by Calculated Risk on 9/20/2008 03:34:00 PM

Update: While I was writing, Dr. Krugman wrote: No deal. Definitely worth reading!

The underlying problem is that house prices are still too high and no one knows how much further prices will fall. The value of the troubled assets is dependent on future house prices (note: house prices are the key factor for foreclosures and loss severity). Some people don't seem to understand house prices will continue to fall, from the NY Times: But Will It Work? (hat tip Yal)

“It’s easy to forget amid all the fancy stuff — credit derivatives, swaps — that the root cause of all this is declining house prices. If you can reverse that, then people start coming out of their foxholes and start putting their money in places they have been too afraid to put it.”
Alan S. Blinder, Princeton economist, and former vice chairman of the Federal Reserve Board of Governors, Sept 20, 2008

This plan isn't intended to reverse the house price trend - and it shouldn't be intended for that purpose. The Paulson plan is intended to keep the banks lending to credit worthy borrowers. Other goals are to minimize the burden for taxpayers, and minimize moral hazard - but the primary goal is to keep the banks lending to minimize the impact of the credit crisis on the economy.

There are private investors willing to buy these troubled assets right now, but the banks do not want to sell at those prices. Why? Some banks believe the assets are worth more than the current bids (it all depends on future house prices, and different banks and investors have different projections). And many banks are unwilling to accept the current bids because the banks would then be insolvent. See Professor Krugman's: Doubts about the rescue and Uneasy feelings. Also, even solvent banks would probably have to recapitalize (dilute shareholders) or reduce lending if they sold at current bid prices.

So how does the Treasury plan help? It isn't clear yet. The first goal should be transparency of the troubled assets. What do the banks own, and what are the assets really worth? Transparency is surprisingly difficult: each RMBS and CDOs - even within the same asset class and origination year - can have significantly different values depending on the orginator and other factors. If the Treasury conducts a reverse dutch auction on a broad asset class, they will probably end up with certain New Century and Bear Stearns deals that are basically worthless.

To facilitate price discovery, it would probably be better to bid for individual mortgages from RMBS pools, but analyzing each mortgage would be a monumental task. We definitely do not want the Treasury to buy RMBS and CDOs at anywhere near the value on the bank's books. Buying at those prices would help keep the banks lending, but it would also severely impact the taxpayers, it would be a transfer of wealth from the many to the few, and it would also encourage future excessive risk taking.

So determining price will be difficult. And what happens if a price can be determined? How does this help keep the banks lending?

As I noted when the plan was announced, buying impaired assets at a steep discount reduces regulatory capital as losses are realized, and therefore will lead to less lending unless the banks are recapitalized.

Perhaps with a clean balance sheet, the banks can attract private capital (with significant dilution of current shareholders). Or perhaps something similar to the Depression era Reconstruction Finance Corporation (RFC) can be part of the plan to invest capital in the banks.

If an investment from the Government is required anyway, why bother buying the impaired assets?

This suggests a different approach: First, a recognition phase with complete transparency. Have private investors bid on some assets to establish market prices (some portion should be sold to the private investors to encourage bids), and then let the banks argue for their own valuations. Based on an analysis of these valuations, have the Treasury make an RFC type investment in the bank with a convertible debenture that would count as regulatory capital. This capital infusion would keep the banks lending (the primary goal) and the amount required would be far less than the amount needed to buy the troubled assets.

If a bank can pay off the debentures with interest - possibly because the assets perform as the bank expects, or perhaps by bringing in private capital - then there would be no dilution from the debentures. Otherwise the debentures convert into preferred shares and significantly dilute the shareholders - and then the government can sell the shares on the public market. Ideally the debtholders would take a haircut too (before the taxpayers), but that is probably too complicated. This alternative would keep the banks lending, minimize the cost to the taxpayers, and reduce moral hazard.

Just my two cents as we wait for more details ... Best to all.