years:




By no means is there even remotely a 1:1 relationship (see e.g. the late 1990s) but in general a waxing/waning in real consumer spending tends to lead to a waxing/waning of employment growth over the next 3-6 months.

We've had a little spurt in consumer spending in the last several months:


You can also see that upward jag in the YoY graph above.  So that suggests an increase in monthly employment growth in the next few periods.

Secondly, a much tighter leading relationship exists between initial jobless claims and the unemployment rate.  Here's the graph going back 50 years:



The leading relationship is pretty obvious.  Now here is a close-up of this economic expansion (you can disregard the green line for now): 



Again the relationship has been pretty consistent.

Now here is the same relationship expresses in YoY%s:



This is a remarkably tight correlation.  Once again, here is this economic expansion:



We can see that as the YoY change in jobless claims turned nearly flat beginning late last year, and the unemployment rate similarly stopped declining a few months later.  Since initial jobless claims have fallen to 40 year lows in the last few months, there's a very good chance that the unemployment rate will similarly decline, at least a little further. perhaps to the 4.6%-4.7% range.

Not only do several leading relationships suggest that monthly employment growth is likely to pick up a little, and the unemployment rate further decline, but that in turn suggests that we may see a little more wage growth, at least nominally.

Coming out of recessions, wage growth tends to decline until the underemployment rate U6 falls to roughly 10%. Once it goes below that number, wage growth tends to increase.  Here's what that looks like:



We're presently at 9.7% underemployment, and we've seen a noticeable if small increase in wage growth.  Since I'm expecting unemployment to decline at least a little more, that suggests that nominal wage growth will also pick up at least a little bit in the coming months.

All in all, there is a pretty good chance that the first half of 2017 will be Indian Summer for employment, with employment growth averaging over 200,000, a slight decline in the unemployment rate, and an increase in nominal wage growth.