Wednesday, August 27, 2014

Regression toward the mean / Pendulum swings back


Hussman has a very interesting graph.




We are currently at extremes never reached before.   We should expect regression toward the mean or for the pendulum to swing back past the mean.   We should not expect to stay at the current extreme levels for low interest rates or for high monetary base to NGDP ratio for a long time.

So how do we return to normal levels?   With  US deficit spending levels, and Yellen as Fed Chair, I don't think there is any real chance of reducing the monetary base.   Over a long enough time, the real GNP might grow by a factor of 2 and bring us back to normal levels of monetary base to NGDP, but I don't think we have that much time before something happens.  So what will happen?   The easy way is for prices to go up so that NGDP goes up.   That is what I expect.   Prices going up.

Prices going up by a factor of 2 or 3 gets us near the mean for the above data.  However, after such extremes, I think we should expect a pendulum swinging to extremes at the other end of the graph.   So prices should go up well over a factor of 3 and we should have high interest rates.

5 comments:

  1. SeekingAlpha.com posted a few of my posts over there and one got a bunch of comments:

    http://seekingalpha.com/article/2451485-how-we-know-high-inflation-is-coming

    ReplyDelete
  2. This guy's not a Hussman fan, have you seen this?:

    http://www.etf.com/sections/index-investor-corner/23004-swedroe-why-care-what-hussman-forecasts.html?fullart=1&start=3

    ReplyDelete
  3. ... also you might be interested in this Smith response to Beckworth on inflation and inflation targeting wrt LSAPs:

    http://informationtransfereconomics.blogspot.com/2014/08/smooth-move.html

    ReplyDelete
  4. Thanks Tom!

    Krugman is forever saying nobody has a model of how too much debt leads to trouble an I am forever commenting that I do and getting ignored. Did it again:
    http://krugman.blogs.nytimes.com/2014/08/30/day-of-imfamy-2/#commentsContainer

    ReplyDelete
  5. Hi Vincent,

    The graph you posted above indicates that log r ~ 1/log MB, which is exactly the relationship here:

    http://informationtransfereconomics.blogspot.com/2014/08/are-interest-rates-good-indicator-of.html

    There is an overall rising and falling trend, so there might not be any mean reversion -- at least in this model.

    ReplyDelete

Looking for polite debate on ideas. Never attack a person. Be nice.