May 4, 2005

General Glut on the New 30-year

He really read my mind with this one:
As I type the 10-year is up to 4.23%, rising over 1% today. Now this is still nothing compared to where it was in late March (~4.6%), but one has to wonder if the ultra-low ten-year yields can last with the 30-year on the way. And more importantly, what will this do to mortgage rates which tend to move in tandem with the 10-year? After all, it was late March when the 30-year FRM topped 6.0% for the first time since July 2004.
A while back Brad Setser pointed out the tight supply of U.S. long maturity debt as a reason for the yield curves flatness and I commented:
While the limited supply of long-maturity bonds is an interesting data point I am not sure it explains a whole lot about the low yields. There has been no supply limit on the corporate side of debt issuance.
If the restricted supply of treasuries is keeping yields down then corporate spreads should be wide and not historically tight as they are. Maybe it is a second tier effect of some sort where the low treasury yields create the illusion of stability which leads to spread tightening but that is bit tougher to believe.
Probably easier just to say that bond investors expect inflation to remain low for a long time.
Sitting here today I was probably wrong and Brad was right. My comment puts way too much faith in the capacity of investors to act rationally and widen out spreads as non-UST debt grew.

I did not show up today with bond shorts on but have been looking for steepening (and been wrong) for quite a while. I am not planning to chase yields here and expect some retracement of this mornings move. Unless the Treasury announces in August that we don't need a long bond after all, this announcement looks like a green light to sell the long end of the yield curve.

1 comment:

  1. in my general analysis, i suspect i put too little emphasis on the relatively subdued corp. issuance (net) in 2004, which I certainly do think is part of the story. net issuance in 04 of all long-term bonds seems to be much smaller than in 03, per the BMA tables.