Saturday, June 10, 2017

Weekly Indicators for June 5 - 9 at XE.com


 - by New Deal democrat

My Weekly Indicators post is up at XE.com.

No big changes from the recent paradigm.  We are all waiting on the Fed this week.  If it raises rates, what will happen to the yield curve?

Friday, June 9, 2017

Is the employment situation weaker than we thought? Probably not


 - by New Deal democrat

Via Menzie Chinn of Econbrowser Wells Fargo describes The Quarterly County Employment and Wages report as "a detailed count of employment and wages derived from the unemployment insurance tax rolls and serve as the basis for the annual revisions to the monthly employment series."

And the big news is, it declined -- substantially -- in the last quarterly report, released this week.  Is the employment and wage situation a lot weaker than we thought?  Probably not.  Here's why.

Since it isn't a small sample, but is derived from the total data, I would expect it to show a smooth progression, even at turning points.  But that's just not the case.

To show you how volatile the YoY measure is, here is the last eight quarters of the YoY change in average weekly wages, with the nominal $ amount: 

Q1 2015 +2.1% $1048
Q2 2015 +3.0% $968
Q3 2015 +2.6% $974
Q4 2015 +4.4% YoY $1082
Q1 2016 -0.5% YoY $1043
Q2 2016 +2.2% YoY $989
Q3 2016 +5.4% YoY $1027
Q4 2016 -1.5% YoY $1067

I could do a similar exercise with previous years.  The simple fact is, there is just simply enormous volatility that is difficult to reconcile with reality in this series. [There is huge seasonal volatility, which is why the YoY change is so important. While FRED does have this info graphically for each of the approximately 400 metro areas in the survey, they don't carry the national average!]

By contrast, here is a graph of average weekly wages from the monthly employment report:



And here is the graph of the quarterly report of "usual median weekly wages":



Bottom line: the QCEW is and has been for years a volatile outlier both to the upside and downside.  Until someone can explain why the QCEW is so volatile, I am discounting it.  In the meantime, should any Doomer trumpet the most recent report (as a few did with the Q1 report), google their reporting on any of the positive numbers.  Hint: you won't find it.

Wednesday, June 7, 2017

No, record job openings in JOLTS do not mean that everything is Teh Awesome!


 - by New Deal democrat

Once again most of the commentary on yesterday's JOLTS report for April was that job openings jumped, so everything is Teh Awesome!

<  Sigh  >

To recap one more time... 

In the one and only complete business cycle that we have for this data:

  • First, hires peaked. They started a long plateau in 2005, making a 3 month peak in late 2005, with no meaningful progress thereafter. 
  • Second, quits peaked. They started to plateau in early 2006, making a 3 month peak in spring 2006, with no meaningful progress thereafter.
  • Finally, openings peaked in Q1 2007
Hires and quits are the only *hard* economic data in the series. "Openings" can be aspirational trolling for a future bank of resumes or, worse, designed to fail and lay the groundwork for cheap  H1-B foreign slaves.

So, here is the entire history of hires and quits:



Noisy, but it sure looks like both have established plateaus again.  Here's the YoY look:



Hires are flat YoY, and quits have decelerated from roughly +10% to +5%.

Here's the close-up of monthly data for the last 24 months, with both hires and quits set to "zero" for April 2015:



Hires made a peak in December 2015, and are actually lower now than they were 24 months ago. Quits recently peaked in January. Beginning next month, they are going to have much more difficult YoY comparisons. For example Quits in April were only 1% higher than 11 months ago in last May.

The pattern looks very much like 2006.

Finally, here are job openings:



New record! Whoop-de-doo. At least we aren't in the equivalent of mid or late 2007. 

The good news is, there is nothing in April's JOLTS data suggesting any imminent economic downturn.  The bad news is, it adds to the accumulating evidence of late cycle deceleration in the jobs market.

Tuesday, June 6, 2017

In which I disagree with Prof. Tim Duy


 - by New Deal democrat

You should NEVER rely on any one single indicator. No indicator is perfoct -- in this case, the yield curve.

This post is up at XE.com.

