Tuesday, August 19, 2008

Hypothetical Scenario For The Mortgage Nuts....

Something has always bugged me about credit default swaps and mortgage/bond insurers. Imagine a community with one bank, and that bank has 10 million in loans. Now the usual rules dictate that the bank should have at least a 1 million in equity as a cushion because, face it, lending has a tendency to have its bad cycles. The other 9 million is financed by deposits.

Now let's say another party comes to the bank and offers to take the credit risk on all 10 million of the bank's loans, in what is essentially a blanket credit default swap on their loans. The bank now views its operation as essentially risk-free (ignoring liquidity risk), so their equity cushion doesn't matter. But shouldn't the new counter-party now be required to have a 1 million equity cushion in case things go bad, since they are supposed to be the first and only line of defense? I think the answer to that is yes.

The problem is I haven't seen that. Look at Fannie Mae. At September 30, 2007, they had 2.8 trillion in total mortgage assets and guarantees. But their balance sheet had total capital of only 49 billion, or 1.75% of total obligations. If it were looked at like a bank, that would be an asset-equity ratio of over 50-1. That doesn't leave much margin of safety at all.

Am I missing something?

3 comments:

ragnar said...

I seem to remember hearing in a class once that in the depression, people were able to buy stock on margin, and use the stock as the collateral for the debt (well, this is oversimplified).

Basically, this is the same thing with the mortgages. After all, for a long time, there was absolutely NO risk in originating mortgages. With rising housing prices, in the event of default, the house would be sold at the commissioners sale, and more than cover all the owed amounts, generally, the defaulting party would actually make some money. Obviously, this couldn't have lasted, though, I suppose that few really saw it. After all, wouldn't they have more of a debt coverage if there was more risk with the notes that the banks held?

With all this said, my buddy just bought a house with nothing down. And guess who it was through:FHA! Ironic, but it seems that the government is the last to catch on to this sort of thing!

Sivaram Velauthapillai said...

There are two issues here. One is insurance; and the other is leverage.

I think the answer to your question lies in the nature of insurance. Replace mortgages with fire insurance (or property insurance or whatever else) and think about it. Does a fire insurer hold enough to cover the damage for every single house it insurers? Nope (if they were required, they would have to have trillions in capital). The same thing with bond insurers, mortage insurers, and the like.

The key thing is that the losses are expected to be only a small percentage of the assets. The fact that you are responsible for trillions worth of houses doesn't mean much if only, say, 1% of houses catch fire.

As for Fannie and Freddie, they are not insurance companies but a similar concept, along with the notion of leverage, applies. The reason the GSEs own trillions in mortgages while having very small capital is because the losses are expected to be small. The historical loss rates are quite small.

But, due to a bunch of issues, including weak underwriting standardards, fraudulent mortgages, and hard to analyze securities, we are ending up with much higher loss rates. (In the case of bond insurers, the real culprit--especially for CDOs--is that one of the big assumptions of mortagage defaults being independent from one another, turned out to be completely wrong and losses are showing up everywhere.)

In the end, the bond insurers, mortage insurers, and the GSEs end up in a situation with high leverage with greater than expected losses. Not a good place to be. The bond and mortgage insurers have super high leverage (50x to 100x) because they wrote with zero-loss assumptions and thought that AAA-rated super-senior securities won't default appreciably. The GSEs were able to employ high leverage due to the market expectation of US government backing.

Although it's easy to blame leverage, the real problem was a mistake with their modelling. The companies, along with rating agencies, didn't price in the weakening mortage standards, fraudulent mortgages, liar loans, and so forth.

Nnejad said...

Sivaram, your right that their capital should depend entirely on the amount of expected losses, not the amount of total liability. We should be able to gauge the proper amount of capital for the GSE's by looking at the banks, which at least in general originate fairly similar products. I mean, most banks (usually) haven't been out there writing loans that were far inferior to the GSEs. And even when they did, the GSEs began lowering their guidelines right alongside them. So if banks were historically always required to have about an 8-10% equity cushion, is it right that Fannie had less than 2%? Especially during a period of excessively high risk taking? I think your right that their modeling was entirely out of whack,. And if that 10% rule has any value, then they were far too over-leveraged and I don't think it is likely that the GSE shareholders will make it out of this one.