The main costs from TARP will come from bailing out the auto companies and insurer AIG, not banks.
NYTimes: Even as voters rage and candidates put up ads against government bailouts, the reviled mother of them all — the $700 billion lifeline to banks, insurance and auto companies — will expire after Sunday at a fraction of that cost and could conceivably earn taxpayers a profit.
... The Treasury never tapped the full $700 billion. It committed $470 billion and to date has disbursed $387 billion, mostly to hundreds of banks and later to A.I.G., to the auto industry — Chrysler, General Motors, the G.M. financing company and suppliers — and to what is, so far, an unsuccessful effort to help homeowners avoid foreclosures.
When Mr. Obama took office, the financial system remained so weak that his first budget indicated the Treasury might need another $750 billion for TARP. The administration soon dropped that idea as Mr. Geithner overhauled the rescue program and the banking system stabilized. Still, by mid-2009, the administration projected that TARP could lose $341 billion, a figure that reflected new commitments to A.I.G. and the auto industry.
The Congressional Budget Office, which had a slightly higher loss estimate initially, in August reduced that to $66 billion.
Now Treasury reckons that taxpayers will lose less than $50 billion at worst, but at best could break even or even make money. Its best-case scenarios, however, assume that A.I.G. and the auto companies will remain profitable and that Treasury will get a good price as it sells its corporate shares in coming years.
“We’d have to be very lucky to have both A.I.G. and the auto companies pay us back in full,” Mr. Elliott said.
... By any measure, TARP’s final tally will be less than even its advocates expected amid the crisis. But the program remains a big loser politically.
See this post from September 2008, during the heat of the TARP debate:
... The following false conundrum has been stated recently by numerous analysts, including Paul Krugman: "if Treasury wants to recapitalize banks it has to overpay for toxic assets, to the detriment of taxpayers; if it wants to pay fair prices for the assets then banks won't benefit." There is no conundrum if markets, at this instant in time, are systematically underpricing mortgage assets.
When the Internet bubble burst in the early years of this century, investors were so gun shy and under so much pressure that they would not pay even rationally justifiable prices for stakes in technology companies. Smart investors who were willing to put capital at risk could buy assets at fire sale prices and made huge profits. This is nothing more than fear and herd mentality at work. If herd thinking can lead to overpricing of assets, why not underpricing immediately following a collapse?
Markets overshoot on both the up- and the down-side!
These points are obvious to any trader... it's the academics with equilibrium intuitions who are struggling to understand!
... To understand, it helps to have seen the collapse of a financial bubble firsthand. If you haven't (as, I suspect is the case with most academic economists), you are likely to cling to the idea that the market price of an asset is a good forecast of its actual value. However, this is completely wrong in the wake of a collapse. (And, certainly, the predictive power of the market price cannot hold at all times -- it is likely to be most wrong at the peak and in the aftermath of a bubble.)