Tuesday, July 7, 2015

June 2015 Employment

We now have real wages and JOLTS through May and national employment data through June.  June looked like a decent continuation of labor recovery, with the unemployment rate falling to 5.3% and establishment data estimating 200,000+ job growth.  Markets seemed to react to Thursday's employment figures with some disappointment, which continues to suggest that the focus has shifted from unemployment to wages.  I don't think that we should expect wage growth to cause inflation, but wage growth does seem to correlate with real interest rates, so that may be a distinction without a difference, if we think of Fed policy in terms of nominal interest rate targets.

Some of the disappointment may have been related to the decline in labor force participation, which may have led to a muted reaction to the decline in the unemployment rate.  But, I think the details generally back up a positive reaction.  The unemployment rate moved back to the range we might expect with current levels of unemployment insurance claims.  It should continue to trend in that direction.

This doesn't look like a fluke.  Much of the decline in unemployment was from long duration unemployment.  Using data from the unemployment durations table (A-12), I have estimated that unemployment durations have moved into a regular distribution, but with a group of Very Long Term Duration unemployed (VLTUE) workers (durations typically over 2 years) that had been declining by about 0.05% per month, but had leveled off in recent months.  This had consisted of about 1% of the labor force in early 2014, and had been hanging around 0.7-0.8% this year.  With the large drop this month, my estimate of VLTUE workers fell below 0.5% for the first time.

Suddenly, the VLTUE issue looks to be less of a defining factor, and the relationship between insured unemployment and total unemployment looks a lot more like the labor market of the 1980s, where there wasn't an unusual level of long term unemployment, but short term unemployment remained slightly elevated for several years, relative to insured unemployment.

Flows from unemployment to employment were strong.  Unusually high flows between unemployment and Not-in-Labor-Force (NLF) have been associated with the unusual levels of very long duration in this cycle.  There was not an unusual flow of unemployed workers leaving the labor force this month.  There was an unusually low flow from NLF to unemployed.  This is a flow that would be associated with a strengthening labor market, even though it means that it caused a decline in labor force growth.

Most of the decline in labor force came from a sharp increase in employed workers leaving the labor force.  This is also a flow that is related to a good trend, even though it reduced the labor force this month.  So, I think the fall in labor force should not be a concern.  The declining unemployment rate should be sustainable.

Looking at the net flows (as weighted moving averages, to reduce noise), this month the recent fall in net UtoE flows stopped, which suggests that employment growth remains strong.  Here we see again that EtoN flows were the big mover this month, responsible for most of the decline in labor force.  These flows (the green line) have been strong recently (similar to the 2004-2006 period) and moved back to more normal levels this month.  I suspect we will see this move back up to strong levels again, leading again to recovery in labor force, but with no direct effect on unemployment.

Source: http://www.frbsf.org/economic-research/files/JEP-slides.pdf
JOLTS data (through May) confirm this, with hires and quits continuing to trend slightly higher.  These will probably level out, since it looks like we are settling into a mature recovery in labor markets.  Job openings continue to exhibit unusually strong behavior.  This might be related to a rightward shift in the Beveridge Curve.  This could be partially demographic - older workers tend to have more difficult re-employment frictions.  The Beveridge Curve shifted right in the late 1960s and 1970s also, with persistently high job openings.  The late 60s and 1970s saw large permanent increases in Federal transfer payments.  And, we have seen a persistent jump in transfer payments with this recession, also.  Possibly, persistently higher job openings are related to these developments.  The historical Beveridge Curve shown here is not up to date.  Job Openings are up to about 3.6% now, which corresponded to an unemployment rate in the late 1990s of about 4%.  The Beveridge Curve shifted left in the 2000s, but the current trajectory looks like it is headed more in the direction of the late 60s.

If that is the case, the high Job Openings level could be good news, cyclically, but bad news, secularly.  Cyclically, this means that employers are expanding.  But, secularly, higher job openings could be related to public policies that remove potential workers from the labor force.  The rightward shift of the Beveridge Curve in the 1970s came with declining relative compensation, declining real wage growth, declining real GDP growth, and higher unemployment.  Maybe a continued sharp increase in Job Openings should cause us to shade lower our expectations for economic growth in the coming decade, especially given the demographic headwinds, the stagnant mortgage market, a new wave of novel employer mandates, and the hawkish Fed.  Could a sharp increase in Job Openings be the harbinger of a larger zero lower bound problem than we fear?

Finally, here is a chart comparing real wage growth to unemployment.  There are many complaints of low wage growth, but this seems to be a product of low inflation.  Real wage growth has been very strong during this period of high unemployment, and today appears to be roughly where we would expect it to be - around 1% or a little less, Year over Year.

Wage growth is actually stronger than it appears, even now, because about 0.7% of inflation is due to housing supply constraints.  Without that unusual supply situation real wages would be rising by 1.5% or so, annually.  (Actually, they would be rising even more than that, because additional building would be lowering the unemployment rate, which is associated with stronger wage growth.)  So, in effect, real wages are rising by about 1.5%, annually, but households are choosing to use some of that wage growth to bid up rents on a stagnant housing stock.

This is probably pulling down productivity, too.  Instead of building million dollar apartments in Manhattan, consisting of $500,000 in building expenses and $500,000 in transfers to real estate owners, we are building million dollar houses in the hinterlands, consisting of $900,000 in building expenses and $100,000 to real estate owners.  There is a lot of make-work required to build housing in sub-optimal locations.  (These specific numbers are hypothetical.)


  1. Hey Kevin, is there anyone writing interesting stuff about Greece that you have read? I hate to be a misanthrope, but it seems like nothing I read has any real depth to it.

    1. Steve Waldman at Interfluidity (in my blogroll) had a couple of passionate posts on it. It's a populist perspective (not mine), but it seemed to come from a more detailed and personal place than other writers I have seen.

      It's not a topic I am that interested in, though. It's more politics than economics.

    2. Maybe thats why everything seems thin to me, I keep reading for the econ, and get politocon.