Why Banks Write Down but Don't Sell Subprime Loans

The market doesn’t like the Obama stimulus plan, it doesn’t like the Geithner non-plan, and it just generally doesn’t like anything.

I think the problem with all these plans is they are too complicated (even the Obama housing plan, which is relatively simple), and every Tom, Dick, and Harry (or is it Dennis, Steve, and Rick?) has an opinion as to why they won’t work. What we need is a simple plan that people can understand (bad bank) and to remove any problems (what to pay for the assets).

First of all, I want to say that the way to pay for the assets is a relatively simple thing to figure out. There is no 'second of all'; this is so painfully simple it’s annoying that it’s even discussed as an issue.

Assume that on any foreclosures you get $0, go buy a Bic lighter ($0.99, and it can be used on multiple homes) and light any delinquent homes on fire. So there are now just 3 variables: Income stream, default rates, and cost of capital.

Cost of capital should include whatever rate of return you want. In the case of the banks, something around 6% would probably provide them with a 3% net interest margin (which means their cost of borrowing is 3% and their return on the borrowed money is 6%). If it was the government you could use a lower cost of capital, since they are hypothetically not in the money making business. But 6%-- they are getting 5% on TARP-- is not an unreasonable assumption.

For income stream, I just assumed mortgage payments based on 7% interest and a $250,000 mortgage ($1,767/mo, or if there were 1,000 similar mortgages that would be $1.8mm/mo).

Default rates is where it gets tricky. In JP Morgan’s (we can still trust their data right?) latest quarterly filing they had subprime default rates at 28% and prime default rates at 7%.

Plug those numbers into a simple Excel spreadsheet (literally 3 formulas) and you can come up with the fair value for a basket of these mortgages today. Assuming 28% default rates (and that all 28% default right away), that basket of mortgages is worth 79 cents on the dollar.

If we assume things get a lot worse, and we jack the default rate up to 50% then the pool is worth 55 cents on the dollar. Keep in mind this is assuming zero recoveries from foreclosure (which may or may not be reasonable, but it is as conservative as possible) and that all defaulters default right away (again may or may not be reasonable, but it is the most conservative way to run the numbers).

What about prime mortgages? At 7% default rates those mortgages actually have a value of $1.02 per dollar of principal (this is because the interest rate is 7% but the cost of capital is only 6%). At a more aggressive 10% default rate, the mortgages are worth $0.99 on the dollar.

What this means for any individual bank, I don’t know. What I do know is it took me less than 15 minutes to figure out (obviously this is very general) how to calculate a ‘fair’ price to pay for these assets. I’m sick to death of hearing how these things are impossible to price and how we are staring into a black hole of losses for the banks. If held to maturity (which incidentally is what the guarantees for Citigroup (C) and Bank of America (BAC) are designed to let them do), these assets could reasonably be worth $0.55 on the dollar (for subprime) and $0.99 (for prime).

Timothy Geithner should put out an all or none order, meaning all banks participate. This will reduce the risk that individual baskets may have for all subprime mortgage backed securities at $0.55 on the dollar (or whatever figure he deems appropriate). Problem solved.

Incidentally, the above math is exactly why the banks do not want to sell their exposures. They don’t mind writing them down to zero (or near zero, at least on subprime) but selling them would actually crystallize a loss that they probably don’t expect to have happen. As the loans run off (especially now that refinancing will be less of an issue with the housing stimulus plan) the bank may actually book gains relative to marks (depending on how aggressive those marks were).

So we end up stuck in no man’s land: No one is willing to buy them (the cost of capital, including return, for people buying this would be closer to 15-20%) at the price the banks know they are worth.

If the private/public partnership were to take place, it’s likely that cost of capital for buyers could be dramatically reduced. The government provides financing at say 3% for 90% of the value, then the blended cost of capital would be closer to 5-7%. But I’m not sure why that’s better for taxpayers.

There is a nice 10% cushion on downside, but the taxpayer would be giving up the vast majority of any upside (unless you count the low interest rate).