Housing Is Rebounding

After looking at the new-home sales this morning, quite strong at 337,000, and knowing that the rates are going lower still, and that the national average price of a home is down 18%, you are looking at what might be the single greatest moment to buy a home in the last 10 years. (I sure wish home prices were in the CPI, because you would see the deflation.)

When the collateralized debt obligations were first built, there was a constant, and that was house price appreciation. When that peaked in 2006, the models all went haywire. These CDOs were not supposed to "go bad" until unemployment spiked, because the models never showed that housing could decline 18%, or 40%-50% in the hottest areas that probably had close to 50% of the CDOs from 2005 to 2007.

When you combine the 100% loan-to-value with the fraud, with the added 20% to value -- making 120% -- and if you used home-equity loans to close the deal (more fraud in some cases) without documentation, you know that many of the CDOs will never come back. Many are probably worth zero.

Let's do the math. We are dealing with a universe of 17 million homes sold between 2005 and 2007 first quarter -- the worst vintages -- of which all but 20% were securitized, so we are using a universe of 13.6 million homes in CDOs (80% securitized of total number).

Half of these were fixed. That's 6.8 million homes in CDOs that are fixed. The fixed default or defaulting rate is 10%. That's 680,000 defaults in CDOs.

Half are variable. You have a 30% default or defaulting rate on the variables, that's 2.04 million homes.

Now, you have 680,000 fixed that are bad in CDOs and 2.04 million variable that are bad in CDOs. Let's call it 3 million defaults or defaulting in CDOs.

The average home price during that period for California (using the highest for worst case) was $330,000. So let's use that. We have 3 million homes in default or defaulting at $330,000 a pop, which yields a total of almost $10 trillion in CDOs.

That's the true size of the problem.

The CDOs are all screwed up. They are diversified among all sorts of geographies, including lots of second-lien mortgages within them.

If you can get the "defaulting" part, not the default part, halted, you can see the magnitude of the problem diminishing.

That's what might be happening.

Any diminishing is great news and makes Geithner's job easier. Of course, you could easily argue that all of these are worth pennies on the dollar. But that presumes they are all stuffed with all defaulting loans, and that's a mistake.

I think that a 2005 vintage now be something worth owning, as you have lots of equity in those homes now.

I think that the 2006 vintage as it goes on gets worse, with the first quarter of 2007 being horrible.

When I see these housing sales numbers -- these are new, but they help burn off inventory -- I would be willing to gamble with federal money on some of the vintages, maybe as much as a third of them; the 2005s have real value.

That's at the crux of the issue.

Remember, there have been lots of write-downs and write-offs already. Some would say we have written off worldwide about $3 trillion.

Now you are getting somewhere. You have about $4 trillion in 2005 that might be worth a lot. You have $4 trillion in 2006 that could have real value. And then let's just write off $1 trillion in 2007.

With these housing sales numbers and the new low rates and the money to buy CDOs, you could argue that what happens is the problem becomes "manageable." And that's where we are getting: a manageable problem, because it can be capped at some price.

Lots of businesses are going from bad to worse. Unemployment is going from bad to worse.

But the housing problem is going from worse to better. We are going toward house price stabilization.