Like many people, I like to believe that I prefer the government policies that I prefer because they’re a Good Thing for the world, not because they advance my self-interest. But as a relatively new homeowner, I break down a little on mortgages. Sure the mortgage interest deduction is a crazy and inefficient boondoggle, but it’s my crazy and inefficient boondoggle, and I don’t really want my apartment to lose (more) value if the deduction goes away.
Similarly, I’m pretty psyched about the plan that’s been kicking around, and that made it in vague form into the president’s jobs proposal, to allow people to refinance mortgages even if their houses are underwater or their income wouldn’t support the new payments. When I took out my mortgage I had a bit over one turn of leverage, as it were, which had my mortgage bankers congratulating me and asking if maybe I wanted to take a little more money just in case. Whereas now I make TXU look like a strong credit. Because um blogging pays less than banking you see. So I like the idea of being able to reduce my mortgage payment without actually having to try to convince a banker that it's a good idea for me to keep this much debt. Even though I’m not entirely convinced that it’s good for the world.
Others are also skeptical. The Congressional Budget Office did a rough-cut analysis of a version of the plan, which is mildly critical. The CBO sums up the tradeoffs as follows:
A well-designed and well-executed large-scale refinancing program with relatively broad eligibility criteria would have benefits both for borrowers with above-market interest rate mortgages and for the enterprises providing the credit guarantee on those same loans. Those benefits would come at a cost to those who invested in the MBSs backed by the loans. Some of that cost would be borne by the federal entities (the GSEs, the Federal Reserve, and the Treasury) that have amassed portfolios of those securities.
How would those investors lose money? Well, they’d get paid down at par – and their MBSs aren’t trading at par. Specifically they’re trading like this:
This shows the trading price, as a percent of par, of Fannie Mae 30-year mortgage backed securities with a coupon that’s about 150 basis points above the coupon of the 30-year Fannie Mae MBS that trades at par. So on a day where a 4.5% MBS trades at par, this shows the trading price of a 6.0% MBS. The number got close to 102 in January of 2009, but has gotten close to 108 this year. If you have a bond marked at 108 that gets paid down at par, you lose 8 points.
You can make moral arguments about this. Some dude told the FT:
“The study recognises the enormous losses private investors would suffer in a transfer of wealth to borrowers,” said Joshua Rosner, a housing finance expert and managing director at independent research firm Graham Fisher & Co. “While such a transfer would be acceptable to some in Washington ... it would result in the unwillingness of investors to buy mortgage-backed securities without charging an exorbitant risk premium to compensate for the event the government does this regularly.”
On the other hand, Felix Salmon at Reuters says:
But what’s happening right now is that mortgage bonds are trading well above par just because investors are well aware that refis are hard to come by for many homeowners. They’re basically taking unfair advantage of the fact that homeowners are locked into above-market mortgage rates. If the value of their bonds came down towards the face value of the bonds, that would be a good thing. It’s not good when mortgage bonds trade well below par, but it’s not good when they trade well above par, either — it’s a sign of market failure. Remember, there would be no default involved here. So bondholders really couldn’t complain much.
Now this Joshua Rosner person has a theoretical point: if you eliminated all barriers to refinancing, refinancing would be more common, you could never see any principal appreciation on your MBS, and you’d have to charge more in interest to make up for the fact that your bonds can only lose value and never gain. So implementing this policy would drive up mortgage rates for everyone in the future.
On the other hand – there will still be barriers to refinancing. Barriers like personal laziness, fees (which the plan would sort of risk-normalize but not eliminate), etc. In other words, the barriers that you’d see in a normal economy. The barriers that the plan would try to eliminate are pretty specific to this economy, to wit (1) massive unemployment and (2) massive home value deterioration.
But the CBO's chart actually lets you have sort of real, data-driven thoughts about these questions. And by "you" I don't necessarily mean "me" - I know nothing about mortgage-bond math except that it involves copulas, and that those aren't as sexy as they sound. I haven’t gotten to that section in the CFA materials yet. (Wait, is that actually a section?) But still, you can be a human and think about these things with rough numbers. Two things suggest themselves.
First, the 10-year data provides some benchmark for what refinancing risk is "supposed" to look like. In a normal economy, the free-refinance option flattens a 150bps coupon premium to just under 4 points of bond price. If it trades at 106, or 108, or whatever, in the midst of a once-in-a-generation recession, then those extra 2 or 4 points are probably *due to* that recession and are, as Salmon says, “a sign of market failure.”
A policy goal of getting these bonds back to the 104 area would help borrowers into efficient refis – while not crushing bondholders in a way that would reduce further extensions of credit. Future mortgage buyers would not say to themselves "every time rates drop, the government is going to screw me out of all my profit." Rather, they'll say to themselves "if rates drop 150bps, I'll make about 4 points of capital appreciation, just like I would have in the really quite robust housing market of 2000-2005." That's presumably a target that you could manage toward if you wanted to, e.g. by optimizing how the refinance plan deals with fees, by adjusting the "market rate" on the new loans, and/or by means-testing the plan so that it helps people who might otherwise default more than it does wealthy interest-rate speculators.
Second, and again disclaiming any insight into mortgage math, I’m going to guess that a trading level of 104 means that a 150bps-overcouponed security is expected to refinance, on average – for some loose value of “average” – in a bit under 3 years, so you get to clip a little under 450bps of excess coupons. (Right? If your entire portfolio was one loan and everything happened with 100% probability, and you picked some reasonable discount rate, you'd earn about 4 points more in the next three years than you would on a par bond. Construct bell curves around that, etc.) A level of 106 means that it’s going to refinance, on average, in a bit over 4 years, so you get to clip a little over 600bps of excess coupons. 108 means close to six years.
On that very stupid basis, this suggests that mortgage bonds say something about expectations for the length of the recesssion. A 106 trading level suggests an extra year or so before the "typical" homeowner refinances a way-above-market mortgage. A 108 trading level suggests an extra 3 years. Put differently, these levels suggest that the market sees 1 to 3 years before employment and housing prices stabilize to the point where people start refinancing normally.
Of course there's noise in the data, plus my math is not to be trusted. Some of the difference between normal times and now is not due to refinancing risk but just duration - a 4.5% coupon is a bigger premium to a 3% par bond than a 7% coupon is to a 5.5% par bond. But also, I suspect that speculation about potential refinancing relief is keeping premiums down. In other words, markets aren't just trying to predict when people will go back to refinancing normally on their own - they're baking in the probability that the government will be refinancing for them.
And even so mortgage bonds trade well above par, and well above historical levels. Which to me suggests that a broad refinancing plan would be much reason - either on a moral basis or an ex-ante-incentives basis - to shed tears for the owners of these bonds. A lot of whom are Fannie and Freddie anyway.