Friday, August 20, 2010

Treasuries: Bubble or Accurate Reflection of Slow Growth?

Sean over at Dead Cats Bouncing details:

While another economic crash landing remains unlikely given the inventory and corporate funding backdrop, there won't be much room for policy error either politically or at the Fed in coming months. Monthly headline US CPI has now fallen for three consecutive months, which has only happened a handful of times since the data series began in 1947. If you take the rough and ready rule that a 10 year government bond yield should equal the long term growth rate plus the long term inflation rate, then it's clear that a near 2.5% 10 year Treasury yield is pricing in a grim growth scenario.
Well, I for one was surprised just how strong the relationship has been.



Update: An anonymous reader commented:
I suspect both nominal GDP growth and treasury yields are mostly driven by inflation, since both real yields and real GDP growth vary less than inflation did.Which means all this graph is telling you is that inflation plus some noise is correlated with inflation plus some different noise
While inflation has indeed dominated the change in nominal GDP over the long run, real GDP has actually been more volatile than CPI over the course of the past ten years. Also of note; ten year annualized real GDP has dipped to 1.62%... the lowest level since the early 1950's.


6 comments:

  1. I suspect both nominal GDP growth and treasury yields are mostly driven by inflation, since both real yields and real GDP growth vary less than inflation did.

    Which means all this graph is telling you is that inflation plus some noise is correlated with inflation plus some different noise :)

    ReplyDelete
  2. Great graphs; will have to digest those for a bit.

    ReplyDelete
  3. Jake,

    The old classical theory on natural interest rates vs. GDP has fallen out of favor -- in general nominal GDP growth and 5-10-year Treasury yields track each other.

    As an aside, that is a reason to worry about GDP growth. :(

    David

    ReplyDelete
  4. Jake,

    I love the graphs. I particularly like how they show the period of the late 1960s as an era of supernormal growth, and I wonder if it would ever be possible (or desireable) to replicate that record.

    You note that 10 year annualized real GDP growth has dipped to 1.62%, the lowest on record for over 50 years. I'm a little concerned becaue I think the 1.62% number is, in part, an artifact of your measurement methodology. It just so happens that the previous 10 years just encompass 2 recessions entirely, so that the measurement process includes the subnormal years of 2001-2004 and 2008-2010. You see a similar dip in 1983, with the previous 10 years encompassing the recessions.

    Would the results differ if you considered different time windows (5 year, 7 year, 15 year, 20 year)? The comparison to a Treasury note of bond may not be quite as natural, but it might give you a better idea of the robustness of your growth measure.

    ReplyDelete
  5. i get your point regarding using different time periods to see the results, but i don't see why we'd want to back out recessionary figures. these figures are what they are over that time.

    ReplyDelete
  6. Jake,

    I understand your position that the number of recessions is relevant to the average growth rate over a period of time -- I do. But what might be more relevant is the average growth rate between recessions. By using different length filters, one might be able to tease out the between-recession growth rate from the within-recession growth rate.

    ReplyDelete