Tuesday, November 10, 2009

Where are Long Bond Yields Going?

The Case for Higher Rates

  • Massive debt issuance (over supply)
  • Central bank diversification (less demand)
  • Rebounding economy will bring inflation (decrease value of nominal debt)
  • Near historical (recent history at least) low yields
The Case for Lower Rates
  • Pension (liability driven investing) demand
  • Flight to safety
  • Excess capacity and labor supply will bring deflation
  • When the yield curve has been this steep, rates have rallied (in recent history)
Lets take a deeper look into the last bullet points for each case (i.e. the low yields vs. the steepness of the yield curve). The blue line in the chart below shows the one way direction of the Long Bond's yield since the early 1980's, which on a stand alone basis points to much more "downside risk" than "upside potential", while the red line shows how cyclical the spread between the yield of the 3 month T-Bill and 30 year Treasury bond has been.

In the second case, if one were to believe in mean reversion, history would indicate the yield curve will flatten in coming months / years. A yield curve can flatten (i.e. the spread becomes smaller) due to a fall in the long-end or due to a rise in the front-end (there are two camps on whether the Fed will hike rates anytime soon... I personally don't see that happening).

So what does history show has been the cause of flattening... the long bond rallying or the short-end rising? To see, below is a colorful chart showing the steepness of the yield curve on the x-axis and the change in the 30 year yield, one year forward, on the y-axis. The various colors show different time frames to show that each "lumping" isn't all from one period (each point represent the end of month figure going back to January 1986).

Notice any patterns?

The first pattern I notice is that the change in the Long Bond's yield becomes much more unpredictable when the yield curve is mildly steep (the first chart shows it doesn't stay 'mildly steep' long) and tends to be more stable in periods when the yield curve is either very flat or very steep.

But the money shot is that when the yield curve has been VERY steep (i.e. more than 4%) the Long Bond's yield has rallied in all cases (19 out of 19 times) over the next year.

Why? The only thing I can think of is that the yield curve becomes this steep due to massive easing by the Fed. And why would the Fed cut rates so dramatically? A slow deteriorating economy, which has historically been the opportune time to own long dated Treasuries.


  1. "But the money shot is that when the yield curve has been VERY steep (i.e. more than 4%) the Long Bond has rallied in all cases (19 out of 19 times) over the next year."

    This is a great observation! One caveat though: the data is taken during a period of time where the 30-year bond yield DECLINED consistently. This observation would be more reliable if it includes 1960-1980 where the 30-year bond yield ROSE.

    If we're experiencing a regime change today, then the relationship likely won't hold.

  2. Having a tough time digging data out that goes back further than the mid-80's, but if I can find it I will post it...

  3. You just underlined the reason why Robertson is buying curve caps, Einhorn is shorting long bonds and Pellegrini says T-Bond short is the next big play.