MBS yields are probably bottoming

Barring an actual double-dip recession, which I think is quite unlikely, mortgage rates are most likely at or very near their lows. As this chart shows, the spread between MBS yields (which are typically about 50 bps below what homeowners pay on a 30-yr fixed rate mortgage) is about as low relative to 10-yr Treasury yields as it has ever been. 10-yr Treasuries, in turn, are down to levels that are consistent with very low inflation and economic recession. Treasury yields aren't likely to go lower unless all heck breaks loose, as it did in late 2008.

For those unfamiliar with how MBS work, they have what is called "negative convexity." That means that the duration of MBS falls as yields fall, and rises as yields rise. Most other bonds have positive convexity, which means their duration increases as yields fall, and falls as yields rise. (Duration is a measure of how sensitive a bond's price is to changes in yields. A duration of 5 means that a bond's price will rise by 5% if it's yield falls by 1 percentage point.) Institutional investors (the vast majority of whom hold significant amounts of MBS in their portfolios) seeking to keep their overall portfolio duration at a given level must therefore use Treasuries and Treasury futures to hedge their MBS exposure. That means adding exposure to Treasury bonds as interest rates fall, so that the duration gain experienced by Treasuries offsets the duration loss experienced by MBS. If Treasury yields start to rise, the process must be reversed by selling Treasuries. This negative convexity dynamic can at times increase the market's volatility, exaggerating the up and down swings in interest rates, because buying increases as Treasury bond prices rise, and selling increases as bond prices fall. The long and short of it is that if this indeed proves to be the low in yields—and I think it will because I really doubt the economy is rolling over—then the rebound in yields could be rather dramatic.

Yields and MBS spreads are both extraordinarily low right now, and to remain at this level or to go lower would require a rather dramatic deterioration in the economy's growth prospects. Yields are already quite low in anticipation of a double-dip; so the market would at least need a double-dip to happen to keep yields from rising. I for one just don't see the reason for the economy to collapse.

In response to a reader's question, here is the sort of thing I would need to see to be worried about the future: I think it would have to be some unforeseen and unexpected government action that threw a curve ball at the markets and the economy. A round of Smoot Hawley-like tariff wars, for example, that could shut down global trade. A big change in the laws governing the financial system, in which the law of unintended consequences could step in and turn what politicians thought was a sensible "solution" into a new nightmare. A big hike in corporate income taxes, which are already the highest in the developing world. A decision to impose capital controls, which might result in a sharp reduction in foreign investment.