Monday, May 25, 2015

Unemployment by Duration and Interest Rates

Here is a graph of unemployment durations over the past several years, as of December.  For durations under 27 weeks, we are back to 2007 levels (which were similar to 2005, 1999, and 1995 levels).  (The unusual approx. 1 million remaining in the unemployed workers in the "over 26" category have an average duration of over 100 weeks.  The normal approx. 1.5 million workers in that category should have average durations of around 56 weeks.  The remaining additional workers are marginally attached to the labor force, and likely have little effect on general trends in the labor market.  That is not to minimize their challenges or to ignore that they will probably slowly re-enter the labor force.  But, they should have a very diminished influence on things like interest rates and wage growth.)  In other words, employment is back to levels we would normally see at the end of an economic recovery, with high short term interest rates and a flattened yield curve.

What the Fed does with short term interest rates between here and 2% isn't going to matter much, as long as we can keep out of a highly deflationary context.  What will matter for the business cycle (besides attracting capital back into the housing market) is how the Fed manages the money supply when the yield curve flattens.

The year 2017 could be the year 1997, or it could be the years 2001 or 2007.  Whether they begin raising rates in September 2015 or in 2016 shouldn't have much bearing on that.  Right now there is more than enough liquidity for everything, but just not enough ways to employ it as housing capital.


  1. Yes, I concur that the Fed raising interest rates by a quarter point will almost be a non-event. But I do worry that that Fed, by not conducting QE, will essentially create in Japan- like economy.

    1. I agree. But if they would take the clamps off of mortgage lending, we wouldn't need QE because euthanizing half the banks balance sheets is what is keeping QE from becoming cash.