Quantifying the "Geithner Put"

The plan is out. The Treasury will buy the toxic assets in an auction, but will allow participants to leverage their investment with non-recourse loans from the FDIC on a six-to-one basis. The taxpayer gets some of the upside because the Treasury invests equity dollar-for-dollar with the bidders. But, because of those non-recourse loans, the taxpayer gets all the downside below the 14% of the purchase price and shares losses dollar-for-dollar with private investors up to 14%. I like the clear way Nemo spells it out.

Is it a good idea? Will it work?

Brad DeLong says 'Yes', Paul Krugman says 'No'. There are many other views, but these two guys usually see eye-to-eye. No ideological wrangling here. Hmmmm.

Let's do the math.

We can only answer this question with some measure of uncertainty about true value of the underlying asset. If uncertainty were small, say a standard deviation (SD) equal to 6% of the purchase price, then the subsidy implicit in those non-recourse loans is pretty small. The odds the price falls below the 14% public/private-equity skin in the game is small, t-stat of 7/3. That's less than a 1% chance of FDIC losing some of its principal.

Nemo has an extreme case where the SD = 50% of the face value and nearly 60% of the purchase price. THAT's a lot of uncertainty. But he also takes the price as fixed.

A 7% SD seems too small to Paul Krugman. It doesn't seem altogether crazy to me.

But suppose the SD is actually much greater than 7%. That risk means the upside potential is huge. With all that government-financed leverage and protection from the downside, investors will bid up the price.

Let's look at a range of possibilities. In this exercise I consider the expected return on the underlying assets assuming the marginal investor's return is fixed at 10% (due to competitive bidding). I'm assuming a normal distribution (no fat tails) and no risk premium

SD : (Expected Return Payoff/Price - 1) x 100%

6% : 1.4%
9% : 1.3%
12% : 0.8%
15% : 0.05%
18% : -1.0%
21% : -2.2%
24% : -3.8%
27% : -5.5%
30% : -7.4%

Expected returns of the underlying asset go down with risk. Expected returns for the taxpayer would be somewhat less than that of the underlying asset since the investor gets 10% in expectation. A seriously raw deal for the taxpayer kicks in around an SD of 15% of the purchase price.

I don't know how much uncertainty there is. But my guess is it's less than an SD of 15%.

Krugman, as usual, has a good point. But I think Krugman may be a little too pessimistic on this one.

Update: I realize this is a bit abstract. I'll try to be clearer. The figure on the right-hand-side is an estimate for the expected return of the toxic assets (aka, mortgage-backed securities) starting from the market-clearing auction price. The SD is a measure of uncertainty around that expectation. In words, the SD is a "typical" deviation from the expected return, whether positive or negative. Expected returns go down as the SD increases because private investors are protected from downside risk from non-recourse loans ("Geithner's put"). Private investors determine the price of the assets at auction such that they always earn a 10% return in expectation. The taxpayers' returns are expected to be slightly less than the number on the right-hand-side of the column.