The economists at Goldman Sachs modeled how the Geithner plan would work for the banks' subprime loans and came up with some useful conclusions (not online, alas). The first is that government financing could increase the price investors are willing to pay for the "toxic" loans by as much as 50 percent, if not more. Second, because the banks still value these loans at 91 cents on the dollar (on average), even a 50-percent increase in willingness to pay won't result in a sale in many cases. Third, if the banks do end up selling, it will have to be at well-below 91 cents on the dollar, which will create huge losses and require the banks to seek a lot more capital, probably from the government.* (As I've suggested before, this could be precisely what Treasury wants--it'll quantify the size of the hole and make it easier to make the case to Congress.)
Then there's this point, which is very important. In a nutshell, Goldman shows why the Geithner plan won't be as bad a deal for taxpayers as all the teeth-gnashing suggests:
We have read numerous op-eds decrying the PPIP [this portion of the Geithner plan] on the grounds that the government is taking on enormous risk. Typically, these criticisms focus on “binary” scenarios: either the loans are worth a lot or they are worth zero, and in the latter case the government loses a fortune. [See here, for example.] While these types of examples are generally good for building intuition, they exaggerate the potential cost to the government. For first-lien residential mortgages, which represent a good chunk of the PPIP-eligible assets, there is almost no probability that the value will be near zero...
For example, suppose an investor pays 70 cents on the dollar for a book of subprime mortgage loans. With 6 to 1 leverage ... [the government would be on the hook for a loan of about 60 cents]. Even if all the mortgages default, the investor will probably receive 10-15 cents of cash flow from borrowers before they stop payment, and should get 25-35 cents from the value of the foreclosed properties. So, leaving aside issues of discounting cash flows, even if the investor overpays vis-