Monday, June 5, 2017

YoY job growth and the Fed funds rate


 - by New Deal democrat

Continuing with a more detailed look at the current employment situation, one of the least noisy series - that also has a very long history - is the YoY% change in payrolls:



In general YoY payroll growth increases until about midpoint in an expansion, and then decelerates until finally turning negative shortly into the next recession.

So the first thing to note is that YoY employment growth peaked in early 2015, and has been decelerating ever since.

But there are several counterexamples -- notably the 1960s, 1980s, and 1990s -- where YoY employment had several peaks. That's what I want to take a further look at today.

All three of the above counterexamples coincided with expansions in which there were also multiple Fed interest rate cycles.

Just how close is the match?  Take a look at YoY payroll growth (blue) compared with the YoY change in the Fed funds rate (red):



There were at least 3 notable peaks in the YoY change in the Fed funds rate in the 1960s, and two each in the 1980s and 1990s. And they correlated quite well -- within several months -- of both the peaks and troughs in YoY job growth.

In fact the relationship is decent enough that, as a general rule, when the YoY change of the Fed funds approaches or exceeds YoY job growth, the expansion is in trouble (with 1994 being the only exception).  Note that in all but two of the above cycles, the YoY change in the Fed funds rate was +2% or higher:



The exceptions were 2000, with a 1.75% YoY increase in Fed funds, and 1955, with a 1.5% increase. The 1955 exception is particularly noteworthy, since that was also a case where the yield curve never inverted.  This underscores the point that one cannot rely upon and increase in short term interest rates as the primary determinant of when an expansion in a deflationary era will end.

To return to my main point, I emphasize that I'm not claiming any causal relationship here, just that the same underlying reasons that give rise to the Fed tightening and loosening have historically also similarly affected employment growth.

Which brings me to the current expansion.  Here's a close-up:



The current situation violates the paradigm since 1955 in that there was no increase in the Fed funds rate as the rate of job growth rose.  Now, in a negative development, we have increasing interest rates in the face of declining YoY employment.  But we still haven't had enough of an increase that would have correlated in the past 60 years with an oncoming recession.

If the Fed should raise rates again later this month, it will certainly be problematic.  Will the Fed reverse quickly enough should there be a further stumble in employment, to extend the expansion, as in the 1980s and 1990s? Frankly, I doubt it, but we'll see.

A Day in the Life of the jobs market, May 2017


 - by New Deal democrat

Fifty years ago, when I was a little teenybopper, this album came out and blew me away:



Why bore you with this ancient Boomer reminiscence?

Because the unemployment rate has only been lower than last month's 4.3% in only six of the last 50 years, and only two of them in the last 46 years:



Since February 1970, the only time the unemployment rate has been less than it is now is from 1999 into 2001.

That's not trivial.  If you are under retirement age, then if you are unemployed, your odds of finding a job now are better than almost 90% of the entire time since you reached working age -- or better.

While we all know this isn't the entire story, I want to take a slightly different look by starting with the U-6 underemployment rate, and comparing the "slices" of the jobs market from there.  Here are the "slices:"
  • employed full time
  • employed part time voluntarily
  • employed part time because no full time job is available
  • unemployed but looking
  • not in the labor force but want a job
  • not in the labor force and not interested in a job
The U6 underemployment rate in May was 8.4%.  Let's see when it has been that rate before (red in the graph below. note that most of these statistics are only available back to 1994):



The underemployment rate now is where it was in 1997 and the end of 2005.

Further, the ratio of full time to part time workers (including both voluntary and involuntary part-timers) is also where it was in 1997(blue in the graph below, compared with red = U6)):



as is the proportion of those who aren't even looking, and aren't in the labor force, but want a job now:



What is higher is, among those who are part-timers, the percentage of those who are so employed because they couldn't find a full-time job:



This is more like it was in 1996.

Turning to the U3 unemployment rate from the 1990s, we see that it was in the 5.0%-5.5% range in 1996, and the 4.5%-5.0% range in 1997:



The bottom line is that, when we leave aside those who tell the Census Bureau that they *aren't* interested in working now, we see that the situation now is roughly equivalent to the time when there was a 5.0% unemployment rate in the booming 1990s economy.

On top of that, there is some number of people, mainly currently on disability, or caring for their children or parents at home, who would probably re-enter the jobs market and be interested in a job, if wages were better.

That's a Day in the Life of the jobs market in May 2017.