By James Kwak
My previous post on Fannie/Freddie had two major parts. In the first part, I questioned whether the thirty-year fixed-rate mortgage would really go away (or become much more expensive) without Fannie/Freddie, as some people have argued. In the second part, I said, who cares?
The first part has gotten a fair amount of good criticism, for example from Arnold Kling and John Hempton (by email), and also in comments. My position, simplified, was that a thirty-year fixed-rate mortgage includes three kinds of risk: credit risk, interest rate risk, and prepayment risk. Credit risk can be diversified, interest rate risk can be hedged, and Fannie/Freddie didn’t do anything about prepayment risk anyway. This is the kind of theoretical argument people make all the time, and the obvious question is whether the world actually works that way.
I think there are two important criticisms. One, which Kling makes, is that while there are thirty-year fixed-rate bonds (like Treasuries) floating around out there, there just aren’t that many compared to the volume of U.S. thirty-year fixed-rate mortgages. So there might not be enough buyers, and without enough buyers the yields could go way up. One response is that the interest rate risk can be hedged, but that means you have to find a lot of people willing to take the other side of the interest rate swaps, and maybe that would be too hard. So the real question is how much demand there is for thirty-year fixed-rate assets.
The second criticism, which Kling and Hempton make, is that the big issue isn’t the thirty-year maturity; the big issue is prepayment risk. A thirty-year fixed-rate mortgage gives the borrower the right to refinance and pay off the loan at any time, which means that even if you want a thirty-year asset, you can’t count on it. When interest rates go down, you’re likely to get your principal back, and then you’ll have to reinvest it at lower rates. Now, the classical response is that this is just an embedded option (for the borrower), and you should be able to price the option into the mortgage. So the real question is how many people are willing to write those options, and maybe the markets just aren’t deep enough.
I’m not completely convinced by this, though, because Fannie/Freddie didn’t actually hold onto most of their mortgages. Instead, they created pools that issued mortgage-backed securities, and those things had both interest rate risk and prepayment risk (see pages 11-16 of this prospectus for a long list of prepayment risk factors). So people buying the MBS issued by Fannie/Freddie ($5 trillion of them), it seems to me, were happily taking on both interest rate risk and prepayment risk. This seems obvious to me, so I’m almost certainly missing something.
For this reason, I’m still not entirely convinced that the private sector couldn’t take this on. Sure, interest rates would have to be higher because the private sector wouldn’t have the Fannie/Freddie implicit government guarantee* (although, as one commenter pointed out, part of the private sector has something like it). And with higher interest rates, thirty-year fixed-rate mortgages might become less popular compared to adjustable-rate mortgages. But that’s not the end of the world.
Relatedly, John Hempton has another proposal for Fannie/Freddie: simply raise the fee they charge for guaranteeing credit risk. At some point the private sector will step in and take on the job, and in the meantime the government will lose less money and distort the economy less.
* Without the implicit guarantee, the credit risk would be properly priced into the MBS issued by Fannie/Freddie, because Fannie/Freddie already charge a fee for guaranteeing the MBS.