Editor's Note: This post has been edited. Some of the data I had used on Lehman indices was incorrectly posted on Bloomberg, and therefore misreported in my post. Anyone using the Bloomberg index ticker LUMSER should be aware that data set is incorrect. I am working with Lehman to have it fixed. Use LD10ER instead. Anyway, the core of what I was writing stands, but now the data is corrected. My sincere apologies for the incorrect facts, and my thanks to reader FI PM for pointing out the problem.
Under-reported amidst all the sub-prime problems was the terrible performance of agency MBS during 2Q 2007. In fact, in terms of excess return, it was the worst quarter for the Lehman MBS index since 2003. Bear in mind we're talking about the AAA/Aaa rated Fannie Mae and Freddie Mac-backed securities. Not the credit-risky "whole loan" market that is making so many headlines.
Various reasons for the underperformance have been kicked around. In my opinion, it boils down to this. Prices are set by supply and demand. We know from various sources that supply growth has actually been slow this year. Mortgage debt increased by 1.54% in 1Q 2007, which was the slowest pace in 10 years.
So the problem is most likely on the demand side. I argue demand for bonds is made up of three simple factors:
- System-wide liquidity. The more liquidity in the system, the more in demand bonds will be generally.
- Yield competitiveness. A bond will always react to shifts in yield spreads for competing products. In other words bond sectors compete for the same dollars, so if one sector is widening more than others, money will tend to drift toward the widening sector, causing other sectors to widen as well.
- Principal repayment certainty. For corporate bonds, this is default risk. For optionable bonds (like MBS), this is the timing of principal repayment.
So one or more of these factors is moving against MBS. Obviously #2 is a problem. Pretty much all credit-risky bonds have moved wider, which is forcing MBS spreads to widen in order to compete. Also #3 is a problem. Forward interest rate volatility is rising, due in part to uncertainty over whether stubborn inflation or growing sub-prime contagion will cause the Fed to move on way or another. With interest rates so uncertain, the path of MBS principal repayments is hard to gauge.
There is also the problem that we are in the midst of a truly unique housing market. MBS investors are rightly questioning the validity of prepayment models that have not dealt with a sustained period of negative HPA. Even if rates stay relatively constant, how mobile are home buyers going to be in the next 5-10 years? That's a more difficult question than it has been in the past.
All that being said, MBS represent an excellent buying opportunity, in my opinion. Historically, MBS have underperformed Treasuries by at least 50bps 9 times since 1990. All 9 times, MBS rebounded to produce positive excess return the following quarter, with average outperformance of 53bps.
Why would this be? Its more than just mean reversion. First of all, if its increased volatility that pushes MBS wider, the fact is that vol can't logically keep rising indefinitely. Second, increased vol is often an indicator of heightened risk premia. Since the downside for a corporate (default) is much worse than MBS (pays too slow or too fast), its logical that if risk premia continue to rise, corporate bonds will eventually widen more so than MBS. If the volatility subsides without any economic weakness, then MBS are going to revert, and since they initially underperformed they are likely to outperform during the reversion.
Ultimately, very few long-term bond investors want to own too many Treasuries. If credit fears persist, these buyers will look to MBS. Until this happens, I'm happy to clip the coupon.
Fair disclosure. I have a very large portfolio of MBS. Of course, its a $10 trillion market. Oh, and I get like 600 page views/day. Maybe a few more on aggregators. So while I am indeed speaking from position, don't over estimate my powers.