One hypothesis about the banking crisis is that a secondary market for mortgages "does not now exist" (quote from Mr. Geithner's oped introducing the plan) because there is a "lack of clarity about the value of these legacy assets [which makes] it difficult for some financial institutions to raise new private capital on their own." (quote from U.S. Treasury fact sheet).
I guess the story is: a bank owner thinks a pool of mortgages is worth $10 million, and therefore refuses to sell for less than that. Potential buyers of that pool think it is worth $5 million, and therefore refuses to buy for more than $5 million. The market does not exist because the sellers value the assets less than potential buyers do.
A simple subsidy will not solve the problem. Suppose that the government said that it would pay for $5 million of the purchase price. Then potential buyers would be willing to pay $10 million ($5 million for the pool itself and another $5 million for the government subsidy). The problem is that the bank owner (presumably aware of the subsidy) may not want to sell for less than $15 million. The reason is that he might be able to buy the pool from himself -- in which case he would pay $10 million for the asset (that's what he thinks it is worth) and another $5 million for the subsidy. So if he can sell to himself for $15 million, why should he sell to another buyer who will pay only $10 million?
You might say that the Treasury or FDIC would not allow a sham transaction like I just described. Well then you are saying that the Treasury or FDIC will micro manage things -- good luck with that!
The Geithner-Summers plan is more complicated than the simple subsidy described above. But my intuition is that even the complicated version misses the point that a subsidy raises every one's valuations of the item subsidized, rather than (as supposed needed) raising the valuations of some relative to the valuation others. Thus, the Geithner-Summers plan looks like giving taxpayer money to banks and creating a flurry of activity without really changing any of the fundamentals (reminiscent of what I wrote last fall about the Paulson plan).
Commenters: do you know of anyone who has worked out an explicit lemon's model and then added the Geithner-Summers plan to it?