Friday, September 6, 2013

1970s vs. 2000s: Gender Effect

Steve Waldman at says that demographics caused inflation in the 1970's, not monetary policy.  Scott Sumner pushes back a little.

I think there are some semantics going on here, but if I understand Steve correctly, he is saying that the Fed was goosing the money supply in order to accommodate the quickly rising labor force brought on by the sheer number of baby boomers and the new tendency for young women to work.  I'm not sure I understand the mechanisms at work there, or whether it was a matter of the Fed meeting a true need or being fooled into over-inflating.

But, the discussion prodded me to think some more about the differences between these decades.  If the housing market and low real interest rates are the result of parallel demographics, why did the 70's see skyrocketing labor force participation while the 2000's saw plummeting LFP?

Jumping off from this post, I decided to make another adjustment to LFP.  Here is a graph comparing the actual LFP (and a forecast LFP that accounts for the aging labor force), to a modeled LFP that assumes the pre-baby boom LFP's had remained stable (in other words, no bump from working women):

The difference is working women!  Without them, the 1970's might just have looked like the 2000's, with declining LFP, and, if Steve Waldman is correct, low inflation.  But, now that female LFP has peaked, we don't have this factor to save us any more.  In fact, the problem is made worse.  My counterfactual LFP is based on stable age-group LFP's, but male LFP's across most age groups have been declining, slightly but persistently, for many decades.  Female LFP's peaked in the 90's and are now declining at a rate similar to the male rates.  So, not only are we lacking a fresh new female labor force, but women are actually leaving the labor force now.

This has tremendous implications for long term interest rate forecasts.  If there really is some reliable relationship between inflation and labor force growth (and Japan is another obvious anecdote in favor of this relationship), then we have several decades of low inflation ahead of us, and several business cycles where "zero lower bound" dynamics will be in play.

The odd thing is that the Fed should clearly be able to set long term inflation targets wherever they want, so if this is a factor, it would have to be playing out as some sort of institutional behavioral bias that is persistent through different eras of monetary thought and technique.

Addendum:  Steve Waldman responds, then Scott Sumner responds, with some very good comments by Mark Sadowski.  I think Waldman has some interesting points to consider, but Sadowski and Sumner have strong counter-evidence.

No comments:

Post a Comment