Sunday, February 15, 2009

How to recapitalize the banks with no money down!

Many people have come up with proposals to recapitalize the banks with distressed assets. Since the latest plan from Treasury Secretary Geithner still leaves a lot of room for extrapolation, let me propose my ideas for how this might be done without expending government funds up front.
This idea is based loosely on Ashby's Law: using Variety to absorb Variety. In other words, since the cost of a government guarantee is uncertain, the government can charge an uncertain but correlated back-loaded amount for it.


Banks have distressed loans or securities backed by loans with high and uncertain delinquency rates. Market assigns a high risk premium to this valuation uncertainty, thus bid/ask spread is too big and no trading takes place. Without trading, there is no generally accepted market value for these assets and no confidence in any valuation models. Thus, these assets count for very little in terms of bank capital.

Problem statement:

How can government provide certainty to these distress assets so that they will count substantively toward bank capital?

My idea is as follows:
1. As per original TARP idea, hold reverse auction for pools of such distressed securities or loans.

2. Offer term is as follows, government will sell a 5y Put on this pool at $X in exchange for a 5y Call on this pool at $Y. Bank will propose the strike prices: X and Y.

3. Government will value these calls and puts using internal models and accept an exchange when the Call is valued at Z % or more of the value of the Put.

4. The Call is detachable and sellable on the market to private investors, but the Put is nontransferable.

5. Repeat above steps.

Now why should this work without the bad side effect of triggering more writedown of similar securities on other banks' books?

1. Most importantly, the Put sold by the government put a floor on the value on the pool of assets and changes the psychology of investor toward these assets: i.e. the uncertainty in the value of this pool+Put is greatly reduced. Thus, the uncertainty premium drops and the standard option models for this pool+Put combination become a better approximation to reality. In contrast, the standard option models would not describe well similar pools held by other banks without the floor guarantee.

2. The danger of model error is mitigated due to the well-known Put/Call Parity relation:

For any underlying, the value of the Put = value of the Call + strike price with riskless return - fair value of the pool, assuming the Call and Put strike prices and maturities are the same.

A VERY important property of this relation is that it does NOT depend on modeling assumptions at all. Thus, with Call and Put strike prices that are not too far apart and riskless rate driven to near zero. A reasonably fair exchange offer can then be constructed. Coming up with workable Cut and Put strike prices should be much easier than coming up with a fair value for the Put option alone.

3. Since the government already control several banks via capital infusion or conservatorship, the government can mandate these banks to submit exchange offers with reasonable strike prices to jump start the auction process. As the number of bidders, information availability, and market conditions changes, the hurdle rate of Z can be adjusted in subsequent auctions.

4. There is NO cash outlay by the government initially. There may be a cost when the Put is excercised at maturity, but there is also opportunity to sell the Call before or at maturity at a profit.

5. Valuation is done on a pool level, not individual securities. So if banks submit pool with varied loans and securities, there may be some diversification benefit and it would be difficult to impute prices to individual positions.

An Example in Numbers

A numerical example should make this even more clear.
Say the pool is booked at $60, but the best market offer is $20. The bid/ask spread is at $60-$20=$40. With no trading, the pool may contribute nothing to the bank capital. Let's say the bank offered a Call at $65 in exchange for a Put at $45 and the exchange is accepted. Bank will write down $15 for this pool. Now the bid/ask spread is $60-$45=$15 and the pool can be counted at $45 as part of the bank's capital. This lower uncertainty actually is input into the valuation of the Put and Call on this pool.

At th end of 5 years, if the pool has market price of $35, the government loses $10 and the bank loses $10. If the price would be $55, the bank would make $10 and the government might have made money by selling the call sometime before the end of the 5 years. If the price would be $65 or above, the bank would make $15 max, and the government would make $5 or above on the call.

Market professionals will readingly recognize this as selling a Collar on the pool to the bank. Though I think this scheme is straight-forward enough to explain to the public and the US Congress. In simplest terms, the government is offering to give a floor guarantee for these assets in exchange for a piece of the upside profit. And by providing the floor guarantee, the government is increasing the chance that it will profit on the upside along with the bank.

I would be most interested to hear of weaknesses or problems with this approach. Please feel free to add your criticism or suggestions for improvement.
All contents on this blog copyrighted by respective authors (2008). No rights have been conferred to any other parties by the act of publication. You are most welcome to link to this blog or make fair use of it.