Monday, July 28, 2014

A Brief Look at Unemployment Persistence

I have been using this graph, which compares unemployment to insured unemployment, to guide expectations of labor trends.  I am afraid that we have a "new normal" where there will be strong pressures for pro-cyclical labor policies similar to those we have seen in the last 5 years, that continue to create persistence in unemployment as we move out of recessions.



Source: http://www.frbsf.org/economic-research/files/JEP-slides.pdf
There appear to be long term trends in the Beveridge Curve (job openings vs. unemployment), which are probably a product of policy and cultural changes and demographic factors.  It occurs to me that there may be some status quo bias in this thinking, though.  So, I extended the graph of total vs. insured unemployment back further.  I can get back to 1971 with data from Fred.  (I used the 12 month moving average of continued claims as a proportion of Civilian Labor Force for both better consistency in relative long term levels and noise reduction.)

The Beveridge Curve suggests that there were frictions in the labor market in the 1970's and 1980's that inflated the unemployment rate.  I would have expected this to be related to more persistent unemployment, higher unemployment durations, and thus higher unemployment compared to insured unemployment.  But, the recessions in this period don't exhibit high unemployment relative to insured unemployment.  And, later in the 1980's when unemployment remains high compared to insured unemployment, the Beveridge Curve implies a more robust labor market.

I take this as a sign for optimism.  Maybe short term trends aren't destiny...Of course short term trends aren't destiny.  The current cycle has been unusual in the level of total unemployment and in the persistence of high insured and uninsured unemployment.  The cycles in the early 1970's and 2000's had fairly low total unemployment, compared to insured unemployment.  The 1980 recession also was proceeding well until the second wave came.  Then the we moved through the worst of it pretty quickly.  (The red dots represent each month.)  But, after the recession ended, there was persistence in high unemployment for several years, and this persistence continued through the early 1990's recession, considering the very tame levels of insured unemployment. (The red dots are above the normal range of the relationship and are close together.)

This relationship might be an interesting one to watch as unemployment peaks during the next downturn.  Have pro-cyclical policy and cultural changes meant that this relationship been much steeper since the 1970's, but low relative peak levels of insured unemployment kept total unemployment low until the most recent downturn?  Or, is there a chance that the next time insured unemployment gets to 3% or 4%, that a healthy labor market will help push total unemployment down quickly as insured unemployment declines?

Maybe my interpretation of the Beveridge Curve is backwards.  Maybe instead of thinking the unemployment rate was high compared to vacancies in the 1970's, maybe we should think that the vacancies rate was high compared to unemployment, and this is related to the healthy recovery in total unemployment as insured unemployment declined from the peaks.  I have been thinking of the Beveridge Curve on a slope, and that a more robust labor market is signaled by a movement toward the origin (movement number 1).  But, maybe a better functioning labor market is signaled by a shift to the left, but also a flattening, in the Beveridge Curve (movement number 2).

Maybe right/left shifts are related to labor supply and up/down shifts are related to labor demand.  To the extent that the relationship moves cyclically, it moves diagonally along the curve as both supply and demand change through the sentiment changes of the business cycle.  But, possibly, the unusual position of the 1970's Beveridge Curve was a combination of loose monetary policy that kept vacancies high during cyclical downturns and other policies or cultural shifts that were keeping unemployment high during recoveries.  Maybe a little looser monetary policy with a less regulated labor market is the best of both worlds, but of course I would say that.

No comments:

Post a Comment